Dynamic mortality modelling: fitting a parametric model to Kenyan insurance mortality data and parameter estimation in a VAR/VECM system

Abstract

This is among pioneer studies on Modelling Kenyan Mortality using a parametric model. Kenyan insurers have not embraced the use of complex approaches to modelling mortality. They have concentrated on modelling expected returns on investments which they believe is a more significant assumption. Lack of sufficient mortality data has also hindered modelling future mortality. This paper uses four years Kenyan Insurance Mortality data (2007-2010) to assess the appropriateness of the Heligman Pollard Model in modelling Kenyan Mortality. The rates between 2007 and 2009 are used in a VAR system to obtain 2010 rates and a comparison of the HP output and the real rates done. Further, parameter estimates for 2011 are done and the general trend in the movement of parameters over the 5 years discussed. The Heligman Pollard is used in this paper because it captures the whole mortality curve well and also it has recently been used to model South Africa mortality.