Malaysian Water Utilities Performance With the Presence of Undesirable Output: A Directional Distance Function Approach

Abstract

Non-Revenue Water (NRW) is water losses in the distribution process and it affects water supply management worldwide. Malaysia is not excluded and the authority has put a high priority on NRW as it affects the revenue collection. Consequently, NRW is established as one of the Key Performance Indicators (KPIs) to assess the efficiency of Malaysia water supply industry. However, the current policy is impractical; the assessment of all the water utilities is against a single NRW target. Moreover, NRW should be considered as an undesirable product in the water supply system. Therefore, an alternative to Data Envelopment Analysis (DEA)-based approach called Directional Distance Function (DDF) is applied to measure the performance of the integrated production of desirable and undesirable outputs. The result shows that the measurement of water technical efficiency is more explicit using DDF model, where the potential reduced level of NRW for each inefficient water utility can be determined in order to improve their performance. This is in line with the government’s aim to reduce the NRW level countrywide.

Evaluating the Efficiency of Water Distribution Network Sectors Using the DEA-Weight Russell Directional Distance Model: The Case of the City of Valencia (Spain)

“…Recent studies have used complex variants of the DEA, such as super-efficiency (e.g., to investigate the potential for efficiency improvement [4]), DEA-scale efficiency (e.g., to examine the efficiency of leakage-management [14]); DEA model with statistical tolerance (e.g., to assess the efficiency of water and sewerage companies (WaSCs) [38]); shared input data envelopment analysis model (e.g., to separately measure the efficiency when the same operator delivers more than one service, as water and wastewater services [39]); the directional metadistance function (e.g., to revisit the relationship between ownership and performance [13]; Bootstrap DEA (e.g., to identify the determinants of efficiency of water provision services [40,41] or to evaluate the influence of the management nature (private vs. public) on efficiency [10]); DEA double bootstrap (e.g., to overcome the limitation of deterministic method that does not allow identifying environmental factors influencing efficiency scores [42]); the DEA-based approach on Directional Distance Function (DDF) (e.g., to measure the performance of the integrated production of desirable and undesirable outputs [43]); or Network DEA models (e.g., to overcome the shortcomings of the standard AED for relative performance assessment (which do not allow setting clear guidelines for improvement) [44]).…”

“…In this sense, the aforementioned models do not provide information on the efficiency of each of the variables (inputs and desired and undesired outputs) of each of the DMUs, or that they only incorporate information on the generation of desirable outputs but not on undesirable ones, or that they are not directional. Only one of the models cited in that section incorporates information on evils and also considers the directional distance function [43], but it does not include inputs in the DDF (unlike the DEA-WRDD model). In other words, DEA-WRDDM assigns a specific inefficiency value for each variable and each DMU, while the DDF chooses only one for each DMU, not for each variable (therefore, the information provided by the WRDDM is much richer).…”