Report No. 18114 India Impact Evaluation Report Comparative Review of Rural Water Systems Experience Rajasthan Water Supply and Sewerage Project (Credit 1046-IN) Maharashtra Rural Water Supply and Environmental Sanitation Project (Credit 2234-IN) Karnataka Rural Water Supply and Environmental Sanitation Project (Credit 2483-IN) June 30, 1998 Operations Evaluation Department Document of the world Bank Currency Equivalents Currency unit = Indian Rupees (Rs.) June 1980 US$1.00 Rs 8.40 March 1991 US$1.00 Rs 17.80 March 1993 US$1.00 RS 31.00 December 1977 US$1.00 Rs. 39.00 Abbreviations and Acronyms ERR economic rate of return FYDP Five-year Development Plan GNP gross national product GOI Government of India IDA International Development Association IER Impact Evaluation Report LGU Local government unit MWSSB Maharashtra Water Supply and Sewerage Board MRD&PR Ministry of Rural Development and Panchayat Raj (Karnataka) lcd liters per capita per day NGO nongovernmental organization OED Operations Evaluation Department O&M operation and maintenance PHED Public Health Engineering Department RWS rural water supply W&S water supply and sanitation Director-General, Operations Evaluation Mr. Robert Picciotto Director, Operations Evaluation Department Ms. Elizabeth McAllister Acting Manager, Sector and Thematic Evaluation Group Mr. Roger Slade Task Manager Mr. Tauno Skytta FOR OFFICIAL USE ONLY The World Bank Washington, D.C. 20433 U.S.A. Office of the Director-General Operations Evaluation June 30, 1998 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Impact Evaluation Report on India Comparative Review of Rural Water Systems Experience Rajasthan Water Supply and Sewerage Project (Credit 1046-IN) Maharashtra Rural Water Supply and Environmental Sanitation Project (Credit 2234-IN) Karnataka Rural Water Supply and Environmental Sanitation Project (Credit 2483-IN) More than 70 percent of India's 990 million people live in half a million rural villages where water-borne diseases are a major health problem. To reduce this problem, over the past two decades India has implemented major investment programs in rural water supply and sanitation (RWS). World Bank lending for RWS in India grew from a total of approximately US$120 million in the 1980s to about US$260 million in the 1990s. Three IDA-funded RWS projects have been completed and two are ongoing. Two of the completed projects, Rajasthan Water Supply and Sewerage (Cr. 1046-IN) and Maharashtra Rural Water Supply (Cr. 2234-IN), and one ongoing project, Karnataka Water Supply and Environmental Sanitation (Cr. 2483-IN), are the subjects of this Impact Evaluation Report (IER). The OED study was done in partnership with the Center for Institutional Reform and the Informal Sector (IRIS) at the University of Maryland. IRIS staff carried out the data analysis, contributed sections to the report, and prepared a background paper on social capital. The evaluation was conducted in two phases. During a pilot phase the study team reviewed available background data, developed and field tested the data sheets and questionnaires, elaborated household selection criteria, identified the appropriate sample size and composition, and conducted pilot focus group and community interviews. In the second phase, the study fielded teams of local researchers to administer questionnaires on the household and village levels (in two states, Karnataka and Maharashtra), collect secondary data, and fill out technical data sheets on each district and village visited and the operations of its water system. Data were collected from 60 Indian communities (50 project villages and 10 control villages) through a survey of about 1,100 households. Villages were randomly selected from among those that had been in operation a minimum of six months. In Karnataka and Maharashtra the team also conducted a technical evaluation. In Rajasthan the study team only conducted two simultaneous participatory workshops for stakeholders. In all three states the study team met with state and district officials and visited facilities constructed by the projects. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. -2- The evaluation shows that when pre-project social capital levels are high, communities generally have better performing schemes. The economic benefit of time saved in collecting water is estimated to produce a potential increase in household income of 30 percent. Finally, the impact on public health in those communities with the best-performing schemes is stronger than in communities with the worst- performing ones. The evaluation also finds that Bank support for the rural water supply schemes in India is significantly improving the quality of life in project villages. The rural water infrastructure is in place and functioning, at a per capita cost that compares with cost indicators from other developing countries. The achievements and impacts of the projects are considerable. Time required for water collection has been reduced by an average of 85 percent, system operation has improved about 50 percent, and system hours of operation have become much more regular, and the quality of water has improved. Over 80 percent of consumers in Karnataka and over 50 percent in Maharashtra rate the design of their new systems satisfactory. Sanitary conditions have improved substantially. While the various schemes are helping to reduce the deficit in rural water supply and increasing water coverage, they have done little to foster local organizational capacity. The results of this study show that involvement of users in the development, implementation, and O&M of water supply systems has not yet been sufficient to achieve a desirable level of sustainability. Women, the primary water collectors in most rural households, have the most interest in ensuring that the water and sanitation service is designed to match their needs and performs well, and they need to be more involved than has been the case to date. The study finds that project designers need to adjust the approach to service delivery according to existing levels of social capital. Special efforts may be necessary to motivate and mobilize community members if water systems are to operate sustainably in weaker communities. Besides the general lessons generated by the study (inter alia on the importance of adequate social mobilization, project rules, and hygiene education; and the efficacy and effectiveness of the community-based.approach), policy recommendations include the advisability of respecting beneficiary aspirations for appropriate levels of service. Factors such as participation in multiple community activities by water committee members, participation in system design, participation in system construction, improved beneficiary health, and satisfactory participation in the water group by women are all positively related to above-average system performance and should therefore receive special attention during water and sanitation project design. 1 Contents P refa ce ..............................................................................................................................................3 1. Study M ethodology and Project Context............................................................................. 5 The Impact Evaluation Study........................................................................................... 5 Country Context.................................................................................................................. 5 Rural Water Supply in India before the Project: One Safe Source per Village.............. 5 State-specific Pre-project Conditions: W ater and Sanitation.......................................... 7 GOI Commitment to the W ater Sector and Strategy........................................................ 8 Study M eth od s .................................................................................................................... 9 Questions Studied............................................................................................................. 10 2. Evolution of the Rural W ater Paradigm in India............................................................. 11 RW S in the India Portfolio............................................................................................... 1 Rajasthan W ater Supply and Sewerage Project............................................................... 11 Kerala W ater Supply and Sanitation Project.................................................................. 13 The Karnataka and M aharashtra Projects........................................................................ 13 The M ost Recent Project.................................................................................................. 15 The Evolving Paradigm.................................................................................................... 16 3. Project Achievements............................................................................................................ 16 Village Surveyed.............................................................................................................. 17 Coverage Improvements................................................................................................... 18 Pre- and Post-Project Access........................................................................................... 18 Functioning of Schemes................................................................................................... 19 Cost of W ater Service....................................................................................................... 20 Tariffs and Cost Recovery................................................................................................ 22 Institutional Achievements............................................................................................... 22 Sustainability of Service................................................................................................... 25 Social Capital and Institutional Performance ................................................................. 25 4. Project Impacts......................................................................................................................27 Health Impacts .................................................................................................................. 28 Economic Impact.............................................................................................................. 32 S o cial Im p act..................................................................................................................... 3 5 Institutional Impact.......................................................................................................... 36 This study was conducted under the supervision of Mr. Tauno Skytta, Task Manager. Mr. Ronald Parker (Evaluation Officer) prepared the report. Ms. Reno Dewina and Mr. Gary Wu (Consultants) provided research support. Ms. Satu Kahkonen and Mr. Jonathan Isham of the Center for Institutional Reform and the Informal Sector (IRIS) at the University of Maryland, College Park carried out the data analysis, contributed sections to the report, and prepared a background paper on social capital. The survey was carried out by ORG-MARG in India. Mr. William Hurlbut edited the report, and Ms. Helen Watkins provided administrative support. 2 5. Findings, Conclusions and Recommendations ................................................................. 37 F in din g s .............................................................................................................................3 7 C on clu sion s .......................................................................................................................3 8 R ecom m endations............................................................................................................. 40 Annexes 1. Objectives and Components of the Surveyed Projects................................................... 43 2. Perform ance indicators: India.......................................................................................... 47 Text Tables 3.1 Status of Randomly Selected Schemes............................................................................ 17 3.2 Improved Access to Service, Time Saving..................................................................... 19 3.3 M ost Com m on Schem e Problem s.................................................................................... 19 3.4 Improved Functioning of Schemes: Service Hours & User Satisfaction...................... 20 3.5 System Com ponents and Costs........................................................................................ 21 3.6 Effectiveness of Village-level Institutions...................................................................... 22 3.7 Training and Capacity-Building Achievements.............................................................. 24 3.8 V illage Participation Indicators....................................................................................... 24 3.9 Selected Indicators of Social Capital in Maharashtra and Karnataka........................... 26 3.10 Comparing Group Means for the Best and Worst Performing W&S Services............ 27 4.1 Diarrhea: Incidence, Treatment Sought and Mortality in Indian Project Sites............ 28 4.2 Improved Health and Current Water Source................................................................... 30 4.3 Time Saving and Current Source of Water...................................................................... 34 Boxes 1.1 P roblem V illages................................................................................................................. 6 2.1 Project R ules: M aharashtra ............................................................................................... 14 2.2 Project R ules: K arnataka................................................................................................... 15 4.1 Water Supply Improvement Caused an increased Density of Rural Areas in Kerala.. 34 Figures 4.1a Improved Health and Previous Water Source: Karnataka.............................................. 29 4. lb Improved Health and previous Water Source: Maharashtra............................................ 29 4.2a Increased Water Consumption and Previous Water Source: Karnataka........................ 31 4.2b Increased Water Consumption and Previous Water Source: Maharashtra.................... 31 4.3 Average Reduction in Time a Household Spends Daily Collecting Water................... 32 4.4a Time Savings and Previous Water Source: Karnataka................................................... 33 4.4b Time Savings and previous Water Source: Maharashtra................................................ 33 4.5 Impact on Social Interaction in Communitie................................................................. 35 4.6 Share of Women in Water Committees........................................................................... 36 3 Preface This Impact Evaluation Report (IER) assesses the medium- and long-term impacts of the rural water infrastructure constructed with the proceeds of three International Development Association (IDA) credits. The purpose was to determine the impact of Bank lending for rural water and sanitation in India, as well as to compare areas that had benefited from previous interventions with areas yet to be served. Basic Credit Data (actual) US$ Board Credit Number Project Name Million Approval Completion 1046 Rajasthan Water Supply & Sewerage 80.0 06/25/80 12/31/88 2234 Maharashtra Rural Water Supply & Environmental Sanitation 109.9 06/05/91 06/30/98 2483 Karnataka Rural Water Supply & Environmental Sanitation 92.0 06/04/93 12/31/99 The Operations Evaluation Department (OED) study team visited India in November and December 1997 and conducted on-site inspections in 32 villages (12 in Karnataka and 20 in Maharashtra). The selected villages covered the entire implementation period (from 1981 to the present). In each village researchers interviewed the local water users' group and administered questionnaires to randomly selected households. The study mission also met with state and district officials, visited facilities constructed by the projects, and conducted a technical evaluation. The study team extends its gratitude to Brian Grover (peer reviewer) for his helpful comments on the work in progress. The survey was conducted by ORG/MARG India and was supervised by the Center for Institutional Reform and the Informal Sector (IRIS) at the University of Maryland, College Park. Research in the field was conducted over a period of about three months. The study team particularly wishes to acknowledge the participation and facilitation during the field work of the UNDP/World Bank Regional Water and Sanitation Group for South Asia, Messrs. K. M. Minnatullah and A. K. Sengupta, and World Bank Task Managers Mr. G. V. Abhyankar and Mr. S. Sarkar. Copies of the draft IER were sent to the officials and agencies of the Government of India for review and comment, but no comments were received. 5 1. Study Methodology and Project Context The Impact Evaluation Study 1.1 The main objectives of this impact evaluation are to examine the relevance and assess the impacts of the Bank's assistance program on the performance of the rural water supply and sanitation sector in India. The study focused its field research on two projects: the Maharashtra Rural Water Supply and Environmental Sanitation Project and the Karnataka Rural Water Supply and Environmental Sanitation Project. It also selectively examined the experience of the Rajasthan Water Supply and Sewerage Project through limited field visits and a stakeholder workshop. The results of a recent OED evaluation of the Kerala Water Supply and Sanitation Project' also contributed to the study. Country Context 1.2 With a GNP per capita of US$380 (in 1996) and a third of its people living below the official poverty line, India ranks among the world's poorest countries. Life expectancy is 62 years, but infant mortality is 68 per 1,000 live births, one of the highest in the world. Half of the nation's adults-including two thirds of the women--are illiterate; the adult labor force has an average of 2.4 years of schooling, even though India has one of the world's largest pools of internationally qualified professionals, engineers, and scientists. More than 70 percent of the nation's 990 million people live in half a million rural villages, where the poverty is worst and illiteracy and infant mortality are highest. Improving the supply of water and sanitation services to those villages is essential to bettering the health and well-being of India's rural population: water-borne diseases are a major health problem and affect some 80 percent of infants and children nationwide. Rural Water Supply in India Before the Projects: One Safe Source Per Village 1.3 Over the past two decades India has implemented major investment programs in rural water supply and sanitation. By the early 1990s (when the Maharashtra and Karnataka projects were under preparation), the Government of India (GOI) could confidently estimate that about 95 percent of all rural villages in India had access to at least one safe water source. A safe water source, as defined by the GOI, provides potable water and is free of contaminants such as excessive salinity, iron, fluorides, and guinea worm (this definition is used throughout this report). Another criterion the GOI set for public investments was accessibility: to be considered for use with a village system, a water source had to be no farther than 1.6 kilometers from the village.2 Villages not meeting these criteria, the so-called problem villages (Box 1.1), were given priority for water supply development. 1. OED audited the Kerala project (Cr. 1622-IN) in 1998. 2. In a piped water supply system, the government uses the following consumption standards in system design: standpost yields of 40 liters per capita per day (lcd); house connections (or yard taps) should receive 70 lcd. 6 Box 1.1. Problem Villages Nationally, problem villages are those meeting any of the following conditions: (1) there are no wells; (2) wells are only seasonal; (3) wells produce water that does not meet drinking water standards; or (4) the distance to an adequate water source (according to the central government definition) is more than 1.6 kilometers. States differ slightly in the way they define problem villages. The responsible sector ministry in each state establishes a definition based on the availability of water resources and the most common health problems. For example, the criteria for the Government of Rajasthan are based on the incidence of diarrheal disease plus the normal considerations (distance from water sources, salinity, presence of guinea worm). The Government of Maharashtra has set a maximum distance of 0.5 km between the farthest house and the water source, excluding sources where the depth of water exceeds 15 meters and groundwater is contaminated by salinity, iron, fluoride, and several other toxic elements. Furthermore, villages with endemic cholera or guinea worm receive preference. The Government of Karnataka has reduced the acceptable distance to source from 1.6 km to 1.0 km and reduced the elevation differential from 100 meters elevation to 50 meters elevation for hilly areas. 1.4 During project preparation, Bank and GOI staff were aware that, despite important improvements in water quality and availability, potential health benefits had been undermined by maintenance problems, underutilization, and the occasional use of unsafe sources.3 In the early 1990s, infant mortality rates in rural areas were still some 70 percent higher than in urban areas. This was due, in part, to low usage of water from the "one safe source." Low usage follows from insufficient supply, distance, and social pressures, which sometimes inhibit lower-status groups from sharing a water point used by the more privileged. Some studies indicate that the usage of safe sources diminishes if the distance exceeds 150 meters and alternative sources (even unsafe ones) are available . 1.5 At project preparation, then, it was clear that the supply of water from "one safe source" was inadequate to meet demand in most villages, even when their schemes were fully operational. At any given time, though, about one fourth of the schemes was out of operation. Therefore, the potential impact from higher service levels, improved sanitation, and hygiene/health education was expected to be great. 1.6 Before the projects, several programs had been initiated to promote hygiene education and appropriate sanitation in rural areas, but they proved ineffective because of supply interruptions and shortfalls, as well as because the demand for improved sanitation facilities remained unmet. Less than 3 percent of the rural population used sanitary latrines at the time of project preparation. Improved drainage and solid waste disposal were not common. 3. These issues were identified in a Bank study "Strategy Paper, RuralWater Supply and Sanitation", February 1989. The strategy paper (prepared with the assistance of the UNDP/World Bank Water and Sanitation Program) established the basis for formulation of objectives and the strategy for the next generation of rural water supply and sanitation projects in Maharashira and Karnataka. 4. Churchill, A., D. de Ferranti, R. Roche, C. Tager, AA. Walters, and A. Yazer. (1987) Time for a Change. World Bank Discussion Paper No. 18. Washington D.C. 7 State-specific Pre-project Conditions: Water and Sanitation Rajasthan 1.7 Rajasthan has a total population of 44 million (1991 census) , of which about 77 percent live in rural areas. Within the rural population, 18 percent are scheduled castes and 15 percent are scheduled tribes. Rajasthan's literacy rate of about 39 percent is the second lowest among all the states in India and well below the all-India rate of about 52 percent. 1.8 By 1979 (when the Rajasthan project was under preparation), about 3,800 villages (almost all of them classified at project appraisal as problem villages), with a population of about 6.3 million, had been provided with piped water supplies, mainly through standposts. The state- wide service coverage was about 24 percent in 1979. The unserved population relied on shallow wells, remote surface water sources, or rainwater collection. About 3 percent were served by water supplied through public handpumps. The service levels were similar to those in other rural areas throughout India. 1.9 Sanitary facilities were generally lacking in the rural areas, and no program had yet evolved to improve the situation. Karnataka 1.10 Karnataka has a total population of about 45 million (1991 census), about 31 million (68.9 percent) of which live in rural areas. Scheduled castes make up about 16 percent of the rural population, scheduled tribes about 4 percent. Karnataka was one of the first states in India to provide at least one safe source for each village. 1.11 In the early 1980s, about 70 percent of the state's villages did not have water supplies meeting the GOI criteria. By mid-1991, after implementation of a large-scale program all villages had at least one safe water source. Per capita consumption levels were quite low, however, partly because of long hauling distances, particularly during the dry seasons when many seasonal water sources go dry. Later, when Karnataka reduced its distance criterion (see Box 1.1), in the schemes constructed according to the old criterion, 100 percent safe source coverage was no longer the norm. 1.12 Karnataka experienced three successive droughts in the 1980s, which increased the competition for the limited available supply of water. Before the project, some 35 percent of the 27,000 villages were served entirely by handpumps drawing on either borewells or shallow wells; 65 percent were served by combined piped water supply and handpump well schemes (about a fifth of these were mini-schemes consisting of a cistern with four taps). Knowledgeable people in the sector estimate that about 40 percent of the rural population was consuming less than 20 liters per capita per day (lcd) and that consumption levels rarely went beyond 30 lcd, even among the villages served by better-functioning piped schemes. 1.13 At project preparation, about 2 percent of the rural population had access to a latrine that met minimum sanitary conditions. Diarrheal diseases were among the main health problems, making hygiene education and sanitation improvements high developmental priorities in Karnataka. 8 Maharashtra 1.14 Maharashtra is the largest state in the country, and its literacy rate is about 10 percent above the national average. The 1991 census estimated its population at about 79 million-with 61 percent living in rural areas. About 25 percent of the state's inhabitants are either scheduled castes or tribes. 1.15 By about 1990, the rural water situation in Maharashtra was nearly the same as in Karnataka. Public officials claimed that access to the one safe source was better than the all-India average of 95 percent. There are no statistics in Maharashtra for consumption levels before the project, but groundwater is less common in Maharashtra than in Karnataka. Hence the state has more piped systems-which often are regional-conveying water from distant sources to (often) multiple villages. The state has no water supply mini-schemes. Of those families that had some water access, about 52 percent were served by handpump wells and 48 percent were served by piped schemes. Only about one fourth of the families served by piped systems had house connections; the rest were supplied at public standposts. Again, though, usage was low because of the long distance to the source; and during the dry season many handpump wells went dry. About 5 percent relied solely on traditional (unsafe) sources, though a far greater percentage used them sporadically. 1.16 Sanitation standards in the state were above the all-India average but still generally low. Only about 5 percent of the rural population had access to sanitary latrines and there was a consequent high incidence of diarrheal diseases. Local Institutional Arrangements 1.17 Before the project, water systems in Karnataka were mainly operated by local village governments with assistance from district government units (Zilla Panchayat)' as necessary on major technical repairs and equipment replacement. Maharashtra had a similar arrangement that involved different types of local governments6 and the Maharashtra Water Supply and Sewerage Board (MWSSB). In Rajasthan, rural water and sanitation (RWS) was solely the responsibility of PHED. In all states, funding came mostly from the central and state governments. GOI Commitment to the Water Sector and Strategy 1.18 The GOI has normally spelled out its commitment to water supply and sanitation priorities in a sector strategy paper updated during preparation of its five-year development plans (FYDPs). The seventh FYDP (1985-90) particularly focused on RWS and identified the following key issues and strategic objectives: (a) Community participation. Communities need to be given an opportunity to participate in planning, implementation, and operation and maintenance (O&M) of their water supply and sanitation systems through village committees. 5. Water systems were usually operated by Mandal Panchayat (local government units that manage 10 villages). Zilla Panchayats operate under the Ministry of Rural Development and Panchayat Raj (MRD&PR). 6. Gram Panchayat, local governments that cover just one village. 9 (b) O&Mof water and sanitation systems. Although greatly improved during the 1980s, the institutional and financial demands of O&M continue to be major burdens as coverage and service levels increase. (c) House connections (yard taps). The GOI's policy in the past discouraged the increase of house connections in the name of better service equity. As demand for private connections has grown considerably, future schemes will need to accommodate more such connections. (d) Cost recovery. Cost recovery-even generating funds to cover the O&M costs of RWS schemes-has received little attention. Handing over O&M responsibilities to rural villages is expected to improve this situation, but the capital costs of new investments will continue to come from central and state funds for some time. 1.19 The eighth FYDP (1992-97) further emphasized the above strategies, and project designs included specific actions to attain these strategic objectives. Particularly in the Karnataka project, the design included testing the effectiveness of community participation. It did so through the implementation of 12 pilot schemes (one in each district), encouraging private (house) connections, and demanding that the local government (Mandal), together with water committees, establish tariffs that would fully cover the O&M costs. The eighth FYDP also required that sanitation be linked with water supply improvements and health and hygiene education activities to create a greater demand for latrine construction. 1.20 The funding of water and sanitation sector investments by central and state sources has dramatically increased over the past decade. By the early 1990s, the total funding of the sector investments was about 3 percent of the national budget. Of this, some 60 percent was allocated to RWS investments. Although RWS is a state subject, central funding continues to increase and is currently about 40 percent of all RWS investments. Study Methods 1.21 In November 1997, OED began a series of activities in India that included fielding teams of researchers to administer questionnaires at the household, village, and district levels; collecting secondary data; and filling out technical data sheets on each district and village visited and the operation of its water system. Data were collected from 60 Indian communities through a survey of 1,100 households and 25 water committees in Karnataka and 25 local government units in Maharashtra. 1.22 - The study started with a pilot phase during which researchers reviewed available secondary data, tested data sheets and questionnaires, developed household selection criteria, identified the appropriate sample size and composition, and conducted pilot focus group and community interviews. The study team compiled lists of all villages that have had access to potable water through the projects for about six months or more and selected survey villages at random. Neighborhoods and households to be polled were selected randomly from voting lists of all village adults. The household survey was conducted by carefully trained interviewers; during the pilot testing, the interviewers were under constant supervision to ensure consistency in the way the questionnaire was administered. The interviews were conducted at times convenient to the villagers to ensure maximum participation by both women and men. Ten villages that never 10 benefited from a Bank-financed water project and have characteristics similar to served villages were included as control villages. 1.23 To complement the data collected with the questionnaires, the study team met with local water committees and beneficiary groups. Participants were asked about the quality of service, problems with system design or social interaction, and factors that contribute to the success or failure of rural water schemes. A workshop held in Rajasthan addressed similar questions and provided useful qualitative information. Questions Studied 1.24 The study addresses questions related to outcome and impact in three areas: project preparation, project design, and implementation. The principal research questions are as follows: Project Preparation * What has been the involvement of local institutions in the preparation process? * What are the characteristics of the project institutional framework? To what degree was the project supply or demand driven? How are these factors associated with success or failure? * What has been the participation of local stakeholders? * How was the water extraction technology chosen? Was the participation of local stakeholders in technology selection adequate? Project Design * What is the relationship between eligibility criteria, nature of village contributions, the delegation of work and responsibility, cost recovery, and observed project outcomes? * To what degree was post-project follow-up and technical support included in design? Implementation * What have been the projects' achievements, and what have been their impact? * What are the key ingredients for a successful project or sustainable operation of a water point? * How significant are project requirements for guiding field actions and communication with beneficiaries? * What are the reasons for failure of rural water projects? 11 1.25 Pilot open-ended interviews were conducted with Bank staff early in the research process. During the field research, the study team interviewed representatives of borrower agencies. Bilateral agencies and nongovernmental organizations (NGOs) working in India have also shared their experiences with the researchers. 2. Evolution of the Rural Water Paradigm in India RWS in the India Portfolio 2.1 Bank lending for rural water and sanitation in India has grown significantly during the 1990s, increasing from US$120 million in the 1980s to US$260 million in the 1990s. Three RWS projects have been completed and two are ongoing. The completed projects are the Rajasthan Water Supply and Sewerage Project, Kerala Water Supply and Sanitation Project, and the Maharashtra Rural Water Supply and Environmental Sanitation Project. The Karnataka Water Supply and Environmental Sanitation Project and the Uttar Pradesh Rural Water Supply and Sanitation Project (Ln. 4056-IN) are ongoing. These projects show a clear line of progression from a supply-driven, top-down development strategy to a more demand-driven, community- based approach. 2.2 This chapter looks at the degree to which India's Bank-funded RWS projects have characteristics of community-based water supply and sanitation and whether the projects were intended to be old- or new-style undertakings. In a demand-responsive approach, project rules tend to emphasize the provision of "appropriate" (low-cost) options to the community, assuming that many communities could not afford higher levels of service such as house connections. Community-based systems in rural areas7 are invariably characterized by beneficiary participation in system administration and direct local responsibility for all or certain aspects of operation and maintenance. Rajasthan Water Supply and Sewerage Project 2.3 The RWS component of the Rajasthan project, within what was largely an urban water supply and sanitation (W&S) project, sought to improve water supply for about 2,000 villages. The government, in this case the Rajasthan PHED, assessed the needs and made the technology and cost decisions without consulting the beneficiaries, clearly an old-style exercise. The RWS component consisted of handpump well schemes and motorized pumped schemes (often using tubewells as a source), with storage tanks to provide service during hours when the pumps were not in operation and to accommodate periods of high use and limited distribution mains. What makes this project a first step toward the new paradigm is, for the first time, the project rules emphasized consumer responsibility for covering all O&M costs. 2.4 After the infrastructure for the Rajasthan project was completed in 1988, without fully enforcing the cost-recovery conditions, the Rajasthan PHED took responsibility for the O&M of 7. Community-based systems are also becoming common in peri-urban areas. 12 the new schemes. PHED has also gradually begun to expand and upgrade the schemes with grant funding wherever the need has arisen. 2.5 An OED evaluation of the Rajasthan project in 1990 concluded that more than 4,200 villages were better served because the physical infrastructure was built. Further, the schemes were largely operating as planned, except for the proposed tariffs, which were never imposed in most rural villages. The evaluation went on to note that financial targets were never attained, so the daily delivery of water was still being subsidized. Also, the institutional development had been minimal as the implementation capacity of the various units was compromised by political pressures; the design process had been inadequate because of its rigid approach and overly strict adherence to prototype designs; and a policy on finance was left for future development. 2.6 As part of this study, OED organized The Impact Evaluation Workshop on the World Bank-Assisted Rural Water Supply Project in Rajasthan in Jaipur on December 16, 1997.8 PHED staff and other stakeholder groups that attended evaluated the project achievements and identified factors that enhanced and constrained the achievement of desirable outcomes. During the workshop, there was widespread agreement that the project's benefits fell in the following areas: Greater Access andAvailability * Clean, safe water available for the first time * Twice-daily service has permitted new uses for clean water in the homes * Scheduled castes and tribes have adequate access for the first time Health * Certain diseases were eliminated, the prevalence of others was greatly reduced * Health is generally better, especially among children * Greater awareness of environmental hygiene issues Social and Economic * Poor have access to clean water for the first time * Reduction in untouchability * Water takes a smaller share of household resources, including time 8. The participatory evaluation workshop was the first of its kind in India. PHED, in its report on the event, noted that "the response was very enthusiastic and the ideas liberally flowed ... the views will be taken care of in future projects so that the sentiment of the user group is accommodated and the future projects are more in tune with the expectation of the user group." 13 2.7 The workshop participants noted new challenges facing project-funded systems. Among these were rising levels of expectation in several areas. For example, people are now aware of drainage problems, conflicts over wastage sometimes arise, and in the more urbanized rural areas the increased supply of water turns the streets into mudholes. Additionally, many users are demanding sanitation; parents particularly want improved water and sanitation services in the local schools. The small contribution toward O&M has lead to strong demands for increased community participation in a range of decisions affecting the operation of the water schemes. These include demands for public consultation on decisions regarding the capacity of storage tanks, commissioning of civil works, and major purchases; insistence on stiffer penalties for poor performance by contractors; demands for increased participation of local government in system operation; and numerous requests for the authorization of local water committees. This conjunction of rising expectations and increased awareness bodes well for future community involvement. Other common complaints included the lack of adequate participation in system repair decisions, the inadequacy of water storage capacity, and the number of unresolved disputes regarding location of key infrastructure components and land rights. Kerala Water Supply and Sanitation Project 2.8 This project, like the one in Rajasthan, took atop-down approach. In this case the implementer was the Kerala Water Authority, an agency the state government had created out of its PHED. 2.9 Although this project also attempted a cost-recovery scheme, it was almost completely dysfunctional. The tax levied to cover costs rarely reached KWA, which then had to cover costs from its state-subsidized budget. The Karnataka and Maharashtra Projects 2.10 Two more-recent projects (in Maharashtra and Karnataka) clearly attempted to decentralize the responsibility of O&M, that is, to involve the beneficiaries of the new supply facilities in their O&M and upkeep. In this undertaking, however, there were serious conflicts and contradictions between the aims of the project and its design (rules, management structure, budget). Although early Maharashtra project documents highlighted the importance of community involvement and committed project staff to a participatory approach, in practice the communities' roles and the functions they were expected to perform were inadequately detailed at the design phase, and they received scant attention thereafter (see Box 2.1). 14 Box 2.1 Project Rules: Maharashtra Funding * Water supply facilities were to be financed through grant funds from state and central government sources with no contribution anticipated from consumers. * Environmental sanitation facilities were to be funded jointly by the district government and state government for drainage works (50 percent each). * Funding for household latrines was to come from the district and state governments (this involved a direct subsidy of 50 to 75 percent) with the remaining 50 to 25 percent to come from beneficiary contributions.9 Operation, Maintenance, and Tarifs After completion of the schemes, the MSSWB was to transfer the schemes to the district governments and local governments for O&M, as appropriate: District governments were to be responsible for the O&M of all regional water supply schemes. Local governments were to be responsible for the O&M of individual village-based water supply schemes, including handpumps, as well as for O&M of sanitation facilities. The costs of O&M were to be covered as follows: District governments were to transfer sufficient funds to a separate "Fund for O&M of RWS Works," (including those funds generated from bulk water charges) to cover the regional schemes' O&M costs. (The cost component for O&M up to the boundaries of connected villages normally represents 20 percent of district governments' regular income.) Local governments were to set and collect water tariffs to cover the O&M costs of those segment, of the schemes that lie within the village boundaries, as well as the O&M of handpumps. In addition, this tariff should cover bulk water charges in cases where the village is connected to a regional scheme. Water and sanitation committees were to be established in each village, but project documents fail to specify their role in the O&M of the schemes. As envisaged, NGOs were to have a role in supporting the PPMU in community mobilization, hygiene, and health education promotion; implementation of pilot schemes with a special emphasis on promoting women's involvement; and training. 2.11 The Karnataka project, designed just two years after Maharashtra, was a major step forward for community involvement. Its design included clearly defined attributions for local water committees and basic rules that governed the manner in which the communities are expected discharge their duties (see Box 2.2). 9. In addition, the subsidy was limited by an upper ceiling which was increased to Rs 3,500 per unit in 1995, thus making the sanitation program practically free and supply driven. 15 Box 2.2 Project Rules: Karnataka Funding * Water supply facilities were to be financed through grant funds from state and central government sources with no contribution anticipated from consumers. * Environmental sanitation facilities were to be funded as follows: For drainage works, communities would fund 30 percent and state government would fund 70 percent. The state government would provide subsidies for household latrine construction at Rs. 1,650 (69 percent) for low-income and Rs. 1,200 (33 percent) for higher-income groups, the remainder of the costs would be beneficiary contributions. Operation, Maintenance, and Tariffs After completion of the schemes, the state government and project implement unit were to transfer the schemes to beneficiary communities for O&M. The costs of O&M were to be covered fully by the beneficiary communities, including bulk water charges (for villages that get their water from a regional scheme). The Mandals, a local government unit that governs a cluster of 10 villages, set tariffs for the various categories of households, charging one rate for families with house connections, and another (lower) rate for those using standposts or handpumps. These tariffs typically include a connection fee (ranging from Rs. 500 to Rs. 600) and a charge for water consumption to be set at a level sufficient to cover the costs of O&M. In addition: * the costs of O&M for environmental sanitation facilities were to be covered by the communities (for drainage works) through a maintenance charge; * households had to pay off the loans they took out for private latrines; * the state government was to issue regulations (Mandal bylaws) for establishment of VWSCs and, through local governments, to empower the VWSCs to operate the schemes and levy charges to cover operating costs; and * the state government was also to issue contracts with local NGOs to pay them for: (i) organizing VWSCs, (ii) assisting authorities with village planning, (iii) monitoring the initial performance of VWSCs, and (iv) promoting sanitation and health education. The Most Recent Project 2.12 The ongoing RWS project in Uttar Pradesh (UP, approved in 1996) is the latest iteration of community-based RWS and a significant milestone. The project design is based on the experience of earlier Indian RWS projects as well as on the lessons learned in other countries with similar RWS situation. Incorporation of these lessons in the project makes it the most demand- responsive endeavor in the subsector to date.10 10. India has had many achievements in the RWS sector. These were accomplished for instance through an accelerated RWS program financed and coordinated by the Rajiv Ghandi National Drinking Water Mission as well as through the two latest projects reviewed by this study. The lessons from these achievements are documented in detail in the "Rural Water Supply and Sanitation Sector Review", which was carried out jointly by the GOI and the World Bank in 1997. 16 The Evolving Paradigm 2.13 By world standards, the Indian projects reviewed by the study are still in transition from one approach to the other. The earlier ones have certain demand-responsive characteristics (primarily, emphasizing beneficiary participation in cost recovery) but include no mechanism that will permit meaningful local involvement. Change is slow. Implementing agencies are staffed by utility engineers long accustomed to making important decisionsfor the beneficiaries rather than with them. OED evaluations of early water projects in the region showed clearly that a new approach was needed for rural communities. The villagers, who know most aspects of their own situation better than anyone else, choose the cost and level of service they need. Moreover, they are more willing to repair systems that respond to their requirement, repairs can be accomplished more quickly when local bodies (village institutions) handle them, and costs can be kept at affordable levels. 2.14 During the seventh FYDP, the "top-down" orientation began evolving into a community- based approach. The first attempt to design such a project in India was the Maharashtra project. Looking at the India water portfolio in broad terms, projects slowly shifted toward the "new" model and away from the "old" approaches over the 15-year period from Rajasthan to the present. Stated differently, the more recent the project the more elements of the "bottom-up," demand- based approach are likely to be present in the project's rules and approach (in general terms). Water projects such as those in Kerala and Rajasthan can be classified as bureaucratic. Government agencies made all the service level and design decisions, put the physical infrastructure in place, O&M is handled by government agencies, and expansions need to be covered with grant funding. The Rajasthan projects was implemented by PHED; Kerala was designed by PHED but managed by a PHED spin-off. More recent projects, such as Maharashtra, and Karnataka have a role for a local body (albeit governmental in the former), but local involvement in design and O&M is weak, and not much attention has been given to finance. The Karnataka project is the first to require establishment of a water committee that is responsible for O&M, has a role in training, and must make a cash contribution for environmental sanitation (latrines and drainage) to receive funding for water supply improvements from the central and state governments. 2.15 As none of the projects visited by this study team can be categorized as full community- based water supply and sanitation, we have tried to use each project as a separate window on the transition process. One question the following chapters address is: as projects move along the spectrum from no local participation toward full participation, is there a noticeable difference in results that can be attributed to changes in the level of stakeholder involvement? To give a more project-specific perspective, the achievements and impacts of those achievements will be reviewed in detail for two projects, Karnataka and Maharashtra (Chapters 3 and 4). 3. Project Achievements 3.1 The two projects subjected to in-depth review by the study were designed to improve rural water supply and sanitation coverage in all districts in the two states (12 districts in Karnataka, 10 districts in Maharashtra). The Karnataka project is still under implementation. Work in the field has been under way for about four and a half years; the project is intended to be 17 completed by the end of 1999. The Maharashtra project was implemented over seven years starting in 1991. See Annex 1 for a discussion of project objectives and components. Villages Surveyed 3.2 Table 3.1 lists the surveyed villages (25 in each state)", dates each'scheme was commissioned and handed over, and the period of operation from commissioning through the time of the survey (December 1997). The range in operation times is between four months and 45 Table 3.1. Status of Randomly Selected Schemes Karnataka Maharashtra Village Date of Date of Months in Village Date of Date of Months in Commissioning Handover Operation Commissioning Handover Operation Since Since Commissioninga Commissioninge through 12/97 through 12/97 Adaki Aug-97 Aug-97 5 Asangaon Aug-97 5 Arkerab Oct-95 27 Bamanwadi Aug-95 Jan-96 28 Bendigere Sep-97 4 Bhopala Aug-95 29 Bilgumba Aug-97 5 Bhose Jan-95 Jan-96 36 Bylanarsapura Sep-97 3 Bhoyara Aug-95 29 Chandakavadi Jul-97 nothanded 6 Chata Jul-96 18 over Daginakatta Feb-96 22 Dedale Aug-97 5 Garalgunji Jul-97 Aug-97 5 Dimbe Apr-96 21 Hally Aug-97 Sep-97 4 Eksal Apr-96 21 Hiremyageri Sep-97 Sep-97 4 Gharod Mar-97 10 Hombadimandoddi Jun-95 Mar-96 31 Khatwal Jan-96 Apr-96 24 Kembliganahally Jan-96 24 Nimbori Apr-97 Nov-96 8 Kemmale Sep-97 Sep-97 4 Osarwadi Aug-97 5 Kinihal - Sep-97 Sep-97 4 Patgaon Jan-95 Jan-96 36 Kudremani Jul-97 Aug-97 5 Pimpalwadi Mar-96 Aug-96 21 Kyathanahally completed not handed Rangaon Apr-97 Apr-96 8 over Madlur Jan-95 Sep-95 36 Salgara Jul-96 Oct-96 18 Manihal Jan-97 Feb-97 11 Sawargaon Aug-95 29 Nagathihally May-95 32 Shahiahanpur Mar-96 Jun-96 21 Saidapur Aug-97 Aug-97 5 Sulewadi Mar-96 Nov-96 21 Sathnurb Aug-97 Sep-97 4 Taharpur Apr-96 21 Screedaragadda Aug-97 Sep-97 4 Tarapur Apr-96 21 Thyamagondl Apr-94 Jun-94 45 Waghivali Apr-96 21 Udigala not commissioned - Waksai Oct-97 1 Veerapura Aug-97 Aug-97 4 Walipur Mar-96 Mar-97 21 a. When commissioning date is not available, duration is calculated from hand over through December 1997. b. No Water Committee. months for Karnataka and between five months and 36 months for Maharashtra;12 the average operational experience was about 13 months and 19 months, respectively. Although this operational experience is substantial in technical terms, the length of time each water committee 11. The study also surveyed five control villages in each state, thus the total number of villages studied is 60. 12. In Karnataka one scheme had been in operation only three months at the time of the survey; in Maharashtra one scheme had been in operation only one month. 18 or local body has had O&M responsibility is much shorter. The maximum is 43 months (Thyamagondi in Karnataka), but many schemes have been operated by the responsible group only for a couple of months. Coverage Improvements 3.3 The Karnataka project is about 80 percent through its implementation cycle. In only 22 of the surveyed villages has the scheme been fully handed over to the water committee; in two other villages, although the system has been handed over, there is no functioning committee yet. 3.4 Interviews with VWSCs in the villages where the survey was conducted indicate that, at least in 65 percent of the surveyed villages, the entire population now has access to project water. In the remainder, while the amount or quality of potable water has improved, full coverage has yet to be obtained. The household survey shows that about 45 percent of the households in the newly served villages get their service through house connections (including yard taps); another 42 percent get service from standposts. The remaining 13 percent of the population depend on handpump wells for safe water. The survey also shows that less than 10 percent of the house connection and standpost users rely on the new system exclusively. This means (and the survey shows) that a large majority (90 percent) supplement the water supplied by the project infrastructure with water from traditional sources-generally for laundry, bathing, and the care of domestic animals. 3.5 The Maharashtra project is now completed and is assumed to have achieved its physical objectives. In the surveyed villages, the systems have been handed over to the respective district government or local government units. 3.6 In study interviews district chief executive officers indicated that the entire population of the target villages should now have access to project-supplied water. Of the randomly selected households, about 27 percent said that they have a house connection (or a yard tap), another 48 percent obtain their water from public standposts. The remaining population (25 percent) uses handpump wells. The survey also shows that only about 17 percent of the households that have house connections (or yard taps) use the new system as their exclusive source; over 80 percent of the households still supplement their project supplied water from non-project sources. Pre- and Post-project Access 3.7 While the study team was concerned that in some instances social rather than technical or geographic reasons might impede access to water in the project area, this was not reflected in the survey results. Nonetheless, the higher service levels were probably distributed as they were for social as much as for economic reasons. 3.8 The survey indicates that the distance traveled to collect water has been reduced by nearly 90 percent (from 2.5 km to about 280 m after the project) in Karnataka. The time required for water collection has also dropped dramatically: from about six hours to nearly one hour (a decrease of over 80 percent). The survey of a group of households who are not users of the new system reveals that the distance to collect water remains at about 3.5 km and the daily time spent is nearly four hours (Table 3.2). 19 Table 3.2. Improved Access to Service, Time Saving Indicator Karnataka Maharashtra Reduction in time spent daily collecting water (minutes) 262 138 Reduction in distance to collect water (meters) 2,220 2,070 3.9 In Maharashtra, the project reduced the distance traveled to collect water by over 85 percent (from 2.4 km to about 330 meters). The time used daily for collecting water also dropped dramatically from slightly over three hours to less than one hour a day, a reduction of about 70 percent. The distance and time indicators for non-users in Maharashtra remain at 2.7 km and nearly three hours. Functioning of Schemes 3.10 The survey and interviews collected information on various aspects of system functioning such as prevalence of leakages, system failures, service problems at critical points, and quality of water. Table 3.3 summarizes the results. Table 3.3. Most Common Scheme Problems Indicator Karnataka Maharashtra Leakage in network (% of communities) 54.5 60 Leakage in standpipes or house connections (% of communities) 0.4 35.1 Frequent system failure (% of communities) 25.0 57.1 Inadequate water at critical points (% of communities) 50.0 50.0 Poor quality of water: - colored (% of communities) 14.3 11.8 - turbid (% of communities) 25.0 27.8 3.11 About 55 percent of Karnataka system users and 60 percent of Maharashtra users report major system leakages. In contrast, less than 1 percent of users in Karnataka and 35 percent of users in Maharashtra reported leakages at the yard tap and standpipe. This indicates that small repairs are more likely to be carried out in a timely fashion than larger ones. While this is understandable because small repairs are easier to do, leaving water mains carrying large volumes of water unrepaired poses a serious threat to an adequate supply. Interviews with water committees in Karnataka found that their water systems go completely out of service three to four times a year; in Maharashtra, village local government units also report occasional system breakdowns. The beneficiaries confirmed this: 25 percent in Karnataka and 57 percent in Maharashtra reported frequent system failure. Major repairs, when they are carried out, tend to take about a week (the average for all the cases about which the study has data). The longest service interruptions lasted two weeks. 3.12 About 13 percent of users complained about excessive or unpleasant coloration of their water, and about 26 percent complained about excessive turbidity. Opinions about the overall water quality were mixed: almost half of the respondents felt that water quality had improved after the project while the other half felt that it had stayed the same. Water quality testing was initiated at the outset of the Karnataka project as a regular activity, but testing has become sporadic in some villages and has been discontinued in others. In Maharashtra, water quality is tested regularly only in about 40 percent of the villages. An additional 20 percent tested the 20 quality at the time of system installation; the remaining 40 percent report that the water quality was never tested. Laboratory reports documenting serious water problems are not known to exist in either state, nor did committees allege that any existed. However, the survey shows that consumers suspect (rightly) that water may get contaminated due to system leakages. In water mains-especially when there is a long time lapse during the repair of the leakages- contaminants are more likely to infiltrate when systems are not under pressure due to intermittent supply. 3.13 Table 3.4 presents improvements in water supply service measured by the number of service hours and user satisfaction. The number of hours the service is available remains very low in both states. Table 3.4. Improved Functioning of Schemes: Service Hours and User Satisfaction Indicator Karnataka Maharashtra Water availability during wet season (hours per day) 1.5 1.6 Water availability during dry season (hours per day) 1.5 1.5 Satisfaction with the design of the service (% of households) 81 53 3.14 Before the projects, water systems were operated at irregular hours for short periods in the morning and evening for a total of about one hour a day. The survey found about a 50 percent improvement over this baseline-although the current average is about 1.5 hours a day. Even in the best-served villages the number of reported service hours is about three to four hours a day. Despite this, the hours of operation and water quality (especially during the dry season) have at least become more regular. In that sense, then, the reliability of water service has improved. To put the before and after situation in perspective: over 80 percent of the consumers in Karnataka, and over 50 percent in Maharashtra, express their satisfaction with the design of the new systems by rating them "good" or higher. The survey also shows that the current consumption levels in both schemes are quite low-between 10 and 17 lcd-well below official consumption standards (see footnote 2 and para. 4.7). Cost of Water Service 3.15 The capital costs of the water systems vary according to the complexity of the infrastructure. The schemes in Karnataka consist mostly of tubewell/piped systems; in Maharashtra, most schemes are based on a supply of water from regional bulk supply schemes. Table 3 .514 gives village-by-village capital costs for the new systems constructed so far in Karnataka and Maharashtra. 13. Per the cost data available to the study, the most expensive system in Kamataka was a borewell/piped scheme in Thyamagondlu costing nearly Rs. 3.5 million. The most expensive in Maharashtra was a regional scheme in Pimpalwadi costing more than Rs. 2.8 million. The cheapest schemes were Rs. 0.6 million (Kembliganahalli in Kamataka) and Rs. 0.9 million (Salgara in Maharashitra). 14. The table presents only the schemes where information was determined to be reasonably complete and reliable. 21 Table 3.5. System Components and Costs Location Type of System Cost Shallow Bore Wells/ Piped, Total, Rs. Cost Per Capita, Cost Per Capita, Wells Piped Regionalb Rs. US$ Karnataka Kembliganahalli 573,000 477.5 12.3 Bilagumba 1,090,000 463.2 11.9 Kemmale 1,098,000 784.3 20.2 Bylanararapura 1,823,000 446.0 11.5 Sathnur 3,272,000 673.3 17.3 Thyamagondlu x 3,493,000 334.8 8.6 Ankanahalli x 1,100,000 366.7 9.4 Chandakavadi x 1,202,000 267.1 6.9 Udigala x 1,574,000 276.1 7.1 Madlur x 1,910,000 341.1 8.8 Maharashtra Salgara x 900,000 666.7 17.1 Bamanwadi x 980,000 1068.7 27.5 Sulewadi x 1,080,000 1384.6 35.6 Pimpalwadi x 2,837,000 1773.1 45.6 Shahjahanpur x x 1,607,000 1267.4 32.6 Walipur x 2,122,400 4890.3 125.7 a. Only the villages with available information. b. Applicable to Maharashra only. c. US$ 1= Rs. 38.9 as of January 1998. 3.16 The average per capita cost to date in Karnataka has been about US$11; in Maharashtra the average cost has been about US$47, which compares favorably with cost indicators from other developing countries. The large difference in per capita costs is because all Maharashtra schemes in the table are regional systems that involve long transmission mains (which are relatively high cost). The per capita costs of the Karnataka schemes are lower than is typical for this type of schemes because many of them are simple tubewell systems with limited distribution networks; some schemes, in fact, only expanded or rehabilitated existing facilities. 3.17 The project rules did not require any particular contribution by consumers toward the capital cost of the schemes. The rules in Karnataka at the outset, though, required that consumers finance on-site sanitation facilities and drainage (loans were subsidized) through their local water committee.'s Project staff considered village borrowing for sanitation an indicator of village commitment to the water system and, more important, a commitment to accept the responsibility for water system O&M. The requirement that beneficiaries invest in improved sanitation has largely been met by all villages participating in the project. The rules in Maharashtra are structured in principle in the same way, but the government subsidy for both sanitation and drainage was higher than in Karnataka. 15. In order to foster increased local ownership, the project rules were changed after the mid-term review in 1996 so that there was also a capital contribution for water. From that point onwards, villagers were required to pay 50 percent of their contribution up-front as a prerequisite of participation in the project and prior to the award of any water supply contract in their village. Thus, water supply was more clearly brought under the capital cost sharing rules and more sensitive to real demand for improved services. 22 Tariffs and Cost Recovery 3.18 The collection of periodic water charges is not yet accepted practice, but initial steps are being taken in some villages. The survey shows that (of those whose schemes are currently charging) some 95 percent in Karnataka and about 70 percent in Maharashtra do pay something for either water or sanitation or both. The study was unable to confirm this finding with accounting data on the cost recovery performance of various water committees in Karnataka and the local bodies in Maharashtra. The reason is that even where records are kept, bookkeeping systems provide very little useful information. In some of the villages surveyed, consumers make in-kind contributions (mostly labor) to system O&M. The survey indicates that, on average, households contribute less than one day of labor a year. The tariff for water consumption is generally a flat rate per month per household, ranging from Rs. 10-20 for standposts to (rarely) more than Rs. 100-individual house connections almost invariably costing more. 3.19 The project rules require that water committees (Karnataka) and local bodies (Maharashtra) become fully responsible for the O&M of the schemes. With locally generated funds they are expected to cover the full O&M costs within the consumption area. Here again, most villages did not have useful data. Of the few that were able to provide the needed information, a small minority has achieved a positive revenue/expenditure ratio-most are still far from achieving it. Accordingly, the collection ratios (as percentage of billing) are low; the highest level observed was 60 percent (in one village in the Satara district of Maharashtra), and most range between 10 and 30 percent. This low level of performance is discussed further in para. 3.29. Institutional Achievements Functioning of Village-level Groups 3.20 Government agents, in collaboration with local NGOs, have implemented the evaluated projects. In addition, many committees formed for project purposes and other existing groups gained the capacity to handle water-related activities. During the design, construction, and management of water schemes and related services, these three groups were responsible for (or trained to handle) a number of tasks. The degree to which these tasks were accomplished is discussed below. Table 3.6. Effectiveness of Village-level Institutions Karnataka Maharashtra (water committees) (LGUs) Service operation and maintenance O&M participation (%) 29.7 1.5 O&M monitoring (%) 31.7 8.1 O&M sanctions (%) 29.5 27.0 O&M dispute resolution (%) 29.7 1.5 Good water group index-' 5.66 2.60 Female committee members (%) 13.7 20.3 Good committee attendance (%) 83.3 55.0 a!/ See para 3.21. 3.21 The study team asked survey participants how effectively the operating body-water committee or local government unit (LGU)-is handling the day-to-day challenges of scheme 23 operation (Table 3.6). There was likely to be a relationship between the exercise of administrative control by scheme managers over member labor and material contributions and whether contributions continued to be made. The study found that O&M participation was far superior in Kamataka and that o & M monitoring was taking place in 32 percent of Kamataka schemes but only in 8 percent of those in Maharashtra. The percentage of households that report sanctions are imposed on non-contributors to O&M was nearly equal in the two states. Kamataka is far ahead on the percentage of households reporting existence of a mechanism for resolving disputes among water users. Thegood water group is an additive index based on the percentage of households that rated the functioning and effectiveness of the community water group as "good" or better in terms of a series of questions on management ability, system functioning, inclusiveness, and good committeeattendance. The table shows that the functioning of Karnataka committees is rated more highly than the functioning of organizations in Maharashtra.6 The two water committee indicators, female membership on the committee and whether more than 50 percent of all members attend meetings, again indicate that Karnataka has achieved better performance, although female participation is somewhat lower." Training and Capacity-building by Local Government Units and NGOs 3.22 The performance of govemmental and NGO institutions can be evaluated by the degree to which they built capacity in the villagers to make the best use of project-constructed infrastructure. As Table 3.7 shows, respondents in Karnataka more often than respondents in Maharashtra said they received administrative training. This finding is consistent with the differences in institutional framework in the two states, as Karnataka has most of the fully functioning committees and has had committees since the project began. Health benefits from potable water depend heavily on beneficiaries receiving hygiene training: Maharashtra respondents more often said they had received such training. Unfortunately, the household survey question does not distinguish between training provided as part of the project and other hygiene training (for example, from NGOs working in the area but not contracted to the project). Furthermore, hygiene education began only recently in Karnataka and it is still ongoing. 3.23 The increase in the degree to which villagers have access to tools and spare parts is an important institutional achievement. In Karnataka the percentage of service operators with access to tools was much higher than in Maharashtra. This is probably because fewer tools were supplied in the first place. In contrast, a higher percentage in Maharashtra noted that the village had adequate access to spare parts; this may be because Maharashtra is ahead of Karnataka in the implementation cycle. 16. After the mid-term review in 1996, the community development activities in Karnataka were begun in advance of engineering planning and at least one year in advance of starting construction of water supply systems. 17. Since women are the primary water collectors in households surveyed, the participation of women indicates that the primary users of the service are able to influence decisions about its operation and maintenance. 24 Table 3.7. Training and Capacity-building Achievements Karnataka Maharashitra Administrative training (%) 41.7 5.0 Hygiene training (%) 7.9 41.8 Access to tools (%) 75.0 55.0 Access to spare parts (%) 33.3 45.0 Beneficiary Participation 3.24 In Karnataka project staff and NGOs helped to establish a functioning water committee in each community, ensuring that communities were consulted and able to make an informed decision about the type of service they wanted and overseeing the quality of construction of the service. So far the project has established functioning committees in the project villages quite effectively. In 23 of the 25 villages surveyed, the committee was already functional as the implementation of the water systems was nearing completion. Establishing the committee early, allowing for significant participation in design, and providing ongoing technical assistance help to make a committee strong and enhance a sense of ownership. Table 3.8 reviews indicators relevant to participation and ownership. As noted earlier, no capital contribution toward the construction of the water systems was expected or obtained from consumers. 3.25 In Maharashtra, other than minimal efforts in a few pilot endeavors, the whole attempt to involve communities early in the project cycle failed. The project did establish a "task force" that included representatives of local governments and selected (experienced) NGOs to work out the details of how the communities were to participate in the planning and management of project activities. After a slow start and a couple of years of work in these areas, the task force was dissolved without resolving the issue. Later, some water committees were nominally established, but on the whole the systems are operated and maintained by the local government units. As pointed out in (Box 2.1), project rules are somewhat contradictory and unworkable in the sense that they require a demand-driven approach within a top-down structure. O&M responsibility for the state's projects remains with local government (per the project rules) but attempts are being made to establish water committees. This community mobilization task has been complicated by the fact that the committees' roles and responsibilities are not yet fully defined. Table 3.8. Village Participation Indicators Karnataka Maharashtra Participation in areas likely to enhance ownership Early water committee (%) 83.3 25.0 Technology choice (%) 33.3 25.0 Service choice (%) 50.0 15.0 Cost choice (%) 50.0 15.0 Location choice (%) 41.7 25.0 Design performance 2.58 1.05 a/ An additive index based on the previous five indicators. 3.26 Overall, the participation indicators are clearly better in Karnataka than in Maharashtra. This finding is logical as the preparation of the Karnataka project followed the community-based approach early and more systematically than was the case in Maharashtra; see also the discussion in para 4.20 on institutional impact. 25 Sustainability of Service 3.27 The quality of the water service provided to the villages in the project states has certainly improved in the short term through the implementation of the two projects. Several positive technical achievements discussed above-such as improved access to water, increased reliability of service, and better quality of water (which, in many villages, is safe for the first time)- contribute to a significant quality-of-life improvement for project beneficiaries. The question remains, however: how well can this quality of service be sustained? The quality of construction was generally viewed to be good in many communities-more respondents in Karnataka (75 percent) than in Maharashtra (30 percent) rated the quality of construction as "good". 3.28 However, the survey results reveal gaps and overlaps in important responsibilities. When asked about who is responsible for the O&M of the scheme, 40 percent of Karnataka's water committees responded that no one was. Further, in all communities both the household respondents and the water committees said that critically needed major repairs have been neglected. In Maharashtra the situation is even worse. 3.29 Lack of stakeholder commitment to sound financial principles is another cause for concern. In the communities visited, the collection of fees and tariffs has been ad hoc." Only an extremely small number of water committees and local bodies were able to provide solid information on system operating costs and O&M cost recovery. When collection ratios and the inability of scheme managers to access system financial performance data are juxtaposed with the remaining or pending uncertainties about O&M responsibilities, and the generally abysmal financial performance, serious doubt is cast on the long-term sustainability of the schemes in Maharashtra and Karnataka. The same challenge is now being addressed in the UP project. A new partnership seems to be coping more proactively with these and other problems (see para. 2.12), and the results can be expected in the next couple of years. Social Capital and Institutional Performance 3.30 Table 3.9 presents selected social capital indicators for Maharashtra and Karnataka. Indicators of associational activity are based on questions households were asked about their participation in groups and associations. Although the picture is mixed when comparing the indicators for the two projects, interviews in project villages indicate that community mobilization was more timely in Karnataka and it occurred on many fronts. Karnataka therefore enjoys a slightly stronger level of social capital. 18. In some communities, users fees have not been collected because the scheme has not yet been officially handed over to the community or LGU. 26 Table 3.9. Selected indicators of social capital in Maharashtra and Karnataka Kamataka Maharashtra Association activity Group membership (%)a 14.4 27.8 Number of groups per household 0.2 0.5 Group characteristics' 1.1 1.9 Civic activityd One community activity (%)' 36.9 36.6 Two community activities (%)f 12.4 10.2 Road maintenance (%)g 6.9 7.8 Households participated (%)h 16.6 11.7 Cooperative conununity (%)i 83.3 70.0 a. A dichotomous variable for households that report being a member of a farmers' group, women's group, credit/finance group, political group, youth group, religious group, Panchyat Raj, death donation society, or other community group. b. The number of groups to which a household belongs excluding water users' groups. c. An unweighted average of various characteristics of each household's most important group, including heterogeneity of members by caste and religion, gender, and occupation; decisionmaking mechanisms; and effectiveness of group functioning. d. Indicators of civic activity are based on questions to households and to water committees about community-wide activities to build or maintain local public works (building schools, constructing roads, maintaining roads, constructing religious buildings). e. Households reporting at least one such activity in their community in the previous year. f. Households reporting at least two such activities in their community in the previous year. g. Households reporting road maintenance in their community in the previous year. h. Households reporting participation in any of these activities. i. Water committees reporting that their community is characterized by cooperation and civic interaction. 3.31 Among the indicators that measure civic activity, the first two show that communities in both states have similar (low) levels of community activities, including about the same amount of community road maintenance. Small minorities in Karnataka and Maharashtra report being involved in a community activity. Two of the water committees in Karnataka and six in Maharashtra said that their community is not cooperative. 3.32 In the analytical work conducted by the Center for Institutional Reform and the Informal Sector (IRIS), the communities surveyed were ranked according to the performance of W&S services. In Table 3.10, the top half of the table summarizes the results for the 16 best performers, and the bottom half the 16 worst performers. The functioning of schemes was measured by an index.19 For each category, the table reports selected significance-tested indicators of social capital, governmental and NGO institutions, service-level institutions, performance, and impact. 3.33 The study shows that in those communities where the pre-existing social capital (measured by association and civic activity) was high, as well as in the communities where social capital increased through participation in the project, the schemes are performing better. Governmental, NGOs and local institutions have performed better and produced better outcomes in communities with the best groups than in communities with the worst ones. A comparison of 19. An additive 10 point index based equally on consumer satisfaction and lack of observed or reported technical problems. The absence of the following factors were each assigned a seventh share of 5 points: leakage in the network, leakage in the standpipes, frequent system failure, neglect of key repairs, excessive coloration, high turbidity, and service problems at critical areas. The remaining (maximum of) 5 points were awarded according to the percentage of households surveyed that expressed satisfaction with the technology choice they were given. 27 the Putnam index0 reveals that households in communities with the best performing services belong to more groups, and that membership in these groups is also more heterogeneous regarding religion, occupation, and gender. Because better organized communities were more informed about their rights-and more insistent on being treated as they should be-government officials and NGO representatives more often completed their assignments in these communities as planned. Table 3.10. Comparing Group Means for the Best and Worst Performing W&S Services Indicator Mean Best Performers Social Capital Indicators: Number of groups 0.55 Putnam index 3.47 Two community activities 0.16 Indicators of Government Institutions: Design performance 1.81 Hygiene training 0.25 Indicators ofService-level Institutions: Design participation 0.23 Satisfactory women participation 0.15 Construction participation (> 2 days) 0.04 Construction monitoring 0.29 O&M any payment 0.59 Good water group 3.82 Performance Indicator: Functioning index 7.77 Impact Indicator: Improved health 0.56 Worst Performers Social Capital Indicators: Number of groups 0.26 Putnam index 1.76 Two community activities 0.10 Indicators of Government Institutions: Design performance 1.44 Hygiene training 0.21 Indicators ofService-level Institutions: Design participation 0.15 Satisfactory women participation 0.05 Construction participation (> 2 days) 0.05 Construction monitoring 0.15 O&M any payment 0.46 Good water group 3.69 Performance Indicator: Functioning index 3.63 Impact Indicator: Improved health 0.40 3.34 Annex 2 presents additional performance indicators that are relevant to understanding the achievements in institutional capacity building at the village level. 4. Project Impacts 4.1 The project achievements reviewed in the previous chapter had important impacts on the lives of beneficiaries, their communities, and the institutions that continue to provide them water 20. The Putnam index is an index of social capital that is the product of the number of groups a household is a member of and the characteristics of the group. Unlike the results with two community activities, the difference between the means of this measure of social capital is not statistically significant. 28 and sanitation services. Foremost among these consequences were health improvements. The number of families reporting improvements, a comparison between the health status of users and non-users, and the relationship between those whose health improved and the water source that they previously used are discussed below. The economic impacts of the projects were primarily time savings; few families launched new productive enterprises because of their improved access to safe water. This chapter also reviews social and institutional impacts, analyzing interaction patterns and how they changed with the introduction of water committees and systems. The impact on the lives of women is also discussed. Health Impacts Self-reported Improvements in General Health 4.2 Study researchers interviewed more than 1,000 rural households (in the villages listed in Table 3.1) about recent health problems .21 Slightly more than half (54 percent) of the households in Kamataka using the new water schemes and 63 percent of them in Maharashtra reported that they thought the health of their family had improved since the water scheme started to function. However, health impacts vary greatly across communities. In some communities no households noted health improvements, while in other communities all the households did. 4.3 The survey team sought detailed information only about the two-week period immediately preceding the day of the interview because self-reported health data become increasingly unreliable over longer periods of recall. The results show marked differences in the incidence of water-bome diseases among the users and non-users of schemes. As Table 4.1 indicates, at the time of the survey 5 percent of households using the new schemes in Karnataka and 15 percent in Maharashtra had suffered from diarrhea in the past two weeks. During the same period, the incidence of diarrhea among households not using the schemes was higher in both states. However, in Karnataka, a larger share of users than non-users reported that a family member had been hospitalized or died due to diarrhea in the past year. 4.4 Overall, the survey results indicate that the incidence of diarrhea and hospitalization and mortality due to diarrhea are more common in Maharashtra than in Karnataka. Table 4.1. Diarrhea: Incidence, Treatment Sought and Mortality in Indian Project Sites Incidence of diarrhea in Hospitalization due to diarrhea Death in household due to the past two weeks in the past year diarrhea in the past year (% of households) (% of households with diarrhea) (% of households) Among Users Karnataka (n=240) 5.0 5 0.8 Maharashtra (n=438) 15.0 10.3 0.9 Among Non-users Karnataka (n= 180) 13.0 1.1 0 Maharashtra (n= 151) 23.0 16.6 0.7 21. In Kamataka, 240 individuals in households with access to water systems, 180 in those without. In Maharashtra 438 individuals in households with access to water systems, 151 in those without. 29 Relationship ofPre- and Post-project Water Source to Health Improvement 4.5 In Maharashtra, households whose previous source of water was a pond, wadi, river, or stream reported health improvements most often.22 As Figures 4. la and 4. lb indicate, all households that had relied on a pond or wadi and 72 percent of those households that previously used a river or stream said that their health had improved. By contrast, in Karnataka, where surface water is less common, households that had used either a tubewell with a hand pump or a shallow well as a source of water noted most often that their health had improved. About 60 percent of previous users of tubewells with hand pumps or shallow wells thought the project has had a positive impact on their health. Figure 4.1a. Improved Health and Previous Water Source: Karnataka 100- 59.6 59.6 =60- - 51.6 5 a) 5 x40 3 0 20- Tubwell Tubewell Shallow Spring Rier or Pond or w/Hand w/Motor Well Stream Wadi Pump Pump Figure 4.1b. Improved Health and Previous Water Source: Maharashtra 100 100 - 80-. 72.2 65.9 65 .59.6 0- z40 3. 0- 0 1 Tubwell Tubewell Shallow Spring River or Pond or w/lHand w/Motor Well Stream Wadi Pump Pump 22. The small sample size of users of some sources makes some of the results statistically insignificant. 30 4.6 Further, the higher the level of service, the greater the self-reported improvement. Especially in Maharashtra, households that currently receive water through private piped connections reported health improvements most often (Table 4.2). Table 4.2. Improved Health and Current Water Source (% of households) Private Piped Shared Piped Connections Connections Karnataka 54,.93 53 n=122 Maharashtra 66 ,n=io 60n-,3 Health and Water Consumption Levels 4.7 In areas where water is scarce, health improvements due to the introduction of safe water are usually a function of increased availability leading to one or more new uses (it becomes possible to use clean water for cooking, bathing, and/or laundry). In Karnataka about 70 percent and in Maharashtra about 24 percent of households that reported improvement in their health had increased the consumption of water since the new water schemes started functioning. The quantity of water consumed per household and per capita is higher in Karnataka than in Maharashtra. On average, a household using the project-supplied service consumes 78 liters of water per day in Karnataka and 46 liters in Maharashtra. The per capita consumption in Karnataka ranges between 8 and 17 liters a day, the average being 11 liters. In Maharashtra, the average daily per capita consumption of water is 8 liters.2 4.8 Increased usage can be expected to be highest where pre-project demand was most in excess of supply. This proved to be true in Karnataka, where households whose previous source of water was a spring most often increased their water usage. In fact, all households that previously used water from springs in Karnataka have raised their water usage, as can be seen from Figures 4.2a and 4.2b. In Maharashtra households that had relied on a pond, wadi, river, or stream for drinking water have most often raised their water consumption. Since surface water is more plentiful in Maharashtra, increased usage might reflect the popular perception of these sources as unsafe (and indeed it does, see following paragraph). In Maharashtra, all households interviewed whose previous source of water was a pond and about 46 percent of households that had used water from a river for drinking noted that they consume more water now than before the project. 23. Compared with experience from other RWS projects, although increased, these quantitities are very low and only barely meet the need for potable water. For other household necessities, water needs to be extracted from non-project sources (see also paras. 3.4 and 3.6). 31 Figure 4.2a. Increased Water Consumption and Previous Water Source: Karnataka 100 100- 80 1 67.3 0 60.6 60-- 51.6 40-- 20 0 - Tubwell Tubewell Shallow Spring River or Pond or w/Hand w/Motor Well Stream Wadi Pump Pump Figure 4.2b. Increased Water Consumption and Previous Water Source: Maharashtra 100 100_. U,080 . so 0 4 XE 40o-- 32.9 30.6 33.3 20 0 Tubwell Tubewell Shallow Spring River or Pond or w/Hand w/Motor Well Stream Wadi Pump Pump 4.9 The awareness of safe sources of water and practices of water handling partially influence the health impact. Thus, the impacts of water systems are usually maximized by parallel health education programs. A recent OED evaluation in Africa, for example, showed that the introduction of handpumps did not lead to health improvements because the parallel effort in health education and sanitation was too feeble and not sustained long enough.2 About 88 percent of households that had obtained hygiene training as part of the project in Karnataka and 74 percent in Maharashtra considered water from a river or stream as unsafe. When households using the water service were asked to classify different water sources as safe or unsafe, about 89 percent in Karnataka and 68 percent of households in Maharashtra classified water from a river or a stream as unsafe for drinking. 24. See OED Precis 154 "Sustaining Rural Water Systems: The Case of Mali." November 1997. 32 4.10 About 41 percent of households surveyed in Karnataka reported that they purify water before drinking and about 68 percent of them reported the same in Maharashtra. About 35 percent of households in Karnataka that claimed improved health also purify water before drinking. In Maharashtra, 71 percent of households that reported improved health also stated that they purify their water. Economic Impact 4.11 As was noted above--and in addition to many health improvements that have economic impacts--the main economic impacts of the water and sanitation projects in both states have been time savings to individual households. The average time spent daily by a household collecting water decreased by more than four hours in Karnataka and by more than two hours in Maharashtra (Figure 4.3). Before the project, households in both states used to travel on average 2.5 km in a day to haul water. The new water schemes reduced the average distance each household travels daily to collect water to 279 meters in Karnataka and to 326 meters in Maharashtra. Figure 4.3. Average Reduction in Time a Household Spends Daily Collecting Water 300 - - 262 250 -- 2001. S150 100 . 50 .. 0- Kamataka Maharashtra 4.12 In Karnataka, the average daily water collection time dropped most for households whose previous source of water was a spring, while in Maharashtra households that had used tubewells with hand pumps experienced the biggest time savings. As Figures 4.4a and 4.4b indicate, the average time spent daily collecting water dropped by about four hours on average for households in Karnataka, and by nearly three hours for households that had relied on tubewells with hand pumps in Maharashtra. 33 Figure 4.4a. Time Savings and Previous Water Source: Karnataka (Average Reduction in Time Spent Collecting Water) 1000- 800-- 690 a 600-- -c400 . 40- 233 208 191 200 . 1 120 * 0 Tubw ell Tubew ell Shallow River or Pond or w /Hand w /Ivbtor Wel Stream Wadi Fimp Pump Note: A few users of distant springs may have benefited even more. Figure 4.4b. Time Savings and Previous Water Source: Maharashtra (Average Reduction in Time Spent Collecting Water) 200 -- 173 150-- 144 126 10 . 93 95 90 '~100- 50- 0 Tubwell Tubewell Shallow Spring River or Pond or w/Hand w/Motor Well Stream Wadi Pump Pump 4.13 Not surprisingly, households receiving private piped connections obtained the largest time saving-on average 6.5 hours in Karnataka and 3 hours in Maharashtra (Table 4.3). The benefits of these time savings primarily accrue to women, who are the main water collectors in households surveyed in both states. 34 Table 4.3. Time Saving and Current Source of Water Average Reduction in Time Spent Daily Collecting Water (minutes) Private Piped Connections Shared Pipe Connections Karnataka 395,3 184 ,m=22 Maharashtra 214,6 95,r253 4.14 The substantial time savings represent in a typical village some nine additional (10-hour) working days a month. Using an average field work daily rate of Rs. 202 as a basis, this represents a potential increase in household income of over 30 percent; the economic rate of return of such a gain is about 17 percent. 4.15 Only a fraction of households surveyed in the two states has initiated new income- generating activities that use water. In Karnataka, 1.7 percent of households using the schemes have put water to a new economic use; in Maharashtra 0.9 percent have done so. These activities include silk processing and cultivation of flowers, and the production of refreshments. Another example of an economic impact was observed in Kerala where OED has found that densification may follow the introduction of potable water in rural areas (Box 4.1). Box 4.1. Water Supply Improvement Caused An Increased Density of Rural Areas in Kerala A recent OED study in the state of Kerala found that the introduction of potable water led to an increased density of settlements because there was a strong pent-up demand for housing (due to lack of available water). Site visits to project areas revealed that as water was being introduced into the various local government wards, many more homes were built. Land records were reviewed in two local governments to confirm this perception. Average annual growth in the number of dwellings in wards for which eight years of data exist was 35.3 percent in one area and 16.1 percent in the other. Population growth in the communities that benefited from the water schemes was greater than in the surrounding district. Another significant project impact turned out to be that plots of land served by project water schemes appreciated substantially in value. Under the project rules the plots had to be within easy walking distance of standposts, and house connections were usually permitted (allowing the installation of sinks, showers, and other improvements). In two Puthencruz communities focus group meetings with the local government executive council members and other municipal employees sought information on the increase in land values. In one community the average increase in price for one house plot was reported to be 185 percent, in the other the groups estimated 225 percent. It appears that families who lived in crowded circumstances before there was water occupied additional dwellings when those could be served with safe water. But relatives living where water was still scarce tended to move to join families that had a better water supply because of the project. The impact on poverty of these windfall gains was minimal, however, in that most families did not sell land, and so realized no actual income benefit other than an increase in net worth. 25. Agricultural Wages in India, 1993-94. Directorate of Economics and Statistics, Ministry of Agriculture. 35 Social Impact 4.16 About 67 percent of water committees interviewed in Karnataka and 35 percent in Maharashtra said that social interaction in their community has increased since the project came to the community. This difference between the two states demonstrates the effectiveness of community promotion activities in Karnataka over those in Maharashtra. According to these water committees, members are now socializing and participating in community activities more than before the project. Figure 4.5. Impact on Social Interaction in Communities Karnataka Maharashtra No No Change Change 65% 33% Ieased - - - Increased SocialSocial InteractionInteraction 35% Gender Impacts 4.17 Women in India are still not full participants in decisions about water and sanitation, which affect them most of all. Water committee members are predominantly male in both states. In Karnataka, on average only about 14 percent of water committee members are women. In Maharashtra, the average is about 20 percent (Figure 4.6). This low participation by women in water committees is consistent with the low share of women in community leadership positions. In 92 percent of communities surveyed in Karnataka and in 95 percent in Maharashtra, all or most of community leaders are men. The share of female members in water committees is high in those few communities where the majority of community leaders are women and low in those where all or most leaders are male. Respondents to the household survey were highly critical of the lack of opportunities provided for women in the management of the water schemes: only 3 percent in Karnataka and 14 percent in Maharashtra said that the current level of women's participation was satisfactory. 36 Figure 4.6. Share of Women in Water Committees 25- 20.3 201 C.' 5. 0 Kamataka Maharashtra 4.18 Women are in charge of providing water for home use, household cleanliness, and sanitation-related maintenance. In focus groups held in the villages women noted that safe and more convenient water supply has certainly improved their quality of life. It has also enhanced the productivity of those rural women that use project-supplied water in their homes. They noted that easier access to safe water (and an assured supply) allows them to provide a cleaner environment in their homes with less effort. It is now easier and safer to wash their children. Women wash their own and their children's hands more often, and healthier children allow the mothers to be away from the house doing productive activities more days per year. Safe water close to the home helps women to keep food preparation areas clean. Washing clothes and cooking are simplified and less time consuming. Children's education is also greatly enhanced by good health (and nutrition is improved by adequate hygiene). All the above can also lead to higher self-esteem. 4.19 Improved sanitation has impacts beyond eliminating the health impacts of open defecation. The focus group interviews reveal, that in some villages, women were limited to attending to their bodily functions during hours of darkness because it is not proper for them to be seen doing so by day. Having a clean place to attend to personal hygiene is particularly welcome during menstruation. Institutional Impact 4.20 When the institutional impact of project interventions is positive, the expected outcomes are that O&M of the system will be more effective than in the past, and that institutional development has enhanced the systems' long-term sustainability in some way. As discussed in Chapter 3, the gains in O&M effectiveness and community participation have clearly been better in Karnataka than in Maharashtra. This is explained by the stronger pre-existing social capital base, which has interacted with the social impact of the project and the way it was carried out in Karnataka. It is also partly due to the early promotion of community involvement and more deliberate preparation of water committees for their future role. Similarly, government and NGO performance was stronger in Karnataka. 37 4.21 Village water committees have been set up, and based on their achievements so far (at least in Karnataka) they can be considered a modest success. However, to go beyond the creation of new local organizations-which require careful follow-up support if they are to have an impact-and to achieve longer-term sustainability, the focus of project staff and their NGO partners should be on scheme administration. Committees and local government units need stronger financial management, more and better training of operational staff, and more women's participation. In Maharashtra any future project should move,away from local government-run O&M and encourage the formation of committees using the lessons already learned and still to be learned from Karnataka. It is probable that as income levels rise there will be decreasing returns to increased participation. In the future, villagers may prefer to get water without having to attend regular meetings and doing maintenance work on their village water system. 4.22 The Karnataka and Maharashtra projects provide evidence that, given current conditions in rural India, the level of community participation and demand-responsiveness of a project does affect results, and that increasing local involvement in planning, technology selection, and scheme operation leads to improved service. In 1996 a new project was launched in Uttar Pradesh. Its new approach to expanded government, NGO, and community partnership-which stretches over to full responsibility by the NGO and community even in project implementation-sets new targets for demand-responsiveness and community participation. The Uttar Pradesh project undoubtedly will provide further lessons about the advisability of broader participation, and will set new standards for subsequent projects in the subsector. 5. Findings, Conclusions and Recommendations Findings 5.1 Although the hours of water supply have increased only slightly, the access to water is much better than before the projects and the supply is now more regular. Bank support for the rural water subsector in India generally has been relevant and is making a meaningful difference in the quality of village life. The rural water infrastructure is in place and functioning. The various schemes are helping to reduce the deficit in rural water supply, and increasing water coverage; in the best villages coverage has reached 100 percent. Sanitation in the project areas has improved substantially because of better water availability, and use of pour-flush latrines has greatly increased.26 5.2 One question needs to be asked, however. Did the differing degree of demand- responsiveness of the projects enhance or constrain their operations and the longer-term sustainability of the service delivery? The preponderance of the evidence is that the more community-based and demand-responsive the project (or even project village), the better it did in a full range of activities. 26. Although the sanitation program is still underway, much of the demand for improved sanitation is already met in the surveyed villages. 38 5.3 The Rajasthan project experience illustrates the severe difficulty of creating a sustainable operation without community participation. In that project, a community contribution and O&M cost recovery were explicitly called for during project design but neither materialized during implementation. And the lack of real ownership demonstrated by the ultimate beneficiaries was due to the fact that the systems cost them nothing, and they had neither control over any type of resource nor influence on decision-making. Even now, the State PHED continues to handle the O&M and finance system expansions. Thus, the continuity of service and post-project expansions of coverage depend heavily on governmental budgetary allocations and subsidies. In Kerala the ownership and sustainability situations were largely the same, the only community contributions regularly made were repayments of loans for household sanitary infrastructure (local governments are generally not passing along any funds collected from the rental value tax to the Kerala Water Authority). And the autonomous Kerala Water Authority designs and finances expansion without local input when government grants are made available. 5.4 In Maharashtra, an attempt was made to introduce more meaningful community participation with substantial assistance from NGOs. The project was designed in such atop- down exercise, however, that the respective roles of communities, NGOs and local government were never actually defined-although in practice local governments were made responsible for O&M. Since the role of village water committees was not clearly established during project preparation and the top-down structure did not facilitate their subsequent empowerment, the results were largely as expected: poor cost recovery, the project-provided opportunity for community development was wasted (the social capital indicators are still quite low), and the sustained operation of the systems is unlikely or uncertain at best. Consequently, the delivery of water and sanitation services in Maharashtra was not very demand-responsive and village participation in the ensuing implementation phases was not as active as originally anticipated. 5.5 Karnataka, the later of the two projects, was characterized by greater attention to participation and scheme design was more demand-responsive. Communities were mobilized with the help of NGOs, but the creation of opportunities for community members to participate began only well into implementation. With hindsight, it is clear that faster progress and a greater development impact could have been achieved had participatory activities commenced before and not during system construction. Zilla Panchayats are still strongly in charge of implementation, and (perhaps) some officials do not fully understand and appreciate a community-based way of working together with villagers. In Karnataka the results seem to be better than in Maharashtra (and the social capital indicators are generally higher) but cost recovery indicators are still low, critical financial data are generally not available, political interference by local government is a frequent complaint, and sustainability is far from ensured. The GOI in its most recent RWS project (in Uttar Pradesh) is approaching these issues in a fundamentally different way, i.e., by introducing direct capital contribution for the first time, and encouraging stronger community involvement in decision-making through a more dynamic partnership between the community, the NGOs, and the state project implementing unit. Conclusions 5.6 The evolving RWS paradigm in India has not yet come to a point that many experts in the sector would consider an effective demand-responsive model. Nevertheless, its experience with ever more participatory approaches provides a rich basis for a detailed evaluation of social capital (and other) factors that contribute to positive project results and the sustainability of services. Now that community mobilization in Karnataka begins at least a year before system construction 39 (see footnote 15), it would be worth researching during the Implementation Completion Report process (in 1999) whether the increased investment in organization is producing better results. 5.7 Responding to the priorities set out in the eighth FYDP (para. 1.19), the Maharashtra and Karnataka projects are showing that health benefits are enhanced by integrating water, sanitation, and hygiene education interventions (see para. 3.22); an approach typical of community-based projects. While women's participation is not equal to men's in all areas, the study has shown that women play a vital role in the local water users' groups, and the status and lives of women are better than they were before the project owing to better access and more regular availability of water. 5.8 The study yields several important policy conclusions: * When beneficiaries make a regular and significant financial contribution to the operation ofa water scheme, they feel a sense ofownership and demand more ofa say in scheme operation. In communities where a substantial portion of O&M costs is covered by users, the users become more concerned about those management aspects involved with keeping costs low. This was also true in Rajasthan (during the workshop, village representatives strongly insisted on greater participation in decision-making) and, to a lesser degree, in Kerala as well. * Involve women in the design and management of the W&S service. The results of this study provide strong evidence that involving users in all aspects of the development, implementation, and O&M of water supply systems can improve sustainability. Involving women makes particularly good sense; women's involvement in system management is critical for performance: women are the primary water collectors in most rural households and have the most interest in ensuring that the W&S service matches their needs and performs well.27 * Community-based W&S services are likely to perform well and have strong impacts in communities with high levels ofsocial capital.28 Project designers need to pay close attention to existing levels of social capital in communities, and adjust the approach to service delivery accordingly. The existence of social networks improves group organization and service functioning, as community members are accustomed to working together as a group. Also, social ties among community members deter free-riding and encourage community members to meet their commitments. Depending upon the way RWS projects are implemented, there are indications that social interactions can increase, and social capital growth is possible (see Figure 4.5).Therefore, in the design of projects that finance community-based W&S services-particularly in the design of social mobilization efforts-the existing levels of social capital in communities need to be taken into account. 27. These results complement the results obtained by Sara and Katz (1998). They showed that demand-responsive community-based W&S services are likely to have sustainable impacts on poverty alleviation and that sustainability of services was markedly higher in communities where households had made informed choices about whether to build a water system and about the type and level of service. 28. This extends the previous results on social capital obtained by Narayan and Pritchett (1997). They found that villages in Tanzania with high levels of social capital have better public services and more communal road construction and maintenance activities than villages with low social capital. 40 The effectiveness of water committees influences the performance of the W&S service. Committees provide users with incentives to contribute the required inputs to the design, construction, and O&M of water and sanitation services-and, together with other project partners, they develop rules. Rules need to be clear and well understood and accepted by all stakeholders, especially those rules that govern the use, operation, and maintenance of W&S services by the water users. Whether such rules exist-and whether they are properly implemented and enforced- depends on the social mobilization efforts of the project implementers. For example, whether women participate in service design and management, and whether households contribute to construction and pay for water (as agreed), depends on whether committees feel they own the project rules. Recommendations 5.9 The Key recommendations resulting from the study are as follows: * The Karnataka implementation completion report (ICR) due next year should be combined with the UP mid-term review (scheduled for 1999). Combining these two exercises will provide a marvelous opportunity to further improve the RWS program's modus operandi as well as inform the basic framework and rules for future investments. * The project implementer's role, the communities' role, and the functions of local committees and other groups (and the rules that govern their behavior) need to be specified clearly and designed early in the project cycle to facilitate broad participation in all aspects of project development. A dialog in which all groups participate during planning stages facilitates active community participation in the implementation stage. Government and NGO participation is also governed by project rules. Whether implementing agencies and NGOs carry out their tasks as planned and provide accurate and helpful information to community members about technology and service options, and teach community members about better hygiene practices, critically affects the delivery of decentralized W&S services. * In project design, aim at a high recovery level for O&M costs. Full recovery of O&M costs is the sine qua non of longer-term sustainability. Because O&M eventually becomes the responsibility of the users' groups, factors such as participation in multiple community activities by user group members, adequate participation in system design and construction, improved beneficiary health, and satisfactory participation in the water group by women are important indicators- they are all related to above-average cost recovery. Consequently, during project design, special attention needs to be given to these and related areas. * Design for full coverage and equitable distribution. Many systems have difficulty providing adequate service to critical areas, such as hilltops and furthest reaches of the system. A common pattern is that systems are designed for only the original number of users that express interest and pay their quota. Once the rest of the community sees that a system was really built, however, they begin applying pressure on the water committee to include more families. Systems usually expand 41 until the quality of service approaches the intolerable. For that reason, it is better to design for full coverage and to levy a significant surcharge for each system expansion (to include those who turned down the opportunity to join once before). * Pay close attention to existing levels of social capital in target communities. Adjusting the approach to service delivery according to existing levels of social capital should prove beneficial. Community-based W&S services are likely to perform better and have stronger impacts in communities with high levels of social capital. The existence of social networks improves group organization and service functioning. In communities with low levels of social capital, special efforts, and perhaps more time, are necessary to motivate and mobilize community members. Success in one community activity often leads a village to success in a subsequent activity. 43 Annex 1 Objectives and Components of the Surveyed Projects Project Objectives Karnataka Rural Water Supply and Environmental Sanitation Project. The main objective of the project was to raise the standard of living in rural villages by improving peoples' access to potable water and providing better environmental sanitation-thereby improving health and productivity. A secondary objective was to develop replicable models for building rural water systems, coordinating water distribution, promoting environmental sanitation and health, and improving the communication of water and health-related programs with rural villagers. Additionally, the project was to enhance village-level planning and strengthen local institutional capacity. As noted in Chapter 2, the Karnataka project was clearly more detailed than the Maharashtra project in terms of community participation: although participatory elements were only considered when the project was quite far along in its development process, the design took the role of the community more into account in project preparation and implementation, especially regarding the responsibilities of the committees in facility O&M. Maharashtra Rural Water Supply & Environmental Sanitation Project. The main and secondary objectives of the project were largely the same as those of the Karnataka project. The design of the Maharashtra project was, however, quite vague in the area of communities' actions on the local level and on their role in various aspects of the project cycle. Project components Karnataka The project comprises the following components: (1) Rural Water Supply. Including construction of new and rehabilitation of existing water supply schemes in about 1,200 villages, assistance for PHED's borewell program, leakage repair works, a water quality monitoring program and measures for groundwater recharge; (2) Environmental Sanitation. Including construction of environmental sanitation facilities in all the project villages based on demand, covering sullage drainage, pit latrines, washing platforms, cattle troughs, bathing cubicles, street bins and biogas plants; (3) Institution Building and Project Support. Including community development programs for creation of village water supply and sanitation committees, comprehensive training programs and strengthening of the project management, PHED and ZPs; and (4) Health communication. for creation of greater community awareness and demand for improved hygiene and environmental sanitation 44 Maharashtra The project includes the following components: (1) Rural Water Supply. The component is expected to include a total of about 75 regional piped water supply schemes and about 170 small individual schemes. The piped schemes range in size from one village serving a design population of 1,000-2,000 to large, regional schemes covering up to 24 villages and serving a design population of over 85,000. Water sources for these schemes are from percolation wells, minor irrigation dams, river intakes, Bombay and Pune Municipal Corporation water mains, and municipal reservoirs. The hand pump program would include drilling, installation of 1,700 new pumps, rejuvenation of existing wells through hydrofracturing and the replacement of 1,300 India Mark II hand pumps which are beyond repair. (2) Environmental Sanitation. Combined with the health communications component, it would have the objective of developing new strategies to increase demand and usage, and to implement those strategies through innovative programs in a flexible manner to allow the implementing agencies to adjust the program on the basis of lessons learned during implementation. This component would include small scale surface drains for sullage, soak pits, individual latrines and street bins: - The construction of surface drains to carry sullage will be undertaken in all villages with design populations over 10,000 and in selected villages with populations over 5,000 in which the soil conditions or population congestion preclude the use of soak pits. The drainage systems would include main line drains, branch lines in by-lanes, cross drains and disposal pits. - The sanitation subcomponent would install on a demand basis low cost latrines in villages with piped water supply schemes. First priority would go to villages with populations above 1,000 which are receiving piped water supply schemes under the project. Second priority would go to other villages in the districts with populations over 1,000 which have existing piped water supply schemes. - Refuse bins for solid waste removal would be provided as an incentive program to the large villages in which drainage schemes have been constructed and where 50% of the households have adopted low cost latrines. (3) Health Communications Program. This component has four main objectives. - Its first objective is to create greater community awareness and demand in the following areas: (i) personal and fanily hygiene; (ii) promotion of latrine usage; (iii) consumer awareness of the requirements and importance of O&M of water supply schemes; and (iv) community awareness of their responsibility for and the importance of the maintenance of drains and soak pits. - The second objective would be to redefine the job descriptions and responsibilities of District Extension and Media Officer (DEMO), Block Extension Officer, Health (BEOH) and the Multipurpose Worker, to give them a greater role in promoting environmental sanitation and related to health education. - The third objective would be to create greater awareness of water related and environmental sanitation problems among health personnel through revised training curriculum. 45 - The fourth objective is to strengthen the water quality monitoring system in rural areas. This would include the strengthening of the staff and facilities of the state, regional and district Public Health Laboratories (PHL) in all 30 districts in the state. (4) Institutional Strengthening. This component consists of four main subcomponents: - Project management; includes the employment of consultants to support the Project Planning and Monitoring Unit (PPMU). - Community participation; involves the hiring of NGOs to undertake implementation of aspects of the environmental sanitation and health communication components. - The strengthening of MWSSB and the Zilla Parishads would be through the provision of office buildings, equipment and materials. - Training; would cover personnel from MWSSB and Zilla Parishad Engineering Wings, village mechanics, village masons, and NGOs' staff. 47 Annex 2 Table of Performance Indicators 1 2 3 4 5 6 7 8 9 10 11 12 13 Karnataka A. Negathihalli 0.47 3.59 0.53 3.00 0.12 0.12 0.00 0.00 0.59 0.71 4.00 7.89 0.65 Chandakavadi 0.12 0.73 0.35 2.00 0.12 0.00 0.00 0.00 0.17 0.74 3.00 7.22 0.46 Homad mandoddi 0.00 0.00 0.00 1.00 0.00 0.20 0.00 0.00 0.60 1.00 0.00 6.72 0.28 Madhur 0.28 2.12 0.08 5.00 0.04 0.08 0.00 0.00 0.46 0.96 6.67 10.60 0.76 Daginakatte 0.08 0.48 0.20 3.00 0.00 0.12 0.00 0.00 0.28 0.84 8.50 7.40 0.44 Bylanarsipura 0.08 0.48 0.12 4.00 0.08 0.00 0.04 0.11 0.44 0.79 0.00 6.20 0.32 Kembliganahalli 0.16 1.12 0.04 0.00 0.04 0.27 0.12 0.05 0.41 0.77 7.50 8.88 0.76 K. Ankanahalli 0.04 0.27 0.12 5.00 0.08 0.04 0.00 0.00 0.29 0.25 4.00 7.75 0.27 Thyamagondlu 0.24 1.76 0.08 3.00 0.00 0.00 0.04 0.00 0.43 0.86 17.75 3.48 0.12 Kemmale 0.10 0.48 0.00 1.00 0.10 0.08 0.00 0.00 0.33 0.33 4.00 6.74 0.52 Bendigere 0.52 4.16 0.08 3.00 0.12 0.24 0.08 0.06 0.63 0.88 7.00 6.44 0.20 Sreedaragadda 0.56 4.00 0.08 1.00 0.04. 0.47 0.04 0.00 0.93 0.87 5.60 5.92 0.08 Median 0.14 0.93 0.08 3.00 0.06 0.10 0.00 0.00 0.44 0.81 4.80 698 0.38 Mean 0.22 1.60 0.14 2.58 0.06 0.13 0.03 0.02 0.46 0.75 5.67 7.10 0.41 1 = Number of groups 8 = Construction participation (> 2 days) 2 = Putnam index 9 = Construction monitoring 3 = Two community activities 10 = O&M any payment 4 = Design performance 11 = Good water group 5 = Hygiene training 12 = Functioning index 6 = Design participation 13 = Improved health 7 = Satisfactory women participation 48 1 2 3 4 5 6 7 8 9 10 11 12 13 Maharashtra Shahajahanpur 0.24 1.92 0.16 0.00 0.48 0.28 0.04 0.16 0.17 0.84 0.00 5.76 0.76 Pimpalwadi 0.16 1.32 0.00 0.00 0.28 0.00 0.00 0.12 0.00 0.24 5.00 6.72 0.32 Sawargaon 0.04 0.32 0.00 0.00 0.12 0.00 0.00 0.05 0.00 0.00 0.14 1.76 0.52 Bhopala 0.31 2.15 0.00 0.00 0.31 0.04 0.00 0.08 0.00 0.28 0.00 5.98 0.54 Bhoyara 0.17 1.29 0.13 0.00 0.13 0.18 0.04 0.06 0.00 0.00 0.00 2.85 0.54 Salgara 0.26 1.63 0.07 0.00 0.41 0.13 0.04 0.09 0.09 0.78 0.00 2.16 0.52 Chata 0.46 3.04 0.21 0.00 0.13 0.06 0.42 0.00 0.00 0.29 5.71 3.10 0.42 Asangaon 0.16 1.16 0.00 3.00 0.28 0.09 0.04 0.00 0.04 0.96 0.00 3.04 0.40 Eksal 0.87 4.60 0.13 1.00 0.33 0.38 0.07 0.08 0.08 0.00 11.83 0.20 0.27 Dedale 0.13 1.13 0.00 1.00 0.00 0.08 0.07 0.00 0.00 0.00 7.50 4.60 0.33 Tarapur 0.40 2.88 0.24 4.00 0.08 0.65 0.28 0.04 0.04 0.87 7.56 3.44 0.52 Waghivali 2.07 0.80 0.00 0.00 0.47 0.36 0.33 0.21 0.25 0.00 8.37 8.20 0.20 Dimbe 0.07 0.33 0.33 0.00 0.00 0.00 0.00 0.00 0.13 0.00 2.50 0.00 0.00 Waksai 1.12 6.96 0.08 0.00 0.40 0.29 0.28 0.05 0.10 0.19 3.05 7.92 0.84 Bamanwadi 0.60 3.67 0.13 4.00 0.67 0.67 0.47 0.00 0.08 0.93 0.00 10.20 0.80 Sulewadi 1.27 8.33 0.20 1.00 0.73 0.67 0.67 0.07 0.29 0.67 0.11 7.60 1.00 Bhose 1.88 1.44 0.72 1.00 0.44 0.52 0.40 0.05 0.24 0.92 0.00 6.64 0.88 Rangaon 0.00 0.00 0.00 0.00 0.33 0.07 0.00 0.00 0.00 0.00 0.00 7.40 .0.60 Nimboori 0.00 0.00 0.00 2.00 0.87 0.00 0.00 0.20 0.00 0.00 0.41 5.00 0.87 Gharod 0.16 1.44 0.00 4.00 0.16 0.00 0.00 0.00 0.00 0.75 0.00 4.00 0.12 Median 0.25 1.44 0.08 0.00 0.32 0.11 0.04 0.05 0.04 0.26 0.12 4.80 0.52 Mean 0.52 2.22 0.12 1.05 0.33 0.22 0.16 0.06 0.07 0.39 2.61 4.83 0.52 1 = Number of groups 8 = Construction participation (> 2 days) 2 = Putnam index 9 = Construction monitoring 3 = Two community activities 10 = O&M any payment 4 = Design performance II = Good water group 5 = Hygiene training 12 = Functioning index 6 = Design participation 13 = Improved health 7 = Satisfactory women participation