Stochastic Retirement Modeling

Specialeforsvar: Marc Danakil Sørensen

Titel: Stochastic Retirement Modeling - in Life Insurance

Abstract: We explore a way to extend the usual Markov Chains in Life Insurance Mathematics, to include discontinuous transition distributions. This is done to achieve a flexible framework, suited for modeling policyholder options, particularly retirement, with potentially complex exercise patterns. Through careful construction of these Markov Chains, we are able to derive extended versions of important classical, actuarial results such as Kolmogorov and Thiele type equations. We then consider simple life insurance contracts, in order to construct benefit sizes that depend on the uncertain time of retirement, and
examine the affects of including stochastic retirement in insurance contracts. Finally, a numerical study highlights some of the computational differences, particularly regarding reserve calculations, between the new model and more traditional contracts.

Vejleder: Jesper Lund Pedersen
Censor:    Jeppe Woetmann Nielsen, AkademikerPension