Comparison of Minimum Bias and Maximum Likelihood Methods for Claim Severity

The objective of this study is to compare the methods of minimum bias and maximum likelihood by using a weighted equation on claim severity data. The advantage of using the weighted equation is that the fitting procedure provides a faster convergence compared to the classical procedure introduced by Bailey and Simon [1] and Bailey [2]. Furthermore, the fitting procedure may be extended to other models in addition to the multiplicative and additive models, as long as the function of the fitted value is written in a specified linear form. In this study, the minimum bias and maximum likelihood methods will be compared and fitted on three types of claim severity data; the Malaysian data, the U.K. data from McCullagh and Nelder [3] and the Canadian data from Bailey and Simon [1].

Keywords: Minimum bias; maximum likelihood; claim severity; multiplicative; additive