Measuring water utility performance using nonparametric linear programming

摘要

In this study, a cutting-edge methodology for measuring the performance of water utilities based on two-stage Data Envelopment Analysis (DEA) was applied to individual districts of a California-based water utility. A bootstrap technique involving the construction of confidence intervals was implemented to overcome the deterministic nature of conventional DEA, and a number of exogenous variables were incorporated into the model to help identify the factors affecting technical efficiency. Results indicated high overall performance achieved by the utility on average (92%). The number of connections and precipitation were found to be statistically significant exogenous variables, and both were determined to have a negative impact on efficiencies. The findings of this study are expected to be useful for guiding subsequent managerial improvement initiatives.