Banks Adopt a Fast-Follower Approach to Alternative Data, But Pitfalls Remain

(New York, NY):  Until now, banks have taken a cautious approach to alternative data – non-traditional credit scoring information like rent and cell phone payments – as questions around consumer privacy and disparate treatment swirled.

But after ceding a key competitive advantage to fintech companies in recent years, several mass market and regional banks are testing alternative data for use in marketing, pre-screening, and underwriting consumer loan products like credit cards and personal loans, and exploring emerging use cases in fraud prevention and other areas.

While most lender decisions still use traditional sources, such as credit file data managed by the consumer reporting agencies, the potential for alternative data to transform lending is clear. Rent and utility payments, employment information, and behavioral data paint a clearer picture of a consumer’s creditworthiness and could allow lenders to reach underserved populations. In fact, alternative data has allowed Lending Club, a marketplace lender, to extend lower-priced credit to borrowers that would otherwise be classified as subprime, according to the Federal Reserve.

“Banks are looking for new avenues to increase approvals and make better credit decisions, and alternative data is increasingly seen as a means to do both,” said Diana Middleton, who manages several of Auriemma’s business-to-business information services groups. “There are emerging use cases across the full credit process, not only for underwriting but also to better inform prospecting and marketing.”

Later in the credit lifecycle, banks are exploring the use of alternative data for fraud prevention, account management, and credit line increases. For example, alternative data is a potential solution to thwart synthetic identity fraud, a form of application fraud that costs lenders billions of dollars per year, according to Auriemma. An identity with employment information, payroll accounts, and utility records, for instance, is much more likely to be legitimate.

Banks are at different stages of development and adoption. For example, some subprime specialist issuers have been using alternative data to underwrite “thin-file” applicants for years, while mass market issuers are just beginning to analyze the predictive value of alternative data. However, increased competition from fintechs, which were early adopters, is expected to spur market acceptance. As traditional banks increasingly go up against non-bank lenders, particularly in the personal lending space, the use of alternative data in the credit process may become table stakes.

Still, even as the industry works to harness alternative data in new ways, questions around regulatory compliance and consumer privacy remain. Integrating more complex data and modeling into the credit process could make it more difficult to explain what factors led to a decision. Lenders are required to provide such explanation to applicants in certain circumstances.

The use of alternative data also carries fair lending risks. For example, variables that predict a consumer’s loan performance may also be highly correlated with race, ethnicity, or education level, potentially leading to disparate impact. Meanwhile, the revelation that Facebook mishandled user information for 87 million accounts has raised fresh privacy concerns and questions over how companies use consumer data.

“The use of alternative data involves risk and opportunity,” Middleton said. “But lenders are working to overcome these challenges.”

For more information, call Diana Middleton at (212) 323-7000.