Optimal, near-optimal and rule-of-thumb claiming rules for a protected bonus vehicle insurance policy

Abstract Most motorists involved in an accident will claim from their insurance company only if the cost of repair exceeds a certain amount. In this paper, a dynamic programming model is presented for determining no-claim limits for a protected bonus policy. A simple rule of thumb which works well for an unprotected bonus policy is shown to work less well for the protected bonus policy. A greatly simplified dynamic programming formulation is introduced which works almost as well as the much more complex fully optimal model.