An adaptive premium policy with a Bayesian motivation in the classical risk model

In this paper, we consider an extension of the classical risk model in which the premium rate policy is adaptive to claim experience. We assume that the premium rate is reviewed each time the surplus reaches a new descending ladder height. A choice between a finite number m of rates is then made depending on the time elapsed between successive ladder heights. We derive explicit expressions for the probability of ruin in this model, assuming claim sizes have a mixed Erlang distribution. We then motivate further the idea behind this adaptive premium rate policy by using a mixed Poisson process for the claim arrival, and propose a method to fix the parameters of the policy in this setting. Finally, we discuss other applications of this method.