The Data Is Clear: Auto Lenders Discriminate

We can't let Congress weaken the ability of the CFPB to address discrimination in auto financing.

U.S. News & World Report

Auto Lenders Discriminate by Race

Millions of Americans will hit the road over Thanksgiving Weekend this year and benefit from relatively low national gas prices.

Dealers are twice as likely to add a markup to the loans of African-Americans.(iStockphoto)

The House of Representatives is about to vote on H.R. 1737, the Reforming CFPB Indirect Auto Financing Guidance Act, which would weaken the ability of the Consumer Financial Protection Bureau to address discrimination in auto financing. That would be a mistake.

The CFPB and Department of Justice have both concluded that auto financers' policy of giving dealers discretion to mark up the interest rate of auto financing results in discrimination against minority borrowers. In enforcement actions against Ally Bank, American Honda Finance Co. and Fifth Third Bank, the agencies found that borrowers of color paid higher interest rates than white borrowers with similar creditworthiness. The CFPB has urged all auto financers to move to a method of compensating auto dealers that does not result in discrimination.

These findings of discrimination have come under attack because they were based on proxy analysis. Loan files do not disclose the race of individual borrowers, so the CFPB and DOJ conducted statistical analyses based on consumers' last names and geographic location. This analysis conducted on data from millions of auto finance transactions can find patterns that almost certainly reveal actual differences based on race.

In fact, a few years ago, the National Consumer Law Center proved the same conclusion in courts of law based on data that did reveal the race of individual borrowers. In the late 1990s, we co-counseled class action lawsuits against all of the major auto finance companies challenging the use of discretionary dealer markups. In discovery, we obtained data on individual loans, and we hired an expert witness to match the loans to drivers' license data in states that collected the drivers' race. With millions of loans to analyze, we also could find the race of many borrowers who financed a car in a state that does not collect racial information but previously lived in a state that does.

The results were overwhelming: Dealers were twice as likely to add a markup to the loans of African-Americans than to loans taken out by comparable white borrowers. Furthermore, when African-American and compatible white borrowers both were marked up, the African-American borrowers paid significantly more. For example, in Wisconsin, black Ford buyers paid an average $1,041 markup, whereas white buyers paid $156. In Alabama, black GMAC buyers paid markups that averaged $836, but the markup for white buyers was only $276.

We found statistically significant racial disparities in every state with sufficient data and in every region of the country. We also observed disparities for Hispanics on a national level, but Hispanic origin was not coded on enough loans to analyze state by state.

This powerful evidence convinced the courts that "the plaintiffs have proved their case": Permitting discretionary markups led to unacceptable racially disparate impacts. All of the auto finance companies settled with us, paid millions in compensation and agreed to limit discretionary auto dealer markups for five years or more.

Unfortunately, these agreements all expired by 2012. Freed from the constraints, under pressure from dealers who play one finance company off of another, the financers have nearly all returned to the same practices that led to our lawsuit.

This is not to say that auto dealers charge higher rates to minorities intentionally. Many subconscious factors go into the negotiation over a loan. Dealers may be subliminally letting race affect their calculation of how much money they can extract. Whatever the cause, the racially disparate results are clear and unacceptable.

So why can't the CFPB and DOJ do what we did, and analyze the race of individual loans? Unlike the race of home-mortgage borrowers, the race of car finance borrowers is not recorded or reported after the transaction is completed. Fewer states now collect race data on drivers' licenses, and even where available, requiring lenders to turn over data on individual borrowers and matching it to drivers' license data is quite burdensome. Race coding based on proxy is a recognized, reliable way for statistical experts to produce robust analyses. Modern statistical tools are quite powerful when applied to millions of data points, and the correlation with our own findings only reinforces the validity of the findings.

Some in Congress are calling for the CFPB to repeal its warning to auto financers about the risks of discretionary markups. But rooting out racial discrimination is just too important to allow this dangerous practice to continue. Auto dealers deserve to be compensated for their efforts to originate an auto loan, but we can find a better method – one that does not result in higher charges for minority borrowers. A flat rate based on the size of the loan and not the race of the borrower would be a win-win for auto dealers and all their customers.