Does having bad credit make you a worse driver or a riskier homeowner?

No, but your premium bills might make you believe that your insurance company thinks so.

According to a study by Conning and Co., more than 90 percent of auto insurance companies, and an increasing number of home insurers, use your credit information, filtered through a formula to create an "insurance risk score," to determine how likely you are to file a claim on an insurance policy. More than half of those insurers use that information to determine how much to charge you in premiums.

Insurance risk scores are similar to credit risk scores — used by lenders to determine whether or not to approve a loan or line of credit — because both look at your credit information, but the two are not the same thing, says Craig Watts, a spokesperson for Fair, Isaac, and Co., whose insurance risk scores are used by about 300 insurers nationwide.

"Consumers are becoming more familiar with credit risk scoring, but insurance risk scoring is still fairly arcane," says Watts.

While both insurance scores and credit scores look at the same five characteristics of a person's credit report (see list at right), the data are weighted differently. This difference in weighting can swing 5 to 10 percent in each category.

"The biggest difference is that insurance risk scores look for stability, but credit risk scores look for a reliable pattern," says Watts. "Insurance scores are also more interested in how regularly you pay than in how much you already owe."

Insurers use these insurance scores to try to identify consumers who are consistent and reliable, as well as those who show a pattern of demonstrating common sense with money. Insurers say these people are less likely to file a claim on an insurance policy.

"We've studied millions of records and have found that there is a clear and reliable correlation between credit history and insurance risk," says Watts.

Allstate Insurance Co. and State Farm Mutual Auto Insurance Co., the nation's two largest auto and home insurers, have also noted this correlation and have developed their own insurance risk-scoring systems that incorporate credit information.

"We went in and looked at our data . . . to see if we could find differences between groups of people," says Mike Trevino, a spokesperson for Allstate. That analysis showed that people who have better creditt — at least as reflected in their insurance scores — tend to file fewer claims, thus costing the insurer less money.

"Our feeling is that using credit information allows us to more fairly price our insurance," says Trevino. "Those that have better credit pay a lower rate."

Dick Luedke, a spokesperson for State Farm, which uses credit information only in deciding whether or not to issue an insurance policy, points out that in some cases use of credit information has allowed State Farm to cover people that wouldn't ordinarily have qualified.

"Study after study has shown that credit history can be correlated with the likelihood that someone will file a claim," says Luedke. "We don't claim to have the definitive answer as to why there is a correlation, but we believe one exists."

Correlations can't explain why insurance scoring works

That reasoning for using insurance risk scoring infuriates Georgia Insurance Commissioner John Oxendine, who is also a member of the Consumer Protection Working Group of the National Association of Insurance Commissioners.

"I hear a lot of talk about correlation, but no talk about causation," says Oxendine. "Insurers don't have any reason for why scoring works, they just say 'correlation' over and over."

While Oxendine acknowledges that credit information — which has been used by some insurers for more than a decade — can be useful to insurance companies for avoiding insurance fraud-motivated arson and similar hazards, he places little faith in computer-modeled insurance scores and statistical relationships.

"If you punch enough numbers through a computer you can get anything," says Oxendine. "It's time we learn about how it works and make sure the criteria they use are in the best interest of the public."

So far Oxendine has had little success in getting insurers to divulge their methods for calculating insurance risk scores. "The information currently available to consumers seems designed to limit their understanding of [insurance] scores," says Oxendine. "If you don't know the rules of the game, you can't protect yourself."

It isn't quite that simple, says Dan Kummer, an expert on credit information issues and the director of auto insurance for the National Association of Independent Insurers, a property/casualty trade organization that has been an outspoken advocate for the use of credit-based insurance risk scores for selling and pricing insurance.

The computer models used to generate insurance scores from credit information represent a tremendous investment of time and money for insurers and they don't want that proprietary information to be leaked to other companies, says Kummer.

"This is very disturbing — it's like a black box," counters consumer advocate J. Robert Hunter, the director of insurance for the Consumer Federation of America (CFA) and a former Texas Insurance Commissioner. "They haven't verified that minorities, people with disabilities, and the poor aren't discriminated against by these systems."

According to Kummer and Watts, insurance risk-scoring models do not discriminate. "In the studies we've done, we looked specifically at the scoring of low- to moderate-income and high minority areas," says Watts. "People in those areas score the same as in areas of higher income. We didn't see a pattern of indirect discrimination."