|
Misrepresentation, a type of insurance fraud that occurs when insureds make an untrue statement on the rating factors, is a pervasive issue in the health insurance industry. Tobacco consumption is an important rating factor for many health insurance products. Indeed, it is one of the only few factors that are allowed to be used for determining health insurance premium since the enaction of the Patient Protection and Affordable Care Act in 2014.
In this project, we aim to develop a set of user-friendly yet theoretically rigorous statistical tools for measuring and pricing misrepresentation risk. Special attention is paid toward the self-reported tobacco status in health insurance rate-making. Our proposed models assess the misrepresentation risk based on regular rate-making data, without requiring any formal investigation/screening on the misrepresentation. We use two-stage models to capture the statistical signals of misrepresentation from the occurrence and severity of insurance claims. The use of two-stage models not only coincides with the data structure commonly observed in the health insurance industry but also guarantees the full use of available data.
|
