Credit-Scoring Alternatives Sought

P&C personal lines carriers' recent reliance on credit scoring is increasingly threatened by legislation in a large number of states aiming to restrict and, in some cases, even completely ban its use.

"From a defensive perspective, credit scoring is our number-one issue," says Joseph Annotti, assistant vice president, public affairs, National Association of Independent Insurers (NAII, Des Plaines, IL). This year 28 states have considered some type of credit scoring-related legislation, Annotti says.

The industry's educational and lobbying efforts have helped limit restrictions, in Annotti's opinion, but 2003 promises to bring more anti-credit scoring legislation. In this climate, Annotti concedes, insurers will take a hard look at alternatives to credit scoring that permit accurate prediction without provoking concerns about discrimination.

'Black Box' Approach

Not sharing Annotti's perspective, Jack Gohsler, senior vice president, Conning & Co. (Hartford), sees the way technology has been used as characteristic of the way the industry has handled the credit scoring controversy in general. "When I tell you I know more about you than you know yourself because I've got this little 'black box' that takes data, grinds it up and spits out a credit score—but I won't tell you anything about it—that spells trouble," he says.

Solutions that brings both greater transparency and independence from the credit scoring controversy exist, according to John Lucker, senior manager, Deloitte & Touche Advanced Quantitative Services (D&T AQS, Hartford). He says the proprietary models developed by D&T AQS "are not provided to our clients as 'black boxes.' We work collaboratively, end-to-end, with our insurance clients to create multivariant, non-credit-based predictive models."

By "multivariant," Lucker means his firm's solution taps into up to 60 external, electronically available variables unrelated to credit, unlike traditional solutions. For new business underwriting, D&T's models fall short of the predictive power of credit-based models by "a very acceptable margin," Lucker says. But for renewals, he adds, "we can be at least as good, if not better."

While it insists that it hasn't given up on credit-based insurance scoring, San Rafael, CA-based Fair, Isaac & Co.—the leading provider of traditional credit-based predictive modeling—is taking appropriate measures in "the very small number of states" enacting significant credit-scoring restrictions, according to Lamont Boyd, the firm's insurance solutions manager. He says Fair, Isaac has always offered "custom-built models," and contends there is misunderstanding about the firm's offerings. "There is no 'black box,'" he says. (See related article on pg. 18.)

Wanted: New Vision

What's needed to address the credit-scoring problem is a new vision, according to Gary Preysner, director of Avon, CT-based Ironwood Consulting. For example, he suggests, anyone who understands how to use SAS's (Cary, NC) technologies, understands recent mathematical breakthroughs and has access to trained statisticians can develop models every bit as effective as credit-based models using "other specific demographic information available that can serve as an equivalent proxy."

Insurance companies' dependence on "what's been given to them by Fair, Isaac and others that have prepackaged information" amounts to a failure of vision, Preysner declares. "They haven't done a good job of taking raw data and making it proprietary," he claims. "They've just taken the same information as everyone else and integrated it into their underwriting, which guarantees they'll be no better than anybody else who can buy the data."

Conning & Co.'s Gohsler agrees, but says that such solutions are still vulnerable to privacy concerns. "When people recognize that you can put harmless pieces of information together and get useful stuff, pretty soon there's going to be a public furor to stop making that it available," he says.

Anthony O'Donnell has covered technology in the insurance industry since 2000, when he joined the editorial staff of Insurance & Technology. As an editor and reporter for I&T and the InformationWeek Financial Services of TechWeb he has written on all areas of information ... View Full Bio