A user-friendly approach to stochastic mortality modelling

This paper proposes a general approach to stochastic mortality modelling. The logit transforms of annual survival probabilities are modelled by a linear combination of user-specified basis function of age. The model is easy to calibrate using the maximum likelihood method. The flexible construction and tangible interpretation of the underlying risk factors allows for an easy incorporation of population-specific characteristics and user views into the model. We fit two versions of the model into Finnish adult (18-100 years) population and mortality data, and present simulations for the future development of life spans.