Currently, the Compound Poisson model with independent and identically distributed Gamma claim sizes (CPG model) is among the most commonly used regression models in non-life insurance pricing. This model involves fitting two separate generalised linear models; namely a Poisson GLM to the claim counts and a Gamma GLM to the claim amounts. Although this procedure is now standard practice within the insurance industry, the use of Tweedie's Compound Poisson (CP) model, as an alternative way to model insurance claims, has been the focus of several papers. This thesis is based on the paper by Delong, Lindholm and Wüthrich, titled "Making Tweedie’s Compound Poisson Model More Accessible". Within the thesis the CPG model and Tweedie’s CP model are explained in detail and the relationships between them are explored. A case study is then undertaken by programming both models in R and applying each of them to an Australian automobile insurance claims dataset, in order to analyse and compare the fit of the models. This leads to a discussion on the suitability of each of the models for this particular dataset.     «

Currently, the Compound Poisson model with independent and identically distributed Gamma claim sizes (CPG model) is among the most commonly used regression models in non-life insurance pricing. This model involves fitting two separate generalised linear models; namely a Poisson GLM to the claim counts and a Gamma GLM to the claim amounts. Although this procedure is now standard practice within the insurance industry, the use of Tweedie's Compound Poisson (CP) model, as an alternative way to mod...     »