An insurance company was sued by a fair housing group over the underwriting criteria it had used in the past to determine which applicants would get access to homeowners policies with replacement cost. These policies would insure the homeowner against losses up to the replacement cost of the home, as opposed to market value policies, which covered losses only up to the market value of the home. Plaintiff alleged that the restrictions violated the Fair Housing Act in that they had a disparate impact on minority homeowners in five metropolitan areas.

In the past, the defendant had a number of restrictions on access to replacement cost policies. For example, it had required that homes be insured for at least $60,000 and be under 30 years of age. The defendant removed these restrictions in stages, and the last were removed three years before the filing of the suit. Nevertheless, plaintiff argued that the restrictions had a disparate impact on minority homeowners. Plaintiff supported this claim with statistical evidence indicating that the defendant insured a lower percentage of homes in neighborhoods with high minority representation than in neighborhoods that were predominately white.

One of the counterarguments to this claim was that economic factors were primarily responsible for the disparity in policies sold in high-minority versus primarily white neighborhoods. Replacement cost policies cost more than market value policies. On average, minority homeowners had less income and wealth than white homeowners and would therefore prefer to purchase lower-cost insurance.  

To test this proposition, we compared the percentage of policies in minority neighborhoods before and after the removal of the underwriting restrictions at issue. The change in each of the five metro areas is shown in the chart below. The results show no consistent pattern of increases in business in minority neighborhoods after the removal of the challenged guidelines. Only in MSA 1 did the percentage of policies in minority neighborhoods increase to a statistically significant degree. In MSA 4 the percentage decreased significantly.  In the other MSAs the changes were not statistically significant.

We did similar comparisons for 19 other MSAs that were comparable to the five identified by the plaintiff.  In only two of those did we find a statistically significant increase in the percentage of business in minority neighborhoods. Thus, we concluded that the criteria challenged by the plaintiff did not have a systematic effect on minority purchases of homeowners insurance from the defendant.