Three essays on the impact of political and economic shocks during childhood on health outcomes : evidence from developing countries

Abstract

The dissertation consists of three essays which attempt to capture causal relationships between shocks during childhood and before birth, and later health outcomes. Exogenous shocks such as the experiences of war and political upheaval are treated as natural experiments which minimize problems of endogeneity and selection that are present in most association studies. The first essay examines how exposure to civil war during childhood affects females’ outcomes including age at first marriage, fertility, and second generation infant mortality using the Biafra war which took place in Nigeria between years 1967 and 1970. The study uses difference-in-difference analysis to show that females that witnessed war during early adolescence got married younger than their peers not exposed to the war, and were more likely to have higher fertility and second-generation infant mortality. The second essay uses the same shock, the Biafra war, to test if males’ and females’ exposure to community-level violence results in higher risk of experiencing domestic violence in their marital relationships in the long-run. The study conducts difference-in-difference analysis on females and males separately to show that the males’ exposure to the war at ages 13 and older is the main mechanism behind females being victims of domestic violence in the long-run. The third essay examines the impact of acute prenatal stress on birth weight using the 2011 Egyptian revolution fatalities as an indicator for exposure to violence and stressful events. Results show that higher prenatal stress resulting from political conflict during the first and second trimesters of pregnancy has a significant negative impact on birth weight. This finding is robust to restricting the sample to siblings’ data and using mother fixed effects, suggesting that neither observable nor unobservable characteristics of mothers are driving the results.