New Writers, but Same Old Story

You may recall our discussion of a paper released in October 2000 by the Association of Community Organizations for Reform (ACORN). That paper relied on data from the Department of Housing and Urban Development (HUD) to suggest that mortgage lenders discriminated against minorities in granting mortgage loans and loans to refinance existing mortgages. Now, the Center for Community Change (CCC) has published another study alleging racial discrimination in mortgage lending using the same database and methodology. The Center's website notes that the Center has existed for 30 years "and has been nationally recognized for its work helping people build organizations and create better communities and policies." The study alleges discrimination based on evidence that a disproportionate number of African Americans receive subprime (i.e., higher-priced) mortgage loans.

A critical problem is that the authors of both studies relied entirely on the HUD database. However, the HUD data reports only on mortgage loans actually made to consumers. It provides no data relating to the credit risk posed by the applicants that were granted mortgage loans, nor on those applicants who were rejected.

Few of us would bet on a racehorse without knowing its record. In the HUD database we know only that a lender "bet" on a credit applicant. But the authors of the ACORN report and the CCC report did not and could not know what the lender knew about the borrower's track record. Just as bettors at a racetrack, mortgage lenders view the credit records and other information provided by the credit applicants. Taken together, these data are a key element in deciding whether or not to grant a mortgage or to refinance an existing mortgage. Yet, in both the ACORN report and the CCC report, the authors had no way of knowing what the lenders knew about applicants' "track records," making an evaluation of the resulting lending decisions problematic. Let us now explore the issue in greater detail.

Missing data

What is the key basis for any lender when making the approve/disapprove decision? The essential elements are the ability and willingness to repay the loan, which are assessed by reviewing the applicant's credit record, much of which is reflected in the applicant's "credit score." This information is not available in the HUD data.

The same problems that plagued the previous ACORN study also afflict the CCC study. As in the ACORN study, the authors assert racial discrimination by comparing the number of subprime loans granted to minorities and whites. For example, let us examine the finding cited on page vii. "Lower-income African-Americans received 2.4 times as many subprime loans as lower income whites." For the sake of discussion, assume that we have the following table, which shows that African-American borrowers received 2.4 times the number of subprime loans as whites (2,400/1,000 = 2.4).

Now let us add some hypothetical information that the authors did not have and could not have obtained from the HUD data: the number of applicants. Such information allows us to calculate the approval rates for the two groups of applicants. Observe that in the example, the approval rate is higher for white applicants than for African Americans.

Approval rates largely depend upon applicants' credit scores, which reflect the applicant's credit record. However, the authors did not have that highly critical piece of information, which is also absent from the HUD data. Suppose that we observe that whites have higher credit scores, on average, than do African Americans, as indicated in the hypothetical table below (Indeed, the ACORN studies showing that African-Americans rely more heavily on subprime mortgages than do whites is consistent with this observation).

Nonetheless, because there were more applications from African-Americans, they obtained more mortgage loans than did whites: 2.4 times as many to be exact. This result is a product of two partially offsetting forces: (1) the higher number of applications from African-Americans, tempered by (2) the higher acceptance rate for white applicants, reflecting their higher credit scores.

As in the case of the ACORN study, the basic problem with this study is that the HUD data do not include the variables that are essential to determine whether applicants were given loans on the basis of the probability of their repaying the loan. Without knowing the percentage of loans approved and the credit scores of the applicants, the ACORN report and the CCC paper tell us nothing about the extent of racial discrimination, if any, in mortgage lending.

Assertion that subprime loans are bad for consumers

Another basic problem with this study, as well as the ACORN study is the implicit assumption that subprime loans are "bad" for consumers, especially low-income consumers. The U.S. Department of Housing and Urban Development would not agree.

Subprime mortgage lending provides a source of funds for credit-impaired borrowers and other borrowers that are unable to obtain credit in the prime market. While subprime lending often involves lending to borrowers with past credit problems, the subprime market is more than that. A borrower may not qualify for a prime loan, yet may be eligible for a subprime loan, for reasons other than past credit performance. For example, a borrower may have a good credit history but have high monthly debt payment relative to income. Or, a borrower may have few or no assets other than current income to support loan payments. Borrowers may also be self-employed, have variable income, or simply want to limit disclosure of their financial situation. These conditions may make it difficult for a borrower to obtain credit at prime rates, yet a subprime lender may be willing to take the added risk in exchange for a higher interest rate. (U.S. Department of Housing and Urban Development, Curbing Predatory Mortgage Lending, August 2000, p. 25.

Furthermore, the Center's criticism of subprime lenders also appears to be inconsistent with its own mission statement on page ii of its study: "The Neighborhood Revitalization Project of the Center for Community is based in Washington, D.C. The project provides training and technical assistance and also advocates for public policies on behalf of community groups and others seeking to expand the availability of responsible lending and banking services for lower-income, minority and other underserved urban and rural areas." (Emphasis supplied) Would not these "lower-income, minority" consumers welcome subprime loans that they needed to buy homes or to remodel existing homes? Put another way, what good is a low-priced, prime loan to such a borrower, if the borrower can't qualify?