Author: Daniel Snedaker Bennett

Publisher:

Published: 2019

Total Pages:

ISBN-13:

The research presented in this dissertation answers questions about some of the most important and popular public services provided by the federal government and local governments in the United States. At the local level, I examine issues that determine how efficiently police services are provided, and at the federal level, I provide evidence to help explain why Social Security has been so hard to reform. In Chapter 2, I measure how the degree to which police are fragmented impacts police efficiency, and how the characteristics of the communities that police serve impact the efficiency of service that those communities receive. I assembled a unique dataset of every event recorded by the Computer Aided Dispatch systems of 40 different police agencies from December of 2015 to January of 2016 to measure police efficiency by their response times to public calls for service. Using the detailed geographic information provided by these systems, I geocode the calls, match them to the census block group from which they originate, and calculate a predicted response time based on the optimal placement of police response units using a Maximum Covering Model with capacity constraints. I find that minority communities can expect slower response times on average for lower priority calls, but there is considerable heterogeneity across jurisdictions. In areas with less fragmented law enforcement, response times are significantly faster for calls requiring an immediate response. Chapter 3 investigates how public scrutiny following some recent high profile lethal encounters with police impacts police behavior. Since 2014, there has been wide spread speculation that negative media coverage of instances in which police use force has a ``Ferguson Effect, " or a de-policing effect in which officers exert less effort causing an increase in crime rates. I the use unique data described in Chapter \ref{chapter:FergussonEffect} to directly measure how officer initiated events change in response to high profile events. I also use the most recent data from the FBI's Uniform Crime Reports to compare estimates from studies examining shootings in 2014 with other high profile events in the following years. The balance of the evidence suggests that there is a modest de-policing effect with minimal to no impact on the aggregate crime rates. Chapter 4 was co-authored with Michael Boskin and Diego Perez, and examines the political economy of Social Security Reform. It has been known for decades that Social Security is facing a financial crisis and soon will only be able to pay 71\% of benefits. We calculated the expected net present value of Social Security taxes, benefits and transfers for 112 worker-beneficiary groups under several proposed reforms covering the unfunded liabilities. We identified which reforms receive support when people vote in their financial self-interest under alternative economic and demographic projections and voting proclivity assumptions. While 40\% of voters have a negative lifetime net transfer, less than 10\% have a negative future transfer under the unsustainable status quo. So framing the problem as a choice between reforms is necessary for any reform to receive support. Delayed reforms are often preferred, but immediate reform is possible, whether tax hikes or slowing benefit growth depends upon whether SSA or CBO assumptions are used. Alternative representations of the interests of younger voters strongly affect preferred reforms. Intergenerational AND intragenerational heterogeneity of economic interests combine to enable coalitions employing side payments between voters to affect which reforms are blocked and which are feasible.