Rating Factors Identification using Claim Frequency Approach: The Malaysian Experience

This paper identifies the rating factors of the Malaysian motor insurance experience using the claim Frequency approach. In the recent years, the Poisson regression has been widely used by the insurance practitioners for modeling claim frequency. However, the Poisson regression assumes that the mean and variance of the dependent variable is equal. In insurance practice, claim count or frequency data often display over-dispersion or extra-Poisson variation a situation where the variance exceeds the mean. Inappropriate imposition of the Poisson may underestimate the standard errors and overstate the significance of the regression parameters, consequently misleading the inference for the rating factors. Therefore, the Negative Binomial and Generalized Poisson regressions are suggested for handling over-dispersion in the claim Frequency model. In this study, the Poisson, Negative Binomial and Generalized Poisson regressions are considered for the claim frequency model to identify the rating factors in two types of Malaysian motor insurance data third party property damage and own damage.