ABSTRACT
This chapter computes Data Envelopment Analysis (DEA) and Free Disposal Hull (FDH) measures of technical efficiency and returns to scale, for public services provided by municipalities of the Brazilian Northeast and Southeast regions. It evaluates the sensibility of those computed efficiency scores when the decision-making units are highly heterogeneous, as is the case with the Brazilian municipalities. The chapter presents the data-base and discusses the physical indicators chosen as proxies for the supply of public services as well as the dispersion of the data set. It discusses non-parametric efficiency calculations for FDH and DEA reference technologies. The chapter covers the relationship between efficiency, heterogeneity of the data set and returns to scale. It shows that diversity among the data set influences considerably efficiency measurements. In particular, DEA-based efficiency estimates are quite sensitive to the dispersion in the data set. Compared to the DEA techniques, FDH-based estimates are much less vulnerable to the impact of heterogeneity.
