The Force of Mortality at Ages 80 to 120

The purpose of this monograph is to see whether new data make it possible to re-assess the relative merits of various contending models for the way in which the probability of dying changes with age, at least in the range of ages from 80 to 120. Thirteen countries have a sufficiently long run of reliable data to be useful for the specialised purposes of the present analysis : Austria, Denmark, England and Wales, Finland, France, Germany (West), Iceland, Italy, Japan, the Netherlands, Norway, Sweden and Switzerland. These countries include almost 40 million persons who reached age 80, and over 120'000 who reached age 100, during the period 1960-1990. The analysis covers over 32 million deaths at age 80and over this period. The results show that the Gompertz, Weibull, and Heligman & Pollard models give estimates of mortality which are far too high above age 100. The other three models studied - the logistic, Kannisto and quadriatic models - were all far closer to the observed values, and it is not easy to choose among them. Based on various theoretical and pragmatic considerations we conclude that the logistic model and its Kannisto approximation are the best of the six models. The analysis indicates that the probability of dying reaches a value between about 0.5 and 0.65, for both males and females, at age 120.