Bureau for Research and Economic Analysis of Development, Working Paper No. 219 (April) Meng, Xin and Nancy Qian (2009)

An empirically strong study of China’s 1959-1961 Great Famine finds that children who survived the disaster suffered in weight, height, educational attainment, and labor supply thirty years later.

Famines were responsible for more twentieth-century deaths than both World Wars.  And as the 2007-2008 food price crisis illustrated, food security is far from guaranteed in the opening decades of the twenty-first.  The World Food Programme calculated in 2002 that 38 million Africans live in danger of famine. 

Despite the continued grave importance of food security issues and the now substantial policy attention devoted to this topic, little consideration has been devoted to the effects of famine on survivors.  Malnutrition has proven health impacts, but malnourished children have also been shown to regain health quickly if care and nutrition is available.  It has been theorized that famines could have long run impacts on labor supply and economic growth if they substantially damage the health of the population, but whether this long run effect actually materializes remains unclear.  Part of this murkiness is due to the methodological difficulty of studies of the impact of famine on long run health and labor supply outcomes. 

Quantifying famine exposure is difficult, famine is usually accompanied by conflict or other disasters, and selection for survival distorts studies of the impact of famines on health. 

This study avoids these three problems of measurement error, endogeneity, and selection with a novel methodological approach.  Xin and Qian examines the long run impact of famines by conducting an analytically rigorous evaluation of the effects of China’s 1959-1961 Great Famine on child survivors thirty years later. 

Exposure is measured for the average county level cohort of survivors born in the famine.  Exposure is estimated only for the top 90th percentile of survivors, therefore eliminating the concern about selection bias.  Measurement error and omitted variable bias are accounted for with instrumental variables regression. 

Because the famine was largely caused by policy choices—specifically, grain over procurement in high grain production areas—the authors are able to “use this institution driven variation in cross-sectional famine intensity (measured as non-famine grain production or as geographic and climactic suitability for rice or wheat production) in combination with birth cohort variation to instrument for exposure.” 

The interaction between grain production in non-famine years and birth year is used to instrument for famine intensity.  This China-specific context allows for the determination of causality, which has been difficult to estimate in other famines. 

Data are drawn from China’s 1990 Population Census, the 1989 China Health and Nutritional Survey, the 1997 China Agricultural Census, and the FAO Global Agro-Ecological Zones database.

Results indicate a striking impact on adult health outcomes even among the most robust famine survivors.  The authors find that “in-utero exposure to famine on average caused a 1.7% (2.8 cm) reduction in height, a 2.3% (1.4 kg) reduction in weight and a 8.6% (6.8 years) reduction in educational attainment.  For those exposed during early childhood, the famine on average reduced height by 1.6% (2.7 cm), weight by 5% (3 kg), WFH [weight for height] by 1.2% (0.004 kg/cm) and labor supply by 13.9% (12.6 hours per week).”

These effects are larger than those estimated by earlier studies, principally because this study corrects for the bias caused by selection and measurement error.  The authors suggest that conservatively, the estimates from using the 90th percentile of survivors are applicable to those above the 30th percentile.  Regarding future research, the health impact of famine may be non-linear in age, so the authors plan to monitor the health of the survivor cohort in their 50s and 60s with future China Health and Nutritional Surveys.  They comment that applying their novel technique of analyzing the health outcomes of only the strongest survivors to other famines could very well increase the estimates of the health consequences of those famines, because China’s 1959-1961 famine was less severe than many of its historical counterparts.

Severe child malnourishment is prevalent today, affecting 30% of children under five according to the 2004 World Development Indicators.  This paper is significant because it shows that, as the authors conclude, “these deprivations can have large adverse long run effects even amongst a sample of ‘strong’ individuals selected to survive severe famine.  And this could in turn affect long run economic development.”