Evaluating the performance of urban water utilities: Robust nonparametric approach

This paper empirically analyzes the efficiency of urban water utilities using state-of-the-art methodology combining data envelopment analysis (DEA) and a two-stage double bootstrap procedure. In the first stage, robust efficiency estimates are obtained with an improved DEA analysis. In the second stage, a truncated regression model and a double bootstrap procedure are used to estimate the effect of a set of environmental variables on unbiased DEA estimates. The findings suggest that the efficiency scores obtained after bias correction are significantly different to the original efficiency estimates. The results also show the existence of a significant relation between efficiency and several environmental variables including the proportion of water sourced from groundwater, customer density, and residential consumption.