Public Sector Performance: Move or Monitor?

• Author: Hayes, Kathy

Kathy Hayes [*]

1. Introduction

   Governmental performance is a concern of citizens and politicians alike. A specific interest is public sector waste, which is thought to be pervasive, and efforts to improve public sector performance are believed to be minimal. Economists have studied the topic extensively. This paper discusses options on measuring public sector performance in urban areas, and it also suggests certain public policy options that I believe deserve additional attention.

   From a citizen's point of view, when public service performance declines, one can leave the jurisdiction. This exit option has a substantial historical precedent in the United States, going back to at least the landing of the Mayflower. [1] In the urban public finance literature, this is a consequence of the Tiebout (1956) model, in which there are a sufficient number of jurisdictions among which individuals may choose to live and people vote with their feet. The model suggests that exit or the threat of exit leads to competitive forces that improve efficiency in the public sector.

   An alternative to exit comes about through monitoring and voting. Monitoring and voting are attractive ways to effect change for individuals who do not want to exit. As suggested by Davis and Hayes (1993), a fairly simple model can demonstrate the demand for good government. There is a literature that examines the motives for exit (Dowding and John 1996), which points out and concludes that people may not move in response to the fiscal decisions of local governments. Perhaps voting and monitoring are more attractive options, at least to a proportion of the population. [2]

   This paper explores the move-monitor trade-off in influencing urban public sector performance. Although heuristic arguments can be made concerning the relative importance of moving or monitoring, few researchers have investigated when, if ever, move and/or monitor are effective means of improving public sector efficiency. In this paper. I would like to suggest that there are gaps in the literature proposing topics or future research.

   An understanding of the current literature is necessary before one can consider future research possibilities. Next, I will discuss measurement of public sector output in the local government public finance setting. To determine public sector performance, one must put it in the context of the appropriate objective functions for the provision of governmental services. A short examination of alternative empirical approaches to measurement of performance or benchmarking is provided. Finally, this paper explores the relationship of Tiebout sorting and monitoring to public sector efficiency.

2. Output Measures

   For urban areas, the public sector typically is expected to protect from crime, educate the young, protect from fires, maintain streets, collect trash, and so forth. These are all services that may be provided with various degrees of quality. However, precise measurement of public sector output is elusive for at least two reasons. First, as a service, the production outcome is difficult to measure. Data availability on public sector output is often dismal. Second, we know little about citizens' evaluation of the various attributes of the service produced. Public sector demand studies have tried to determine the preferences of the citizenry using various proxies for output. But little if any of this literature has been acknowledged in explorations of public sector production.

   Historically, economists have relied on expenditures or public sector employment to measure output. [3] As the literature has become more sophisticated, so have the techniques of measuring output. Bradford, Malt, and Oates (1969) argue that what is produced might be different than what is consumed based on community characteristics. Output should be adjusted for characteristics of the jurisdiction. For example, a unit of police protection in the south side of Chicago might be different than one in rural Illinois. Many studies have examined congestion in the provision of public services. Other innovations have included a latent variable approach, using more than one measure of output as well as hedonic methodologies. A particularly interesting example is the value-added approach used in the education literature proposed by Hanushek and Taylor (199O). [4] Future research should consider the merging of the demand-and-supply-side studies to better identify public sector output. [5]

   Often data availability limits the scope of services one can consider. However, the researcher must still be cognizant of the implications for the results of the study. When one attempts to benchmark, what does the measure mean if not all outputs are included? I can make observations from my own work; in a study of police departments, do you include all crimes and traffic violations (see Grosskopf, Hayes, and Hirschberg 1995)? How should safety be measured? When studying primary education (see Grosskopf et al. 1997), must you consider special education classes? Additional examples would be services provided by parks and libraries, where quality is an important factor.

3. Appropriate Objective Functions for Public Service Providers

   The researcher's responsibility is to determine a benchmark by which public sector managers can be evaluated. For this research, we need to specify the goals the decision maker ought to have, given the structure of the public service provision. If in fact the public sector managers have personal goals that are different from serving constituents, this will be revealed as inefficiency. [6] This section examines the alternative objective functions the researcher might encounter and the data requirements for establishing the benchmark.

   Starting with an analytical framework, let's say there are y outputs or services that are produced by a public sector agency and that there are x inputs that are used to produce the...