Underwriting Individual Drivers - A Sequential Approach

Adaptive decision models have been used with great success in many fields. This paper shows the value of the adaptive approach in underwriting individual automobile risks. Dropkin’s model of the accident process serves as the basis by which adaptive and non-adaptive decisions are compared. The expected value of information about past or future driving experience is explained and developed to illustrate why adaptive and non-adaptive decisions may differ. Further insight into the adaptive model and the underlying accident process is developed by evaluating the value of information from stage m and the “true” value of stage m. The paper concludes by studying the adaptive model with discounting for the probability that a policy may lapse prematurely.