Analysis of Possibility of Application of Higher-Order Hierarchical Credibility Estimators in Automobile Insurance

Abstract

The work concerns the applicability of credibility estimators of order higher than one, called hierarchical estimators, to posterior estimates of claim ratio in motor insurance systems based on the history of the insured. The first part presents the idea of the credibility estimation method. In the second part the linear credibility estimator is discussed. The third part introduces appropriate for risk models in the form of claims distributions, which are used to derive the corresponding hierarchical estimates of the first order. The fourth section expands the discussion of the hierarchical estimators of the order two. In the empirical fifth part previously presented solutions are used to estimate the loss ratio for the sample insurance portfolio, results are discussed and conclusions concerning the possibility of application of second order hierarchical estimators into motor insurance are formulated.