par
Weihong NI(Department of Statistics and Actuarial Science, The University of Hong Kong)
→
Europe/Paris
Séminaire 2 (Bâtiment Braconnier)
Séminaire 2
Bâtiment Braconnier
Description
Bonus-Malus (BM) systems are common merit rating schemes serving auto-mobile insurance companies. Basically, such systems impose certain rules to set premium levels based on one's historical claims with the purpose of providing a fair share of risks. Incorporating the features of a BM system, we have attempted to model pricing strategies as well as the risk surplus process for an insurer. Bayesian adjustment acts as the main tool to characterise a BM system so that premiums are adjusted posteriorly. Other than that, bearing the no-claim-discount (NCD) property into a risk surplus process, we have built modied risk models and found the associated ruin probabilities. Several premium-adjusted models have been proposed. One of them directly applies Bayesian adjustment on premium rates. We also transfer the change of premium rates to alterations in inter-arrival times so that a Markov Additive process could be identified. In addition, letting premium rates depend on the increments over the previous n random observed periods, we are able to obtain ruin related quantities by Gerber-Shiu functions. Analytical results of these studies will be demonstrated in this talk together with a few numerical examples.
