Alternative data model boosts credit access, says CFPB NAL recipient
On August 6, the CFPB published a blog providing an update on credit access and the Bureau’s first-issued No-Action Letter (NAL), and reporting that use of alternative data in underwriting may expand access to credit. In 2017, the CFPB announced its first NAL to a company that uses alternative data and machine learning to make credit underwriting and pricing decisions. One condition for receiving the NAL required the company to agree to a model risk management and compliance plan, which analyzed and addressed risks to consumers and the real-world impact of its service. Through specific testing, the company worked to answer two key questions: (i) “whether the tested model’s use of alternative data and machine learning expands access to credit, including lower-priced credit, overall and for various applicant segments, compared to the traditional model”; and (ii) “whether the tested model’s underwriting or pricing outcomes result in greater disparities than the traditional model with respect to race, ethnicity, sex, or age, and if so, whether applicants in different protected class groups with similar model-predicted default risk actually default at the same rate.”
According to the Bureau, the company reported that in the access to credit comparisons, the alternative data model approved 27 percent more applicants as compared to a traditional underwriting model, and yielded 16 percent lower average APRs for approved loans, with the expansion in access to credit “occur[ing] across all tested race, ethnicity, and sex segments.” For the fair lending testing, the company reported that no disparities were found in the approval rate and APR analysis results provided for minority, female, and older applicants. Additionally, the company reported significant expansion of access to credit for certain consumer segments under the tested model, including that (i) “consumers with FICO scores from 620 to 660 are approved approximately twice as frequently”; (ii) “[a]pplicants under 25 years of age are 32 [percent] more likely to be approved”; and (iii) “[c]onsumers with incomes under $50,000 are 13 [percent] more likely to be approved.” The Bureau noted that the testing results were provided by the company, and the simulations and analyses were not separately replicated by the Bureau.