Childhood Malnutrition and Mortality in a Changing Climate

Climatic variability and change disrupt environmental and socioeconomic systems, with widespread implications for human health and wellbeing. A growing demographic literature examines the relationship between climate variability and children’s health and nutritional security in low- and middle-income countries. These impacts may operate through multiple biological and socioeconomic pathways, which provide a strong conceptual basis for expecting climate effects but make the formation of directional hypotheses difficult. These issues are amplified by the potential influence of several selection biases, including from climate-induced mortality. We contribute to this literature by measuring the effects of temperature and precipitation variability on the weight and height of 0-23 month-old (n=362,352) and 24-59 month-old (n=532,497) children, respectively, across 57 developing countries. We measure overall temperature and precipitation effects across the sample and by world region, and test for differences in effects within regions by sex, birth order, mother’s education, and rural (urban) residence. We supplement our analysis of nutrition by fitting models of child mortality, which is an important outcome and a potential source of bias in our health models. We find that recent temperature anomalies are negatively associated with children’s weight-for-height (WHZ) in most regions of the world but early-childhood temperature exposures are positively associated with height-for-age (HAZ). These effects vary only modestly between and within regions, although the lack of overall climate effects on health in Latin America and the Caribbean is notable. Importantly, our models of childhood mortality show that heat exposures are positively associated with mortality risk at all ages and may be especially strong among children ages 24-59 months. These results suggest that estimated climate effects on WHZ and HAZ may be upwardly biased and imply that the apparently positive effects of temperature on HAZ may be in part an artifact of selective mortality.