A study of centenarians in Sardinia detected the existence of an area where the number of centenarians is higher than elsewhere, in particularly for men. There is a widespread opinion that the existence of a greater or a lesser number of centenarians largely depends on mortality features between 80 and 100 years. This study aims to cast light on our knowledge of elderly mortality differentials, total and by cause of death, in Sardinia, and attempts to verify this hypothesis. To do so, an analysis is conducted of age and sex mortality trends over time at province and municipality level. Results fully confirm the underlying hypothesis.

No abstract

This paper considers the issue of actuarial fairness of the new Italian public pension system in view of the recent trends in old-age mortality and the survival differences by gender, birth cohort and region of residence. After reviewing the secular trends in elderly mortality in Italy, and the evolution of regional differences in survival over the last three decades, we evaluate the impact, on the conversion factors introduced by the Dini reform, of a further decline in elderly mortality over the next few decades. We compute the conversion factors using a close approximation to the unknown formula employed in the Dini reform but allowing for gender-and region-specific survival probabilities. Our results leave no doubt about the importance of frequently updating the conversion factors in the light of the rapid increase in elderly survival. The paper also quantifies to what extent gender-and region-specific conversion factors may differ from their currently legislated values, that only vary by age. Finally, we recognize that the actuarial fairness of the system introduced by the recent reform can only be guaranteed on average and that, in the presence of a heterogeneous population of individuals that differ considerably in their mortality prospects, the current system implies a substantial degree of redistribution from high-mortality groups (typically characterized by low income and low wealth) to low-mortality groups (typically characterized by high income and high wealth).LABOUR 17 (Special Issue) 45-78 (2003) JEL J11, J14, N33, N34Note: Some differences in life expectancy are generated by inevitable approximations in applying Pollard's formula (1982) adapted to estimate the contributions made. Figure 2.Ratio of male to female mortality probabilities by age and calendar year from 1887-88 to 1994-95

No abstract

No abstract