March 12, 2003
University of Texas Study Links Poor Credit, Insurance Losses
|
Austin – A study released from the Bureau of Business Research at the University of Texas at Austin shows a "statistically significant" relationship between a person's credit history and tendency to incur losses on an auto insurance policy. "In general, lower credit scores were associated with larger incurred losses," reported the study, which Texas State Sen. William Ratliff commissioned in June 2002 to examine the relationship between credit scores and incurred losses.
The practice of using credit histories to adjust insurance rates has stirred controversy across the U.S. Legislators in Kansas, Nevada, and Texas are exploring curbs on the practice. In January, the state of Washington issued a change in insurance rules forbidding the use of a person's credit history to judge risk.
The Texas study adds to the debate. Researchers initially collected data on policies written in 1998, including credit scores of named insured on 153,326 policies from five participating insurance groups doing business in Texas. They then tracked actual losses from claims associated with the policies.
The 10 percent of policyholders with the lowest credit scores had loss ratios (the ratios of losses to premiums) about 53 percent higher than expected. Overall, all three groups with the lowest credit scores had claim losses above the targeted loss ratio. The 10 percent of policyholders with the highest credit scores had loss ratios about 25 percent lower than expected.
Since premiums already reflect characteristics such as age, gender, and make and model of car, the results show that credit scoring yields significant new information about loss propensity not already included in the underwriting process.
The study further revealed that credit scores can help predict the size of claims. The average loss per policy was $695, but the group of policies with the lowest credit scores averaged $918. The group with the highest credit scores averaged $558.
The study factored out losses explained under existing underwriting variables such as age, geography and type of car. At the same time, variables not used in underwriting were excluded, such as race, ethnicity or income, nor did the authors attempt to explain why credit scores correlate to insurance losses.
Asked if the results were applicable outside of Texas, co-author Dr. Patrick Brockett noted that, "Texas drivers tend to be representative. We have a mix of urban and rural areas that tends to be very similar to the nation's market."
The study is online at: http://www.utexas.edu/depts/bbr/bbr_creditstudy.pdf.
Brockett and co-author Bruce Kellison of the Bureau of Business Research testified before the Texas state senate on Tuesday, Mar. 11 in Austin.
Patrick Brockett can reached for comment at brockett@mail.utexas.edu, (512) 471-6028 - office, (512) 293-1009 - cell (recommended).