Mortality risk via affine stochastic intensities: calibration and empirical relevance

In this paper, we address the mortality risk of individuals and adopt parsimonious time-homogeneous affine processes for their mortality intensities. We calibrate the models to different generations in the UK population and investigate their empirical appropriateness. We find that, in spite of their simplicity, non mean reverting processes with deterministic part that increases exponentially - which generalize the Gompertz law - seem to be appropriate descriptors of human mortality. The proposed models prove to fulfill most of the properties that a good model for stochastic mortality should have. Empirical results show that the generalization is worth exploring. Indeed, the variability of number of deaths may increase considerably due to the randomness of the mortality intensity. We show that the models are suitable for mortality forecasting and mortality trend assessment.