In recent times, the solutions to water supply problems in India have increasingly focused on improving the service efficiencies of existing services. This implies that the performances of water supply operations are analyzed to bring about improvements in the water supply services. These performances are measured in terms of specific attributes of an individual utility, and therefore the determining factors underlying the changes in various measures of efficiency assume significant importance.
This Paper employs the use of a Regression analysis to determine the nature of relationship between the input and output variables for performance evaluation carried out using the Stochastic frontier analysis (SFA) based approach to evaluate efficiencies of water supply utilities. This paper details the importance of these determinants on performance measures with reference to policy implications and outcomes.
Purpose
Increased emphasis on performance improvements can bring benefits of large internal savings which may be beneficially deployed for greater expansion and coverage of services especially to the poor, and for furthering service quality improvements. Hence it becomes imperative to measure efficiencies, and to identify those operational variables that play the most significant role in the determination of costs.
Key Issues
In the developing country context, the efficiencies of water supply utilities often remain unmeasured in India. Neither are the determinants of efficiencies known. The present work identifies gaps between actual and existing best practices, and gauges the extent of savings possible if inefficiencies are mitigated. Impact of various operational parameters on efficiency measurements such as length of distribution network, number of labor per connection, percentage of connections metered, percentage of surface water and the nature of ownership are also determined and discussed in detail.
Methodology
Stochastic Frontier Analysis (SFA) was applied for data from a total of 20 cities in the urban areas of India. For a utility with quantity of Output (Q), the selected inputs are Staff per thousand connections (L/’000), Length of the network (L), Density (Population served/Area covered by the service) (Dens) and Non-revenue Water (NRW %) as represented by the production function, Q = ƒ{(L/’000), Len, Dens, NRW(%)}.A total of 3 models were analyzed using various combinations of inputs-output variables in different functional form to choose an appropriate SFPF model. The assumptions of the “isotonicity” relationships between the input and output factors were examined. A regression analyses was further carried out to aid an understanding of the systematic causes of variation in the efficiency of water utilities, and to derive the extent of importance of various operational parameters on efficiencies of utilities.
Analysis, results and implications for policy and /or research
Typical results reveal that the metro cities have efficiencies (79.16%) higher than mean sample efficiency of 65.69%, indicative of the fact that larger urban centers get greater attention from the service managers. The Mean efficiencies for larger cities (82.45%). also exceeded mean efficiencies for smaller city group (61.92%). The Regression analysis indicates that if losses increase by 10% then inefficiencies increase by 1.5%.The paper also determines the scope of total additional average daily production of water if best practices are followed. The regression results reveal that percentage of metered connections and length of distribution network are major determinants of performance of water utilities.