There was a notable absence at the American Financial Services Assn.’s annual vehicle-financing conference in San Francisco this year.
No one from the Consumer Financial Protection Bureau showed up, unlike the last couple of years. I got the impression they weren’t invited, and wouldn’t have come anyway.
The federal regulator wants to eliminate dealer reserve, the practice of dealers adding a percentage point or two to an auto loan interest rate as compensation for arranging the loan.
Last year, CFPB’s Patrice Ficklin addressed a packed room at the AFSA event. In 2013, Richard Hackett spoke, attempting to explain the agency’s thinking.
Since those appearances, AFSA commissioned a scholarly study conducted by the highly regarded Charles River Associates.
The study isn’t kind to CFPB’s suspect methodology called Bayesian Improved Surname Geocoding (BISG). CFPB used it in an analysis to “prove” disparate impact or unintended discrimination against minorities obtaining car loans through dealerships.
One gets the impression CFPB zealots see discrimination behind every tree, and will stop at nothing to “prove” it.
The bureau used BISG in an attempt to determine borrowers’ race and ethnicity based on information from automotive lenders, who had no idea about the borrowers’ ethnicity. BISG uses last names and zip codes to make its determinations.
AFSA’s commissioned study calculated BISG probabilities against a test population of mortgage data, where race and ethnicity are known. Among the findings:
- When the proxy uses an 80% probability that a person belongs to an African American group, the proxy correctly identified their race less than 25% of the time.
- Applying BISG on a continuous method overestimates the disparities and the amount of alleged harm and provides no ability to identify which contracts are associated with the allegedly harmed consumers.
It’s a huge stretch to say race and ethnicity is known in the case of mortgage data, because mortgage applicants are asked to “self-describe” their race and ethnicity without any objective standard.
For example, if a person is one-fourth Native American, one-fourth Asian, and one-half African American, what objectively is his or her race? People have a variety of motivations for how they answer such a question. With no objective standard, the margin of error is impossible to calculate.
Yet, no lender will stand and fight CFPB in court. CFPB knows they have an immense intimidation element working in their favor, and they leverage it at every turn.
That said, CFPB has done a lot of good in some areas. But attempting to prove discrimination where there is none could cause a backlash that overwhelms the good.
David Ruggles is an automotive consultant and former dealership general manager. He can be reached at [email protected].
