Using a Bayesian hierarchical model for fitting automobile claim frequency data

Abstract

In this article, we consider Bayesian statistical models in which prior distribution of the risk parameter is to be specified in a hierarchical fashion. The model obtained is shown to be over-dispersed and competitive with other models in the literature for fitting automobile claim frequency data. We obtain analytical forms of the distribution, which presents excellent properties. The methodology is illustrated by practical examples in setting automobile third party liability insurance, with which its sensitivity to the structure function is tested.