Three Essays on Child Health and Economic Wellbeing in Four Developing Countries

Abstract

This dissertation contributes to literature in three areas of child and adolescent health and economic wellbeing. All three chapters use data from the Young Lives Study of International Child Poverty, a longitudinal study of two age cohorts (1-year olds and 8-year olds) that began in 2002 in Ethiopia, Andhra Pradesh state in India, Peru, and Vietnam. The first chapter evaluates children's height trajectories in rural and urban areas. I find these trajectories are largely parallel, with urban children on average significantly taller than their rural counterparts. Observable characteristics such as mother's education and epidemiological environment explain these differentials in most cases. Where unexplained variation remains, I find that rural-urban differentials in birthweight suggest that very early endowments - such as a mother's nutritional intake during pregnancy - may play an important role in explaining "residual" rural-urban height inequalities. In the second chapter, I evaluate another metric of health: self-reported health (SRH). Very little is known about this measure in young people and in low resource settings. I find that adolescent SRH and a parent's report of the child's health are both reasonably well associated with a variety of physical health indicators. This second chapter provides some encouraging evidence for the validity of SRH among young people in low resource settings. The third chapter maintains the focus on household survey questions, but shifts substantively to the measurement of economic wellbeing, comparing the two most commonly used measures of socioeconomic status in developing countries: the wealth index and consumption expenditure. Trends over time in these measures have never before been empirically investigated, and I find that the correlation of household rankings produced by the two measures is moderate but highly stable over time. Regression models treating the data as 1) multiple cross-sections, and 2) a panel (using fixed effects) produce different results. These results suggest that the more limited variability of the wealth index may reduce its predictive capability in some specifications. The findings of this dissertation are intended to inform policy and public health practice as well as survey methodology in low resource settings.