Applying methods in credibility theory to the Maltese motor insurance context

Abstract

The following study tackles various topics in Credibility Theory, in particular the Limited Fluctuation approach and the Greatest Accuracy approach, where mainly, for the latter, the Buhlmann model and non-parametric Bayes' estimation shall be discussed. Both these approaches give the means for the insurer to develop a net premium based on the policyholders' own claim history, and hence determine whether an individual's claim experience is credible. Several examples will be demonstrated where the problem of estimating the number of claims, and the credibility premium for individual policyholders, as well as a group of policyholders sharing similar risk characteristics, is tackled, where real data as obtained from local insurance companies shall be considered.