An industry group known as the Insurance Information Institute is analyzing the role racial bias plays in calculating insurance premiums. Explicit racial bias, i.e.., setting premiums directly based on race (known as “redlining”) has been illegal since the mid 20th century. But rates continue to bet set based on criteria that indirectly reflect racial bias. One study found persistent rate increases for homeowners’ insurance in minority neighborhoods that exceeded legitimate risk differentials.
Rate criteria reflecting implicit racial bias include credit scores and occupations. The insurance industry has long defended these criteria as reliable predictors of risk. But the new working group pushes back on those assumptions:
Research shows that average credit scores for white and Asian customers are better than those for Black and Hispanic customers…Insurance credit scores reflect and perpetuate historic racism and unfairly discriminate against Black and Hispanic communities.
Other facially neutral rate setting policies can have a discriminatory impact. Motor vehicle records (e.g., traffic tickets) can reflect systemic racism on the basis that affluent white drivers are better able to afford hiring lawyers to dismiss or downgrade citations.
The industry group is also investigating whether the use of computer algorithms to analyze so-called “big data” about drivers can reflect implicit racial bias. This mirrors concerns in other fields (e.g., facial recognition software) that computer programs inadvertently perpetuate existing biases.
This new report shows the insurance industry as a whole is following up on efforts from state regulators to limit discriminatory premium rates. New York’s Department of Financial Services recently prohibited using education and occupation to price car insurance. The rule only applies in New York. Hopefully this pushback will become more widespread as other groups take note.