Dividend maximization and stochastic investments with stochastic process arising in non-life insurance processes

Specialeforsvar: Mikkel Vaabengaard Pedersen

Titel: Dividend maximization and stochastic investments with stochastic process arising in non-life insurance processes

Abstract: In a classical paper by de Finetti (1957), an alternative approach to classical ruin theory is proposed, where dividends are maximized. This idea has gained some popularity over time, and today there is a lot of work devoted to expanding upon this idea. This thesis begins with a theoretical discussion on how to find the optimal dividend payment using ruin theory, and ideas from stochastic control theory to identify the form of the optimal strategy. We survey some extensions of these ideas, and consider the problem of ruin minimization following the paper of Gaier, Grandits, and Schachermayer (2003), and attempt to synthesize these ideas. The thesis ends with a simulation study, where different modeling of dividend strategies and investment plans can be compared. Furthermore, we use a statistical analysis to determine if we can estimate dividend payout and ruin probability, by studying an independent and identical distributed (iid) ”block” of the larger model. This estimation is based on the dividend case, which we will expand to include different investment plans.

Vejleder: Jeffrey Collamore
Censor:   Camilla Schaldemose, TopDanmark