Water PDF

STANFORD CENTER FOR INTERNATIONAL DEVELOPMENT
Working Paper No. 224

Household Water Delivery Options in Urban and Rural India
by
David McKenzie * and Isha Ray **
Updated December 2005
Stanford University
579 Serra Mall @ Galvez, Landau Economics Building, Room 153
Stanford, CA 94305-6015
Development Research Group, The World Bank

Energy and Resources Group, University of California at Berkeley
**
McKenzie & Ray; Drinking water options for India

Revised 8/29/05
HOUSEHOLD WATER DELIVERY OPTIONS IN URBAN AND

RURAL INDIA*
David McKenzie1
Isha Ray2
Introduction
In over 50 years of political independence and economic development, India has not been
able to ensure the most basic of human needs safe drinking water for all its citizens. In
April 2002, then Prime Minister Vajpayee confronted this problem openly and clearly.3
Scarcity of water is compounded by its unequal, irrational, and unjust distribution in both
rural and urban communities, he said. Therefore, the situation is forcing us to recognize
water security as an overriding national objective Accordingly, the 2002 National Water
Policy of the Government of India (GoI) states that provision for drinking water should be
a primary consideration in water resource development projects and that drinking water
needs of human beings and animals should be the first charge on any available water (GoI,
2002, Sections 6.1, 8).
In a further acknowledgment of the lack of access to water in India, in December of 2002
Mr Vajpayee launched the Swajaldhara programme, with the aim of providing drinking water
to every village by the end of 2004.4 This is an ambitious World Bank aided programme that
marks a departure from traditional government water provision. Instead, it works on the
90/10 financing model -- if communities come up with 10% of the cost of provision, the
government will provide the other 90%. This provision is in keeping with the spirit of the
73rd Amendment to the Indian Constitution, which turned over the responsibility for local
water supplies to local level governments (Gram and Zilla Pancahayats).
These recent and potentially far-reaching policy changes frame our paper on drinking water
options for urban and rural India. Given the primacy of drinking water as a national
objective, and the policy of decentralization through community ownership, private sector
participation and devolution to local governments, we ask: How can India alleviate its householdlevel drinking water deprivation, in the near-to-medium term, and in cost-effective ways? We take the goal
of reform in the drinking water sector to be better public health, in particular a reduction in
We thank our discussants Mary Shirley and Scott Wallsten, along with participants at the 5th Stanford
Conference on Indian Economic Development, for their comments. Niny Khor provided excellent research
assistance.
1 Economics Department, Stanford University. Email: [email protected]
2 Energy and Resources Group, University of California at Berkeley. Email: [email protected]
3 Prime Minister Shri Atal Behari Vajpayees speech to the Fifth Meeting of the Water Resources Council, New
Delhi, April 1, 2002.
4 Indian Express, December 6 2002. PM bday gift: water to drink.
*

Revised 8/29/05
water-borne diseases. Therefore the primary focus of this paper is on universal access to a
minimum daily quantity of safe water for drinking, cooking and basic hygiene. The threshold
for minimum varies from 20 litres per capita per day (lpcd) (UNICEF 1995); to 40 lpcd
(GoI norm) to a more generous 50 lpcd (Gleick 1998). In this paper, we work with the
Government of India norm.
Our second focus is the cost-effectiveness of different modes of water delivery and access.
In general, national goals should be efficiently rather than inefficiently achieved and the
parlous financial state of rural and urban water agencies prevents them from extending
access to those who are unserved. There are also many competing claims on the national
budget, therefore we have to accept the reality of limited outlays for water and sanitation.5
Cost-effectiveness and a measure of cost-recovery should allow more people to be served
with the same outlays.
Household- and community-level water options represent the downstream end of the
drinking water sector. We focus this paper on conditions and policies for drinking water
provision at the downstream end, although we recognize that at the upstream end and in the
long term, universal access to safe drinking water depends on overall water sector priorities
and policies. In particular, there are several important issues in the Indian context which we
cannot address in this paper for reasons of scope and space constraints. The first is sectoral
allocation of water between industry, domestic use and environmental protection. In 2000,
92% of Indias usable water went to agriculture, mainly in the form of irrigation
(Development Alternatives 2001). While inter-sectoral water allocation decisions will
determine the long term supply of water, we argue that, in the near-to-medium term, there is
plenty of scope for reform even using the share of water that domestic use currently
receives. The household options discussed in our paper will remain important even if the
share of water for domestic uses goes up. A second issue is that in many cities and villages,
aquifer overdraft and poor groundwater quality pose great threats to a safe water supply. We
discuss methods of water provision to cope with this problem, but do not address in detail
measures to deal with groundwater depletion from the supply-side. Thirdly, we do not
discuss large-scale water projects such as dams and river-linking schemes, which are not
primarily meant to solve drinking water problems. Lastly, we recognize the strong and welldocumented links between improved sewage and sanitation services, safe drinking water and
public health. Sewage is a major source of water pollution and the most effective way to
keep water clean is to protect it from human and animal wastes.6 In this paper, we cannot do
justice to the many ways in which sanitation, sewage and wastewater treatment can be
provided, so we focus solely on the drinking water aspect. However, many of the
institutional barriers and political economy issues discussed here apply in equal measure to
the provision of sanitation and sewage services.
This paper is divided into three parts. In Part I we provide an overview of the state of access
to drinking water in urban and rural India. In Part II we discuss urban water delivery
options, with examples of ongoing attempts to reform the urban water sector. In Part III we
analyze decentralized technological institutional options for extending access to water in
Of course the limited budget for water and sanitation does not reflect merely Indias low per capita income,
but also its national priorities.
6 We thank Irena Amundson for raising the point on the connection between sewage and water.
5

Revised 8/29/05
peri-urban and rural areas. Finally, we draw on the literature and on our own field
experiences to recommend fruitful directions for data collection, research and policy reform.
PART I: ACCESS TO DRINKING WATER
I.1 Nature and extent of access to water
To examine patterns in household water delivery across different geographic and
socioeconomic dimensions we employ data from the National Family Health Survey
(NFHS), taken in 1992-93 and 1998-99.7 This is Indias version of the Demographic and
Health (DHS) surveys taken worldwide and is a nationally representative survey directed
towards ever married women of child-bearing age. Approximately 90,000 households were
surveyed in each wave. In addition to data on household characteristics and on demographic
outcomes, the survey asks several questions of direct interest to household water delivery
including the source of the households drinking water, the time taken to get to water
sources for households who do not receive water on their premises, and the method used by
the household to purify their water.
Table 1 provides a breakdown of the various water delivery mechanisms for all of India, and
separately for urban and rural areas. Urban areas are broken down further into large cities,
and small cities/towns. Looking first across all of India, one notes substantial heterogeneity
in water delivery, with no dominant mechanism. In 1998-99, only 21 percent of all
households have piped water, while 18 percent use public taps, 15 percent private
handpumps, 24 percent public handpumps, and 19 percent rely on wells. There are large
differences between urban and rural areas in the source of drinking water. Piped water
supplies 69 percent of households in large cities, 45 percent in smaller cities and towns, and
only 9 percent of rural households. Public taps are used more in smaller cities than either in
rural or in large cities. Handpumps are the predominant source of drinking water in rural
areas, with 47 percent of households in 1998-99 receiving water from either private or public
handpumps. Wells also supply a substantial number of rural households with their water.
Very few households use tanker trucks, rivers, streams, and the other such sources.
Comparing the 1992-93 data with the 1998-99 data allows some insight into areas where
there were most rapid changes in water provision during the 1990s. Piped water expanded
in the large cities, to cover 12 percent more of the population. There was much less
expansion in the rural areas and smaller cities, and so national coverage only increased by 2.6
percentage points. This expansion in piped water access is seen to result in less large city
urban residents relying on public wells, public handpumps, and public taps. In contrast,
public taps expanded in rural areas and smaller cities. Use of public wells is seen to have
fallen across both rural and urban areas, with the decline largest in rural areas. Public well
usage fell 9 percentage points among rural households, with increased handpump, public tap,
and piped water use accounting for this change.
The World Health Organization (WHO) defines access to water supply services as the
availability of at least 20 litres per person per day from an improved source within one
Data from the ongoing Measure DHS+ benchmark survey 1998-2003 are not yet available. Available surveys
can be downloaded from http://www.measuredhs.com/.
7

Revised 8/29/05
kilometre of the users dwelling.8 Improved sources are those likely to provide safe water
such as household connections, public standpipes, protected dug wells, rainwater collection,
boreholes, and protected springs. Not improved sources include unprotected wells and
springs, vendor provided water and tanker truck water. The foot of Table 1 shows the
proportion of Indian households receiving water from these improved sources. Note that
the survey does not provide any data on the quantity of water obtained from these sources,
so these numbers are likely to overestimate the proportion of households with improved
water. In 1998-99, 98 percent of households in large cities, 92 percent in small cities, and 75
percent in rural areas are estimated to receive water from improved sources. This represents
an improvement over the 1992-93 survey, especially in rural areas.9 Overall, 71 percent of
Indian households in 1992-93, compared to 80 percent in 1998-99, were estimated to receive
their drinking water from improved sources.
Data from the 1981 Census show that 75 percent of urban households and 27 percent of
rural households had access to improved sources of water, while the 1991 Census shows 81
percent of urban households and 56 percent of rural households with improved sources
(WHO-UNICEF 2001). The 94 percent of urban and 75 percent of rural households with
access to improved sources in the 1998-99 NFHS therefore represents a substantial
improvement in access to safer sources of water during the 1980s and 1990s.
There is considerable variation in the sources of drinking water across Indian states. Table 2
shows sources used in urban and rural areas by state. Figure 1 plots access to piped water
against per capita state domestic product. There is a strong positive relationship between
state incomes and access to piped water. As a result, access to piped water in urban areas
ranges from 18 percent in Kerala, 21 percent in Orissa and 26 percent in Bihar to over 75
percent access in urban areas in New Delhi, Sikkim, Gujarat, Jammu and Rajasthan. There is
a great deal of cross-state variation in use of other methods, with different mechanisms for
supply in different states. In urban areas, use of public taps ranges from only 2 percent in
Punjab to 43 percent in Andhra Pradesh; use of private handpumps ranges from less than 1
percent in Goa to almost 40 percent in Uttar Pradesh; and wells range from less than 1
percent in several states to 60 percent in Kerala. Rural areas show just as much cross-region
variation, with rivers, springs and ponds, which supply drinking water to less than 3 percent
of rural households in the nation as a whole, supplying 25 percent or more of households in
Jammu, Manipur, Meghalaya, and Nagaland.
Access to piped water is therefore seen to be more prevalent in richer states. The NFHS
does not collect income data at the household level, but does ask households questions on
ownership of a number of durable assets. Filmer and Pritchett (2001) show that the first
principal component of these indicators provides reasonable estimates of wealth effects.
Since households in rural and urban areas may hold different portfolios of assets, we define
asset deciles separately for rural and urban households, as well as for all of India. Table 3
then reports our estimates of the percentage of households with access to piped water by
asset decile. One finds a strong relationship between asset wealth and access to water. Even
8See
the website of the WHO/UNICEF Joint Monitoring Programme for Water Supply and Sanitation
http://www.wssinfo.org/en/welcome.html.
9 The 1992-93 survey does not separate wells into open and covered, and so our calculations for 1998-99
assume that the ratio of covered wells to total wells was the same as in 1998-99.

Revised 8/29/05
within large cities, in which 70 percent of households have piped water, we see considerable
variation across asset levels. Only 26 percent of households in the bottom decile and 42
percent in the second decile have piped water, compared to 87-90 percent in the top three
deciles. Also notable is the levelling off of access at the top of the distribution; even among
those with the highest level of assets 12 percent of households do not receive piped water.
Households that do not receive water supply at their housing structure encounter several
costs in obtaining water from other sources. In addition to the direct costs of storage
containers, households spend time travelling to the water source, and then potentially
waiting in line to use community supplies. Table 4 details the breakdown of time taken to get
to the alternative water supply for households without water on their premises. While
roughly 50 percent of such households spend 10 minutes or less getting to the water supply,
21 percent of these urban households and 26 percent of these rural households spend 20
minutes or more getting to the water source. Households may then pay financial costs for
transporting water from standpipes to their homes.
I.2 Quality of drinking water
The WHO Guidelines for Drinking-water Quality (WHO 1993, 1997, 1998) assess the health risks
posed by contaminants in drinking water. The guideline values for biological and chemical
pathogens in drinking water safety are not mandatory the WHO recognises that on-paper
mandatory limits are not useful for most developing countries. Rather, the guidelines are
intended to develop risk management strategies, and national or regional standards in the
context of environmental, social and economic conditions. The WHOs primary health
requirement is a sufficient water supply, aside from quality, which the Government of India
takes to mean 40 lpcd.
The second requirement is that the water be microbiologically safe, since these waterborne
pathogens are the leading cause of those diseases that kill or seriously debilitate the affected
(Esrey et al, 1991). Children under 5 years of age are especially vulnerable to acute diarrhoeal
episodes. In most developing countries, India included, the primary contaminant of surface
and ground waters is human and animal waste.10 The WHO guidelines suggest that E. coli
(the indicator organism for bacterial contamination) should not be detectable in a 100-ml
sample of water. In practice, with fewer than 10 coliforms in a 100ml sample, the water is
considered to be of moderately good quality. The Government of India accepts these
guidelines, but has generally been unable to ensure that they are met. Water-borne diseases
from faecal contamination are one of the biggest public health risks in India -- diarrhoeal
diseases are the largest killers of children. Repeated bouts of diarrhoea also cause stunting
through malnourishment children with diarrhoea are unable to absorb the nutrients from
any food they eat. It has been argued that India loses 90 million days a year due to
waterborne diseases, costing Rs 6 billion in production losses and treatment (Chaudhuri
1998). Another study estimated that each year India lost 30.5 million disability-adjusted life
years because of poor water quality, sanitation and hygiene (WSP 1999a).
For this reason, most water and sanitation practitioners think that it is a fundamental error to treat water,
sanitation and hygiene separately however this separation is still the norm in the literature and in policy (Jolly
2003).
10

Revised 8/29/05
The third requirement of the WHO is that the water be chemically safe, although there are
different levels of risk associated with different chemical parameters. Though chemical
residues in water are not as widespread as biological pathogens, chemical contamination
through agricultural runoff and unregulated industrial pollution is a rapidly growing problem.
The most widespread and significant naturally occurring waterborne toxics are arsenic and
fluoride (guideline maximum concentrations of 10 g/l and 1.5 mg/l respectively). Fluoride
problems exist in 150 districts of 17 states in the country 11 -- with Orissa and Rajasthan the
most severely affected. Excessive fluoride in drinking water causes fluorosis, manifested in
weak bones, weak teeth and anaemia. The calamity of arsenic a poison and a carcinogen
in the groundwater of the Gangetic delta, affecting 35 70 million people in West Bengal,
Bihar and Bangladesh, is by now well known (WHO 2001). It has been hypothesized that
the naturally-occurring arsenic leached into the water from underground geological strata
because over-pumping lowered the local water tables. There are no national standards for
chemical concentrations in water, a point which caused much public outrage when pesticide
residues were found in bottled sodas and waters (CSE 2003). However, the central
government intends to introduce an ordinance that classifies drinking water as a food, with
its own norms for biological, chemical and other additives (Rastogi 2004).
Standards for drinking water that are actually enforced could have enormous positive
impacts on public health, but for this to occur, the procedures for water testing and data
sharing have to be made regular, standardized and public. As of today, multiple
organizations including Public Health Engineering Departments, Central and State
Groundwater Pollution Boards, the Central Water Commission, and other government and
research groups are all involved in assessing water quality. However these groups rarely
coordinate their activities, and overlapping responsibilities result in inefficiency, negligence
and delays (Sridhar 2003). Additionally, published information on much of this water testing
is not readily available. GoI (2001) reports that much of the information, if available at all,
can only be obtained upon request at the regional office level (and informal reports from
water researchers confirm that some state and regional offices jealously guard these data). As
a result, standardized comparison of water quality across locations or over time is not
possible, and baseline information is not reliable.
PART II: URBAN WATER SUPPLY OPTIONS
Authority over urban water services lies with each state, with some decentralization to the
municipal level. The institutional arrangements vary across states. Some state water supply
boards, such as Karnatakas, are responsible for operation, maintenance, and capital works,
while in other states operation and management is run by municipal bodies. In some of the
larger cities, such as Delhi, Chennai, Bangalore and Hyderabad, separate municipal
corporations have been created. These are responsible solely for water and sewerage
facilities, in contrast to municipal bodies which also carry out other local government
activities. The World Bank (2000) reports that although some service providers are semiautonomous in theory, there is considerable political interference in operations, managerial
decision-making and tariff-setting.
II.1 The current state of urban water provision
11
Source: Department of Drinking Water Supply, Ministry of Rural Development, http://ddws.nic.in

Revised 8/29/05
Municipal provision of water is the status quo in most parts of India. While only half of all
urban households have a piped water connection, even those with a connection generally do
not receive a regular supply of good quality water. The municipal water supply in most
Indian cities is only available for a few hours per day, pressure is irregular, and the water is of
questionable quality.
Irregularity
Table 5 summarizes some key indicators of irregularity, wastage, and poor performance of
water suppliers in the largest Indian cities. For comparison purposes, these same indicators
are also provided for Lahore, Kathmandu, Bangkok, Beijing, and an average of 50 cities
surveyed by the Asian Development Bank in 1997 (ADB 1997). No major Indian city has a
24 hour supply of water, with 4 to 5 hours of supply per day being the norm. This compares
to the Asian-Pacific average of 19 hours per day supply, with both Bangkok and Beijing
having 24 hour coverage.
These averages conceal a great deal of heterogeneity within cities in terms of the regularity of
supply. In a 1995 survey of Delhi households with in-house connections, Zrah (2000) finds
that only 40 percent had 24 hour supply of water, while more than 25 percent had less than 4
hours a day of service. Moreover, there were large differences in supply from floor to floor
of a house or apartment. Ground floor apartments averaged 16 hours per day, compared to
8 hours on the first floor and 4-5 hours on second and third floors. Fewer than 20 % of
households surveyed received water at uniform pressure during different days of the week.
McIntosh (2003) refutes the assertion that there is not enough water for 24-hour supply.
Much of the available water is wasted, both through leakages, and a lack of incentive to
conserve water due to low tariffs. He gives the example of Mal which has achieved 24-hour
piped supply, along with a high tariff, and strict metering, billing and collection. He notes
that consumers without 24-hour supply tend to use more water than those with 24-hour
supply. Because they can never be sure when they will next be served, consumers store
water, which they then throw away to replace with fresh supplies each day. Additionally, in
India, household use accounts for only a small fraction of overall water usage. Agriculture is
the prime user, with 92 percent of utilizable water going to this sector (Development
Alternatives 2001).
Urban Water Quality
An unreliable water supply, characterized by intermittent water supply, insufficient pressure
and unpredictable service, imposes both financial and health costs on Indian households.
Based on a survey conducted in Delhi in 1995, Zrah (2000) estimated that each household
on average spent around 2000 Rupees annually in coping with unreliable supply of water,
which is 5.5 times as much as they were paying their municipality for their annual water
consumption. Most households with in-house connections were found to have undertaken
long-term investments in the form of acquiring storage devices such as water tanks or
buckets, or acquiring handpumps or sinking tubewells. Households with water tanks would
install electric booster pumps on the main water line itself and pump water directly to water
tanks. This increases risks of contamination of the general water supply and reduces the
pressure in the network for other users, leading them also to install motors on the main line.
In addition to the contamination that can occur through household use of booster pumps,
the vacuum conditions created when supply stops can draw in foul water (McIntosh 2003).

Revised 8/29/05
Monitoring of water quality is haphazard. While municipal boards claim to conduct regular
tests of water supply, the results of these tests are generally not made public.12 The
information that does exist points to generally poor quality of water in a number of cases.
The most reliable testing information we have encountered comes from the Sukthankar
Committee (2001) report to the Government of Maharashtra, which reported results from
136,000 daily tests of bacteriological contamination carried out on water samples from the
various municipal corporations in Maharashtra over the course of 1999. Ten percent of
samples were found to be contaminated, with 14 percent of samples from Mumbai being
contaminated. The Committee noted that no testing is done for chemical or physical
contamination, so it is likely that the actual level of poor quality samples exceeded 10
percent. Water quality in Mumbai remains poor, with a study by the Society for Clean
Environment (SoCleen) in July 2003 finding that a significant percentage of water in many
parts of the city was not potable and contaminated with excessive bacterial pollution.13
Thermotolerant coliform organisms, an important indicator of water safety, were several
hundred times higher than the norm. Water quality was poor across socioeconomic groups,
with 70 percent of samples from middle-class housing societies found to be not potable.
Water monitoring conducted in January-March 2003 by Clean India in 28 cities found that
ground water in most areas exceeded permissible limits in terms of fluoride, ammonia and
hardness. Municipal water supply in some cities also contained high numbers of
contaminants. For example, in Jaipur in the state of Rajasthan, permissible limits of nitrates,
ammonia, coliform bacteria and hardness were exceeded in more than 50 percent of samples
from the municipal water supply.14 A 2003 survey of 1000 locations in Calcutta found that 87
percent of water reservoirs serving residential buildings and 63 percent of taps had high
levels of faecal contamination.15 Even bottled water is not completely safe. A 2003 study by
the Centre for Science and Environment (CSE) in Delhi found that most of the most
popular brands of bottled water had high levels of pesticides, some of which were over 100
times EEC limits (CSE 2003).
The demand for and cost of improved quality
There are a range of mechanisms households can use to improve the quality of the water
they receive before drinking it. These range from low-grade technologies such as straining
with a cloth, using chlorine and safe storage vessels, to relatively more sophisticated
technologies like electronic filters. Based on their own cost survey in Delhi, Jalan,
Somanathan and Chaudhuri (2003) estimate the annual average cost per person of different
purification methods. The cheapest method is straining with a cloth, which provides some
protection against cholera, but does not filter out most free-floating bacteria. Alum tablets
are estimated to cost Rs. 73 per person per year, and are effective at reducing turbidity and in
reducing bacteria, but do not remove all pathogens of concern. The two most effective
methods are also the most expensive. Electronic filters which first filter particulates and then
irradiate the water with ultraviolet light are very effective provided they are properly
For example, the BWSSB in Bangalore now collects 1200 samples a month from several parts of Bangalore
City and tests them in a Central Lab to measure water quality (see http://www.bwssb.org/BSWW_news.cfm).
However, results from these tests are not readily accessible.
13 See Sridhar (2003) for a report on this study.
14 See results reported in Clean India July 2003 Newsletter Watch that water you are drinking! Available
online at http://www.devalt.org/newsletter/jul03/ [accessed January 2004].
15 See Warning over Calcutta water quality, BBC News UK edition, 29 August, 2003.
12

Revised 8/29/05
maintained, and are estimated to cost Rs. 791 per person per year. This is still about half the
cost of boiling, which is estimated to cost Rs. 1635 per person per year.
Using the 1999 NFHS, Jalan, Somanathan and Chaudhuri (2003) report that 47 percent of
households do not use any purification method, with 32 percent of the top wealth quartile
also not purifying their water. While the cost of electronic filters and boiling may preclude
their use for the poor, cost should not be a factor in determining whether to strain with a
cloth, while the alum tablets and ordinary filters are also relatively inexpensive. Lack of use
of these methods must therefore arise from household perceptions that there is no need to
filter, or not large benefits from doing so. Some households receiving piped water may
believe it to be of sufficient quality not to require filtering. However, Jalan and Ravallion
(2003) find that while provision of piped water does reduce the incidence of diarrhoea
among children in rural households in India, health gains are lower for children with less
well-educated women in the household. One explanation might be that more-educated
households enjoy better quality piped water, while another is that even piped water needs to
be treated or purified, and stored safely, due to irregular supply.
Jalan, Somanathan and Chaudhuri (2003) provide support for the latter explanation.
Controlling for wealth, they find that higher female education and more exposure to the
mass media, such as newspapers, increases household willingness to pay for purification.
They conclude that a lack of awareness of the adverse health effects of poor water quality is
a significant barrier to the adoption of different home purification methods. As a result,
public awareness campaigns to educate the population about the health consequences of
poor water quality are needed to accompany policies designed to provide and promote the
use of in-home water purification methods.
II.2 Inefficiencies in the supply of water
One of the prime reasons for the poor quality of the water supply is inefficiencies in the
water delivery suppliers. Prime symptoms of this are the large amount of unaccounted for
water, overstaffing, and low cost recovery.
Unaccounted For Water
A standard indicator of inefficiency and wastage is the percentage of water produced which
does not reach water board customers. Unaccounted for water results from both leakages
and illegal connections. In addition to the financial costs to the water utility, high levels of
unaccounted for water are also a major reason for intermittency in the supply of water, since
leaks and illegal connections lower water pressure in the distribution system (McIntosh
2003). Table 6 shows that unaccounted for water accounts for 25-40 percent of water
produced by utilities in the main urban areas in India. While this is no higher than the AsianPacific average, McIntosh (2003) claims that the large number of obvious leaks means there
is still substantial scope for improvement. He cites the example of Singapore, which has
reduced unaccounted for water to 6 percent as an example of what can be achieved by
metering, leak surveillance, proper billing procedures and strong management.
Overstaffing
A second indicator of inefficiency is high staffing levels. A good utility will have two staff for
every 1,000 connections (McIntosh 2003). Such levels have been attained by water utilities in
Taipei, Kuala Lumpur, Singapore and Seoul, and can be considered as best practice

Revised 8/29/05
10
numbers. Table 6 shows that the Asian-Pacific regional average is around 12 staff per 1000.
Hyderabad and Bangalore are around this level, but staffing levels are double this in Chennai
and Delhi, and higher still at 33 per 1000 in Mumbai. While staffing levels are high, the
average quality of workers in many utilities is low. Based on visits to different water utilities
in Maharashtra, the Sukthankar Committee (2001, p. 90) reported that most of the
operating staff was not qualified to work in water works installations. Many were unskilled
labourers who had worked on the construction of the works and stayed on to work as
operators. The Committee is even more damning of staff dealing with the water treatment
processes and operations, remarking that such staff were found to be hardly having even
the basic understanding of the unit process and unit operations which they were handling
(Sukthankar Committee, 2001, p. 90).
Low cost recovery
The consequence of overstaffing, underpricing, and high levels of unaccounted for water is
that most urban water utilities in India are unable to even cover operating and maintenance
costs out of revenues from tariffs, let alone provide for expansion and improvement of the
network. Table 6 shows that only Chennai has managed to cover operating costs, while
Bangalore and Mumbai come close. The situation is worse in Hyderabad and Delhi where
only 66-68 percent of costs were met, and severe in Kolkata, with only 15 percent of
operational costs being recovered. The Kolkata Municipal Corporation (KMC) itself reports
that in 2001 57 percent of water generated by the KMC did not fetch any revenue at all.16
II.3 Pricing of water
State Governments in India are responsible for choosing urban tariff structures. Municipal
boards and corporations can propose tariff structures, but these are subject to State
Government approval. The result is a wide variety in pricing practices. Raghupati and Foster
(2002) surveyed water charging practices in all 23 metropolitan areas (cities of over 1 million
population) and 277 smaller cities of populations 50,000 to 1 million. They find that most
cities operate a mixture of measured and unmeasured tariffs, due to relatively low coverage
of metering (see Section II.4). For unmeasured areas, a flat rate is the most common form of
tariff. Other options used include a ferrule-based tariff, which charges according to the
diameter of the connection, and water rates based on the value or size of the dwelling. Two
tariff structures are found for metered connections. The most predominant are uniform
volumetric charges, which charge a constant rate per Kilolitre of water. This system is found
in 77 percent of smaller cities and 58 percent of metropolitan areas. In the remaining 42
percent of metropolitan areas and 23 percent of smaller cities an increasing block tariff (IBT)
is used.
Under an increasing block tariff, a low rate is charged for the first few units of water each
month, and then higher amounts of use are charged at higher marginal rates. Boland and
Whittington (2000) outline the arguments made by supporters of Increasing Block Tariffs.
Public health externalities and the idea that water is a merit good are used to argue that all
households should have access to some baseline level of water at low rates. Since water is
assumed to be a normal good, increasing rates on higher amounts forces wealthy households
to subsidize poorer households, which is seen as desirable for equity reasons. Prices can then
be kept low for the poor, while the higher prices at increased levels of use can allow
16
See http://www.kolkatamunicipalcorporation.com/water.html

Revised 8/29/05
11
revenues to cover costs. Proponents also argue that charging higher rates for more use can
act to promote conservation and sustainable use.
Table 7 provides examples of the prevailing tariff structures in Hyderabad, Bangalore, Delhi
and Chennai, which are the cities for which we could obtain recent data. All four cities use
IBTs for metered customers. The size of the blocks or slabs can be seen to vary across cities,
with the initial monthly block being 10 KL in Delhi and Chennai, 15 KL in Bangalore, and
30 KL in Hyderabad. To compare prices across cities, Table 8 reports the effective price for
10-50 KL in each of the four cities. Delhi has the lowest rates by far for all quantities, with
the cost of 50KL only one-third of that in Hyderabad, and less than one-sixth of that in
Chennai. Among the other cities, Chennai has the lowest price for 10KL, but then the rates
for subsequent blocks increase more rapidly, so that larger amounts of water are more
expensive there.
Raghupati and Foster (2002) note that relatively little information is available as to the level
of tariffs needed for full coverage of operating and maintenance costs. They provide results
from several consultant studies which suggest operating and maintenance costs average rates
of Rs. 13 per KL for Chennai, Rs. 16 per KL for Bangalore, and Rs. 17 per KL for
Hyderabad. As outlined above, there is substantial inefficiency in terms of excess staffing
and wastage of water, which may reduce the tariffs needed in an efficient company.
However, incorporation of capital costs, which no public information is available for, would
raise the tariffs needed. Comparing these rates to those in Table 7, one sees that only the
metered tariffs above 50 KL in Bangalore, above 200 KL in Hyderabad and above 15 KL in
Chennai are likely to cover operating and maintenance costs. However, very few households
in Chennai are metered (see Table 6) so these higher rates do not cover many households
there. These results are not unique to this selection of cities; Raghupati and Foster (2002)
find that the typical price charged in their larger sample of cities is Rs. 1.5 per KL, or only
approximately 10 percent of operating and maintenance costs.
The main rationale used for subsidizing the initial block of water is to provide some lifeline
level of access to water. However, as Brocklehurst, Pandurangi and Ramanathan (2002) note,
the size of the first block in all the Indian cases exceeds what would generally be considered
a true lifeline level. They give the example of a family of five using 40 litres per capita per
day for 30 days, which would require 6 KL a month.17 This is the size of the lifeline block in
South Africa, whereas the blocks in India are two to five times this level. Brocklehurst,
Pandurangi and Ramanathan (2002) report the proportion of customers consuming in
different blocks in 2001 in Bangalore and Hyderabad. In Hyderabad they find that 70
percent of customers consume in the 0-15 KL/month range, 20 percent in the 15-25
KL/month range and only 10 percent consume 25 KL/month or more. In Bangalore, 66
percent of customers are in the 0-25 KL/month range, 28% in the 25-50 KL/month range,
and only 7 percent consumer more than 50 KL/month. Coupled with the low tariffs in the
first couple of blocks, the result in Bangalore is that 93 percent of the customers account for
less than 15 percent of revenues, and are paying less than 41 percent of the cost of providing
them with water.
Boland and Whittington (2000) report that international standards for basic water needs are in the range of
25-30 litres per capita per day, which is 4-5 KL/month for a household of five.
17

Revised 8/29/05
12
In practice then, the increasing block tariff has resulted in a large initial block, with the
majority of households consuming in the first couple of blocks. Boland and Whittington
(2000) find this to be a general phenomenon experienced in many places across the world
where IBTs are used. They remark that there is large political pressure for politicians to
increase the size of the first block, since all households with piped connections benefit
directly from this (whereas the costs of having to subsidize the underfunded public
corporation remain opaque). The result is that the IBT does a poor job of targeting subsidies
for water towards poor households. There are large numbers of non-poor households who
benefit from the large initial block, while poor households with larger families may actually
be pushed into the higher-priced block. Moreover, since poorer households are less likely to
have a metered connection in the first place, many poor households do not benefit from the
increasing block structure. Foster , Pattanayak and Prokopy (2003a) calculate that altogether
the state and federal governments in India spend US $1.1 billion or 0.5% of GDP in
subsidizing water, but find that 70 to 80 percent of these subsidies fail to reach the poor.
Moreover, the average amount of the subsidy received by the top decile of the population is
2-3 times as high as the average subsidy received by the bottom decile.
These low prices for the majority of consumers are one of the prime determinants of the low
cost recovery ratios seen in Table 6. The extremely low prices in Delhi make this a particular
problem, and there has been no change in Delhi water prices since 1998. The gap between
revenue and expenditure has been growing. The shortfall in 2003 was Rs 5 billion, on annual
revenue of less than Rs 2 billion. The Delhi Jal Board put forward a proposal to increase
rates in January 2004, with the new rates proposed between 2.5 and 6 Rupees per Kilolitre,
however this proposal was put aside by the state government due to election pressures.18
II.4 Scope for Pricing Reform
Average tariffs in India are clearly low relative to costs. A cross-region study by the Asian
Development Bank in 1997 found average rates in Calcutta and Delhi of 1-3 US cents per
KL, 6 cents per KL in Mumbai and 25 US cents per KL in Chennai. In comparison, rates
were 5 cents per KL in Beijing, 9 cents per KL in Dhaka and Karachi. 20 cents per KL in
Lahore, 23 cents per KL in Manila, 31 cents per KL in Bangkok, 34 cents per KL in Kuala
Lumpur and 55 cents per KL in Singapore. With the exception of Chennai, where few
customers are metered and therefore paying this rate, Indian cities therefore tend to have
much lower prices than other Asian cities.
Existing evidence suggests that even poor households in India can afford to pay more for
their water service, particularly if the increase in prices is accompanied by better service.
Under the current system, households may pay several times the municipal charges in coping
costs arising from the irregularity and unreliability of supply. For example, Zrah (2000)
estimates that households in Delhi in 1995 were paying 5.5 times the cost of their municipal
water rates in coping costs, while WSP (1999b) reports that a 1996 survey in Uttar Pradesh
found these coping costs averaged Rs 10 per KL, five times the prevailing tariff rate.
Willingness to pay (WTP) studies summarized in WSP (1999b) also suggest that households
are willing to pay more for adequate and safe water supplies.
source: The Times of India, January 5, 2004 Water watch,

http://timesofindia.indiatimes.com/articleshow/406554.cms [accessed April 11, 2004].
18

Revised 8/29/05
13
Raghupati and Foster (2002) calculate that even the poorest households could afford to meet
their subsistence needs with water tariffs several times higher than current levels. Based on a
World Health Organization definition, in which water is affordable if it constitutes less than
5 percent of a households budget, they calculate that even a typical five-member family
living under the poverty line with a per capita monthly budget of Rs 350 could afford to pay
up to Rs 6 per KL for a subsistence block of up to 10 KL per month. Given current income
levels, prices above Rs 10 per KL would make the service unaffordable for a significant
proportion of the population. Given estimates of operating and maintenance costs in the
range of Rs 15 per KL, some subsidization therefore appears inevitable, although more
efficient operation could substantially lower this gap.
Although some subsidization seems inevitable, the current system effectively directs most of
the subsidies towards the non-poor, while many poor do not benefit from low initial tariffs
due to a lack of access to the piped connection. Foster , Pattanayak and Prokopy (2003b)
provide evidence that a geographically-targeted subsidy system would result in significant
improvements in performance. For example lifeline rates could be set at lower levels in
designated slum areas. Additionally, subsidizing connection rather than use appears to be a
better way of directing subsidies towards the poor. It is also desirable to spread the
connection fee over several years rather than requiring a one-off advance payment for
connection from the poor.
Boland and Whittington (2000) suggest replacing the increasing block tariff with a system of
uniform price with rebate. Under this method, a volumetric charge is set equal to marginal
cost and coupled with a negative fixed charge, or rebate. In practice they note that a
minimum charge would have to be set to avoid zero or negative bills, and so at very low use
households would effectively face zero prices. However, they note that water demand is
likely to be very price inelastic at very low levels, and so this feature is unlikely to result in
large inefficiencies. The advantage of such a method would be that it is simple, transparent,
easy to implement, and avoids several of the problems inherent in the IBT.
In sum, there appears to be considerable scope for reform of the existing water pricing
system in India. Two key elements needed for reform are the ability to charge in proportion
to use, which requires metering, and political will for reform. We consider these features
next.
II.5 Metering
Some form of metering is necessary in order for suppliers to be able to charge in proportion
to water used. Metering has other advantages as well, such as helping suppliers keep track of
how much water is being used in different parts of the system, allowing detection and
reduction of illegal connections, and providing for identification of leakages. Despite these
advantages, not all municipal corporations have implemented widespread metering. Table 6
reports that the percentage of customers with metered connections ranges from less than 5
percent in Chennai in 2001 and 5 percent in Kolkata in 2004, to approximately 70 percent in
Delhi and Mumbai and 100 percent coverage in Bangalore. Only Bangalore has metering
above the Asian-Pacific average of 83 percent.
Of course in order to benefit from the advantages of metering, the meters must be
operational. Data on whether or not the meters are functioning is even scarcer than the

Revised 8/29/05
14
scattered data on metering itself, but available evidence shows very high-levels of nonfunctioning meters. Zrah (2000) reports that officially three-quarters of all connections in
Delhi in 1994 were metered, but that approximately one-third of the meters were not
working due to breakdowns and theft. The current situation in Delhi is no better, with
approximately one half of the 75 percent of meters not working.19 ADB (1997) reports that
70 percent of meters in Mumbai were not functioning in 1997, while 81 percent of meters in
Mumbai in 2000 were reported to be non-functional (Mathur 2001). Even with functioning
meters, bureaucratic inefficiency in the bill collecting process means that a good number of
customers do not pay their bills regularly. For example, in February 2004, only 50 percent of
metered customers of the Hyderabad Metropolitan Water Supply and Sewerage Board
(HMWSSB) were paying their bills regularly.20
One might question whether metering the poor is cost-effective for the water utilities. As
Noll, Shirley and Cowan (2000) note, externalities may be a substantial part of the cost in
urban water systems, so that even if the direct benefits of metering are low, there may still be
a strong argument for charging in proportion to usage in order to curtail wastage and
prevent spills that cause community health problems. Municipal water boards in India
generally charge either a security deposit on the meter (Rs 400 for a 15mm meter in Delhi),
or a monthly rental rate (Rs 15 per month for a 15mm in Hyderabad), which allows for the
costs of metering to be born by customers. Even among very poor households, such as slum
households, there is scope for metering with some adaptation. For example, the private
concessionaire in Manila put banks of meters in poor areas, from which families who wanted
a private connection could install and run a rubber piping connection to their own home. A
meter was provided for each private connection. (WSP 2001). This resulted in an increase in
coverage from 67 percent to 80 percent of households, providing direct connections to one
million urban poor.
II.6 Political Economy of Reform
In light of the widespread inefficiencies in the Indian water sector, there is clearly ample
scope for reform. However, given the predominantly state- and municipal-run water system,
any major reform needs to survive the political process, while even small changes in prices
require political approval.
Although the poor state of the existing system may suggest that reforms would be popular,
in practice reform has been difficult and several reform efforts have been killed by political
opposition. As Noll, Shirley and Cowan (2000) argue, several features of urban water
systems generally make reform politically difficult. In particular, they note that the political
benefits of water reform are often low, while, in contrast, reform will often involve giving up
political command of employment and investment in the public enterprise, and raising
prices. Changes in prices and layoffs are more transparent to the public than improved
operating efficiency, a reduction in the need for state subsidies, and small improvements in
quality, while larger long-term benefits of reform may take longer than an electoral cycle to
materialize. As a result politicians may not find the political benefits of reform outweigh the
costs.
19
20
Reference: private communication with Bunty Gupta, PriceWaterhouse consultant on water in New Delhi.
Source: The Hindu, February 6, 2004 Illegal connections prick water board, www.hindu.com.

Revised 8/29/05
15
Promoting reform therefore requires finding a way to raise the political benefits of reform
efforts, or of increasing the political costs of not-reforming. One key element in this regard
is increased information and public awareness. The Sukthankar Committee (2001) suggests
that local bodies should be made to publish fact sheets containing data on operations
periodically and subject themselves to public review.21 This would appear to be an important
first step in increasing the transparency of the existing system, and would allow for
benchmarking of local water utilities with other cities. Taking this further, it would seem that
benchmarking additionally against neighbouring countries, such as Nepal and Pakistan,
might be a good way of increasing public dissatisfaction with the status quo.
Public awareness could be further increased by explicit reporting of regular quality testing of
the water supply, along with information on hours of service and pressure, again
benchmarked against other states and neighbouring countries. This information should be
coupled with explicit information on how much water subsidies take from the state budget.
Reform goals in terms of quality improvement and expansion of access could then be easily
compared to the existing state of affairs, and price increases could be explicitly linked to
targeted improvements in key factors.
The lack of reliable, up-to-date, and publicly accessible information on many of the key
elements of municipal water supply acts as an important constraint on the transparency of
reform efforts. Additionally, benchmarked information is necessary in order to be able to
provide a baseline against which to evaluate reform efforts, and in order to help determine
where reforms seem to be having the most effect.
II.7 Private sector participation in urban water delivery
Although some improvement in public utilities may be possible, Harris (2003) finds that
there seems to be little evidence of the types of rapid improvements that have been
documented following private participation in many inefficient and poorly performing
utilities around the world, and that improvements which do occur under the existing
framework tend to be unsustainable. As a result, Governments have increasingly used
private sector participation to improve the efficiency and service of water provision.
Although widespread in much of the developing world, especially in Latin America, private
sector participation in water delivery is still rare in India. However, the 2002 National Water
Policy of the Government of India for the first time called for the encouragement of private
sector participation in water resources. Section 13 reads:
Private sector participation should be encouraged in planning, development
and management of water resource projects for diverse uses, wherever feasible.
Private sector participation may help in producing innovative ideas, generating
financial resources and introducing corporate management and improving
service efficiency and accountability to users. (GoI 2002, page 5).
We outline the arguments put forward for private sector involvement and summarize the
evidence from other countries, before highlighting the areas of debate in India.
Several NGOs such as Samaj Pragati Sahayog do this already. As a result of a campaign led by Mazdoor
Kisan Shakti Sangathan in Rajasthan, the right to information on public spending is now the law.
21

Revised 8/29/05
16
The argument for private sector participation

The above Section of the 2002 National Water Policy captures many of the arguments made
for involving the private sector in water provision. Supporters argue that the private sector
can introduce higher levels of management and technical skills, increase operating efficiency
and therefore lower costs, and provide investment financing to provide for expansion of
infrastructure (ADB 1997).
Private sector participation (PSP) incorporates a wide range of levels of private sector
involvement. At one end lies contracting out of services to the private sector, such as mains
repair, billing and collecting, and meter reading. Such arrangements are relatively straightforward and usually involve short-term renewable contracts. Under a management contract,
operation and maintenance of the water network is carried out by a private firm in exchange
for a management fee, with a typical contract lasting 3-5 years or more. Leasing and
Affermage contracts are similar to management contracts, although the private firms revenue
is determined by tariffs collected, resulting in more commercial risk being shared than under
a straight management contract. Typical contracts last 8-15 years. More private involvement
occurs under concession and Build-own-transfer (BOT) contracts. Under a concession, a
private firm manages and operates the whole utility at its own commercial risk and is also
required to finance new investment. Contracts tend to be for longer periods, such as 25-30
years. BOT contracts are used for major investment in new facilities. A private firm is
selected to build new infrastructure, such as a treatment plant, operate it under license for
the contract period, and then transfer it to the Government at the end of the contract. As
with concessions, contracts are generally for periods of 20-30 years. At the other end of the
public-private spectrum lies full privatization or divestiture, whereby the Government sells
the assets of the water supply company to a private firm, who runs it on a permanent basis
subject to government regulation.22
The term privatization is often used in both the literature and in public debate to encompass
both divestiture, and many of the other forms of private sector participation, such as
concessions and BOT schemes. Examples of full privatization, or divestiture are rare, limited
mostly to the United Kingdom (UN-HABITAT 2003), so studies which examine the effects
of privatization will generally include these other forms of private sector participation in
their analysis.
Evidence from other countries
Evaluation of the effects of private sector participation on urban water supplies has been
severely constrained by the overall poor quality of data available and the small number of
cases from which to draw general conclusions. As in India, in many parts of the world the
public utility does not release regular and informative information on costs, water quality, the
size of the network, etc. Moreover, with a change in ownership, any such information is
often lost as the private firm begins its own set of records. These features make it difficult to
measure the pre-privatization trends in access, prices, costs, and quality, which are key
elements in determining a counterfactual of what would have happened in the absence of
private involvement.
See ADB (1997) and UN-HABITAT (2003) for an expanded discussion of the pros and cons of each
arrangement, and examples of where each type of system has been applied.
22

Revised 8/29/05
17
As discussed above, privatization is not a yes/no decision, and there are several levels of
private sector participation. Moreover, the success of privatization is likely to depend
critically on the precise features of the contract itself, the extent of regulation and
competition, and the criteria set out for contracting. To take one example, Governments
which privatize in situations of fiscal distress may be primarily contracting on the basis of
highest Government revenues realized, whereas other contracts may be awarded on the basis
of least expensive network expansion coupled with lowest tariff.
Much of the literature has focused on examining the effects of privatization on the efficiency
of the former public firms (see Megginson and Netter 2001 for an excellent summary), with
some empirical support for greater efficiency with private operation.23 However, much of the
criticism regarding water privatization in particular is that it will result in large increases in
water tariffs, making the service unaffordable to the poor. This is one of the chief fears
raised by critics of moves towards greater private involvement in the Indian water system,
with the experiences in Cochabamba (Bolivia), Manila and Ho Chi Minh City provided as
evidence of this (see for example Sridhar 2003b).
Several recent studies have attempted to evaluate the effects of privatization of water and
other utilities on the welfare of the poor, explicitly attempting to control for existing trends
and using non-privatized firms or regions as control groups where possible. Most studies
focus on access to the water network and prices, since detailed information on water quality
and regularity is not generally available. However, the one careful study which does look at
quality does find significant gains. Galiani, Gertler and Schargrodsky (2002) compare
privatized to non-privatized municipal water systems in Argentina, and find child mortality
fell in the privatized areas through fewer water-borne diseases.
Three sets of studies which examine how water connections change with privatization are
Shirley (2002), which examines Buenos Aires, Santiago, Mexico City, Lima, Conakry
(Guinea) and Abidjan (Cte dIvoire); McKenzie and Mookherjee (2003) which examines
Argentina, Bolivia and Mexico; and Clarke, Kosee and Wallsten (2003), which examines
Argentina, Bolivia and Brazil. All these studies find increases in access following
privatization, although only part (or for Clarke et al. (2003), almost none) of this increase can
be attributed to privatization rather than other existing trends. The poor are the main
beneficiaries of this expansion in access, since the upper part of the income distribution in
most cases already have a piped connection.
The evidence on prices is more mixed. Shirley (2002) reports prices falling in Abidjan and
rising in Conakry. McKenzie and Mookherjee (2003) find that water prices fell in Buenos
Aires following the introduction of a concessionaire, and increased less under the water
concession in La Paz than in other areas of Bolivia. However, the most infamous water
privatization case, in Cochabamba, Bolivia, resulted in an average tariff increase of 43
percent. The result was the so-called water war, whereby violent strikes and demonstrations
resulted in the cancellation of the concession and reversion of the water network to the old
public utility.
See also the literature survey in Clarke, Kosee and Wallsten (2003) for more studies explicitly focused on the
water sector.
23

Revised 8/29/05
18
McKenzie and Mookherjee (2003) provide a method to calculate the effects of water
privatizations on poverty and inequality by valuing the effect on consumers of both the
increase in access and the change in prices. Using this method in Bolivia, Barja, McKenzie
and Urquiola (2003) estimate that gaining access to the water network is valued at
approximately 11-25 percent of per capita expenditure for the poorest deciles, whereas
relative price changes are found to have much more modest effects due to the low share of
water in total expenditure. As a result, the water concession in La Paz is estimated to have
resulted in a 2 percent reduction in headcount poverty and a small decrease in inequality.
Even the price increases during the Cochabamba failed privatization are estimated to result
in at most a 1 percent increase in headcount poverty.
The experience in other countries therefore provides some support for the view that
appropriate private sector participation in the urban water sector can both improve
efficiency and provide better service to the poor. While it is too soon to evaluate the effects
of private involvement in water in India, several projects are now underway.
Examples of Private-Sector Participation in India
1.Leak reduction in Bangalore
An example of contracting out services to the private sector can be seen in a pilot project
which began in June 2003 in Bangalore. The private firms Larsen and Toubro (L&T) and
Thames Water-UK in a 70:30 joint venture received a Rs. 500 million contract from the
Bangalore Water Supply and Sewerage Board (BWSSB) for a project designed to reduce
leakage and unaccounted for water through district metering, replacing consumer meters,
and relaying of supply lines. The project is funded by the Japan Bank for International Cooperation (JBIC). Competitive bidding was used to select the contractor, which was awarded
a contract for a pilot area of 35,000 house water connections. In 18 months the two
companies have to cut unaccounted for water for these connections from 31 percent to
under 15 percent, and then operate and maintain the area for 18 more months, at which
stage the BWSSB will take over management. Upon successful completion of the pilot
projection, the BWSSB proposes to then invite global tenders to expand the project to cover
400,000 household connections.24 To date there has been little controversy regarding this
project.
2. The Sonia-Vihar Water Treatment Plant in Delhi
In 2001, after an international call for tenders, the Delhi Jal Board (DJB) (the Delhi
Government water supply department) awarded a 10-year Build-Operate-Transfer (BOT)
contract to Ondeo Degrmont, a subsidiary of the French company Suez Lyonnaise des
Eaux. The contract covered the construction and maintenance of a 635 million litres per day
water treatment plant, intended to supply 3 million Delhi residents with drinking water, at a
cost of Rs 1.8 billion (approximately 50 million dollars). The plant itself has been completed
and is now waiting for the water supply to reach the plant. It is scheduled to become
See D.S. Madhumathi in The Hindu Business Line, December 26, 2002 Project to plug leakages: BWSSB to
shortlist bids (http://www.blonnet.com/2002/12/26/stories/2002122600811100.htm), Larsen and Toubro
announcements on the National Stock Exchange of India website http://www.nseindia.com/marketinfo/companyinfo/eod/announcements.jsp?symbol=L%26T, and The Hindu (2004).
24

Revised 8/29/05
19
operational in June 2004.25 Water for the plant will be supplied from the upper Ganga canal
of the Tehri Dam project, treated in the plant, and then distributed by the DJB.
This project has proved contentious with opposition from several groups. Critics such as
Shiva et al. (2002) suggest that details of the contract and tendering process remain oblique,
with allegations of corruption linked to claims that the contract was not in fact awarded to
the lowest cost bidder. Farmers contend that the use of water from the Ganga river will
deprive them of irrigation water, while there is also religious opposition to privatizing water
from the Ganga, a holy river for Hindus (Ninan 2003). The Delhi BJP recently claimed that
the project was a surreptitious government plan to privatize water in Delhi through the
backdoor, and would result in huge price increases once privatization was completed.26
These claims were denied by the Delhi Jal Board, but a lack of information regarding the
specifics of Degrmonts contract continues to fuel speculation.
3. Bangalore Operation and Management
In September 2000, the Bangalore Water Supply and Sewerage Board (BWSSB) signed
agreements with Vivendi Water and Northumbrian Water Group (NWG), a subsidiary of
Suez Lyonnaise des Eaux towards operating Indias first delegated water management. It was
proposed that each company would be given a 5 year contract for a city zone, each of
population 1 million. In their pilot zone, each company would be responsible for all of the
services of the board, including water supply and waste water management, revenue
improvement measures, billing and collection, and customer relations. The companies were
expected to reduce leakage and improve distribution, and it was suggested that successful
completion of the pilots would lead to 25-year contracts. However, political opposition and
allegations of corruption in the awarding of the contracts have meant that as of January
2004, no follow-up action has been taken on the contracts. Northumbrian Water, having
changed its name to Ondeo Water services, exited in January 2003, citing political instability
and uncertainty as reasons for its withdrawal.27
4. The Tirupur Area Development Program
Indias furthest step towards full privatization of water supply is the Build-Own-OperateTransfer (BOOT) contract carried out by the New Tirupur Area Development Corporation
Limited (NTADCL). Tirupur, in the state of Tamil Nadu, is Indias largest producer of
cotton knitwear and the textile exporters depend on water for their production. The
industrial sector played a lead role in the development of a project to supply water to the
dyeing and bleaching industries in Tirupur along with domestic consumers in Tirupur Local
Planning Area. A BOOT contract was awarded after an international competitive bidding
process to a consortium of Indias Mahindra and Mahindra, Bechtel, and United Utilities
JVC of the United Kingdom. These companies became equity owners in the project along
with the Governments of India and Tamil Nadu, and the Tirupur Exporters Association.
Sources: Delhi Jal Board (www.delhijalboard.com), The Hindu (2004), Sasi and Jain (2004).
Reported in the NCR Tribune, February 24, 2004 BJP resents foreign origin of Sonia Vihar water project,
http://www.tribuneindia.com/2004/20040224/ncr1.htm
27 Sources: The Hindu, August 2, 2002 Row over privatization of water supply,
http://www.thehindu.com/2002/08/02/stories/2002080204030600.htm; The Hindu (2004).
25
26

Revised 8/29/05
20
The project was first conceived in 1991, and after a long inception phase, construction began
in November 2002. By February 2004, 47 percent of construction was completed, and the
project is scheduled to complete construction in November 2004, and begin supplying water
in April 2005. The project will supply 185 million litres per day, with 125 million litres
destined for the knitwear industry, 25 million litres to domestic users in Tirupur
municipality, and a further 35 million litres to surrounding villages and towns. The planned
price will be Rs. 45 per Kl for industrial users and Rs 3.5-5 per Kl for domestic consumers.28
Although the Tirupur project has not met with as much resistance as the Sonia Vihar
project, anti-privatization activists have still raised several objections. The first is that the
prevailing groundwater in the area is of poor quality due to pollution from effluents from the
textile producers, and now these producers will come to control the communitys water
supply. The involvement of Bechtel has also led to comparisons with the involvement of a
Bechtel subsidiary in the failed privatization in Cochabamba, Bolivia, which resulted in large
price increases (AID 2003).
II.8 The challenge ahead
The current system of municipal water provision in urban India has not succeeded. Many
poor households are excluded from the water network, while a piped water connection still
means irregular pressure and water supply, and questionable water quality for a large majority
of customers. Moreover, in most cities water is underpriced, with much of the subsidy not
being well-targeted to the poor. The result is that revenues do not cover operating and
maintenance costs in most areas, let alone providing financing for needed quality
improvements. The next few decades are predicted to place more pressure on urban water
supplies, with large increases in the urban population adding to increased demand from
growing incomes. The urban population of 244 million is projected to expand to about 660
million by 2025 (Pitman 2002). As a result the status quo is not sustainable, and reform is
clearly needed.
Water is currently priced well below cost. Water rates in Delhi are currently less than onesixth of operating and maintenance costs for many consumers, while at least a doubling of
prices is required in many other cities. Such rates mean that even with efficiency
improvements, water prices need to rise in order to cover costs. This provides both the
reason for substantial reform and the biggest challenge. Successful pricing reform requires
functioning meters, which allow prices to be tied to units of use, and political will.
Experiences in several cities suggest that near full coverage of metering is possible. This
seems to be an area where contracting out services to the private sector under a competitive
bidding process could lead to improvements. The leak reduction project in Bangalore
provides one such example.
Sample sources: http://www.water-technology.net/projects/tirupur/; Government of Tamil Nadu

(http://www.tn.gov.in/policynotes/maws2003-04-12.htm); FIRE(D) (1999); The Hindu, February 16, 2004
Tirupur water supply may begin by April 2005
http://www.hindu.com/2004/02/16/stories/2004021604850400.htm
28

Revised 8/29/05
21
Concessions, Management and Leasing agreements will be difficult without prior pricing
reform. The proposed operation and management contract in Bangalore has hit resistance,
while a construction management project was also cancelled in Pune, in part due to pricing
concerns. Opponents also cited a lack of transparency and corruption allegations in both
projects. While there are possible gains from private management of utilities, given current
pricing levels, it will be easy for critics to argue that privatization causes rises in price in such
a situation. Successful reform therefore requires public awareness campaigns of the true
costs of the current policies, and for price rises to be clearly linked to goals such as
improvements in access, quality and service, and to the tax implications of reduced subsidies.
There is perhaps more scope for private involvement in providing new service, such as the
BOT contracts in Tirupur and Sonia-Vihar. Providing a new service provides more visible
benefits than improvements in management of an existing service. Nevertheless, if the new
service operates alongside or in addition to an underpriced municipal service, the higher
prices are again likely to be a political barrier to reform. Again general pricing reform is
needed to make such projects politically viable in most areas. While increases in price are
required in most cities, subsidies will still be needed to provide for the needs of the poor.
However, the current system effectively subsidizes almost all connected users, while many of
the poor do not benefit due to a lack of connection. Geographical targeting of subsidies
along with increased emphasis on connection subsidies should be used to provide for lifeline
water needs of the poor while increasing prices for other users.
Pricing reform is likely to be politically unpopular in the short term, while the benefits of an
improved system will take time to materialize. For this reason it is vital to increase the
amount of information provided, both in terms of the failing state of existing water systems,
and in terms of transparency of the reforms sought. The World Banks South Asia Water
and Sanitation Program is in the initial stage of a regional water utility benchmarking project
which should provide some useful information in this regard. However, much more
information needs to be collected and made publicly available in India in order to make
reform politically more popular.
While it is important to provide transparent and clear goals of the reform process in order to
generate public support for reforms, it is important not to oversell the reforms. Unrealistic
goals and public statements can cause the public to become quickly cynical and disillusioned
above the reform process, and frustrate further efforts. Barja, McKenzie and Urquiola (2003)
posit that overselling of the capitalization process in Bolivia was one reason behind public
discontent with the process, despite overall results which appear quite positive from an
economic perspective.
PART III: PERI-URBAN AND RURAL WATER SUPPLY OPTIONS
In this part of the paper, we start with a brief overview of water services in rural and periurban regions, and make the case for intermediate technological options and multiinstitutional partnerships for these regions. We then review low- to intermediate-cost
options for water delivery, water augmentation and water treatment. In each case, we review
the technology, give examples of the capital and operational costs (from field situations in
India wherever possible), discuss the challenges to extending access through this means, and
cite case studies showing how access was extended to previously underserved populations.

Revised 8/29/05
22
Unfortunately there are few peer-reviewed articles evaluating drinking water interventions in
rural India. Case studies from the grey literature, on which we have drawn, tend to be brief,
optimistic, and not easily comparable to one another therefore it is not always clear how
to distinguish the features of a case from its lessons. Consequently, our discussion of costs
and of successful cases should be considered illustrative rather than representative. In our
concluding remarks, we summarize the recurrent themes in the literature, draw out the policy
implications that these have, and point out some important themes on which the literature is
scanty. We suggest directions for research and data collection that should make it easier for
future researchers to understand what is working or not working, where and why.
III.1 Overview of water services in peri-urban and rural India
Rural India (which contains the largest number of people without access to safe water) and
peri-urban India (which contains the fastest growing segment of such people) are different
from urban India in that household water and sewage connections from and to centralized
piped systems are not the norm. Investment in water supply and sanitation has been for
decades skewed towards core urban areas rather than rural or peri-urban ones, and to water
supply rather than sanitation (Christmas and de Rooy 1991). Patterns of access naturally
reflect these investment decisions.
The consensus among water policy experts is that, as living standards rise, the demand for
household water connections will increase. However, with the exception of the better-off
residents of large slums and well-connected villages, in-house piped water and sewage will
remain inaccessible to most poor people in the near- to medium-term. Community-based
(shared) water facilities, therefore, which have always been the mode of access in Indian
villages, have to be extended and upgraded in ways that are affordable, relatively convenient
and sustainable for poor and unserved populations. Given the enormous diversity of
income, community capacity, local and state government capacity, topography and rainfall in
India, the range of feasible technological, institutional and financing solutions for access to
drinking water will also be diverse. Where decentralized technologies and decentralized
management are possible through local entrepreneurs, local governments and community
management they have to be recognized as mainstream rather than alternative methods of
service (Ruet 2004).
Water services in peri-urban regions
For each major city in the developing world, India included, there is the city proper where
roads, piped water, and electricity are relatively accessible. But either on its fringes or in
slums and tenements at its centre, informal settlements spring up, populated mostly by
migrants in search of work. These peripherally urban zones are called peri-urban areas.
Peripheral urbanization is growing fast -- the Government of India estimates show that 20
25% of the urban population lives in slums. This is likely an underestimate because only
official or notified slums are counted. Peri-urban incomes are low, because of the lack of
job opportunities and the low skills of many of the migrants. 32.4% of the urban population
lives below the poverty line (Reddy 2002). The haphazard build-up of the settlements, the
lack of funding for infrastructure, the unwillingness of governments to legitimize such
settlements, and high population densities combine to make these places water-stressed and
disease-prone. Diarrhoea is the most common disease, and human waste is the most
common contaminant. Indeed, sanitation is generally a more pressing need than water
supply in these slum conditions.

Revised 8/29/05
23
Peri-urban areas are rarely connected to their main city water supply or sewage systems. The
settlements are often illegal, so municipal governments are unwilling to legitimize them by
providing or extending their water and sanitation services. The Brihanmumbai Municipal
Corporation (BMC), for example, recognises only those peri-urban settlements that were in
existence before 1995. In other instances, the poor financial state of the water agency
prevents the central network from being extended to these unplanned areas.
Peri-urban residents without in-home connections meet their water needs in different ways.
Where groundwater conditions permit, residents have constructed shallow wells with
handpumps. Some urban slums in India have communal standpipes, fed by the city mains,
and shared by groups of families (from a few households to several hundred people). But
such water is often of poor quality. Where there is no potable groundwater, or the residents
have not been able to construct communal wells, or to arrange for standpipes, the two most
common options are illegal water siphoning and water vendors. Illegal water siphoning
occurs by breaking into the city pipes or storage systems, diverting the needed water, and in
the process allowing biological and other contaminants to enter the citys water supply.
Newspaper reports on how, for example, the BMC repaired water pipes that had been
broken only to find that in a few days the holes had reappeared are not uncommon. Water
vendors are also common in India, especially in the unserved areas of water-short cities such
as Rajkot, Ahmedabad and Chennai. Such privately vended water -- which seldom has any
quality controls -- sells for from 5 to 50 times the price of piped city-supplied water.
Water services in rural regions
Rural areas are characterized by traditional activities such as fishing, farming, herding and
crafts-based production such as food processing, textile weaving and printing, pottery etc.
Populations in rural areas are more dispersed than in urban centres, though South Asian
rural habitations are denser than in Africa and Latin America, and the boundaries between
villages and towns are uncertain and shifting. In developing countries, rural areas typically
account for between 60% and 70% of the population, but only 30% - 40% of the gross
domestic product. For India these figures are 70% and 40% respectively (GoI 2001).
Extreme poverty (defined as earnings of under $2 a day) is concentrated in such areas.
The simplest way to capture the state of rural access to drinking water is to quote from the
10th Five Year Plan of the Government of India. After explaining that only a third of rural
households have water within their premises, and therefore the rest have to fetch water, the
report says:
About 60 per cent did not have to go beyond 0.2 km for this. Seasonal
disruption of supply was common, especially in the summer months.
Households still depended on supplementary sources, especially where tubewell
or handpump was the main source. Practices of filtering or boiling water before
drinking were almost non-existent. (GoI 2002 - 2007, Chapter 5.5).
The national cost of fetching water has been estimated at 150 million person-days a year,
costing the Indian economy Rs 10 billion in (potential) lost production (Chaudhuri 1998).
The task of fetching is overwhelmingly female a study of the Konkan region of
Maharashtra estimated that each woman spent an average of 1.8 hours per day fetching
drinking and cooking water (Ahmed 1999). Common methods of carrying water are a large

Revised 8/29/05
24
pot on the hip or two nested pots on the head, with a total capacity of between 20 and 30
litres (Reddy and Rathore 1993).
The majority of the rural population (67%) draws its drinking water from open wells and
handpump systems, and aquifers supply almost 80% of rural Indias drinking water.
Handpumps have been used in India for decades over 3 million have been set up by the
Rural Water Supply Programme alone so they have been exhaustively refined and fieldtested. Modern handpumps are widely manufactured in India and are designed for easy
maintenance using few tools. Well drilling has also become cheap and efficient, especially in
the alluvial North. Despite these advances, a third of South Asias rural water infrastructure
is estimated to be dysfunctional (Parker and Skytta 2000).
In rural areas of India, the primary reasons for inadequate access to drinking water are
aquifer overdraft, diversion of surface water sources towards agricultural and urban uses,
biological contamination of (especially surface) water sources from human and animal waste,
terrible sanitation conditions (a major cause of biological contamination), groundwater
pollution, the failure of water services to keep up with population growth and the failure to
maintain existing water delivery systems such as handpumps. The Tenth Plan cited above
shifts the responsibility for drinking water and maintenance of community assets from
state level entities to Panchayati Raj institutions (PRI), which, being closer to their
constituencies, are presumably more accountable to them.
III.2 The case for mixed and intermediate delivery systems
As with urban systems, the shortage of investment capital and the lack of cost recovery are
major constraints to extending water and sanitation access in rural and peri-urban areas.
Centralized piped delivery systems are not the norm here, and in this section we make the
case that, to extend access within a reasonable time frame, low- to intermediate-level
technologies and multi-institution collaborations will play a significant role (Ruet 2004).
These are technically and financially more feasible than in-home piped water supply and
sewage, given the costs of the latter, and given the likely future outlays for water and
sanitation in India.
If full access (defined by Indian national standards) were to be extended to all the currently
unserved using traditional, government provided systems, public expenditures on water
supply and treatment would have to go up by several orders of magnitude. A joint GoI /
World Bank estimate, reported in WSP (1999a), concluded that Rs 29 billion a year would be
necessary just for operation and maintenance (O&M) in rural areas, plus investments of Rs
200 billion to rehabilitate and repair existing but malfunctioning systems, and Rs 450 billion
to achieve full rural coverage.29 As against these numbers, the actual O&M expenditures
were Rs 2.5 billion a year. For urban areas, if O&M could be fully covered through higher
tariffs, an investment of Rs 284 billion over five years would be needed for full coverage.
The report also suggested that the maximum expected outlay would be Rs 30 billion annually
for rural and urban water supply. New technical, financial and institutional options would
clearly have to be sought.
All costs in 1999 rupees, when $1 = Rs 43. We have quoted costs both in Indian Rupees and US dollars,
depending on the source. Consistent conversion to one currency was difficult because some source documents
do not give the base year for their cost and price information.
29

Revised 8/29/05
25
We have found cost comparisons of technical and institutional alternatives extremely

difficult to carry out. First, they are not always available. When available, they are rarely
complete so they might include upfront capital costs and perhaps the annual cost of
consumables. They almost never include transactions costs such as capacity building,
contract negotiations etc. Nor do they include the costs of community education etc when
these have been facilitated by NGOs. Even reports written by economists do not
consistently impute an opportunity cost to donated labour. Aggregated cost estimates with
missing components make it hard to see who is paying the cost of what, relative to who is
receiving the benefits. Nevertheless, we report below the best existing estimates, which
should be taken as indicative of the relevant orders of magnitude,
In a study of the global water situation, UNICEF argued that, to cover those without
access to safe water, it was imperative to focus on cost efficiency, including low-cost but
effective technical choices (Christmas and de Rooy 1991). The study estimates were that
for urban-type systems with treated, piped, in-home delivery, the capital and recurring
costs would be on average $200 per capita (in 1990 dollars); minimally treated piped
water delivered to public standpipes would cost $100 per capita; and low cost options
such as borewells with handpumps, rainwater harvesting or gravity-flow piped systems
would cost only $30 per capita. Similar (though somewhat higher) estimates can be found
in Briscoe (1995).30 In comparison, per capita GDP in India in 2002 was US $480 (World
Bank 2004).
Many low-cost (and low-maintenance) water delivery, augmentation and treatment
technologies have been, and are being, developed to serve the rural and peri-urban poor.
Many of these have remained at the demonstration level, others have been more widely
implemented but are of variable effectiveness in real-world settings, and yet others have
failed outright. We focus on technical options and institutional partnerships that are:
1. already in existence in India, but should either be more widespread, or made more
effective and reliable; or
2. not common in India, but tried and tested in other low-income settings.
III.3 Water delivery mechanisms
A water delivery mechanism is a method by which water that is already available to a water
agency or community can be accessed by individual users and households. These include
rural piped water systems that serve one or several villages from a small reservoir or
overhead tank; shared standpipes in slums and small towns; spring-fed gravity flow systems
(very common in hilly areas); water sold from tankers or lorries by small-scale private
entrepreneurs; and, of course, the handpumps on which the majority of the rural population
We have found cost comparisons extremely difficult to carry out. First, they are not always available. When
available, they are rarely complete so they might include upfront capital costs and perhaps the annual cost of
consumables. They almost never include transactions costs such as capacity building, contract negotiations etc.
Nor do they include the costs of community education etc when these have been facilitated by NGOs. Even
reports written by economists do not consistently impute an opportunity cost to donated labour. Aggregated
cost estimates with missing components make it hard to see who is paying the cost of what, relative to who is
receiving the benefits.
30

Revised 8/29/05
26
still depends. We discuss below the role of non-networked (or mini-networked) piped water
schemes, water vending and handpumps.
Small-scale piped water systems
Piped water schemes serve about 9% of rural India and 45% of small-town India (Section I.1
above). Rural piped systems are either single village or multi-village schemes. Water is
pumped into an overhead holding tank, ostensibly managed by the gram panchayat(s), from
which a small network of pipes carries water to communal collection points in the served
villages. Very few rural homes have individual connections. Recognised (or notified) slums
are served by piped water from municipal water agencies, when they are served at all. Periurban piped systems usually terminate in standpipes shared by a large number of households.
The Government of India considers a household to be served if there is a water point no
more than 1.6 km from the dwelling, shared by no more than 250 people, and capable of
delivering 40 lpcd.
Piped water is considered more desirable than shallow well water or surface water because it
goes through at least secondary, and usually tertiary, treatment.31 However the smallnetworked piped system in India is in very poor condition. In 1994 the Government of India
estimated that 26% of rural piped schemes needed repair (WSP 1999a). Many standpipes are
defunct they are either dry or have lost their faucets and run continuously. Municipal
agencies or gram panchayats are nominally in charge of standpipe repairs, but in reality there is
little follow-through. There is generally no arrangement to recover costs from piped systems
except through the relatively few in-house connections. Where the electricity supply is
intermittent, the large number of users per tap and restricted hours of operation make for
long queues, mostly of women and girls. In addition, as mentioned above, low pressure in
the pipes allows sewage to enter the water system. In short, public standpipes are a classic
public good that display the classic symptoms of lack of ownership and free-rider neglect.
As individual households become better off, and able to afford individual connections, the
demand for public standpipes is likely to fall. For example, SEWA, the trade union for selfemployed women, offers micro-credit and capacity building services so that its smallentrepreneur members can afford in-house water connections and toilets.32 SEWAs
Parivartan Slum Networking Project in Ahmedabad, in partnership with the World Bank and
the Ahmedabad Municipal Corporation, has been particularly successful in improving piped
water deliveries to Sanjay Nagar and Sinheshwari Nagar (see WSP n.d. for a detailed account
of SEWA Banks innovative and accountable micro-credit structure). However, shared
standpipes remain important sources of drinking water for those slum dwellers and rural
residents who cannot afford, or do not want to pay for, in-home connections, or whose
slums and homes are not legitimate. They will also remain important for pavement dwellers
and the rural landless who do not have pukka homes.33 In the near- to medium-term,
Secondary treatment separates solids from liquids and includes processes to remove microbiological
contaminants. Tertiary treatment removes additional solids and nutrients and inactivates pathogenic organisms.
Chemical coagulation, filtration and disinfection are all tertiary processes. (Gleick 2000, Ch 7)
32 SEWAs drinking water campaign in rural western Gujarat encompasses individual as well as shared piped
water (Joy and Paranjape 2003); its slum upgrading programme, however, is a response to the growing demand
of its members for private water supplies and toilets.
33 A pukka home is one that is built out of sturdy materials (clay, brick, thatch, sheet metals) and can be
expected to last some years.
31

Revised 8/29/05
27
therefore, small-scale piped water systems will likely comprise both private connections and
shared taps. The question is, how can they be made more reliable and effective for rural and
peri-urban residents?
We discuss in some detail a multi-village case study which elucidates both the financial and
organizational arrangements through which piped systems can be effective in the mediumterm, and some sustainability-related concerns for the long-term. The case is a four village
piped water scheme from Kolhapur district in the state of Maharashtra.
For over 20 years now, the joint water management body (mandal) of the villages of Lat,
Latwadi, Shiradwad and Shivnakwadi has run its own multi-village piped water supply
scheme, and has produced a healthy revenue surplus in the process. The water committee
came into being with the active encouragement of the Irrigation Department engineer and
the gram panchayats in particular the sarpanch (headman) of Lat. In this scheme, water is
pumped from the river to a settling tank, particulate matter is allowed to settle, and then the
water is released through pipelines at fixed times. Piped water supply is given to individual
households and to public standpipes (for poorer people and for common use). Individual
connections cost Rs 520 (in 1999), of which Rs 200 is a recurring annual fee;34 water from
the standpipes is free. The operation and maintenance of the public standpipes is paid for by
the gram panchayats, and they in turn collect a small Panchayat charge from all the village
residents. The annual water charges and the Panchayat charge cover the costs of repairs,
electricity and salaries, plus surplus revenues have been used to create common village assets.
Essentially, the Kolhapur scheme has found a way to use individual connection revenues to
cross-subsidize the public water points.
The World Bank assessment (WSP 1999c) of Kolhapurs multi-village system highlights its
three important lessons (presumably for scaling up and/or replication). First, it brings out
the role of advocacy and mobilization, in this case performed by government officials (we
see throughout the case studies in this paper the enabling and facilitating role of the state).
Second, it shows how innovative cross-subsidies can make public standpipes financially
viable (a problem that bedevils most public piped systems in the country). And third, the
small size of the scheme enabled multi-village cooperation to be sustainable (this is a critical
insight, well understood in the common-pool resources literature, to which we shall return).
The Kolhapur scheme also reveals some problems the source water in the river is getting
polluted, the pumps are aging and need replacement etc. We focus here on two aspects that
are relevant to the sustainability of, and to access from, piped systems everywhere, not just in
Kolhapur (and that are just as much lessons as the positive lessons cited in the literature).
1.
Declining water availability per capita over time
With growing populations and growing demands, the scheme can no longer sustain its
designed allocation of 70 lpcd. The 1999 assessment projected an availability of 31 35 lpcd
by 2004 (lower than the GoI norms of 40 lpcd). In these circumstances, those with private
connections will almost certainly withdraw more than 40 lpcd while those at the public
standpipes will get less.
To contextualize these numbers, it may be helpful to note that in 1998 the minimum agricultural wage in
Maharashtra ranged from Rs 35 Rs 40 per day. Minimum wage laws, however, are not consistently followed.
34

Revised 8/29/05
28
2.
Low water pressure in the pipes as a result of more household connections
With more in-home connections, the withdrawal of water is spread out through the day. The
standpipes, however, have fixed hours of operation. The staggered withdrawal lowers the
water pressure in the pipes for everyone, allowing surrounding contamination to enter
through cracks and leaks in the system. The declining pressure also increases the waiting
time at the standpipes, disproportionately affecting the worse-off.
One option in such cases is to increase the water supply overall, perhaps through a larger
allocation from the state, or through re-allocating a fraction of agricultural water, or through
treating and recycling water. But this may not be possible (for scarcity or political reasons).
Another option would be to meter the individual connections with steep charges beyond a
minimum first tier (see section II.5 above). This is not the norm today, even in the
successful piped water schemes, and the viability of individual metering in rural and periurban areas is yet to be determined. A third option for water stressed regions would be to
increase the number of standpipes and keep the charges for in-home connections high. The
standpipes could serve fewer than, say, 150 people each, and either be run as common-pool
resources (if the user communities are able and willing), or on the ration-card model (where
each households withdrawal is monitored to ensure that everyone has access to an
affordable first tier of water), or on the kiosk model where water is sold to the users or
perhaps some combination of these. But we do not know of any piped water projects that
incorporate these or similar safeguards. Finally, local water agencies could resort to tankersupplied water for the poor, even in areas with piped water (see below).
The general problem is what works well for the principle of cost recovery does not
necessarily work well for the principle of 40 lpcd for everyone. A casual line in the Kolhapur
report from WSP reveals that: The mandal reduced the number of public standposts to
promote revenue-yielding household connections within the villages. That is, rather than
make the standpipes on which poorer people depend more accessible, the mandal made them
less convenient in order to encourage more household connections. We see from numerous
case studies in the literature that people are willing to pay for in-home connections rather
than walk 1.6 km to a shared tap (naturally); but we do not see from these studies how to
ration water (in water-short regions or in water-short years) so that the public health goal of
40 lpcd can sustainably be met. We do not know if and what people would be willing to pay
for a standpipe system, but within only 0.2 km of their dwellings Our argument here is not
against in-home connections or cost recovery, but against institutional and financial
arrangements through which these goals could undermine, rather than support, the goal of
wider access.
Water vending
Water vending encompasses many different kinds of water transactions that occur outside
of, on the one hand, municipal distribution of tap water, and, on the other, direct collection
of water by a community from a standpipe or handpump. The vendors play an intermediary
role, which can involve re-selling water from a municipally supplied standpipe or obtaining
water from a groundwater source and transporting it by tanker or cart to slum areas where
residents purchase it. As a rule, vendors sell water for many times the rate charged by the city
to its networked customers.

Revised 8/29/05
29
Though the bhishti (water carrier) is a traditional figure in India, the nature and extent of
water vending has not received enough attention in the Indian context. If vendors are
discussed, they are generally vilified for exploiting the poorest. However, Crane (1994)
reports that the legalization of water re-selling from municipal connections in Jakarta acted
as a costless extension of the citys piped system (and as a transfer from those without
connections to those with). Zaroff and Okun (1984) also argue that vending is a low-cost
alternative form of water delivery. Kjelln (2000) shows that, given the inadequate state of
Dar-es-Salaams water infrastructure, small-scale water providers complement the public
distribution system and do not provide poorer quality water to the slums than the city does
to its official customers. Similar points were made by a comprehensive study of water
vending and re-selling in ten African countries (Collignon and Vzina 2000).
Though few urban residents in India (1% according to the National Sample Survey, 54th
Round) depend exclusively on water vendors, during periods of scarcity they are the
mechanism of water service provision to the poor (and in some cases the rich as well). In
peri-urban areas the exclusive dependence on vended water is greater than at the urban core.
In water-short cities and rural areas, e.g. in and around Rajkot and Ahmedabad, city
governments themselves operate tankers to serve areas without piped connections or clean
groundwater. But the government may not own enough tankers to cover the unserved areas,
and private water vendors operate in addition to these municipally approved tanker suppliers
(Matzger and Moench 1994). We argue that, given the significant infrastructure investment
needed to extend piped connections to the urban unserved, the operating deficits of most
Indian utilities (Ruet et al 2002), the increasing focus on cost recovery (including capital
costs), the inability of the majority of slum dwellers to contribute to capital -- though not
necessarily operating costs (Bajpai and Bhandari 2001), and the unattractiveness of periurban areas to the formal private sector (Derby and Gadgil 2003), more cities should
consider recognizing, contracting with, and regulating local water entrepreneurs. In the
absence of official recognition and regulation, water vendors will continue operating anyway,
but without quality controls, price monitoring or accountability. Because access to water (or
the lack thereof) has clear public health externalities, the current lack of quality and price
regulation is not tenable. Chennais Metro Water Board, as reported by Ruet et al (2002),
offers an example of devolution to local contractors as an interim way of coping with the
joint scarcity of water and funds.35
Chennai provides tap water for an average of just 4 hours a day, and has a slum population
of about 400,000, but 97% of its residents are covered via tap and tanker services (MIDS
1995). Complex water markets have developed in Chennai, with a mix of illegal re-selling of
municipal supplies, private bore-wells, and recognised as well as unofficial water vendors.
The Water Board, which enjoys more autonomy than its equivalent in most other cities, has
contracted with 500 private contractors to supply various parts of the city, including slums
that do not have public standpipes. Exnora, a Chennai-based NGO, works with the Board
to organize the slum dwellers into committees that distribute the tanker waters. The tanker
transporters buy water from farmers outside the city centre,36 paying them just over Rs 3 per
We are grateful to Veena Subramaniam for bringing us up to date with the Chennai water markets.
These groundwater withdrawals are unregulated, but groundwater withdrawal for domestic use is not a threat
to the water table. Tamil Nadus dire aquifer situation is almost wholly attributable to pumping for rice paddies
and the resulting salt ingress along its coast.
35
36

Revised 8/29/05
30
cu m, and being paid Rs 15 per cu m by the Water Board (this includes the cost of transport
and maintenance, in addition to the cost of water). Approximately 10% of the Chennai
Water Boards annual expenses go towards hiring and monitoring these tankers. The
contract is monitored by the Water Board -- the tanker owners attend regular vigilance
committee meetings (at which Exnora is also represented) and train their drivers. In return,
they are guaranteed daily business and prompt payments. Private tankers that are not under
contract to the Board, in contrast, also buy water from farmers and sell the water in bulk to
consumers. But the Board has a chance to monitor the quality of water and service provided
by those under formal contract.
Ruet (2004) suggests that such public-private-civil partnerships, some of which are already
under way, should be expanded and formalized in several Class I cities. He argues that cities
could start with delivery arrangements through local entrepreneurs, and only then consider
contracting out to national or global service companies. This would lower the risk borne by
the companies (who would not have to reach distant and unconnected households), better
target the citys investments, and reduce the chance of cutting out the poor and unconnected
at least where institutional safeguards have been built in. This implies that exclusivity
provisions in concession contracts, such as those granted to the privatized Aguas del Ilimani
in Bolivia, should generally be discouraged, lest they close down the possibility of
employing competition in the market to reduce prices, improve service, or provide
alternative service options (Komives 2000).
Wells and handpumps
Well and handpump technology can be divided into three general categories:37
1.
Open wells: The basic village well. New open wells are inexpensive to construct and
require a minimum of external expertise and equipment. They are still appropriate in
locations where the water table is high and regularly recharged, such as in canal command
areas. Even where the lower layers of groundwater are brackish or saline, canal seepage
forms a layer of fresh water above the brackish layer, which can then be accessed via a well
(van der Hoek et al 2001). Simple open wells are an ancient technology, but are susceptible
to biological contamination. Sealed bore-wells fitted with handpumps are therefore a public
health improvement.
2.
Shallow well handpumps: can lift water from depths of up to 7 m. Shallow-well pumps
operate by suction, which limits their effective pumping depth, but means that all the
moving parts (that eventually need replacement) in the system are above ground-level,
making maintenance and repairs relatively simple. Many different Village Level Operation
and Maintenance (VLOM) shallow handpump designs exist and are currently in operation.
The choice for a given village / community would depend on the availability of the pump
and the supply chain of spare parts and technical expertise in a given area.
3.
Deep-well handpumps: are generally referred to as lift pumps, since they use a pump
mechanism placed underground rather than using suction at surface level. This design
enables deep well handpumps to lift water from 50-60 m below the surface, but the
configuration makes maintenance more difficult in comparison with shallow-well designs.
International and Indian research and development efforts over the past 25 years have
The technology descriptions and some of the cost estimates for handpumps, rainwater harvesting and water
treatment methods borrow heavily from Downing and Ray (2002). We thank Jim Downing for generously
allowing us to draw upon this work.
37

Revised 8/29/05
31
produced several VLOM deep well handpump designs, including the India Mark III and
Afridev. These pumps are widely manufactured in India (and Pakistan) and have been field
tested in thousands of locations.
Handpump technology has two upfront costs: digging the bore, and purchasing and
installing the pump, which includes pouring a concrete pad. The generally-accepted $25
average cost per capita rule of thumb for water supply and sanitation projects in Asia (Nigam
and Ghosh 1995) is based on (shared) handpump installation, because the vast majority of
community water projects in developing countries, including India, have been of this type.
Nigam and Ghosh (1995) estimate that per-capita best practices costs (BPC) for deep well
boring and handpump installation in rural areas are of the order of $6 in Asia.38 Actual costs
vary depending on the availability of pumps and spare parts, drilling equipment, road
conditions, and local management expertise. The price of a high quality VLOM deep well
handpump is, roughly, $500-$800, while well drilling can cost anywhere from a few hundred
dollars to over $10,000, depending on the depth of the well, soil conditions, the drilling
equipment used, labour costs, and the degree of community participation (Carter and Ball
2002). In hardrock regions of India (where it is naturally more expensive to drill), Talbot
(1997) estimates that an India Mark II pump, of 5 diameter, drilled to a depth of 60 m and
cased to 10m would cost about $1300. 39 For shallower bores upto 30m the cost of a pump
unit is only $170 (1996 dollars). In neighbouring Pakistan, competition among producers of
a smaller and local version of the Afridev has brought down capital costs of the pump unit
from $320 to $130 in 10 years (WSP 2000).
Handpumps and wells are a suitable technology for any rural or peri-urban location where
the groundwater is not contaminated and the water table is not lower than about 60 metres.
Because of high population densities, lack of sanitation and industrial activity, peri-urban
groundwater in many, but not all, cases is too contaminated for a handpump to provide safe
water. Aquifer levels are also dropping as water is withdrawn for municipal and industrial
use, and as the natural recharge area is reduced through paving over. In some areas,
however, there are several layers of groundwater, the deeper of which is usually less
contaminated. So residents using a shallow-well pump to draw water from a polluted nearsurface aquifer would benefit from additional investment to tap their deep aquifer. Many
sites also simply have broken handpumps that could be replaced the Government of
Indias own report in 1994 estimated that 22% of handpumps needed repair, and 12% were
totally defunct (cited in WSP 1999a). Finally, as with standpipes, walking times and waiting
times are a daily burden on (mostly) rural women and girls (Ahmed 2000). The UNICEF
standards of one water point for 250 people is probably not adequate for most communities
(as we argue in our conclusion), and in any case is not met where populations have risen
over time but the number of water collection points has not.
Groundwater depletion and aquifer pollution (from natural sources, e.g. arsenic, or human
ones e.g. agricultural and faecal contamination) are the main threats to sustainable water
extraction for basic domestic needs. In most locations, groundwater extraction for domestic
South and South East Asia have the lowest costs the per capita BPC is $15 in Africa (because almost all
supplies spare parts have to be imported) and $30 in Latin America (because of relatively high labour costs).
39 The cost of drilling in hardrock has been brought down to about $20 per metre in India this is probably the
lowest figure in the world.
38

Revised 8/29/05
32
consumption alone will not deplete the local aquifer. Unfortunately, groundwater withdrawal
for irrigation is almost totally unregulated, even in developed countries, and aquifer overdraft
in India is effectively promoted by state governments through energy subsidies for pumping.
With increasing industrial activity and agricultural chemical use, aquifer contamination is
increasing, with little prospect of near-term regulatory or management intervention. Despite
these problems, thousands of villages remain where inexpensive handpumps and working
wells are urgently needed.
Setting aside the aquifer depletion problem for now, Indias handpump infrastructure can be
revitalized if the systems are convenient to access, if they can be repaired and maintained
without delays, and if the costs of new installations and maintenance can (at least partially)
be recovered. Two innovative efforts from two different states offer some insights into how
these goals might be achieved.
1. Swajal Project; Uttar Pradesh (WSP 2001b)
The World Bank assisted Swajal Project is a US$ 63 million effort, to cover 1,200 villages
with a population of 1.2 million in 19 districts in the Hill and Bundelkhand regions of Uttar
Pradesh. The main goals are to provide water through a combination of technologies (piped
water, handpumps, rainwater harvesting and open wells); to introduce sanitation and
composting; and to increase environmental and health education. All the projects under this
scheme require 10% of the capital costs and 100% of the operating costs to be paid by the
users. Village water and sanitation committees are elected, are empowered to operate the
project bank account, and to make decisions about technology choice and the contracting of
goods and services. The committees work closely with their facilitating NGOs and their
Project Management Units, which are autonomous and registered at the state level. Swajal is
an ongoing project so it is too early to evaluate its overall effectiveness and sustainability. An
interim report from the village of Laxmipur relates how the water committee visited private
firms that had submitted competing bids for the construction of a tubewell with an overhead
storage tank. Another report from Khankatiya a hill village in Pithorgarh shows how the
village committee, which had decided that they needed a gravity flow water system, found a
skilled mason from a distant village to construct the reservoir. The Uttar Pradesh Jal Nigam
(the official water agency) plays little role in Swajal, but the UP state government has
supported it in principle and through allocating funds.
The Swajal Project is not a handpump project as such it covers too wide an area to use any
one means of water delivery but we note that villages that do use handpumps contract for
an average of eight handpumps per village. A crude per-capita calculation reveals that 125
150 people share a handpump (that is, the water points are almost twice as dense as
UNICEF standards). Waiting times, walking times, and free-rider induced neglect are all
likely to be reduced with smaller user groups, making these community-level schemes more
expensive in the short run but potentially more sustainable in the long run. We return to this
point later in the paper.
2. Samaj Pragati Sahayog; Madhya Pradesh (SPS 1999)

Revised 8/29/05
33
Samaj Pragati Sahayog (SPS) is based in Dewas district, Madhya Pradesh, a drought-prone
and heavily adivasi region.40 In 1993, a survey found that 70% of the areas handpumps were
in disrepair, and that government-trained mechanics were unwilling to travel to these remote
villages to carry out repair and maintenance activities. SPS sent 4 tribal men to Tilonia in
Rajasthan (which has a nationally respected Social Work and Research Centre) to be trained
in handpump repair and maintenance. Upon their return, they formed a core team of
mechanics, repaired the handpumps in 14 villages (previously declared irreparable by
government mechanics) and trained others through SPS-organized handpump workshops
and camps. By 1997, over 50 adivasis, including 11 women, had been trained to repair
handpumps. They worked in 14 hand-pump circles, each circle comprising 2 3 panchayats,
with a trained mechanic in every circle. Each circle today is self-reliant in parts and tools, and
all the handpumps are working. This has ensured a basic supply of safe drinking water.
The SPS model is very different from the Swajal model in that formal cost-sharing
arrangements do not exist between the government and the user groups. Nevertheless, it
exemplifies the principle of cost-sharing in the form of labour from the users, organization
from the facilitating NGO, and a small cash grant and legal recognition from the Madhya
Pradesh government. The government has officially turned over the responsibility of
handpump maintenance to the circles, and it was the government that suggested expanding
out from the original 14 villages to the 90 villages that are now served by the circles.
The Madhya Pradesh effort is small, while the Swajal Programme of Uttar Pradesh is much
more ambitious. What they have in common is that the relevant NGOs or donors partnered
in some fashion with the local governments in question instead of trying to bypass the
government altogether which could be essential for scaling up successful pilot efforts
(World Bank 2004; Chapter 2). This makes Swajal and SPS potential models for replication
in other communities that face similar problems but have the capacity to self-provision.
A note on handpump history
We have seen that approximately a third of Indias handpumps are either defunct or in need
of repair a depressing statistic that is a regular feature of the literature on the failures of
government water policy. But it is worth mentioning that, when looked at through the lens
of history, the largely publicly-funded handpump programme has been a phenomenal
success. Access to safe water increased from less than 10 per cent in 1966 to 31 per cent in
1980 to 82 per cent in 1996 (Talbot 1997). This remarkable achievement was a result of
NGOs experimenting with technologies of drilling boreholes and developing pumps strong
enough for community (rather than family) use; of the Government of India generating
demand so that private companies stepped in to produce and improve the handpumps and
spare parts; of training GoI engineers and mechanics to use and repair these new
technologies; of UNICEFs sustained support at the level of national policy as well as onthe-ground implementation; and of NGOs continuing to work with rural communities. For
instance, the India Mark II model was designed by an NGO in Maharashtra, modified and
improved by an engineering company with UNICEF funding, and launched as a massive
effort by the GoIs Rural Water Supply programme.
Adivasis are Indias indigenous peoples. They are commonly referred to as tribals but many scholars prefer
the term adivasi (literally: first residents)
40

Revised 8/29/05
34
Of course, today it is clear that the state-led maintenance system has broken down and that
the handpump programme needs some mode of cost-sharing to remain viable. It is also clear
that declining water tables and water quality pose the most serious threats to the handpump
network. As Talbot (1997) reminds his readers, The India experience has taught us that
every problem has a solution; but that each solution creates another set of problems.
Nevertheless, the history of Indias handpump programme is a history of multi-institution
coordination towards a single goal that remains instructive in the light of todays problems.
III.4 Water augmentation mechanisms
A water augmentation mechanism is a method by which water that would normally run off into
rivers or seas, and so would not be accessible for drinking or agriculture, is captured and
stored so that it can be used. Augmentation methods include storing water in underground
tanks for use in the dry season; collecting rainwater on the rooftops of homes, schools etc;
and watershed-scale rainwater harvesting which can directly be tapped or can recharge the
surrounding aquifer. Of these methods, watershed-scale rainwater harvesting has received
the most attention in the literature and on the ground in India, and is part of official water
policy in at least five states. We focus this section on this option.
Rainwater harvesting
The simplest and most common method of watershed-scale rainwater harvesting (RWH) is
the construction of a checkdam across a seasonal drainage. During heavy rains the ground
becomes saturated and rainwater flows quickly along the surface instead of percolating into
the earth,41 flowing into drainage channels and then into streams, rivers, and ultimately the
ocean. A checkdam built across a drainage channel prevents the water from flowing
downstream, creating a small reservoir. The water in this reservoir has two uses -- directly as
surface water for watering animals or for bathing, washing, or, with protection and
disinfection, drinking. The water in the reservoir also seeps into the ground, recharging the
local aquifer, to be tapped over the rest of the year. Maintenance of a checkdam and
reservoir is relatively simple, the major task being periodic silt removal.
Watershed RWH on its own is not a complete supply solution. A surface reservoir is subject
to biological contamination as is any pond and will need protection and treatment to be safe
for drinking. Rainwater harvesting provides a reliably clean water source most easily and
inexpensively when it is used to recharge an aquifer that a community can tap via an existing
well. Many locations in India suffer from aquifer depletion, most often as a result of irrigated
agriculture. In such locations, a well and handpump are often already in place, and a wellsited aquifer recharge project can make those wells useful again. An important benefit of
watershed RWH is that it can provide additional water for livestock watering, and potentially
for agriculture. A rural community thus gets water for productive activities important to the
sustainability of its lifestyle, rather than just a subsistence water supply. But one problem
with RWH in such settings (as with pump projects in general) is that, unless groundwater
This is why rainwater harvesting is necessary where topography permits, and why wells alone are not enough.
The soil cannot store the moisture when the rain falls over a few days in heavy torrents especially if there is a
hardrock layer a few meters below the ground. This rainfall and bedrock combination is common in Indias
drylands. In addition, where groundwater levels are depleted or polluted, harvested rainwater is a cleaner, more
accessible source of water and of shallow aquifer recharge (Shah et al 1998).
41

Revised 8/29/05
35
extraction by agriculture is regulated, there is a significant risk that the recharge project will
just be supplying water to farmers, and not to the villagers that need the drinking water.
The capital and start-up costs of RWH are hard to generalize. The major cost is in checkdam
construction, and the size of the dam needed to catch a given volume of water will vary
substantially with topography, and with whether the goal is just drinking water or additional
productive activity. Local availability of construction materials, equipment, and construction
expertise also influence the cost. So do the arrangements for construction labour certain
communities donate labour while other pay wages in cash and/or food. In the publiclyfunded Gujarat program mentioned below, the average checkdam construction cost was
about US $3200, and the average reservoir retained about 15 million litres of water. So, very
roughly, the cost per litre of harvested water was $.00023 per litre, or 23 cents per thousand
litres -- approximately two orders of magnitude less than the costs cited for rooftop
harvesting systems (Shingi and Asopa 2002). At this rate, catching enough water to give one
person 20 litres per day for one year would cost only $1.69 (though to make sure that
enough recharged water makes it to wells where it can be withdrawn and used, more water
would need to be harvested, so costs will be higher). Another set of costs was provided by
Sanjit Roy,42 from RWH efforts in Rajasthan and Sikkim. These systems store rainwater (for
drinking only) in underground tanks that are treated with chlorine and fitted with
handpumps. The labour component in Rajasthan comes to $0.03 / litre in rocky soils and
even less in sandy soils. The water is sold (by tanker and at the pump itself) for an average
price of $0.02 / litre. The Sikkim structure with the tank carved into the hillside -- has a
storage capacity of 160,000 litres and a total per litre cost of $0.03.43These costs are only for
the investment in the checkdam, and do not include well re-boring, handpump replacement,
capacity building, NGO facilitation, etc. that may additionally be necessary.
Despite impressive results from many RWH efforts, and the near-zealous advocacy of its
many proponents, RWH cannot be implemented in every town and village. The lower
reaches of watersheds are better catchment areas for harvesting rain than the upper reaches
in general (Kerr, Pangare and Pangare 2003). Uncertain subsurface hydrology makes it
difficult to predict the rate, direction, and extent of the movement of water underground. It
is therefore difficult to predict the annual yield of a watershed rainwater harvesting project,
in terms of the accessible groundwater that it generates. Small checkdams are simple
engineering structures and can be built from earth, rocks, masonry or concrete. Larger
checkdams introduce a series of complications -- the engineering is more complex, and the
danger of collapse more substantial. Further, because water rights are poorly defined (or
simply defined as being the property of the state), large checkdams that produce large
reservoirs can generate conflict with the state, other water authorities and downstream water
users. We shall re-visit this point later in the paper.
Major NGOs and environmental organizations such as Samaj Pragati Sahayog (Madhya
Pradesh), Centre for Science and the Environment (Delhi), Barefoot College and GRAVIS
Sanjit Bunker Roy of Tilonia, Rajasthan, is a nationally known rainwater harvesting promoter and facilitator,
who runs an innovative village-based programme of extension and outreach in the fields of health, education,
rainwater harvesting, handpump repair, veterinary care etc. (www.barefootcollege.org)
43 Cost breakdown from Barefoot College Training Campus, Sikkim. Costs were unsubsidized, and included
material costs of cement, sand, iron rods, bamboo etc as well as labour, transport and sales taxes.
42

Revised 8/29/05
36
(both in Rajasthan) and Utthan (Gujarat) have been active in promoting and implementing
RWH for many years. Government programmes in several Indian states have also led to the
construction of large numbers of watershed RWH structures. The state government of
Gujarat in 2000-1 funded a cost-sharing program whereby farming village residents, with
NGO assistance, constructed more than 10,000 checkdams in less than a year to recharge
agricultural and domestic wells (Shingi and Asopa 2002). Several NGO efforts have focused
on developing community capacity and resource mobilization, without any connection to the
state water authorities. However, as RWH goes to scale, it is essential to work with relevant
water agencies or Irrigation Departments for reasons of property right allocations, third
party effects of numerous (though individually small) hydrological interventions, resource
mobilization, and overall integration into mainstream water policy. We present an example
which is NGO-led, but state government-partnered, as a model for advancing access to
water through RWH, where conditions permit.
Samaj Pragati Sahayog (SPS) is based in Bagli tehsil in Dewas, Madhya Pradesh. It has
pioneered efforts to provide water, and also year-round employment, through decentralized
rural water systems and watershed protection.44 In addition to the hand-pump project
mentioned above, SPS has facilitated rainwater harvesting and water storage through earthen
bunding and terracing structures. Central and Western India have seen repeated failures of the
monsoon and severe drought conditions over the last decade. Rainfall deficits have led to a
severe shortage of drinking water for people and water for livestock. To drought-proof
villages against future lean years, SPS-organized teams of adivasis constructed protected
ponds to hold rainwater and to recharge the local aquifers. They organized labour sharing
and cost sharing schedules (most labour was paid the minimum wage, with less than 10% of
it donated, and the financial costs were shared by donors, community members as well as
the Madhya Pradesh state government). The teams also designed and implemented rules of
water allocation that were transparent and (apparently) perceived to be equitable. In short,
the water system was run on the common-pool resource model, which is not a feature
shared by most (community-level) piped or handpump systems.
Today, after 10 years of work in the area, trained teams of adivasis frequently plan, choose the
sites, and design the RWH structures for new projects (not necessarily initiated by SPS). The
team accompanies SPSs technical experts from village to village, planning the project,
explaining to the local people its purpose and mobilizing them to participate in the process.
SPS has spread its work to over 100, 000 hectares in 4 Indian states, in most cases with
financial and technical support from the relevant government agencies. SPSs cost estimate
for a system of 100 rainwater ponds and 8 earthen bunds, for 7000 people in 15 villages, is
approximately $300,000 (in 2002 US dollars). 10% of this sum goes to administration,
training, education and community organization. The remaining 90% goes into the works.
This is many times more than the number calculated for the hypothetical 20 lpcd water
harvesting cost for the Gujarat checkdam project, because much more than 20 lpcd are being
It is quite normal for RWH projects to have multiple goals (drinking water, livestock maintenance,
emergency irrigation and employment) this is a feature that makes cost comparisons with purely drinking
water systems impossible. While these multiple goals increase the costs per litre of water harvested, arguably
they also increase community commitment and sustainability.
44

Revised 8/29/05
37
collected to recharge the aquifer. We note that the average cost of checkdam construction,
about $3000, was roughly consistent across both projects.45
In a country such as India there can be no single blueprint for scaling up innovative water
supply delivery and financing mechanisms. The SPS model of indirectly forging technical
and financial partnerships between poor communities and their local governments is a way
of delivering services in which the government does not provide the service, but pays a more
efficient provider to do so. Morris (2004) argues that most private providers and NGOs
enter the field of water and sanitation to compensate for state failure, but some innovations
can in fact be precursors of larger social changes. This was true of the Anand dairy
cooperative which became the seed for Operation Flood. SPSs watershed development and
SEWAs micro-banks, which have grown through partnerships with international financing
and local governments, but have retained their focus on community needs and community
autonomy, may have similar potential to be leverage points (Morris 2004).
III.5 Water treatment mechanisms
If water exists but is not safe to drink then it needs secondary and tertiary treatment but
treatment of water to make it potable has not been given the same attention as making a
minimum quantity of water available. Where it is not possible (for reasons of cost or political
resistance) to address the environmental source of contamination, on-site treatment is
necessary. As a rule, the implementation of water treatment has been less successful than the
development of water supply projects. Water supply is an obvious necessity for life, but
water treatment could involve substantial changes in behaviour to achieve often difficult-toperceive benefits. Water agency staff are typically better trained in managing repairs and
delivery schedules than in water treatment options (as noted by the Sukthankar Committee
2003; p 9 above). And community capacity building and education for water treatment
projects is, if anything, more important than that for water supply efforts.
In this section we focus on particle removing techniques and disinfection. Particle removal
(settling and filtration) is the first step towards drinking water acceptability, and many
disinfection methods will not work if the water is too turbid. Disinfection is next in
importance the WHO is clear in its recommendations that microbiological quality of water
should be given priority over the chemical quality of the water when the health budget does
not permit both (Gadgil 1998). Good sanitation practices and adequate methods to dispose
of human and animal waste are, of course, inextricably tied to microbiological quality. We
do not discuss methods of chemical removal, such as arsenic or fluoride treatment, because
most of the relatively inexpensive treatments, that could conceivably go to scale in rural and
peri-urban areas, are still in the field-testing phase.46
Particle removal techniques
Nivedita Banerji and Mihir Shah of SPS; personal communication.
The focus of most research and action in this area is now on arsenic removal in rural Bangladesh and West
Bengal, India. Several arsenic treatment technologies for developing country application, including coprecipitation, ion exchange and activated alumina filtration are now being field-tested. The household-based 2
Kolshi model of Bangladesh is under extensive testing. Another option is to bypass the groundwater altogether
with rainwater harvesting or protected pond systems. Many foundations, research groups, and multilateral
organizations are working on research and field testing in this critical area (WHO 2001; Murcott and Lukacs
2002).
45
46

Revised 8/29/05
38
Well-tested low-cost filtration approaches include sand filters and ceramic filters, both of
which have been used for hundreds of years. Slow sand filters incorporate both biological
and physical filtration properties. The BioSand filter developed by the Center for Affordable
Water and Sanitation Technology in Calgary (www.cawst.org) is a household sand filter unit
designed to be produced in developing countries. 20,000 BioSand filters have been installed
worldwide.47 Costs vary between $15-$30 per household unit. Flow rates are high, on the
order of 30 l/hour. The BioSand (or any sand filter) has not gained traction in India, but the
Dhan Foundation based in Tamil Nadu has just begun to promote these devices.
Ceramic filters, locally produced from clay using indigenous pottery techniques, provide a
physical barrier that prepares the water for disinfection, and may even be sufficient to
disinfect water alone. Ceramic filters such as the Terafil from Orissa, and the colloidal silver
enhanced filters (Filtron) developed in Guatemala, may cost only $5 - $10 for a household
unit. They are less durable than sand filters (their lifespan is 1 5 years) and they have low
flow rates (1 2 litres per hour) (Murcott and Lukacs 2002; Sobsey 2002). But they are easy
to use, and are easy to produce (and therefore a local supply chain can be maintained).
International Development Enterprises (www.ideorg.org) has piloted a 1000-household
effort to test and promote a locally-produced Filtron in Cambodia.
As of today, the thrust of NGO-led and commercially produced filtration techniques for
developing countries is on point-of-use (household) devices.48 Both the BioSand and the
Filtron are point-of-use filters, designed to be combined with an overall water delivery
network and intensive educational efforts aimed at improving water hygiene in marginalized
communities throughout the world.49 This makes them cheap, but it is generally harder to
cross-subsidize the poorest with, and to ensure the rapid adoption of, household-level
treatments. The intensive education advocated for the proper and continued use of point-ofuse devices is uncommon except for pilot projects. Regular cleaning and maintenance
activities (that ceramic filters may need) are often inconvenient.50 A comprehensive survey of
household preferences in Bangladesh indicated a strong preference for centralized filtering
and arsenic removal systems over household filtering technologies (Ahmad et al 2002).
Finally, the effectiveness of these filters depends on the quality of the source water in the
first place, so where the water is really turbid or coliform organism counts are really high,
sand and ceramic filters will be of limited value.
Disinfection techniques
Boiling and chlorination
Boiling water is a widely used, simple and reliable way to disinfect water. At sea level, boiling
water for one minute will effectively disinfect even turbid water. But energy costs of boiling
The specific countries are Haiti, Uganda, Honduras, Guatemala, Mexico, and the Dominican Republic.
Community-based slow sand filters are being tested in India by the CLEANIndia campaign
(http://www.cleanindia.org/jalpur_system1.htm). Improved slow sand filters (Jal-TARA), capable of
removing both particles and most pathogenic bacteria, and with a yield of 2000 3000 litres per day, have been
installed in schools, temples and villages. The Jal-TARA is very much a pilot project, and details on its use and
costs are not readily available.
49 Quoted from the literature of Potters for Peace (www.potpaz.org/pfpfilters.htm).
50 The inconvenience of daily treatment coupled with the inability of crude filtration methods to reduce many
diarrhoeas could be alternative reasons for the findings of Jalan, Somanathan and Chaudhuri (2003), cited
above in Section II.1.
47
48

Revised 8/29/05
39
water for the poor (who do not have access to, for instance, efficient gas stoves), are very
high. In a typical biomass cookstove, boiling water for an entire family just for drinking will
more than double the familys daily fuel needs.
Chlorine is the most widely-used chemical disinfectant in both the developed and developing
worlds. The amount of chlorine required for effective disinfection varies widely with water
characteristics including pH, temperature, particulate concentration, and common chemicals.
Thus it can be difficult to determine the dose required (overdosing has health dangers and
makes the water taste unpleasant). In-home disinfection with chlorine tablets has all the
advantages and disadvantages of point-of-use treatments discussed above. Community water
system treatments, in contrast, do not need such extensive education but do need trained
operators to administer the dose and test the water source. Chlorine degrades over time
(over weeks or months, depending on how it has been stored); and maintaining a supply
chain can be difficult in rural areas. However, the Barefoot College initiated RWH efforts
have used chlorine effectively to disinfect their community-level water storage. As a pointof-use example, the Centers for Disease Control (USA) have developed and tested the
common-salt based Safe Water System. By 2000, there were 1 million users of the SWS in 15
countries, with the supply chain funded and maintained by international donors,
multinational companies, NGOs, Ministries of Health and of Water, and door-to-door
salesmen and kiosk operators (Lukacs 2003).51
Ultraviolet light disinfection
Ultraviolet (UV) light with wavelength in the 240 280 nanometre range damages the DNA
of micro-organisms and is thus an effective water disinfectant. Compared to chlorine
disinfection, its disadvantages are that it imparts no residual protection to the treated water,
and needs a source of (grid or PV) electricity for the UV lamp. Its advantages are that it is
scale neutral with respect to cost, needs a very short contact time with the raw water (a few
seconds rather than 30 60 minutes) and over-dosing is not possible (Gadgil 1998).
Recent innovations for low-cost community-scale UV disinfection have been extensively
field-tested and are in use in Mexico, the Philippines and Uganda. A 40 watt UV lamp in a
unit such as Water Health Internationals UVWaterWorks disinfects water at a rate of 15
litres per minute (www.waterhealth.com) high enough to serve a community of about
2000. Energy requirements for UV disinfection are thus three to four orders of magnitude
less than that for boiling. With grid electricity, the cost of disinfection is approximately $0.10
per 10 litres of water, including the amortized cost of capital as well as all annual costs.
Although per-litre operating costs for UV systems are lower than those for chlorine
disinfection, the capital cost is higher about $1,500 for a basic UVWaterWorks unit in the
USA, and less for locally-manufactured alternatives.52 So the costs must be spread over many
community members. UV disinfection could be an especially good option for peri-urban
areas where populations are dense, and where piped supplies are non-existent or untreated.
As expected with these individual units, the CDC cautions: Maintaining long-term use of the SWS requires
sufficient resources to continue project activities such as social marketing, promotion and product
distribution. (www.cdc.gov/safewater)
52 This particular design is expensive because it is fail-safe and very low maintenance less expensive but less
reliable UV designs, for community use as well as individual use, and based on the original model, are also
being tested. The UV Tube, an in-home system that can be constructed using local materials, is being fieldtested in Sri Lanka and Mexico (http://ist-socrates.berkeley.edu/~rael/)
51

Revised 8/29/05
40
Small-scale vendors operating kiosks using UV systems could also provide a safer alternative
to existing vended, handpump, or standpipe supplies.
Burch and Thomas (1998) have estimated the average operating costs of different in-home
and small community-scale water disinfection methods in developing countries. Their cost
comparisons indicate that a filter plus chlorine-treated plant would cost $0.07 per m3; a filter
plus grid-powered UV would cost $0.03; a filter plus PV-powered UV water would cost
$0.15 per m3; chlorine for home treatment without prior filtration would cost $0.09; and
boiling with purchased fuel $20.83. (All costs are in 1998 US dollars, and indicate orders of
magnitude rather than precise costs).
We now discuss a case study of rural water purification using UV from the state of Guerrero
in Mexico.53 The arrangement between the private producer (WaterHealth Mexico), local and
also private maintenance engineer, the Department of Health in the state of Guerrero (who
paid the capital costs) and the rural community-based health clinics (who perform minimal
maintenance and track the health outcomes), offers an example of multi-institution
coordination that could be feasible for at least some Indian states. But the Guerrero case
also suggests potential problems of regulation, maintenance and cost recovery that are
instructive.
Guerrero, near the south central Pacific coast of Mexico, is one of its poorest states. In 1998
the Department of Health of the State of Guerrero purchased 58 community systems from a
WaterHealth Mexico (WHM) and installed them in dispersed rural communities 1999/2000.
The communities were largely without in-home piped water. WHM incorporated the basic
UVWaterworks into a six-step community level drinking water treatment plant, which
processes water from a protected spring or a bore well and delivers filtered and disinfected
drinking water for pick-up at the village health clinic. The total capital cost including
pumps, settling tanks, filtration units, the UV system itself and the tanks which store the
purified water was just over $3000. Each installation can serve 2000 2500 people daily.
The weekly maintenance (cleaning the filters and the holding tanks) is relatively simple and is
supposed to be carried out by a community member, often an employee of the health centre.
Quarterly maintenance such as changing the filters, cleaning the system and minor repairs is
carried out by the engineer employed by WHM. Annual maintenance (primarily checking the
system thoroughly and changing the UV bulb) is also the responsibility of WHM. The
Health Department pays for the maintenance costs (approximately $20 per community per
year). The users themselves collect the water for free. Though no rigorous before-and-after
studies of diarrhoea exist, anecdotal evidence and weekly clinic records suggest that the rate
of diarrhoea has fallen dramatically among children under age 5 in several (but not all) of the
UV villages. How much this decline is a result of these particular water systems, as opposed
to a secular decline in diarrhoeal disease in Mexico overall, is hard to gauge.
Unlike the other studies in this paper, there is no official documentation on the Guerrero case. All the
information in this section comes from visits to several Guerrero sites by one of the authors (Ray) and Dr.
Ashok Gadgil; data collected from the Department of Health authorities in Chilpancingo, Guerrero, by Dr.
Yvonne Flores; discussions with Dr. Mario Manjarrez in the clinic of Los Mogotes, Guerrero; a 2003 site visit
by a student team ( Sharon Karlsberg, Pablo Seminario, Jason Stein and Michelle Thomas) from Haas Business
School, UC Berkeley; and a survey of 22 sites by Anita Milman and Carol Kolb in 2004. We are grateful to all
of them for sharing their unpublished data.
53

Revised 8/29/05
41
By 2004, approximately one-third of the installations were still in use.54 The Guerrero sites
show that, where in-home connections are not an option but a communal source of water
exists, water community-scale water treatment systems can be installed and sustained with
workable collaborations between local governments, the private sector and community
members. However, three issues remain unresolved.
1.
There appears to be no formal mechanism for the Health Department to monitor and
regulate the maintenance performance of WHM and its engineers. As a result, quarterly
maintenance is irregular, especially in the more remote villages.55 Some of the Guerrero
communities are pro-active in maintaining their systems, but not all communities are able or
willing to do so.
2.
There is no mechanism for cost recovery from the users. The operating costs of the UV
systems are so low that they can easily be recovered. Cost-sharing may not be necessary at
the 60-village scale, but will almost certainly be necessary if any state decides to scale this (or
equivalent) programme up.
3.
Finally, many target users live too far from the safe water systems. The rate of use of
UV water is higher where populations are densely located close to the installation sites.
Access is therefore inconvenient for many households, though technically in accordance
with UNICEF guidelines. Several families buy bottled water at steep prices rather than walk
to the free but distant UV systems. The systems would undoubtedly be more effective if
there were e.g. two per village rather than one the costs shared by 1000 rather than 2000
people would still be low relative to the costs of frequent diarrhoea. Another option would
be to construct a small piped network from the system, terminating in shared water
collection points throughout the residential areas. This is how many rural piped water
schemes in India are designed. The water could even be sold at modest prices at these
collection points, partially paying for the operating costs of the system.
III.6 Some conclusions, some concerns
The literature on drinking water in India is characterized by an overall sense of policy failure
and barriers to access, punctuated with numerous examples of successful cases. But what is
each case a case of? Cases tend to be written up as examples of what their particular authors
are interested in of community participation, of private sector participation, of appropriate
technology, of willingness to pay, of equitable sharing among all users. It is therefore a
challenge to answer the key question for rural and peri-urban India i.e. how feasible is it to
scale up access to drinking water through decentralized technical and institutional options? In more
concrete terms, what are the lessons from these diverse experiences and how do we know
that these are the lessons? In this section we bring together the vast but varied literature to
highlight some recurring themes of relevance to an affordable, sustainable and replicable
drinking water programme. We also identify themes that do not recur, but should.
Sharing the costs
Sharing the cost of a new water system or of upgrading and maintaining an existing water
system is a consistent theme in the literature. Many NGO-led rainwater harvesting efforts
The other installations had either fallen into disrepair, or were no longer in use because more communities
had access to piped water or because there was no longer enough water at the source. It is interesting to note
that none of the unused or dysfunctional systems were so because of the lack of availability or high cost of
parts a common objection to water and sanitation systems that are not locally built.
55 This pattern of providing or (in this case) financing the supply mechanism but not following through on
repairs and maintenance is all too common in developing country contexts (World Bank 2004; p 74).
54

Revised 8/29/05
42
work only in those communities that are willing to contribute up to 20% of the capital costs,
in addition to 100% of the annual costs. The Prime Ministers drinking water initiative,
Swajaldhara, has made a 10% cost contribution from the community a condition for releasing
the remaining 90%. The World Bank has conducted and commissioned many willingness-topay studies that show that even poor people will pay for water if it is conveniently and
reliably supplied.
The cost-sharing mechanisms that both communities and lenders find acceptable will vary.
These mechanisms include micro-credit for infrastructure improvement (SEWA); operating
costs recovered through community contributions (Swajal; Kolhapur multi-village system);
and combined cash and labour contributions (SPS; Barefoot College). Some of these
arrangements are variants of Public Financing of Infrastructure, or PFI, whereby the
government finances essential infrastructure through providers who are more efficient than
the government itself (Morris 2004). Mukhopadhayay (2004) persuasively argues that the
governments existing water budget can go much further with creative PFI contracts.
Three questions remain under-explored in the literature on cost-sharing. First, how can the
poorest people or the poorest regions be cross-subsidized? Targeting specific populations or
specific areas is complex, and targeting methods have both direct and hidden costs for the
administration and for the poor (van de Waal 1998). Few studies on rural cost recovery
contain thoughtful discussions on cross-subsidization, though they admit that it is necessary
(WSP 1999a; p8).56 Second, what have been the public health impacts of specific cost-sharing
arrangements? With few exceptions, cost-sharing is treated as a goal in itself rather than as a
means to extending access for reasons of public health. And third, are parallel efforts under
way to withdraw urban subsidies over time? In a fine report on water services in India, the
Water and Sanitation Program claims that, People, including poor people, can afford and
are willing to pay for safe water: if they get a reliable supply; if they get the service levels and
other features they want; and if they are sure other people are paying their fair share too
(WSP 1999a; p11). This last consideration has, however, dropped out of most studies on
cost-sharing and cost-recovery.
Sharing the water
Relative to the rules of cost-sharing, the rules of water-sharing have not been given much
importance in the literature. Water supply connections for which users are willing to pay may
not assure access for all just because cost recovery has been assured. As mentioned earlier,
the more extensive the piped network for a given volume of water, the more difficult it is to
maintain water pressure. The larger the number of un-metered private connections, the
harder it is to ensure a minimum ration of 40 lpcd for all especially in times of scarcity. In
short, demand-led supply solutions have to take seriously the near-term growth of demand,
and to protect access to the national goal of 40 lpcd. But there is little evidence that they do
so at present.
One possibility is to have publicly funded community-level services (such as wells and yard taps) and recover
full costs from private services beyond that (meaning, no subsidies for the added cost of in-home
connections) (Briscoe and Garn 1995). This rule of thumb has not been followed in recent slum upgrading
initiatives.
56

Revised 8/29/05
43
As for community-scale water systems these are designed to be shared, and are therefore
common-pool resources (CPRs). Yet the rich body of literature on CPRs, much of it from
the irrigation sector in India itself, is all but ignored in the literature on drinking water. In the
absence of guidelines for sharing water and labour (not just costs), common-pool resources
become open access resources owned by everyone and therefore no one (Ostrom 1990;
Seabright 1993). Groundwater as well as public standpipes in India display this characteristic.
Empirical and theoretical work on CPR management has shown that shared resources can
be used efficiently and cooperatively under certain (necessary, not sufficient) conditions.
These include a small number of users, clear criteria for membership in the group of users,
simple rules of allocation that can be followed (and violations of which are easy to detect),
and repeated interactions among the users (Wade 1988). Standpipes that serve over 300
people or borewells that serve an entire village are unlikely to fulfil these conditions.
However, a water point that is shared by a small user group, living within a short distance of
the source, might be easier to maintain as a CPR. It would also be possible to devise and
monitor rules of water allocation during times of stress as the participatory irrigation
literature has shown (Datye and Patil 1987; Bardhan 2000). We do not assume that tubewells
and standpipes that are spatially located in accordance with CPR principles will actually be
maintained in good repair or will ensure water for all. However, we note that the Swajal
study above showed that there were on average eight handpumps per village, and the
Kolhapur study also emphasized the small size of the management unit. The implication
here is that shared water points in rural and peri-urban India should be more densely located,
certainly less than 1 km away from homes, despite the additional investment that that would
require. Effective low-cost options are not necessarily the cheapest options.
Community-based water management is hardly a panacea. In heterogeneous user groups the
resource is vulnerable to elite capture (Platteau and Gaspart 2003); proponents of CPR
management often romanticize community and ignore the larger political context (Mohan
and Stokke 2000); and there are dangers to cementing a two-tier system where the rich are
subsidised while the poor participate (Jaglin 2002). However, we have seen that in rural and
peri-urban areas common water sources are already the norm. Increasing the number of
collection points would make them more convenient to use and potentially more amenable
to sustainable management. Under conditions of water stress, a larger number of accessible
water points with guaranteed minimum rights might better ensure access for all than mixed
systems with in-home water for some plus inconveniently located shared water for the rest.
Institutional partnerships
There is remarkable consensus in the literature that governments should not be in the water
provision business but should be responsible for ensuring that the poor receive a basic level
of essential services (World Bank 2004). Governments should then ensure that private and
NGO providers are regulated and monitored with respect to price structures and water
quality, and should provide incentives for these providers to serve the poor. This new role
for government translates to developing partnerships with the private sector and with civil
society for water delivery. Such partnerships provide much-needed separation between the
service provider and the regulator.
Public-private partnerships (PPP) have received the most attention, and have worked well in
some rural and peri-urban areas (Chennai Water Board tanker vendors; Guerrero Health

Revised 8/29/05
44
Department Water Health Mexico). Public NGO partnerships (such as with SPS in
Madhya Pradesh) are also effective and may be more committed to serving the underserved
than traditional PPPs. Public private NGO partnerships are exemplified by SEWAs
housing improvement loans57 and the Indian handpump programme of the 1970s and 1980s.
Finally, as the central government devolves responsibilities to lower levels, as the 73rd
Amendment commits it to do, public public partnerships between different levels of
government (such as the Kolhapur scheme) may become increasingly common. All these
mechanisms are potentially viable in the Indian context.
Two aspects of this partnership mode of operation need further consideration. First, relative
to regular discussions of community capacity-building, empowerment etc, water agency staff
capacity-building is under-explored in policy documents. Yet most water agencies are not
geared towards decentralized and distributed services nor are they adept at working with
local populations and independent NGOs. This lack of capacity may well be a binding
constraint to extending access (Biswas 1996). Second, the consequences of devolving
responsibility to Panchayati Raj institutions for water and other services are very poorly
understood. Such decentralization is no guarantee of equity, efficiency or accountability
and in fact many PRIs lack the capacity to select local user groups, allocate funds and
organize village level services (Litvack, Ahmad and Bird 1998). Much more empirical
research and comparative institutional analysis are needed on both water agencies and
Panchayati Raj institutions in order to re-organize and deploy them effectively.
Scaling up
Physical constraints
Decentralized water projects on the same watershed or within a single catchment area are
linked by hydrology, so it is not feasible to scale them up in a decentralized way.58 Boreholes
with electric or diesel-operated pumps deplete the common aquifer unless their numbers and
spacing are regulated. Rainwater harvesting is considered environmentally low-impact -- even
beneficial -- because of its ability to recharge local aquifers. However, part of the water
captured at any one site would have, if un-captured, recharged downstream aquifers or reentered the ecosystem. The state of Rajasthans controversial claim to harvested water in
Alwar was unpopular with RWH proponents,59 but as watershed level interventions multiply,
negative third-party effects are inevitable. Preliminary water balance studies from Andhra
Pradesh and Karnataka suggest that RWH interventions have already impacted water use
patterns with positive consequences for some and negative consequences for others
(Batchelor et al 2002). A recent review of watershed programmes finds that almost no
attention has been paid so far to these unintended consequences (Joy and Paranjape 2004).
Property rights in this area remain unclear, or by default, belong to the state. We suggest that
the safest way to go to scale would be regularly to monitor basin-level water balances (which,
for example, the National Institute of Hydrology60 could spearhead) as new RWH projects
SEWA is, strictly speaking, a membership-based trade union rather than an NGO.
This cautionary point does not apply to water treatment interventions, only to water supply interventions.
59 See www.cseindia.org for details of the debate on who owns harvested rainwater does it belong to the
community whose labour created the reservoir or does the government own the rain?
60 NIH is Indias premier institute for hydrology and water resources research. Its 6 affiliated centres conduct
basin studies, flood and drought studies, water quality monitoring, hydrologic modelling, conjunctive use
57
58

Revised 8/29/05
45
are implemented or boreholes are deepened. Such monitoring would help to anticipate the
watershed level consequences of potentially thousands of small interventions, and to
understand where small-scale technologies can deliver the greatest benefits. Water is a State
subject in India, and with the passage of the 73rd Amendment, local water supplies are the
jurisdiction of local level governments. But state level data hoarding and uncoordinated local
decisions are not sustainable modes of operation where watershed boundaries do not
coincide with political boundaries.
Context constraints
A sustainable system is one that can continue to provide 30 - 40 lpcd of safe water for the
majority of its intended users into the foreseeable future. Well-intentioned pilot projects that
do not survive their donors withdrawal, and government funded pipes and pumps
projects61 that languish from subsequent neglect are, by definition, not sustainable. Scaling
up of sustainable technologies, institutional mechanisms and financing arrangement is critical
if universal access is to be achieved within a reasonable time period. However, generic
solutions are much easier to replicate than are context-sensitive solutions. The consensus
that participatory and context-specific approaches are the only sustainable ones would seem
to pose a problem for scaling up. Clearly not all communities today are capable of taking
charge of their water supply or treatment through the 90/10 financing model, or through
joint water management. Nor are many water agencies capable of adapting to the social and
economic contexts of their clients.
One of the more honest evaluations of the potential trade-off between scaling up and
sustainability is presented in Davis and Iyer (2002). The authors argue that though there are
numerous successful rural water supply projects, they do not go to scale because of human
and financial resource constraints; failure to incorporate the goal of scaling up into the
design of the pilot; resistance to scaling up by key stakeholders; and cherry-picking the site of
a typical pilot so that many real world problems do not have to be confronted. They
conclude and we concur -- that a set of detailed case studies, focusing on strategies to
overcome real world barriers, would be extremely helpful in bridging the gap between
scaling up and sustainability.
If faced with a choice between sustainability of water interventions and scaling up of
possibly unsustainable models, we would have to side with sustainability as the more
important goal. It may therefore be some time before universal access in rural and peri-urban
India becomes a reality.
Concluding remarks
Despite achievements in improving access to safer sources of water in the 1980s and 1990s,
the current system of household water provision in India is unable to provide convenient
and safe access to drinking water for a large sector of the population. Urban water delivery is
characterized by irregular delivery and pressure, provided by inefficient and heavily
subsidized municipal boards which are unable to maintain and expand the existing system.
studies, and GIS and remote sensing modelling for groundwater assessment. It is especially known for its work
on hard-rock aquifers. www.nih.ernet.in.
61 This is the term used by the late Anil Agarwal, Centre for Science and Environment, to describe government
funded water projects that focus on supply targets but ignore sustainability and community buy in.

Revised 8/29/05
46
The situation is somewhat worse in peri-urban and rural areas with declining water
availability, groundwater depletion and poor maintenance of existing infrastructure.
Substantial reform is needed in order to cater for existing needs, and to meet increasing
demand for water fuelled by both population and income growth. We have already discussed
potential avenues for reform in urban and rural areas, and here will only comment on a
couple of implications for further research and policy.
A major barrier to research and the design of appropriate policies is the lack of reliable, upto-date and publicly accessible information on many aspects of the Indian water system.
Baseline information is necessary in order to be able to evaluate various reforms in progress,
and in order to allow for benchmarking against Government targets, other states and nearby
countries. A lack of transparency over the true costs of underpriced and inefficient municipal
systems dampens public support for the major reforms needed. Incomplete and difficult-tofind information on groundwater withdrawals makes rural drinking water interventions
unsustainable. In this paper we have attempted to lay out the current state of knowledge
available from public sources in India, and in doing so, have revealed many areas where
more information is clearly needed. While efforts are underway to carry out some
benchmarking of financial performance of several large utilities, regular and comparable data
needs to be made available on, inter alia, water quality, subsidization, metering and broken
meters, groundwater levels, handpump maintenance, and other such information.
The past few years have seen some experimentation with private sector involvement in the
urban water sector and an expansion of decentralized community-based systems in rural
areas. We have provided examples of the types of reform efforts underway, but insufficient
time has passed to evaluate even the short term effects of many of these projects, while
longer term issues such as sustainability require even more time to analyse. Further research
is needed to study the effects of efforts underway, and to delve deeper into the institutional
conditions and political circumstances which allowed these limited reforms to take place.
The past few years have also seen many innovative technical and institutional solutions to
scarce drinking water or unsafe drinking water. Protecting the quality and sustainability of
harvested rainwater, allocating and monitoring access to community-based water systems,
and developing effective collaborations among local water agencies, NGOs, the private
sector and communities are all areas that deserve serious research support and study. Forbes
(2003) argues that, in a developing country with limited research resources, research
priorities should be guided by what is most useful and appropriable within the country of
origin. In particular he argues that some topics are critical to the well-being of developing
country citizens, but are of relatively little relevance to richer countries, so developing
countries cannot free ride on the international research commons. They have to pursue
those research areas themselves. Most of our suggestions for investigation and data
collection in Indias drinking water sector fulfil these criteria.
Finally, while our focus has been on drinking water delivery, over 90 percent of fresh water
in India goes to agriculture. Long-term expansion and sustainability of access to drinking
water will therefore require conservation and reallocation from the agricultural to the
domestic sector. Many of the issues discussed here in terms of cost recovery, institutional
coordination and common resources policy will also have relevance in research and policy
reform in this area.

Revised 8/29/05
47
REFERENCES
Ahmad, Junaid, BN Goldar, M Jakariya and Smita Misra (2002). Fighting Arsenic, Listening
to Rural Communities: Findings from a Study on Willingness to Pay for Arsenic-free,
Safe Drinking Water in Rural Bangladesh. World Bank, Water and Sanitation
Program-South Asia. Downloadable at www.wsp.org/pdfs/sa_arsenic_learning.pdf
Ahmed, Sara (1999). Changing Gender Roles in Irrigation Management : Sadgurus Lift
Irrigation Co-operatives, Economic and Political Weekly, Vol. XXXIV, No. 51.
Ahmed, Sara (2000). In Search of a Drop: Women and Drinking Water Management in
Rural Gujarat (India) - The Case of SEWA, Working Paper 137, Anand: Institute of
Rural Management [cited in Joy and Paranjape 2003]
AID (2003) Dishaa: The Association for Indias Development newsletter, June 2003August 2003, available online at
http://www.aidindia.org/hq/publications/pdf/dishaa/jun03.pdf [accessed April
2004].
Asian Development Bank (ADB) (1997). Second Water Utilities Data Book: Asian and
Pacific Region, edited by Arthur C. McIntosh and Cesar E. Yiguez. Asian
Development Bank: Manila
Bajpai, P. and Laveesh Bhandari (2001). Ensuring Access to Water in Urban Households.
Economic and Political Weekly, 29 Sept 2001: 3774-8.
Bardhan, Pranab (2000). Irrigation and cooperation: an empirical analysis of 48 irrigation
communities in south India. Economic Development and Cultural Change 48: 847866.
Barja, Gover, David J. McKenzie and Miguel Urquiola (2003) Capitalization and
Privatization in Bolivia: An Approximation to an Evaluation, Center for Global
Development Working Paper.
Batchelor, CA, A Singh, MSRM Rao and J Butterworth (2002). Mitigating the Potential
Unintended Impacts of Water Harvesting.
Biswas, Asit (1996). Capacity building for water management: Some Personal Thoughts.
Water Resources Development 12: 399 405.
Boland, John and Dale Whittington (2000) The Political Economy of Water Tariff Design
in Developing Countries, pp. 215-236 in Ariel Dinar (ed.) The Political Economy of
Water Pricing Reforms, Oxford University Press: New York.
Briscoe, John and Harvey Garn (1995). Financing water supply and sanitation under Agenda
21. Natural Resources Forum 19: 59 70.
Briscoe John (1995). When the cup is half full: improving water and sanitation services in
the developing world. Environment 35: 7 37.
Brocklehurst, Clarissa, Amrit Pandurangi and Latha Ramanathan (2002) Tariff Structures in
Six South Asian Cities, Water Tariffs and Subsidies in South Asia Paper 3, Water and
Sanitation Program-South Asia.
Burch J, and K Thomas (1998). An overview of water disinfection in developing countries
and the potential for solar thermal water pasteurization. Nat. Renew. EnergyLab. Rep.,
NREL / TP-550-23110. [cited in Gadgil 1998]
Carter, R. C. and P. Ball (2002). Private Sector Participation in Low Cost Water Well Drilling
in Africa. http://www.thewaterpage.com/cranfiel.htm
Centre for Science and the Environment (CSE) (2003). Analysis of Pesticide Residues in
Bottled Water [Delhi Region], CSE/PML-6/2002, online at www.cseindia.org.

Revised 8/29/05
48
Chaudhuri, Rajat (1998). Water: What are our rights to it? Safety Watch, Calcutta, India.
[cited in World Bank UNPD WSP October 1999]
Christmas, J and C. de Rooy (1991). The water decade and beyond. Water International 16:
127 134.
Clasen, Thomas (2003). Safe Household Water Treatment and Storage using Ceramic
Filtration. Downloaded April 2004 www.lshtm.ac.uk/dcrbu/ehg/tom_clasen.htm
Clarke, George R.G., Katrina Kosee and Scott Wallsten (2003) Has Private Participation in
Water and Sewerage Improved Coverage? Empirical Evidence from Latin America,
paper presented at CREDPR Latin America Conference, Stanford University,
November 2003.
Collignon, Bernard and Marc Vzina (2000). Independent Water and Sanitation Providers in
African Cities. Working Paper, UNDP-World Bank Water and Sanitation Program.
Washington, DC.
Crane, Randall (1994). Water markets, market reform and the urban poor: Results from
Jakarta, Indonesia, World Development 22: 71 83.
Datye, KR and RK Patil (1987). Farmer Managed Irrigation Systems: Indian Experiences. Centre
for Applied Systems Analysis in Development, Pune.
Davis, Jennifer and Parameswaran Iyer (2002). Taking Sustainable Rural Water Supply
Services to Scale: A Discussion Paper. Bank-Netherlands Water Partnership and
Water and Sanitation Program. Washington, DC.
Development Alternatives (2001) Too much, too little: The characteristics of water in India,
http://www.devalt.org/water/WaterinIndia/characteristics.htmw [April, 2004].
Downing, James and Isha Ray (2002). Safe Drinking Water for Everyone: A Review Paper
for the Blue Planet Run Foundation. ERG Working Paper. UC Berkeley.
Esrey S.A., J.B. Potash, L. Roberts, C. Shiff (1991). Effects of improved water supply and
sanitation on ascariasis, diarrhoea, dracunculiasis, hookworm infection,
schistosomiasis and trachoma. Bull. WHO 69: 609 621.
Esrey, S.A. (1996). Water, Waste, and Well-Being: A Multicountry Study. American Journal of
Epidemiology 143: 608-623.
Filmer, Deon and Lant Pritchett (2001) Estimating wealth effects without expenditure dataor tears: An application to educational enrollments in states of India, Demography
38(1): 115-32.
FIRE(D) (1999) Project Notes: The Tirupur Area Development Programme, Indo-US
Financial Institutions Reform and Expansion Project Debt Market Component,
Note No. 13.
Forbes, Naushad (2003). Higher education, scientific research and industrial
competitiveness: reflections on priorities for India. Fourth Annual Conference on
Indian Economic Policy Reform, June 6 7 2003, Stanford University.
Foster, Vivien, Subhrendu Pattanayak and Linda Stalker Prokopy (2003a) Do current water
subsidies reach the poor? Water Tariffs and Subsidies in South Asia Paper 4, Water and
Sanitation Program-South Asia.
Foster, Vivien, Subhrendu Pattanayak and Linda Stalker Prokopy (2003b) Can subsidies be
better targeted?, Water Tariffs and Subsidies in South Asia Paper 5, Water and Sanitation
Program-South Asia.
Gadgil, Ashok. 1998. Drinking water in developing countries. Annual Review of Energy and
Environment, v. 23, pp. 253 286.

Revised 8/29/05
49
Galiani, Sebastian, Paul Gertler and Ernesto Schargrodsky (2002) Water for Life: The
Impact of the Privatization of Water Services, CREDPR Working Paper No 154,
Stanford University.
Gleick, Peter. 1998. The Worlds Water 1998 1999. Island Press, Washington, DC.
Gleick, Peter. 2000. The Worlds Water 2000 - 2001. Island Press, Washington, DC.
Government of India, Ministry of Environment and Forests (2001) The State of Environment
India 2001, available online at http://envfor.nic.in [accessed January 2004]
Government of India, Ministry of Urban Affair and Employment (2001). Urban India. GoI:
New Delhi
Government of India (2001). Census of India 2001. GoI: New Delhi.
Government of India, Ministry of Water Resources (April 2002). National Water Policy. GoI:
New Delhi.
Government of India (2002 2007). Tenth Five Year Plan. GoI, New Delhi.
Harris, Clive (2003) Private participation in infrastructure in developing countries: trends,
impacts and policy lessons, World Bank Private Sector Advisory Services
Department Working Paper No. 5.
The Hindu (2004) BWSSBs plans to privatise water go awry, online edition, January 7,
2004, http://www.hindu.com/2004/01/07/stories/2004010711810300.htm
Jaglin, Sylvy (2002). The right to water versus cost recovery: participation, urban water
supply and the poor in sub-Saharan Africa. Environment and Urbanization 14: 231
245.
Jalan, Jyotsna and Martin Ravallion (2003) Does piped water reduce diarrhoea for children
in rural India? Journal of Econometrics 112: 153-173.
Jalan, Jyotsna, E. Somanathan and Saraswata Chaudhuri (2003) Awareness and the Demand
for Environmental Quality: Drinking Water in Urban India, Indian Statistical
Institute, Delhi, Planning Unit Discussion Paper 03-05.
Jolly, Richard (2003). Wanted: 3 million people alive. World Water Day message from
Chairman of the WSSCC, March 2003, www.wsscc.org [accessed April 2004].
Joy, KJ and Suhas Paranjape (2004). Watershed Development Review: Issues and
Prospects. Centre for Interdisciplinary Studies in Environment and Development,
Bangalore. January 2004.
Kerr, John, G Pangare and VL Pangare (2003). Watershed Development Projects in India:
An Evaluation. Research Report 127, International Food Policy Research Institute,
Washington DC.
Kjelln, Marianne (2000). Complementary water systems in Dar es Salaam, Tanzania: the
case of water vending. Water Resources Development 16: 143 154
Komives, Kristin (2000). Designing Pro-Poor Water and Sewer Concessions: Early Lessons
from Bolivia. Private Participation in Infrastructure; Private Sector Development
Division; World Bank, Washington D.C.
Litvack, Jennie, Junaid Ahmad and Richard Bird (1998). Rethinking Decentralization in
Developing Countries. Sector Studies Series, World Bank, Washington, DC.
Lukacs, Heather (2003). From appropriate to sustainable: Can we identify traits that
promote a technologys diffusion? CAWST Seminar presentation, Toronto 2003.
Mathur, Om Prakash (2001). Coming to grips with issues of pricing urban water and intracity bus transport, paper presented at India: Fiscal policies to accelerate economic
growth conference in New Delhi, May 21-22, 2001, http://www.fiscalconf.org/
Matzger, Helen and Marcus Moench (1994). Ground Water Availability for Drinking in
Gujarat. Economic and Political Weekly, March 26: A31 A41.

Revised 8/29/05
50
McIntosh, Arthur C. (2003) Asian Water Supplies: Reaching the Urban Poor, Asian Development
Bank and IWA Publishing: London.
McKenzie, David J. and Dilip Mookherjee (2003) The Distributive Impact of Privatization
in Latin America: Evidence from Four Countries, Economia 3(2):161-218.
Megginson, William L. and Jeffrey M. Netter (2001) From State to Market: A Survey of
Empirical Studies on Privatization, Journal of Economic Literature 39(2): 321-389.
MIDS (1995). Water Allocation and Management in the Chennai Metropolitan Area.
Madras Institute of Development Studies [cited in Ruet, Saravanan and Zrah 2002]
Mohan, Giles and K. Stokke (2000). Participatory development and empowerment: the
dangers of localism. Third World Quarterly, v. 21, pp. 247 268.
Morris, Sebastian (2004). Overview. Chapter 1 in 3iNetwork, 2004.
Murcott, S. and H. Lukacs (2002). Household Water Treatment in Nepal.
Mukhopadhayay, Partha (2004). In the Name of the Poor: What Role for PFI? Chapter 3 in
3iNetwork, 2004.
Nigam, Ashok and Gourisankar Ghosh (1995). A model of costs and resources for rural
and peri-urban water supply and sanitation in the 1990s. Natural Resources Forum
19:193 202.
Ninan, Ann (2003) Private water, public misery, CorpWatch India,
http://www.corpwatchindia.org/issues/PID.jsp?articleid=3623 [accessed 01/2004].
Noll, Roger G., Mary M. Shirley and Simon Cowan (2000) Reforming Urban Water Systems
in Developing Countries, pp. 243-294 in Anne O. Krueger (ed.) Economic Policy
Reform: The Second Stage, University of Chicago Press: Chicago.
Ostrom, Elinor (1990). Governing the Commons: The Evolution of Institutions for Collective Action.
Cambridge University Press, Cambridge.
Parker, Ronald and Tauno Skytta (2000). Rural Water Projects: Lessons from OED
Evaluations. World Bank Operations Evaluations Department, Working Paper
Series 3. Washington, DC.
Pitman, G.T. Keith (2002) India: World Bank Assistance for Water Resources
Management, A Country Assistance Evaluation, World Bank Operation Evaluation
Department Working Paper.
Platteau, Jean-Philippe and Frederic Gaspart (2003). Elite Capture Problem in Participatory
Development. Unpublished. University of Namur. [cited in World Bank 2004].
Raghupati, Usha P. and Vivien Foster (2002) Water: A Scorecard for India, Water Tariffs
and Subsidies in South Asia Paper 2, Water and Sanitation Program-South Asia.
Rastogi, Anupam (2004). The Infrastructure Sector in India. Chapter 2 in 3iNetwork, 2004.
Reddy, B. Sudhakara (2002). The Urban Challenge. Chapter 11 in India Development Report
2002. Oxford University Press: New Delhi.
Reddy, V. Ratna and MS Rathore (1993). Bias in Social Consumption: Case of Residential
Water in Rajasthan. Economic and Political Weekly, August 7 14: 1645 1648.
Ruet, Jol, VS Saravanan and M-H Zrah (2002). The Water and Sanitation Scenario in
Indian Metropolitan Cities. Occasional Paper, Centre de Sciences Humaines: New
Delhi.
Ruet, Jol (2004). Water in Urban India: The Scenario, Energy Linkage and Private
Participation. Chapter 21.4 in 3iNetwork, 2004.
Saleth, R. Maria and Ariel Dinar (1997) Satisfying Urban Thirst: Water supply augmentation
and pricing policy in Hyderabad City, India, World Bank Technical Paper No. 395.
Samaj Pragati Sahayog (SPS) (1999). Samaj Pragati Sahayog: An Introduction. Photo-essay
available from SPS, Bagli, District Dewas, Madhya Pradesh 455227, India.

Revised 8/29/05
51
Sasi, Anil and Sunil Jain (2004) Water Woes: Part III, A fifth of Indians happy with water
supply, Business Standard online, March 27, 2004, http://www.businessstandard.com/today/story.asp?Menu=19&story=37323 [accessed April 3, 2004].
Seabright, Paul (1993). Managing local commons: theoretical issues in incentive design.
Journal of Economic Perspectives 7: 113134.
Shah, Mihir, D. Banerji, PS Vijay Shankar and P Ambastha (1998). Indias Drylands: Tribal
Societies and Development through Environment Regeneration. Oxford University Press: New
Delhi.
Shingi, P.M. and V.N. Asopa (2002). Independent Evaluation of Checkdams in Gujarat: Strategies
and Impacts. Report: Centre for Management in Agriculture, Indian Institute of
Management, Ahmedabad.
Shirley, Mary M. (2002) Thirsting for Efficiency: The Economics and Politics of Urban Water System
Reform, Pergamon, Elsevier Science: Oxford.
Shiva, Vandana, Radha Holla Bhar, Afsar H. Jafri and Kunwar Jalees (2002) Corporate Hijack
of Water: How the World Bank, IMF and GATS-WTO rules are forcing water privatisation,
Navdanya: New Delhi.
Sobsey, M. G. 2002. Managing Water in the Home: Accelerated Health Gains from Improved Water
Supply. Geneva, World Health Organization.
Sridhar, Lalitha (2003a), Piped dreams, www.indiatogether.org, accessed 2 Feb 2004.
Sridhar, Lalitha (2003b), Water: the privatization debate,
http://www.indiatogether.org/2003/nov/env-wtrdebate.htm [accessed March
2004].
Sukthankar Committee (2001) Sukthankar Committee Report on Operation, Maintenance
and Management of Rural and Urban Water Supply Scheme submitted to the
Government of Maharashtra, Water and Sanitation Department, Government of
Maharashtra.
Talbot, Rupert (1997). Goals, Roles and Innovations in India. 23rd WEDC Conference on
Water and sanitation for all: partnerships and innovations. Durban, South Africa.
Downloadable from http://wedc.lboro.ac.uk/conferences/pdfs/ 23/Talbot.pdf
3iNetwork (2004) India Infrastructure Report 2004: Ensuring value for money. Oxford University
Press: New Delhi.
UNICEF (1995) Strategies in Water and Environmental Sanitation. New York, UNICEF.
United Nations Human Settlements Programme (UN-HABITAT) (2003) Water and Sanitation
in the Worlds Cities: Local Action for Global Goals, Earthscan Publications Ltd: Sterling,
VA.
Van de Waal, Dominique (1998). Targeting Revisited. The World Bank Research Observer 13:
231 248.
Van der Hoek, Wim, Flemming Konradsen, Jeroen H. J. Ensink, Muhammad Mudasser and
Peter K. Jensen (2001). Irrigation water as a source of drinking water: is safe use
possible? Tropical Medicine and International Health, 6: 46 54.
Wade, Robert (1988). Village Republics. Cambridge University Press, Cambridge.
Water and Sanitation Program (WSP) (1999a). Water for Indias Poor: Who Pays the Price
for Broken Promises? Report, October 1999. UNDP-World Bank Water and
Sanitation Program-South Asia http://www.wsp.org/
Water and Sanitation Program (WSP) (1999b). Nagari Willing to Pay but Unwilling to
Charge: Do Willingness-to-pay Studies Make a Difference? Field Note, June
1999, UNDP-World Bank Water and Sanitation Program-South Asia.

Revised 8/29/05
52
Water and Sanitation Program (WSP) (1999c). Sustainable Community Management of a

Multi-village Water Supply Scheme in Kolhapur, Maharashtra, India. Field Note,
December 1999. UNDP-World Bank Water and Sanitation Program-South Asia.
Water and Sanitation Program (WSP) (2000a). Afridev Handpumps in Pakistan: Private
demand emerges for an affordable and reliable pump. Case study, December 2000.
UNDP-World Bank Water and Sanitation Program-South Asia.
Water and Sanitation Program (WSP) (2000b). The West Bengal Pilot Project: Responding
to community demands for safe drinking water in an arsenic affected area. Field
Note, August 2000. UNDP-World Bank Water and Sanitation Program-South Asia.
Water and Sanitation Program (WSP) (2001a) Nagari Thirteenth Meeting of the Urban
Think Tank: Private Sector Participation in Urban Water and Sanitation: Managing
the Process and Regulating the Sector, December 5-6, 2001, UNDP-World Bank
Water and Sanitation Program-South Asia.
Water and Sanitation Program (WSP) (2001b). Community Contracting in Rural Water and
Sanitation: The Swajal Project, Uttar Pradesh, India. Field Note, June 2001. UNDPWorld Bank Water and Sanitation Program-South Asia.
Water and Sanitation Program (WSP) (2002) Nagari, Fourteenth Meeting of the Urban
Think Tank: Willingness to Charge for Water and Sanitation Services: Outcomes
from Three Workshops, UNDP-World Bank Water and Sanitation Program-South
Asia.
Water and Sanitation Program (WSP) (no date). Credit Connections: SEWA Bank. UNDPWorld Bank Water and Sanitation Program-South Asia & Department for
International Development (DfID).
World Bank (2000a) India: Country framework report for private participation in infrastructure, The
World Bank: Washington D.C.
World Bank (2000b). World Development Indicators. The World Bank: Washington DC.
World Bank (2004). Making Services Work for Poor People. World Development Report 2004.
The World Bank: Washington DC and Oxford University Press: New York.
World Health Organization and UNICEF (WHO-UNICEF) (2001) WHO/UNICEF Joint
Monitoring Programme for Water Supply and Sanitation Coverage Estimates 19802000: Access to Improved Drinking Water,
http://www.wssinfo.org/en/pdf/country/india_water1.pdf [accessed April 3, 2004].
WHO (1993) Guidelines for drinking-water quality: Volume 1, Recommendations. 2nd edition, World
Health Organization, Geneva.
WHO (1997) Guidelines for drinking-water quality: Volume 3, Surveillance and control of community
supplies. 2nd edition, World Health Organization, Geneva.
WHO (1998) Guidelines for drinking-water quality: Addendum to Volume 2, Health criteria and other
supporting information. 2nd edition, World Health Organization, Geneva.
World Health Organization (WHO) Fact Sheet 210, May 2001. Arsenic in Drinking Water.
Downloadable from www.who.int
Zaroff B, and Daniel A Okun (1984). Water Vending in Developing Countries. Aqua,
International Water Supply Association (UK) 5: 289 295.
Zrah, Marie-Hlne (2000) Water: Unreliable Supply in Delhi. Manohar, Centre de Sciences
Humaines, Delhi.
Table 1: Source of Household Drinking Water

Percentage of Households Receiving their Drinking Water from each source
All India
Source of drinking water
Piped into residence/yard/plot
Public tap
Handpump in residence/yard/plot
Public handpump
Private Wells
Well in residence/yard/plot
Open well
Public Wells
Public covered well
Public open well
Spring
River, stream
Pond, lake
Dam
Rainwater
Tanker truck
Other
Water from improved source (excluding wells)
Water from improved source
(including wells in residence and covered wells) 1
1992-93
18.4
14.8
13.6
21.6
7.2
Urban
1992-93
48.1
21.5
9.6
8.5
4.7
0.7
1.9
1.2
0.1
0.0
0.4
1.6
1998-99
21.0
17.6
15.4
23.8
7.4
1.2
6.2
11.4
1.0
10.4
0.5
1.2
0.8
0.1
0.0
0.2
0.5
68.3
71.0
18.6
Rural
1992-93
7.1
12.2
15.1
26.5
8.2
0.1
0.5
0.3
0.0
0.0
1.0
1.1
1998-99
51.6
22.7
9.9
8.3
3.8
0.7
3.1
2.2
0.3
1.9
0.1
0.1
0.3
0.0
0.0
0.3
0.7
77.9
87.7
80.0
89.2
4.6
Small City/Town
1992-93
43.3
20.7
12.1
8.5
6.3
0.9
2.5
1.5
0.1
0.0
0.1
1.8
1998-99
9.3
15.7
17.5
29.7
8.7
1.4
7.4
14.8
1.2
13.6
0.7
1.6
1.0
0.1
0.1
0.1
0.5
92.5
60.9
93.5
64.1
23.9
Large Cities
1992-93
57.6
23.1
4.6
8.5
1.6
0.1
0.7
0.4
0.0
0.0
0.9
1.4
1998-99
45.3
24.9
11.7
8.9
4.7
0.8
3.9
2.9
0.4
2.5
0.1
0.1
0.4
0.0
0.0
0.4
0.7
0.0
0.3
0.1
0.0
0.0
1.2
0.5
1998-99
69.1
16.9
4.9
6.5
1.5
0.4
1.1
0.4
0.1
0.3
0.1
0.0
0.0
0.0
0.0
0.2
0.5
72.3
84.7
90.8
93.7
97.3
74.9
86.5
92.0
94.6
97.8
5.5
2.7
Source: Population weighted values calculated from the National Family Health Survey
Urban areas consist of large cities, small cities and towns.
1. 1992-93 figures calculated assuming that wells in residence constitute same proportion of private wells as in 1998-99, and that public covered wells constitute
the same proportion of public wells as in 1998-99.
Table 2: Source of Drinking Water by State 1998-99

Percentage of Households Receiving their Drinking Water from each source
STATE
Andhra Pradesh
Assam
Bihar
Goa
Gujarat
Haryana
Himachal Pradesh
Jammu
Karnataka
Kerala
Madhya Pradesh
Maharashtra
Manipur
Meghalaya
Mizoram
Nagaland
Orissa
Punjab
Rajasthan
Sikkim
Tamil Nadhu
West Bengal
Uttar Pradesh
New Delhi
Arunachal Pradesh
Tripura
Piped
Public
Water
43.8
35.9
26.4
62.9
75.6
63.7
71.3
87.9
55.6
17.8
43.8
68.9
35.5
40.8
47.6
26.6
21.0
67.6
78.5
89.6
35.8
36.0
38.8
78.7
50.6
25.1
Tap
42.9
11.8
8.3
12.1
14.0
8.1
19.4
8.3
35.1
19.8
28.9
26.2
27.7
42.6
25.3
9.8
25.9
2.0
10.8
2.4
37.2
30.0
4.1
8.9
25.3
26.7
URBAN AREAS
Private
Public
handpump handpump
1.9
6.8
30.0
4.8
35.4
15.0
0.2
0.5
2.8
3.6
22.5
5.2
1.6
3.2
0.8
0.5
2.0
3.5
1.3
0.9
4.8
13.6
0.3
1.6
0.9
2.0
0.4
0.4
0.3
1.4
3.6
1.2
5.7
28.5
29.6
0.7
1.9
4.3
0.6
0.6
5.3
10.3
6.5
19.6
39.6
15.6
8.6
2.5
8.6
10.3
28.3
8.8
Source: Own Calculations from National Family Health Survey 1998-99
Wells
3.3
16.9
14.2
23.9
0.9
0.5
4.4
0.7
3.8
60.0
8.1
2.3
1.9
10.6
11.4
39.6
17.6
0.2
2.8
0.0
6.9
6.6
1.7
0.2
1.1
11.2
Springs, rivers,
streams &
ponds
0.0
0.2
0.3
0.2
0.1
0.0
0.2
1.8
0.1
0.0
0.1
0.0
28.0
5.2
12.9
19.2
1.1
0.0
0.8
6.7
2.1
0.0
0.1
0.0
1.1
0.0
Piped
Public
Water
9.3
3.3
2.8
33.5
37.8
11.2
26.6
27.4
10.4
3.6
5.8
22.5
8.3
8.2
1.5
14.1
0.9
16.4
16.0
79.9
9.1
5.1
2.8
65.3
16.8
4.1
Tap
30.7
3.8
1.6
18.5
17.9
24.5
39.1
19.7
44.7
7.8
5.1
24.8
23.7
19.1
46.0
26.0
3.0
5.3
12.4
2.4
55.7
5.4
2.7
10.3
48.3
18.0
RURAL AREAS
Private
Public
handpump handpump
2.5
30.4
29.0
21.5
37.8
32.0
0.0
0.0
2.9
17.3
24.5
22.4
1.4
8.0
10.1
5.2
1.7
25.0
1.4
0.8
2.6
40.6
1.4
21.6
1.4
7.5
3.7
0.8
0.0
2.7
0.1
0.1
5.2
54.3
72.2
4.5
3.9
28.8
0.8
0.0
1.9
16.4
19.1
58.6
46.1
30.6
18.8
5.1
7.5
5.4
8.5
24.6
Springs, rivers,
Wells
22.9
30.4
23.5
41.3
17.4
17.0
18.6
11.8
16.2
83.6
42.0
26.1
4.6
34.0
22.0
22.4
27.8
1.3
32.5
0.5
13.3
9.9
15.7
0.5
9.6
35.8
streams & ponds

3.5
11.7
1.6
6.6
3.8
0.1
5.7
25.2
1.9
1.1
3.8
2.5
54.2
33.8
26.9
37.0
8.7
0.1
4.6
15.7
3.3
0.5
2.0
0.0
12.2
6.0
Table 3: Access to Piped Water by Asset Deciles

Percentage with Access by relative asset levels
Location-specific
Asset Decile
1
2
3
4
5
6
7
8
9
10
All India
2.8
3.0
5.8
8.9
13.9
19.1
29.5
40.8
54.2
71.3
Urban
15.9
31.0
34.9
43.3
53.7
60.5
68.1
77.1
81.2
83.7
Rural
2.4
5.5
3.2
4.6
6.9
10.3
13.1
19.4
22.9
35.7
Small Cities
& Towns
12.3
23.4
28.4
35.6
45.4
48.7
60.0
66.8
74.9
78.3
Large
Cities
26.4
42.1
50.2
60.2
66.3
78.3
85.1
86.5
90.0
88.6
Small Cities
& Towns
23.3
1.4
14.2
12.3
0.7
53.4
0.6
Large
Cities
32.0
1.3
17.9
17.4
2.2
39.6
0.7
Notes: Asset Deciles Based on First Principal Component for location

Source: NFHS 1998-99
Table 4: Time taken to get to Water Sources 1998-99
for households without water on their premises
Time taken
5 minutes or less
6-10 minutes
11-20 minutes
20 minutes or more
Percentage of Households in
Rural
Urban
34.3
25.5
25.7
26.7
18.8
22.4
21.2
25.5
Source: Own calculations from NFHS 1998-99

Table 5: Methods of Purifying Water 1998-99
percentage of households using given method
Purification Method
Strain with cloth
Alum
Water filter
Boil
Electronic purifier
Nothing
Other
All India
18.5
1.3
5.9
8.1
0.4
68.5
0.7
Urban
25.5
1.4
15.2
13.6
1.1
49.9
0.6
Source: Own calculations from NFHS 1998-99
Rural
15.9
1.2
2.5
6.1
0.1
75.3
0.8
Table 6: Indicators of Irregular Supply and Wastage

City
Source
4
year
2002-03
5
4
1
2
2004
2002-03
1997
1993
6
1
2001
1997
1
3
2
Mumbai
Bangladore
Calcutta/Kolkata
Chennai
Hours per day

of water supply
4
9
6 to 7
10
Unaccounted
for water (%)
34-44
Metering (%)
100
30-40
50
36
Staff per 1000

connections
8
Cost recovery
(% of operating costs)
95
Accounts
receivables (months)
5
very low
0
14
17.1
15
19
1.5
20
<5
1
25.9
106
5.8
1997
1995
1993
4
5
26
73
21.4
68
4.5
1
2
1997
1993
18
24
67
33.3
93
19.7
Hyderabad
2002-03
0.5 to 4
33
13
66
Comparison cities
Lahore
Kathmandu
Bangkok
Beijing
Asian-Pacific average
1
1
1
1
1
1997
1997
1997
1997
1997
17
6
24
24
19
40
40
38
8
35
5.7
15
4.6
27.2
11.8
141
139
112
77
95
Delhi
30
Notes: cost recovery is annual billing as percentage of operation and maintenance costs
Sources:
1: Asian Development Bank (1997)
2. Zrah (2000) reports of Asian Development Bank 1993 figures
3. Zrah (2000) reports of her own 1995 survey.
4. WSP (2002)
5. Own communication with Kolkata Municipal Corporation
6. Brocklehurst, Pandurangi and Ramanathan (2002)
24
83
100
100
83
7
4.5
2
0.1
4
Table 7: Water Tariffs in Several Large Indian Cities
Metered Connections:
Hyderabad 1993
Consumption
Rate
Slab (Kl/month)
(Rs/Kl or Rs)
Up to 15
40 (flat rate)
15-25
3
25-500
5
over 500
10
Hyderabad 2003-04
Consumption
Rate
Slab (Kl/month) (Rs/Kl or Rs)
Up to 30
6
30-200
10
200-500
25
over 500
25 on all units
Minimum monthly tariff
90
Bangladore 2003-04
Consumption
Rate
Slab (Kl/month) (Rs/Kl or Rs)
Up to 15
6
15-25
8
25-50
12
50-75
30
over 75
36
90
Unmetered connections
all
120 (flat rate) 15mm pipe

20mm pipe
25mm pipe
40mm
50mm +
90 (flat rate)
270 (flat rate)
600 (flat rate)
1500 (flat rate)
3200 (flat rate)
Metered Connections:
Delhi
Consumption
Slab (Kl/month)
Up to 10
10-20
20-30
over 30
1995-96
Rate
(Rs/Kl or Rs)
0.46
0.46
0.91
0.91
Chennai 2003-04
2003-04
Rate
Consumption
Rate
(Rs/Kl or Rs) Slab (Kl/month) (Rs/Kl or Rs)
0.53
Up to 10
2.5
1.50
10-15
10
2.25
16-25
15
4.50
over 25
25
1997-98
Rate
(Rs/Kl or Rs)
0.53
1.50
2.25
4.50
Minimum monthly tariff

Unmetered connections
all
9.75 (flat rate)
30
30
30 (flat rate)
30 (flat rate)
50
all
50 (flat rate)
Notes:
Flat rates are monthly rates in Rs. $1US = Rs 43 (April, 2004)
Sources:
Hyderabad 1993 rates are from Saleth and Dinar (1997) and 2003-04 rates are from
Hyderabad Metropolitian Water Supply and Sewerage Board (MWSSB), http://www.hyderabadwater.gov.in/
Bangladore rates from Bangladore Water Supply and Sewerage Board (BWSSB), http://www.bwssb.org/
Chennai rates from Chennai Metropolitan Water Supply and Sewerage Board (CMWSSB), http://www.chennaimetrowater.com/
Delhi 1995-96 and 1997-98 rates are from Zrah (2000, p. 74) and include 30 percent and 50 percent surcharges
Delhi 2003-04 rates from Delhi Jal Board (DJB), http://www.delhijalboard.com [accessed April 2004]
Table 8: Comparison of Metered Prices of First 50 Kl 2003-04
Kl per month
10
20
30
40
50
Hyderabad
90
120
180
280
380
Bangalore
90
130
230
350
470
Chennai
50
150
350
600
850
source: own calculations from block and minimum tariffs in Table 7.
Delhi
30
30
42.8
87.8
132.8
Data
% with Piped Water by State, 1998

(weighted)
STATE water
Andhra Pradesh
Assam
Bihar
Goa
Gujarat
Haryana
Himachal Pradesh
Jammu
Karnataka
Kerala
Madhya Pradesh
Maharashtra
Manipur
Meghalaya
Mizoram
Nagaland
Orissa
Punjab
Rajasthan
Sikkim
Tamil Nadhu
West Bengal
Uttar Pradesh
New Delhi
Arunachal Pradesh
Tripura
18.24
6.28
5.26
44.91
53.97
27.39
31.79
41.84
26.54
6.79
15.69
41.75
15.43
14.39
26.97
15.89
3.08
31.64
31.50
80.81
18.47
11.92
10.37
77.64
23.55
7.20
Per capita Net State Domestic Product at Current Prices

SDP
13993
8826
4474
40248
18815
19716
16144
11591
15420
16029
10682
20356
10504
11090
13479
12408
8324
21184
12348
13158
17613
13641
8633
33870
13129
11012

Water PDF

Uploaded by

Copyright:

Available Formats

Water PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Water PDF

Uploaded by

Copyright:

Available Formats

STANFORD CENTER FOR INTERNATIONAL DEVELOPMENT

Working Paper No. 224

Updated December 2005

Development Research Group, The World Bank

McKenzie & Ray; Drinking water options for India

HOUSEHOLD WATER DELIVERY OPTIONS IN URBAN AND

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

Source: Department of Drinking Water Supply, Ministry of Rural Development, http://ddws.nic.in

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

source: The Times of India, January 5, 2004 Water watch,

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

The argument for private sector participation

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

Sample sources: http://www.water-technology.net/projects/tirupur/; Government of Tamil Nadu

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

We have found cost comparisons of technical and institutional alternatives extremely

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India

McKenzie & Ray; Drinking water options for India