Peer effect and dynamic ALM games among insurers

Technology advances have enhanced competition in insurance industry. This paper investigates a class of dynamic asset and liability management(ALM) games among insurers with mean-variance of relative log return performance. We obtain the unique time-consistent equilibrium ALM strategies explicitly. The results show that competition leads to the increases of insurers risky asset investment and the insurance liability. The equilibrium solution implies that social utility(relative concerns) leads to social learning(herd effect) in the complete information market. In particular, the peer relative concern can induce preference interaction. We find the Granger causal relationship between industry competition and herd behavior in the empirical data. Finally, we conduct sensitivity analysis of the competition outcome with respect to risk parameters.