Content area
Full Text
INTRODUCTION
William A. Niskanen's (1971, 1975) theory of bureaucracy has been a source of much debate and research.l The theory assumes that utility-maximizing bureaucrats will seek to maximize their agency's discretionary budget. A central implication of this theory is that bureaus will supply output inefficiently or in quantities greater than that desired by citizens.
The authors test Niskanen's hypothesis by examining the decision of cities to contract out for services. Their premise is that city managers fit the characteristics of the greedy bureaucrat Niskanen describes. If his theory is valid, then city managers should be expected to avoid contracting out because it would expose their discretionary budget and make it more difficult to use for their own purposes.
At odds with Niskanen's theory in this case is the municipal reform literature that argues that professionalism in city management will seek to perform services in the "best" or most efficient way, presumably, a manner in which the discretionary budget is minimal or zero. In this case, the authors might expect to observe contracting out in professionally managed cities. These opposing hypotheses thus allow for a test of two competing theories of bureaucracy. Contracting out is one form of privatization. Privatization increases the service production and delivery options available to government managers. Contracting out may enhance cost-efficiency and service quality (Kettl, 1988) or it may lead to cost overruns, shoddy workmanship, and corruption (Hanrahan, 1983). While there has been much debate about the merits of contracting out, there is little evidence about the propensity of different cities to contract out for services. In this article, the authors examine the difference in contracting out between city manager cities and cities with other forms of governance.
THE GREEDY BUREAUCRAT HYPOTHESIS
Niskanen derives a model that assumes the bureaucrats maximize utility based on their income and their perquisites. These, in turn, are based on a bureau's output and its discretionary budget. Niskanen (1975: 618-619) defines discretionary budget as "the difference between [the bureaucrat's] total budget and the minimum cost of producing the expected output." It is not the total agency budget. In the first case, where it is assumed that utility is totally based on the discretionary budget, the result is that the bureau will produce the optimal output inefficiently....