Stochastic Modeling of Mortality Risks in Nigeria Using Lee-Carter and Renshaw-Haberman Model

Abstract

The reductions in mortality rates experienced during the last decades and the resulting increases in life expectancy show that longevity risk, arising from unexpected changes in mortality, cannot be ignored.  The study therefore explained the mortality improvements for males’ aged 40-65 using Nigeria available data using two stochastic mortality models- Lee Carter Model (M1) and Renshaw-Haberman model (M2). The fitting methodology was applied to the model using the Poisson model; the calibration was done using Life metrics R-code software.

The Lee-Carter class of models allows for greater flexibility in the age effects. On the BIC ranking criterion, the model M2 for the data dominates. However, if we take into account the robustness of the parameter estimates, then model M1 is preferred for the dataset. This model fits the dataset well, and the stability of the parameter estimates over time enables one to place some degree of trust in its projections of mortality rates. The model also shows, for the dataset, that there have been approximately linear improvements over time in mortality rates at all ages.

Keywords: Mortality, Stochastic, Modelling