Simulations of a New Model for Stochastic Transition Rates in Life Insurance Based on Generalized Cox Processes

In this thesis we simulate a new model, formulated as a nonlinear filtering problem with a generalized Cox process, for the modeling of transition rates in life insurance. This allows for a time-continuous model which may capture both Gaussian and non-Gaussian noise effects. In addition, for the non-Gaussian noise we introduce a jump component with intensity which is subject to random effects. These elements make for a very flexible model with great opportunities for capturing many different effects such as e.g. regime switching effects or mean reversion. Since we are able to incorporate such elements into our model, we can capture effects from regulatory changes in the insurance market, changes in government activity, and different kinds of "shocks" which affect transition rates, such as natural disasters. In addition, a completely specified model with an in depth discussion on simulations will be given. With our new model we simulate future mortality rates. Where in the simulation stage, the focus has been laid on capturing mean reversion through the intensity of the jump component in the generalized Cox process.