Abstract:

This dissertation is an examination of the production function for infant health. In particular, I investigate the contribution of three inputs--maternal age, health insurance, and maternal education--to infant health.

The first chapter analyzes the effect of maternal age on infant health. Previous cross-sectional studies reveal that both older and younger women are higher risk of an adverse birth outcome. However, the main drawback of such estimates is the inability to control for unobserved differences across mothers. To control for this unobserved heterogeneity, I create a unique panel dataset of Texas mothers. Using these data I can examine differences in outcomes across births for a fixed mother, rather than relying on comparisons across mothers. My results suggest that older women face a higher risk of having a premature birth, having a newborn with an abnormal condition, and giving birth to an infant who dies within the first year of life while younger women have an increased likelihood of a preterm delivery.

The second chapter exploits a policy change resulting from the 1996 welfare reform to explore the relationships between Medicaid coverage, prenatal care usage, and infant health. The 1996 reform altered Medicaid eligibility requirements for non-citizens who recently immigrated at the national level but allowed states the opportunity to counteract this change using state-level funds to cover those newly ineligible. Given that some states chose this option, the reform created variation across states in eligibility for otherwise similar individuals. Using this variation across states and population groups, I find that the impact of the legislation was modest. Non-citizens affected by the policy change experienced the expected fall in Medicaid coverage but overall health insurance coverage and prenatal care utilization exhibited small changes. Meanwhile, birth outcomes were unaffected.

The third and final chapter, coauthored with Justin McCrary, closely examines the effect of maternal education on infant health using exogenous variation in maternal education induced by school-age entry laws. These laws mandate the minimum age at which a child may enter school. For example, in Texas, one of the states we study, an individual must be 5 by September 1 st . As such, among those still in school, similarly-aged children born on either side of the threshold will have different levels of education. Using this regression discontinuity design, we find that maternal education has little influence on infant health.