Use of an “alternative data” model for determining credit worthiness has resulted in 27% more applicants being approved, and 16% lower interest rates, than traditional models have recorded, the federal consumer financial protection agency said Tuesday.
In an update on the results of its 2017 “no action letter” to a company that uses alternative data and machine learning in making credit underwriting and pricing decisions, the Consumer Financial Protection Bureau (CFPB) reported the approval and interest rate results. (A “no-action” letter is intended for “innovative financial products or services that promise substantial consumer benefit where there is substantial uncertainty about whether or how specific provisions of certain statutes implemented or regulations issued would be applied.” For example, it may apply, according to the bureau’s policy statement on the practice, if, because of intervening technological developments, the “application of statutes and regulations to a new product is novel and complicated.”)
The bureau said the letter, to Upstart Network Inc. (which uses traditional underwriting data and various categories of alternative data, including information related to borrowers’ education and employment history, in underwriting and pricing) required the company to provide information about how its model, in use, compared to more traditional models of underwriting and pricing. In exchange for providing the information, CFPB noted, the company was given leeway in the application of the Equal Credit Opportunity Act (ECOA) and its implementing regulation, Regulation B, to Upstart’s use of alternative data and machine learning for its model.
CFPB said that the reported expansion of credit access was reflected across tested race, ethnicity and sex segments, resulting in increased acceptance rates of 23% to 29%, and decreasing interest rates (annual percentage rates [APRs]) between 15% to 17%.
“In many consumer segments, the results provided show that the tested model significantly expands access to credit compared to the traditional model,” the bureau stated. In particular:
- “Near prime” consumers with FICO scores from 620 to 660 are approved approximately twice as frequently.
- Applicants under 25 years of age are 32% more likely to be approved.
- Consumers with incomes under $50,000 are 13% more likely to be approved.
“With regard to fair lending testing, which compared the tested model with the traditional model, the approval rate and APR analysis results provided for minority, female, and 62 and older applicants show no disparities that require further fair lending analysis under the compliance plan,” the CFPB stated.