Testing for moral hazard in reinsurance markets

Purpose - The purpose of this paper is to test for the existence of residual moral hazard in the three largest US reinsurance markets over the period 1995-2000 and examine the effectiveness of retention limit, experience rating and long-term contracting relationship in controlling for moral hazard. Design/methodology/approach - This paper focuses on one peculiar feature in the insurance industry, group affiliation, and tests the presence of residual moral hazard in reinsurance markets. This approach may enable moral hazard to be separated from adverse selection. Moreover, two different econometric methods are employed for the empirical tests: the non-parametric matching estimators method and the parametric fixed effects model, which may enhance the robustness of the results. Findings - The author finds that, over the period 1995-2000, residual moral hazard does not exist in the private passenger auto liability and product liability reinsurance markets, but might exist in the homeowners reinsurance market. This finding suggests that the US reinsurance markets are efficient overall and moral hazard is not a serious issue over this period of time. In addition, the author finds that retention limit is effectively used by reinsurers to mitigate the moral hazard problem, whereas experience rating and long-term contracting relationship are either not used or not effective in controlling the loss experience of reinsurance. Practical implications - US reinsurance markets are efficient overall and moral hazard is not a serious issue. Originality/value - The significance of this paper is multifaceted. First, it investigates moral hazard in reinsurance markets by examining internal and external reinsurance jointly. Second, instead of directly examining the correlation between risk and coverage, this paper tests for the presence of residual moral hazard in reinsurance markets. Moreover, the author employs two different econometric methods: the non-parametric matching estimators method and the parametric fixed effects model, which may enhance the robustness of the results. Third, the use of panel data makes it possible to explore the roles of retention limit, experience rating and long-term contracting relationship in mitigating the moral hazard problem.