Water‐Energy Nexus: Benchmarking the Energy Efficiency of Drinking Water Treatment Plants Accounting for Heterogeneity

ABSTRACT The assessment of energy efficiency in water facilities is receiving increasing attention as a key strategy in the transition toward an energy‐neutral urban water cycle. This study evaluates the energy efficiency of 146 drinking water treatment plants (DWTPs) in Chile by applying a latent class stochastic frontier analysis (LCSFA) to account for both observable and unobservable sources of heterogeneity. Unlike traditional models that assume homogeneity across units, the LCSFA approach identified two distinct groups of DWTPs with significantly different operational characteristics, pollutant loads, and efficiency profiles. Group 1 exhibited an average energy efficiency score of 0.534, while Group 2 showed a significantly higher average of 0.737. In Group 1, efficiency scores estimated under the pooled frontier are higher than those obtained from the latent class model for 73% of DWTPs, whereas in Group 2 the pooled frontier yields lower efficiency scores for 58% of plants. The analysis also revealed substantial energy‐saving opportunities. Group 1 facilities could reduce energy use by an average of 0.112 kWh/m 3 , compared to 0.044 kWh/m 3 in Group 2. These findings demonstrate the importance of adopting heterogeneity‐aware methods for fair and accurate benchmarking. The study offers methodological innovation and practical insights for regulators and utility managers aiming to design targeted energy efficiency interventions and performance‐based incentives in the water sector.