Demographic effects of famines

Armed conflict and famine often have large-scale population
consequences. Yet the measurement of the mortality consequences of
such crises is challenging, given that accurate data at the population
level before, during and after such crises is often lacking and
classical demographic methods and model life tables are ill-suited for
such situations. In this paper, I explore the flexible,
two-dimensional mortality model recently proposed by Wilmoth et
al. (2012) and consider its adaptation to crisis mortality
situations. Specifically, I draw on high quality data from a diverse
array of demographic surveillance sites, population censuses,
longitudinal surveys, and series from the Human Mortality Database to explore the suitability of this mortality
modeling approach to famine and armed conflict situations. These
situations include a diverse array of armed conflict and famine
situations from both contemporary and historical populations. I
contrast this approach with existing approaches used by the United
Nations Population Division and the World Health Organization and
conclude with an assessment of this indirect mortality modeling
approach and outline future research directions to advance crisis
mortality estimation approaches.

From September 1870 to May 1871, the siege of Paris, first by the Prussian army then by 'legalist' French troops, resulted in a harsh famine: Parisians were forced to eat rats and even the elephant of the zoo. We study the impact of this shock on both child mortality and the height stature of survivors, with a specific focus on the selection effects linked to mortality as well as to fertility and migration. To this aim, we collect original data on 20 year-old military conscripts born in one of the poorest district of Paris between 1855 and 1875; conscripts born in the poorest district of Lyon, France's second largest city, are used as a comparison group. The analysis of height stature evolution across birth cohorts reveals a rather unusual pattern: a large drop in height of 2 centimeters is observed for cohorts who were 7 to 9 year-old by the time of the siege, whereas lower age children seem unaffected. Using data from the 1866, 1872 and 1876 population censuses, we examine whether selection effects can account for this specific timing, and conclude preliminarily it is not the case. We tentatively argue that early-age children were protected from the famine at the expense of their older brothers or sisters who could not catch up.

China’s great famine is known as ‘three-year natural disaster’.For the lack of continuous statistical data from 1950 to 1982 in China, traditional research methods were powerless to reconstruct the life table and mortality trends. Calculation of excess deaths during the Great Famine is complicated and confusing.Data used in this study was from China’s six censuses on age distributions, annual age-specific fertility and mortality rates which were also obtained from one-percent population samples survey initialed in 80’s.Modified Lee-Carter model is used in the study. Based on modified Lee-Carter prediction model of morality, this study did reverse prediction with reference to China’s official age-specific mortality data to re-construct and calculate single-age morality and normal deaths if there were no great famine in the last century. In the end, with regard to total deaths estimation calculated by scholars home and abroad, we did estimation research of abnormal deaths led by famine.
It is shown in our result that population dynamics, especially abnormal deaths, during great famine, mainly occurred between 1958 and 1961. If there were no famine, normal deaths would be 43.39 to 43. 85 million. With famine, abnormal deaths are between 16.24 and 23. 37million.

Investigating the effect of mortality shocks on humans is difficult in the absence of the possibility to set up laboratory experiments. However, some events in the human history serve as natural experiments.
This paper aims to analyze whether sudden changes in external conditions affect the slope of the mortality curve or shift the curve upwards proportionally at all ages.
Two cases of natural mortality experiments are presented and used for the analysis: Australian civilian prisoners during WWII in a Japanese camp and the Ukrainian Famine in 1933.
The death rates of the POWs were higher during the imprisonment but the slope of the curve appeared not to have changed compared to the normal mortality regime. During the Ukrainian Famine, instead, the curves for different years of famine converged at old ages. The results found evidence that selection could be the cause of the convergence.