Ordering properties of the smallest and largest lifetimes in Gompertz-Makeham model

TL;DR

This paper studies the stochastic ordering properties of the smallest and largest lifetimes in Gompertz-Makeham models, providing conditions for comparing different groups of lifetimes in actuarial science.

Contribution

It establishes new stochastic ordering results for smallest and largest Gompertz-Makeham lifetimes, including conditions for hazard rate and ageing faster orderings.

Findings

Sufficient conditions for stochastic ordering of smallest lifetimes.

No reversed hazard rate ordering exists for largest lifetimes.

Comparison of lifetimes under random shocks using stochastic orders.

Abstract

The Gompertz-Makeham distribution, which is used commonly to represent lifetimes based on laws of mortality, is one of the most popular choices for mortality modelling in the field of actuarial science. This paper investigates ordering properties of the smallest and largest lifetimes arising from two sets of heterogeneous groups of insurees following respective Gompertz-Makeham distributions. Some sufficient conditions are provided in the sense of usual stochastic ordering to compare the smallest and largest lifetimes from two sets of dependent variables. Comparison results on the smallest lifetimes in the sense of hazard rate ordering and ageing faster ordering are established for two groups of heterogeneous independent lifetimes. Under similar set-up, no reversed hazard rate ordering is shown to exist between the largest lifetimes with the use of a counter-example. Finally, we present sufficient conditions to stochastically compare two sets of independent heterogeneous lifetimes under random shocks by means of usual stochastic ordering. Such comparisons for the smallest lifetimes are also carried out in terms of hazard rate ordering.