Individualized Offers and the Measurement of Consumer Inertia: Evidence from Auto Insurance

Abstract Demand estimates can overstate consumer inertia when researchers do not observe the alternatives and prices available to each consumer. We study this issue using auto insurance renewal data that record both the chosen contract and all insurer quotes returned for the policyholder, including the premiums on alternative offers. We exploit these data by estimating the same demand model as we sequentially remove information on alternative prices and consumer-specific availability. With the observed offers, the premium-equivalent incumbency effect is about 106 euros or 19% of the mean annual premium. Replacing alternative premiums with predicted prices while holding each consumer's observed offer set fixed raises this measure to about 183 euros. Replacing the consumer-specific offer set with a common market-level set raises it further to about 232 euros. The first step mainly reduces estimated price sensitivity, while the second raises the estimated incumbency effect. Missing offer-level information can therefore substantially affect measured price responsiveness and consumer inertia. JEL Codes: D12, D83, G22, L11, M31