Explaining differences in productive efficiency: An application to Belgian municipalities

The purpose of this paper is to measure and explain variations in productive efficiency of municipal governments in Belgium. Technical efficiency is evaluated using a non-parametric method based on the Free Disposal Hull (FDH) reference technology. We first calculate input, output, and graph Farrell efficiency measures for a cross-section of all 589 Belgian municipalities. In a second stage of the analysis we explain the calculated differences in efficiency in terms of variables related to the structural characteristics of municipalities and to the institutional environment.     

Sizing the government

Is there such a thing as an optimal government size? We investigate the so-called Armey curve, which claims an inverted U-shaped relationship between government size and economic performance, using non-parametric Data Envelopment Analysis (DEA). The DEA scores are linked to control variables, such as initial per capita income, openness, population density, urbanization, country size and family size. For 23 OECD-countries we estimate the country specific efficiency scores, which reveal the extent to which a country uses excess public resources to achieve the observed growth rate of GDP.     

The need for performance auditing in the public sector and the best-practice frontier

The article is organized as follows. In Section 2 the arguments for performance auditing in the public sector are reviewed along the lines of the public-choice approach. An international comparison of the functions of supreme auditing institutions (Section 3) reveals that the BelgianRekenhof-Cour des Comptes lags behind, lacking a mandate for performance auditing. In Section 4 the notions of efficiency and effectiveness are clarified. Section S compares the concept of the best-practice frontier with a standard production function. In the public sector one is often confronted with a weak knowledge of the underlying technology. This poses problems of specification for a parametric estimation. Therefore a nonparametric frontier may be recommended. In Section 6 the Farrell method and the free disposal hull are presented. Moreover, when several inputs are transformed into a multiple output (instead of one single homogeneous output), data-envelopment analysis offers a solution (Section 7). These latter methods are also illustrated by an international empirical comparison of the efficiency of public enterprises: postal services in Section 6 and railway companies in Section 7. Some conclusions are reported in the final Section 8.     

The Status of the Budget Constraint,Federalism and the Relative Size of Government: A Bureaucracy Approach

We develop a model along the lines of Niskanen,articulating that under a soft government budgetconstraint the full production cost of the public goodis not reflected in the tax price as perceived by theconsumer-taxpayer-voter. Various proportions ofnon-tax financing and different degrees of votermyopia with respect to discounting the future taxliabilities are taken into account. It can be shownthat both the actual level of public output and theamount of slack resources are lower under a hardbudget constraint than under a soft budget regime.Lower levels of government typically operate under ahard budget constraint when compared with the federallevel since they have only limited (public) borrowingopportunities and no access to money creation(seignorage). In a federalist setting more governmentdecisions are taken under a hard budget constraintthan in a unitary state. Hence one would expect thatthe overall size of government is relatively smallerin a structure with fiscal federalism. An empiricaltest for 19 OECD-countries (1990–1992) seems tosupport this hypothesis.