Two-part models for assessing misrepresentation on risk status (Q2066782)

The paper focuses on claims modeling is the actuarial framework and suggests an innovative class of two-part aggregate loss models. This class enables consideration of the semi-continuous feature of aggregate loss data, as well as of the adjustment for misrepresentation risk in insurance ratemaking; moreover, the class presented can incorporate an arbitrary number of correctly measured risk factors. After presenting the theoretical guidelines for the two-part loss models adjusting for misrepresentation, the study deepens the statistical estimation and inference for the models presented. Then, this analysis is validated by a simulation study. The theoretical models are applied on two sets of real data from a group health insurance provided by a large public institution in the US. Technical proofs and discussions are contained in Appendix.