LAS VEGAS — There’s a new auto finance compliance risk: unintentional data “redlining.”
Brian Donohue, executive vice president of analytics and marketing at Exeter Finance Corp., warned that auto finance lenders should be wary of unintentional redlining of customers when adopting alternative data sources.
“Redlining” is defined as the practice of denying services — generally loans — either directly or through selectively raising prices, to particular people because of their race or ethnicity.
“The CFPB has really pushed all of us to increase our oversight of vendors, and really understand who we are doing business with,” he said during a panel at the AFSA Vehicle Finance Conference here yesterday. “You have to keep that in mind, when establishing relationships with data providers.”
Donohue suggested first testing out a new system to see whether it generates a skew in population.
“It has to be anonymous, as far as ethnicity, and it has to be geographically indifferent, meaning, when you start looking at geographic characteristics like average income, or population, you start creating lines that you are not supposed to create,” he said. “Stay clear of that, and work with your legal departments.”
In most cases, alternative data does help lower underwriting risk and make prudent decisions. Some of the recent trends in the space, Donohue told Auto Finance News, are data points pulled from payday lending.
“There is a large segment of population that uses that product [payday lending], and it doesn’t make it into your traditional credit bureau data sources,” he said. “That product became very prevalent, and the fact that some data providers collect that information to provide alternative credit history for that population is just a large trend.”