Are credit scores obsolete?
Big Data can do a better job of assessing credit risk than a credit report can: That is the premise of Affirm, arguably the most prolific of the silicon valley startups seeking to replace the credit report with Big Data fueled predictions about a loan applicant’s likely behavior. As the NY Times explained in an article in yesterday’s paper Affirm and other similar startups “ envision consumer finance fueled by abundant information and clever software — the tools of data science, or big data — as opposed to the traditional math of creditworthiness, which relies mainly on a person’s credit history.”
Big Data means many things to many different people but at its core it’s about finding correlations between varied and seemingly unrelated data sets. Computers are now big enough and fast enough to process huge amounts of information and find correlations that can be used to predict everything from voting behavior to potential terrorist activities. It makes sense that these same techniques can be used to predict who is more or less likely to pay back a loan. None of us is as unique as we think we are.
All of these startups have investment bankers anxious to apply Money Ball concepts to consumer lending. Investment banking is all about figuring out pockets in the market where the price of a good doesn’t reflect its true value. There are people out there that are great credit risks but don’t have the credit history to prove it. Think of how much safer student loans could be if lenders had some way of figuring out an 18 year-old’s repayment prospects.
It’s not a coincidence that one of Affirm’s first lending products the company is partnering with a bank to provide is an installment loan product that the USA Today described as follows: A young person needs a mattress but doesn’t have or want to use a credit card “Affirm’s answer is to let the person pay for the mattress over 3, 6 or 12 months, and while there is interest tacked on here too, the customer knows exactly how much he or she will have to pay, with no compounding interest. Loans are underwritten and serviced in association with Cross River Bank, a New Jersey bank.” Once they got their models perfected you can bet that Affirm won’t use credit scores at all when making these and other loans.
Affirm’s leadership includes Rajeev V. Date, formerly the CFPB’s Deputy Director, who argues that the new technology will make lending cheaper for millions of Americans. Affirm believes that there are many potential borrowers, such as the recent college graduate, out there whose credit scores don’t do them justice
The catch is that under existing law a lending system that has a disparate impact on a protected class is illegal even in the absence of discriminatory intent if a less discriminatory method could have been used to make lending decisions. One of the reasons why Mr. Date’s former employer wants to expand the amount of HMDA data lenders have to disclose is that it wants to scrutinize what criteria lenders are using on a much more granular level. Big data may very well allow lenders to make more accurate and subtle distinctions between borrowers but these subtle distinctions may be harder to justify than traditional credit distinctions. Does your loan officer really want to explain to a member that they didn’t get a loan because of the way they signed the loan document?
One more thing: This is not an abstract discussion analytics are getting cheaper and as they do all but the smallest institutions will be able to adopt them to their lending models. Here are links to the article’s from the Times and USA Today.