Efficiency, Conflicting Goals and Trade-Offs: A Nonparametric Analysis of the Water and Wastewater Service Industry in Italy

This paper presents a benchmarking study of the water and wastewater industry in Italy. A three-stage modeling approach was implemented to measure the efficiency of 53 utility operators. This approach is based on the implementation of network and conventional data envelopment analysis (DEA) to model the production process of the water service utility operators. In comparison to the conventional black-box or one-stage production model generally adopted in previous studies, the proposed approach provides information relative to the different efficiency components of the stages and blocks of the water service production process and its overall efficiency. Further, by shifting the efficiency analysis to a two-dimensional performance space, i.e., resource and market-efficiency, it offers a more comprehensive view of the water service industry and allows accounting for different business goals at the same time and for an investigation of industry trade-offs. Results show that the operators’ efficiencies in the Italian water service industry are generally variable and low. There are no water service utilities which are 100% efficient from the resource-efficiency perspective, and the maximum efficiency score is 0.545. Efficiency measurements suggest that there is a general orientation of the Italian water industry to not invest in upgrading and improving the infrastructure assets, and achieving an acceptable efficiency in the operations is critical to delivering water services to market in an efficient way. Only one utility operator is 100% efficient from the market-efficiency perspective. The low tariffs adopted by the water service operators do not allow the gaining of satisfactory service remuneration and the achievement of long-term business sustainability. The joint analysis of the resource and market efficiency scores indicates that there is a trade-off between the corresponding business goals.