Maximizing Customer Lifetime Value in a simple insurance model

Specialeforsvar ved Isabella Munk Billing

Titel: Maximizing Customer Lifetime Value in a simple insurance model

Abstract: In this thesis it is examined if it is possible to determine a premium strategy that optimizes the expected future value added to the firm by a prospective customer. A model describing the expected future value is established by linking customer-specific preferences with claims costs, other expenses, and premiums. The process is a Markov Decision Process, as the outcomes of the expected future value are partly controlled by the insurer and partly stochastic due to the preferences of the customer and the claims cost. The optimization approach used for identifying the premium strategies is dynamic programming. Two premium strategies are compared and their performance and usability in reality is evaluated. One premium strategy is identified by optimizing the weights in the traditional linear Bayes estimator and the other premium strategy is identified by optimizing without any restrictions. It is concluded that the premium strategies perform excellent, but unfortunately the presented model is unable to capture the full complexity of the customer's preferences. Hence the model should be improved before the premium strategies are implemented in reality. 

Vejleder:  Thomas Mikosch

Censor:    Søren Asmussen, Aarhus Universitet