Abstract
In my first chapter, I investigate the impact of the minimum wage on the schooling decisions of teenagers. While the possible disemployment effect of the minimum wage on teenagers has been the subject of contentious debate, comparatively little attention has been paid to the impact of the minimum wage on teen educational outcomes. This is surprising given that education is more directly linked to the later-life success of teenagers than is teen employment. In this paper, I investigate the educational effects of changes in the minimum wage, looking specifically at high school dropout decisions. I identify the effect of the minimum wage using two sources of variation (within state over time and cross-border at one point in time) and three individual-level datasets (ACS, CPS, and SIPP). I consistently find that a 10% increase in the minimum wage lowers the likelihood of dropping out for low-SES teenagers by 0.5-0.9 percentage points, roughly 4-10% of the group’s dropout rate, but has no effect on higher-SES teenagers. Additionally, I find that an increase in the minimum wage has a negative effect on hours worked that is concentrated at the upper tail of the hours distribution (not at the employment margin) for low-SES teens, but not for other young and similarly low-skilled groups. Taken together, these findings suggest that an increase in the minimum wage generates an income effect on low-SES teens, which leads them to shift their allocation of time to school-related activities and away from paid work.
In my second chapter, I investigate the long-run impacts of universal pre-kindergarten (UPK) on criminal activity. There is little evidence to date regarding the long-run effects of statewide universal preschool programs, only studies of programs targeted at more at-risk populations (e.g. Head Start and Perry Preschool) that are often more resource-intensive. I estimate the impact of Oklahoma’s universal prekindergarten program (UPK) on later criminal activity, an outcome that accounted for 40-65% of the large estimated long-run benefits of Perry Preschool. I assemble data on criminal charges in the state of Oklahoma and identify the effect of UPK availability using a regression discontinuity design that leverages the birthdate cutoff for UPK in the program’s first year of implementation. I find significant negative impacts of UPK availability on the likelihood that black children are later charged with a crime at age 18 or 19 of 7 percentage points for misdemeanors and 5 percentage points for felonies. I find no impact on the likelihood of later charges for white children. The results suggest that universal Pre-K can, like more targeted programs, have dramatic effects on later criminal outcomes, but these effects are concentrated among more at-risk populations.
In my third chapter, I investigate the impacts of performance-based compensation systems with co-authors James Wyckoff and Thomas Dee. We examine systems of differential teacher pay based on student test performance and classroom observations implemented by districts across the country as part of the Teacher Incentive Fund (TIF), a large competitive grant program operated by the U.S. Department of Education. We use student-level data from six TIF-participating districts in the state of Oregon to provide early evidence on how these high-profile reforms influenced measures of student performance. We estimate this impact of TIF by leveraging the discontinuous rule that defines a school’s TIF-eligibility in a regression-discontinuity design. Our approach provides an effective proof of concept for TIF in two ways. First, we capture the treatment contrast between implementing the full bundle of TIF-reforms and the status quo. Second, we focus on districts with high fidelity of implementation compared to many of the participating districts. We find significant effects of TIF in improving student reading achievement but not math achievement.
