Statistical Methods for Reserving Individual Claims

Most actuarial methods for reserving in property and casualty insurance utilize aggregate data. Yet a wealth of information at the individual claim level is available at most large companies and is generally ignored when estimating reserves. In this session, statistical models for estimating the reserves for individual claims will be presented. The application of artificial intelligence and neural networking techniques will be covered. The advantages of this approach compared to using claims adjusters’ case estimates will be discussed. These procedures are utilized to improve the accuracy and adequacy of individual case reserves and do not address reserves for IBNR claims. However, the impact on IBNR reserving of shifting to a model based approach for setting individual case reserves will be discussed.