Card Banks Making Ever-Finer Risk Assessments

The most successful credit card banks have long been admired for their savvy use of information technology.

And in this time of record personal bankruptcies, rising delinquencies, and more intense competition, the top credit card issuers are looking for ways to keep that edge.

This year, for example, three of the top four monoline credit card banks - First USA Inc., Capital One Financial Corp., and Advanta Corp. - have begun to install a risk management tool that can be used to measure and predict interest rate exposure.

The system, from Risk Management Technologies in Berkeley, Calif., is not aimed at analyzing credit quality. Rather, it can gauge the value of a card portfolio based on changes in interest rates, balances, and attrition rates.

The three banks would not comment. But Jay Wheeler, a senior consultant at Risk Management Technologies, said the system can be used to determine the potential profitability and value of card solicitations.

He cited as an example an offer of a 5.9% introductory rate for one year. "They fully expect to lose money on these cards in the first year," said Mr. Wheeler. "But what they are trying to predict is the overall value of these cards. How many of them are going to stay once the teaser period ends? What's the true economic value?"

The system, called Radar, runs so-called Monte Carlo simulations, an analytical technique in which changes in cost or value are calculated simultaneously for numerous interest rate or market scenarios.

"You need to be able to look at your portfolios and understand the rate sensitivities and be able to segregate these things at very fine levels so that you are not applying gross assumptions across" the entire portfolio, said Mr. Wheeler.

First Data Corp. is also stepping up its offerings of collection and profitability tools through its subsidiaries and alliances with other suppliers like Fair, Isaac & Co. and HNC Software Inc.

One recently introduced tool employs a proprietary model to predict the likelihood a cardholder will declare bankruptcy.

And First Data Card Services Group in Omaha is marketing a sophisticated new profitability scoring tool developed by HNC.

The system, which uses neural net technology, is designed to give card issuers a more complete picture of customer profitability. The technology combines historical data, actual customer behavior, and revenue information to predict the profitability of each account.

Gary Rutledge, senior vice president of issuer risk management at First Data, said, "That allows banks to manage to what they believe might be the likelihood that customers will either attrit, go delinquent, or charge off at some point in time."

For reprint and licensing requests for this article, click here.
MORE FROM AMERICAN BANKER