Stochastic mortality models: an infinite-dimensional approach (Q471180)

Deterministic mortality tables constitute a rather substantial simplification of reality. In fact, demographic projections of future mortality rates involve a high level of uncertainty. The current paper investigates stochastic forward mortality models driven by a Wiener process and a compensated Poisson random measure. The authors of the paper provide a mathematically rigorous and transparent treatment of the theory of such models. They introduce a family of processes called forward mortality improvements which constitute a tool for the simple construction of stochastic forward mortality models. They prove a conditional law of large numbers which is the basis for the use of forward mortality models in practice. They describe consistency conditions for the forward mortality rates and for the forward mortality improvements. In order to obtain a unified framework, the authors transform the systems of Itō processes which describe the forward mortality rates and forward mortality improvements jointly. The obtained theoretical results are illustrated by a Lévy process driven version of the Gompertz-Makeham mortality model.