Content area
Full text
INTRODUCTION
According to one of America's leading criminologists, Wilson [2011], greater efficiency of police departments is one of the four major causes that have driven down the level of crime in the past 40 years. In the face of a growing fiscal crisis, it is more important than ever that policing be carried out as efficiently as possible. For example, Oregon's State Police had to reduce its force by 35 percent in 2003 solely due to budget cuts, which decreased officer effectiveness [DeAngelo and Hansen 2010]. 1 In New York State, Governor Cuomo proposed a US$60 million budget cut to the New York State Police [Smart, WGXC Hands-on Radio Newsroom 2011], while other upstate New York cities such as Rochester unveiled a spending plan that includes laying off 51 full-time positions in the city police department -- a story that is echoed throughout upstate New York cities [Voorhees and George 2011]. In this paper, we investigate the comparative efficiency in minimizing serious crimes of 50 local police departments in New York State. An output-oriented Data Envelopment Analysis (DEA) is used to investigate the potential for each police force to reduce crime while holding constant budgeted resources. The results are used to make recommendations about how to increase police efficiency and effectiveness.
The DEA method of mathematical programming was first introduced by Charnes et al. [1978]. Among its advantages, this method does not need a specific parametric function and allows for zero values of inputs or outputs. Its principal shortcoming is the possibility of measurement errors that can lead to outlier observations resulting in a false efficient frontier. A careful examination of the raw data is therefore necessary when DEA is utilized.
The data for our analysis is taken from the 2003 Law Enforcement Management and Administrative Statistics (LEMAS) survey [Bureau of Justice Statistics 2003], the US Census, and the Crime Index for New York State. The data is discussed in more detail below.
This paper breaks new ground in several areas. First, we believe that we present the first logically coherent model of the production of police services in contrast to the somewhat ad hoc framework found in the existing literature on measuring police efficiency. This model informs both the choice of technique (data...