Three Essays on the Economics of Education

Despite large measured returns to completing a high school diploma, the US has a dropout rate among the highest in the industrialized world and, while there are many studies of dropout, we still have limited knowledge of the process through which students dropout and the causal evidence of effective dropout prevention strategies is sparse. This dissertation studies the decisions students make while in high school, and how those decisions are related to dropout, academic achievement, and health. Specifically, it focuses on the relationship between high school attendance behavior, academic achievement, and dropout and the relationship between health care provision, teen pregnancy, and dropout. Chapters 1 and 2 focus on how attendance and truancy affect students' ability to graduate from high school. I use a novel dataset of daily level student attendance data for multiple cohorts of ninth graders in a Large Urban School District (LUSD) to follow individual students' attendance decisions through their high school career. Chapter 1 provides the first detailed description of the daily patterns of high school attendance. The chapter examines how measures of attendance intensity, aggregate attendance, and attendance type evolve over the course of the school year and their relationship to student achievement. I find that, even given a total number of absences, the number of truant absences is the most predictive of negative educational outcomes. Truant absences become more frequent over the course of the school year, more persistent, and students spend longer spells outside of school. These patterns are more intense for students in lower performing high schools and who have lower grades and test scores. Truancy predicts dropout and low test scores as early as the first month of high school and can serve as a valuable early warning indicator for school administrators. These patterns also provide suggestive evidence that attendance is an input into the education production function and insight into the dynamics of high school disengagement before students dropout. In Chapter 2 I model dropout as the outcome of a series of small investment decisions, specifically the decision of a student to attend school or be truant on a given day, rather than dropout as a one-time decision based on the patterns of attendance detailed in Chapter 1. Students' daily attendance decisions not only affect their ability to graduate, but also the costs they face on future days and the quality of the diploma they can earn. I find that, once students are subject to the plausibly exogenous timing of an opportunity cost increase, they are truant more frequently and the accumulation of truancies occurs at a faster rate. A student who barely misses the school cut off age is 12% more likely to be truant in 12th grade, and a student with existing attendance problems is up to 69% more likely to be truant after only 3 months of increased schooling cost. Students with higher levels of truancy each year, especially truancies late in the school year, are less likely to return to school the following fall and if they do return are more likely to transfer into a non-traditional diploma program. These results help explain the persistence of the US's high dropout rate and suggest that policies to reduce the daily attendance costs faced by students, even late in high school, may be most effective at dropout prevention. Chapter 3 turns its focus away from high school attendance to high school students' access to health care. I, along with coauthors Michael Lovenheim and Randall Reback, explore whether teenagers' access to primary health care influences their fertility and educational attainment. We study how the significant expansion of school-based health centers (SBHCs) in the United States since the early 1990's has affected teen birth and school dropout rates. Our results indicate that school-based health centers have a negative effect on teen birth rates: adding services equivalent to the average SBHC reduces the 15-18 year old birth rate by 5%. The effects are largest among younger teens and among African Americans and Hispanics. However, primary care health services do not reduce dropout rates by very much despite the sizable reductions in teen birth rates