Date of Award

Summer 8-11-2020

Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

James Marton

Second Advisor

Tim Sass

Third Advisor

Charles Courtemance

Fourth Advisor

Tom Mroz


This dissertation evaluates the impact of various institutions and policies on health and educational outcomes.

The first chapter estimates the effect of a school stipend program for girls’ education in Bangladesh on the health of their children. For this study, I use five rounds of a repeated cross-sectional survey, which has a rich set of objectively measured child height and weight information. Using the geographic and cohort variation of the program implementation, I find that the stipend program led to a lower probability of stunting as well as higher height for age and weight for age for the children of the stipend-eligible women. The effects are larger for first births and larger still for women who were induced by the program to acquire more years of schooling.

The second chapter, co-authored with Todd Jones and Jonathan Smith, explores the phenomena of taking an Advanced Placement (AP) course but not the exam. Using data from four large school districts in the metro Atlanta area, we find that 15 percent of the AP courses do not lead to an exam. Traditionally disadvantaged populations have higher rates of missing the AP exam. Using a rich set of individual-level academic and demographic variables, we predict that a large portion of these courses would lead to a credit-granting score (3 or higher) if the exams were taken. We find evidence that students are more likely to take the AP exam when school districts offer a more generous AP exam subsidy. We find no evidence that a female AP student being paired with a female teacher leads to higher rates of exam-taking, even for courses where females are underrepresented.

The third chapter, co-authored with Charles Courtemanche, James Marton, Benjamin Ukert, Aaron Yelowitz, and Daniela Zapata, estimates the effect of the major components of ACA on various types of insurance coverage during the first two years of the Trump administration (2017 and 2018). During this time, the implementation of different components of the ACA that would impose a burden have been tabled, budgets for outreach have been cut, and a vote to repeal the ACA narrowly failed. All of these things may have affected the impact of the ACA on insurance coverage during 2017 and 2018. In non-expansion states, which should be impacted by the national components of the ACA alone, we see statistically significantly smaller coverage increases in 2017 and 2018 as compared to 2016. In expansion states, which should be impacted both by the national components of the ACA and the Medicaid expansion, we estimate similar gains in coverage in 2016 through 2018. This difference between expansion and non-expansion states is due to smaller year over year increases in coverage because of the national components of the ACA and larger year over year increases because of the Medicaid expansion.