Did you know that our credit scores affect our insurance rates? The Subcommittee on Oversight and Investigations of the House Committee on Financial Services is holding hearings on credit-based insurance rates. The FTC released a report this past July that, according to an FTC press release,
found that [credit] scores are effective predictors of the claims that consumers will file and that, on average, African-Americans and Hispanics tend to have lower scores than non-Hispanic whites and Asians, and so the use of scores is likely to increase the amount they pay for automobile insurance relative to the amount that non-Hispanic whites and Asians pay.
The methodology of that study as been questioned, including by dissenting commissioner Pamela Jones Harbour. The gist of the methodological challenge is that the FTC used data supplied by the industry, and those data are not reliable. Today Commissioner J. Thomas Rosch of the FTC is testifying on the issue and defending the July report. Among other things, Commissioner Rosch notes that other studies also correlated low credit scores and high numbers and amounts of insurance claims. But are credit scores merely a proxy for race? Rosch says no. The bottom line, he claims, is that "the use of effective risk prediction techniques, including credit-based insurance scores, decreases premiums for less risky consumers and increases premiums for more risky consumers."
This practice is part of a wider set of practices in which credit reports are used for "off-label" purposes. As noted by Katie Porter at Credit Slips:
Credit reports are also widely used as a screen for employment. Again, employees won't generally be loaning their employers money. Instead, the report is used as a rough measure of whether the potential employee is under financial strain, since this is perceived to increase the likelihood that the employee engaging in theft. (I suspect the real harm to an employer is actually that the stress and anxiety that the employee suffers that distracts them from their job duties). What's next? Using credit reports as part of college admissions? As part of E-Harmony's 20 gazillion point matching process? I predict "off-label" use of credit reports will become increasingly common and contentious in the next few years. The practice raises many of the same ethical and practical issues that trouble this practice in the pharmaceutical context. The existing legal regime, FCRA, wasn't designed to prevent the potential harms from this practice, and it may not be reasonable to expect consumers to understand the myriad impacts of their credit behavior on their life. Further, the only way that new uses for credit reports are discovered is by "experimenting" and looking for correlations. Who wants to be a guinea pig for this practice? Are there privacy concerns implicated in sharing a credit report--say with a bunch of delinquencies to hospitals reported--with Blue Cross and Blue Shield?
Food for thought.
TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8345157d569e200e54eeed0f88833
Links to weblogs that reference Credit Scores and Insurance Rates:
1. Posted by Julie Hill on October 2, 2007 @ 15:51 | Permalink
Congress did not seem that concerned about insurers’ use of credit information when it passed the Fair Credit Reporting Act. The Act specifically allows a credit reporting agency to provide a credit report to “a person which it has reason to believe intends to use the information in connection with the underwriting of insurance involving the consumer.” 15 U.S.C. § 1681b(a)(3)(C). States, on the other hand, have been more receptive to the idea that insurers' use of credit scores may unfairly discriminate against racial minorities. Some state legislatures already prohibit insurers from denying insurance based on poor credit scores (but they continue to allow the use of credit scores in setting rates). Hawaii has gone even further. It prohibits insurers from basing “any standard or rating plan, in whole or in part, directly or indirectly, upon a person’s … credit bureau rating.” Haw. Rev. Stat. § 431:10C-207.
2. Posted by Gordon Smith on October 2, 2007 @ 16:54 | Permalink
Thanks for the comment, Julie. This is not an area I know well, so perhaps you could help me. With regard to the states that prohibit the use of credit scores in setting insurance rates, are these states asserting that credit scores are not a good predictor or insurance risks, or is the practice of linking rates to credit scores simply viewed as unseemly?
3. Posted by Scott Moss on October 2, 2007 @ 21:07 | Permalink
Use of credit scores for screening employees is a practice that often has a disparate negative impact on African-Americans, which makes it unlawfully discriminatory under Title VII of the civil righhts act of 1964 -- unless the use of credit scores for that particular job has been validated by actual analysis, which never has been done (to my knowledge, and I've looked).
I don't know that disparate impact alone (i.e., without discriminatory intent) is enough to render it unlawful for insurance companies to rely on credit scores; the civil rights law that covers contractual relationships (as opposed to employment) requires a showing of discriminatory intent.
But I don't think there's much real dispute that credit score correlates with race even when you control for income, etc. I think that relationship is pretty well-established.
4. Posted by Julie Hill on October 3, 2007 @ 14:08 | Permalink
In response to Gordon’s question, the states that have enacted legislation restricting the use of credit scores by insurers were motivated by their views of fairness. They thought it was unfair to use credit scores because, as Scott notes, credit scores are probably correlated with race. Some states were also concerned that the use of credit scores by insurers would unfairly burden other groups with low credit scores like people without credit cards, people who recently divorced, and people who suffered catastrophic events (like medical problems or property damage). Occasionally a politician will assert that credit score is not a statistically valid underwriting variable, but most statistical analyses seem to disagree. See, e.g., Texas Department of Insurance, Use of Credit Information by Insurers in Texas: The Multivariate Analysis 5 (Jan. 31, 2005) (finding that “credit score provides insurers with additional predictive information, distinct from other rating variables, which an insurer can use to better classify and rate risks based on differences in claim experience”).
5. Posted by Gordon Smith on October 3, 2007 @ 15:07 | Permalink
Thanks, Julie. Very interesting. Based on what I have read over the past couple of days, that would have been my supposition.
6. Posted by E.B. P. on October 13, 2007 @ 8:18 | Permalink
My husband died in recent years and the next Insurance offering from our company contained a "You did not receive our best rate due to Credit score" Rate was not bad though. I asked for and received copy of credit report which contined 4 store accounts & 1 bank account open. Noted never a late payment all good reports. No real estate mos. payments or rent all that long ago paid off. A check of the credit scoring shows that I would be penalized for not having a long term monthly pay off loan etc. Was I also penalyzed because my husband died and now I am single. I have always been the book keeper and know how to pay bills. The reality makes me angry.
7. Posted by Ted Vandenberg on December 25, 2008 @ 11:18 | Permalink
My recent reading of the FTC Commissioner's statement on their auto insurance study is that Credit Scores are not a proxy for race or ethnicity. I cannot see how this translates into the assertion that insurance underwriting practices using credit scores are discriminatory against protected classes. Especially when race and ethnicity data are not used in the models. A credit score alone is not a predictor of loss costs. Multivariate models used in insurance are by definition using the combined effect of varied data, including credit scores and the underlying credit data.
| Sun | Mon | Tue | Wed | Thu | Fri | Sat |
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | |||
| 5 | 6 | 7 | 8 | 9 | 10 | 11 |
| 12 | 13 | 14 | 15 | 16 | 17 | 18 |
| 19 | 20 | 21 | 22 | 23 | 24 | 25 |
| 26 | 27 | 28 | 29 |





