Racial Discrimination Alive and Well in Finance Biz

Posted by Kathy G, The G-Spot at 7:10 AM on July 21, 2008.

Responding to this post, a Certain Someone writes:

Credit is one of those weird areas where there is a lot of belief in discrimination, but as far as I can tell, not all that much evidence.

[. . .]

Yes, I've seen the research arguing that people in black communities get worse loan terms than their credit score suggests. As far as I can tell, this research failed to control for some pretty major factors, like assets. . .

[. . .]

Most of the aggregate research I've seen fails to reject the null hypothesis that there is no discrimination in loan markets . . .

Would that that were true. But actually, the bulk of the academic literature on this subject suggests that there is a significant degree of racial discrimination in loan and credit markets.

To be fair, it's not easy to determine the extent to which discrimination occurs. The available datasets are incomplete. Researchers don't necessarily know which variables the lenders and creditors are looking at when considering credit or loan applications, or how those variables are weighted.

Nevertheless, researchers have been able to get their hands on some unusually rich datasets, which they've examined using the most plausible specifications as to the lending criteria.

There are basically two types of credit discrimination that occur: discrimination based on the race of the applicant, and discrimination based on the racial composition of the neighborhood where the applicant resides (the latter type of discrimination is known as "redlining").

In terms of discrimination based on the race of the applicant, the best evidence we have is a series of studies based on data from the 1991 Home Mortgage Disclosure Act, with supplementary data supplied by the Federal Reserve Bank of Boston. Among the factors looked at were individual applicants' financial, employment, and property background variables -- and yes, this data contain info about assets, both liquid and total (which is contrary to the assertion made above that factors like assets are not controlled for in these studies). The original paper by Alicia Munnell et al., which was published in the American Economic Review, the premier academic journal in the economics field, found that applications from blacks and Hispanics were significantly more likely to be rejected than similar applications from whites.

Now, as the economist Kevin Lang points out in his invaluable book Poverty and Discrimination -- a book, I might add, that has received strongly favorable reviews by such right-leaning economists as Arnold Kling (who wrote, "I heart Kevin Lang") and Tyler Cowan Tyler Cowen --  Munnell's original study has been subject to "extensive reanalysis." Critics have pointed out that even though the dataset used in the study is probably the richest we have on the subject, it doesn't, in Lang's words "fully capture all the information available to the banks." This is true;  but based on chapter 3 of this book, which reanalyzes the Boston Fed data, Lang concurs with the assessment that the Boston Fed studies "create a presumption that discrimination exists in mortgage lending."

Lang, citing the discussion of audit studies that occur in chapter 2 of the Urban Institute book, also notes that audit studies of the pre-approval loan process "confirm that blacks receive less encouragement to apply and are more likely to be encouraged to apply to other lenders."

There's more: a review of the literature on racial discrimination in credit markets in the current issue of the Annual Review of Sociology

concludes that, although the number of mortgage loans given to blacks and Hispanics has substantially increased over the past two decades, significant discrimination persists:

Nevertheless, the evidence indicates that blacks and Hispanics continue to face higher rejection rates and receive less favorable terms than whites of equal credit risk.

Thus far, the best research that has been done on the subject of racial discrimination in credit markets concerns mortgages. But studies on other forms of consumer credit have just begun to emerge. In a review of this emerging literature, economist Gary Dymski writes that initial findings suggest that "specific minority borrowers are significantly more likely to be sold a subprime or predatory loan than are nonminority borrowers with similar risk profiles." A recent, very interesting paper by the Federal Reserve Bank of Boston looks at credit card redlining; here's a summary of the results:

Using a unique and proprietary database of credit histories from a major credit bureau, this paper links location-based information on race with individual credit files. After controlling for the influence of such other place-specific factors as crime, housing vacancy rates, and general population demographics, the paper finds qualitatively large differences in the amount of credit offered to similarly qualified applicants living in Black versus White areas.

Again, these findings must be taken with caution. Even though the dataset used for this study is unusually rich, it is far from complete, and the researchers did not have information on the criteria by which these financial institutions made their lending decisions. On the other hand, the researchers did have access to thousands of individual credit reports from a geographically stratified random sample, and the results are suggestive.

Even acknowledging that the available data is far from perfect, the evidence is overwhelmingly clear: there is a significant degree of racial discrimination in credit markets. Honest brokers who have reviewed this literature, such as economists like Lang and Dymski and the sociologists in the Annual Review of Sociology acknowledge the ambiguities but come to the basic conclusion that the evidence on this score shows that discrimination exists.

The argument that this kind of discrimination can't happen because it would mean markets are behaving irrationally is dubious at best. Human beings and the institutions they create are not always rational. Norms, institutions, and psychological factors play a role in economic decisions, and sometimes they cause economic actors to behave in ways that are not entirely rational. 

In the same post where McCardle argued that credit discrimination doesn't exist, she nevertheless admitted that "the evidence for discrimination in the labor market seems strong--nay, nearly incontrovertible." Well, surely racial discrimination in labor markets is irrational behavior as well -- a firm would almost certainly be better off by hiring the most productive job applicants, regardless of race. Why would firms act irrationally when it comes to employment discrimination, yet behave completely rationally when it comes to issuing credit?

If racial bias is strong enough to cause firms to act in irrational ways when it comes to hiring decisions, it seems highly plausible to me that it might cause them to act irrationally when it comes to credit decisions as well. And indeed, the evidence strongly suggests that racial discrimination is no less absent in credit markets than it is in so many other features of American life.

Tagged as: finance, discrimination

Kathy G Runs The G-Spot blog.