UC Berkeley

This dissertation examines how the institutional organization of local governments in the United States shapes the provision of public safety. Despite internationally high levels of incarceration, violent victimization remains elevated relative to other advanced democracies and is highly concentrated across both space and race. Although there are many points of entry for studying this paradox, this project centers the role of local institutions.Each paper in the dissertation identifies distinct mechanisms through which breakdowns in local institutional arrangements may contribute to these outcomes, with an emphasis on potential avenues for reform. Specifically, I examine how electoral institutions shape the translation of citizen preferences into demand for public safety, how exposure to violence interacts with participatory inequality in whose concerns are articulated, and how institutional capacity shapes the ability of local governments to respond. Taken together, the findings suggest that the modern criminal justice crisis may partly reflect weak local institutional capacity and uneven electoral incorporation within a highly decentralized system. As a result, many high-violence communities experience both limited access to effective state protection and patterns of state intervention that are often disconnected from the problems they face most acutely.Chapter 1, Enclaves of Isolation: Violence and Participation in U.S. Cities, examines how exposure to nearby homicide affects political participation. Using fine-grained spatial data on homicides and voter turnout, I show that proximate exposure to violence depresses electoral participation, particularly in high-violence, predominantly Black neighborhoods. The findings suggest that repeated exposure to nearby violence generates fear, uncertainty, and social withdrawal that undermine the conditions necessary for democratic engagement. Rather than reflecting disengagement or apathy alone, lower participation may emerge from the psychological and social consequences of living in environments where serious violence is persistent and proximate. Violence, therefore, not only reflects failures of the state's capacity to provide safety, but also feeds back into the political system by shaping who participates and mobilizes, potentially reinforcing inequalities in whose priorities are represented in local crime policy.Chapter 2, Can Local Electoral Reforms Improve Police Responsiveness?: Evidence from California examines another mechanism underlying the public safety paradox: weak democratic accountability in local governance. Specifically, it evaluates whether strengthening links between elected officials and voters can improve public safety outcomes. Leveraging quasi-experimental variation from electoral reforms in California, I examine the effects of election timing and district-based representation on violent victimization and homicide clearance rates. I find little evidence that increasing voter turnout through election consolidation affects policing outcomes. By contrast, district-based elections modestly improve measures of public safety, particularly homicide clearance rates, suggesting that tighter geographic links between constituents and elected officials can enhance responsiveness under certain conditions. However, these gains are uneven, with improvements in clearance rates concentrated primarily among White homicide victims rather than Black victims.To probe the mechanisms underlying this disparity, I introduce a newly constructed dataset of participation in city council and public safety committee meetings across California cities. Participation in these forums is disproportionately dominated by White residents relative to their population share. At the same time, groups do not differ substantially in how often they raise public safety concerns. Instead, differences emerge in how those concerns are framed: White speakers are more likely to emphasize crime-related issues, while Black speakers focus more heavily on police conduct despite higher average exposure to violence. This pattern could suggest that the individuals most likely to participate in these forums may not necessarily reflect the perspectives of residents living in the most violence-affected communities, even within racial sub-groups. Together, the results point to the limits of electoral reforms that do not fundamentally alter underlying inequalities in political incorporation and organizational capacity. In this sense, participatory institutions may reproduce, rather than mitigate, inequalities in political voice and state responsiveness.Chapter 3, (In)Capacity, Violence, and the Local Roots of Mass Incarceration, examines long-run variation in the relationship between violence and punishment across U.S. counties. Using a multi-decade panel of county-level data on violent crime, prison admissions, court administration, and public opinion, I challenge dominant accounts that portray mass incarceration as a uniformly excessive response to crime by highlighting substantial variation in incarceration patterns across local jurisdictions operating under identical state laws. I show that prison admissions frequently failed to keep pace with serious violence, particularly in large, high-violence jurisdictions. Smaller and less institutionally strained counties were often better able to translate offenses into confinement, while large urban counties faced persistent bottlenecks in policing, prosecution, and courts that constrained this process. I argue that the prison boom reflected uneven local institutional capacity within a highly decentralized system: while downstream prison infrastructure expanded rapidly, the local institutions responsible for investigating and adjudicating violence developed unevenly across places. As a result, many high-violence communities experienced both elevated exposure to serious crime and limited state capacity to respond effectively.