Water supply services in India are currently not regulated and remain essential monopolies under the strict control of government. These services are mostly perceived as inefficient, reflecting insufficient municipal revenue generations which result in loss-making municipalities that need to be subsidized by the government. The efficiencies and productivities remain largely unmeasured and undetermined, and therefore the perception of inefficient services is not backed by scientific studies. This paper therefore sets up a framework for measuring the performance and productivity of Indian water supply operations. The study utilizes panel data on 21 municipalities observed over the financial years 2005 and 2010, to analyze the relative efficiencies and productivity growths of utilities. Data envelopment analysis is used to obtain efficiency scores and productivity measures such as the Malmquist index. The results indicate the presence of large relative inefficiencies and decreasing productivity of water supply services over time, alluding to a failure to revise tariffs upwards despite increasing input costs over time, thereby driving the need for sector regulation and tariff revision in accordance with X-factors, which was found to have a mean value of 2.4% for possible price cap regulation in the sector. The paper highlights related policy implications.

INTRODUCTION

Regulatory frameworks can have significant influence on making water services accessible to the poor, and on giving quality service to consumers. A typical regulatory framework would consist of a set of rules and processes that bind the water service providers, including formal rules (laws, contracts, bylaws, etc.) and informal rules (personal commitments, financial incentives, reputation, etc.) (Tremolet 2006), besides defining how the main regulatory functions get allocated to various institutions, which can include an autonomous regulatory agency, a Ministry, an asset-holding company, a customer group, an independent expert, and so forth. Regulatory arrangements further increase and improve access to water services in order to benefit consumers in terms of the availability, affordability, and sustainability of services. It allows the tariffs to be set at appropriate levels, i.e. levels that permit recovering the costs of service provisioning, with reasonable contribution towards further investments, and without jeopardizing the financial stability of the operating utilities. The structures of tariffs, i.e. the balance between the connection and consumption charges, as well as the structure of the consumption charges, are defined so as to provide the right incentives to all customers even as allocating subsidies to those who need them most (Tremolet & Halpern 2006).

A typical regulatory framework is often managed by an independent regulator, such as the economic regulator of the water sector in England and Wales, Ofwat, which remains a non-ministerial government department working for the benefit of customers of water and wastewater services and wider society in England and Wales. Most developing countries, including India, however, do not have an independent regulator for water-supply services, resulting in the formation of government managed monopolies that are far from being professionally managed (Kulshrestha et al. 2012). As a result, sector performances and efficiencies remain unmeasured, and determination of tariffs remains under the whims of local self-governing municipalities whose decisions are forced by political compulsions due to elected representation in local self-governing municipalities.

In the present work, an effort has been made to present a framework wherein tariffs are set using a scientific basis. The performance of urban water supply services is analyzed using data envelopment analysis (DEA), which provides an estimate of the relative efficiencies and productivity growths of some of the urban water supply services in India. The aim of the study is to measure the performance of the sector in order to calculate productivity indices, which assist regulatory authorities in setting Consumer Price Index based CPI-X price paths that might help in improving the performance of these services. The paper also discusses the issues related to data shortcomings that exist in the sector, and the extent to which performance measures should be used in any regulatory planning.

The present work focuses on DEA-based measurement of efficiencies and productivities of water supply services in 21 Indian municipalities, endeavoring to answer the following questions:

  • 1. What is the extent of inefficiencies in the water supply operations of the Indian municipalities?

  • 2. Are the Indian water supply operations improving over time (i.e., do they have positive productivities with respect to time?), and if not, then what can be the possible causes?

  • 3. How should the present tariffs be restructured in order to overcome losses in the sector?

  • 4. What are the quanta of savings that are possible if best practices are followed?

  • 5. What are the policy outcomes in the context of a possible sector regulation?

The paper is organized as follows. Subsequent to this introductory section, an overview of the Indian water supply sector is detailed, which is followed by a brief review of relevant literature in the third section. The following section describes the DEA method used in this paper before presenting and discussing the results in the subsequent section. The concluding remarks are made in the final section.

INDIAN WATER SUPPLY SECTOR

Under the Constitution of federal India, water supply is largely a State subject, and following the 74th Constitutional Amendment, the States assign the responsibility to the urban local bodies (ULBs) for managing water supplies in urban areas. The Central government allocates funds, besides ensuring that funds are also provided in the State budgets. States generally plan, design and execute water supply schemes (and often continue to operate) through their State Public Health Engineering Departments, Water Boards and similar bodies, and through municipalities. The entire water supply sector has therefore so far been government managed, without any participation of the private sector. To the government's credit, progressively increasing allocations have been made for the water supply sector in the various Five Year Plans. However, these allocations fall short of what actually is required (MoUD & PA 2004) due to competing allocations in other sectors such as health, education, infrastructure etc.

India is increasingly facing the problem of rapid urbanization, which itself is a result of rapid economic growth in an otherwise traditional agrarian economy. In India, the share of urban areas in net domestic product has been on the rise: from 31% in 1971 to 41.1% in 1981, and is estimated to rise to over 50% in coming years (HSMI 2003). High productivity of urban areas is however, subject to the expansion and improvement of urban infrastructure services. It is therefore certain that infrastructure services will have to grow proportionately to make up for the backlog as well as to cater for future needs, since urban centers would continue expanding, constituting an unenviable task for the urban planners and policy-makers. In this context, Kulshrestha et al. (2012) have identified a host of sustainability issues in the urban water supply sector of India owing their roots to low-tariffs and poor governance that would be of interest to policy-makers.

Water tariffs in India (as well as in the whole of South Asia) are too low to ensure that water be regarded as an economic good (Kulshrestha & Mittal 2005). In urban areas they were set at the equivalent of about US$0.10 per cubic meter in 2007 and recovered about 60% of operating and maintenance costs. Since India has a federal structure, the Central and State governments of India together spend an estimated US$1.1 billion per year on subsidizing the water sector; accounting for 4% of all government subsidies in India and amounting to 0.5% of GDP. As many as 69% of households in larger cities, 45% in smaller cities, and only 9% in rural households have piped water supplies available. Water metering should ideally be the precondition for billing water users based on volumes consumed, but in India, the share of metering is only 50% according to the survey carried out by the Indian Ministry of Urban Development in 2006 in 28 cities. Most customers pay a flat tariff independent of consumption (WSP 2010). The average tariff for all customers, including industrial, commercial and public customers, is US$0.073 per cubic meter (ADB 2007). These tariffs are typically uniform linear tariffs except in a few cities where increasing block tariffs apply.

Water tariffs are decided by ULBs, with members being elected as representatives for fixed tenures. The towns are divided into wards, and ULB representatives are elected from each ward, and hence there exist considerable political interference and constraints in deciding tariffs, as votes are to be garnered in the election process of ULB representatives. A chief officer, along with officers like engineers, sanitary inspectors, health officers and education officers, etc., is appointed by the State government to control the administrative and technical affairs of the municipal bodies. These municipal bodies mostly look after not just the water supplies, but also the collection of solid waste and wastewater treatment. However, the current study confines itself to water service operations of ULBs only, which is the focus area of reforms being envisaged.

One of the major thrust areas of reforms being undertaken in India is the induction of efficiency enhancement steps in ULBs. The problems of sustainability of water availability, maintenance of the supply system, and dealing with the issue of water quality are recognized as the major challenges in the National Eleventh Five Year Plan (Planning Commission 2007). The solution to water supply problems in India has traditionally related to capacity addition or quality improvement, rather than improvements in service efficiencies of the existing services. Efficient operation of existing water supply schemes is the theoretical first step to make the schemes operate viably. Such internal efficiency measures may incorporate increasing the availability of water, and reducing the operational expenditures. Since the low quality of service is the biggest obstacle to the levying of reasonable user charges, efficient and better operation are likely to improve the acceptability of higher user charges. For accessing financial resources from the market and for induction of the private sector, the urban water supply utilities need to be made creditworthy and efficient. Hence, the performance of the sector needs to be analyzed in order to bring about improvements in the water supply service (Kulshrestha 2009).

LITERATURE REVIEW

In this section, a review of studies related to economic analysis of urban water supply services is presented in the context of sector regulation.

In CPI-X price regulation, the regulated sector is permitted to increase its prices over a particular period (usually five years) by varying the consumer price index (CPI) minus an X-factor value. The X-factor is generally called a productivity offset, because it reflects the degree to which the regulator believes the sector can improve its productivity (i.e. reduce or increase its costs in real terms). This generally involves the computation of average annual productivity growth using historical data and the computation of relative efficiencies, which are calculated relative to an expected production frontier using methods such as DEA or stochastic frontier analysis (SFA). These methods have benefits of less persistence and better incentives for efficiency improvements (Coelli & Walding 2005).

In a study that included more than 50 regulators worldwide, Marques et al. (2011) analysed the use of benchmarking in water regulation. They found that 72% of the regulators applied benchmarking either in the economic regulation or in the quality of service regulation for the setting of prices and tariffs. The study further revealed that as far as economic regulation is concerned, around 41% (20 countries) adopted benchmarking methodologies to determine X-factors, which helped in price cap regulation. Berg & Marques (2011) carried out literature surveys of quantitative studies of water and sanitation services (WSS) and identified 190 articles that were published between 1969 and 2009, and used cost of production functions to evaluate the performance of WSS utilities. The literature exposed that about 58% of the quantitative studies used parametric methods, while only 34% used non-parametric methods. Nearly 28% of the studies used Malmquist Productivity index and total factor productivity (TFP) indices for measuring productivities and efficiencies in the sectors, while most of the studies used DEA for calculating efficiencies.

Studies conducted in the USA

Econometric analysis of water supply utilities first started in the USA to evaluate relative efficiencies of public vs. private utilities. Zardkoohi & Crain (1978) estimated a Cobb-Douglas cost function involving a (logarithmic) regression of cost on input variables like the labor price, capital price and an ownership dummy variable. The output variable used in the analysis was the volume of water delivered. The results showed that the public firms had significantly higher costs, relative to private firms. The cost measure used in the study was the sum of operating, maintenance and depreciation costs.

Bruggink (1982) found that private firms reduced operating costs and limited potential rate hikes, and thereby left customers happy as they also handled customer service operations more effectively. The approach used was similar to that employed by Zardkoohi & Crain (1978), but this was criticized by Feigenbaum & Teeples (1983), who asserted that there were no major differences in the costs of public and private firms. Nevertheless, the capital costs were excluded from the cost measure, which seems eccentric given that capital costs commonly exceeded operating costs in nearly all network utilities. Byrnes et al. (1986) used DEA to calculate technical efficiency (TE) for each firm in the model to address this issue. The model used one output variable, i.e. volume of water delivered, and seven input variables, i.e. length of pipeline, ground water, part-time labor, surface water, purchased water, full-time labor, and storage capacity. The results obtained showed that there were no significant differences in the TE scores (nor the scale efficiency (SE) scores) of public versus private firms. The study can be criticized for not counting more output indicators as in the Feigenbaum & Teeples (1983) study. However, the input variables used are expected to control these differences in output characteristics.

Teeples & Glyer (1987) used data on water utilities in California and provided the model limitations, which were left undeclared in the earlier papers of Feigenbaum & Teeples (1983) and Bruggink (1982). Bhattacharyya et al. (1994) calculated econometric cost functions with several approaches such as the use of first order equations, SFA etc.

Lambert et al. (1993) used DEA to assess 32 private firms and 238 public firms. He used water delivered as the output variable, while the inputs comprised value of capital, value of materials, value of labor and value of energy. The study did not find a significant difference between the performances of public and private firm management.

Studies conducted in the UK

Sawkins & Accam (1994) studied the Scottish water industry, and provided some initial strategy on the likely choice of input and output variables. Cubbin & Tzanidakis (1998) compared DEA over regression analysis using data from the UK water industry for the financial year 1994/95. The output variables used in the analysis were water delivered, length of the mains and proportion of water delivered to non-households. The study concluded that DEA was the best technique to be used when large samples are available. Besides, Thanassoulis (2000a, 2000b) used Ofwat data to analyze performances for UK water distribution through DEA, using operational expenditure as input, and number of properties connected, length of the mains, and volume of water delivered as outputs, and established the use of DEA in the Regulation of Water Companies.

Studies conducted in India

Vishwakarma & Kulshrestha (2010) assessed the performance of urban water supply utilities in the state of Madhya Pradesh, India, and estimated their cost efficiencies using SFA. Total cost was taken as the output, and total volume of delivered water, price of labor and price of electricity as inputs. Four models were adopted: Cobb-Douglas cost function, Translog cost function truncated normal Error components model, Translog cost function truncated half normal Error component model and Translog cost function truncated normal TE model. The analysis was done for 18 municipalities in Madhya Pradesh, India, based on 2005 data.

Gupta et al. (2006) used an output-oriented DEA approach to estimate the efficiency of 27 urban water utilities.

Kulshrestha & Vishwakarma (2013) evolved a framework for evaluating the relative efficiencies of the water supply sector and applied DEA to 20 urban centers in the state of Madhya Pradesh, India. They employed three models, with output variables being the number of connections, length of distribution network and average daily clear water production. Model I used operating expenditure as the input, while Model II used staff per 1,000 connections and operating expenditure as inputs and Model III was framed using staff per 1,000 connections, unaccounted-for water (UFW) and operating expenditure as inputs. The overall results indicated that the water supply services were quite inefficient compared to the best practices. The metro cities were found to be most efficient, owing to the fact that larger cities were getting better services compared to smaller ones, which calls for policy interventions. The results also showed that significant savings would be possible if best practices were adopted with respect to operating expenditure, number of staff and UFW.

Studies conducted in various parts of the world

In addition to these US and UK studies, a handful of further studies have appeared in recent years. For instance, Marques (2006) studied the Portuguese water and sewerage services to regulate the sector. He employed DEA to develop a standard model. A comparison between public and private water utilities was carried out by Marques (2008) and water services under private management were found to be more productive than public ones. Tupper & Resende (2004) studied the Brazilian water and sewage sectors for regulation. Coelli & Walding (2005), published a working paper series on performance measurement in the Australian Water supply industry for 18 utilities over an eight-year period using DEA on the panel data available.

In Asia, Aida et al. (1998), employed DEA to evaluate the efficiencies of 108 water supply services in the Kanto region of Japan. Marques et al. (2014) analyzed Japanese water utilities for many factors affecting efficiency, governance, monthly water charge, water losses, production etc. The study derived conclusions regarding efficiency patterns in Japan, updating previous studies that had fewer observations. The study also extended the literature by studying the impact of various exogenous variables on the efficiency of water utilities.

MATERIAL AND METHODS: PERFORMANCE MEASUREMENT USING DEA

DEA is a multi-factor productivity analysis for measuring the relative efficiencies of a homogenous set of decision-making units (DMUs), and can be applied to analyze multiple outputs and multiple inputs Charnes et al. (1978). There are a number of DEA models, which can be either input or output oriented. For water utilities, input quantities act as decision variables that need minimization, as the output is often fixed. Hence, the DEA models discussed below have an input orientation as also suggested by Kulshrestha (2009) and Kulshrestha & Vishwakarma (2013).

Efficiency measurement using DEA

The CRS model was suggested by Charnes et al. (1978), and assumes constant returns to scale (hence the name CRS model). The chosen sample z utilities are named as DMUs, each with m inputs and n outputs, the relative efficiency score of a test DMU p is obtained by solving the model proposed by Charnes et al. (1978). 
formula
 
formula
 
formula
where

i = 1 to z,

j = 1 to m,

k = 1 to n,

yki = amount of output k produced by DMU i,

xji = amount of input j utilized by DMU i,

vk = weight given to output k,

uj = weight given to input j.

The above fractional program is subsequently converted to a linear programming format and a mathematical dual is employed to solve the linear problem, as it reduces the number of constraints from z + m + n + 1 in the primal to m + n in the dual, thereby rendering the linear problem easier to solve (Charnes et al. 1978).

minθ,λθ 
formula
 
formula
 
formula
where θ = efficiency score and λ = dual variables (weights in the dual model for the inputs and outputs of the z DMUs).

θ is the input TE score having a range of value 0 ≤ θ ≤ 1. If the value of θ = 1, it indicates that the region is on the frontier. The vector λ is an N × 1 vector of weights that define the linear combination of peers of the ith utility. The linear programming problem needs to be solved N times (i.e. for each DMU) and a value of θ is provided for each utility in the sample.

The variable returns to scale (VRS) DEA model is defined by adding the convexity constraint (Banker et al. 1984), ∑λi = 1 
formula
 
formula
 
formula
 
formula
 
formula
The VRS DEA model is more elastic and envelops the data in a tighter way than the CRS DEA; the VRS TE score is equal to or greater than the CRS or ‘overall’ TE score. Under the VRS model, a utility with an efficiency score of less than unity is considered as neither technically efficient nor scale efficient. On the other hand, the utility with a unit efficiency score is considered purely technically efficient but it may not be scale efficient.
The SE of the ith utility can be represented as: 
formula
where SE = 1 implies scale efficiency and SE <1 indicates scale inefficiency.

DEA allows computing a reference unit for inefficient DMUs and the essential improvements required in the inefficient unit's inputs and outputs to make it efficient are noted. It should be distinguished that DEA is primarily a diagnostic tool and does not lay down any re-engineering strategies to make inefficient units efficient. Such development strategies need to be identified and implemented by water managers and administrators understanding the operations and management models of the efficient units.

Productivity measurement using DEA

If one has access to suitable panel data, Fare et al. (1994) have shown that DEA frontier construction methods can be used to obtain estimates of Malmquist TFP index numbers. This approach also has an advantage relative to PIN (Price based index numbers) TFP methods with respect to the fact that Price data are not required and the TFP indices obtained may be decomposed into components like technical change (TC) (frontier-shift), and technical efficiency change (TEC) (catch-up).

The TFP change (TFPC) between two data points is measured using the Malmquist TFP index, which calculates the ratio of the distances of each data point relative to a common technology. The Malmquist (input-orientated) TFPC index between period s (the base period) and period t can be written as follows if the period t is used as the reference technology. 
formula
1
On the other hand, if the period s is used as reference technology the Malmquist TFP index can be written as: 
formula
2

The notation represents the distance from the period t observation to the period s technology. This distance is equivalent to the TE scores when t = s as defined earlier. A value of Mi (Malmquist TFP index) greater than one indicates positive TFP growth from period s to period t; at the same time, a value less than one indicates a TFP decline.

These two (period s and period t) indices are only equivalent if the technology is output neutral as noted by Fare et al. (1998), i.e. if the output distance functions may be represented as . The Malmquist TFP index is often defined as the geometric mean of two indices, namely TEC and TC, to avoid the necessity to either impose this constraint or to randomly choose one of the two technologies. 
formula
3
A corresponding way of writing this productivity index is 
formula
4
The ratio outside the square brackets measures the change in TE between periods s and t, i.e., the efficiency change is equivalent to the ratio of the Farrell TE in period t to the Farrell TE in periods. The remaining part of the index in Equation (4) is a measure of TC. It is the geometric mean of the shift in technology between the two periods, evaluated at xt and also at xs. Thus the two terms in Equation (4) are: 
formula
5
and 
formula
6
The required linear programs (LP) s to solve four DEA LP problems to calculate the four distance measures in Equation (5) is: 
formula
 
formula
 
formula
 
formula
 
formula
 
formula
 
formula
 
formula
 
formula
 
formula
 
formula
 
formula
 
formula
 
formula
 
formula
 
formula
where
  • yi is an M × 1 vector of output quantities for the ith water utility (M: number of outputs),

  • xi is a K × 1 vector of input quantities for the ith water utility (K: number of inputs),

  • Y is an N × M matrix of output quantities for all N utilities,

  • X is an N × K matrix of input quantities for all N utilities, is an N × 1 vector of weights, and

  • is a scalar.

This study calculates the productivity change between two points, t and t + 1, which represents a comparison of the current year with a year later. Thus, the period s mentioned in the theoretical section can be replaced by t + 1. The output used in the measurement is annual water supply volume. The input consists of three indicators: amount of labor available, tangible fixed assets (such as the underground pipes, and water pumping and storage facilities etc.), and other costs. The solution to these linear programming problems, equations, indicates a level of TE under the conditions of constant return to scale technology, because this is believed to be more reliable than a variable return to scale when calculating Malmquist TFP indices (Coelli et al. 2005; Fox, 2002). From the TFP indices obtained from the measurement, a second-stage estimation to determine the impact on the TFP productivity change as a dependent variable may be necessary, as indicated in the equation: 
formula
7
where denotes the vector of coefficients; is an unobserved individual effect; is an error term; and X is the vector of independent variables that include subsidy growth, private consignment growth, supply population growth, and length extension of the water main. Notations i and t represent the ith water utility and time t, respectively.

Data employed

The data used in this exercise are taken from Indian government organizations such as the Central Pollution Control Board (CPCB, 2010) and the Central Public Health and Environmental Engineering Organization (CPHEEO, 2005). The data used for the analysis are divided into two sections. The first is the most recent available information on annual data of 21 municipalities from the 2009/10 financial year. This is used to calculate the TE and SE scores. The second data set is the panel data, containing statistics on these 21 municipalities in the year 2005/06.

A simple model for the analysis is framed based on the cost drivers available in the sector. The input and output variables of the model are as follows:

Input variables Output variables 
1. Operational expenditure (OPEX) (Rs. Millions) 1. Total quantity of water supplied in Million liters per day (MLD) 
  2. No. of connections 
Input variables Output variables 
1. Operational expenditure (OPEX) (Rs. Millions) 1. Total quantity of water supplied in Million liters per day (MLD) 
  2. No. of connections 

RESULTS AND DISCUSSION

Table 1 gives the descriptive statistics of the data used for the calculation of CRS TE scores for the periods 2010 and 2005, respectively. The variation between maximum and minimum values for all the parameters is observed to be very large, and the standard deviations for the selected parameters are also large. The scales of operations are ignored in the present study, therefore CRS models are used in the analysis of the data, and the results are plotted in Figure 1. The TE scores are calculated for the recent financial year 2010 and are tabulated in Table 2.
Table 1

Descriptive statistics of data for sample utilities

Variable description Mean Standard deviation Maximum Minimum 
Year 2005 
 Total quantity of water supplied (MLD) 127.10 160.24 682.00 4.00 
 Operating expenditure (Rs. Millions) 9.7864 21.4895 100.5200 0.180 
 No. of connections 58611.14 65532.45 253614.00 940.00 
Year 2010 
 Total quantity of water supplied (MLD) 175.85 180.10 577.76 8.01 
 Operating expenditure (Rs. Millions) 26.3572 29.7792 136.5500 1.2539 
 No. of connections 108767.43 164654.76 742263.00 6218.00 
Variable description Mean Standard deviation Maximum Minimum 
Year 2005 
 Total quantity of water supplied (MLD) 127.10 160.24 682.00 4.00 
 Operating expenditure (Rs. Millions) 9.7864 21.4895 100.5200 0.180 
 No. of connections 58611.14 65532.45 253614.00 940.00 
Year 2010 
 Total quantity of water supplied (MLD) 175.85 180.10 577.76 8.01 
 Operating expenditure (Rs. Millions) 26.3572 29.7792 136.5500 1.2539 
 No. of connections 108767.43 164654.76 742263.00 6218.00 

Data sources:CPHEEO (2005), CPCB (2010).

Table 2

DEA efficiency scores for the year 2009–10

S. no. Name of the utility City/District State TE-CRS TE SE Scale 
Dharmavaram Anantapur Andhra Pradesh 0.25 0.51 0.50 IRS 
Vijayawada Krishna Andhra Pradesh 0.08 0.09 0.87 IRS 
Madanapalle Chittor Andhra Pradesh 0.33 0.53 0.64 IRS 
Srikakulam Srikakulam Andhra Pradesh 0.12 0.36 0.33 IRS 
Hyderabad Hyderabad Telangana 0.34 1.00 0.34 DRS 
Rajahmundry East Godavari Andhra Pradesh 0.07 0.09 0.83 IRS 
Rajkot Rajkot Gujarat 1.00 1.00 1.00 CRS 
Surat Surat Gujarat 0.54 1.00 0.54 DRS 
Bellary Bellary Karnataka 0s.16 0.18 0.90 IRS 
10 Davanagere-Harihara Davanagere Karnataka 0.06 0.07 0.85 IRS 
11 Kolar Kolar Karnataka 0.20 0.54 0.37 IRS 
12 Gadag- Betigeri Gadag Karnataka 0.72 1.00 0.72 IRS 
13 Dewas Dewas Madhya Pradesh 0.07 0.08 0.78 IRS 
14 Aurangabad Aurangabad Maharashtra 0.30 0.31 0.97 IRS 
15 Nashik Nashik Maharashtra 0.41 0.42 0.99 IRS 
16 Amritsar Amritsar Punjab 0.56 0.57 0.98 IRS 
17 Pathankot Pathankot Punjab 0.17 0.22 0.76 IRS 
18 Ghaziabad Ghaziabad Uttar Pradesh 0.16 0.16 0.98 IRS 
19 Kanpur Kanpur Uttar Pradesh 1.00 1.00 1.00 CRS 
20 Varanasi Varanasi Uttar Pradesh 0.17 0.18 0.97 IRS 
21 Gwalior Gwalior Madhya Pradesh 0.21 0.22 0.95 IRS 
 Mean     0.33 0.45 0.78   
S. no. Name of the utility City/District State TE-CRS TE SE Scale 
Dharmavaram Anantapur Andhra Pradesh 0.25 0.51 0.50 IRS 
Vijayawada Krishna Andhra Pradesh 0.08 0.09 0.87 IRS 
Madanapalle Chittor Andhra Pradesh 0.33 0.53 0.64 IRS 
Srikakulam Srikakulam Andhra Pradesh 0.12 0.36 0.33 IRS 
Hyderabad Hyderabad Telangana 0.34 1.00 0.34 DRS 
Rajahmundry East Godavari Andhra Pradesh 0.07 0.09 0.83 IRS 
Rajkot Rajkot Gujarat 1.00 1.00 1.00 CRS 
Surat Surat Gujarat 0.54 1.00 0.54 DRS 
Bellary Bellary Karnataka 0s.16 0.18 0.90 IRS 
10 Davanagere-Harihara Davanagere Karnataka 0.06 0.07 0.85 IRS 
11 Kolar Kolar Karnataka 0.20 0.54 0.37 IRS 
12 Gadag- Betigeri Gadag Karnataka 0.72 1.00 0.72 IRS 
13 Dewas Dewas Madhya Pradesh 0.07 0.08 0.78 IRS 
14 Aurangabad Aurangabad Maharashtra 0.30 0.31 0.97 IRS 
15 Nashik Nashik Maharashtra 0.41 0.42 0.99 IRS 
16 Amritsar Amritsar Punjab 0.56 0.57 0.98 IRS 
17 Pathankot Pathankot Punjab 0.17 0.22 0.76 IRS 
18 Ghaziabad Ghaziabad Uttar Pradesh 0.16 0.16 0.98 IRS 
19 Kanpur Kanpur Uttar Pradesh 1.00 1.00 1.00 CRS 
20 Varanasi Varanasi Uttar Pradesh 0.17 0.18 0.97 IRS 
21 Gwalior Gwalior Madhya Pradesh 0.21 0.22 0.95 IRS 
 Mean     0.33 0.45 0.78   

IRS = Increasing returns to scale; DRS = Decreasing returns to scale; CRS = Constant returns to scale.

Figure 1

DEA CRS efficiency scores for 2010 and 2005.

Figure 1

DEA CRS efficiency scores for 2010 and 2005.

The sample mean efficiencies are as low as 45%, indicating that the water supply services are quite inefficient. In fact, as shown in Table 2, a large number of municipalities have efficiency scores less than 50% therefore indicating that the services are quite inefficient. This confirms the relative lack of infrastructure facilities over municipalities. It is evident from Table 2 that most of the municipalities show increasing returns to scale (IRS) which means that these municipalities give increasing returns for outputs when there is an increase in input. On the other hand, two municipalities show decreasing returns to scale (DRS), implying that these utilities give decreasing returns for outputs when there is an increase in input level. A restructuring of their area under operation is needed for improving efficiencies, and this can be done by dividing the area into two or more parts.

It is also evident from Figure 1 that the efficiencies of most of the municipalities (17 out of 21) are consistently decreasing over time instead of the normal expectation of better performance over time, which shows that the municipalities are functioning poorly and there is deterioration in their service levels over time. If this continues, the productivity growth of the sector will be negative with respect to time. This has implications for possible future privatization or for induction of competition in the sector, as water services are natural monopolies and need lump investments, and no private company would like to invest in utilities that are not just inefficient, but also record negative productivity growths.

As the productivity analysis has difficulty in the choice of price deflators, one must understand that OPEX and capital (CAPEX) are proxies for the quantities of inputs used and therefore considered as fiscal measures. In the analysis of productivity growth, there is an additional complexity of the choice of price deflators. With this point in view, Malmquist DEA TFP measures are calculated, and the TEC, TC and TFPC are summarized in Table 3.

Table 3

TFP results

S. no. Name of the utility City/District State TEC TC TFP 
Dharmavaram Anantapur Andhra Pradesh 0.593 0.786 0.466 
Vijayawada Krishna Andhra Pradesh 0.092 1.124 0.103 
Madanapalle Chittor Andhra Pradesh 0.440 1.095 0.482 
Srikakulam Srikakulam Andhra Pradesh 0.266 0.761 0.202 
Hyderabad Hyderabad Telangana 1.700 0.895 1.521 
Rajahmundry East Godavari Andhra Pradesh 0.133 0.786 0.104 
Rajkot Rajkot Gujarat 1.119 0.714 0.799 
Surat Surat Gujarat 0.893 0.845 0.755 
Bellary Bellary Karnataka 0.160 0.722 0.116 
10 Davanagere-Harihara Davanagere Karnataka 0.070 1.425 0.100 
11 Kolar Kolar Karnataka 0.325 0.835 0.272 
12 Gadag- Betigeri Gadag Karnataka 2.094 0.724 1.517 
13 Dewas Dewas Madhya Pradesh 0.178 0.719 0.128 
14 Aurangabad Aurangabad Maharashtra 0.502 1.044 0.524 
15 Nashik Nashik Maharashtra 1.636 0.893 1.460 
16 Amritsar Amritsar Punjab 0.914 0.800 0.731 
17 Pathankot Pathankot Punjab 0.283 0.892 0.252 
18 Ghaziabad Ghaziabad Uttar Pradesh 0.160 0.942 0.151 
19 Kanpur Kanpur Uttar Pradesh 1.011 1.136 1.149 
20 Varanasi Varanasi Uttar Pradesh 0.206 1.126 0.232 
21 Gwalior Gwalior Madhya Pradesh 0.525 0.907 0.476 
 Mean     0.633 0.913 0.550 
S. no. Name of the utility City/District State TEC TC TFP 
Dharmavaram Anantapur Andhra Pradesh 0.593 0.786 0.466 
Vijayawada Krishna Andhra Pradesh 0.092 1.124 0.103 
Madanapalle Chittor Andhra Pradesh 0.440 1.095 0.482 
Srikakulam Srikakulam Andhra Pradesh 0.266 0.761 0.202 
Hyderabad Hyderabad Telangana 1.700 0.895 1.521 
Rajahmundry East Godavari Andhra Pradesh 0.133 0.786 0.104 
Rajkot Rajkot Gujarat 1.119 0.714 0.799 
Surat Surat Gujarat 0.893 0.845 0.755 
Bellary Bellary Karnataka 0.160 0.722 0.116 
10 Davanagere-Harihara Davanagere Karnataka 0.070 1.425 0.100 
11 Kolar Kolar Karnataka 0.325 0.835 0.272 
12 Gadag- Betigeri Gadag Karnataka 2.094 0.724 1.517 
13 Dewas Dewas Madhya Pradesh 0.178 0.719 0.128 
14 Aurangabad Aurangabad Maharashtra 0.502 1.044 0.524 
15 Nashik Nashik Maharashtra 1.636 0.893 1.460 
16 Amritsar Amritsar Punjab 0.914 0.800 0.731 
17 Pathankot Pathankot Punjab 0.283 0.892 0.252 
18 Ghaziabad Ghaziabad Uttar Pradesh 0.160 0.942 0.151 
19 Kanpur Kanpur Uttar Pradesh 1.011 1.136 1.149 
20 Varanasi Varanasi Uttar Pradesh 0.206 1.126 0.232 
21 Gwalior Gwalior Madhya Pradesh 0.525 0.907 0.476 
 Mean     0.633 0.913 0.550 

From the results it is shown that less than 20% of municipalities (4 out of 21) have a TFP index greater than one, indicating increasing productivities with time irrespective of differences in their input levels. On the other hand, a wide majority (exceeding 80%) of the municipalities (17 out of 21) have TFP index values less than one, reflective of their inability to use the available resources effectively. The productivity of these municipalities decreases with time if best practices are not followed, and they may gradually fall in the negative zone.

This serious observation of decreasing productivities over time may be ascribed to the fact that the present tariffs for water in most cities have little relation to the cost of production, and are set up on the whims of political parties who have a vested interest in keeping tariffs low as they need to seek votes time and again, and populist measures surely are vote garnishes. The gap between the costs of production widens over time due to non-revision of tariffs leading to widening net financial losses for most of the utilities.

Under such a situation, it becomes imperative to ensure that tariffs are periodically reset at appropriate levels, i.e. levels that permit recovering the costs of service provision and include some contribution towards investments, whilst generating considerable profits with the help of an independent regulator.

This is done by calculating the X-factor, which is a measure of productivity offset that helps in improving the present price regime. The shift in the frontier represents TC as shown in Table 3, and TEC can be calculated as the ratio of Malmquist TFP index and TC (Frontier shift). These factors further help in the calculation of X-factors, which in turn help in the regulation of the existing tariff structure. The regulator may require all the municipalities to achieve the weighted mean annual productivity growth of 1.5% in order to prevent the municipalities from remaining in the negative zone. Furthermore, it requires municipalities with DEA TE scores below one to catch up 10% of the way to the frontier over the next five years. This is a conservative approach, designed to reflect the fact that a few municipalities find it difficult to make effective savings due to their low productivity growth, which is a result of excess capacity in their areas. Considering the above rules, calculations for X-factors are made for 21 municipalities. The TE scores are calculated from 2010 data and provide X-factors for each municipality, which are reported in Table 4.

Table 4

Illustrative calculation of X-factors

S. no. Name of the utility City/District State TE TFPC Catch up X-factor 
Dharmavaram Anantapur Andhra Pradesh 0.51 1.5 1.05 2.59 
Vijayawada Krishna Andhra Pradesh 0.10 1.5 1.08 2.62 
Madanapalle Chittor Andhra Pradesh 0.53 1.5 1.04 2.59 
Srikakulam Srikakulam Andhra Pradesh 0.36 1.5 1.06 2.60 
Hyderabad Hyderabad Telangana 1.00 1.5 0.00 1.54 
Rajahmundry East Godavari Andhra Pradesh 0.10 1.5 1.08 2.62 
Rajkot Rajkot Gujarat 1.00 1.5 0.00 1.54 
Surat Surat Gujarat 1.00 1.5 0.00 1.54 
Bellary Bellary Karnataka 0.18 1.5 1.07 2.61 
10 Davanagere-Harihara Davanagere Karnataka 0.10 1.5 1.08 2.62 
11 Kolar Kolar Karnataka 0.54 1.5 1.04 2.59 
12 Gadag- Betigeri Gadag Karnataka 1.00 1.5 1.00 2.54 
13 Dewas Dewas Madhya Pradesh 0.10 1.5 1.08 2.62 
14 Aurangabad Aurangabad Maharashtra 0.31 1.5 1.06 2.60 
15 Nashik Nashik Maharashtra 0.42 1.5 1.05 2.60 
16 Amritsar Amritsar Punjab 0.57 1.5 1.04 2.58 
17 Pathankot Pathankot Punjab 0.22 1.5 1.07 2.61 
18 Ghaziabad Ghaziabad Uttar Pradesh 0.16 1.5 1.07 2.62 
19 Kanpur Kanpur Uttar Pradesh 1.00 1.5 0.00 1.54 
20 Varanasi Varanasi Uttar Pradesh 0.17 1.5 1.07 2.61 
21 Gwalior Gwalior Madhya Pradesh 0.22 1.5 1.07 2.61 
  Mean     0.46 1.5 0.86 2.40 
S. no. Name of the utility City/District State TE TFPC Catch up X-factor 
Dharmavaram Anantapur Andhra Pradesh 0.51 1.5 1.05 2.59 
Vijayawada Krishna Andhra Pradesh 0.10 1.5 1.08 2.62 
Madanapalle Chittor Andhra Pradesh 0.53 1.5 1.04 2.59 
Srikakulam Srikakulam Andhra Pradesh 0.36 1.5 1.06 2.60 
Hyderabad Hyderabad Telangana 1.00 1.5 0.00 1.54 
Rajahmundry East Godavari Andhra Pradesh 0.10 1.5 1.08 2.62 
Rajkot Rajkot Gujarat 1.00 1.5 0.00 1.54 
Surat Surat Gujarat 1.00 1.5 0.00 1.54 
Bellary Bellary Karnataka 0.18 1.5 1.07 2.61 
10 Davanagere-Harihara Davanagere Karnataka 0.10 1.5 1.08 2.62 
11 Kolar Kolar Karnataka 0.54 1.5 1.04 2.59 
12 Gadag- Betigeri Gadag Karnataka 1.00 1.5 1.00 2.54 
13 Dewas Dewas Madhya Pradesh 0.10 1.5 1.08 2.62 
14 Aurangabad Aurangabad Maharashtra 0.31 1.5 1.06 2.60 
15 Nashik Nashik Maharashtra 0.42 1.5 1.05 2.60 
16 Amritsar Amritsar Punjab 0.57 1.5 1.04 2.58 
17 Pathankot Pathankot Punjab 0.22 1.5 1.07 2.61 
18 Ghaziabad Ghaziabad Uttar Pradesh 0.16 1.5 1.07 2.62 
19 Kanpur Kanpur Uttar Pradesh 1.00 1.5 0.00 1.54 
20 Varanasi Varanasi Uttar Pradesh 0.17 1.5 1.07 2.61 
21 Gwalior Gwalior Madhya Pradesh 0.22 1.5 1.07 2.61 
  Mean     0.46 1.5 0.86 2.40 

The results showed the average municipality has a TE score of 46% and the annual average TFPCs rapidly reflecting the better performance of the larger municipalities in the sample. The X-factors measured considering the above conditions lie in a range of 1.54–2.64. This means that the municipality with highest TE score should increase their existing tariffs at a rate of 1.54% per year for a period of five years and the municipalities with the lowest TE scores should increase their prices at a rate of 2.64% per year for a period of five years. There is a possibility that all the municipalities can collectively increase their existing tariff prices by a mean value of 2.4% per year for a period of five years. The analysis should be carried out for each five-year period, and the tariffs should be revised accordingly in the presence of the regulator. If this continues, in the long run most of the municipalities having efficiency scores less than one should stand a reasonable chance to catch up significantly on the way towards the efficient frontier by following best practices, leading to effective services and better management policies, reflecting in satisfied consumers. Once the municipality manages to generate revenues that can cover cost expenditures, and tend towards gaining profits, the X-factor tends to decrease over the next regulatory period due to improvements in productivity growth. However, in reality, this is difficult to achieve in India, as water and sanitation, being essential services touching the lives of the common man, often invite political intervention, either to keep prices as low as possible irrespective of the sector's financial needs or in the form of meddling with personnel management and recruitment policies. Therefore, any future regulator would also need to focus on protecting the service providers from politically driven unexpected changes in operating conditions that might be imposed in defiance of long-term sector development objectives.

Table 5 shows the annual potential savings possible in operational expenditure for the financial period 2010. The results indicate that the maximum saving possible in the sector is Rs. 3,997 million, which is nearly 22.57% of the operational expenditure if the municipalities were to adopt appropriate polices and management tools. The operational expenditure of the utilities can thus be minimized theoretically even with the existing level of input resources by replicating the management policies of the best practices.

Table 5

Annual potential savings for the year 2010

S. no. Name of the utility Potential savings 2010
 
City/District State TE-CRS Operational expenditure (Rs. Millions) 
Dharmavaram Anantapur Andhra Pradesh 0.25 18.454 
Vijayawada Krishna Andhra Pradesh 0.07 534.393 
Madanapalle Chittor Andhra Pradesh 0.33 29.529 
Srikakulam Srikakulam Andhra Pradesh 0.12 30.263 
Hyderabad Hyderabad Telangana 0.34 902.194 
Rajahmundry East Godavari Andhra Pradesh 0.07 261.366 
Rajkot Rajkot Gujarat 1.00 0.00 
Surat Surat Gujarat 0.54 175.285 
Bellary Bellary Karnataka 0.16 139.374 
10 Davanagere-Harihara Davanagere Karnataka 0.07 272.653 
11 Kolar Kolar Karnataka 0.20 19.044 
12 Gadag- Betigeri Gadag Karnataka 0.72 3.502 
13 Dewas Dewas Madhya Pradesh 0.07 174.991 
14 Aurangabad Aurangabad Maharashtra 0.30 148.219 
15 Nashik Nashik Maharashtra 0.41 138.682 
16 Amritsar Amritsar Punjab 0.56 61.436 
17 Pathankot Pathankot Punjab 0.17 121.712 
18 Ghaziabad Ghaziabad Uttar Pradesh 0.16 392.353 
19 Kanpur Kanpur Uttar Pradesh 1.00 0.00 
20 Varanasi Varanasi Uttar Pradesh 0.17 388.481 
21 Gwalior Gwalior Madhya Pradesh 0.21 185.680 
 Total savings      3997.611 
 Mean    1903.62 
 SD    2616.48 
 Max    9021.94 
 Min    0.00 
 % Savings    22.57% 
S. no. Name of the utility Potential savings 2010
 
City/District State TE-CRS Operational expenditure (Rs. Millions) 
Dharmavaram Anantapur Andhra Pradesh 0.25 18.454 
Vijayawada Krishna Andhra Pradesh 0.07 534.393 
Madanapalle Chittor Andhra Pradesh 0.33 29.529 
Srikakulam Srikakulam Andhra Pradesh 0.12 30.263 
Hyderabad Hyderabad Telangana 0.34 902.194 
Rajahmundry East Godavari Andhra Pradesh 0.07 261.366 
Rajkot Rajkot Gujarat 1.00 0.00 
Surat Surat Gujarat 0.54 175.285 
Bellary Bellary Karnataka 0.16 139.374 
10 Davanagere-Harihara Davanagere Karnataka 0.07 272.653 
11 Kolar Kolar Karnataka 0.20 19.044 
12 Gadag- Betigeri Gadag Karnataka 0.72 3.502 
13 Dewas Dewas Madhya Pradesh 0.07 174.991 
14 Aurangabad Aurangabad Maharashtra 0.30 148.219 
15 Nashik Nashik Maharashtra 0.41 138.682 
16 Amritsar Amritsar Punjab 0.56 61.436 
17 Pathankot Pathankot Punjab 0.17 121.712 
18 Ghaziabad Ghaziabad Uttar Pradesh 0.16 392.353 
19 Kanpur Kanpur Uttar Pradesh 1.00 0.00 
20 Varanasi Varanasi Uttar Pradesh 0.17 388.481 
21 Gwalior Gwalior Madhya Pradesh 0.21 185.680 
 Total savings      3997.611 
 Mean    1903.62 
 SD    2616.48 
 Max    9021.94 
 Min    0.00 
 % Savings    22.57% 

CONCLUSIONS

This study provides measures of performances and productivities for 21 urban municipalities in India between the financial years 2005 and 2010. At present, the Indian water supply sector is not regulated by an independent regulator, and hence the illustrated framework may comprise a first step in the direction of bringing regulatory reforms in a sector which otherwise remains an essential monopoly of government utilities having large operational inefficiencies, as revealed by the results.

The DEA results showed that water supply services exhibit marked inefficiencies, and the relative performance gaps amongst the utilities is rather large, indicative of the fact that poorly performing municipalities need to adopt management and governance models similar to the best practices to improve their performance.

Given the model inadequacies, the results may be notional. Yet, large relative inefficiencies need to be addressed for effecting internal cost savings and for making the sector efficient, and the data over the period 2005–2010 indicate that this has not been done and relative efficiency gaps remain large over time. Hence, the actual scope for enhancing efficiencies in the Indian water supply services may be quite substantial and, if realized, the resultant internal cost savings may be used to expand services in rapidly expanding urban areas and to bring improvements in operation and maintenance of existing water services, resulting in incremental increase in efficiencies and savings. The current results already demonstrate the possibility of saving nearly 22.6% of the operational expenditure if the municipalities were to adopt appropriate polices and management tools of the existing best practices.

The results also demonstrate that most of the municipalities have a Malmquist TFP index of less than one, which means that the productivity of these municipalities decreases with time. In fact the majority (17 out of 21) of the municipalities showed performances decreasing over time, indicating negative productivity growth. This has implications for possible future privatization or for induction of competition in the monopolistic sector, as water services need lump investments, and no private company would be forthcoming to sink its investments in utilities that are not just inefficient, but also have been having negative productivity growth.

The X-factors calculated in the study, if implemented, may help the municipalities with the lowest efficiency scores to catch up nearly 10% of the way towards the frontier over the next five-year period. This may help municipalities to improve their annual productivity growth in accordance with the changes made in their input levels and generate revenues that cover cost expenditure incurred by the government, ultimately leading to financial sustainability of the services and reduction of the burden on administrators as well as customers.

It may be noted that in the Indian context, raising tariffs is often a sensitive issue intricately linked to local politics as the tariffs are decided by ULBs that comprise elected politicians who have vested interests in keeping the tariffs low. It is no wonder that raising tariffs to recover the cost of water supplies is not only difficult, but even the recovery of operating expenses remains a challenge, resulting in reliance on subsidies, year after year financial losses, and sector inefficiencies. Therefore, from a policy perspective, any future sector regulator would have to focus on limiting the financial role of government and bring about policies that foster a financial strategy that ensures sustainability of water supplies in urban areas without political intervention.

This study is limited by the scarcity of data available in India, a trend prevalent across most developing countries. The absence of a sector regulator implies that utilities do not bother to collect and record data or to make it public, thereby limiting analysis and severely affecting policy decisions needed to improve sector efficiencies. Also, lack of appropriate water sector price deflators for use in the TFP calculations is an additional concern. In most cases, proxies for the input quantities are used for analysis in the absence of data. Similarly, data breakup on water supply input variables like labor, electricity, chemical costs and water supply assets like pipes, pumps, construction services is not available, thereby limiting analysis and constraining and disabling the initiation of appropriate policy measures and sector reforms.

ACKNOWLEDGEMENTS

The authors are truly grateful to the three anonymous referees whose comments have contributed immensely to this paper by way of improved presentation and valuable insights.

REFERENCES

REFERENCES
ADB
2007
2007 Benchmarking and Data Book of Water Utilities in India
.
Asian Development Bank
,
Manila, The Philippines
.
Aida
K.
Cooper
W. W.
Pastor
J. T.
Sueyoshi
T.
1998
Evaluating water supply services in Japan with RAM: a range- adjusted measure of efficiency
.
Omega, The International Journal of Management Science
26
(
2
),
207
232
.
Banker
R. D.
Charnes
A.
Cooper
W. W.
1984
Some models for estimating technical and scale inefficiencies in data envelopment analysis
.
Management Science
30
(
9
),
1078
1092
.
Bruggink
T. H.
1982
Public versus regulated private enterprise in the municipal water industry: a comparison of operating costs
.
The Quarterly Review of Economics and Business: Journal of the Midwest Economics Association
22
(
1
),
111
125
.
Byrnes
P.
Grosskopf
S.
Hayes
K.
1986
Efficiency and ownership: further evidence
.
Review of Economics and Statistics
68
(
2
),
337
341
.
Charnes
A.
Cooper
W.
Rhodes
E.
1978
Measuring the efficiency of decision making units
.
European Journal of Operational Research
2
(
6
),
429
444
.
Coelli
T.
Walding
S.
2005
Performance Measurement in the Australian Water Supply Industry
.
School of Economics; University of Queensland
,
Brisbane
,
Australia
.
Coelli
T. J.
Rao
D. S. P.
O'Donnell
C. J.
Battese
G. E.
2005
An Introduction to Efficiency and Productivity Analysis
,
2nd edn
.
Springer
,
New York, NY
.
CPCB
2010
Status of Water Supply Waste Water Generation and Treatment in Class I Cities and Class II Towns of India
.
Central Pollution Control Board, Ministry of Environment and forests
,
Series: CUPS/70/2009-10, Government of India
.
CPHEEO
2005
Status of Water Supply, Sanitation and Solid Waste Management in Urban Areas
.
Central Public Health and Environmental Engineering Organisation, Ministry of Urban Development
,
Research study series No. 88, Government of India
.
Fare
R.
Grooskopf
S.
Norris
M.
Zang
Z.
1994
Productivity growth, technical progress, and efficiency changes in industrialised countries
.
The American Economic Review
84
(
1
),
66
83
.
Fare
R.
Grosskopf
S.
Roos
P.
1998
Malmquist Productivity Indexes: A Survey of Theory and Practice
.
Kluwer Academic Publishers, Springer
,
Dordrecht, The Netherlands
, pp.
127
190
.
Feigenbaum
S.
Teeples
R.
1983
Public versus private water delivery: a hedonic cost approach
.
Review of Economics and Statistics
65
(
4
),
672
678
.
Fox
K. J.
(ed.)
2002
Efficiency in the Public Sector
.
Springer
,
New York, NY, USA
, pp.
173
174
Gupta
S.
Kumar
S.
Sarangi
G. K.
2006
Measuring the Performance of Water Services Providers in Urban India: Implications for Managing Water Utilities
.
National Institute of Urban Affairs (NIUA)
,
WP 06-08
,
New Delhi
.
HSMI
2003
Public Private Partnership in Urban Water Supply and Sanitation Sector in India
.
Human Settlement Management Institute
,
New Delhi, India
.
Kulshrestha
M.
2009
Efficiency evaluation of urban water supply utilities in India
. In:
Proceedings of the 8th International Conference on Applied Infrastructure Research, German Institute for Economic Research, 9–10 October 2009. Workgroup for Infrastructure Policy (WIP), Berlin University of Technology, Berlin, Germany
.
Kulshrestha
M.
Mittal
A. K.
2005
Water and sanitation in South Asia in the context of millennium development goals
.
South Asia Economic Journal
6
,
99
115
.
Kulshrestha
M.
Vishwakarma
A.
Phadnis
S. S.
Thakur
T.
2012
Sustainability issues in the water supply sector of urban India: implications for developing countries
.
International Journal of Environmental Engineering
4
(
1/2
),
105
136
.
Lambert
D. K.
Dichey
D.
Raffie
K.
1993
Ownership and sources of inefficiency in the provision of water services
.
Water Resources Research
29
(
6
),
1573
1578
.
Marques
R.
Simões
P.
Pires
J.
2011
Performance benchmarking in utility regulation: the worldwide experience
.
Polish Journal of Environmental Studies
20
(
1
),
125
132
.
Marques
R.
Berg
S.
Yane
S.
2014
Nonparametric benchmarking of Japanese water utilities: institutional and environmental factors affecting efficiency
.
Journal of Water Research Planning and Management
140
(
5
),
562
571
.
MoUD & PA
2004
India: Urban water and sanitation services – guidelines for sector reform and successful public-private partnerships
.
Ministry of Urban Development and Poverty Alleviation, Government of India
,
New Delhi, India
.
Planning Commission India
2007
Report of the Steering Committee on Water Resources for Eleventh Five Year Plan (2007–2012)
,
Planning Commission, Government of India
,
New Delhi, India
.
Sawkins
J. S.
Accam
B.
1994
Comparative Efficiency Measurement in the Scottish Water Industry: An Application of Data Envelopment Analysis
.
Discussion paper
,
University of Aberdeen
,
UK
.
Teeples
R.
Glyer
D.
1987
Cost of water delivery systems: specification and ownership effects
.
The Review of Economics and Statistics
69
(
3
),
399
408
.
Thanassoulis
E.
2000a
The use of data envelopment analysis in the regulation of UK water utilities: water distribution
.
European Journal of Operations Research
126
(
2
),
436
453
.
Thanassoulis
E.
2000b
DEA and its use in the regulation of water companies
.
European Journal of Operational Research
127
(
1
),
1
13
.
Tremolet
S.
2006
Adapting Regulation to the Needs of the Poor: Experience in 4 East African Countries
.
Building Partnerships for Development (BPDWS)
,
BPD Research Series
,
UK
.
Tremolet
S.
Halpern
J.
2006
Regulation of Water and Sanitation Services: Getting Better Service to Poor People
.
OBA working paper series, paper No. 8
.
World Bank
,
Washington, DC, USA
.
Vishwakarma
A.
Kulshrestha
M.
2010
Cost efficiency evaluation and performance assesment of urban water supply utilities in the state of Madhya Pradesh
.
International Journal of Civil and Structural Engineering
1
(
2
),
234
244
.
WSP
2010
The Karnataka Urban Water Sector Improvement Project: 24 × 7 Water Supply is Achievable
.
World Bank Water and Sanitation Program
,
India
.
Zardkoohi
A.
Crain
W. M.
1978
A test of the property-rights theory of the firm: water utilities in the United States
.
The Journal of Law and Economics
21
(
2
),
395
408
.