Three Essays on the Provision of Local Public Goods

Abstract

This dissertation consists of three papers on the provision of local public goods. Chapters 1 and 2 consider the criminogenic and social consequences of New York City's ``Stop and Frisk'' program, respectively. In these papers, co-author Andrew Bacher-Hicks and I leverage the quasi-random movement of NYPD police commanders across precincts to identify the causal impact of this policing strategy on crime and long-run educational attainment. We first find that a commander’s predicted effect on stops—--based on data from previous precincts—--is highly predictive of changes in observed stops after that commander enters a new precinct. In Chapter 1, we find that high propensity-to-stop commanders decrease low-level offenses within their precinct. However, we find no corresponding decrease in more serious felony offenses, and find suggestive evidence that within-precinct misdemeanor crime reductions are offset by crime displacement to adjacent neighborhoods. Contrary to the broken windows theory of policing, our findings suggest that stop-and-frisk tactics do not deter more serious criminal behavior. In Chapter 2, we estimate the social impact of stop-and-frisk policing. We find that exposure to high propensity-to-stop commanders in middle school has negative effects on students' high school graduation, college enrollment, and college persistence rates. These effects are concentrated among black students, who belong to the racial group overwhelmingly targeted by police stops. We find evidence of improvements to school safety and positive spillovers for white and Asian students, who are less likely to directly interact with police. Finally, Chapter 3 estimates the impact of top-down local public goods provision on citizen engagement. I leverage the fact that routine street maintenance in Boston is determined according to a well-defined set of criteria---underlying street quality and city geography---to estimate the causal impact of traditional, bureaucrat-driven local public goods provision on citizen requests. I find that exogenous shocks to seasonal street paving decrease requests for future street repairs, while at the same time increasing future requests for non-street public goods. Reductions in requests for future street repairs are primarily driven by the mechanical improvement paving brings to street quality. On the other hand, results for non-street public goods are consistent with the notion that top-down local public goods provision sends a positive signal to citizens about both government effectiveness and responsiveness.