Partners (Summer 1998)

Photograph by Adria Wiggins Age 16
Photograph by Adria Wiggins Age 16
Do Minority-Owned Banks Treat Minorities Better?
An Empirical Test of the Cultural Affinity Hypothesis
Raphael W. Bostic, Economist, and Glenn B. Canner, Senior Advisor,
Board of Governors of the Federal Reserve System

National data on the disposition of applications for home mortgages reveal wide disparities in rejection rates among racial and ethnic groups. One theory offered to explain these disparities is the "Cultural affinity hypothesis," developed by Calomiris, Kahn, and Longhofer (1994). This hypothesis posits that lenders find it easier, and therefore less costly, to evaluate applicants "like them" than applicants with different backgrounds. Regarding mortgage rejection differentials, then, the cultural affinity hypothesis suggests that lenders at white-owned banks reject minority applicants more often because they have more difficulty precisely estimating the credit risks they represent.

Recently, Black, Collins, and Cyree (1997) found that black-owned banks are more likely to reject black applicants than white-owned banks with similar characteristics. On its face, these results appear inconsistent with the cultural affinity hypothesis. However, these findings do not necessarily refute the notion of cultural affinity. Rather than occurring only on the part of lenders, cultural affinity might play out among applicants as well.

Minority applicants may feel more comfortable applying for mortgages at minority-owned banks, which could result in a relatively large volume of marginally qualified minority applicants at minority-owned banks. In such a case, minority-owned banks could have higherrejection rates of minority applicants than white-owned banks, even if only minority lenders exhibit cultural affinity or if lenders of both races applied the same underwriting standards.

This paper provides empirical tests which identify and separate the potential effects of applicant-driven and lender driven cultural affinity. Applicant-driven cultural affinity is tested by searching for differences in the applicant pools of minority-owned and white-owned peer banks.

Our methodology for studying lender-based cultural affinity is similar to that used in Black, Collins, and Cyree (1997), but it differs from theirs in important ways. First, we use a sample selection methodology which avoids the potential confounding of the two cultural affinity effects which might arise if applicants across banks differ systemically in their general characteristics. This approach thus permits a more direct test of the existence of lender-based cultural affinity.

Second, we test for whether the theory applies differently across races by conducting the analysis for both blacks and Asians. Lender-based cultural affinity among Hispanics could not be evaluated due to a small number of Hispanic-owned banks.

We find strong evidence suggesting that cultural affinity operates through applicants as, after control ling for non-ownership bank characteristics, black-owned and Asian-owned banks are significantly more likely to receive applicants from blacks and Asians, respectively. Interestingly, the data suggest no affinity between Hispanics and bank owners at either black-owned or Asian-owned banks.

The evidence also suggests that affinities are individual-based rather than geography-based, as no relationship is observed between application rates and the percentage of the local population that is minority. In contrast to the result for applicant-based cultural affinity, we find no evidence that lender-based cultural affinity exists. After accounting for differences in the applicants' pools, whites, blacks, and Asians all largely face the same likelihood of rejection at black-owned, Asian-owned, or white-owned peer banks.


The cultural affinity hypothesis has been offered as one explanation for observed racial disparities in mortgage lending. According to this theory, lenders are better able to assess the riskiness of applicants with similar backgrounds, which results in these applicants facing more favorable underwriting conditions relative to members of other ethnic groups and backgrounds.

Additionally, although the cultural affinity hypothesis was introduced in the context of lender behavior, it may also be more relevant for applicants. In particular, applicants may be more likely to apply to banks owned or operated by members of their ethnic group, for comfort reasons or because they believe their chances for acceptance are enhanced at such institutions.

In this paper, these arguments are evaluated by comparing the characteristics of applicant pools and application approval rate patterns for minority-owned banks and white-owned peer banks. The evidence is consistent with the view that cultural affinity occurs among mortgage applicants, as minorities are significantly more likely to file applications at minority-owned banks than at white-owned peer banks. This affinity appears to be quite specific among minorities, as only Asians are more likely to apply to Asian-owned banks and only blacks are more likely to submit applications at black-owned banks.

No "across-race" affinity is observed. Also, we find no evidence of "neighborhood-based" applicant-driven cultural affinity. Application distributions at minority-owned and peer banks do not vary significantly with the minority composition of the neighborhood. Regarding lender-driven cultural affinity, however, the evidence is not supportive of the theory.

Like Black, Collins, and Cyree (1997) we find no evidence that black-owned banks treat black applicants better than their white-owned peer banks do or that Asian-owned banks treat Asian applicants better than their white-owned peer banks do. Unlike those authors, however, we find no evidence that they systemically treat minorities any worse either. Consistent lender-driven cultural affinity does not appear to exist.

Overall, our results differ from those of Hunter and Walker (1996), who find evidence consistent with the existence of lender-based cultural affinity. However, their approach differs from ours in that they do not attempt to control for differences in the applicant distributions or in lender characteristics, both of which we do using matching procedures. As we have shown, these differences can have an important effect on observed outcomes. In particular, they tend to attenuate observed behavioral differences.

An important ex ante assumption underlying our interpretation of the results is that underwriting standards for institutions operating in the same market do not vary with the race of the bank's owners. The evidence offers general support for this assumption. Within a given market, applicants with similar characteristics received the same treatment on average, regardless of the characteristics of a bank's ownership. This is consistent with the notion that bank underwriting standards are based on objective measures of risk and that more subjective factors which might come into play, such as cultural affinity, do not significantly influence underwriting decisions.

For copies of the research papers identified in this article, please call Community Affairs at 404/498-7200.

Return to Index