Credibility theory and the Kalman filter (Q2266333)

Three classical credibility models [namely, \textit{H. Bühlmann} and \textit{E. Straub's} model, Mitt. Verein. Schweiz. Versicherungsmath. 70, 111-133 (1970; Zbl 0197.465), \textit{C. A. Hachemeister's} regression model, Credibility, Theory Appl., Proc. Actuarial Res. Conf., Berkeley 1974, 129-163 (1975; Zbl 0354.62057) and \textit{W. S. Jewell's} hierarchical model, Giorn. Ist. Ital. Attuari 38, 1-16 (1975; Zbl 0392.62086)] are reformulated in terms of Kalman filtering problems, using well-known connections between Bayesian models and the Kalman filter. This formulation provides a new derivation of earlier results on the recursive forecasting of premiums and the corresponding forecasting error, as well as a unified and convenient approach to their computation in practical situations.