A report saying Missouri drivers face “a subtler form of redlining” in buying basic auto insurance was itself “fundamentally flawed,” an insurance industry trade group said Wednesday.
The report from Consumer Reports and ProPublica had looked at data in Missouri, Texas, Illinois and California. It concluded that drivers in ZIP codes with largely minority populations pay higher premiums for basic liability insurance than could be accounted for by risks insurance companies face in those areas.
“It is fundamentally flawed,” James Lynch, chief actuary for the Insurance Information Institute, said of the Consumer Reports/ProPublica report. “You cannot draw the conclusions that they draw from the data and analysis they have performed.”
Lynch said the insurance group had worked with ProPublica extensively to explain its objections. It also obtained the data and analysis ProPublica relied on and provided it to an independent actuarial firm. Lynch said a draft of that firm’s report cited a fundamental flaw.
Lynch raised the complaint in a conference call with reporters from selected media outlets, including The Kansas City Star. ProPublica was not invited to the call.
Julia Angwin, a ProPublica senior reporter and co-author of the report, said the publication had been aware of the issues Lynch raised and stood by its report and its methods.
ProPublica’s analysis had requested premium quotes for a fictitious 30-year-old female driver in each ZIP code and provided the same information each time about miles driven, occupation and other factors that can impact insurance premiums. Still, they received different prices for basic liability coverage in different ZIP codes, typically higher ones in ZIP codes with predominately minority populations.
ProPublica then compared the differences in premiums between minority and non-minority ZIP codes to insurance claims losses in those ZIP codes. The comparison measured the risk insurers were taking in each location.
Often, premiums remained at least 10 percent higher in ZIP codes with largely minority populations even after accounting for differences in the risks, the Consumer Reports/ProPublica report said.
Lynch said the report’s “basic error” came from comparing premiums for the fictional female driver with aggregate data on claims losses for all drivers in the ZIP code. He made the same charge, as well as others, in an article published by the Insurance Journal.
“They have information upon the wide variety of risks, but they only compared it to a single driver. This is the problem,” Lynch said in the conference call.
Angwin responded to questions about Lynch’s comments by pointing out that ProPublica ran its analysis using 44 different profiles that varied by age, gender, marital status, occupation and other factors.
“We did our analysis on this one safe driver in order to control for all factors other than geography,” Angwin said. “We repeated this analysis across all 44 profiles, and the results were similar.”
She said the analysis did not prove insurance companies had intentionally discriminated against minority drivers, but that it puts a burden on insurers to explain why there were differences in premiums independent of risks.
Consumer Reports and ProPublica had included comments from Lynch in their report. These included his statement that insurance companies do not collect information about the race or ethnicity of consumers and do not discriminate on the basis of race.