An Exploratory Analysis of Insurer Groups

Grouping is a widespread and interesting phenomenon of the insurance industry, among both life‐health insurers and property‐liability insurers. Recognizing the potentially important implications of group membership for insurer behavior and characteristics, numerous academic researchers using insurance company data have included a dummy variable in their regression analysis to control for group membership. However, it has never been clear exactly what is being controlled for when such a variable is included. This article attempts to shed light on this question. Results indicate that group affiliated insurers tend to be larger than unaffiliated insurers, are more likely to be licensed in New York, are more likely to be stock firms than mutuals, and are likely to be less geographically concentrated.