Nouvelles observations sur la longévité humaine

With the fall of mortality, the frequency distribution of ages at death has been shifting towards older ages. The ages at death around the mode have become more compressed illustrating the scenario of compression of mortality proposed by James Fries in 1980. However, this homogenization of the adults? life durations could not prevent the explosion in the numbers of oldest-old, nonagenarians and centenarians, which we have witnessed in recent decades. Some reasons for this phenomenon are investigated, and results are illustrated by historical data from Euro-pean countries (i.e. England and Wales, France, Italy, Sweden, and Switzerland), the United States of America, and Japan. The most recent results, especially for Japan, prompt us to re-examine the scenario of compression of mortality. It seems to us that the homogenization of adults? life durations, which is the driving force of the compression of mortality, may meet limits today ? beyond remaining social inequalities ? inherent in all biometric quantities. Such a scenario ? shifting the whole distribution of adults? individual life durations towards older ages without decreasing their dispersion ? would increase more and more the number oldest-old as well as the number of centenarians and persons 105 or 110 years old. Is the lengthening in healthy life duration able to follow such a rate of increase? The few studies available abroad suggest not. It thus becomes an urgent need to appreciate the evolution of the health status of centenarians whose number has multiplied by nearly three in the last 10 years in France, from 7 175 individuals on 1st January 1998 to 20 115 on 1st January 2008. The consequences of such a demographic change may be of considerable importance for financing and organizing the burden of dependency if the new nonagenarians and centenarians are frailer and have a poorer functional health status than their elders. Code JEL : J11, J14