Abstract:
Grouping insureds in clusters such as joint-life annuities imposes statistical
dependence. In this paper, we propose the shared compound frailty approach
in collective valuation of joint-life annuity products where most applications
have been in bio-statistics. The positive stable compound process used entails the
frailty mixing distribution with the weighted exponential, generalized exponential
and weighted Weibull as the base force of mortality distributions calibrated on a
large Kenyan insurer joint-life last-survivor dataset. The findings shows that the
positive stable generalized exponential model addresses time-varying heterogeneity
effects positively and negatively associated with dependence.