The Role of Memory in Long-Term Contracting with Moral Hazard: Empirical Evidence in Automobile Insurance

This paper tests the efficiency associated with the role of memory in long-term contracting. Bonus-malus schemes in automobile insurance are examples of contracts that use memory. During the eighties different contributors (Lambert, 1983, Rogerson, 1985, Boyer, and Dionne, 1989) showed how multi-period contracting under moral hazard improves resource allocation. In particular, it was demonstrated that multi-period contracts with memory outperform those without memory under full commitment. However, Allen (1985), Fudenberg et al. (1990), Rey and Salanié (1990) and Chiappori et al. (1994) stressed the fact that the above models did not consider the possibility of savings. Indeed it can be shown that the optimal level of action (or safety in automobile insurance) can be a function of the agent's saving activity. In the absence of full commitment, the presence of savings can eliminate all the potential gains of multi-period contracting with memory when wealth effects are significant. Consequently, it is not clear that introducing a bonus-malus scheme in automobile insurance will work efficiently to reduce moral hazard. Our empirical results show, however, that the introduction of the new bonus-malus scheme in the Quebec automobile insurance industry reduced accidents and traffic violations. This structural change was a transition from a contract regime without memory to a regime with memory and can be interpreted as a laboratory experiment to test for the efficiency of the role of memory in reducing moral hazard.