Development of water utility performance index using hybrid aggregation technique for water supply systems in India

Water utility performance indices (WUPI) are a simple method for performance assessment which involves summarizing and accounting data of several indicators into simple words or numeric values for easy understanding and reporting to policymakers and the general public. The development of such indices involves the use of aggregation functions such as additive, multiplicative, and harmonic. Unfortunately, use of these aggregation techniques results in the problem of eclipsing and ambiguity, wherein the final index value does not represent the actual overall performance. Currently, it is unclear as to how the selection of appropriate aggregation function be made so as to resolve this problem of eclipsing among different variables. This study attempts to fill this gap by developing water utility performance indices (WUPI) using a hybrid aggregation technique (combination of additive and geometric methods). Specifically, it proposes a framework for development of WUPI based on the use of principal component analysis (PCA) and a hybrid aggregation technique. The parameters that have a very strong correlation are first clustered into groups using the principal component analysis. Thereafter, the sub-index values for each group are aggregated using the equal additive method. Lastly, the sub-indices are aggregated using geometric mean to arrive at the final WUPI value. The methodology is then applied to rank a total of 78 million-plus cities, 81 municipal corporations and 553 cities/municipalities in BIMARU states in India based on nine performance indicators. These indicators are: coverage connections, per capita supply, metering of connections, non-revenue water, continuity of supply, quality and treatment, redressal of customer complaints, cost recovery and efficiency in collection of charges. Results reveal that Nashik in Maharashtra ranks first among all the million-plus cities and municipal corporations and ranks in "very good" category. In fact, seven out of the top ten million cities are in the state of Maharashtra and rank in "very good" or "good" category. This study is probably the first attempt, wherein a PCA and modified aggregation have been used for development of water utility performance indices. The findings of this study will have implications for the water supply utilities in evaluating their performance for improvement.