CFPB Wants To Know How 'Alternative Data' Changes Credit Scores


New methods of determining credit scores, which might involve an applicant's web history, could harm consumers.

The new data sources some financial services providers are now using to calculate credit scores could ultimately harm consumers, the Consumer Financial Protection Bureau fears.

CFPB is looking for more information about how modern credit processes—credit authorization, debt collection and others—incorporate "alternative data" instead of the standard factors such as credit limit information, debt repayment, credit inquiries and income.

Looking beyond those traditional data sources to alternative ones, such as an applicant's online shopping behavior, social media connections or utility payment history, could allow financial institutions to assess the credit risk of consumers who previously didn't have enough information to fill out a traditional credit file. 

» Get the best federal technology news and ideas delivered right to your inbox. Sign up here.

"Potentially millions of consumers previously locked out of mainstream credit could become eligible for credit products that might help them buy a car or a home," a CFPB request for information said. While newer calculation methods "hold the promise of potentially significant benefits for some consumers," they also may "present certain potentially significant risks."

Roughly 45 million people in the U.S. don't have files with major credit bureaus or are "unscorable" because they lack the traditional data required to calculate a score; disproportionate number of those people are black, Hispanic or low income.

Financial institutions also might be able to assess the risk of those applicants by using big data analytics techniques, tapping into the large amounts of information about similar applicants instead of requesting those numbers from the applicant directly.

Traditional modeling techniques might include algorithms and regression models; "alternative modeling" might refer to "artificial neural networks" and other newer approaches to decision-making.

But those more advanced techniques also bring a "lack of transparency, control and ability to correct data that might result from their use; potential infringements on consumer privacy; and the risk that certain data could dampen social mobility, result in discriminatory outcomes, or otherwise disadvantage certain groups, characteristics, or behaviors."

A White House report on big data last year warned against "algorithmic discrimination," which might lump an individual into a larger group of subjects just because they share one characteristic, such as shopping at the same grocery store or living in the same neighborhood.

"An increasing ability for lenders to accurately assess risk could reduce the price of credit for those who are shown to be good risks," CFPB noted, though it could also "increase the price of credit for those shown to be worse risks."

Eventually, CFPB aims to be able to mitigate the risks of these new techniques and data sources, according to the agency.