Strategic group formation, return, and risk in the property and liability insurance industry

Empirical research indicates that subgroups, referred to as strategic groups , within an industry exhibit different performance consequences for similar strategic actions and environmental changes. A strategic group is defined as a set of firms within an industry that compete alike. This research effort applies the strategic group concept to a study of the structure of the property and liability insurance industry. The purpose of this study is to improve the understanding of competition and performance in the insurance industry from a strategic perspective. Data for fifty-seven property and liability insurance fleets were gathered from A. M. Best Company sources. The time period under study encompassed the nine years 1981 through 1989. Strategic variables define strategic positioning in terms of similarity of strategic scope and resource deployment commitments. These variables describe geographic and product diversification, financial commitments to functional areas, and scale of operations. Variance-covariance matrix equality tests evaluate strategic changes over time. Four distinct subperiods of stable strategic structures are evident during the time period studied. Within each stable strategic subperiod, cluster analysis procedures identify strategic group membership. Group membership proved to be transitory across the nine years studied. A large number of membership changes occurred and new groups emerged routinely. The number of groups in each stable strategic time period ranged from eight in the first subperiod to thirteen in the third subperiod. Following the establishment of group membership, hypotheses are tested for differences in the levels of economic impact. The performance dimensions analyzed are return on assets, risk, and risk adjusted return. Significant differences along the return dimensions are found between strategic groups in all subperiods. These findings are consistent with the theory. Risk differences, however, are noted strongly only in the first subperiod and weakly in the second subperiod. Significant differences between groups also are not indicated during any of the subperiods when returns are adjusted for varying levels of risk