As the number of wastewater treatment plants (WWTPs) has increased, the economics associated with their management have become more relevant. The efficiency assessment is therefore a useful tool fo cost reduction. For this purpose, Data Envelopment Analysis (DEA) is a highly suitable technique, since it is a holistic approach that aggregates performance indicators into a single index. However, one of the most common criticisms of DEA models is that information on uncertainty estimates is not provided. To overcome this limitation, we assess efficiency by using a DEA model with statistical tolerances for both
inputs and outputs. This model is applied to a sample of Spanish WWTPs. The results show that WWTP efficiency scores change when data modifications are incorporated. In addition, we verify that not all WWTPs have the same sensitivity with respect to changes in the inputs and outputs. Moreover, WWTPs are ranked in terms of efficiency, allowing the identification of facilities with the best practices, which will serve as a reference for minimizing operating costs at other plants. This empirical application illustrates that the combination of the DEA model with uncertainty assessments provides more robust results, leading to more reliable conclusions than traditional DEA. From a policy perspective, the incorporation of uncertainty in the DEA model with tolerances allows future performance of the WWTPs to be predicted and ranked, demonstrating the usefulness of this approach.
