Measuring the eco-efficiency of the provision of drinking water by two-stage network data envelopment analysis

The water production system is a complex network and policy makers can get a better insight into how the system operates if they know how efficient each sub-process is. In this study, and for the first time, we employ a two-stage network Data Envelopment Analysis model to evaluate the eco-efficiency of water services in England and Wales, integrating greenhouse gas (GHG) emissions as an undesirable output. We then use regression techniques to determine the impact of environmental variables on a water company's efficiency. The results indicate that from an economic perspective (first stage), companies need to reduce the costs of running their business by 22.3% on average to produce the same level of services. From an operations and environmental point of view (second stage), companies need to curtail down the levels of inputs and GHG emissions by 32.6% on average to generate the same level of output. Thus, the mean overall eco-efficiency was 0.514, which means that the potential for input and GHG emissions savings among companies was approximately 48.6%. Moreover, the complexity of water treatment and average pumping head could lead to lower eco-efficiency. Several policy implications are finally discussed.