Stochastic Mortality Models. Application to CR Mortality Data

Abstract

The ageing process is a great challenge for many European countries, not excluding Czech Republic (CR) and it brings financial risk in areas such as social policy, pensions and health care. The motivation for this paper is to compare various mortality models. We have attempted to explain mortality improvements for males aged 62–90 in CR using a several stochastic mortality models. We compare quantitatively number of stochastic models explaining improvements in mortality rates in CR. It is clear that mortality improvements are driven by an underlying process that is stochastic. Numbers of stochastic models have been developed to analyse these mortality improvements. We will deal in models such as Lee-Carter model, Renshaw and Haberman model, Aged-Periodic-Cohort model (APC), Cairns-Blake-Dowd model (CBD) and their extensions. Each model is fitted to the male data between 1968 and 2011. Our analysis focuses on mortality at higher ages (62–90), given our interest in pension-related applications. By the Bayes Information Criterion (BIC) we find that an extension of the Cairns-Blake-Dowd (CBD) model fits the Czech Republic male’s data best.