Bankruptcy

Credit Card Redlining

By Adam Levitin

Several months ago, when I was a scarcely tolerated guest blogger, I wrote a post that asked (What Determines) What's In Your Wallet? The point was to highlight how little we know about what determines what credit card offers a particular individual receives. I suggested that there was a danger of red-lining in the credit card industry based at least on what solicitations one received. (I got a bunch of indignant e-mails about this emphasizing that federal law prohibits discriminatory lending...as if no one ever violated the law. Ah, Camelot.)

Well, now comes an empirical study from Ethan Cohen-Cole, an economist at the Boston Federal Reserve that indicates that there is redlining in the credit card industry. Residents of black neighborhoods are less likely to less consumer credit than residents of white neighborhoods, all things being equal:

This paper?s principal observation is that remarkably, in spite of identical scores and identical community characteristics, our individual in the Black neighborhood receives less consumer credit (e.g. fewer credit cards) than the individual in the White area. That is, in spite of the fact that both have been assessed to have similar risks of nonpayment, as determined by the credit score, the person living in the Black area has less ability to access credit.

And if there is less available credit card credit, where do people in black neighborhoods turn for credit?

To be sure, a single empirical study is just that and one can quibble about methodology, but wow! Talk about opening up a new front in credit card regulation. I've got to think that if this study gets some attention it is going to cause Congress to ask some questions. Of course, the study says nothing about the terms of the credit, but if I had to take a guess...well, what would you think?

The study includes a number of appropriate caveats about its findings. It also notes that

given the degree of regulatory scrutiny over the credit decision itself, one suspects that if any disparity exists in the provision of credit, it likely originates in the pre-screening (marketing) efforts.

I'm less sure of this. Certainly the marketing is a big factor; there are certain cards that are specifically marketed at particular minority communities, e.g., the Freedom Card (not to be confused with the Chase Freedom Card), which is marketed to poor blacks in Philadelphia. Moreover, the study can only speak to aggregate levels of credit, not to individual issuers' lending solicitation and lending decisions.

But the regulatory scrutiny of the credit issuance decision--what exactly is that? Who (if anyone) is actively screening for Equal Credit Opportunity Act (ECOA) violations and how are they doing it? We don't have anything like HMDA data for credit card lending. If the issuer never records the race of the borrower, but only uses a proxy variable, it would be pretty hard for an examiner to pick up systematic ECOA violations. And is it a ECOA violation is the discrimination is not on race per se, but on a proxy variable?

The card industry has carefully homed in on all sorts of variations in consumer behavior to fine-tune its profit model. In a competitive industry that does such careful data mining, do we really think that every issuer would shy away from discriminatory lending if it were (a) profitable and (b) not explicitly discriminatory because of use of a proxy variable? Or is racial discrimination different and just out of bounds? I'm not sanguine.

What this all speaks to is the need for greater transparency in the credit card industry, including the collection of more publicly available data. Government agencies are charged with enforcing the ECOA, but they have limited resources. Making collecting data and making it publicly available allows the academic community to contribute to ECOA monitoring, a win-win for the public and data hungry academics.

Full post as published by Credit Slips on April 03, 2008 (boomark / email).