Insurers Say Data Shows Low Credit Scores Predict Risky Driver Behaviors

The insurance industry says it now has hard data proving that if you have a bad credit score, you’re more likely to slam on the brakes or step hard on the gas.

Also if you’re less educated, unmarried, or sell real estate for a living.

The American Property Casualty Insurance Association this week released a white paper that it says validates auto insurance pricing decisions based on credit scores, marital status, education level and occupation. The research paper says telematics data collected from January 2017 to June 2019 shows that two behaviors that are known to correlate with increased claims cost — hard braking and hard accelerating — also correlate with insurers’ pricing practices.

Dr. Robert Hartwig, the University of South Carolina professor who wrote the paper, said previous studies have shown that credit scores accurately predict losses. Hartwig is a former director of the Insurance Information Institute.

“The contribution to the literature here is now there is a behavioral linkage that is demonstrated between two hazardous activities to be involved in and loss among individuals who have specific credit attributes,” he said during a telephone interview on Tuesday.

Consumer advocates say that pricing decisions based on credit scores are inherently unfair because they penalize people for events they cannot control, such as loss of a job or a death in the family. The use of credit-based insurer scores is banned in California, Hawaii, Massachusetts, Michigan and Washington.

“My view is that one’s credit score is often beyond one’s control and can be adversely impacted by things like being born to a low income household, losing one’s job, getting sick, or just not being ‘in the system,’” said Amy Bach, executive director of United Policyholders, in an email. “My 23-year-old employed daughter is having a very hard time getting a credit card for no discernible reason other than she’s 23 and never had one before.”

Bach pointed to a paper by the Consumer Financial Protection Bureau, which happened to be released Tuesday, showing that consumers in majority Black and Hispanic neighborhoods are most likely to have disputes appear on credit reports.

The insurance industry, on the other hand, points to a body of research that shows lower credit scores are associated with a higher risk of loss.

The National Association of Insurance Commissioners reported in 2020 that insured drivers with the lowest credit scores had the greatest amount of losses. The Vermont Department of Financial Regulation reached a similar conclusion after studying the issue at the request of state lawmakers in 2016, as did the Texas Department of Insurance in 2004 and the Federal Trade Commission in 2007, according to the APCIA.

The APCIA’s white paper shows graphs that depict the “hard brake relativity” and “hard acceleration relativity” for drivers with low credit scores and high credit scores. Drivers with the lowest credit scores cost 28% more to insure than drivers with the highest scores, based on those behaviors.

Education levels showed a similar correlation. People with graduate degrees had the lowest relative cost to insure, based on hard braking behaviors, at about 0.9. College graduates rated about 0.95. People with only a high school education, didn’t graduate high school or had some college clustered together at about 1.1.

Occupation also predicted driving behaviors. Librarians were low risk with a relative cost to insure about 0.9, while real estate agents had a relative cost of 1.1.

Population density had the greatest impact on driver behavior, with people in the least dense areas enjoying a relative cost to insure of about 0.8, compared to about 1.2 for drivers in the most dense areas, the paper says.

Hartwig says the telematics data adds “a missing behavioral link” to the other studies that shows the reasoning that goes into charging certain groups higher premiums. He said the credit-based insurance score is only one of many variables that will be used to determine the amount of premiums that any individual will pay.

“The rationale for maintaining these rating variables is that people pay a premium that more closely resembles their actual losses,” he said. “That is unequivocally fair to everybody.”

Bach, the policyholder advocate, said she’s skeptical because numbers can be used to show correlations in just about anything. But she said she can’t refute the APCIA’s data. Her main point continues to be consumers shouldn’t pay more for insurance simply because they have less access to credit than others.

“Use something else please, because it’s just not fair,” she said. “Since insurers seem to have access to tremendous amounts of data, they can use their sources, but not this.”