Essays in health economics

This dissertation studies topics of health economics, mainly focusing on mental health outcomes, such as the effects of educational transitions, school start time delay, and the surprising news of an electoral outcome. Each chapter of the dissertation examines one of these topics.In the first chapter, I answer the question: Does the transition from middle school to high school change students’ behavior and mental health outcomes? South Korean students transitioning from middle school to high school face longer school hours, increased academic pressure, and an older peer group. All of these have the potential to cause mental health problems. I estimate the effects of transitioning to high school from middle school on sleep patterns, risky behaviors, and mental health using a regression discontinuity design on date of birth. I find that students lose 40 minutes of sleep as they advance from middle school to high school, and their probability of trying alcohol or cigarettes goes up by 3.7 and 2.2 percentage points, respectively. Despite the reduction in sleep and the increase in risky behaviors, I find little evidence of statistically significant effects on various mental health measures, including stress, depression, and suicidal ideations or attempts. In fact, some measures of mental health improve with the transition to high school, which is surprising as there is substantial literature linking reduced sleep and substance abuse to worse mental health outcomes.In the second chapter, I evaluate the effects of a 2014 school policy in South Korea that delayed start times to 9:00 a.m. for all public middle and high schools in one of the country’s largest regions. Using a difference-in-differences design and nationally representative data from the Korea Youth Risk Behavior Survey (KYRBS), I estimate the impact of the policy on students’ sleep duration and mental health outcomes. The reform led to an average increase of approximately 16 minutes in weekday sleep, primarily through later wake-up times. Despite this measurable gain, I find no significant changes in mental health indicators, including stress, depression, suicide ideation, and self-reported happiness. Two-stage least squares estimates, using the policy as an instrument for sleep, reinforce these null results, suggesting that modest increases in sleep may not be sufficient to yield meaningful psychological improvements. These findings underscore the limits of light-touch sleep interventions and emphasize the role of broader social and institutional contexts in shaping adolescent well-being.The third chapter, written jointly with Albert Young-Il Kim and Dongyoung Kim, investigates the impact of the 2016 U.S. presidential election on infant health, leveraging the unexpected outcome of the election as a quasi-natural experiment. Using restricted-use individual-level natality data from the CDC, we implement a difference-in-differences design that compares birth outcomes for infants exposed in utero during the election result period across counties with varying political orientations. We find that birth weights declined modestly—but statistically significantly—in more Democratic-leaning counties following the election, with estimates robust to alternative definitions of partisanship. However, similar patterns emerge in a placebo analysis using the 2012 election, and we find no differential effects among Hispanic mothers, a group thought to be more directly affected by the political climate. These results suggest that while political shocks may coincide with changes in fetal development, the evidence does not support a uniquely causal interpretation of the 2016 election. Our findings underscore the complexity of isolating sociopolitical stress effects in population health using observational data.