Cross-country comparisons avoid the unsteady equilibrium in which regulators have to balance between economies of scale and a sufficient number of remaining comparable utilities. By the use of data envelopment analysis, we compare the efficiency of the drinking water sector in the Netherlands, England and Wales, Australia, Portugal and Belgium. After introducing a procedure to measure the homogeneity of an industry, robust order-m partial frontiers are used to detect outlying observations. By applying bootstrapping algorithms, bias-corrected first and second stage results are estimated. Our results suggest that incentive regulation in the sense of regulatory and benchmark incentive schemes have a significant positive effect on efficiency. By suitably adapting the conditional efficiency measures of Daraio and Simar (2007) to the bias corrected estimates of Simar and Wilson (1998), we incorporate environmental variables directly into the efficiency estimates. We firstly equalize the social, physical and institutional environment, and secondly, deduce the effect of incentive schemes on utilities as they would work under similar conditions. The analysis demonstrates that in absence of clear and structural incentives the average efficiency of the utilities falls in comparison with utilities which are encouraged by incentives.
