Discussion Paper no. 06-05


Increasingly, bio-demographers are turning to infancy and childhood to gain a better understanding of old age mortality. However, evidence of a link between early life conditions and survival until old age is fragmentary, and the intervening mechanisms remain unclear. Drawing from data on a cohort of French-Canadian children born in the 17th and 18th centuries, we study the effects of infant exposure to infectious diseases (as revealed by the infant mortality rate in the year of birth) on later life mortality. A series of Cox proportional hazard models are used and we control for other familial and environmental conditions prevalent in childhood, as well as in adulthood. Results point to a slight, but not significant effect of a disease load in infancy for females born in years of exceptionally high infant mortality. The results are also not conclusive for males. More generally, a trend of increasing infant mortality over time correlates with general decreases in post-reproductive mortality rates, which are probably due to period improvements in later life conditions. Our study supports the view that period changes have stronger relevance than cohort effects in the study of historical variations in old age mortality.