Regressions Methods: A Unified Approach to Loss Reserving Using Adaptive Filters

The presentation will concentrate on loss reserving techniques which apply the Kalman filter to regression models of the claims process. An effort is made to model each of the major aspects of the claims process: movement of case reserves, settlement of claims, and their payment. Superimposed inflation is considered. Attention is given to the merging of separate models of these aspects of the claims process to produce a single estimate of mean liability. Bootstrapping provides estimates of higher moments or of the distribution of the liability.