Here's an alternative to credit scores
WASHINGTON — Digital Risk, a mortgage analytics firm, is mounting an unusual frontal assault on one of the lending industry's sacred cows. It argues that credit scores such as FICO failed to predict large numbers of defaults during the mortgage bust years — notably thousands of "strategic" walkaways by borrowers with high scores — because they could not anticipate homeowners' reactions to economic stress. Unless lenders use more sophisticated assessment tools that incorporate far more than credit histories, Digital Risk says, they may be misjudging not only many of today's high-risk borrowers but other applicants who are safer bets than their credit scores suggest.
"The mortgage industry is relying on outdated methods to determine risk," says Peter Kassabov, chairman and chief executive of Digital Risk, which is based in Maitland. "During the mortgage crisis, high-FICO borrowers encountering distress defaulted in huge numbers, yet we still depend heavily on that one score along with (down payments) to make lending and loan modification decisions."
According to one study conducted in 2009, 588,000 homeowners walked away from their homes strategically during 2008. That's 18 percent of all serious defaults that year and shocked the mortgage industry. Fair Isaac, creator of the FICO score, acknowledged the problem, and last year released an "analytic tool" that lenders can use to detect potential strategic defaulters — high-scoring, credit-savvy borrowers primarily — before they stop paying.
Digital Risk describes itself as the "nation's largest provider of mortgage risk, compliance and transaction management solutions," and claims to have seven of the top 10 mortgage lenders as active clients. In early August it introduced a multidimensional risk evaluation system it calls "Veritas," which it claims integrates borrower credit characteristics with property and local real estate market data along with proprietary behavioral prediction models. The behavioral component includes what the firm calls statistical "clusters" of borrower, property and market situations — 123 in all — that give lenders a better idea of how an applicant will react to financial problems.
The system is based on analyses of more than 5 million loans originated between 2006 and 2011, plus a separate study of how 100,000 borrowers performed after having their loan terms modified, according to the company. Alex Santos, president and co-founder of Digital Risk, said in an interview that the models have been "validated" on large batches of clients' loan files. Veritas separated out applicants destined to default in a future period of financial stress from those likely to keep paying on time, even when credit scores and other data were similar.
The value of this for mortgage applicants whose scores don't meet today's high requirements is significant. For example, according to Santos, two home buyers with identical 690 FICO scores and down payments might be rejected — Fannie Mae and Freddie Mac both have average FICOs in the 760 range. Yet using the Veritas system, one of them could be identified as a safe bet and the other a future disaster.
Fair Isaac isn't taking critiques. Anthony Sprauve, a FICO spokesman, said, "We continually work with our customers to make sure the FICO score is the best predictor of a person's likelihood to repay a debt. Our customers vote with their feet since, according to (research firm) Tower Group, lenders ask for FICO scores more than 90 percent of the time when buying scores from the big credit bureaus."