A thrift executive was attending a conference on credit scoring here this week with a single purpose: to find out whether his institution could pool loans with high credit scores into securities that would be inviting to investors.
He reasoned that, given investor acceptance, his thrift could get a higher price than by selling to Fannie Mae or Freddie Mac. And that extra margin - perhaps 10 basis points - could make a huge difference in the thrift's profitability.
This is just one example of the rising interest in credit-scoring technology in the mortgage industry for purposes that go well beyond the loan underwriting, the original application.
Patrick G. Culhane, executive vice president of Fair, Isaac & Co., said scoring technology was getting attention in areas such as mortgage collection, prediction of prepayment likelihood, evaluation of securities portfolios, and in targeting of cross-selling efforts.
The San Rafael, Calif., firm, a pioneer in credit scoring for consumer finance, sponsored the four-day conference, which ends today.
"Prepayments have financial ramifications for lenders and servicers, and they are driven heavily by external factors, but also by some internal ones as well," Mr. Culhane said. He added that sensitivity to rate changes could be measured with scoring techniques, and that a current credit profile could also be suggestive.
"There are opportunities to develop attrition or repricing models for mortgages," he said.
In securitizations, Mr. Culhane noted, Fair, Isaac is seeing increasing demand for scoring as a way of evaluating portfolios. Companies that rate mortgage securities, as Fitch Investors Service and Standard & Poor's, are pursuing the technology, he said.
Perhaps one of the rich but relatively untapped areas for the use of credit scoring is in cross-marketing. The normal procedure would be for a lender to do test mailings to generate response that could then be analyzed to provide the basis for targeting larger mailings. While this is similar to standard direct mail procedures, the scoring capability provides tighter targeting because it is based on a larger number of customer traits.
"One big area is predicting utilization of the financial services sold," Mr. Culhane said.
It's not just response to the offer, but "it's how likely it is that the person will use" the credit card, home equity line, or other service, Mr. Culhane said.
Fair, Isaac is also developing the ability to go beyond numeric values and do profiling that takes into account household demographics. "We can help the marketer arrive at a picture broader than what's provided by behavioral analysis," Mr. Culhane said. "We want to let marketers know not only who to mail to but what to tell them."
Mr. Culhane pointed out that credit bureau scores were of little use in targeting offers of first mortgages because of legal restrictions. "If you use credit scores to prescreen, then you must make a firm offer of credit," he explained. But firm offers of mortgage credit are obviously impractical.
While the appropriate use of scoring in making lending decisions remains a subject of debate, the migration of the technology into every segment of the mortgage business seems inevitable.
