Although nearly half of the states have passed laws allowing insurance companies to use credit-based insurance scores to assess and price risk-albeit with some restrictions-a new study reveals that the practice disproportionately harms low-income and minority policyholders."There was a tremendous demand to know what happened to low-income folks and protected classes like minorities when you use credit scoring," says Randy McConnell, communications director for the Missouri Department of Insurance (MDI), which conducted the study. "It struck people as disingenuous whenever the credit-scoring vendors were saying, 'There's no relationship between credit scoring and income.'"
In fact, supporters of credit-based insurance scores cite two studies when making this claim:
- The American Insurance Association-now Des Plaines, Ill.-based Property Casualty Insurers Association (PCI)-in 1998 released a study that found no correlation between income and credit scores.
- The Virginia Bureau of Insurance in 1999 claimed a study it conducted found nothing to conclude that income or race alone is a reliable predictor of credit scores.
Opposite conclusion
This newest study from MDI, however, came to the opposite conclusion. MDI used credit score data from 12 automobile and homeowners writers in Missouri for the period from 1999 to 2001. The study found that insurance credit-scoring produces significantly worse scores for residents of high-minority ZIP codes in Missouri (see chart, page 9), as well as for residents of low-income ZIP codes.
Specifically, the average credit score rank in "all minority" areas was 18.4 out of a possible 100, compared with 57.3 in "no minority" neighborhoods-a gap of 38.9 points. And, the average credit score rank in the poorest 5% of ZIP codes was 40.9 out of 100, compared to 53.6 in the wealthiest 5% of ZIP codes-a gap of 12.8 points.
MDI's research also showed the relationship between minority concentration in a ZIP code and credit scores exists even after eliminating other socioeconomic variables, such as income, educational attainment, marital status and unemployment rates.
"The concern is that credit scoring is unfairly penalizing low-income citizens with inflated insurance prices, with much of the burden falling on African Americans and Hispanics with higher insurance prices," said Democratic Gov. Bob Holden, who urged the Missouri Legisla-ture to prohibit the use of credit scoring in establishing the price of auto and homeowners insurance.
However, in 2002 and 2003, the Missouri legislature rejected bills that proposed banning the use of credit scores for insurance purposes.
Instead, it enacted a law that allows insurers to use them, but not as the sole reason to cancel, non-renew or refuse to issue a policy. The Missouri law also requires insurers to provide state regulators with actuarial justification for rate increases or decreases based on a consumer's score.
Flawed results
The Property Casualty Insurance Association (PCI) contends the results of the Missouri study are flawed because of its overly simplistic methodology. According to PCI, the study only examines insurance score data aggregated by ZIP code and does not take into account policyholders' loss experience.
In fact, several studies have shown a strong correlation between credit history and the likelihood of future loss. One study conducted by the Casualty Actuarial Society, based on 170,000 premium records, showed that policyholders in the worst credit category had an average loss ratio of 101.4%, while those in the best category had a 57.4% loss ratio.
"We believe that the action taken by state legislators in Missouri-and in dozens of other states (allowing insurers to use credit scores with some limitations) is in the best interest of all consumers because it protects policyholders and upholds the fundamental principle of risk-based pricing," says John Lobert, senior vice president of state legislative affairs at PCI.
PCI also questions the intent of the Missouri study. "The study was clearly intended to support the Department of Insurance's view that insurance scores should be banned," Lobert says.
But MDI's McConnell says the study's methodology is not flawed. Its purpose was to focus on the impact of insurance scores on low-income and minority communities-not on risk of loss. Furthermore, he says, MDI made no recommendation to ban insurance scores when it released the report.
In addition, he notes, this is a public policy issue-not just an insurance issue, McConnell says. "Back in the 1960s, it was actuarially sound to charge blacks more than whites for life insurance because they didn't live as long," says McConnell. "And we decided that's not the way we wanted our society to operate. All of us will end up paying a lot more if we drive low-income and minority folks out of the auto insurance market. We have areas of the state that have tremendously high uninsured vehicle rates."
Despite the controversy, both PCI and MDI representatives don't expect the Missouri legislature to ban the use of the scores as a result of this study.
"The General Assembly has not been hospitable to consumer protection in regard to insurance matters," says MDI's McConnell.
At PCI, "our sense is that the (Missouri) state legislature feels it has already dealt with this issue," says Joseph Annotti, PCI vice president of public affairs.
