Essays on the Economic Determinants of Health

This dissertation consists of three chapters on the economic determinants of health, specifically those related to the pathways of nutrition and government nutrition assistance policy. I evaluate the Community Eligibility Provision’s (CEP’s) effects on child weight outcomes in my first two chapters, a program that allows certain schools to offer universally free school meals to all students. My first chapter uses child-level data from a nationally representative survey which follows a single sample of children from Kindergarten to fifth grade. I use these data to identify the effect of attending a CEP school on outcomes of child weight. I find that CEP school attendance increases a child’s Body Mass Index (BMI) percentile score, decreases their likelihood of falling within the healthy weight range, and increases their probabilities of being overweight and obese. In my second chapter, I utilize school-level data for the universe of K-12 schools in the state of Georgia. My data set includes aggregate measures of child weight including average child BMI and the percentage of students attending a school who fall within the healthy weight range. I find that adopting the CEP decreases average child BMI and increases the percentage of healthy weight students. Differences in the results of Chapters 1 and 2 highlight the likelihood that the CEP’s effects on child weight may vary by location and student characteristics. Finally, my third chapter proposes a new model for the measurement of food security. Specifically, I construct a Bayesian Graded Response Model (BGRM) which can be used to measure food security with responses to the United States Department of Agriculture’s core Food Security Module (FSM). I use a simulated data exercise to evaluate the performance of my model in a controlled environment. I find that my model properly retrieves the set of data generating parameters. Comparing the performance of my model to the most commonly used measure of food security, the FSM scale, I find that my model more accurately assigns the food security status of households in all cases