NAII Says Washington OIC Study is Inconclusive on Scoring Debate

January 30, 2003

The findings of a recent survey on the use of insurance scores commissioned by the Washington Office of the Insurance
Commissioner (OIC) are inconclusive and do nothing to dispel insurer
contentions that such scores are a valid and accurate underwriting and rating tool, according to the National Association of Independent Insurers (NAII). Moreover, the methodology of the study reportedly raises serious questions about the report’s results.

The study was conducted as a part of a new law imposing significant
restrictions on the use of insurance scores enacted in Washington in 2002. The OIC was required to report to the Legislature the effects of insurance scoring on auto insurance underwriting and pricing. A follow-up study on how insurers are implementing the new law is due in January 2004.

“The OIC concluded that the results of its study were ‘too varied for a clear pattern to emerge,'” Diana Lee, senior research consultant for the NAII, said. “Our analysis of the survey methodology and results reveals severe limitations that make drawing overall conclusions premature and cast doubt on the validity of the allegations about the impact on insurance scoring made by the OIC. To date, there is no solid evidence to suggest the use of insurance scoring has a substantially disproportionate impact on specific ethnic or economic groups.”

According to the NAII, the danger of this study is that legislators and
regulators in other states may rely on this report – with its limited scope and inconclusive results – to formulate public policy recommendations. “Indeed, the study itself reports that ‘an overall conclusion that credit scoring generally does or does not have a particular, consistent, quantifiable, unequal negative effect on certain demographic groups is premature,'” Lee said.

The stated purpose of the report was to find out whether industry-wide use of insurance scoring has unequal impact on specific demographic groups. Of particular concern in this study was the impact of insurance scoring on ethnic minorities and low-income individuals. After reviewing the findings submitted to the legislature, NAII noted several major shortcomings of the study.

“The report only examined three companies. With the small sample of
companies studied and the variation in how insurance scoring is used among insurers, the report has very little value for projecting its conclusion across the wide spectrum of policyholders and insurers doing business in the state. To their credit, the researchers correctly point out that the small minority population in Washington limits the validity of the findings regarding ethnicity. According to the U.S. Bureau of Census, Asians make up 6 percent of the state’s population, African-Americans 3.5 percent and Native Americans 1.8 percent,” Lee remarked.

“Moreover, even though the three insurers provided different types of
information, the same general statistical methodology was applied to whatever data the researchers obtained,” Lee continued. “In essence, this resulted in the researchers developing three sets of conclusions based on the data from individual companies. This approach makes the conclusions even less credible, since rather than looking at a total of about 3,000 policyholders, the researchers basically conducted three separate studies using only 1,000
policyholders each.”

Regarding income, the report found that there were varied results among the insurers studied. Based on the data from two of the companies, the study states that no definitive conclusions could be reached about possible demographic differences other than age. “The process of developing insurance scores does not consider factors such as age, race and income level. This is a very objective process that uses data that has been proven to correlate with risk of loss,” Lee said.

In its analysis, NAII noted that the report does not attempt to take into account information regarding loss history or driving records of the policyholders surveyed. Insurers were asked to provide only age, gender, zip code, effective date of the policy and insurance score or rate classification.

Other reported inaccuracies and misleading statements in the report according to the NAII include:

· The report claims that data in credit reports are “often” inaccurate without mentioning that many inaccuracies are irrelevant to insurance scoring. One NAII member company reported ordering 17 million credit reports in 2001 with fewer than 0.17 percent of the reports disputed by consumers.
· The report claims the dispute process is “cumbersome and time consuming,” but in fact, the federal Fair Credit Reporting Act (FCRA) imposes stringent time requirements and penalties for consumer reporting agencies.
· The report contains anecdotal evidence of Washington residents without any confirmation of the facts or the evaluation of the merits of their allegations. For every anecdotal incident there are literally thousands of other consumers who are paying less for their auto insurance because insurers considered their insurance score.
· The methodology section of the report claims that consumer reporting agencies sell insurance scoring services to insurers. This is not always the case. A 2002 survey of NAII member companies indicated that 20 percent of responding insurers either create their own model or use a customized model developed with outside consultants.
· The report indicates that in 2002, 26 states considered restrictions on the use of credit history. However, the report fails to mention that most states did not impose such restrictions, while others enacted laws allowing insurers to use insurance scores provided the use of such scores were disclosed to consumers. Only Maryland enacted a measure more restrictive than the law passed in Washington.

Was this article valuable?

Here are more articles you may enjoy.