A study by Consumers Union, the non-profit publisher of Consumer Reports magazine, alleges that Insurers charge good drivers living in California's predominantly African-American and Latino zip codes substantially more for automobile insurance than good drivers in predominantly white communities.
Money quote: "Did you know that millions of Californians pay higher auto insurance premiums simply because of the ZIP code they live in?"
Their methodology to test redlining is extremely weak (at best). It consisted of analyzing premiums (using data from insurer filings with the Department of Insurance) charged by State Farm, Farmers, and Allstate to a woman who uses her 1996 Acura primarily to drive to work. She drives the car 20 miles each way to work (16,000 annual miles). She purchases $100,000/$300,000 of bodily injury coverage, $50,000 of property damage coverage, $5,000 of medical payments coverage, and $30,000/$60,000 of uninsured motorist bodily injury coverage. She has a $100 deductible on comprehensive coverage and $200 deductible on collision coverage. The study used this driver profile throughout, and calculated premiums changing only her zip code. Consumers Union then cross-referenced insurers' premium data by zip code with census 2000 data by zip code.
Without really picking on the methodological inadequacies of the study, there are two simple and obvious reasons why prices may have changed based on zip code, irrespective of driving ability. First, the fictional person purchases uninsured motorist coverage (which also covers underinsured motorists), but Consumers Union does not test for whether a relatively greater number of uninsured (or underinsured) motorists reside in the predominantly African-American and Latino zip codes. Second, the fictional person also purchases comprehensive coverage, which covers loss from any of the following perils: falling objects, fire, theft, earthquake, explosion, windstorm, hail, water, flood, malicious mischief, vandalism, riot, civil commotion, contact with an animal, and glass breakage. None of these perils have anything to do with a persons driving ability and have everything to do with where the car resides.
Similarly there is the even more obvious fact that one of the factors used in determining rates is territory, or where the insured vehicle is garaged (from the IJ):
"It is a known fact that the risk of having an accident is greater in urbanized areas of the state that experience heavy traffic congestion," Sam Sorich of the Association of California Insurance Companies said. "A good driver driving rural roads, for example, is less likely to have an accident than another good driver traveling the same number of miles in a major urban area.
For an example of how research on redlining is supposed to be done see either Harrington and Neihaus (1998) or Grace and Klein (Texas Homeowners market - klein_grace_2001.pdf ). Both studies find no evidence of redlining.
Scott Harrington and Greg Niehaus conducted a thorough economic examination of redlining in the Missouri auto insurance market following similar accusations of race discrimination in the late 1980's and early 1990's(Race, Redlining, and Auto Insurance Prices, Journal of Business, 1998, pdf). Harrington and Niehaus's empirical models rely upon Gary Becker's (1993) suggestion that tests for discrimination should attempt to infer whether profits differ for products sold to minorities and nonminorities.This is the essence of Consumers Union's accusation, i.e. that profit-maximizing insurers, for some reason, are collectively willing to forgo profits in an effort to discriminate against certain segments of the population. In order for this to legitimately happen, it requires collusion on the part of all auto insurers in the state of California. Now, just for the sake of argument, lets suppose this type of collusion actually occurs. The question then is why aren't new insurance companies entering these markets and providing insurance at slightly below the collusion price, and reaping large profits?
It all just doesn't add up. Much more can be said, but that is all for now.
P.S. Commisioner John Garamendi, I will gladly help you with the consumer groups and other public leaders that are probably sending you petitions: just send me data.