Effect of exposure to conflicts and childhood mortality: Analyses of pooled cross-sectional data from 105 surveys from 52 countries

This study analyses the adverse impact of different types of conflicts on child mortality and its mechanisms across a large group of countries. Our analysis pools data from multiple countries and time-points, to provide robust evidence on the relationship between conflict and child health. Geo-referenced data on different types of conflict are linked with the unit level information from Demographic Health Survey datasets, to construct a unique database of 943139 pre-school age children across 52 developing countries over the period 1993-2018. Our analysis exploits the within-country differences in children’s exposure to conflict from in utero to age 5, to estimate its association with mortality outcomes. Our baseline estimate shows the association between conflict exposure and childhood mortality where we found excess mortality in children exposed to conflict. The Cox proportional hazard regression model estimates show that even after controlling for an extensive array of socio-economic and demographic characteristics time and entity-fixed effects, conflict exposure is associated with excess child mortality, across all our measures of conflict. Additional analyses show that probability of dying in the childhood increases with increasing the intensity of exposure to conflict events. Proximate mechanisms explaining excess mortality in children exposed to conflict include poor maternal health outcomes (ANC and Institutional delivery care), child health outcomes (height-for-age and weight-for-age z-scores), and immunization status. These findings are robust across alternative measures of conflict, and sub-samples. Our main findings sustain even after controlling for unobserved heterogeneity and migration history.