-. A . - 38294 ; I' l j, j'~' ' ; xpf 4 't i . t .",e'4 . ', ,, b 5 i . ;' ; @t- S ,. ;w- ^ ! r - - .r; ., . + - .'T . - . .-;.- S. -- -'- ' '44-, ** -' 4-.~ *1,- a.>V * o 2.- 4 -- ...i 1F; __ t '. - -'- .-&. -. r- - - - | ''4 - I - *A - - 4 0 - N !lMA 4F 44-X - >L44 - " ;! - - , ; '' i.- - l :£ 7: s.~' - ;-*-"t g ; .K - -_ -3 - -or * -- ------ -- Environmentally Sustainable Development Series Proceedings Culture and Development in Africa: Proceedings of an International Conference ESD Proceedings Series no. 1 (Also in French) Valuing the Environment: Proceedings of the First Annual International Conference on Environmentally Sustainable Development ESD Proceedings Series no. 2 Overcoming Global Hunger: Proceedings of a Conference on Actions to Reduce Hunger Worldwide ESD Proceedings Series no. 3 Traditional Knowledge and Sustainable Development: Proceedings of a Conference ESD Proceedings Series no. 4 The Human Face of the Urban Environment: A Report to the Development Community ESD Proceedings Series no. 5 The Human Face of the Urban Environment: Proceedings of the Second Annual World Bank Conference on Environmentally Sustainable Development ESD Proceedings Series no. 6 (Forthcoming) The Business of Sustainable Cities: Public-Private Partnerships for Creative Technical and Institutional Solutions ESD Proceedings Series no. 7 (Forthcoming) Enabling Sustainable Community Development ESD Proceedings Series no. 8 (Forthcoming) Studies and Monographs (formerly Occasional Papers) The Contribution of People's Participation: Evidence from 121 Rural Water Supply Projects ESD Occasional Paper Series no. 1 Making Development Sustainable: From Concepts to Action ESD Occasional Paper Series no. 2 Sociology, Anthropology, and Development: An Annotated Bibliography of World Bank Publications 1975-1993 ESD Studies and Monographs Series no. 3 The World Bank's Strategy for Reducing Poverty and Hunger: A Report to the Development Community ESD Studies and Monographs Series no. 4 Monitoring Environmental Progress ESD Studies and Monographs Series no. 5 (Forthcoming) T he Contribution of People's Participation Evidencefrom 121 Rural Water Supply Projects Deepa Narayan A Environmentally Suistainable Development Occasional Paper Series No. 1 ESD The World Bank, Washington, D.C. (C 1995 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing July 1995 This report has been prepared by the staff of the World Bank. The judgments expressed do not necessarily reflect the views of the Board of Executive Directors or the governments they represent. The drawing on the cover was prepared to generate discussion around local water problems. Such drawings are used to focus members of a community on what their water problems are and to mobilize them to work together toward solutions. The cover design is by May Eidi. Deepa Narayan is a social scientist in the Environment Department of the World Bank. Library of Congress Cataloging-in-Publication Data Narayan-Parker, Deepa. The contribution of people's participation: evidence from 121 rural water supply projects / by Deepa Narayan. p. cm - (Environmentally sustainable development occasional papers series ; no. 1) Includes bibliographical references. ISBN 0-8213-3043-8 1. Water-supply, Rural-Developing countries-Management-Citizen participation. I. Title. II. Series. HD1702.N37 1994 331.91'009172'4-dc2O 94-34109 CIP Contents Foreword vii Acknowledgments viii Executive Summary 1 CHAPTER 1 Introduction 5 CHAPTER 2 The Concept of Participation 7 What Is Participation? 7 Why Participate? 9 Who Participates? 10 How Do People Participate? 10 Inducing Collective Action 11 CHAPTER 3 Research Methodology and Project Descriptions 13 Methodology 13 Measuring Variables 16 CHATEER 4 Role of Beneficiary Participation in Project Effectiveness 21 Participation and Project Effectiveness 21 Participation and Other Performance Determinants 23 Participation and Other Project Outcomes 25 How Important Is the "Halo Effect"? 28 Proximate Determinants and Participation 29 Project Experience with Beneficiary Participation 33 Policy Lessons 39 iii iv The Contributtion of People's Participation: Evidencefrom 121 Rutral Water Sutpply Projects CHAPTER 5 Factors Affecting Beneficiary Participation 41 Degree and Elements of Participation 41 Determinants of Participation 44 Policy Lessons 62 CHAPTER 6 Translating Lessons into Design Features 65 Selection Criteria 65 Demand Assessment 67 Institutional Framework 67 Funds and Financial Flow 69 Technology and Choice of Service Level 71 Approaches to Planning: Master Plan or Learning Process? 71 Monitoring and Evaluation 73 CHAPTER 7 Conclusions and Recommendations 75 What Do We Know about Participation? 75 Why Hasn't Participation Occurred on a Large Scale? 76 What Can We Do Differently? 79 APPENDIXES 1A Participation with Other Direct Inputs into Water Project Effectiveness 83 1B Participation with Eighteen Direct and Indirect Inputs (Model 3) into Project Outcomes 84 2 Projects Rated High in Overall Effectiveness 86 3 Summary of Differences between Blueprint and Learning Process Approaches 98 4 Centre for Social Research: Indicators for Sociological Monitoring System- Karonga Lakeshore Integrated Rural Groundwater Supply Project 99 Notes 101 Bibliography 105 Tables 3.1 Background characteristics of 121 water projects 16 3.2 Measures of performance outcomes and participation 17 3.3 Nonparticipation determinants of outcomes 18 3.4 Determinants of participation 19 4.1 Overall project effectiveness as a function of overall beneficiary participation 23 4.2 Water project performance outcomes as a function of overall beneficiary participation 25 4.3 Relationship of participation to environmental benefits and equality of access 26 4.4 Outcomes not specific to water systems, as a function of overall beneficiary participation 27 4.5 Tests for intercoder "halo effects" in coding project effectiveness and participation 28 4.6 Bivariate correlations between outcomes and proximate determinants 29 4.7 Impact of overall beneficiary participation on the institutional proximate determinants of project performance 30 4.8 Impact of overall beneficiary participation on the physical outputs related to project performance 31 4.9 Proximate determinants of water project outcomes as a function of beneficiary participation in different stages . 32 5.1 Correlations of elements of participation with overall beneficiary participation 43 Contents v 5.2 Relationships between demand and participation outcomes and elements 48 5.3 Relationships between beneficiary capacity and participation outcomes and elements 54 5.4 Relationships between client orientation and participation outcomes and elements 62 Figures 3.1 Model of the relationship of participation to outcomes and other factors 19 4.1 Model for testing the relationship of participation to performance outcomes 22 4.2 Cross-tabulation of overall project effectiveness with beneficiary participation by number of projects 23 5.1 Model of relationships between beneficiary participation and its determinants 42 Boxes 4.1 Impact of nonparticipation factors on outcomes for water projects, the environment, and equality outcomes 24 4.2 Paraguay: institutional and fiscal reforms 34 4.3 Kenya: community participation with a parastatal agency 35 4.4 Pakistan: community-based rural water systems 36 4.5 Uganda: decentralizing maintenance 37 5.1 Rwanda: government control over community decisionmaking, or a losing proposition? 44 5.2 Indonesia: evolution of a demand-led strategy 47 5.3 Yemen: role of local leaders in community mobilization 50 5.4 Malawi: water committees 51 5.5 Indonesia: the story of Mutis 52 5.6 Nigeria: difficulties in managing incentives 56 5.7 Tanzania: master plan incompatibility with demand orientation 57 5.8 Tunisia: combining local knowledge with Geographic Information Systems 57 5.9 Togo: investing in capacity building 59 5.10 Kenya: community-managed water utilities 61 6.1 Sri Lanka: guidelines on community preparation and community commitment 66 6.2 Nepal: a new funding mechanism 74 I Foreword T his report is the first of a new series of agency problem was reluctance to give up con- Occasional Papers to be issued by the trol or to invest in developing the capacity of Office of the Vice President for Environ- local organizations. mentally Sustainable Development. Since the The results of this study have profound impli- essence of sustainable development is helping cations for the way the World Bank supports its people make their own decisions and take res- partners in planning and implementing develop- ponsibility for their own welfare, I am pleased to ment programs. Among the lessons gleaned from launch this series with a report highlighting the the study are these: obtaining local participation importance of local participation and social orga- in decisionmaking about development is sound nization in the success of rural water supply pro- business practice, and special measures are need- grams. This work clearly establishes the need to ed to ensure that the marginalized are reached in invest in social infrastructure if physical infra- the participatory process. Participation can hap- structure is to be used effectively. pen only in the right policy environment, in which The study is based on systematic quantitative user demand is primary. Even when participation and qualitative analysis of 121 rural water sup- is assured in planning, agencies must listen and ply projects funded by many different agencies learn as projects are implemented. in countries throughout the developing world. These principles are clear, and their implica- The analysis consistently shows that beneficiary tions reach well beyond rural water supply pro- participation was more significant than any jects. The challenge is to act on these principles other factor in achieving functioning water sys- and to place people at the center of development. tems and in building local capacity. The degree of participation depended on local demand and organization, and particularly important were Ismail Serageldin agency autonomy and the degree to which agen- Vice President cies accepted and monitored the goal of achiev- Environmentally Sustainable Development ing local participation. The most common The World Bank vii . - _.,_._._._-_ . Acknowledgments T his study began in 1991 as part of the con- * Phase I-Collection of evaluation reports: Douglas tribution of the United Nations Develop- Keh and members of the Program's Regional ment Programme (UNDP)-World Bank Water and Sanitation Groups, particularly Gunnar Water and Sanitation Program to the Bank-wide Schultzberg and Sunita Chakravarty. Participatory Development Learning Group. The * Phase II-Development of coding methodology: study is a collaborative effort between the United Kurt Finsterbusch, assisted by Warren Van Nations Development Programme-World Bank Wicklin and Elhum Haghihat. In addition to Water and Sanitation Program, and the Social undertaking the coding process, Elhum Haghihat Policy and Resettlement Division of the Environ- performed all the statistical work. ment Department of the World Bank. It was * Phase III-Model building and model testing: Lant financed by the Governments of Sweden and Pritchett, with research assistance from Jon Isham, Norway guided the testing process for the model, using Particular thanks to John Blaxall, manager of multivariate regression analysis. Lant Pritchett the UNDP-World Bank Water and Sanitation also helped organize the statistical evidence in Program, and to Aubrey Williams, coordinator of chapters 4 and 5. Ellen Tynan helped develop the the Bank-wide Participatory Development matrix included in appendix 2 and provided Learning Group, for trusting that something use- assistance with analysis of the qualitative aspects ful would come from the merger of methodolog- of the study. ical tools from economics and sociology. The document benefited from the comments Several people played important supportive and review of John Blaxall, Norman Uphoff, Lant roles in the early days of the project, when the Pritchett, Jannik Boesen, Kurt Finsterbusch, methodology was still evolving: David Beckman, Warren Van Wicklin, and Aubrey Williams. Grace Michael Cernea, David Howarth, Lawrence Sarin and Lidia Tokuda typed and formatted the Salmen, and Michael Garn. manuscript. The manuscript was edited by Laurie The study developed in three phases. I would Edwards and Donna van Lear, under the overall particularly like to acknowledge the contribu- guidance of Alicia Hetzner, David Kinley, and tions of the following people who played invalu- Virginia Hitchcock. The publication was desk- able roles. topped by American Writing Corporation. viii Executive Summary F ield observations have led many people to jects and within the lifetime of individual pro- believe that beneficiary participation in jects, sometimes over a decade or more. The decisionmaking can contribute greatly to quantitative data came from content analyses the success of development projects. When peo- done by two independent coders for each report, ple influence or control the decisions that affect covering 149 variables. Intercoder reliability was them, they have a greater stake in the outcome high on key variables, and testing for so-called and will work harder to ensure success. But the halo effects produced no significant change in the evidence supporting this reasoning is qualita- results. Preliminary correlation analyses and fac- tive-some would say anecdotal-so that many tor analyses to reduce the number of variables led practitioners remain skeptical. Even when they to a model that applied the framework of collec- accept that participation is important, the quali- tive action to management of rural water as a tative approach does not offer much guidance on common property resource. To move beyond cor- how to promote participation in large-scale pro- relations among indicators toward causality, grams. Therefore three questions need to be multivariate regression analysis was used to test addressed systematically: To what degree does the collective-action framework. Additionally, participation contribute to project effectiveness? the twenty projects that were scored as most Which beneficiary and agency characteristics fos- effective were analyzed to determine their key ter the process? And, if participation does benefit design features. project outcomes, how can it be encouraged through policy and project design? How Much Does Participation Contribute To answer these questions, researchers studied to Project Effectiveness? evaluations of 121 completed rural water supply projects in forty-nine developing countries Results obtained from quantifying levels of par- around the world. Eighteen different agencies ticipation establish that participation contributed supported the projects, which employed a variety significantly to overall project effectiveness, even of technical approaches. The results are clear: ben- after controlling for eighteen direct and indirect eficiary participation contributed significantly to determinants of outcomes. The proportion of project effectiveness, even after statistically con- water systems in good condition, overall eco- trolling for the effects of seventeen other factors. nomic benefits, percentages of the target popula- The results are based on quantitative and sys- tion reached, and environmental benefits rose tematic qualitative analyses of data across pro- significantly with participation. Participation T 2 The Contributtion of People's Participation: Evidencefrom 121 Rural Water Supply Projects also helped assure equality of access to facilities, tion details-that is, the what, when, how, and although its effects were less pronounced in this where of implementation. outcome than in the others already cited. * Nongovernmental organizations (NGOs) fig- Not surprisingly, the analysis reveals that par- ured in half of the success sample (that is, ten ticipation fostered individual and community of the twenty most effective projects), empowerment. It also promoted new water-man- although they represented only 15 percent of agement and organizational skills in the commu- the total sample. nity. Finally, it strengthened local organizations, * Issues concerning physical infrastructure and which then went on to undertake other develop- technology have been addressed more effec- ment activities. tively than ones concerning the social organi- One shortfall of the participation process zation necessary for managing the physical deserves mention here. Although most projects works. identified women as a target group, only 17 per- cent of the projects achieved high levels of Which Beneficiary and Agency involvement by women. High levels of benefi- Characteristics Foster Participation? ciary participation did not necessarily mean that women's participation was high. Women in most Participatory processes are nonlinear and itera- rural areas face many constraints to participating tive. The process interacts with intermediate out- in development projects, and unless their involve- comes, which influence the process, and so on. It ment was specifically targeted and resources to is nevertheless possible to identify the critical ele- ensure it were invested, it did not occur. ments of the participation process, as well as the determinants of the eventual participation levels Participation at Different Stages of beneficiaries. These relationships were ana- lyzed through multivariate regression analysis, The study examined how final outcomes were which showed that the three intermediate ele- affected by the quality of outcomes at each of the ments most highly correlated with overall levels stages of a project cycle, namely, design, imple- of beneficiary participation were: mentation, construction, and maintenance. A * User investment in capital costs series of multivariate regression analyses pro- * Local ownership and control duced two major findings: * Agency responsiveness to feedback. * Participation was the single most important Therefore participation is determined by char- determinant of overall quality of implemen- acteristics both of the beneficiaries and of the tation. agency and project. * The impact of participation throughout the At the beneficiary level, the two key charac- project cycle was significantly greater than it teristics determining participation were commit- was during any single stage. ment before construction, or demand, and the degree of organization of the beneficiaries. When ben- In-Depth Qualitative Analysis eficiaries made a commitment before the project was implemented, they became involved in pro- Statistics tell only part of the story. Individual ject activities, including investment in capital and projects had to be analyzed in depth to better recurrent costs. Projects flourished when the ser- understand how participation worked in practice vice provided matched what consumers wanted and to evaluate uncoded agency factors. Four and were willing to pay for. Forms of organiza- findings are worth highlighting: tion-the other beneficiary characteristic impor- * The forms of effective participation varied tant for participation-varied even within tremendously, ranging from representational communities, but common to all forms were committees of users to domination by leaders internalized, self-enforced rules and regulations and elites, and from direct involvement in that governed criteria for membership and access construction to supervision of contractors. to (or exclusion from) the facility. * Effective participation did not result when For agencies, the two most significant charac- agencies retained control over implementa- teristics were relative autonomy and degree of client Execuitive Summary 3 orientation. Three client-orientation strategies oriented, they must become demand-responsive emerged as essential for effective participation: and participatory. Achieving these shifts will * Making beneficiary participation a goal to be require changes at the sectoral policy level and monitored and evaluated, and rewarding staff the community level. for achieving it Research for this study found several steps in * Using local knowledge for designing and the design of large-scale projects to be essential for implementing projects effective outcomes: objectives must be clear and * Investing in building the capacity of local peo- specific; financing mechanisms must be respon- ple, including providing information to help sive to demand; a community self-selection people make informed choices and allowing process must be used to undertake water- time for communities to organize themselves. improvement projects; the planning process must The qualitative analysis revealed also the focus on learning and applying lessons in imple- three most common hindrances to participation, mentation; intermediary organizations that work caused by even some well-meaning agencies that to support communities must be clearly identi- tried to elicit participation: fied; and substantial investments in strengthening * Unwillingness to give up control over imple- the capacity of community-based organizations mentation details must be made. * A lack of incentives for staff to support client Several rural water projects financed by the orientation through new institutional arrange- World Bank are now moving forward; they are ments based on the principle of local participation, con- * A unwillingness to invest resources for build- trol, and authority, and supported by an approach ing the community capacity or social organi- to project design that emphasizes learning over zation to manage the physical infrastructure. blueprints. It is too early to draw conclusions about performance, but two facts are already evi- How Can Participation Be Encouraged dent. First, although large-scale projects can be through Policy and Project Design? designed with a built-in commitment to demand- responsive approaches and active "experimenta- Sector policies, project design, and implementa- tion," there is no standard model for doing so. tion strategies must change fundamentally in Second, in the absence of a standard model-that rural water supply programs. They must encom- is, where learning relies on flexibility, adaptation, pass a shift from centralized ownership of sys- and necessarily short planning horizons-moni- tems to local ownership and control. Approaches toring and evaluating are even more important must change as well: instead of being supply- for success than they are in traditional projects. 7 wl ..,; CHAPTER 1 Introduction P eople's participation in decisionmaking tions-institutions-and organization. The first and local ownership results in effective and challenge for agencies is to create an incentive for sustainable rural water systems. This belief staff to work in partnership with hundreds of has played a central part in the shift in institu- communities. The second task for agency staff is tional strategies from supply-driven to demand- to enable communities to make informed choices, driven approaches, which respond to the felt organize themselves, initiate collective action, needs and aspirations of users, especially the and manage and choose from a menu of water poor. However, quantitative evidence of the effi- supply options (technology and management) cacy of participation in determining project effec- that the agency offers. tiveness, relative to other factors, has been Although the agency task has changed dra- missing. The present study is a step toward fill- matically over the years, that fact has seldom ing this gap. It is based on multivariate analyses been recognized or acknowledged by the agen- of data derived from evaluation reports of 121 cies themselves. Hence, the agencies and their rural water supply projects implemented by dif- competency, organization, structure, and man- ferent agencies around the world. Project reports agement-by-blueprint style have remained were supplemented by in-depth anthropological largely the same. The mismatch between the task and sociological studies and by other project and the mandate, ability, and competence of the reports received after data analyses had been agency has resulted in many unsuccessful gov- completed. ermnent attempts to induce participation. The In most developing countries, public sector key question, then, is: How can organizations agencies provide rural infrastructure. Poor pub- change to induce participation in collective lic sector performance has led to a widespread action? search for institutional alternatives and means to This study examines efforts to induce partici- increase the accountability of the public sector. In pation as a means to create effective rural water the rural water subsector, the search has been for systems and to build the local capacity to manage strategies to increase users' "exit" and "voice" them. Beneficiary participation can be brought options and to restructure the sector so that sup- about in several ways: directly, through participa- pliers have incentives to match the demand of tion in decisionmaking; indirectly, through lead- users. ers; or through representation on committees or The problematic issue therefore is not tech- boards. Participation of beneficiaries can be facil- nology and construction but rules and regula- itated through extension workers, local govern- 5 6 The Contribuition of People's Participation: Evidencefrom 121 Ruiral Water Suipply Projects ment units, nongovernmental organizations ology, describes the coding process and method- (NGOs), and the private sector. Many factors ological limitations, and lists the variables influence beneficiary participation, including the included in the study Chapter 4 addresses the immediate and broader policy context; client contribution of participation to project outcomes, characteristics (including felt need); and agency using multivariate analyses. It examines the con- characteristics, such as flexibility, responsiveness tribution of participation to project outcomes and to clients, and willingness to invest in the sound studies the robustness of the relationship by con- organization of communities.1 This study specifi- trolling for other determinants of project success. cally addresses the following questions: After establishing the importance of benefi- * Does people's participation contribute to pro- ciary participation, chapter 5 examines the deter- ject effectiveness? minants of participation itself, both project * How important is this contribution, relative to characteristics and beneficiary characteristics, other factors? and describes intermediate steps toward partici- * What factors and strategies influence partici- pation. Chapter 6 discusses the implications of pation in collective action? the findings for the design of future World Bank * What are the lessons for the design of large- projects and presents examples of some strategies scale water projects? now being used in recently designed projects. * What are the implications for policy reform? Chapter 7 highlights conclusions and policy rec- Chapter 2 focuses on the conceptual frame- ommendations. The appendixes include a matrix work and defines terminology, including partici- of findings from analysis of the twenty projects in pation; it also outlines the model used in the the study that received the highest ratings for study. Chapter 3 reports on the research method- overall effectiveness. CHAPTER 2 The Concept of Participation articipation means different things to dif- ends, not as an end in itself; the goal therefore is to ferent people. This chapter clarifies the optimize participation to achieve the desired pro- what, why, who, and how questions of par- ject goals, not simply to maximize participation. ticipation and collective action and outlines the The desired goals in rural water supply projects conceptual model used to guide the study. include achieving improved water supply systems and developing the human, organizational, and What Is Participation? management capacity to solve problems as they arise in order to sustain the improvements. Definitions of participation abound (Cohen and The principle underlying participation-to Uphoff 1977; Korten 1980; Paul 1987; and Ghai give people a voice-is constant, yet the choices and Hewit de Alacantara 1990). All of them that people make vary infinitely. Thus, a commu- include in some measure the notions of con- nity may decide to subcontract maintenance to an tributing, influencing, sharing, or redistributing independent mechanic rather than to undergo power and of control, resources, benefits, knowl- training and take turns doing the work. A water edge, and skills to be gained through beneficiary users' group may choose to dissolve the organi- involvement in decisionmaking. There is also zation or to define new goals after the first ones much debate among practitioners and in the lit- have been met. For example, when construction erature about whether participation is a means or is complete, a water committee may transform an end, or both (World Bank 1992; Picciotto 1992). itself to undertake sanitation construction, to For the purposes of this study, participation is build a football field, or to branch into children's defined as a voluntary process by which people, education, depending on the commonality of including the disadvantaged (in income, gender, interests. A large community group may divide ethnicity, or education), influence or control the into smaller, functional subgroups, with the decisions that affect them. The essence of partici- larger group meeting only occasionally. Alter- pation is exercising voice and choice. natively, people may informally nominate lead- This conception does not assume that there is an ers to represent their interests. ideal level of participation to be achieved. The most effective form of participation varies, but over the Levels of Participation long run sustainability will depend on minimizing transaction costs in horizontal and vertical interac- Participation is a multidimensional, dynamic tions. Participation is viewed as a means to defined process, which takes varying forms and changes 8 The Contribuitioni of People's Participation: Evidencefrom 121 Rural Water Suipply Projects during the project cycle and over time, based on exclusively or jointly with the external agency, is interest and need. Samuel Paul (1987) usefully a much more intense level of participation which distinguishes among levels of participation, all often promotes capacity building. Decision- four of which may coexist in a project. The first making may be about policy objectives, project two categories present ways to exercise influence; design, implementation, or maintenance, and dif- the other two offer ways to exercise control. The ferent actors may be involved at different stages levels comprise information sharing, consulta- of the project. thus, the decision to participate in tion, decisionmaking, and initiating action. a project may be made by the community, and the choice of technology may be made jointly, after Infornmation Sharing the costs and benefits of the various technologi- cal options have been explained by the agency Project designers and managers may share infor- and understood by the community. mation with clients to facilitate collective or indi- vidual action. The information flow is one-way, Initiating Action from agencies to communities. Although it reflects a low level of intensity, information shar- Initiating action, within parameters defined by ing can positively affect project outcomes by agencies, represents a high level of participation enlarging clients' understanding of specific that surpasses involvement in the decisionmak- issues (for example, by explaining hygienic prac- ing process. Self-initiated actions are a clear sign tices or how groundwater is polluted). Infor- of empowerment. Once clients are empowered, mation sharing may also be one-way in the other they are more likely to be proactive, to take ini- direction, in the form of baseline or feasibility tiative, and to display confidence for undertaking studies wherein information (but not necessarily other actions to solve problems beyond those opinion) is gathered from beneficiaries. Many defined by the project. This level of participation such studies tap local knowledge but do not con- is qualitatively different from that achieved when sult the local clients. clients merely carry out assigned tasks. Institutional options for rural water supply Consultation depend on whether water is treated as a public, private, or common property good, and on the When project designers and managers not only resultant degrees of excludability (the degree to inform clients but also seek their opinions on key which other users can be excluded) and jointness issues, a two-way flow of information develops. or subtractability (the degree to which use by one This two-way flow presents some opportunities affects the overall production cost of use by some- for clients to give feedback to project designers or one else). Similarly, the most appropriate level of managers, who can then use the information participation depends on who owns the water about preferences, desires, and tastes to develop and on who manages the extraction and distribu- designs and policies that achieve a better fit tion of water. The degree to which water can be between agency programs and community managed collectively depends on the ability to demand. Examples of consultation include meth- exclude some, but not others. The degree of joint- ods that tap indigenous knowledge and organi- ness adds complexity to and determines the par- zational forms, such as socioeconomic surveys, ticipants in the negotiations. (For example, in the beneficiary assessments, and willingness-to-pay development of a system for piped water, users at studies. the top and at the bottom of the distribution lad- der need to be involved in negotiating rules and Decisionmaking regulations for the distribution of the water.) Moreover, the moment external agencies inter- Information sharing and consultation generally vene to improve the quantity and quality of water, do not lead to increased local capacity or empow- or to make water more accessible, issues related to erment of local people and institutions, although rural infrastructure and technology choice come they can lead to more effective programs. Client into play and add another layer of complexity to involvement in decisionmaking, however, either issues of decisionmaking and participation. The Concept of Participation 9 Despite continued government investment, input or an independent variable) can contribute the state of infrastructure has deteriorated, espe- to the achievement of four main objectives: effec- cially in developing countries (Israel 1992). This tiveness; efficiency; empowerment; and equity. decline has led to a renewed attempt to focus on the reliability and maintenance of infrastructure. Project Effectiveness To help identify the role that participation plays in infrastructure effectiveness, it is useful to look Project effectiveness is the degree to which stated at the decisionmaking phases and tasks involved project objectives are achieved. Client involve- in the construction of rural infrastructure. ment, direct or indirect, may result in a better Rural infrastructure is developed in four match between what users want and what an broad phases: agency or project offers. Rural water supply pro- * Design or planning jects are considered effective if they increase * Construction access to and reliability of water sources, so that * Operation people have water in the quantity and quality * Maintenance. and with the reliability and convenience they In practice, work overlaps and shuttles back demand. Effective water projects produce health- t and forth among those tasks. For the purpose of related, economic (time savings), and environ- undertaking a rural water supply project, the cat- mental benefits, among others (Narayan 1989; egories of issues confronting the rural infrastruc- Jaganathan 1992). ture concern: Users can facilitate effective water projects in * Decisions to join, ownership issues, and con- several ways at different stages. They may con- ditionality tribute to redefinition of objectives, better project * Choices of technology and service levels design, redesign, site selection, resource mobi- * Costs and financing lization, construction, implementation, and maint- * Decisions about design and construction enance of facilities beyond the life of the project. * Tariff management Beneficiary ownership and control of the project * Water allocation also are often seen as essential elements in estab- * Operation and maintenance lishing effective projects. l * System expansion and replacement. These categories clearly suggest that clients Project Efficiency and agency personnel can be involved to varying degrees in influencing or determining the many Project efficiency measures the relationship different choices to be made in any given project. between a given output and its cost and inputs. ' The tendency in the past was for agencies to dom- Because anticipatory decisionmaking allows inate over community or client choice, sometimes more timely beneficiary inputs, as well as syn- with disastrous results. chronization of agency and client inputs, it may well lead to greater efficiency. Discussion, consul- Why Participate? tation, and information sharing often produce greater consensus about goals and means and Participation engenders financial, social, and more clarity about roles, authority, and ownership psychological costs as well as benefits. Clients or than would otherwise be possible. Consensus and beneficiaries are likely to participate when their clarity in turn reduce conflict and delays, result- benefits outweigh their costs, just as agencies are ing in smoother implementation and lower over- likely to foster beneficiary participation when the all costs. For example, proper identification of benefits of doing so outstrip the costs to the land and water ownership rights and timely agency. However, knowledge about the costs and acquisition of land permit quick construction and benefits of participation remains limited; little completion of piped-water systems. Community guidance about budget allocations appropriate to management or private sector involvement in fee induce participation is available to those plan- collection may entail high transaction costs, yet ning large-scale projects. Nevertheless, from an either of those options may be much more effi- agency perspective people's participation (as an cient than fee collection by a public agency. 10 The Contribution of People's Participation: Evidencefrom 121 Rural Water Supply Projects Empowerment interest. This is the glue that binds people who may otherwise not have much in common in Empowerment is essentially a political concept terms of geography, wealth, power, leadership, that means more equitable sharing (or, redistrib- degree of organization, social cohesion, ethnicity, ution) of power and resources with those who income, gender, or education. Commonality of previously lacked power. Any activity that leads interest may supersede other distinctions, includ- to increased access and control over resources ing the entity of "community" (or village or other and to acquisition of new skills and confidence, administrative label of convenience). thus, when so that people are enabled to initiate action on water service is not differentiated, and alternative their own behalf and acquire leadership, is an sources are practically nonexistent (as in arid empowering activity. The central argument for zones), the rich or elite are as much affected as the participatory processes is that involvement in poor by lack of water. Although men may be decisionmaking lets people exercise choice and interested in an improved water supply for cattle voice more broadly in their lives, as well as in the and women may want the water for domestic more immediate context of development pro- uses, their common need for water can bring grams that benefit them. Empowerment is thus, them to work together to negotiate change with about the capacity building of individuals and service providers. the organizations that support them. Some important client characteristics that influence participation are gender (women, as the Equlity primary managers of domestic water, are more affected by, and therefore more interested in, A major purpose of development assistance is a improved water supplies); income (the poor have more equitable distribution of the benefits of less ability to pay for water, but their need is development. It is well established that devel- greater because they often live the farthest from opment gains tend to be "captured" by those water sources); education (the assumption is that already better off. Despite the gains made dur- the more educated the group of people, the more ing the International Drinking Water Supply and easily they understand the issues and the more Sanitation Decade, over a billion people are still willing they are to take action); knowledge and without access to safe water. The primary goal of skills (when people have some understanding of rural water supply projects is to reach those peo- water issues, technology, and financial manage- ple. When included in the pursuit of this broad ment skills, they can more easily manage goal, beneficiary participation-which pro- improved systems); and social cohesion and organi- motes transparency and accountability-may zation (trust, loyalty, and reciprocity are impor- lead to less capture by the elites and to more tant in undertaking collective action, achieving equitable access to improved water supply, thus, consensus about objectives and rules, and in helping serve the purpose of development assis- effectively managing conflict). tance in general. Also consistently cited as important in effect- ing participation are leadership qualities and the Who Participates? presence of leaders. The presence of local leaders is important to initiating change, and leadership Participation occurs at global, national, subna- qualities among users are important in bringing tional, community, and household levels. The about and sustaining change on a large scale. primary focus of this study is on the participation of beneficiaries, those who are meant to benefit How Do People Participate? from the change brought about by rural water supply development projects. hence, the charac- Evidence shows that outside agencies and orga- teristics of these users (individuals and groups) nizers can induce collective action when an are important because they influence the type of issue deals with a common interest of commu- participation that occurs. nity groups. This process of organizing the poor The most important characteristic that brings so that their voices are heard is a role that many people together to take action is commonality of NGOs have played effectively in addressing a The Concept of Participation 11 variety of needs, from employment and water to service outcomes. Decisionmaking arrangements marketing and housing (examples include the determine who decides what in relation to Aga Khan Support Program in Pakistan, BRAC whom. Three types of rules govern the success of and Grameen Bank in Bangladesh, Amul in collective action: the conditions that establish the India, WALI and Dian Desa in Indonesia, and ability of a group to act collectively to make deci- KWAHO in Kenya). Increasing evidence also sions common to the group, including obstacles shows that many of the roles currently being ful- or opportunities for individual decisionmaking; filled by public sector agencies can be more operational rules of groups regarding use, effectively and efficiently carried out by the pri- including entry and exit rules; and external deci- vate sector. Various aspects of service delivery sionmaking arrangements that affect the ability can be contracted out, for instance, limiting the of groups to act collectively. These categories public sector role to regulating, fostering com- comprise bureaucratic rules and those made or petition, and monitoring performance (Triche enforced by traditional authority, market arrange- 1990). While it is clear that many NGOs can ments, and legislative rules. induce collective action, financial resources The environment created by these rules and (including external assistance) are channeled decisions influences perceived costs and benefits primarily through government agencies. The of users, and thus, either limits or enhances their central questions, therefore, are these: Can gov- behavior and choices. Successful collective action ernment agencies induce collective action on a or choice of cooperative strategies is dependent large scale? If so, under what conditions, and on reciprocity, a mutual expectation of positive with what mechanisms? The challenge to gov- performance among individuals. Free-rider ernment agencies is to restructure their policies, problems-which occur when one individual institutions, and organizational forms so that does not contribute, assuming that others will- public agencies are accountable for their perfor- affect collective action. In this environment, the mance and have the incentive to respond flexi- degree to which members monitor one another's bly and quickly to people's demand, either behavior, hold one another accountable to certain directly or through involvement of the private standards of behavior, and successfully apply sector, including NGOs. sanctions becomes important. If reciprocity is abandoned, patterns of interaction may include Inducing Collective Action deceit, threats, intimidation, and collapse of col- lective action arrangements. Several factors, which can be divided into three Over time, as learning takes place, poor out- broad categories, affect the success of collective comes influence the pattern of interaction, which action (Bromley 1986; Oakerson 1986; and Wade may eventually lead to changes in rules or deci- 1980). Those categories concern characteristics of sionmaking arrangements. For example, a group the natural resource, of the clients, and of the of water users in Rwanda who managed their agency inducing change. The interaction of peo- own system changed rules when they discovered ple and agencies determines outcomes, and the that women were spending as much time collect- attributes of the natural resource determine the ing fees as they had earlier spent collecting water boundaries, which in turn influence the nature of (Yacoob and Walker 1991). Local government in the interaction. the Kwale district in Kenya instituted a legal Oakerson (1986) provides an analytically use- transfer of handpump ownership to communi- ful framework for understanding the conditions ties when they found that, despite training, com- under which collective action works. Physical munities expected the government to repair (that is, natural) and technical attributes, espe- pumps (Narayan-Parker 1988). In Kumasi, cially those related to degree of jointness, exclud- Ghana, despite a wait of several years, only 300 ability, and differentiation, set the parameters households had been connected to a conven- within which groups and agencies must operate. tional sewerage system meant for 3,000 house- Within these parameters, the most important fac- holds; the city utilities finally adopted a different tors are the rules and decisionmaking arrange- approach-strategic sanitation planning-and ments that affect behavioral interactions and offered different service levels (Wright 1993). a CHAPTER 3 A.. Research Methodology and Project Descriptions T he International Drinking Water Supply had to be physical implementation of water sup- and Sanitation Decade (IDWSSD) brought ply, and the report needed to have sufficient investment in hundreds of small and large information to permit analysis. projects. This study takes advantage of the rich The evaluation reports were based on three- to data base available in the evaluation reports of five-week impact assessments made by teams of agencies involved in the IDWSSD. It is based on two to five experienced evaluators who had con- analyses of 121 completed rural water supply ducted field visits. The teams used a variety of data projects executed by eighteen different agencies sources, including extensive interviews and other in Asia, Africa, and Latin America; the resulting field visits with staff and beneficiaries, workshops, hypotheses were followed by more specific test- and reviews of project documentation. ing in selected countries. The primary method- Reports were not screened for participatory ological advance of this study is that it uses a activity in the projects, because the quality of in- large number of case studies and cases, converts depth information on participation was uneven them into ratings on a range of variables, and sta- (understandably so, since the objectives of the tistically tests a model specifying the relationship evaluation reports differ from those of the present of participation to project effectiveness and local research). Often the reporting on management and capacity building as well as to empowerment. community-level practices was less extensive than Carrying out this study required contacting the reporting on outputs, inputs, institutional rela- many of the major international support agencies, tionships, and external context. The reports were NGOs, water and sanitation reference centers, and generally sketchy on the costs, details of the orga- libraries around the world. In addition, regional nization of the local water system, and role of and country offices of the United Nations Develop- women. Fortunately, project evaluation reports ment Programme and the World Bank Water and were supplemented in many cases by anthropo- Sanitation Program collected reports. Out of an iri- logical research studies and socioeconomic sur- tial listing of over 800 references, about 400 reports veys done within the project context either during were received and screened to determine whether the planning stages or in analyses of impact. they were appropriate subjects for analysis. Many of the evaluation reports dealt with Methodology policies, in-depth studies, and institution-build- ing activities. Projects had to meet two criteria to Within the noneconomic social sciences there has be included in the study: the focus of the project been a long tradition of using qualitative data and 13 14 The Contribution of People's Participation: Evidence from 121 Ruiral Water Slupply Projects small numbers of in-depth case studies. groups of fifty, each of which was scored. When Sociologists, psychologists, and anthropologists the subsamples were compared for all variables have increasingly turned to quantification of by nonparametric statistical tests, the authors essentially qualitative data obtained through found statistically significant differences for less open-ended exploration of issues, however, as a than 5 percent of the variables, which is equiva- way to subject their hypotheses to statistical tests lent to meeting a test of significance at the 95 per- and to explore the causal relations among vari- cent confidence level. These results satisfied the ables. One of the key methodological differences researchers that enlarging the sample by another between economics and the other social sciences fifty cases would not significantly change the is the process used in arriving at the numbers. variance of the data being analyzed. Anthropologists, for instance, start with an open- The large sample size allowed study of a mul- ended process, explore the universe of possibili- tiplicity of variables, the interactions among them, ties, and do not immediately use structured and their effect on local organization perfor- surveys and questionnaires to get at the numbers. mance. Variables fell into five categories: environ- Structuring and ordering of variables are induc- mental, structural, functional, and participation tive and done much later in the process of inquiry variables, and exogenous factors. than they are in economics. The major challenge for Esman and Uphoff This stiudy combines model testing, using mul- was to quantify factors and subfactors so that each tivariate analysis of data, with in-depth qualita- of them could be rated with a high degree of reli- tive analysis of particular cases. The study draws ability by trained researchers. Achieving equiva- particularly upon and adapts methodologies lence of meaning across cultures made the task from two research projects, one by Milton Esman more difficult. After extensive consultations, the and Norman Uphoff (1984) and the other by Kurt project leaders developed a system of coding for Finsterbusch and Warren Van Wicklin (1987). fifty-five variables on a 5-point scale. Pearson- A research project supported by USAID (U.S. product moment correlations were used to test the Agency for International Development) was relationships among variables; regression analy- undertaken in 1975 to clarify the concept of par- sis identified the main factors contributing to the ticipation; it led to a five-year effort to opera- largest variance in performance. The study was tionalize the concept and understand its limits, also informed by the published literature on par- potentials, and applications. Many applied stud- ticipation and primary data collection on that ies grew out of collaborative in-depth work in six topic in selected countries. Although the sample countries. The desire for even more quantified, was not random and the scoring of factors was comparative assessments eventually produced essentially subjective, the project yielded useful some innovative efforts in methodology and a patterns on a subject that had not previously been systematic research project on what determined studied quantitatively The results of this project the performance of local membership organiza- were published in 1984 by Esman and Uphoff. tions (Esman and Uphoff 1984). Local Organizations: Intermediaries in Rural Develop- Realizing the limitations of small samples and ment, the published results of this project, remains purely qualitative assessments, Esman and a milestone in the literature on participation. Uphoff decided to do a large-scale, quantified The study found relatively high correlations comparative analysis. They proceeded to identify between effective task performance and the par- 150 case studies in the literature that contained ticipatory orientation of the organization. Tasks enough information to evaluate at least most of included planning and goal setting, conflict man- fifty-five independent and dependent (perfor- agement, resource mobilization and manage- mance) variables. Research assistants summa- ment, service provision, and claim making. The rized all of the qualitative and quantitative data authors also found that under adverse conditions, from the case studies onto a standard protocol, local organizations experience less-than-average which was scored for the different variables, success unless they are highly participatory and using standardized, specified criteria. To test have effective leadership. whether this sample was large enough for ana- The methodology outlined above, which lytic purposes, the sample was broken into three attempts to explore statistically the relationship Research Methodology and Project Descriptions 15 among variables affecting the performance of issues of interest were listed and a few questions local organizations-including participation- devised to address each issue. The preliminary was further refined by Kurt Finsterbusch and questionnaire thus developed consisted of over Warren Van Wicklin (1987). Rather than focus on 200 different items. Based on interviews and feed- local organization, Finsterbusch and Van Wicklin back from sector specialists, review of the litera- analyzed factors affecting project performance; ture, consideration of the information available in they included participation as one of the vari- reports, and a pilot run on four cases, the ques- ables. An input-output framework guided their tionnaire was eventually reduced to 145 items. analysis. Their model disaggregates input, Multiple items were used to address one concept. process, output, impact, and contextual factors. Most of the items were rated on a scale of 0 The study by Finsterbusch and Van Wicklin through 7, with some of the key performance was based on analyses of evaluation reports for variables scored on a scale of 0 through 10. A code fifty-two USAID projects across sectors. The book was developed, which specified the charac- reports followed a standard format, making it teristics to be considered in assigning a particular easy to develop and apply a coding system. score for each variable. Construction projects constituted most of the Two coders (one male and one female) were sample, but some health, education, and agricul- trained to apply the questionnaire to each project tural research projects were also included. report to produce subjective-but cardinal-rat- The study found only moderate, but statisti- ings until there was a high level of agreement cally significant, correlations between participa- between the coders. The coding of each report tion and project effectiveness. The authors found took between three and four hours, on average. that effective projects involved good communi- Initially the two coders met periodically to com- cation from the project authorities to the public; pare ratings and discuss differences in ratings. the projects were well received by the beneficia- These discussions ensured that the two coders ries, but they were not always participatory. were using the same criteria to arrive at ratings H The main problem with the conclusions of and that particular pieces of information had not Finsterbusch and Van Wicklin is that their study been overlooked. was based on a relatively small sample and included many different types of projects that do Data Limitations not require beneficiary participation to be suc- cessful. The findings therefore may reflect the Although every care was taken to avoid ambigu- nature of the projects chosen rather than the ity to and increase coder reliability through train- importance of participation in its own right. ing, the ratings were subjective and scores do To follow up on the hypothesis that benefi- differ between coders. Since the scores were sub- ciary participation is not equally critical in all pro- jective, coders could agree in general but still jects, Finsterbusch and Van Wicklin expanded the assign different numerical values to a particular sample size to seventy-one projects and did fur- variable. However, as long as a particular coder ther analyses by type of project. They found that was consistent, the difference between coders did the correlations with participation were much not affect correlation analyses. On the other hand, lower for education and health projects (personal if there are differences in direction of rating communication). The expanded study still cov- between coders, the reliability of the rating of that ered a range of USAID-funded projects. variable is called into question. For this reason, two indicators of data relia- Methodology of the Current Sttudy bility were built into the process of data analysis. The first indicator is the intercoder correlation This study borrows from the methodology coefficients for the two sets of coded scores. Most described for the two studies cited above, using of the coefficients were higher than 0.85. The key the conceptual framework developed in Chapter variables of interest in the study had intercoder 2; the aim is to test the contribution of participa- coefficients of 0.9 or above (overall effectiveness, tion to the effectiveness and capacity building 0.96; overall beneficiary participation, 0.92; overall aspects of rural water supply projects. The main community empowerment, 0.90). 16 The Contribuition of People's Participation: Evidence from 121 Rutral Water Suipply Projects The quality of the reports was uneven, and in the 1980s. Overall, 56 percent of the projects information for some of the factors was missing. received some bilateral financing; 26 percent got Coders were therefore asked to assign a confi- multilateral financing; and 15 percent were dence score for each variable as a second measure financed by national and international NGOs; of reliabil-y. Each coder indicated his or her level the remaining 3 percent were completely of confidence in the score assigned to each vari- financed by national governments. Hence, the able on a scale of 1 to 5. Most of these scores, as sample was heavily biased toward externally well as the confidence scores for the key vari- financed projects. ables, were high. Three of the 50 items scored-total costs, pop- Subjective ratings are open to the criticism that ulation reached, and total number of project the "halo effect" colors results. A coder aware of staff-give some idea of size of projects. The costs the hypotheses to be tested is more likely to score of the projects ranged from $500,000 to $250 mil- projects high in participation when the projects are lion; 9 percent of the projects cost more than $25 highly effective, and vice versa-in other words, million; and 17 percent cost less than $1 million.2 all good things go together. To check for the halo Twenty percent of the projects reached more than effect, regression analyses were conducted, using 500,000 people; 47 percent, from 60,000 to 500,000; the input scores of one coder with the output and 33 percent reached fewer than 60,000. Only 76 scores of the other coder. A substantial drop in the projects provided data on the number of staff correlation coefficients and a change in the results would indicate a significant halo effect in opera- Table 3.1. Background characteristics of 121 water projects tion. Regressions were repeated, using nonstan- Frequency dardized and standardized data. The findings Variable Category Number Percent establish that the halo effect does not significantly Region Africa 58 48 alter the results of the coders. (Some findings are Asia 39 32 reported in Chapter 4. For further information see Latin America 24 20 Isham, Narayan, and Pritchett 1994.) Year project ended 1972-80 14 12 1981-85 25 21 The rest of this chapter describes in some 1986-90 72 60 detail the variables included in the model for the 1991-92 10 7 study. Main donor Multilateral donors 32 26 Bilateral donors 68 56 NGOs 18 15 Measuring Variables Government-financed 3 3 Cost of project $0-3 million 39 37 As mentioned before, the questionnaire used to $3 million-10 million 33 32 code the project reports contained 145 items. $10+ million 33 31 Population reached 1,000-60,000 38 33 Based on the number of observations per item, o60,000-500,000 55 47 nature of scale (that is, whether the scale was con- 500,000+ 24 20 tinuous), intercoder reliability, coder confidence, Type of technology Protected springs 2 1 and analysis of principal components, the num- Dug/shallow wells 8 7 Tube wells ber of items included in the model was reduced with handpumps 26 21 to fifty. These variables are discussed fully, fol- Deep-dug wells 25 20 lowing a description of some background char- Gravitysystems 41 317 acteristics of the project, most of which are not lype of distribution Spring 2 2 included in the model. system Well-head handpump 65 53 Community standpipes 31 25 Private connections 24 20 Per capita country $120-340 52 46 GNP in 1989 $360-650 32 28 The 121 projects included in the study were $710-1,000 13 11 located in 49 different countries. Forty-eight per- $1,010-1,760 17 15 cent of the projects were carried out in Africa, 33 village income $0-199 16 1637 percent in Asia, and 20 percent in Latin America $400-599 14 12 (table 3.1). Most of the projects were completed $600+ 10 8 Research Methodology and Project Descriptions 17 involved; approximately 38 percent of these pro- Water-related outcomes covered six perfor- jects had more than 200 staff members. mance variables. Overall project effectiveness is the In keeping with the focus of the study on one key outcome variable used in the study; it was subsector, only projects whose primary objective rated on a scale of 0 through 10. Overall effec- was implementation of rural water systems were tiveness measured all project costs and benefits selected. However, since donors often promul- in the areas of construction, operations and gate an integrated approach to water systems, maintenance (O&M), health and sanitation edu- many projects included components for sanita- cation, extension and community development, tion construction and other primary health care institutional development, and income genera- activities (49 percent). Another 7 percent of pro- tion. Several maintenance measures were coded jects included income-generation activities. and one was selected for inclusion in the model: percentage of water systemsffunctioning and in good Technology type. The projects included a range condition. The measure addressed only down- of technologies, from spring captures to power- time after breakdown during normal times, pumped, piped-water systems. Overall, 36 per- which excluded natural calamities such as floods cent of the projects installed just one type of and drought. technology, 46 percent used two or three differ- Access to and use of new or improved water ent technology systems, and 18 percent installed systems can produce several benefits, namely, four or more different types of systems. Twenty time savings, other income-generating activities, percent of the projects included household con- and improved health. The economic benefits vari- nections; 25 percent, community standposts; and able encompassed all of these measures. Cov- 53 percent, handpumps. The remainder were erage achieved by improved water systems-that spring captures or rain tanks. Only in 10 percent is, percentage of the target population who used the of the cases were communities given any choice improved system-was included as another per- of technology, formance indicator in the analysis. Two other variables of a slightly different Income. Three items focused on income levels: nature, equality of access (whether everyone had per capita gross national product (GNP), per capita equal access to the water system) and environ- client income, and village income. Clearly, projects mental effects, were also included. Although were targeted to the poorest countries: approxi- drinking water projects are generally not mately 75 percent of the projects were from coun- tries with annual per capita GNP of less than $650. Table 3.2. Measures of performance outcomes Intercoder reliability on client and village income Pan pancecipaton was low, particularly for the type of client most Performance outcomes Water-related outcomes served. Hence, unfortunately, these two measures Overall project effectiveness had to be excluded from further analysis. Percentage of water systems functioning and in good condition Economic benefits Measuring the Variables Incluided in the Model Percentage of target population reached Equality of access The model groups variables in six categories: (1) CapEacitybnuimldlnglouftecotmes performance outcomes; (2) beneficiary participa- Community empowerment tion; (3) nonparticipation determinants; (4) proxi- Women's empowerment Capacity and skills related to water systems mate determinants of outcomes; (5) determinants Strengthened local organizations of participation; and (6) intermediate steps to par- Strengthened local leaders ticipation. Each category is discussed below. Beneficiary participation Overall beneficiary participation Overall women's participation Performance outcomes. Performance outcomes Participation in design were of two types, water-related outcomes and Women's participation in design Participation in construction capacity buildig (or empowerment) outcomes. Women's participation in construction (See table 3.2 for a summary of the outcome vari- Participation in operation and maintenance ables that were included in the model.) Women's participation in operation and maintenance able_th _ weeiclddintemoe. 18 The Contributtion of People's Participation: Evidencefrom 121 Rural Water Stpply Projects expected to have a significant environmental major issue, two other intermediate outputs were impact, they do have the potential to improve added to the model: maintenance after one year and reforestation or greenery or to adversely affect maintenance afterfive years. the environment by waterlogging the area imme- diately surrounding the project. Participation determinants. Overall beneficiary Local capacity is essential to achieving sus- participation is determined by characteristics of tainability, so five capacity building (or empow- the beneficiaries and the agency. (See table 3.4 for erment) outcomes at the individual level and at a listing of the determinants and elements of par- the group or organizational level were chosen for ticipation.) If both beneficiaries and agencies per- inclusion: overall beneficiary empowerment, women's ceive the net benefits of participation to be high, empo7verment, increased skills in water-related tasks, participation can occur. An overall measure of strengthened local organizations, and strengthened perceived net benefits was developed after local leaders. coders separately rated agency and beneficiary costs and benefits Beneficiary participation. Several different mea- Net benefits are a function of client and sures of participation were selected for the agency characteristics, which themselves are model. The lower end of the scale represented determinants of overall participation. Client information sharing; at the higher end were deci- characteristics included the demand of clients, or sionmaking and control. The measures included the commitment clients made before implemen- an indicator of overall participation of beneficia- tation of the project; the skills and knowledge of ries in all aspects and stages of the project. The clients; the quality of broad-based leadership; the other measures focused on gender and participa- dependence on strong leaders; the social organi- tion during a particular stage of the project. Table 3.3. Nonparticipation determinants of outcomes Nonparticipation determinants of ouitcomes. Ob- Direct determinants Total cost viously, beneficiary participation is not the only Project complexity determinant of project outcomes. A range of Adequacy of facilities and equipment exogenous variables affects outcomes directly, Difficulty in recruiting and retaining staff Availability of spare parts and technicians and another category of variables affects out- Extent objectives clearly specified comes either directly or indirectly through par- Per capita GNP ticipation. Table 3.3 lists these determinants of Direct and indirect determinants Technology outcomes. Type of technology (sophistication of technology) There were seven variables in the first set of Appropriateness of technology direct determinants, primarily focusing on gen- External agents Support of host government eral project characteristics and per capita GNP. Understanding between agencies The second set of direct and indirect variables Client characteristics were grouped into five subcategories: technology, Average number of users per system r gPresence of other water sources external agents, client characteristics, external cli- Village income mate, and management. External climate Conduciveness of political climate Conduciveness of economic climate Proximate determinants of outcomes. This set of Conduciveness of sociocultural climate variables included those factors through which, it Conduciveness of geological environment Management was hypothesized, beneficiary participation Overall quality of management worked to effect outcomes. In a sense, then, these Skills of staff variables are intermediate ones on the way to Proximate determinants determining final outcomes. Variables in this cat- Institutional outputs Quality of project design egory were classified as institutional or physical. Quality of project implementation Institutional variables covered the quiality of pro- Physical outputs ject design and quality of implementation. Physical Quality of construction Adequacy of operation and maintenance variables included the quiality of construction and Maintenance after one year operations and maintenance. Since O&M is such a Maintenance after five years Research Methodology and Project Descriptions 19 ' Table 3.4. Determinants of participation objectives and means, and the degree to which Determinants of participation the project was driven by physical targets. Client characteristics Commitment of clients before implementation Skills and knowledge Critical elements of the participation process. Quality of broad-based leadership Participation is an iterative process; determinants ltDependence on strong leaders of patcpto c hog nemdaeses Extent of organization of clients partcipation through steps, Extent to which organization is based on traditional which together lead toward high levels of overall Ustructure participation. These steps are important mile- Agency cacteristidg stones that arise as beneficiaries organize and as Extent to which participation is made a goal agencies put participatory plans of implementa- Implementation flexibility tion into practice. The more important of these Autonomy of project and agency Consensus on objectives milestones are the responsiveness of agencies to Degree to which the project is driven by physical targets client feedback; the extent to which clients listen e Net benefits of participation t bto field agents; the extent to which local groups Elements of participation begin to gain control and ownership of resources; Agency-user relations i Responsiveness of agency to clients the users' dissatisfaction with or exit from the Extent to which clients listen to agents system; and the investment in capital and recur- User voice and exit Extent to which control and ownership became local rent costs made by beneficiaries. Extent to which clients exit Dissatisfaction of clients Schematic Presentation of the Model User investment in costs Capital investment made by clients Percentage of recurrent costs paid by users Figure 3.1 is a schematic presentation of the model outlined above. The fundamental rela- zation of clients to undertake water-related tasks; tionship of interest is between participation and and the extent to which organization builds upon outcomes (which are either water-related or more local traditions and structures. Important agency general capacity building outcomes). The model characteristics are use of local knowledge, degree also posits nonparticipation factors that have an to which beneficiary participation was made a impact on outcomes. These are of two kinds: goal, implementation flexibility, autonomy of an direct nonparticipation factors which influence agency to manage its own affairs, consensus on outcomes independently, and direct and indirect Figure 3.1. Model of the relationship of participation to outcomes and other factors I Nonparticipation determinants Direct/ - Proximate determinants .indirect Direct of outcomes I I Outcomes Participation Determinants of participation in collective action Water related Capacity building iAl Critical elements of participation Note: r = correlation; it does not imply a causal relationship. 20 The Contribution of People's Participation: Evidencefrom 121 Rtural Water Suipply Projects nonparticipation factors which exert their influ- Other Notes on Research for the Model ence via the participation process. Finally, the model includes participation determinants, fac- Statistical analysis was conducted in phases. tors that lie behind and cause participation. Frequencies, cross-tabulations, correlations,3 fac- To account for some of the complexity and tor analyses, and a limited number of multivari- iterations of the processes at work, the model also ate regression analyses4 were performed after incorporates two intermediate relationships. The checking the quality of the data. The final round first concerns direct nonparticipation determi- of data analysis, after item reduction,5 consisted nants, working through a series of proximate primarily of multivariate regression analysis for determinants of outcomes. Similarly, the deter- model testing. minants of participation are assumed to work In addition, results of this study were com- through what is called "critical elements of par- pared to those from other studies following an ticipation" in figure 3.1. extensive review of the literature, and several The model and the variables associated with reviews and evaluations of projects received after each are elaborated in chapter 4 (which discusses completion of the coding process were consid- the part of the model relating to outcomes) and ered. Findings from the review of literature and chapter 5 (which addresses the part of the model these additional evaluation reports are high- relating to determinants of participation). lighted when relevant. CHAPTER 4 Role of Beneficiary Participation in Project Effectiveness P roject-specific evidence from around the directly related to water, such as community world suggests that participation of users empowerment. The fourth section examines the in decisionmaking produces more effec- sensitivity of the results to the fact that the data are tive, and more sustainable, projects. There is sur- subjective and therefore may be distorted by the prisingly little quantitative, statistical evidence, "halo effect" (that is, evaluators, because they however, that addresses the following questions: know the hypothesis being tested, may unwit- *Is beneficiary participation an important inde- tingly attribute participation to successful pro- pendent contributor to project outcomes? jects, and vice versa). * What is the path of influence of beneficiary Following these sections, which establish a participation on project outcomes? strong, robust statistical association between par- This chapter answers these questions by using ticipation and performance, are the two that pre- the data on project outcomes, participation, and sent the evidence that this relationship is causal. other factors described in chapter 3. (The perti- The fifth section examines the path of influence of nent part of the model under analysis is dia- participation throughout the stages of the project grammed in figure 4.1.) In short, the chapter to show how participation affects the proximate concludes that the evidence suggests a strong, determinants of project success. The next part causal relationship between greater participation goes beyond the statistical evidence to show, and improved project outcomes. using narratives of specific projects, how partici- The following section describes the measure- pation (or the lack thereof) directly influences ment of overall performance and of overall par- project outcomes. Finally, the concluding section ticipation; it also reports the basic results for the draws policy lessons from the evidence about J0 bivariate relationship between performance and participation and performance. participation. The next section shows that the per- formance-to-participation relationship is strong Participation and Project Effectiveness even after controlling statistically (through multi- variate regressions) for other determinants of pro- Overall project effectiveness (OPE) is a global . ject success. The third section shows that the measure of project performance, which was relationship between participation and perfor- coded from the project evaluation reports. mance holds across several different subjective Although there are many dimensions of project and objective indicators of project success; more- success, they are generally highly correlated with over, the relationship holds true for outcomes not one another, and this overall measure appears to 21 22 The Contribution of People's Participation: Evidencefrom 121 Ruiral Water Suipply Projects capture project effectiveness well.6 Measures of in participation, and no highly participatory other dimensions and aspects of project perfor- projects were low in effectiveness. Only 3 per- mance are discussed in later sections. cent (1 of 37) of projects with low participation Participation is defined here as the extent to were highly effective, in contrast to 31 percent which beneficiaries were involved in information (18 of 58) with medium participation and 81 per- sharing, consultation, decision making, and initi- cent (21 of 26) with high participation. Simply ating action in the program activities. Participa- put, effective projects also tended to be partici- tion was scored on a 1-7 point scale, with 1 being patory.9 little or no participation and 7 indicating high lev- A second test of the relationship of benefi- els of decisionmaking by the community. In addi- ciary participation to project performance is a tion to overall participation, participation at three bivariate linear regression (ordinary least stages of the project cycle design, construction, squares) of OPE and participation.'0 Table 4.1, and operations & maintenance (O&M)-was also column 1, reports the partial correlation coeffi- included; these analyses appear below, in the dis- cient of overall beneficiary participation (OBP) cussion of participation at various stages of pro- on OPE. The estimate of 0.62 implies that the ject implementation. Participation by women impact of increasing participation from a low was included as a separate variable. Overall ben- level-to a high level will improve project perfor- eficiary participation, from both men and mance from between 1.0 to 2.5 points (on a 7- women, correlated highly with the other mea- point scale). A 4-point increase in participation is sures and therefore will be used as the main mea- associated with, on average, a 1.2 point increase sure of participation.7 in project performance. The t-statistic tests Figure 4.2 reports the cross-tabulation bet- whether this estimate is consistent with the ween project effectiveness and beneficiary par- hypothesis that the association is zero. The value ticipation for the 121 projects.8 There was only of the t-statistic at 10.6 implies that the estimate one project that was highly effective and yet low (0.62) is statistically different from zero at essen- Figure 4.1. Model for testing the relationship of participation to performance outcomes Nonparticipation determinants Proximate determinants Quality of project design Direct/indirect Direct Quality of project implementation Appropriateness of technology Per capita GNP Effectiveness of construction Support of host government Total cost Effectiveness of O&M Interagency understanding Project complexity Maintenance after one year Conduciveness of political context Adequacy of facilities Maintenance after five years Economic context Difficulties in staff recruiting Social/cultural context Specified objectives and targets Environmental context Availability of spare parts Average number of users and technicians Competition _ from other water sources Skill of staff Outcomes Overall quality of management For water systems Overall project effectiveness: Percentage of water supply in good condition Economic value of benefits Percentage of target population Participation reached Overall beneficiary participation Environmental effects Overall women's participation > Equality of access Participation in design For capacity building Participation in construction Community empowerment Participation in O&M Water-related capacity and skills Strengthened local organization Strengthened local leaders Note: r = correlation; it does not imply a causal relationship. Role of Beneficiary Participation in Project Effectiveness 23 tially any significance level." Thus the statistic Participation after Controlling implies a significant relationship between bene- for Direct Determinants ficiary participation and overall project effec- tiveness. Based on the model charted in figure 3.1 and elab- orated in figure 4.1, a set of seven project charac- Participation and Other Performance teristics and exogenous factors, which were Determinants hypothesized to directly affect project outcomes, were added to the estimated equation for project Evidence of a bivariate association is just the first effectiveness. These were (1) per capita GNP, (2) step in establishing causality. Before drawing total cost, (3) project complexity, (4) adequacy of any inferences from the evidence of a strong facilities, (5) difficulties in recruiting and retaining bivariate association between participation and staff, (6) specified objectives and targets, and (7) project effectiveness, it must be ascertained availability of parts and repair technicians. Since whether the association is due to the mutual the chapter is focused on participation, the results association of performance and participation for these variables are discussed separately in box with some other variable (such as quality of 4.1 (see also appendixes lA and 1B). management) that may be responsible for per- The coefficient of overall beneficiary partici- formance and participation. There may be a ten- pation is reported in the first row, second column, dency for "all good things to go together" so of table 4.1. A bivariate correlation is likely to that, for instance, projects with strong manage- exaggerate the effect of any single "causal" vari- ment are both more effective and more partici- able with the seven direct determinants. Hence patory. The second step toward establishing the magnitude of the estimated coefficient falls causality therefore is to test for the robustness of roughly in half (from 0.62 to 0.28), as one would the participation variable in the presence of expect from adding controls. The drop suggests other important nonparticipatory determinants that much of the bivariate association of partici- of project outcomes. In order to do so, a series of pation to project effectiveness is attributable to multivariate regressions was performed, adding the mutual association with the direct determi- a wide variety of factors that affect project out- nants of project success.'3 This conclusion is not comes directly (that is, not through participa- surprising, since it is quite likely that one of the tion) as well as both directly and potentially determinants of participation is the beneficiaries' indirectly (that is, factors that directly affect per- expectation of project success. For example, one formance but which may also have an effect on of the direct determinants of project success (see participation).'2 box 4.1) is the availability of spare parts. If clients know that spare parts are not available, they are unlikely to devote much effort to organizing for Figure 4.2. Cross-tabulation of overall project effective- unlikely toevot sem hort to og zn f ness with beneficiary participation by number of projects participaton; the reverse holds, as well. Even though the partial correlation of partic- ipation to project effectiveness is lower after con- Overall project effectiveness 27 Table 4.1. Overall project effectiveness as a function (22%) of overall beneficiary participation Medium 15 34 5 54 Model Regression I Regression 2 Regression 3 Mediun 15 34 5 (450/) Overall project 0.62d 0.28d 0.24c High 1 18 21 40 effectiveness (10.6) (5.3) (3.7) (33%) n=121 n=77 n=68 Total 58 26 121 Note: Reported are the partial correlations and t-statistics (in parentheses) (31%) (48%) (21%) (100%) on OBP from three linearregressions, with OPE as the dependentvariable, including different sets of independent variables: (1) OBP alone Low Medium High Total (bivariate), (2) OBP plus seven direct determinants, and (3) OBP plus seven direct and eleven direct/indirect determiinants. Overall beneficiary participation Significance levels are indicated thus: a = significant at 0.05; b significant at 0.01; c = significant at 0.001; d = significant at 0.0001. Note: Percentages of the total number of projects are given in parentheses Numbers of projects for which the regressions were performed are in the Total row and column. indicated under the t-statistics. 24 The ContributioOt of people's Participation: Evidenicefrom 121 Rutral Water Suipply Projects trolling for direct determinants, the estimated formance may arise from increasing participa- impact is still large and strongly significant. The tion, or that their effect on performance may be estimated impact implies that an increase of 1.0 direct but affected nonetheless by participation. point in participation is associated, all else con- In either case the links are indirect. These vari- stant, with about a 0.3 point increase in effec- ables were added to the model and a third set of tiveness. The t-statistic for the coefficient, 5.3, is multivariate regressions was run. The five sub- still strongly significant. The conclusion, there- sets of variables included in the regression fore, is that beneficiary participation remains a model were the following: significant determinant of project effectiveness 1. Technology (appropriateness of technology) even after controlling for the effects of the seven 2. External agents (support of host government direct factors. and interagency understanding) 3. Client characteristics (average number of users Participation after Controllingfor Both Direct per system and competition from other and Direct anid Indirect Determinants sources) 4. Exogenoius climate (conduciveness of political Project performance is also determined by the climate, conduciveness of economic context, range of factors that exert their influence either conduciveness of the social and cultural con- directly or through their effect on participation text, and conduciveness of the geological and of clients. Factors that could have operated environmental context) either way were classified as "direct/indirect"; 5. Management (skill and motivation of staff and these consisted of eleven variables, organized quality of management). into five subsets. The major distinction of these The results for these variables are discussed in variables is that part of their effect on project per- box 4.1. Box 4.1. Impact of nonparticipation factors on outcomes for water projects, the environment, and equality outcomes Model 2: direct factors countries, so the range of per capita GNP was narrow and at the low end of the income scale. Similarly, complexity Availability of parts and the presence of repair techni- measured the number of different organizations and cians emerged as the most significant determinant of activities to be coordinated, but these projects generally nearly all performance outcome measures in model 2. registered only low or medium complexity; only eleven Clarity of objectives and targets emerged as a signif- projects (9 percent) were rated high in complexity. To some icant determinant of overall project effectiveness, while extent total cost can be taken as a proxy for project size. adequacy of facilities and equipment tumed out to be The findings seemed to indicate that total project size, as only somewhat important. Difficulty in recruiting and measured by total cost, neither has a bearing on project retaining staff was an important determinant of other outcomes nor affects the importance of participation. more narrowly focused project outcomes. As can be seen from appendix LA, the model had Model 3: direct/inidirectfactors lower explanatory power for the percentage of target population reached. Compared to the other variables, Availability of spare parts and presence of repair techni- this variable had the lowest intercoder reliability, .78. cians remained an important input determinant. Other Two other outcome measures were measured, which significant factors were appropriateness of technology were qualitatively different from the water system out- and quality of management. Participation, however, comes: environmental effects and equality of access. remained significant even in the presence of these factors. Beneficiary participation emerged as the only important Depending upon the specific output being considered, determinant of these outcomes in this model. The next adequacy of facilities, difficulties in retaining staff, and set of variables employed have greater explanatory clarity of objectives were important. (See appendix l.B.) power for these outcomes, however. It is equally important to note that per capita GNP The relationship between beneficiary participation and total cost of project, a proxy for project size, and pro- and project outcomes remained strong regardless of vari- ject complexity did not determine project performance. ations in per capita GNP, project complexity, and total cost. For measures of environmental effects and equality A couple of cautionary notes are in order. The projects of access to facilities, beneficiary participation was the included in the study were primarily from the poorest sole critical determinant in this model. Role of Beneficiary Participation in Project Effectiveness 25 Results for the model that includes all vari- good condition, the overall economic benefits, and ables are in the first row, third column of table 4.1. the percentage of the target population reached. It is Controlling for this additional set of inputs likely that these outcomes, because they mea- reduces the estimated partial impact of participa- sure more objectively verifiable quantitative tion only slightly, from 0.28 to 0.24. The effect is achievements, are less subject to coder and still strongly significant statistically. The multi- reporter bias than is the evaluation of overall variate results establish that, controlling for the project effectiveness. effects of eighteen direct and indirect input vari- The results of using these outcome variables ables, overall beneficiary participation remains a as the dependent variable are reported in table large and statistically significant determinant of 4.2. Bivariate results are in column 1 and the overall project effectiveness. results that include the set of direct and direct plus direct/indirect determinants are in columns Participation and Other Project Outcomes 2 and 3. Overall beneficiary participation was found to be strongly significant statistically for A strong association between overall project all three outcomes. Moreover, the estimated effectiveness and participation raises two ques- magnitude of the participation effect is roughly tions. First, to what extent is the association a the same as for overall project effectiveness.14 For result of the subjective nature of the assessment, instance, for percentage of water supply in good both of those writing the project reports on which condition, the estimated bivariate impact is 0.54, the data are based and of the coders who mapped dropping to 0.29 (t-statistic of 2.4) with the inclu- the project evaluations into numbers? Second, sion of all direct and indirect determinants of what are the effects of participation on other project performance, as expected since there aspects of project outcomes, including those not were no controls in the bivariate relationship. For directly measured by effectiveness of water sys- the overall economic benefits and the fraction of tem projects, such as increasing community the target population reached, the results are empowerment? similar, with bivariate (and multivariate) partial !' correlations of 0.53 (0.26) and 0.29 (0.25), respec- Participation and Overall Project Outcomes tively. This strong similarity of results-all esti- mates from column three are between 0.25 and Three additional measures of overall project out- 0.29-from both overall subjective and physical come were selected (based on data quality and indicators of project performance suggests that 4 availability) for testing the basic model (figure the halo effect of coders' subjective ascription of 4.1). These are the percentage of water systems in project effectiveness to participatory projects (or vice versa) does not appear to account for the Table 4.2. Water project performance outcomes positive correlation of participation with overall as a function of overall beneficiary participation project effectiveness. Model Regression I Regression 2 Regression 3 t oParticipation and Other Aspects L-Percentage of water 0.54d 0.30b 0.29a system in good (6.4) (3.1) (2.4) of Water System Outcomes I condition n=98 n=64 n=60 Overall economic 0.53d 0.27d 0.26c Another water-system performance outcome, t benefits (10.3) (4.1) (3.6) n=120 n=77 n=68 equality of access, was recorded.'5 Equality of Percentage of the 0.29d 0.17a 0.25b access measured the degree to which people had target population (5.3) (1.9) (2.5) access to improved facilities, or the degree to .1 actually reached n=118 n=76 n=68 which anyone was denied access for political rea- 4 Note: Reported are the partial correlations and t-statistics (in parentheses) sons or for nonpayment of fees. This measure was on OBP from three linear regressions, with OPE as the dependent variable, including different sets of independent variables: (1) OBP alone included because of its relevance to the important (bivariate), (2) OBP plus seven direct determinants, and (3) OBP plus goal of reaching the poor. Beneficiary participa- seven direct and eleven direct/indirect determinants. Significance levels are indicated thus: a = significant at 0.05; b = tion was a significant determinant of equality of significant at 0.01; c = significant at 0.001; d = significant at 0.0001. access (see row 2, column 1, of table 4.3). The Numbers of projects for which the regressions were performed are 4 indicated under the t-statistics. result is much less strong in the bivariate case, 26 The Contribtition of People's Participation: Evidencefrom 121 Rutral Water Sutpply Projects Table 4.3. Relationship of participation water projects have little dramatic impact on the to environmental benefits and equality of access environment. Nevertheless, the long-term envi- Model Regression I Regression 2 Regression 3 ronmental effects from waterlogging or from Environmental 021d 0.23b 0.23a localized reforestation efforts can be significant. benefits (3.9) (2.8) (2.3) Results from the study show that other determi- n=115 n=74 n=67 nants-the geological and hydrological context, Equality of access 0.23" 0.26" 0.17 and the overall political context, which encour- (3.9) (2.8) (2.3) n=115 n=74 n=67 ages environmental consciousness through legis- Note: Reported are the partial correlations and t-statistics (in parentheses) lation and education-are more significant for on OBP from three linear regressions, with OPE as the dependentvariable, environmental benefits, although beneficiary including different sets of independent variables: (1) OBP alone (bivariate), (2) OBP plus seven direct determinants, and (3) OBP plus participation remains important. The model has seven direct and eleven direct/indirect determinants. low explanatory power for this outcome. Again, Significance levels are indicated thus: a = significant at 0.05; b = o m significant at 0.01; c = significant at 0.001; d = significant at 0.0001. participation might or might not lead to greater Numbers of projects for which the regressions were performed are environmental benefits, depending on where the indicated under the t-statistics, environmental costs fall. One possibility is that local communities might be expected to be less however, and less strong overall than with other sensitive than other decisionmakers to nonlocal project outcomes.16 environmental costs. The relative weakness of the results for equal- ity of access is not surprising, because greater Participation and Capacity Butilding Otutcomes participation could engender either greater or less access. Collective action works to the extent Participation has been conceptualized in this that a group of people can formulate and imple- study as a means to achieving project effective- ment rules and regulations regarding entry and ness, efficiency, equity, and empowerment. The exit in using a common property resource. These limited number of project cases and the poor rules and regulations are critical to preventing quality of the data did not allow testing of the effi- free-rider problems. Participation in decision- ciency hypothesis; this section focuses on capac- making is a means to formulate rules, and to ity building and empowerment. These two achieve consensus on them and on their imple- outcomes are treated as variables, yet they may mentation. Hence high levels of participation well be more properly classified as proximate may mean that access to water systems is suc- determinants. cessfully limited to those who agree to be respon- Participation in decisionmaking is an impor- sible caretakers and owners. For example, in tant capacity building process. As people partic- Honduras those who did not participate in con- ipate in making new decisions and solving struction, but who later wanted to be part of a problems, learning takes place. This learning is piped-water system, had to pay higher connec- internalized, because it is accomplished experi- tion fees. Most community groups do not deny entially rather than by rote. It therefore leads to access to water because of inability to pay fees, changes in attitude, behavior, confidence, and however. The world over, community groups leadership. have developed different criteria, dictated by cul- At the individual level are three social actors tural and social norms as well as by economic of particular importance, namely, women, men, realities, to ensure access by the poor. In and leaders. The question of women's empower- Tanzania, for example, even in poor villages user ment and participation is important, but the issue fees are based on a sliding scale, with widows and is too complex to treat here; to highlight the find- others identified as destitute being exempt from ings and to do them justice, the subject will be user fees. In other countries, the poor are allowed covered fully in a separate paper. Findings on to pay with their labor rather than in cash. overall beneficiary participation and community Environmental benefits were also measured. empowerment follow. The adjusted R-square for environmental effects Empowerment is a result of participation in is low. Compared to large infrastructure such as decisionmaking. An empowered person is one dams or river basin management, rural drinking- who can take initiative, exert leadership, display Role of Beneficiary Participation in Project Effectiveness 27 confidence, solve new problems, mobilize controlling for all eighteen direct and indirect resources, and undertake new actions. Empower- input variables. ment, it is hypothesized, is an important outcome Beneficiary participation is even more posi- of high levels of participation involving control tively related to the other two outcomes not spe- over decisionmaking for a range of activities. cific to water projects: strengthened local Hence empowerment is a leading indicator of organizations and effect on local leaders. Clearly, successful capacity building at the individual and participatory decisionmaking does lead to institutional levels. strengthened local organizations, as evidenced in The second important indicator of capacity several ways: a greater sense of pride and iden- building or empowerment is specific to water tity with the village council or water committee and sanitation. Are the ability and skills of indi- or group; undertaking of new activities, from viduals enhanced in carrying out specific man- latrine building to brick and cloth making; agement and technical tasks related to water and greater political savvy and ability to negotiate sanitation? with (or pressure) political leaders and outside The third indicator is organizational. Decen- agencies to change rules and regulations, or to tralized programs require strong local organiza- mobilize additional resources. tions. When local organizations get the oppor- Evidence also establishes that the net effect of tunity to manage resources and support devel- participation on local leaders is positive, not neg- opment, they can become stronger (Esman and ative. An increasing number of projects work Uphoff 1984; Uphoff 1986). Participation in deci- through local leaders, win their support, and then sionmaking is hypothesized to strengthen the reach out to others in a community, including the capacity of local organizations to carry out activ- poor. Often people prefer that their leaders make ities. Local organizations can be a few people the decisions and negotiate with the outside working on water committees, or a village coun- world. The status of such leaders is enhanced cil, or larger, more formal organizations."7 when they successfully bring resources to the Finally, much has been written and said about community. Many leaders, once they are satisfied the effect of participatory development on local that they are not being bypassed, become either leaders. The general belief is that local leaders personally involved in stimulating collective resist participatory decisionmaking because it action or supportive of others, thus allowing the leads to empowerment, which changes the power emergence of new leaders. balance and jeopardizes their power base. The often accompanying belief is that local leaders should be bypassed because they either resist Table 4.4. Outcomes not specific to water systems, broad-based decisionmaking or capture a dispro- as a function of overall beneficiary participation portionate share of resources. On the other hand, Model Regression I Regression 2 Regression 3 when local leaders get involved they can be effec- Community 0.77d 0.59d 0,55d tive in mobilizing communities and accessing empoweent (15.6) (8.2) (6.7) resources. Although the process may enable com- 0.771 0.70 0.68 Water supply task 0.77d 0.70d 0.63d munity groups to become more powerful, it can capacity building (14.9) (8.3) (6.3) also allow leaders to emerge with more power n=121 n=77 n=68 and respect. For this reason a variable on the net Extent local 0.99d 1.01d 0.98d effect of participation on local leaders was organizations (17.1) (10.0) (8.1) strengthened n=109 n=71 n=63 included in the model. Net effect on 0.24d 0.26b 0.26b Table 4.4 reports the coefficients of participa- local leaders (4.6) (3.0) (2.4) tion for three models, with t-statistics in brack- n=106 n=70 n=64 ets. The results confirm the hypothesis that Note: Reported are the partial correlations and t-statistics (in parentheses) on OBP from three linear regressions, with OPE as the dependent variable, beneficiary participation is deeply embedded in including different sets of independent variables: (1) OBP alone all four capacity building or empowerment indi- (bivariate), (2) OBP plus seven direct determinants, and (3) OBP plus seven direct and eleven direct/indirect determinants. cators, since community empowerment and Significance levels are indicated thus: a = significant at 0.05; b = skills related to water-system tasks are deter-. significant at 0.01; c = significant at 0.001; d = significant at 0.0001. Numbers of projects for which the regressions were performed are mined by beneficiary participation even after indicated under the t-statistics. 28 The Contrbibtion of People's Participation: Evidence from 121 Rutral Water Suipply Projects How Important Is the "Halo Effect"? case that the degree of participation is overstated when project outcomes are good. The case review method uses systematic content The second technique for examining the halo analysis, which is based on subjective evaluation effect is to take advantage of the fact that two by coders of facts and findings reported in doc- coders coded both outcomes and inputs. If it were uments. It assumes that each coder will be able the case that a higher assessment of outcomes led to subjectively judge and quantify the diverse to a higher assessment of participation, then the phenomena. There are two ways in which this "within"-coder results should be stronger that coding could go wrong. Either the codings could the "across"-coder results. That is, if X's assess- be subject to a large degree of random error or ment of OBP is higher when X gives OPE a higher the coders could be influenced in their assess- rating, then the regression of X's OPE on OBP ments of the outcomes by their assessments of should have an upward bias relative to the participation. regression of OPE coded by X on OBP coded by The first problem, random measurement Y (whose assessment of OBP is not influenced by error, is not a major concern for two reasons. First, X's assessment of OPE).20 Table 4.5 shows the all of the variables used in the analysis show a results of exactly this test. The first row repro- high degree of intercoder reliability, in that the duces the results from table 4.1 of OPE on OBP, correlations of the same variable across projects both bivariate and multivariate (for the most for the two different coders are quite high.18 inclusive model). The second and third rows Second, a measurement error, if there were one, regress X's (or Y's) values for OPE on Y's (X's) would lead to an underestimation of the coeffi- values for the independent variables. There is no cients."9 In other words, if the results are strong significant drop in the estimated coefficients. The and statistically significant in the presence of results are the roughly the same, which is consis- measurement error, they would only be stronger tent with no halo effect.21 and more significant if that error were absent. Yet resolving the issue of coder halo effect Hence the possible objection that the measured does not negate the possibility that the writers of variables could be contaminated by a large the project evaluations themselves built the halo degree of measurement error is moot as a criti- effect into the reports that were subsequently cism of the present results. coded. Any halo effect operating on the writers The much more serious concern is that sub- of evaluation reports is likely to be weak, how- jective evaluations are susceptible to two types of ever, since participation was not a focus of those nonrandom measurement error. First, if coders reports and the evaluators did not know that rate both outcomes and inputs, their ratings of their reports would be used as data for a study project outputs may be influenced by their on participation. knowledge of system inputs. This is more likely to happen if the coders know the hypotheses Table 4.5. Tests for intercoder 'halo effects" in coding being tested. Second, if project inputs and project effectiveness and participation selected project outputs tend to be positive, that Bivariate is, good things in their own right, the subjective relationship Multiv'ariate evaluations of overall project effectiveness-by between relationship the original reporters as well as by the coders- Test OBP and OPE (all inputs) to OPE may be upwardly biased. This is known as the Mean values: 0.62d 0.24' halo effect. (X+Y)/2 (10.6) (3.7) Two types of analyses were conducted to Coder X outputs 0.622d 026' on coder Y inputs (10.3) (2.1) check for the pervasiveness of the halo effect. The Coder Y outputs 057d 0.26' first, already discussed elsewhere in this chapter, on coder X inputs (9.3) (2.7) is the use of objective measures of project out- Note: Reported are the partial correlations and t-statistics (in parentheses) come, such as the percentage of the water system onOBP from three linear regressions, with OPE as the dependentvariable, o including different sets of independent variables: (1) OBP alone in good condition. The reporting and coding of (bivariate), (2) OBP plus seven direct determinants, and (3) OBP plus the outputs of such objective measures should seven direct and eleven direct/indirect determinants. Significance levels are indicated thus: a = significant at 0.05; b = reduce the halo effect, although it still may be the significant at 0.01; c = significant at 0.001; d = significant at 0.0001. Role of Beneficiary Participation in Project Effectiveness 29 Proximate Determinants and Participation should lead to a more effective project (although better design quality, as an outcome itself, is not The foregoing sections established the existence of an independent cause of a better outcome). If this a strong, robust association between the level of is true, the correlation coefficients between the set beneficiary participation and project perfor- of proximate determinants and outcomes should mance, but even this evidence falls short of prov- be high and significant. Table 4.6 reports these ing that more participation causes better project bivariate correlations. Looking down column 1, performance. This section and the following one the correlations between overall project effective- move beyond statistical association, to make a ness and the quality of each of the project stages case implying causation and to give meaning to is very high. The same holds true for the other the statistics by highlighting specific project expe- measures of project success, although, not sur- riences. The present section does so by showing prisingly, the correlation is less strong for the that participation is related to success in each of environmental and equality outcomes. the stages of a project, revealing the mechanisms whereby participation causes success. The section Institutional Inpuits on project experiences uses case studies to illus- trate patterns and processes used in linking par- To explore the path through which beneficiary ticipation to improved project effectiveness.22 participation effects outcomes, the proximate The model diagrammed in figure 4.1 identifies determinants were divided into two broad cate- a series of proximate determinants of project out- gories: institutional outputs concern the design comes, which are the stages of a project (design, and implementation of a project, and physical out- implementation, construction, and mainte- puts concern the construction and maintenance of nance). It is hypothesized that beneficiary partic- water systems. Few would dispute that project ipation affects final outcomes by affecting outputs are influenced by overall quality of pro- outcomes at each of these stages. Multivariate ject design and by quality of project implementa- regressions were performed to test these hypoth- tion. Early involvement of project beneficiaries esized relationships between proximate determi- and other stakeholders not only improves project nants and outcomes. design but also begins the process of local own- Success in the stages of a project, or proximate ership of projects, which is critical for achieving determinants, leads to successful project out- sustainability. While the importance of getting comes. For example, better "quality of design" project design right has long been understood, Table 4.6. Bivariate correlations between outcomes and proximate determinants Outcome variable Overall Percentage of Percentage of project water system Objective value target popu- Environmental Equality Proximate determinant effectiveness in good condition of benefits lation reached effects of access Quality of design 0.81d 0.66d 0.70d 0.43d 0.34c 0.32c (n=121) (n=98) (n=120) (n=118) (n=115) (n=112) Quality of implementation 0.92d 0.75d 0.78d 0.50d 0.38d 0.39d (n=121) (n=98) (n=120) (n=118) (n=115) (n=112) Quality of construction 0.61d 0.59d 0.53 0.26b 0.29c (n=121) (n=98) (n=120) (n=118) (n=115) (n=112) Quality of O&M 0.9Qd 0.83d 0.77d 0.42d 0.30' 0.31c (n=121) (n=98) (n=120) (n=118) (n=115) (n=112) Quality of maintenance 0.85d 0.89d 0.74d 0.40d 0.35d 0.29b after one year (n=117) (n=98) (n=116) (n=114) (n=111) (n=108) Quality of maintenance 0.81d 0.87d 0.72d 0.47d 0.37c 0.33b after five years (n=74) (n=70) (n=73) (n=71) (n=70) (n=67) Note: Reported are the partial correlations and t-statistics (in parentheses) on OBP from three linear regressions, with OPE as the dependent variable, including different sets of independent variables: (1) OBP alone (bivariate), (2) OBP plus seven direct determinants, and (3) OBP plus seven direct and eleven direct/indirect determninants. Significance levels are indicated thus: a = significant at 0.05; b = significant at 0.01; c = significant at 0.001; d = significant at 0.0001. The number of projects under consideration are in parentheses. 30 The Contriblution of People's Participation: Evidencefrom 121 Rutral Water Suipply Projects only recently has attention shifted to ensuring Third, there are two ways of making projects quality of implementation to achieve high-qual- responsive to beneficiary demand: beneficiaries ity project outputs. The two proximate inputs can be involved indirectly or directly. This mea- examined here, then, are overall project design sure captures only direct involvement, although and overall quality of implementation. indirect involvement, as evaluated by other tech- niques, is considered in the study's overall find- Quality of design. Table 4.7 lists the coefficient ings. Indirect involvement includes market on participation in each cell, with t-statistics in surveys, beneficiary assessment, contingent valu- parentheses, for the bivariate and for two differ- ation methods to assess willingness to pay, socio- ent multivariate regressions. For quality of logical and anthropological studies, and other design, beneficiary participation is significant consultation techniques which produce informa- only in the bivariate model. Given the general tion on the needs, preferences, capacity, and social belief that early inputs from stakeholders in a and political organization of beneficiaries and project, including beneficiaries, are important in intermediary organizations. Indirect information creating a design that fits the needs of project can also be drawn from lessons learned from clients, the lack of a significant relationship across other projects attempting to provide services to the board is puzzling. the poor, and then fed into the design process. But three issues must be kept in mind when Indirect involvement of beneficiaries in interpreting these findings. First, statistically design was rated low in participation in the speaking, the lack of significance at the 5 percent study; the majority of projects did not involve level is not the only standard. Low precision can beneficiaries in design at all. Only six projects lead to statistical insignificance even when the actually involved beneficiaries directly in deci- qualitative relationship is substantial. Moreover, sionmaking for design, so the spread of projects the coefficient itself is not much lower for design in the highly participatory category was very low. (0.16) than for implementation (0.21), which was Direct beneficiary participation means beneficia- significant at the 0.01 level. Second, clients ries work together with project authorities in include people at the community level-the ulti- evolving the design of the project. This kind of mate project beneficiaries and stakeholders- involvement is rare, except through the involve- agency staff, and others who will be directly ment of NGOs and through pilot implementation affected by the project. The staff of public sector activities during the process of project formula- agencies, through which most large rural water tion; again, lessons from these experiences are fed supply projects have been implemented, were into the design of the project. included in the measure used in this study of ben- What the results suggest, then, is that the crit- eficiaries, but other project stakeholders were not. ical factors in macroproject design may be con- sultative processes and broad stakeholder Table 4.7. Impact of overall beneficiary participation involvement rather than direct and intensive on the institutional proximate determinants involvement of large numbers of beneficiaries. of project performance Besides participation, the factors of importance in Proximate the model were the quality of attention paid to determinant Regression I Regression 2 Regression 3 availability of spare parts and repair technicians Quality of design 0.46d 0.12 0.16 (significant beyond the 0.0001 level); the number (6.9) (1.3) (1.3) n=118 n=76 n=68 of users per water system, as specified by design Quality of 0.53d 0.17b 0.21b criteria (a negative relationship significant at the implementation (9.3) (2.7) (2.7) 0.06 level);23 and the overall complexity of the n=121 n=77 n=68 project (a negative relationship significant at the Note: Reported are the partial correlations and f-statistics (in parentheses) 0.08 level).24 on OBP from three linear regressions, with OPE as the dependent variable, including different sets of independent variables: (1) OBP alone (bivariate), (2) OBP plus seven direct determninants, and (3) OBP plus Qaiy ipeetto."rii seven direct and eleven direct/indirect determinantsa Q,iality of implementation. Beneficiary partici- Significance levels are indicated thus: a = significant at 0.05; b pation is significantly related to quality of imple- significant at 0.01; c = significant at 0.001; d = significant at 0.0001. Numbers of projects for which the regressions were performed are mentahon m the bivariate model and in both indicated under the t-statistics. multivariate equations even after controlling for Role of Beneficiary Participation in Project Effectiveness 31 the effects of other inputs, which means, of course, tive determinant (-0.30), as did number of users that it is a significant determinant of the quality of per system (-0.30). Effectiveness of construction implementation. In fact, OBP is the critical deter- was the most severely affected by the overall eco- minant of overall quality of implementation. The nomic context (0.59). The worse the economic con- implications of these findings are important for text, the worse the quality of construction. This project design and management processes. finding makes sense. When there is high inflation Two other input variables of significance were or dependence on highly taxed imports, or when (1) clarity of objectives and targets and (2) avail- economic conditions are poor and markets are not ability of spare parts and repair technicians. functioning effectively, the quality of physical When the design of individual subprojects is construction is negatively affected. evolving through the process of implementation Although communities actually led the con- and all the details of implementation vary from struction for some projects, historically the norm site to site, it is particularly important to be clear has been that communities provide unskilled on overriding objectives and targets. Strategies labor and participate in the manner dictated by can then be adapted to the local context but still the construction contractors. Generally, the con- achieve defined objectives and targets. tractor is accountable not to the communities but to external, governmental or nongovernmental Physical Outpuits and Participation agents. Contractors therefore have had little incentive to interact with people in local commu- Many factors affect construction and operation nities, whom they see as fonts of free housing, and maintenance, or intermediate physical out- construction assistance, and unpaid labor, not as puts. These two components are treated sepa- decisionmakers who will certify that construction rately because, in the past, the water sector has has been satisfactorily completed. Project experi- equated participation with free labor and materi- ence also establishes that even when construction als or with handing over O&M to communities is led by communities, timely technical assistance after construction had been completed. In both is helpful in assessing quality of construction. scenarios, although communities were asked to "do something," they were not part of the deci- Effectiveness of O&M. Three different measures sionmaking process. Neither approach treats par- of project operations and maintenance were ticipation as a process. coded from the project reports: effectiveness of Effectiveness of construction. The contribution of beneficiary participation to construction effec- Table 4.8. Impact of overall beneficiary participation tivbenesi gfic antypa onlya.in th biaruiat model on the physical outputs related to project performance tiveness iS significant only in the bivariate model (table 4.8). It is not significant in either of the mul- Output Regression I Regression 2 Regression 3 tivariate models. Thus, when controlling for the Effectiveness 0.30d 0.18 0.11 effects of other inputs, beneficiary participation of construction (n46)2 (n=87)7 (n96)8 no longer remains a significant determinant of Effectiveness 0.49d 0.14a 0.11 construction quality, although it is marginally of O&M (7.4) (2.0) (1.1) significant in the limited multivariate results (col- n=121 n=77 n=68 umn 2, p-level of 0.075); also, as discussed else- Maintenance 0.43d 0.16a 0.18 after one year (6.6) (2.0) (1.8) where, the full multivariate results may n=117 n=75 n=66 understate the true total effect. Maintenance 0.46d 0.09 0.25 In the multivariate model that included all after five years (4.9) (0.7) (1.5) direct and indirect inputs, three inputs did emerge n=74 n=49 n=45 as significant for the quality of construction. They Note: Reported are the partial correlations and t-statistics (in parentheses) on OBP from three linear regressions, with OPE as the dependent variable, were project complexity (significant at the 0.01 including different sets of independent variables: (1) OBP alone level), conduciveness of economic context (significant (bivanrate), (2) OBP plu dseve/indirect determinants, at the at 0.003 level), and average number of users per Significance levels are indicated thus: a = significant at 0.05; b = system (significant at the 0.02 level). Project com- significant at 0.01; c = significant at 0.001; d = significant at 0.0001. Numbers of projects for which the regressions were performed are plexity once again emerged as a significant nega- indicated under the t-statistics. 32 The Contribiutionz of People's Participation: Evidencefrom 121 Rutral Water Sutpply Projects O&M, nmainitenance after one year, and maintenance ing out the effect of participation in general. This afterfive years. Since the results for all three were examination is particularly interesting, given similar (as expected, because they are highly cor- that programs reluctant to change their way of related), they will be discussed together. Benefi- doing business typically tagged on beneficiary ciary participation is a significant bivariate participation at the end of the project (the correlate of O&M with all three variables. "handing over" syndrome) or limited participa- Controlling for the seven direct determinants tion to the construction phase and to contribu- of project success again lowers the estimated tions of free labor and local materials. The data impact. But for two of the three measures, par- suggest that neither approach has the desired ticipation remains a significant determinant of effect on any intermediate outcome, institu- success in O&M. As with the outcomes for water tional or physical. system conditions, effective O&M depends heav- To test for the lack of efficacy of piecemeal par- ily on the availability of spare parts and on the ticipation or of participation by command, a presence of repair technicians. series of multivariate regressions were conducted When all input variables are included in the matching participation during a particular stage equations for maintenance, however, the esti- of the cycle with the particular outcome, rather mates of the effect of participation drops to mod- than with overall participation. For example, estly insignificant levels. The estimates range table 4.9 reports multivariate regression testing from 0.11 (for effectiveness of O&M) to 0.25 (for for all models for the effect of participation in maintenance after five years). In the full multi- design (with qutality of design); participation in variate model, the availability of parts remains as construction (on qtiality of construtction); and par- the only significant determinant of maintenance ticipation in O&M (on quality of O&M). success. But here again arise the difficulties of The results establish that participation in one interpreting the results of a "kitchen-sink torture or another stage of a project does not by itself pos- test." Including eighteen variables both reduces itively affect proximate determinants of out- the available sample of projects and decreases the comes. The coefficient values for the bivariate degrees of freedom, either of which can be expect- and multivariate models are lower than when ed to lower the precision of the estimates of the tested for overall beneficiary participation across participation effect, independent of the effect of stages. The findings confirm that participation is the estimate. What is surprising, then, is that so a process whose benefits cannot be fully realized many of the results survived the torture of testing when limited to particular stages of implementa- with a large number of covariates, not that some tion. For maximum benefits, beneficiary partici- of them failed (in a statistical sense) to survive. pation needs to be viewed as a long-term process Examining the impact of participation at var- that necessitates involvement of users from the ious stages of the project is the final step in trac- beginning of a project to its end. Table 4.9. Proximate determinants of water project outcomes as a function of beneficiary participation in different stages Bivariate Limited multivariate Full multivariate Project stage Proximate determinant model I model 2 model 3 Beneficiary participation Quality of design 0.26d 0.12 0.14 in design (3.6) (1.7) (1.5) n=113 n=72 n=64 Beneficiary participation Effectiveness of 0.20c 0.13 0.13 in construction construction (3.3) (1.7) (1.6) n=117 n=76 n=67 Beneficiary participation Effectiveness of O&M 0.39d 0.08 0.05 in O&M (6.5) (1.4) (0.7) n=119 n=77 n=68 Note: Reported are the partial correlations and t-statistics (in parentheses) on OBP from three linear regressions, with OPE as the dependent variable, including different sets of independent variables: (1) OBP alone (bivariate), (2) OBP plus seven direct determinants, and (3) OBP plus seven direct and eleven direct/indirect determinants. Significance levels are indicated thus: a = significant at 0.05; b = significant at 0.01; c = significant at 0.001; d = significant at 0 0001. Numbers of projects for whuch the regressions were performed are indicated under the t-statistics. Role of Beneficiary Participation in Project Effectiveness 33 Project Experience bined with community organization activities with Beneficiary Participation based on community demand. Communities invest in the system and legally own the water The results insofar reported in the chapter estab- systems. lish statistically the positive relationships bet- Box 4.3 describes the Aguthi water supply ween project outcomes and participation. Yet the project in Kenya and the involvement of a paras- statistics, which are themselves derived from tatal organization, which enjoyed a greater project experiences, do not tell the whole story. To degree of fiscal autonomy than is typical of gov- untangle causal linkages, and to understand the ernment departments. The Aguthi community processes and problems in managing rural water initiated the project but turned over responsibil- supply programs, the project experiences have to ity for operation and maintenance to the paras- be examined in depth. tatal agency. This section looks at project experiences over Box 4.4 focuses on the Azad Jammu and the last two decades and highlights three key Kashmir (AJK) province of Pakistan. In the last findings: fifteen years community action, supported by the * Beneficiary participation leads to project local government, has resulted in construction of effectiveness, but the forms of participation 1,200 kilometers of roads and in 40 percent vary from context to context. greater access to safe water. Local leaders and the * To achieve sustainable operation and mainte- rich elite in AJK play leading roles in negotiating nance requires beneficiary participation, with government authorities, in planning sys- spare parts, and repair technicians. Training tems, and in setting tariffs. Almost every com- technicians and distributing spare parts are munity has a functioning system and keeps easier tasks than achieving participation in monthly financial records; many expand the decisionmaking, which requires simultane- water system on their own (S. Khan 1992; ously that beneficiaries be interested and Minnatullah 1993). involved in O&M and that agencies give up control. Sutstainable O&M: Participation, Spare Parts, * For maximum benefits, participation has to be and Agency Control viewed as a process that involves beneficiaries in decisionmaking from the beginning to the Beneficiary participation, availability of spare end of a project. A command-and-control parts and repair technicians, and appropriate- approach to participation does not work. ness of technology were significant determinants of overall project effectiveness. Beneficiary par- Forms of Participation Vary ticipation in decisionmaking was also important in achieving effective operation and mainte- The statistical findings establish that highly nance, but such participation implies that agen- effective projects also tended to be highly par- cies must let go of decisionmaking at the ticipatory. Some of the key characteristics of the microlevel. Agencies, especially government twenty most effective projects have been sum- engineering departments, have found this par- marized in appendix 2. Sixteen of the twenty ticularly difficult to accomplish, as the study most effective projects also scored the highest in results reflect. Of the twenty most effective pro- overall participation. The forms of participation jects in this study, only three were implemented during the various stages of the projects and the primarily by government engineering and water mechanisms used to foster participation varied works departments. widely in the different economic, cultural, and Of course, beneficiary participation cannot political contexts, however. Boxes 4.2, 4.3, and correct for inappropriate technologies or lack of 4.4 illustrate three very different projects, all of spare parts and repair technicians. Yet even if which were highly effective and participatory. spare parts and technicians are available, they Box 4.2 highlights a project in Paraguay, which will not be used, in most situations, unless the achieved a sustainable rural water supply community has an interest and stake in keeping through institutional and fiscal reform, com- systems in good order. 34 The Contrbibtion of People's Participation: Evidencefrom 121 Ruiral Water Suipply Projects These findings are the problems that have Nevertheless, in the absence of meaningful com- plagued the rural water supply sector. By the end munity participation by local people and local of the 1970s there was increasing recognition that responsibility and incentive for undertaking greater access to safe facilities was not the only repairs, the presence of appropriate technology issue; facilities had to be maintained to ensure and spare parts did not solve the O&M problems. their functioning after construction, as well. Three cases illustrate the difficulties with govern- Government, it was also clear, could not sustain ment-led maintenance systems. central maintenance systems. These insights led in the 1980s to the development of different types India: three-tier maintenance systems. The three- of decentralized maintenance systems, ranging tier maintenance system was developed in India from privatized to government-managed two- or in the 1970s in an attempt to create sustainable three-tier systems. O&M for village-level water systems. The central The success of the these variations depended government, which had been largely unsuccessful on creating responsibility for simple repairs at the in achieving this goal, saw decentralization as the community level, through "community participa- key to long-term sustainability of water systems. tion," and on making spare parts easily available. In the three-tier system, responsibility for As a decade of experience attests, technology sim- maintenance is decentralized and tasks are plifications and distribution of spare parts in a sec- assigned to one of three levels. The community is tor dominated by engineers was the easier of the the first tier, which is responsible for basic and two tasks, and much attention was devoted to it. preventive maintenance. The second tier, at the Box 4.2. Paraguay: institutional and fiscal reforms The $12.5 million, World Bank-funded rural water mittee (junta), which is duly recognized as a legal entity supply project in Paraguay had the objective to "promote by the Government of Paraguay. community commitment to, and participation in, the 2. Project description/agreement. The community and project" to help achieve long-term sustainability. To SENASA had to negotiate and sign a project agreement, achieve this goal, the project employed a number of which included a detailed description of each project strategies based on capacity building and establishing component and its quantities and costs. (The contract clearly defined, legally binding responsibilities between also lists all project plans and documents.) the executing agency and each community. 3. Users' contribuition. The junta had to agree to make The capacity building component focused primarily a cash downpayment of 5 percent as a condition for start- on the executing agency, the National Service of ing construction; provide cash, labor, equipment, mate- Environmental Sanitation (SENASA), a subdivision of the rials, land, or a combination thereof, equivalent to ten Ministry of Health. SENASA was a relatively new, untried percent of the project costs; and take out a loan from organization and needed extensive support in a number of SENASA, to be paid back at market interest rates in not areas. The project provided capacity building in such areas more than ten years. as finance, information systems, community organization, 4. Revenuie covenant. Each junta had to set tariffs for tariffs, and design and construction standards. Building water service at a level sufficient to obtain revenues to capacity did take time: the project took four years longer cover routine O&M, debt service on the SENASA loan to implement than had been originally estimated. By stick- to the community, and major repairs and replacements ing to a participatory approach, however, and taking the (at an amount to be determined by SENASA and the time to develop SENASA's institutional capabilities- junta). instead of bringing in an outside organization to fulfill In fact, the project exceeded expectations: communi- short-term construction goals-the project helped to cre- ties contributed 21 percent of total construction costs (6 ate a stronger overall institutional structure and increased percent over original estimates), and the project serves the likelihood of achieving long-term sustainability. almost 20,000 people more than originally estimated. The responsibilities component addressed the sub- Operation and maintenance is satisfactory, and the project negotiations between SENASA and each com- majority of systems provide adequate service. The juntas munity. Before SENASA contracted with a junta for the are well-motivated, manage systems satisfactorily, meet construction of the water system, the community had to most financial commitments, and have little trouble col- fulfill the following legally binding steps: lecting revenues. 1. Juintaformation. The community had to follow pro- ject guidelines for forming a Water and Sanitation Com- Souirce: World Bank (1987). Role of Beneficiary Participation in Project Effectiveness 35 subdistrict or ward level, is responsible for sys- and the second tier become dependent on the tems in several villages; a community calls in the third tier and renege on their own maintenance second tier when minor breakdowns occur. responsibilities, which reintroduces the problems Finally, the third tier, at the level of the district engendered by lack of community involve- govermoent, oversees the second tier and is ment-that is, sustainabiity of the water system responsible for major repairs. is jeopardized. These problems stem in part from The three-tier system looked simple and the design of the three-tier system, which focused straightforward on paper. Unfortunately, when more on structure than on the incentives that put into practice, three-tier maintenance turned would be necessary to get communities to under- out to be much more complex than designers and take repairs. Users have no meaningful role in the program implementers had envisioned. Attempts decisionmaking process and no technical, finan- to implement the system countrywide in several cial, or organizational control. In fact, in this sys- countries in Asia and Africa proved that many tem community inaction is rewarded: others districts do not have the institutional capacity to eventually do the work for the community, at effectively execute their responsibilities. An entire unaffordable cost to the agency. infrastructure (logistical, administrative, finan- In addition, the three-tier maintenance system cial, and technical) is necessary at the district level is part of a bureaucratic structure that leaves little to make the system work. In order to have any room for flexibility if a community wishes to con- success in creating this infrastructure, central gov- tribute or develop a maintenance system outside ernments have to hand over financial and admin- the tier structure. One example from a nationwide istrative control to the districts, which they have study in Zimbabwe proves the point: a commurnity been reluctant to do. had collected $179 toward the improvement of its The most ironic aspect of the three-tier system water supply, but the District Council returned the is that when districts are capable of executing money because, according to an official, the money their responsibilities effectively, communities "could not be used" (Cleaver 1990). The incentives Box 4.3. Kenya: community participation with a parastatal agency Effective community participation takes many different would be no improved water system. Public meetings forms. The Aguthi rural water supply project in Kenya, were held to explain the project fully to community mem- for example, sprung from community initiative; long- bers; villages were given four to six weeks to organize term operation and maintenance is handled by the their participation and discuss any concems or questions. National Water Conservation and Pipeline Corporation Conmunities contributed extensively to the project, (NWCPC), a parastatal agency. The piped-water system, some 93,000 person days, valued at approximately Ksh with metered household connections, serves 68,000 peo- 2-2.5 million in all. Phase II of the project, with the help ple and cost approximately $6.5 million, of this community participation, was completed on Phase I of the Aguthi project started with conventional schedule and within the budget. As agreed, the commu-se contracting methods and no community participation; it nity defined its role as paying monthly tariffs after con- was plagued with problems, including delays in construc- struction was completed. O&M was handled successfully tion, cost overruns, and disagreements over consumer by the NWCPC. payment methods. Eventually construction was halted, NWCPC had a salary structure about 40 percent and it appeared that the project would not move forward. higher than that of the government, with high salary At this point, the Aguthi Water Commdttee met with increases going to the lowest-level staff. In addition, the Danish International Development Agency NWCPC, unlike the Ministry of Water Development (DANIDA), the major funder, and the Ministry of Water (which had to funnel funds through the Treasury), can Development; the committee offered to supply the neces- use revenues directly to meet the costs of the project. sary labor if DANIDA would take charge of project imple- These incentives helped NWCPC develop motivated mentation. The project was redesigned, and the water staff who understood their jobs and performed well. conmittee, working with local leaders and project staff, NWCPC has been successful in collecting revenues mobilized the community. The comnuoittee facilitated com- through its monthly meter reading and billing; about 91 munity organization by explaining the essential role that percent of potential revenue was collected in 1990. community members played in the project-without their participation the project would not go forward and there Soulrce: DANIDA (1991). %A41 36 The Contributioni of People's Participation: Evidencefrom 121 Rural Water Suipply Projects Box 4.4. Pakistan: community-based rural water systems Community-based water supply schemes are common from contributions made by residents and relatives liv- in Azad Jammu and Kashmir (AJK) state, which has a ing overseas. The project was executed as a joint venture population of over two million people.4 These schemes of the community and the local government department. are identified and initiated by communities and devel- Total project cost was Rs 830,000.b The hardware compo- oped on a self-help basis with cost-sharing support from nent of the scheme consists of a turbine pump with a 30- the Local Government Rural Development Department horsepower electric motor and 15,400 running meters of (LGRDD). Guided by local leaders and elites, communi- pipe. ties make technology and service-level choices, and plan Nearly 250 households have water connections. The and design the systems with limited technical guidance security fee for connection is Rs 300, with a monthly fee of from LGRDD. The AJK experience demonstrates both Rs 35 per household. The total monthly contribution cost effectiveness and sustainability. It serves as a model amounts to Rs 8,500, which covers the electricity charges for large scale replication. A recent World Bank loan of and salaries of one operator and one valveman. The chair- $28 million will extend the community-based approach man of the water committee maintains an account register to an additional 1,000 villages. which shows the monthly contribution of each household Bangrila village in Mirpur district (Azad Jammu and and expenses incurred. The register can be examined by Kashmir) is an example of how a community-based, any member of the community upon request. piped-water system works. Bangrila has a population of The scheme has been working without any major nearly 5,000, dispersed along the slopes of a hilly terrain. breakdowns for the last ten years. It was also successfully In 1981, in response to a desperate need for potable expanded during this period. Initially the community water, the local community decided to develop its own built one water tank with a capacity of 10,000 gallons; water supply system. The villagers formed a water com- over time, the community gradually extended the mittee on their own and then approached the LGRDD scheme, which now has five water tanks. through the Union and District Council of the area. The community agreed to share 50 percent of the capital cost a. This project was not included in the 121 cases studied. of the project and the entire cost of operation and main- b. Rs 35 = $1.00 in 1994. tenance. The water committee raised the required funds Soutrce: M. Khan (1992); Minatullah (1993). for communities to take an active role in the sys- showed that communities were dissatisfied with tem are few in such an environment.25 government repair services, and wanted to own the pumps and be trained and equipped to repair Uganda: decentralizing maintenance. The strug- them. UNICEF replaced old pumps with new, stan- gles of the government of Uganda over the last dardized ones, and provided standardized train- thirty years illustrate the difficulties of achieving ing and tools. The central government, however, effective government-managed maintenance sys- would not turn over ownership of the pumps to the tems without fiscal and asset decentralization to communities. In the years after the introduction of the district level. Shortly after the country's inde- the new pumps, the central government's mainte- pendence in 1962, the government began imple- nance program continued to run into bottlenecks. menting its plan to provide safe water and other The government's response time for repairs often social services to the entire population. Water-sys- exceeded six months, and more than 50 percent of tem implementation and maintenance was highly pumps were not in service at any given time. centralized under the control of the government's These problems continued until 1986, when a Water Development Department (WDD). coalition government was formed. Given the new Throughout the next decade, however, escalating government's focus on socialism and self- internal instability wreaked havoc on the water reliance, UNICEF again recommended a commu- systems. About 80 percent of all pumps were nity-based maintenance system for handpumps. working in the 1960s; by the 1970s most were in This time the government responded positively disrepair. According to government statistics, by and project planning with communities began in 1980 approximately 95 percent of the rural popu- the Luwero district in April 1986. Box 4.5 high- lation had no access to safe water supplies. lights the process used in developing the decen- Results of a community survey conducted by tralized systems and the continuing problems in UNICEF and the government in the early 1980s applying the approach nationally. Role of Beneficiary Participation in Project Effectiveness 37 Cote d'Ivoire: niaintenance problems continuie. system of pump repair parts. Nationwide, The experiences of the World Bank Second Rural approximately 460 repairmen were identified Water Supply Project in C6te d'Ivoire point to the and trained, a ratio of 1 repairman to about 30 fact that, even when trained technicians are pumps. Spare parts were made available nation- available and spare parts are distributed, func- wide through an automobile parts distributor. tioning water systems will not be achieved with- Sensitization meetings were held in all villages, out community interest, commitment, and and 94 percent of the village groups agreed to cre- involvement in decisionmaking. A survey in ate water committees and accounts and to prepare 1988 disclosed that, after the government had pump sites before the government rehabilitated spent $115 million to construct 13,000 water the pump free of charge. When the program points, only 52 percent of the handpumps were began, however, 83 percent of the villages did not functioning. This finding resulted in the intro- prepare the pump site, 31 percent did not create a duction of a new program financed by the World village account, and 45 percent of the village com- Bank, called "Programme de Restructuration de mittees were not operational. Government techni- l'Hydraulique Villageoise." The objective of the cians rehabilitated most of the pumps nonetheless. program was to give full O&M responsibility to One year later, according to a survey con- villagers, supported by a privatized system of ducted in late 1992, over 30 percent of the pumps repair technicians and nationwide distribution were again out of order, and half of the village Box 4.5. Uganda: decentralizing maintenance The community-based system of handpump maintenance bility for the community, while still satisfying the sub- evolved through extensive discussions with communities county and district levels by standardizing the system as and administrators at each level in the Luwero district of much as possible. Uganda. Community "mobilizers," provided by the The strategy for the system included the following Luwero district administration, central government, and steps: each community selected a pump caretaker, who UNICEF, explained the project to communities during local was trained in preventive maintenance by the Water religious services and community meetings. Each commu- Development Department (WDD) staff and given neces- nity was then asked to appoint eleven members, plus its sary tools; every group of twenty communities selected nine Resistance Committee (RC) members, to represent it two pump mechanics for training in major repairs and at a series of meetings on the project. (The Resistance maintenance, and each of those mechanics were given Commnittee system, which exists from the village to the tools and a bicycle; each community decided how it national level, was created by the new government to would pay pump mechanics and maintenance people; decentralize social services.) Representatives met to dis- each subcounty RC was given a one-year supply of spare cuss the project with government and UNICEF officials. parts; and sales depots were established to sell spare parts Over 100 such meetings were held before the project began. provided by WDD at the district level. Through these discussions, project implementers The system was monitored closely by the govern- learned from the communities how best to develop a ment and participating communities and refined over the community-based maintenance system for the hand- first two years. Over the eighteen months of the project pumps. The meetings also enabled the communities to cycle and for two years afterward, 98 percent of all pumps participate fully in decisions about pump location and were working at any given time, and the time between installation, maintenance and management systems, and breakdown and repair did not exceed two days, with methods for raising funds to cover maintenance and most repairs made the same day of breakdown. management. Community representatives reported find- Difficulties remain in trying to apply this decentral- ings from the meetings to their communities and gath- ized system throughout the country. The system depends ered additional suggestions for future meetings. heavily on the district level, yet district-level capacity is Representatives were then chosen for a subcounty- very limited after special "project units" are removed. In level committee, from which five members were sent to a addition, although the central govemment has focused five-day workshop at the district level to finalize the sys- on giving districts the administrative, financial, and deci- tem. Representatives from UNICEF, the central govern- sionmaking responsibilities for water systems, it is still ment, NGOs, and the district administration also reluctant to provide districts with the assets and person- attended. After the system was finalized, guidelines were nel necessary to fulfill these responsibilities. written to help communities set up their maintenance systems. This process approach allowed maximum flexi- Souirce: CIDA/SIDA (1993). 38 The Contributtion of People's Participation: Evidencefrom 121 Rural Water Suipply Projects water committees were not functioning, 20 per- In the CARE project, water committees coor- cent of which had already disbanded entirely. dinated construction activities for the community, This unfortunate outcome occurred despite the which contributed a substantial amount of labor availability of well-trained and motivated repair to the project. Using local authority to organize technicians and of spare parts.26 Interestingly, construction also drew successfully on the local though, some villages achieved great success: highland tradition of communal work groups, or committees were functioning well in areas where mingas, used by the indigenous population to the community organization was already strong solve a variety of infrastructure problems. and there was a strong felt need for an improved The CRS project also relied on local water water supply; committees thrived also where committees (juntas) to organize community communities sold water and managed profits members for construction, as well as to resolve themselves (Hino 1993). The program's commit- logistical problems which arose during the con- ment to monitoring its own performance and to struction phase. CRS went even further, however, refining its approach to correct for problems is in giving control to the local community. The esti- now leading to a clearer demand orientation and mated capital cost of each subproject was allo- focus on beneficiary participation. cated to the community junta, which could then purchase materials and equipment and hire Establishinig a Participatory Process skilled labor, as necessary. Both of these projects achieved good opera- The study findings establish that beneficiary par- tion and maintenance of facilities. The CRS ticipation in decisionmaking is an iterative project had especially positive results, with com- process without clearly marked stages (except for munities initiating a number of new community analytical purposes). Hence it is important that projects. It is important to note, however, that beneficiaries be involved in decisionmaking at all beneficiary control of construction was only one stages of a project cycle, and not just carry out cer- part of a whole strategy of participation. tain mechanical tasks, such as gathering local The CARE/Rwanda Byumba Southeast materials, laying the pipes, and hosting out- Water Systems Project is another example of how siders. Beneficiary participation in decisionmak- beneficiary participation in construction, as part ing was the single most important contributor to of an ongoing process, can positively affect oper- overall quality of implementation. ation and maintenance. The CARE project staff Project experiences establish that, when given had extensive dialogues with the community opportunities, communities can take the lead in before construction began-which had not hap- design, redesign, construction, and maintenance. pened with previous projects in the area-and A few projects involved beneficiaries directly in community members themselves worked to con- subproject design. Direct involvement of benefi- struct the water system. After construction was ciaries in determining the overall project design completed, community members expressed a was rare; reliance on indirect participation strong sense of ownership and responsibility for through surveys and other consultation tech- the water systems. Users have since carefully niques was heavy, however. The projects in the maintained the premises and have protected the study employed different approaches to commu- installations from vandalism, in marked contrast nity roles in construction. The level of beneficiary to their treatment of previous water systems con- participation in construction was high for 41 per- structed by outsiders. cent of projects, medium for 29 percent, and low In a demand-based approach, which treats for 30 percent. participation as a process and community people Two projects in Ecuador, one funded by as the clients, the most important issue in con- USAID and implemented by Catholic Relief struction is not contribution of labor and materi- Services (CRS) and the other funded by CARE/ als, or even involvement in all the construction Canada and implemented by CARE and various decisionmaking; it is, instead, shifting account- governmental agencies, are examples of the pos- ability for construction from outside agents to itive effects of local control in the construction internal community groups. In this process the phase of a project. community decides how it will participate. It Role of Beneficiary Participation in Project Effectiveness 39 may decide to contract with an agent for all of the effective projects. Availability of spare parts and construction work, with the community paying repair technicians and use of the right technology for all the services or signing off on government- are particularly important. If promised inputs, provided assistance. Or it may decide something training, or supplies are not delivered, beneficiary else. The point is that community accountability participation cannot transcend their absence. yields positive results. Project designers also need to pay attention to the It is easy to subvert any process. Experience complexity of the system, including technical shows that ensuring accountability for quality complexity and number of users and different construction requires the community at large to activities and organizations to be coordinated. assume a watchdog function. If it does not, the The greater the complexity, the greater the likeli- benefits are captured very quickly by local elites, hood of failure. Given the high rate of unsuccess- who are in a natural alliance with the outsiders. ful rural water supply projects, project design The Orangi Pilot Project in Pakistan is the most should focus on delivering water first and sanita- dramatic documented case of the benefits of local tion later. accountability: system costs dropped to one- Direct, intensive involvement of large num- tenth of their earlier level, largely due to reduc- bers of beneficiaries is not important in influenc- tion in rents and leakages, when accountability ing overall project design, but consultative shifted from distant agents to the community. processes (defined as low levels of participation) While participation of beneficiaries in opera- and broad stakeholder involvement may be crit- tion and maintenance is important, project expe- ical to project design. Moreover, although it may riences show also that injecting participation for not be important for beneficiaries to be directly O&M without any earlier consultation does not involved in project design, it is important that work. The "handing over" syndrome character- project design be such that it enables beneficiary izes this staged approach to participation: typi- participation during implementation, once a pro- cally, a public works agency constructs a system ject has been formulated and financed. without conferring with community members; All projects are made up of subprojects, then, just before leaving, the agency informs the whether they are conceptualized that way or not. community that it is responsible for operation Involvement of beneficiaries in the implementa- and maintenance. Communities, unless they are tion of these location-specific water systems is in extremely dire straits, have no interest in mak- important at every stage, from planning the sys- ing the effort to take care of a system they did not tem through site selection, resource mobilization, ask for and do not own. Experience of project social organization, construction, and operation after project confirms this conclusion. and maintenance. One of the characteristics of participation is Policy Lessons that it cannot be turned on and off like a tap, that is, "now you participate, now you don't." Parti- Beneficiary participation is a key contributor to cipation is an evolutionary process that gathers achieving project effectiveness and maintaining momentum and defies breakdown into neat, self- water systems in good condition, as well as to contained categories, except for analytical pur- enhancing local management capacity and poses. Attempts to establish that participation is empowerment. Beneficiary participation in deci- particularly important in any particular stage, sionmaking by definition implies that a commu- especially in the later stages of operation and nity will share in or control major decisions at the maintenance, are not only misguided but have subproject level. Obviously, this cannot happen if contributed to ineffective projects. agencies continue to control all of those decisions; Participation should be viewed as a process beneficiary participation therefore presupposes that starts with planning and ends with opera- major institutional reform to shift control from tion and maintenance, rather than as an element agencies to communities. that can be injected in the later stages of a pro- Beneficiary participation by itself, however, in ject whenever outsiders determine. Effective the absence of support and linkages to the world participation takes many forms-there is no one oUtside the village community, will not result in ideal form. -7 .1 CHAPTER 5 Factors Affecting Beneficiary Participation T he preceding chapter demonstrated that process, such as agency responsiveness and ben- greater participation is strongly associated eficiary commitment. Another section looks at with enhanced project performance. Creat- those characteristics of projects, agencies, and ing greater levels of beneficiary participation in beneficiaries that are associated with greater projects is therefore an important goal, but par- benefits from participation (for both beneficia- ticipation cannot simply be mandated. It is the ries and agencies), and thus educe higher levels result of decisionmaking rules that affect the of participation. incentives and interactions among beneficiaries, Participation in decisionmaking evolves over within agencies, and between beneficiaries and time, with feedback loops; it is therefore a non- agencies. Two questions are particularly impor- linear process. For example, if the initial interac- tant to answer: tion between agencies and beneficiaries is * Which elements of the participatory process positive, beneficiaries are more likely to attend are critical to producing beneficiary partici- subsequent meetings. If a series of such interac- pation? tions are positive, then trust is established * What are the project, beneficiary, and agency among the groups. Data from purely ex postfacto characteristics that influence both these ele- evaluations cannot, of course, capture the full fla- ments and overall participation? vor and complexity of this evolutionary process, This chapter examines the statistical and case- so this chapter weaves together narrative case specific material on the factors that determine the material and statistics to present a more compre- level of participation. Three of those factors were hensive explanation than could be achieved with found to strongly influence the participation of the data alone. beneficiaries in projects: * Demand for the services the project delivers Degree and Elements of Participation * Organization of beneficiaries * Autonomy and client-orientation of the The rhetoric far outstrips the practice of partici- implementing agency. pation, as analysis of 121 projects shows. Part of this chapter is a review of the evi- Seventy-nine percent of the projects were rated dence on the degree of participation elicited in low or medium in overall beneficiary participa- the projects; this section examines the relation- tion; only 21 percent received high ratings. The ship between overall beneficiary participation situation was worse for women's participation: and specific elements of the participation 83 percent of all projects had low or medium 41 42 The Contributioni of People's Participation: Evidencefrom 121 Rutral Water Suipply Projects ratings in this category. Even nongovernmental The relationship (correlation) between inter- organizations, often assumed to be participatory mediate elements of participation and overall by definition, are not necessarily so. A recent in- participation is presented in table 5.1. All of the depth study of thirty NGOs in Latin America variables are strongly associated with participa- found that a majority of them scored low or tion, but the three factors that stand out are the medium on participation (Carroll 1992). Given investment cost to utsers, the extent control becomes the apparent difficulties in achieving high levels local, and the responsiveness of the agency. of participation-and given participation's criti- cal importance in achieving project effective- User Investment in the System ness-it is important to identify the determinants of participation. Willingness-to-pay studies are an important The difficulty in modeling participation (fig- means of assessing demand, and their results are ure 5.1) is that several events are happening used in designing projects. What people in indi- simultaneously; causality is not easily disentan- vidual communities actually pay for up-front gled. For example, achieving local control and capital costs or operation and maintenance can be ownership is not an all-or-nothing phenomenon. judged only when implementation begins, how- Because it changes over time and interacts with ever. The commitment to "put your money where participation, it could be considered either a sub- your mouth is" on the part of users marks an set of participation or an intermediate outcome, important intermediate step to achieving overall which affects participation. Nonetheless, those participation. who want to induce participation need to know Two indicators of user costs were measured, which key intermediate steps (or elements) indi- the percentage of capital costs and the percentage cate a healthy process of participation, without of recurrent costs paid by users for water systems. implying causality. The correlation between user-paid capital costs Figure 5.1. Model of relationships between beneficiary participation and its determinants Determinants of participation Client characteristics Agency/project design characteristics Commitment of clients before Consensus on objectives t to implementation Implementation flexibility N Extent of organization Autonomy of projects l... . for water system role Extent participation is made a goal Extent organization is based Extent local knowledge is used on traditional collective Extent project is Skills and knowledge of clients physical-target driven Quality of leadership among clients Dependence on charismatic leaders Participation Overall beneficiary participation Overall women's participation Critical elements of participation User voice and exit Extent control becomes local Extent to which clients exit Client dissatisfaction User investment Investment costs paid by users Recurrent costs paid by users User-agency relation Extent agency is listened to by clients Extent agency is responsive to clients Note: r = correlation, it does not imply a causal relationship. Factors Affecting Beneficiary Participation 43 and overall participation was significant (r = 0.75). * To what extent was the local O&M organiza- The percentage of recurrent costs was less tion autonomous? strongly, but still significantly, related to overall * What degree of community/agency sharing participation (r = 0.58). The aphorism "you get existed in management and decisionmaking? what you pay for" has its own corollary, since pay- The three measures correlated highly with ing apparently empowers users to go out and get one another, at about 0.88; therefore only one what they bargained for. item-degree of local control-was included in the model. Extent of Local Control The degree of local control and ownership was significantly correlated to overall participa- When outsiders induce collective action, partici- tion (r = 0.79). Generally, then, it would appear pation can be facilitated to the extent that control difficult to achieve high levels of participation and ownership of rules, regulations, and property without local control over decisionmaking. The becomes local. An inhibiting factor is the degree to experience of the Rwanda Second Water Supply which collective action is controlled and com- Project demonstrates the point (see box 5.1). manded by the external agency (Bromley and The other measurable variable related to the Cernea 1989; McCommon, Warner, and Yohalem extent of local control is the extent to which 1990; and Donnelly-Roark, 1987). People are more clients simply choose to exit the project rather likely to undertake collective action when it than to participate. And exit they do, when they directly increases their access to natural resources. are dissatisfied with the water system, and hence The study asked three questions to address they cannot participate in decisionmaking. The the issue of control and autonomy of the water- study showed a significant negative correlation system organization to ensure that slight seman- between overall participation and dissatisfaction tic differences did not produce false results in the with and exit from the system. quantification process: * To what extent did control and ownership Agency and User Relations become local, that is, to what degree did deci- sionmaking and financial responsibility The interface between the agency and users is of reside in the community rather than in the critical importance when collective action is external agency? induced by outsiders. It determines whether there is a fit between user demand and agency supply, which in turn helps determine project Table 5.1. Correlations of elements of participation supply,vwhich R int thelpsd te nce pr jec with overall beneficiary participation 'effceness. Realizin hei nce e o t interface, many centralized agencies employ a Intetmediate elemnent Correlation variety of intermediaries, which are better suited Investment cost to users 075d than are they for working directly at the commu- (n=113) Recurrent costs to users 0158d nity level. (n=103) But the interface cannot be one-sided. Agencies Extent control becomes local 0,79d have to be responsive to clients, and clients have Client dissatisfaction -0.50d to want to listen to agencies. Two factors were (n=lll) included in the model to capture the interdepen- Client exit from system dent nature of the relationship: agency responsive- Agency responsiveness 0.80d ness to clients and client responsiveness to agencyfield (n=113) staff. The correlation between overall participation Extent field agents listened to clients 0.72d and agency responsiveness was highly significant (r = 0.80), the correlation between participation Note: Reported are the partial correlations and t-statistics (in parentheses) onOBP from threelinearregressions, with OPE as the dependentvariable, and the extent to which users listen to field agents including different sets of independent variables: (1) OBP alone was also significant (r = 0.72). (bivariate), (2) OBP plus seven direct determinants, and (3) OBP plus T seven direct and eleven direct/indirect determinants. The three characteristics just discussed Significance levels are indicated thus: a significant at 0.05; b describe important elements of participatory significant at 0.01; c = significant at 0.001; d = significant at 0.0001. The number of projects analyzed are in parentheses. projects. Beneficiary participation is nearly 44 The Contribution of People's Participation: Evidencefrom 121 Rutral Water Sutpply Projects Box 5.1. Rwanda: government control over community decisionmaking, or a losing proposition? The experience of the Rwanda Second Water Supply Local management was integral to the government's Project highlights the importance of local control in deci- strategy, yet there was no local control in the design or sionmaking and the difficulties in achieving it. A recent decisionmaking process. The guidelines and plan of study of the project (completed in March 1993) showed action for project implementation, including dates of vil- that only 44 of the 144 communities covered by the pro- lage visits, were developed far in advance, without com- ject had functioning Community Water Associations munity input. Committees were established despite the (CWAs), the organizations responsible for managing and fact that the committee structure had been tried and had maintaining water systems. failed in earlier projects. The CWAs took a lot of work The project was implemented by central government and time to establish, yet representatives received no pay ministry staff, who carried out all the activities associ- and, in some cases, no supervision, support, or training. ated with community-based systems. They conducted Incentives for local government officials and com- baseline studies and surveys, developed educational mittee members to cooperate were low. It was often dif- materials, trained communal extension agents and other ficult to convince the niayor and communal authorities assistants, conducted meetings and elections, created to adopt national policies and undertake the plan of Community Water Associations (CWAs), implemented action for the establishment of user groups and associa- user education and construction, and monitored sys- tions. Communal authorities did not accept the principle tems. They did not, however, give up decisionmaking that users should pay, or that they should be required to control. pay in advance. Local government officials did not want All major decisions concerning project components outsiders to interfere in communal mobilization; nor did were made by the World Bank, central government min- they respect elected committee members. istries, and other donors. Communities could not In addition, the project structure was complex, with choose whether they would receive improved services; seven different funders often disagreeing over project what service level they would receive; or how they components. Reaching consensus on an approach to might maintain their water system. Management deci- community management took more time, under difficult sions were made by the communities after construction, communication conditions. but within the framework established by the govern- ment. Souirce: World Bank (1993). always associated with users paying part of the addition to time, costs include not only time but investment costs, with control of the project outlays in cash and perhaps even social costs, becoming local, and with an agency that is such as altered conflict and changes in leader- responsive to client needs. But although these ship However, as experience proves and statis- characteristics (user investment, local control, tics show, participation occurs because it brings and client-responsive agencies) describe ele- benefits to clients. ments of the participation process, they do not Since participation is the result of decisions by cause participation. The next section goes one individuals (members of communities, local step further. It addresses those characteristics of leaders, agency employees, and government offi- projects, agencies, and beneficiaries that are cials, for example), it is more likely to occur when associated with these participatory elements the net benefits to the involved parties are per- and with particular participation outcomes. In ceived to be high. As a matter of arithmetic, the other words, it looks for the determinants of net benefits to potential beneficiaries of water participation. projects are high when either the benefits are high or the costs are low. For agencies the perceived Deterninants of Participation net benefits of participation depend on the inter- nal structure and incentives of the agency, and Participation is not without costs. It requires the especially on how the agency views its objectives. time and skills of beneficiaries, agency represen- The net benefits of participation to agencies tatives, and project employees. From the agency and clients were assessed for each project when point of view, participation increases uncertainty the evaluation reports were coded. Net benefits in project design, and it often delays physical were strongly associated (0.78 correlation) with implementation. From the client viewpoint, in the observed level of participation. Indeed, it Factors Affectinig Beneficiary Participation 45 would be unsurprising to find a perfect correla- only to the extent that clients can and do mobilize tion if net benefits could be perfectly observed; to take control. but for ex post evaluations to identify not only The third set of factors that determine partici- the degree of participation but also the net ben- pation is the client-orientation of the agency, includ- efits suggests that those benefits are associated ing whether the agency has sufficient autonomy to with observable characteristics of any given incorporate participation as well as whether project, including some within the influence of agency goals are determined by client satisfaction policy. or by some other, physical indicator, such as Net benefits of participation should be a func- production targets. The responsiveness of the tion of all the user and agency determinants of agency, a crucial element of participation, depends participation; to test that conclusion, a multivari- heavily on the extent to which the agency has ate regression of net benefits was conducted on incentives to meet client needs. the determinants of participation. Strikingly, only The following sections explore the intercon- two factors, one relating to beneficiaries and one nections among demand, capacity, and client-ori- to agencies, emerged as significant determinants entation by first presenting the evidence from of overall net benefits of participation. These fac- cases and case studies and then by examining the tors were demand or prior commitment of clients and empirical links uncovered in the sample of 121 the extent to which beneficiary participation was an projects between overall participation and its sev- agency goal that was monitored and evaluated and for eral aspects, its net benefits, and the variables which agency staff were rewarded. hypothesized to affect it. Beneficiary participation can be elicited in a wide variety of political, economic, and social Demand contexts, then, so long as incentives exist at both the agency and the community levels, and agen- Felt need is an important determinant of com- cies send strong signals to staff about the impor- munity action, which has long been recognized, tance of generating and supporting beneficiary yet agencies have not systematically accepted felt participation in decisionmaking. need or demand as a key criterion for selecting The findings on net benefits and important communities for water projects. Projects varied elements of the participatory process suggest that widely in their methods of assessing demand, the three classes of factors will be important influ- degree to which they used demand as the key encers of the level of participation. The strength of selection criterion, and the degree to which they demand for the project (that is, the perceived ben- linked demand to service levels. The study found efits by potential clients) will determine the ben- three issues concerning demand to be important: efits of participation. Intensity of demand will be (1) the method of assessing demand, (2) the determined by such factors as alternative sources extent to which projects maintain a commitment of water and the cost and type of proposed water to using demand as the primary selection crite- improvements. The close relationship between rion, and (3) the differentiation of demand by ser- demand and participation is evidenced by the vice levels. strong association between participation and user payments, which at least to some degree Assessing demand. Willingness to pay for ser- must reflect underlying demand. vices is a strong indicator of demand, which has The second set of factors that determine par- been shown to be critical in creating sustainable ticipation concern the beneficiary capacityfor orga- water systems (Briscoe and de Ferranti 1988). nization, which influences the costs to a Many agencies committed to serving the poor community of organizing for effective participa- have been reluctant to adopt "willingness to pay" tion. The existing strength of organization and as the sole criterion for establishing water pro- leadership, the skill and knowledge of clients, the jects; they have instead used poverty to screen degree of consensus on objectives, and other local communities for eligibility. Yet evidence from factors determine whether, for a given level of around the world has shown that the poor are potential benefits from participation, the com- already paying for water, sometimes more than munity will mobilize. Control can become local the rich (Whittington, Lauria, and Xinming 1989), 46 The Contribiution of People's Participation: Evidencefrom 121 Rutral Water Supply Projects and that the poor are willing to pay if reliability Indonesian village, large and small, were not of services is assured (Altaf and others 1992). public property at all; instead, they had well- Demand changes over time, so demand known "ancestral owners" and therefore well- assessment itself cannot be static. Projects established rules regarding access and use gauged demand during implementation by get- (Narayan-Parker 1986). ting communities to demonstrate their interest and commitment before project construction Demand focus in community selection. Despite began. For example, participants displayed com- espousing participation and demand as impor- mitment when they contributed local materials, tant criteria, staff in most projects did not cease signed agreements, put up a certain amount of efforts in those instances when communities cash, held village meetings, and reached group were clearly indifferent to what the project had to consensus on whether and under what condi- offer. In general, government agencies were more tions they would participate in a project and on reluctant to walk away than NGOs. how to proceed. The proclivity to proceed even in the absence Most of the projects, however, simply of community interest is well illustrated in a assumed that demand was high in water-scarce World Bank project in the Mwanza region of areas, as judged by objective criteria such as long Tanzania, which had as a prime objective "full distances to sources, nonavailability of safe participation of villages in all project activities." water, or high levels of morbidity, all of which Once the project began, demand for the wells came from statistics in secondary sources. Such turned out to be low; no requests had been calculations allow some assumptions, of course, received from villagers twelve months into the but experience shows that when all project plan- life of the project. Nevertheless, the detailed ning is based on such rough, nonuser measures, yearly targets for well construction drove the pro- failure rates are high. ject; construction crews eventually did every- In fact, in many of the study projects that thing themselves so that implementation of the scored low in effectiveness, agencies had project would not be "handicapped" (Therkild- defined criteria of need based on poverty, health sen 1988). Echoes of this story are heard all over statistics, and water scarcity in agency-selected the world. A recently completed review of a communities. For a project in Indonesia, for Nigerian government project revealed similar example, government agency staff had done happenings (Boerma 1993). painstaking survey work in villages to establish An important innovation-a system of appli- the severity of water problems, based on objec- cation forms-is now being used to sidestep polit- tive criteria such as population size and distance ical pressure and to help ensure a demand to and quality of water. Four villages judged to orientation. Procedures require communities to be the "most needy"-those with the most request and fill out application forms and to meet severe water problems-were selected to be certain conditions before the external agency will served first. In order to develop project strate- commit itself to a partnership with the community. gies within the selected villages, the executing This system has been tried effectively in several agency, an NGO, began to collect data to assess projects in Indonesia, Paraguay, and Swaziland village needs, including the perceived need for (see box 5.2), and it is being employed increasingly water improvements. Less than 20 percent of the for World Bank projects (see chapter 6). populations reported water to be a high-priority problem. Furthermore, one village with, report- Iniportance of matching demand witlh desired service edly, three water sources turned out to have sev- levels. Even giving more than lip service to assess- enteen, most of which were small springs not ing demand for water is not enough, however. It considered important enough for large-scale has also proven critical to provide the level of ser- development. Since the motivation for change is vice people demand. Several examples, especially not bacteriologically clean water so much as the striking evaluation of USAID assistance in convenience and water quantity, the incentive to Thailand, dramatically establish the importance of participate in collective action in such a situation providing appropriate service levels (Dworkin and is weak. Indeed, all water sources in the Pillsbury 1980). The Thailand experience revealed Factors Affecting Beneficiary Participation 47 that over twenty years, as the service moved from mentation from control and project villages. handpumps to motor pumps to house connections, Community field workers were trained and placed communities that had been apathetic, uninter- in the villages; the Girl Guides project team visited ested, and unwilling to foot the bill for low service the villages, as needed, to provide organizational levels became willing to pay higher amounts for and technical assistance. The review conducted house connections. High levels of service reliabil- after the project had been completed found that the ity also were finally achieved.27 low level of service (improvement of traditional If the service level does not match what people wells) did not generate much enthusiasm and that, want, not even the best-trained agencies can induce despite intense facilitation and training efforts, participation. For instance, the Girl Guides women's involvement in decisionmaking rem- Association of Thailand, with assistance from the ained low (Tunyavitch and others 1987). World Health Organization and the United Nations Development Programme, undertook a Statistical evidence on demand and participation. water and sanitation project aimed at supporting The strong links between the share of investment participatory decisionmaking, especially that of and recurrent costs borne by the users and the women. Anthropologists from Mahidol University degree of participation have already been collected data both before and after project imple- established. This in itself implies that client Box 5.2 Indonesia: evolution of a demand-led strategy CARE's strategy of implementing water supply and san- and through surveys CARE assesses their willingness itation projects in four provinces in Indonesia has and ability to pay. Meetings are held with selected com- changed dramatically since the start of work in the coun- munities to explain and discuss the conditions of the try in the mid-1970s. Initially CARE controlled and man- project. aged all stages of the projects. Over time, however, it 2. Committee formation and negotiation: The commu- became apparent that unless communities took control nity selects a water committee at a meeting. The newly of, and responsibility for, financing and managing water selected committee then negotiates its responsibilities systems, sustainability would not be achieved. with the government and CARE. Subsequent projects, including the current Community 3. Planning: The water committee chooses a technol- Self-Financing of Water and Sanitation project, focused ogy from those that CARE suggests as appropriate. It on community demand as the key selection criterion, designs and prices systems and, with CARE assistance, with control shifted to communities. develops resource mobilization and construction plans. An important indicator of the success of the new These designs and plans are then presented to the com- approach is the shift in the source of cash contributions munity and a formal agreement between the community, over an eleven-year period, 1979-90. In 1979 the com- government, and CARE is signed. bined contributions of CARE and the Indonesian gov- 4. Implementation: CARE trains the water committee ernment constituted approximately 80 percent of project in bookkeeping and control systems and construction of costs. By 1990 combined CARE and government contri- facilities. Once control systems are in place, the commit- butions had dropped to about 30 percent; community tee begins to mobilize cash, human, and material contributions had risen to cover over 70 percent of the resources. Costs are shared and cash contributions by costs. Communities had provided all cash contributions members are graduated, according to ability to pay. for physical construction for more than three-fourths of Credit from pipe suppliers and local banks is a common the projects. Most communities managed to successfully method of mobilizing "outside" cash. Monitoring con- operate and maintain their systems (some for as long as tinues throughout this stage. the period covered by the project review, or up to ten a. Operation and maintenance: The community selects years). In addition, many CARE-assisted communities an O&M committee, which draws up regulations and by- have helped neighboring communities to develop their laws and presents them to the community. The O&M own systems. committee develops a budget, bookkeeping system, and A six-stage implementation strategy, summarized training plan. CARE provides follow-up training. below, is responsible for the focus on demand and capac- 6. Evalutation and monitoring: CARE continues to ity building. assist the O&M committee for approximately one year. 1. Community selection: Government and CARE jointly select potential districts, market the project, and Souirce: Boase and others (1989); Jackson (1988); contact local leaders. Communities apply for a project, McGowan, Soewandi, and Aubel (1991). 48 T1te Contribution of People's Participation: Evidencefrom 121 Ruiral Water Suipply Projects demand, as evidenced by a willingness to share Despite the erosion of common property man- costs, is a determinant of beneficiary participa- agement regimes, rural societies generally offer tion. The strength of the prior commitment of more stability, social cohesion, and personal con- beneficiaries to the project also was explored tact than do urban societies. Rural societies, how- through project data. Table 5.2 shows the links ever, are not necessarily homogeneous. Social between demand, participation, user payments, divisions may be deep, although groups and and a variety of factors influencing participation. social networks typically are marked by reci- The first column of table 5.2 shows the rela- procity, trust, and exchange of goods. Tapping tionship between prior commitment and overall into this social organization is critical for success- beneficiary participation (row 1), the net benefits ful collective action. of participation (row 2), the fraction of user Case analyses revealed three beneficiary char- investment (row 3), and the recurrent costs borne acteristics that are especially important: existing by users (row 4). The prior commitment of clients social organization and cohesioni; client knoz7ledge and is significantly positively associated with each of skills; and leadership roles. these outcomes, even after applying multivariate controls for other influences (column 1). Social organization. Every society has its own Demand, or prior commitment made by ben- organization which influences interactions- eficiaries, therefore determines overall participa- namely, how resources are produced, allocated, tion. Demand also determines the investment and exchanged; how and with whom people beneficiaries made in capital costs and, to a lesser interact (or do not interact); and where people extent, investment in recurrent costs. It also influ- live and what they learn. Acknowledged or unac- ences the overall perceived net benefits of partic- knowledged, good or bad, effective or ineffective ipation. in the eyes of outsiders, the existing social orga- nization is the framework within which projects Beneficiary Capacity are placed. When the social fabric of societies is ignored, projects will run into problems sooner or Organization of beneficiaries is important in later; when the local social fabric is understood, managing participation. Several factors affect strategies that are more or less congruous with success in organizing for collective action, includ- the existing system can be developed (Cernea ing high stakes in a particular outcome; low 1993). transaction costs; and trust, loyalty, and reciproc- The Small Rural Water Systems Project in ity, from which follows elimination of free-rider Yemen, which serves approximately 70,000 peo- problems. To manage a water supply collectively, ple, illustrates how a project can be successfully there must be some organization for treating implemented using a traditional organizational water as a common property good, with some structure, with full cooperation from local lead- degree of excludability and subtractability. ers, sheikhs, and local development councils (see Table 5.2. Relationships between demand and participation outcomes and elements Beneficiary capacity Prior commitment Skills, Leadership of beneficiaries Extent clients knowledge Traditional among Participatony outcome (demand) organized of clients collective clients Overall participation 0.19, 0.23a 0.12 0.10 -0.14 (2.5) (2.1) (1.7) (1.16) (-1.3) Net benefits of participation 0.23a 0.05 0.02 0.04 0.01 (2.2) (0.3) (0.2) (0.3) (0.06) Investment in capital costs 0.63d 0.28 0.01 0.04 -0.01 (4.8) (1.4) (0.1) (0.2) (-0.4) Investment in recurrent costs 0.37 0.37 0.19 -0.11 -0.11 (1.9) (1.4) (1.2) (-0.6) (-0.4) Note: Significance levels are indicated thus: a significant at 0.05, b = significant at 0.01, c = significant at 0.001, and d = significant at 0 0001. Figures in parentheses are t-statistics. Factors Affecting Beneficiary Participation 49 box 5.3). The caveat here is that systems must be Water and Sanitation Committees, is instructive. developed to keep local leaders accountable. Each community was run by a Village Council, but Given the need to work with groups of peo- these had a history of factionalism and were ple, support agencies have often mandated the extremely political. The Village Water and formation of village groups or water committees. Sanitation Committees were designed specifically Unfortunately, such committees have often to circumvent the Village Council and thereby side- degenerated into form without function. step political infighting. Members of the commit- Committees have been created without any tee were largely self-selected. The committee was understanding of the preexisting social fabric, responsible for representing the project in the com- with standard rules imposed from the outside munity, resolving project-related problems and and with members who do not understand their conflicts, maintaining the systems, and requesting functions or responsibilities and who receive no technical assistance when needed. Unfortunately, training to equip them for their tasks. It is not sur- the project did not provide sufficient capacity prising, then, that numerous project evaluations building support. After playing an important role found project-created committees to be nonfunc- in implementation, committees found themselves tional, especially where such groups had no con- unable to fulfill their responsibilities or to help the trol or authority. community with new development tasks when the Projects that have proved to be effective, in project ended; many committee members lost contrast, have created small water-user groups or interest and stopped meeting. pump or tap committees, which have few mem- bers and generally consist of people who are well Traditional collectives. When community known to one another. These are federated groups are based on traditional organizations, upward into groups at the next higher organiza- there is an ongoing basis for cooperation, trust, tional level. The system of committees in the reciprocity, and conflict resolution. The transac- Malawi Self-Help Piped-Water System provides tion costs of transferring these traditions to solv- one such example (see box 5.4). ing a new problem are much lower than they are Beneficiary organization can take many when an artificially created group, with no forms, some based on traditional organizational shared history of working together, attempts the structures and others creating new organizations, same thing. Nevertheless, what is critical to elic- perhaps to get around a problem inherent in the iting participation is whether forms of interaction existing local organization. New organizations and decisionmaking are based on local traditions, are particularly vulnerable to the "project men- not whether the organization itself arises from an tality," however; if they are not provided capac- existing traditional group. ity building support, they dwindle away An externally funded project in Belize, which Functioning of groups. Highly participatory required that participating villages form Village and effective projects were studied in depth to Client orientation of agency Presence Autonomy of strong Use of local Participation of a project! Physical-target Consensus leaders knowledge a goal agency driven Flexibility on objectives 0.2 0.2 0.2a 0.17 0.06 0.02 0.01 (0.4) (2.5) (1.9) (1.8) (0.9) (0.2) (0.06) 0.14 -0.02 0.44b 0.15 -0.15 0.12 0.11 (1.31) (-0.2) (2.9) (1.2) (-1.6) (0.9) (0.8) 0.09 0.19 -0.10 0.25 0.16 -0.05 -0.02 (0.8) (1.4) (-0.5) (1.6) (1.3) (-0.3) (-0.2) 0.15 -0.01 0.26 -0.3 0.3 -0.11 0.22 (0.9) (-0.1) (0.9) (-1.6) (2.0) (-0.5) (0.9) 50 The Contributioni of People's Participation: Evidencefrom 121 Rutral Water Suipply Projects understand the processes used to manage local there was a breakdown. The same was true for organizations effectively. Some findings were written records. Successful groups in Kenya kept common to all of the successful projects: local no written record of money collected, but every- groups had a clear task orientation; projects one trusted the female treasurers, who brought worked through traditional leaders, including all the money to the meetings when asked to do religious leaders, in addition to the formal village so. By contrast, communities in Azad Jammu system; and, over time, leadership shifted to Kashmir, Pakistan, made detailed written records those most interested in the project outcome. available to visitors for inspection. Although agency preconditions and the Whatever their form and composition, func- nature of the water resource imposed bound- tioning groups were marked by clear allocation aries, groups evolved their own rules regarding and acceptance of duties. Additionally, the rules membership, fee payment, graduated fee sched- and regulations regarding membership, dues, ules, sanctions, and disposition of collected sanctions, and use of the facility were well known money. Those groups that started with standard- even if they were unwritten. As tasks were com- ized rules established in accord with project pleted successfully, water committees became guidelines sooner or later changed to rules more less active, unless they took on new development acceptable to them. They also engaged in suc- activities. This was particularly true when project cessful problem solving, as evidenced by the abil- technology required only simple maintenance. ity to solve physical problems (such as pump Government interventions have resulted in malfunctioning, borehole silting, pipeline breaks, the expectation that government will provide all and inadequate funds to buy diesel) and organi- services, a belief actively fostered by local politi- zational problems (such as dealing with changing cians. It would be easy in this environment for politics, resolving conflict, and enforcing rules). agency assistance to breed further dependency In other ways, groups varied widely. For rather than to create user groups that feel respon- instance, some collected money monthly, and sibility for their own water systems. Few agencies some collected money for repairs only when have the courage to walk away when it is clear Box 5.3. Yemen: role of local leaders in community mobilization In Yemen's 168 districts, or nahiyas (made up of clusters and in reviewing the proposed pipeline routes to take of villages), the sheikh, an inherited title, is the final into account long-standing differences among landown- authority and oversees all village activities. Local ers, tribes, or villages. Cooperative Councils for Development (LCCDs) are also LCCDs play a crucial role in facilitating community an important component of the local organizational commitments to the projects, including being a party to structure. In almost all public work and development financial agreements for projects. In cooperation with activities, LCCDs are responsible for identifying village LCCDs, villagers pay roughly one-third of the total con- needs, discussing the needs with villagers and the sheikh, struction cost of a project. Villagers construct and man- and submitting proposals to appropriate government age the distribution system throughout the villages, authorities. Once government approval has been including individual household connections. In addi- received for a public works project, the LCCD collects tion, villagers often extend the overall system, with assis- money from villagers to pay their agreed-upon share. tance from the LCCDs. Each village is represented by one LCCD member, Thus far, in one $20 million, externally financed pro- elected by the villagers for a six year term. ject, every community involved has participated and Working within this system, communities are able to agreed, before implementation, to take responsibility participate actively in almost all phases of a given water- for O&M. Communities assumed all costs for system supply project. After villagers and the LCCDs request a operation and maintenance, and successfully carried "subproject" from the Rural Water Systems Department out all repairs, minor and major, hiring private mechan- (RWSD) of the central government, the RWSD assigns ics when necessary. Through working closely with vil- the subproject to the project implementers. Sheikhs often lagers and implementing the project through the local play an important role in this process, lobbying the organizational system, the project's Phase II output of RWSD intensely on behalf of their communities. 100 systems was reached two and one-half years ahead Sheik/ts also play an extremely important role in help- of schedule. ing project staff conduct before implementation survey. Sheiklis assist the survey team in obtaining personal data Solurce: Laredo, Dawson, and Hashem (1986). Factors Affecting Beneficiary Participation 51 that communities do not perceive improved Jammu Kashmir, O&M is successfully performed water as the answer to their most pressing prob- by a hired mechanic selected by the Village Water lem. But the story can be markedly different if Committee (VWC), which is composed of elites. communities do perceive the water system to be In Indonesia, in Nusa Tenggara Timur, a VWC a key problem. Box 5.5 describes the evolutionary that includes women and the poor seeks a process that develops when communities are left mechanic from the private sector for repairs. In to formulate their own rules and regulations and some regions repair work is still carried out by manage their own development, with limited local government units. In India a government- external technical assistance. supported, two-tier system of maintenance- Organization for operation and maintenance unlike the three-tier system described in chapter was addressed specifically in the study, since 4-works well in some states. poor O&M has emerged as the major problem- atic issue in infrastructure projects. It is impor- Local knowledge and skills. The knowledge and tant to note that the form of organizational skills of clients have two important implications arrangements per se did not seem to have a for water-system projects. First, the rich informa- significant bearing on outcome. What was tion and knowledge systems that local people apparent was that effectiveness decreased on already possess (indigenous knowledge) will government agencies assumed control and determine how those clients evaluate and use responsibility for O&M. water systems. Second, local people's technical Possible organizational forms include opera- skills and knowledge regarding water and water tor/repair persons, with or without village water technologies can be put to good use. The partici- committees; democratic cooperatives; elite-led patory process taps the indigenous knowledge associations; local government units; private and extensive information that only local people enterprise; and the agency itself. Thus in Azad can supply about their environment. Box 5.4. Malawi: water committees The comerstone of the rural piped-water program in ask the main water committee elected for construction to Malawi is its well-organized structure of construction and supervise maintenance as well. maintenance committees. Nearly all of the committees are composed of ten elected members, with men dominating Maintenance committees construction and maintenance and women usually the majority of tap committee members. Local leaders, village The main water committee supervises repair teams, tap chiefs, and party officials are crucial to success, since they committees, and caretakers; raises funds through village determine whether committee authority will be respected. headmen; checks pipelines; reports problems; submits requests for additional taps; organizes self help labor; Construiction committees and settles disputes between tap committees and repair teams. The main water committee is responsible for the overall Repair teams are a technical arm of the main water management of the program; as well as for the setup of the committee. They carry out basic repairs on broken initial work program and the mainline digging program. pipes, and the chairman is responsible for the tools, Section committees, a part of only large projects, are equipment, and spare parts required for maintenance elected from villages located along various sections of work. the pipe that extend from the mainline. They draw up Tap committees are responsible for operation, care, and supervise the daily trench-digging program. and maintenance of a single tap. They organize periodic Branch committees are responsible for organizing cleaning of the tap site and soakaway pit and raise funds labor on branch lines after digging and backfilling on the from users for replacing taps and for repairing the mainline and section lines are completed. apron. Village committees are responsible for selecting stand- Village health committees, organized by field workers pipe sites, supervising the village labor, and ensuring that in the Ministry of Health, educate villagers and promote village attendance is maintained at committee meetings. improved health practices. Once construction is completed, the maintenance committee structure is introduced. Most communities Souirce: Warner and others (1988). 52 The Contribuition of People's Participation: Evidencefrom 121 Rtural Water Suipply Projects Project reports noted several beneficial out- cent of program budgets. Recent World Bank pro- growths of local knowledge and skills: local peo- jects in Asia, using a demand-based approach, ple guided outside surveyors to unplotted water allocate 15-20 percent of project costs to support sources; they helped design pipe distribution so capacity building, including training in partici- as to take into account land ownership and poten- patory approaches. tial social conflict; they were knowledgeable about water purification methods, seasonal dif- Role of elites and leaders. The capture of benefits ferences, rainfall patterns, underground water by the rich has been a troubling issue in the rural flow, and presence of underground rivers (con- water sector, and much emphasis has been placed firmed by hydrological maps); and they had their on creating water committees that include mar- own categories of "good" and "bad" water and ginal groups such as women and the poor. water sources. Experience shows, however, that without the The skills of clients are important in influ- involvement and support of the well-to-do and encing how quickly or how slowly people are the powerful, water committees may be unable to equipped to participate. Clients in successful effect change. The issue, then, becomes one of projects actively negotiated rules and condi- tapping the energy, imagination, skills, and tions among themselves and with agencies power of the rich to serve the poor. (especially rules regarding entry, use, and exit), But experience does not provide a simple for- determining effective management of the com- mula for tapping those assets; it just shows that mon property. Technical know-how among projects need to take account of the interests of all clients also was important in fashioning suc- the major stakeholders in a community. Success- cessful projects. ful projects have evolved three strategies for Some projects incorporate training to support incorporating those stakeholders, including elites: capacity building at both the community and first, co-opt the rich; second, widely disseminate agency levels, but most programs do not give information about the project and about its stip- training the careful attention it deserves because ulated conditions; and third, create small work- it is generally considered a low-status activity. ing groups in the hierarchy below the larger Cost estimates range from 0.5 percent to 25 per- village-wide groups. Box 5.5. Indonesia: the story of Mutis The WAS (Wanita, Air dan Sanitasi) program in water rights with a neighboring group in Kakaana, who Indonesia over a period of two years assisted community agreed to share water from the borehole in return for groups in Timor to initiate and manage their own water Mutis's help with maintenance and repairs. By mid-1987, systems with assistance from the technical govemment Mutis had collected stones, cement, sand, and Rp 35,000. ministries. The growth of groups and their management Mutis members contributed cement and pipes to the and problem-solving skills can best be illustrated by other water group. focusing on one group. It was in 1987, during a second round of intensive data When people in the village Silla heard of WAS in collection, that the group finally accepted that the drillers 1986, those in Cabang, a part of the village consisting of were not coming. They began to explore alternatives. By seventeen households, quickly formed a water-user mid-1988 the group had built three water collection tanks, group. They hoped that finally the little yellow marker with some technical guidance from outsiders, although in placed on the road several years before would become a 1985 they had laughed at the idea of drinking rainwater. borehole. Pak Minggus, a WAS field worker from the They also commissioned a well-digger with a promise to Department of Community Development, told the pay Rp 100,000; the well was 10 meters deep by mid-1988. group several times that they would not receive a bore- The group started holding meetings on the fifth of every hole, they were undeterred. month and tackled other problems. It set up an emergency Copying the other water groups that did receive gov- food fund, and began building one household toilet per ernment assistance, the group named itself Mutis, after month. Eggplants, chiles, and other vegetables were flour- the highest mountain in Timor. Every family contributed ishing in people's yards. A return visit in 1992 found the Rp 250/month and promised to increase that amount to group still functioning. Rp 500 once the water came. Yet, as Minggus had warned, no water came. The group finally negotiated Souirce: Narayan-Parker (1989). Factors Affecting Beneficiary Participation 53 Leaders and the rich are often easy to co-opt nant of participation in the presence of other fac- since their accustomed role is to be prominent in tors, but, as reported later in this book, leadership any project and to bring resources to the com- strongly influenced the extent to which local con- munities. If water problems affect these com- trol was established. munity members (and this is usually true), they too will have a direct stake in changing the water Conflict and factionalism. Leadership, tradi- situation. Their stake in improved water is likely tional or "modern" (official), is important in ini- to be even higher than that of other beneficia- tiating the community organization process. ries, since their needs are greater: the rich own Under any particular leadership, however, com- more livestock, need more water for other pro- munities are not homogenous wholes living in duction activities, and use more water for perfect harmony; rivalries, feuds, and factional- domestic consumption. ism can impede collective action. Follow-up vis- Benefit capture is easy for the powerful since its to four villages in Indonesia, eight years after they usually have good access to information. conclusion of a baseline study of project activi- Thus, in the Indonesian INPRES program, water ties, found that water improvement activities sources regularly ended up near the homes of vil- had ground to a halt in the one village that had lage leaders. Even when community participa- been divided by intense rivalry between the offi- tion is anticipated, as it was in the INPRES cial village chief and other traditional leaders program, communities cannot be meaningfully (Judd 1992). Another example comes from involved, nor can they become effective watch- Panama. For a study of the Panamanian dogs forcing public sector agencies to be account- SANAA-CARE community program, an evalua- able, if they are not informed about available tion team conducted extensive interviews with resources and the rules and regulations regarding community leaders and heads of the Patronato their use. (the community committee). The team found It follows that programs that incorporate the considerable conflicts within the communities rich and disseminate information intensively and (often along lines of political allegiance), which widely also stimulate transparency, accountabil- hindered community solidarity and implemen- ity, and helping behavior. The open nature of par- tation of the water system and watershed con- ticipatory decisionmaking is important: few servation projects. elites would wish to publicly abandon their role of magnanimity. Statistical evidence on beneficiary capacity. As should be obvious from the above discussion, Leadership. Development literature is replete measuring these nuanced sociological distinc- with references to the need for strong leaders to tions is a herculean task. Nevertheless, several start a movement, especially among the poor. Yet dimensions of beneficiary capacity and costs to leadership is not a quality that a program can beneficiaries for organizing for participation are immediately muster. It may be an important cri- reflected in the following measured variables: terion for village selection, however, particularly a Extent to which clients were organized before the if a project does not allow community self-selec- project began tion. Fifteen years into the implementation of * Skills and knowledge of clients projects, CARE identified three factors that pro- * Extent to which traditional collectives were used moted the effectiveness and sustainability of * Leadership of clients water projects: leadership, community organiza- * Presence of strong leaders. tion, and community financial investment in the Table 5.3 examines through multivariate water system (O'Brien 1992). regression analysis the relationship between each In the present study, a distinction was drawn of these independent variables and the variables between two levels of leadership, one being the being determined-namely, overall participa- presence of a strong leader who gets things tion, net benefits of participation, the extent to started and the other being broader-based lead- which control became local, and the extent of ership qualities among the clients. Neither type client exit from and dissatisfaction with the water of leadership turned out to be a critical determi- system. Findings include the following: 54 The Contriblltion of People's Participation: Evidencefrom 121 Rutral Water SLupply Projects Table 5.3. Relationships between beneficiary capacity and participation outcomes and elements Beneficiary capacity Skills, Leadership Presence Participatory Extent clients knowledge Traditional among of strong outcome organized of clients collective clients leaders Overall participation 0.23a 0.12 0.10 -0.14 0.20 (2.1) (1.7) (1.6) (-1.3) (0.4) Net benefits participation 0.05 0.02 0.04 0.01 0.14 (0.3) (0.2) (0.3) (0.1) (1.3) Extent of local control 0.48b 0.48c 0.32a -0.58 -0.04 (2.6) (4.2) (2.2) (-3.4) (-0.3) Client exit 0.20 0.06 -0.12 0.05 0.28 (0.6) (0.3) (-0.4) (0.2) (1.4) Client dissatisfaction 0.09 0.03 -0.25 0.17 0.13 (0.4) (0.3) (-1.6) (0.9) (0.9) Note: Significance levels are indicated thus: a = significant at 0.05; b = significant at 0.01; c significant at 0.001; d = significant at 0.0001. Figures in parentheses are t-statistics. * Extent to which clients were organized was sector, broadening objectives beyond construc- significantly associated with both participa- tion to include sustainability, targeting the poor, tion and local control. empowerment, equity, and cost sharing means * Skills and knowledge of clients was only mod- realigning inputs, rules, structures, and proce- estly significant (at the 10 percent significance dures. Accompanying new objectives, however, level) for overall participation, but was are new problems which have arisen out of lack strongly important for local control. of agreement on objectives, on needed changes, * Use of traditional collectives also had only a and on the consequent restructuring of sector modest association with overall participation; policies and institutions. the variable was more strongly related to the Project experiences reveal five major project/ degree of local control. agency strategies designed to realign structures, * Leadership of clients was strongly associated procedures, and incentives to build an agency with the degree of local control, but less so environment that can support client demand, ini- with overall participation. tiative, and capacity building: * Presence of strong leaders was unimportant. * Agency autonomy This variable is not associated with either * Adaptive planning approaches greater participation or greater local control, * Use of local knowledge and in fact is positively, although not * Changed objectives strongly, associated with client exit from the * Responsive implementation and intermed- system. iation. To summarize, in spite of the difficulties of measuring the client characteristics that relate to Agency autonomy. It is easier for agencies to be the capacity for participation, certain factors (that responsive to communities and support local is, the extent of prior organization, skill and capacity building when they have the mandate, knowledge of clients, the use of traditional col- resources, and authority to carry out planned lectives, and leadership qualities among clients) action independent of other agencies. One of the were found to bear a strong relationship to the problems that has plagued the sector has been extent to which control of projects became local fragmentation of responsibilities within several and, to a lesser degree, to overall participation. different government agencies. For example, pro- ject hardware, expertise, and resources typically Client-Orientation of Agencies reside with the public works ministry, whereas the responsibility for community extension systems Agencies must adopt strategies to respond to lies with the ministry of social welfare and depart- community needs and preferences and to sup- ment of community development, each with dif- port local capacity building. In the rural water ferent work programs and lines of authority. Factors Affecting Beneficiary Participation 55 Demand Client orientation ofagency Use Autonomy Prior of local Participation of a project/ Physical-target Consensus commitment knowledge a goal agency driven Flexibility on objectives 0 l9b 0.20 0.20a 0.17 0.06 0.02 0.01 (1.9) (2.5) (1.9) (1.8) (0.9) (0.2) (0.06) 0.23a -4.01 0.44b 0.15 -0.15 0.12 0.11 (2.2) (-0.1) (2.9) (1.2) (-1.6) (0.9) (0.8) 0.18 0.10 0.20 0.38b 0.01 -0.15 -0.31' (1.5) (0.8) (1.2) (2.6) (0.1) (-1.0) (2.0) -0.17 0.22 -0.50 0.17 -0.31 -0.03 -0.37 (-0.8) (1.0) (-1.7) (0.6) (-1.5) (-1.3) (-1.3) -0.26 0.10 (-0.30) 0.12 -0.17 -0.30a -0.30 (-2.0) (0.8) (-1.6) (0.7) (-1.3) (-2.1) (-1.8) The RUSAFIYA project in Nigeria illustrates rized in box 5.7 is illustrative.) What is needed what happens when a project lacks autonomy instead is a demand-based approach, which and when responsibilities are fragmented at dif- responds to decisions made by hundreds of dif- ferent levels among government agencies ferent communities-at their own pace, in their (Boerma 1993). The project attempted to actively own time-after implementation begins. involve communities in all stages of implementa- In demand-led planning, management adopts tion, and it met with success in some regions. In a learning-process approach, has clear goals and other regions, however, despite the dedication of objectives, expects and tolerates relatively high staff, the project encountered a range of prob- levels of uncertainty and ambiguity, plans for lems, which demonstrates how difficult it is to short time horizons, and operates with indicative create the right incentives environment when costs and budgets rather than within detailed, responsibilities for implementing a participatory inflexible budget lines. (See appendix 3 for a sum- process are fragmented (see box 5.6). Stronger mary of differences between the blueprint and "project units" are not the answer; instead, one or learning-process approaches). two local agencies should be given the resources and authority to implement projects, with Use of local knowledge. One of the important accountability induced through high visibility, benefits of participation is the opportunity to take external pressure, and incremental funding advantage of local knowledge and information based on performance. about the local context. The more the decisions Autonomy in managing tariffs is particularly are made by local people themselves, the less the important. Local councils and divisions of public need for outsiders to be informed about every health engineering departments in various parts detail of community life and preferences. It is of Asia, Africa, and Latin America collect sub- now quite common for projects to undertake stantial revenues. Those revenues are turned over sociocultural studies in preparation for rural to general treasury accounts, however, which water-supply projects, yet such studies are of lit- serves in the long run as a major disincentive to tle value if they do not yield information relevant local collection. to policy choices, or if the information is not used in management decisionmaking. Adaptive planning approaches. The blueprint, or Sociologists and anthropologists both inside masterplan, approach to planning projects and outside the World Bank have long argued for assumes that everything is known, and hence can the inclusion of local knowledge in the design and be planned, costed, and even procured in implementation of development projects (Cernea advance. Developing projects to match client 1993; Salmen 1987,1992; Warren 1991). The extent demand, however, calls for a different planning to which local knowledge is used is important to approach. (The Tanzanian experience summa- supporting beneficiary participation. Local 56 The Contributtion0 of People's Participation: Evidencefrom 121 Rural Water Supply Projects knowvledge can be harnessed directly, by involv- Changed objectives: participation anid capacity ing local people at different levels (especially in btilding. Project experiences show that when the planning and design stages); and indirectly, agencies possessed commitment and a mandate, through data gathering on technical, social, cul- local participation and capacity building were tural, political, organizational, and environ- achieved in a wide variety of cultural, political, mental issues in the local context. Beneficiary environmental and technological contexts. This assessments and other, more traditional social raises the broader issues of planning, priority, science methods are common ways to indirectly clarity, specificity, and consensus on objectives: gather data. Projects in the study sample used if local participation and capacity building are all of these methods. Box 5.8, which describes associated with effective projects-and they the evolution of water-user groups in Tunisia, are-then shouldn't they be project objectives? highlights the importance of sociological Project objectives of the twenty most effective knowledge in developing appropriate local projects are reviewed in appendix 1. The range organizations. of stated objectives for those projects was wide, Box 5.6. Nigeria: difficulties in managing incentives The RUSAFIYA project comprised five key actors: the The project provided intensive training in participa- community, the extension agents, the water and sanitation tory methods for the extension agents, but the pressure for units (WASUs), the local government authorities (LGAs), quick implementation sometimes led extension agents to and the project authorities. Many of the project's difficul- coerce uninterested communities to participate. This may ties arose because these actors lacked, in large part, con- have occurred because extension agents appeared to be trol over their participation in the projects, appropriate judged on how quickly they could get communities to incentives, and accountability for their actions. agree to participate and collect the necessary funds. Lack of control. Rather than allowing communities Finally, during implementation it became clear that control over their participation, project authorities often the cooperation of the LGA chairmen was important for preselected communities for inclusion in the project. In project success. However, the project presented few some cases, communities that did not want to participate incentives to influence these officials to participate effec- were persuaded or coerced into doing so by extension tively. It was politically difficult for local government agents. officials to actively support a project that helped com- LGAs also had little control over their participation munities become less reliant on the patronage of local in the project and little input into project design; LGA government. staff therefore had very little stake in the project's out- Lack of clear accouintability. The lines of accountability come. When problems occurred, LGA staff blamed pro- that were put in practice proved to be complex and dif- ject authorities. ficult to follow. WASU management decisions were over- Project authorities, for their part, also encountered seen by both the LGA and the project authorities, but the "control" problems in perceived conflicts over participa- WASU made financial reports only to the project author- tion and deadlines. Decisions made by project authori- ities. Since LGAs, the higher level, lacked authority over ties (for example, to await community commitment) project funds, they had little interest in the project and were overshadowed by donor insistence on quick con- obviously could not be accountable for fund misman- struction results. Donor insistence also succeeded in low- agement. ering the required community contribution from N 5,000 While the WASU was financially accountable to the to N 1,800. project authorities, the head of WASU, who had chief In Ningi, where the project met with more success, responsibility for implementation, was ultimately communities were allowed and encouraged to pay accountable to the LGA chairman who appointed him, pump suppliers directly. In other regions, however, com- leaving project authorities with few ways to deal with munities were expected to make payments to extension poor performance. In addition, despite a signed agree- agents, which created considerable anxiety. ment, community water committees had no recourse if Lack of appropriate incentives. Since construction often the WASU failed to deliver on its commitments. proceeded regardless of whether communities paid their Similarly, there were no sanctions that could be applied contribution, there was no real community investment in to the communities when they failed to comply with the system. Community interest was also affected by the agreed terms. fact that other government-funded projects were avail- able to the community at no cost. Source: Boerma (1993). Factors Affecting Beneficiary Participation 57 Box 5.7. Tanzania: master plan incompatibility with demand orientation To ensure that villages with the greatest need for water mentation phase began, making much of the data systems were helped first, the Government of Tanzania irrelevant. implemented a series of lengthy and extensive studies to Project implementers discovered that many of the develop Water Master Plans (WMPs) for its rural water- communities on the priority list did not have a high "felt supply program. From 1980 to 1983, project teams con- need." Demand for water systems was low in many of ducted water resources studies, water supply studies, these villages, as was community interest in meeting and socioeconomic studies, and undertook pilot projects project conditions. For example, villagers in one com- in three regions. Research for the socioeconomic studies munity were satisfied with their traditional water alone included circulating a 300-question survey in all source; they had much greater interest in constructing a villages; preparing a detailed study of water use, sanita- road and establishing bus service to a nearby town. tion, and health in 66 villages; developing a village par- Another community felt its need for a milling machine ticipation program; and conducting substudies on such took precedence over its need for a water system. Clearly, issues as water contamination, health education, and the agency-defined criteria had been ineffective in women's involvement. Results of these socioeconomic assessing demand within the communities. studies were not available in time to be included in the As a result, project implementers eventually set aside master plan. much of the WMP "priority list," arguing, among other A priority list of "high need" villages was developed things, that the information collected during the village in accordance with the WMP data that were available. survey in the WMP was insufficient or inaccurate; that High-need status was assigned to 644 of the 1,509 vil- the water supply conditions and needs had changed lages with the poorest accessibility to water, the lowest because of population growth or hydrological circum- source capacity, and/or the highest health risks. stances; or that political pressures had forced inappro- Ironically, the master plan process was so extensive and priate inclusion of some villages on the list. time-consuming that conditions within project commu- nities had changed drastically by the time the imple- Souirce: DANIDA (1987). Box 5.8. Tunisia: combining local knowledge with Geographic Information Systems The development of over 2,000 Tunisian water-user encouraged the expansion of these associations from the 30 associations through the Potable Water User project envisioned in the original project to more than 2,000 today demonstrates the importance of joining local knowledge 2. The water-user associations became instruments and modern information tools.a Together with the for rural democratization as they confronted, one after Central Tunisia Development Authority, the Govern- another, the Ministry of Finance to demand control over ment of Tunisia, and Tunisian geographers and sociolo- their own finances. In general, the Ministry of Finance gists (who assisted with spatial and socioeconomic either accorded such autonomy de jure or acquiesced by analyses), the Institute for Development Anthropology withdrawing objections. Many of the water-user associ- composed a methodology based on geographic infor- ations now have their own bank accounts. mation systems for the siting of thirty new boreholes in 3. Almost all of the water-user associations became the Kasserine and Gafsa governorates; the Institute entrepreneurial. The most common activity is recharging introduced the notion of water-user associations at each of automobile batteries, using step-down transformers new site. The methodology used socioeconomic, demo- driven by well pumps. Other new economic activities graphic, political, and environmental criteria, as well as include women's handicrafts and weaving, as well as conventional financial and hydrologic yardsticks, to bathhouse and greenhouse operation where wells tap a propose sites. thermally heated aquifer. Over time, as strategies evolved, the project pro- duced three unanticipated results: a. This project was not part of the 121-case sample. 1. The Government of Tunisia became an enthusiastic Souirce: Personal communication (1994); International supporter of the concept of water-user associations and Development Association (1992). but, while many projects mentioned participa- Goal of beneficiary participation. Local knowl- tion, capacity building, and empowerment as edge is unlikely to be used, despite its proven specific objectives, many did not. Many of the value, unless managers care about sustainable less effective projects had no clearly stated results and local participation. Did agency staff in objectives. fact have an incentive, either monetary or non- 58 The Contribution of People's Participation: Evidencefrom 121 Rural Water Supply Projects monetary, to facilitate participation? They did in of the participatory objectives should be the those instances where achieving beneficiary par- principal criterion of success, even if in practice, ticipation in decisionmaking was a clear goal or this required a substantial reduction of the pro- objective that was monitored and evaluated. gram" (quoted in Therkildsen 1988). Yet the par- Other incentives for encouraging participation ticipatory objective, as well as the cost-sharing come when local participation is a criterion in objective, was abandoned in favor of speedier personnel evaluations and a measure of project physical construction. success (see appendix 2). Myriad innovative If projects are participatory-and if they make ways were found to monitor and evaluate partic- achieving participation a goal to which they are ipation of beneficiaries in decisionmaking. committed-they do whatever is necessary to Participation is often viewed as a messy busi- achieve it. This includes adopting a learning- ness that inherently defies specificity. Tasks that process approach to management. lack specificity are difficult to undertake and impossible to monitor and evaluate (Israel 1987). Goal of building local capacity. Local participa- Hence, one challenge is to make participation tion occurred when it was supported by a focus itself more definable and specific and less on local capacity building and on strengthening monolithic. It is also true that there are many the local social organization to manage the phys- ways to reach the same goal. Tracking participa- ical infrastructure. Clarity and prioritization of tion throughout the stages of a project can help objectives is essential for those projects that are establish accountability for achieving participa- based on demand and participation and that use tion goals along the way, without forcing prema- a learning-process approach, for two reasons. ture standardization of the means of getting First, the objective of speedy, efficient physical there. Participatory Evaluiation: Tools for Managing construction often conflicts with the objectives of Change in Water and Sanitation (Narayan 1993) capacity building and sustainability. In the clas- details several ways that participation can be sic dilemma, the desire to maximize the produc- monitored and evaluated.28 tivity of drilling rigs must vie with the desire to Increasingly, programs are monitoring partic- wait for community readiness. Most projects in ipation and capacity building at the community the study sample that were highly effective and level and using results to refine strategies. For participatory dealt with such dilemmas by mak- example, the Karonga Lakeshore Integrated ing community organization and capacity a spe- Rural Groundwater Supply Project in Malawi cific, high-priority goal to which they invested complements technical monitoring with socio- substantial money and time. The Togo Rural logical monitoring (see appendix 4). Several Water Supply project is one example. According important insights gleaned from the sociological to project estimates, 25 percent of the project monitoring led to changes in program strategies. budget was spent in supporting local capacity The village water and health committees, which (see box 5.9). were meant to be the main implementing orga- The second reason for clarity and prioritiza- nizations, were found to be less effective than the tion of objectives is that an adaptive learning- pump committees. Surveys revealed that 77 per- process approach by definition means that there cent of the user groups were not even aware of is no implementation blueprint to follow. When the village committees, but all of the user groups dealing with uncertainty, it is important not only knew about the pump committees. The program to be clear on objectives-and to establish bench- subsequently focused its training efforts on the marks and intermediate steps, or to specify out- pump committees. Making participation an indi- puts-but also to allow, and indeed encourage, cator of success and establishing criteria to mon- local solutions to be developed, even though no itor and evaluate participation obviously cannot two communities or field offices may come up make participation occur. Documents for the with exactly the same process. If new objectives World Bank-assisted project of 1977 in Tanzania are not spelled out and emphasized, project staff stated that the Bank and the government of will most likely seek the security of the familiar Tanzania "should refrain from equating success procedures of the past. Even worse, agencies may with quantitative criteria. Genuine achievement standardize methods too early and develop Factors Affecting Beneficiary Participation 59 guidelines on how the community should partic- contrast, successful water committees in C6te ipate, which amounts to the kiss of death for exer- d'Ivoire stepped outside the official guidelines to cising voice and choice. When detailed rules sell water, which gave incentives for careful rather than broad principles are promulgated, monitoring to water minders and furnished participation and sustainability flounder. finances for repairs. The temptation to standardize all procedures Project experiences indicate that when local early, even before implementation starts, is espe- capacity building and decisionmaking are high- cially great in large projects. Staff for large pro- priority goals, effective managers allow time for jects are scattered over wide geographic areas, participation to take root. Typically, a lead time of come from different disciplines, and may have to one to three months is needed so that people can deal with more than one organization. In such a express demand and organize themselves and situation, a common vision, and autonomy (sup- the project can begin to establish a "trust part- ported by financial resources) to pursue what- nership" and dialogue with community groups. ever works in a particular context, is critical to This initial period may extend to a year, depend- project success. Accountability is established ing on the degree of community autonomy being through the monitoring and evaluation-with sought. For example, CARE projects in Indonesia, communities playing active roles-of process which aim for total community self-financing indicators and intermediate outcomes. (including capital costs), have found that initial In the Zimbabwean study mentioned earlier, organizational work takes about a year. Cleaver concluded that national guidelines, Preparatory work also takes longer if demand when painstakingly and unswervingly fol- must first be further generated through educa- lowed, were counterproductive to community tional and awareness raising activities. responsibility and involvement. Examples To summarize, participation is strongly sup- abound. For one, the implementation policy that ported if the design of the project has evolved states that the community should dig the first based on local knowledge systems, if community three meters of a well was often misinterpreted capacity building is supported, and if participa- to mean that the community should dig up only tion is made a goal that is valued, monitored, and three meters. Similarly, guidelines for commit- rewarded and linked to project evaluation crite- tees suggest the posts of chairman, treasurer, and ria. The management challenge therefore is to secretary; these officers are dutifully elected, but create a framework in which the overriding often they have no duties and thus do nothing. objective is clear, higher levels of ambiguity and Informal user groups were frequently found to uncertainty are tolerated (since not all is known), be more effective than the official committees. In and mechanisms and incentives are included to Box 5.9. Togo: investing in capacity building In the $16.7 million water supply project in Togo village at an initiation ceremony. (1980-87), one of the main project objectives was to form, VDC officers; village pump mechanics; women vol- or strengthen, every village development committee unteers responsible for pump maintenance and oral (VDC). Project managers hoped that training communi- rehydration therapy; and latrine and cistern mainte- ties to organize themselves would equip communities nance volunteers received extensive training. VDC offi- for solving not only water-related problems but other cer training included how to open a bank account, development problems that might arise. responsibilities of each office, how to obtain spare parts, Initial organization of the VDC took at least six to and how to run community meetings. twelve months, with Social Affairs and Sanitation field In total, 864 VDCs were established and training pro- agents providing follow-up support to each village. vided to each. Besides participating fully in the con- VDCs were created through a series of four village-level struction and maintenance of the water system, village meetings between the field agents and the villagers, of people, under the direction of their VDCs, went on to which VDC members were selected to represent the construct roads, market centers, pharmacies, and schools community and project activities, objectives, and mutual for their communities. responsibilities were discussed. The seven to thirteen people chosen for each committee were introduced to the Souirce: Roark and others (1988). 60 The Contributtioni of People's Participation: Evidenwcefrom 121 Rural Water Suipply Projects foster client decisionmaking, responsibility, and more community-focused and hence adaptive to capacity. varying community situations. Responsive implementation. Project design ele- An example of the success that can be ments can increase the ease with which partici- achieved when agencies are responsive to the pation can be induced. A key characteristic of a desires and pace of communities, and respect successful demand-based approach is that the community autonomy, comes from the story of agency is responsive to beneficiaries, adjusts and community-managed water utilities in Meru in adapts to client demand, and enhances benefi- eastern Kenya and Ngorika in central Kenya ciary capacity. Responsiveness implies that solu- (Bess 1990). Box 5.10 highlights the process that tions and strategies are tailor-made for each led to sustainable outputs through timely techni- situation, which in turn implies endless variabil- cal, financial, and management assistance to a ity in detail. In this scenario, consensus on over- community water society. The society evolved all objectives and means becomes particularly through the federation of a number of small, self- important, although solutions are open-ended help groups that could not have developed fur- rather than predefined. ther on their own. Engineering agencies have typically had pri- Research for this study indicates that projects mary responsibility for delivering rural water ser- that were high in responsiveness were also flex- vices, yet experience establishes that their technical ible in implementation, allowing time for com- blueprints-the pride of most engineers-and a munity organization. This study cannot "build-and-transfer" approach to projects do not definitively say whether projects achieved effec- foster dialogues and partnerships with communi- tiveness through taking the time to involve ben- ties before construction begins. Changing the focus eficiaries or whether some other method of the engineering departments, however, would produced the involvement. Regardless, since cause most engineers to lose professional interest participation is positively correlated to effective- and pride in their jobs. A young engineer in ness and ultimate efficiency, trade-offs between Pakistan in the Public Health Engineering the time needed to organize participation and Department summed up the sentiment thus: "I the amount of time required for implementation went to engineering school and learned how to are likely to be positive. Inducing participation design water systems; if I am now told to go spend may be time-consuming upstream but may save the day talking to villagers who do not know what time downstream. I know, I feel very frustrated. I would rather spend my time in front of my drawing board; otherwise Clients' listening to agency staff. Agency flexi- what use is my engineering degree?" bility and autonomy are preconditions for adapt- Others have adapted to working in collabora- ing programs to the conditions of a particular tion with extension workers from different min- environment, but several routes are open to agen- istries, but the underlying tension between the cies for responding to the requirements of differ- efficiency orientation of engineers and the stress ing contexts. Without some system of outreach on community readiness remains. As an engineer and connection to clients, either direct or indirect, in Kenya put it: "From working with extension it is impossible for an agency to be responsive and workers for three years I have finally learned how adaptive. to hurry the people slowly." Outreach can be interpersonal or it can be Of the twenty projects with the highest ratings channeled through the mass media, or both. (summarized in appendix 2), almost 90 percent Gathered information can be brought to the were implemented by NGOs and government agency office for planning purposes, or it can be ministries other than the public engineering put to service directly in the field after consulta- departments. The governmental lead agencies tion with local groups. Either way, effective out- included local government offices for community reach would be difficult, if not impossible, in the development, health, social welfare, and absence of agency staff with the requisite skills, women's affairs. NGOs are not necessarily more technology, and mandate. participatory than government agencies, but Many projects used extension workers as their staff and method of operation are generally their contact point with clients, yet, in general, Factors Affecting Beneficiary Participation 61 clients listened more to members of village- efforts, the effectiveness of different combina- based organizations than to extension agents, tions of extension workers and agencies was not intermediary NGOs, or local government units examined. (data were scored separately for each type of Overall, village-based organizations played a organization). Nevertheless, the degree to major role in 20 percent of the projects, extension which clients listened to the extension workers workers in 22 percent, intermediary NGOs in 15 was highly correlated to overall participation (r percent, and local government units in 3 percent. = 0.72), which is not surprising because benefi- Conversely, 75 percent of all projects had practi- ciaries who care about the product being sold cally no involvement with local government; 57 (the water system) are the ones who are more percent had no involvement with village organi- likely both to listen and to participate. The only zations, 48 percent saw no extension workers, factor that was significantly related to client lis- and 73 percent did not use intermediary NGOs. tening was the prior commitment of clients. The The study comprised a wide range of projects, form of the outreach also was analyzed, but including many that had little interest in sup- because projects used a variety of outreach porting participation. Box 5.10. Kenya: community-managed water utilities The communities of Murugi-Mugumango and Ngorika Other penalties play a role in keeping the systems in Kenya stand out as examples of successful commu- commercially viable. Both communities have stringent nity-owned water utilities run along commercial lines.A rules regarding nonpayment of fees or illegal connection Both utilities started out as small, self-help water groups to the water system. In Ngorika a member who bypasses in the mid-1970s and evolved into sophisticated water a meter may be expelled from the society for life, be forced societies with over 1,000 members each. The monthly to pay a large fine, or have to plead her or his case before revenues for the Murugi-Mugumango Water Society are the general members to be reconnected. Nonmembers are approximately Ksh 35,00040,000,b sufficient to cover taken to court. Ngorika has had no cases of illegal con- their monthly expenditures of Ksh 30,000-35,000. While nection since the society's development, as compared the Ngorika Water Society is smaller and less well estab- with the earlier illegal connection rate of 20 percent. lished than Murugi-Mugumango, its monthly revenues Both water societies use disconnection as the penalty of Ksh 8,000-10,000 are also sufficient to cover monthly for nonpayment of fees, but they try to be flexible on this expenditures (Ksh 7,000-8,000). In addition, both com- issue. They will accept partial payments on outstanding munities have consistently expanded the systems and balances, and they refrain from disconnecting if a mem- operations over time, building offices, employing staff, ber explains his or her difficulties before defaulting. and widening distribution. Reconnection fees are high, which discourages members Two factors contribute to the success of these com- from taking advantage of leniency. munities' operations. First, both water societies Both water societies are flourishing. In 1985, when received timely management training from NGOs and Technoserve's (an NGO) management assistance con- technical assistance (for construction design and super- tract with Murugi-Mugumango ended, the water society vision) from the Ministry of Water Development. had Ksh 500,000 in the bank and about 20 km of main and Second, these organizations worked with the water distribution pipeline with over fifty individual connec- societies to develop a commercial framework (based on tions. Since then, 60 km of main piping have been laid paying for water) and a full set of rules, by-laws, and and 1,160 members have obtained metered connections guidelines. Members defined the objectives of the soci- to the system. The society has built an office complex, ety and delineated the rights and responsibilities of employs 18 permanent staff, and has total liquid and each member of the water cooperative and the penalties property assets of almost Ksh 800,000. Ngorika is a for noncompliance. younger organization, but in 1990 it had almost 60 km of These member-created rules foster ownership and main and distribution pipe, as well as 36 km of connect- responsibility to the water system. For example, both ing pipe. It had 320 individual connections serving some water societies require that each member contribute a 2,500 people, as compared to 48 connections serving 456 certain amount of labor as partial payment of fees. Cash people before the society's formation in 1988. payment in lieu of labor is actively discouraged by finan- cially penalizing those who pay in cash. In Ngorika, for a. This project was not part of the 121-case sample. example, an individual would need to pay Ksh 100 in b. US$1.00 = Ksh 54 in 1994. cash to satisfy a day of labor valued at Ksh 25. Soutrce: Bess (1990). 62 The Contribultion of People's Participation: Evidencefrom 121 Ruiral Water SLipply Projects Statistical evidence on agency characteristics. targets reduce the perceived benefits of partic- Table 5.4 presents the relationships between four ipation and (even weaker) evidence that clients outcome variables and the factors discussed are unresponsive in target-driven projects. above. The statistical findings generally support * Greater flexibility is strongly associated with the lessons learned from in-depth analyses of greater agency responsiveness. The reason- individual cases. The outcome variables are par- able conclusion is that an agency cannot be ticipation, net benefits of participation, and two responsive to client demands unless it has intermediate elements or measures of participa- some degree of flexibility. tion, namely, agency responsiveness to clients * The extent of prior consensus on objectives and client responsiveness to the agency The table had a modest influence on agency and client shows the following: responsiveness. * Use of local knowledge contributes to both greater agency responsiveness and greater Policy Lessons overall participation. * Achieving participation, when defined as a The key findings presented in this chapter offer goal of the project, has a net positive influence important policy lessons. To recap, beneficiary on actual participation. It has an even stronger participation contributes to project effectiveness effect on the perceived benefits of participa- and can be induced in a variety of contexts. Five tion, as well as on increased agency respon- factors affect beneficiary participation: siveness to clients and client interest in agency * High demand and financial commitment field agents. made by clients before implementation * Autonomy of the project within the agency is * Degree to which beneficiaries are organized associated, although weakly, with greater par- * Degree to which agencies incorporate local ticipation.29 However, there is no evidence to knowledge into project design and imple- suggest that project autonomy will increase mentation and invest in capacity building agency responsiveness in the absence of addi- * Degree to which agencies make participation tional incentives. The estimated impact of a goal, the achievement of which is rewarded, greater autonomy, controlling for other deter- monitored, and evaluated minants of agency responsiveness, is negative. * Degree to which agencies are relatively * Projects that are physical target-driven are not autonomous and able to function without more (or less) likely to be participatory or to external interference. have better responsiveness from the agency. Since participation is a long-term, iterative, Moreover, there is weak evidence that physical nonlinear process that is difficult to standardize Table 5.4. Relationships between client orientation and participation outcomes and elements Client orientation of agency Autonomy Participatory Use of local Participation of a project/ Physical-target Consensus outcome knowledge a goal agency driven Flexibility on objectives Overall participation 0.20b 0.20 0.17 0.06 -0.02 0.00 (2.5) (1.9) (1.8) (0.9) (-0.2) (0.06) Net benefits participation -0.02 0.44b 0.15 -0.15 0.12 0.11 (-0.2) (2.9) (1.2) (-1.6) (0.9) (0.8) Agency responsiveness 0.29b 0.29a -0.13 0.07 0.36b 034b (2.7) (2.0) (-1.2) (0.8) (3.2) (2.7) Clients listen to field agents 0.01 0.30 -0.08 -0.07 0.22 0.23 (0.1) (1.8) (-0.6) (-0.7) (1.5) (1.6) Note: Significance levels are indicated thus: a = significant at 0.05; b = significant at 0.01; c significant at 0.001; d significant at 0.0001. Figures in parentheses are t-statistics. Factors Affecting Beneficiary Participation 63 or capture within a predecided time framework, the community, for example) is essential at the it is useful to identify important elements, or subproject level. intermediate benchmarks, of the participatory * Roles and functions of different actors at dif- process. Three of those elements were found to be ferent levels must be clear and yet avoid the extent to which communities were achieving overspecification of the details of implemen- control and ownership of the projects, the tation. Roles and functions should dovetail amount of capital and recurrent costs that clients resources, responsibility, authority, and capac- assumed, and the responsiveness of agencies to ity. Implementation details and rules should clients. Local control and ownership was more be worked out with each community likely to come to pass when clients were orga- * Local capacity building is essential to main- nized, when their organization was based on taining physical infrastructure. The degree of local collectives, and when they were skilled and investment needed in local organization displayed leadership qualities. varies, however, depending on objectives, Furthermore, beneficiary participation is preexisting local capacity, and functions to be likely to occur when both agencies and clients see performed. high benefits from such participation. The two * Blueprint master planning, budgeting, and determinants of high net benefits are client procurement approaches cannot be used in demand and agency commitment as signified by demand-based national programs that sup- making participation an agency goal. port community initiatives. The course and The foregoing summary of findings intimates timing of community efforts to organize and that the key issues in achieving sustainable rural the appropriateness of a particular technology water services are institutional rather than tech- at a particular time are not predictable, and nological (although appropriateness of technol- thus are not amenable to a master plan ogy obviously plays a part in effective projects). approach. The policy lessons are several: * Because many engineering agencies have Beneficiary demand and agency client- trouble accommodating a community- orientation are central to achieving sustain- centered approach, it makes sense for agen- able services. Willingness-to-pay studies may cies that are already community-focused to be be useful for setting policy parameters at the the key implementing agencies for projects. macropolicy level, but instituting a commu- The still significant role of engineering agen- nity self-selection process based on "hard cies would then be to offer technical assistance choices" (a substantial cash contribution by and support to the implementing agencies. Demand Beneficiary capacity Skills, Leadership Presence Prior Extent clients knowledge Traditional among of strong commitment organized of clients collective clients leaders 0.19b 0.23' 0.12 0.10 -0.14 0.02 (2.5) (2.1) (1.7) (1.6) (-1.3) (0.3) 0.23a 0.05 0.02 0.04 0.01 0.14 (2.2) (0.3) (0.2) (0.3) (0.06) (1.3) 0.07 -0.01 0.10 -0.06 -0.10 0.00 (0.7) (-0.1) (1.1) (-0.6) (-0.8) (0.0) 0.30b -0.02 0.14 -0.19 -0.17 0.15 (2.6) (-0.1) (1.1) (-1.3) (-0.9) (1.3) I I I I : j CHAPTER 6 Translating Lessons into Design Features T he lessons derived from the study and By asking this question, several rural water sup- analysis of 121 rural water supply projects ply projects in the World Bank's portfolio have are clear. The challenge is to apply these transcended how-to manuals. Although the pro- lessons to the design of large-scale projects within jects are at various stages of preparation or imple- a supportive policy framework. mentation, each has innovated ideas to ensure Two overriding messages emerge from the that participatory projects to reach the poor are study: (1) beneficiary participation is critical for demand-responsive. They highlight workable achieving project effectiveness and building local strategies and offer evidence that lessons can be capacity; and (2) rural water projects have to be fuin- put into practice within the framework of the damentally redesigned to incorporate participation. Bank project cycle. When participation is merely one project activ- The projects, and results from this study, show ity among many, it fails to transpire. When agencies seven design features to be particularly impor- do not offer a menu of options, but instead supply tant for developing demand-responsive partici- one standard technology package irrespective of patory projects: community demand, problems arise. When local * Selection criteria people do not feel ownership of a project or water * Demand assessment points, projects falter and facilities fall into disrepair. * Institutional framework Even when agencies have the best of intentions and * Funds and financial flow highly trained staff, participation will not become * Choice of technology and service level institutionalized unless user demand is the guiding * Planning approach: master plan versus learn- principle permeating the design of projects. ing process The obvious conclusion is that participation is * Monitoring and evaluation. stymied unless the desire for "full community par- This chapter examines these design features ticipation in all phases of a project" is supported within the context of Bank projects.30 by changes in key design features. Efforts must go beyond manuals that detail the standard applica- Selection Criteria tion of "twenty-two steps to achieving participa- tion." Redesign is essential for ensuring effective The most common criteria for selection of districts, participation. subdistricts, and communities are water scarcity, The first question to ask of any project design poor water quality, presence of water-related dis- feature is does it respond to community demand? By eases, difficult hydrogeological conditions, and 65 66 The Contribution of People's Participation: Evidencefrom 121 Rutral Water Suipply Projects poverty. Typically, the villages selected for projects At the outset, the project prepared only a lim- are culled from a master list, in accordance with ited number of engineering designs to provide these criteria or some variant of them. guidance on what costs might be under different As projects move toward a clearer demand circumstances. Community-submitted projects orientation, two criteria become paramount: are appraised on a case-by-case basis. The level at demand and poverty. When the implementing which the project is approved depends on the agency cannot accommodate all of the villages on total cost of the project. All project proposals have the master list, other criteria are used to stratify to demonstrate that certain planning tasks have the sample and to help rank the priority of vil- been carried out and that the project is technically lages. However, if community demand is not sound and ready for implementation. given primacy and if selection criteria are not The criteria applied in a five-step project transparent and supportive of decisionmaking at approval process are described in box 6.1. The the local level, the scope for political selection of preparation process for subprojects will be fur- villages increases dramatically regardless of com- ther simplified based on an examination of the munity demand, as experience bears out. first year's experience. The Sri Lanka Rural Water Supply Project Although some countries still select commu- focuses on community-based planning, using a nities according to externally determined criteria series of selection criteria as the centerpiece of its that indicate the severity of the problem (distance subproject review process. To ensure a demand ori- to a water source is one), several Bank projects entation, communities have to apply to the project have adopted a kind of scoring system, whereby and request inclusion. Hence, the project does not an initial list of villages is compiled and final prepare lists of villages, nor are the villages in which selection is based on community interest and the project will be implemented predetermined. demand. Box 6.1. Sri Lanka: guidelines on community preparation and community commitment In the Sri Lanka rural water supply project, the fol- Stage 3: cost analyses lowing guidelines are used by coimunities to pre- The scheme should be cost-effective. The cost of the pare sub-projects and by the agency to judge scheme to the program will be calculated on a per community commitment and technical appropriate- household basis. There is a built-in bias towards ness of the project proposal. communities paying a higher proportion of costs for Stage 1: technical criteria higher service levels. . Establish that water supply improvements are If the proposal is still unacceptable by standard cri- teria, at this stage it is passed to the national level for necessary * Select and accept CBO (community-based organi- more detailed assessment. zation) for project management Stage 4: cost-benefit analyses * Acquire necessary land When the costs are higher than normal because of * Resolve any water use conflicts (water rights) unusual features-hydrogeological or due to Determine project to be technically sound remoteness-costs have to be justified in terms of * Coordinate village scheme with other villages time savings. * Complete project proposal form. If the proposal is not found satisfactory, it is Stage 5: extreme need appeal returned for revision and resubmission. If costs cannot be justified on time savings, acute Stage 2: sustaintability criteria need has to be established based on: To plan early for long-term maintenance of tech- * Quality of existing water supplies nologies chosen, four factors are given attention: * Quantity provided by existing water supplies * Level of organization of community determined * Community ov erty. * Willingness to pay adequately illustrated * Community poverty. * Environmental sustainability ensured * Reliability of future supply from the source ensured. Souirce: World Bank (1992). Translating Lessons into Design Features 67 Thus, in the Indonesian rural water supply the sole consideration. Self-selection into a pro- project, the government goes through a three- ject is a solid indicator of the community's collec- step process. Step one calls for drawing up a tive demand and its willingness to make hard "long" list of villages-1,300 of them in the cur- choices. thus, in the Sri Lanka project, commuri- rent project-based on water scarcity, poverty, ties (together with a local, nongovernmental remoteness, water quality, and existence of other implementing agency, if needed) have to fill out ongoing projects; a short list is produced after very detailed project proposals, submit evidence screening for poverty (step two). The final selec- of consultation with different segments of the tion of villages (step three) will be based on the community, prove that land and water right con- Village Action Plans (VAPs) submitted by com- flicts have been resolved, and agree to pay the full munities, which must demonstrate the village's cost of O&M and 20 percent of capital costs. As willingness to take full management and finan- the service level rises, the community share of cial responsibility for O&M. capital costs increases. In the Pakistan rural water supply project, Demand Assessment communities must sign a legal document, a "memorandum of understanding," which stipu- There is clearly a need to assess community lates the mutual responsibilities of the different demand in a way that confirms the commitment parties. In Azad Jammu Kashmir province, com- of a group or community. Increasingly, commu- munities make cash and material contributions nities are required to establish their interest in before the local government enters into an agree- and capacity for undertaking a water project by ment with communities to improve the water sit- hurdling a series of tasks specified in the project. uation. The process itself, which has been used Three tools commonly used to assess for the last twenty years, is kept simple and non- demand are willingness-to-pay studies, partici- bureaucratic. patory needs assessment, and beneficiary SENASA, the water and sanitation agency in assessment. Willingness-to-pay studies estimate Paraguay, recognizes that community participa- the value people place on an improved water tion is essential for sustainability. After conduct- supply by determining how much people are ing willingness-to-pay and affordability studies, willing to spend for such improvements. SENASA enters into a contractual agreement Participatory needs assessment involves com- with the junta (a community's water and sanita- munities themselves in assessing their own tion committee), but only after the community needs and in determining whether water supply has fulfilled several specified tasks that establish improvement is a priority around which the its interest in the project. community can-and wants to-organize. For a beneficiary assessment, outsiders speak to, Institutional Framework interview, and observe potential participants to find out whether water is perceived as a problem The lead agency in most countries is usually the and, if so, what the community is willing to do public works or water engineering department. to solve the problem. The three techniques com- Most of these agencies are neither trained nor plement one another and can be used separately interested in working with communities. The or together. Only participatory needs assess- inefficiency and leakage from these large, central ment, however, sets the stage for community- public work agencies have led to a search for initiated planning and action. alternative institutional arrangements and inno- But even when such assessments are under- vations. Another problem stems from the frag- taken, projects will not be demand-based unless mentation of responsibilities among agencies and the studies are used to determine project strate- the need-most frequently unmet-to coordinate gies. Hence, rather than narrowly equating across departments. It is interesting to note that demand with willingness to pay, many projects in countries where large-scale success has been are developing strategies for community self- achieved, the key implementing agencies have selection into a project; willingness to pay is an often been other than public engineering depart- important consideration, of course, but it is not ments. Thus, in Malawi, the rural water supply 68 The Contribution of People's Participation: Evidencefrom 121 Rutral Water Sutpply Projects program is led by the country3s Department of tricts; and (3) prepare a follow-up project, apply- Community Development; in Colombia, by the ing the community-based approach developed National Institute of Public Health; and in Togo, and tested in this project, to complete coverage in by the Ministry of Social Affairs, Public Health the districts cited above and in other districts for and Status of Women. which DDPs have been completed. Different countries are trying out different Since no single agency is responsible for rural strategies, including decentralization (for exam- water supply, the government will create a ple, fiscal decentralization and strengthening of a Community Water Supply and Sanitation central ministry's field branches), greater involve- Program Unit (CWSSPU) based in the Ministry of ment of the private sector (including NGOs), and Housing and Construction. A national steering creation of new units or departments that are bet- committee will advise the CWSSPU, but the new ter equipped than existing ones to reach out to unit will have full financial control and autonomy communities and support community action. to apply project review criteria, approve projects, Although all Bank-supported projects include and disburse funds. The CWSSPU will not directly "institutional strengthening" components, the implement projects itself; it will stead contract ones that aim at radical restructuring of the sector with other, intermediary implementation agen- and policy reform include institutional reform as cies (lAs) for that work. The National Steering a specific objective, and they invest substantial Committee, which will be the policymaking body, resources in capacity building and training for will include members of the CWSSPU and repre- agencies. sentatives of NGOs and community-based orga- The key agencies for rural areas in Sri Lanka nizations. Hence, experience from the project will are the National Water and Sanitation Drainage inform policymaking and produce legislative Board (NWSDB), whose main focus is urban; the changes needed to support the community-based Ministry of Health, for sanitation; and 257 local approach on a national scale. authorities. The Government of Sri Lanka relied The NWSDB together with other specialized primarily on bilateral assistance for investment in agencies will provide support to the CWSSPU to rural areas, and lack of comprehensive policy or build the capacity of the intermediary imple- plans led to the introduction of over fifteen dif- mentation agencies. The NWSDB will assist with ferent handpumps in the country during the drilling; under contract, large NGOs and profes- 1980s. As a result of a review of sector policy, the sional organizations will provide training for government adopted the District Development other technical tasks and help the implementa- Plan (DDP), which employs a planning-by-dis- tion agencies devise community development trict approach. The sector analyses also revealed strategies. inefficiencies in the sector; the government con- Any agency, either an NGO or one in the pri- cluded that sustainability of rural water supply vate sector, or the Pradeshya Sabha (the lowest- systems depended on communities and the pri- level elected authority) can apply to be an vate sector, including NGOs. The govermnent implementation agency for a project. An IA has role was redefined as regulator and promoter. to submit proof of its capacity to undertake the The Sri Lanka rural water supply project, project, its past achievements, and the willing- which is supported by the International Devel- ness of community-based organizations to work opment Association, will be implemented in with it in implementing projects. Thus, all project three districts. The aim is to develop a commu- activities will be managed by the IA, working in nity-based strategy that can be applied nation- close collaboration with a community-based ally. Project objectives include the following: organization or a community group at the village (1) develop systems and institutions for level. Community-based organizations can also community-based planning, implementation, apply directly for grants. operation, and maintenance of cost-effective and These institutional arrangements mark a rad- sustainable water supply, sanitation, and hygiene ical departure from the past, when public sector education; (2) implement community-based engineering departments had the responsibility schemes in some of the rural areas and small for working with communities. The project towns of Badulla, Matara, and Ratnapura dis- review process described earlier, which enables Translating Lessons into Design Features 69 any group with credentials to qualify as an IA, holding company, was established in 1960 to introduces competition in service delivery. At the manage the water supply for Abidjan. In 1978 the same time, large, more experienced agencies are Ivorian government granted a lease contract to tapped to train, support, and build the capacity SODECI to operate and maintain all urban and of the smaller IAs that work directly with the rural water supply outside Abidjan. Con- communities. struction performance was impressive. A total of Similarly, in Nepal a rural water and sanita- 158 cities and towns and 90 villages were tion project now being prepared bypasses the equipped with piped systems. In more than 8,000 Ministry of Public Works. With support from the villages, 13,500 water points had been con- Ministry of Finance and the National Planning structed and equipped with handpumps. Commission, the project will create a "Fund" out- Financial performance for the urban areas was side the Ministry of Public Works. The Fund, also good: collection rates were at 98 percent, and which will be established by an act of parliament, unaccounted-for water was at a low level. will be managed by a board composed of repre- Yet, as striking as these figures seem, perfor- sentatives from the private and public sectors and mance in the rural areas was poor. A majority of NGOs. The Fund will develop a set of transpar- rural water points were broken down. A central- ent criteria for disbursing grants to any agency- ized maintenance system resulted in costly and whether from the community, the private sector, delayed repairs; communities were not orga- an NGO, or the Department of Public Works nized; and villager participation was found to be itself-to provide water supply services for the weak. In general, rural systems were overde- rural poor. The details of Fund operations are cur- signed and expensive. Since SODECI did not rently being worked out. assume any commercial risk for poor financial In Indonesia, the institutional framework for performance and since shortfalls were covered by the rural water supply and sanitation program sector funds, there was little incentive to keep for low-income communities is based on devolv- rural costs low or to monitor performance. ing power, authority, and responsibility to the The water sector was reorganized. The gov- provinces and, further, to the districts in six ernment renegotiated its contract with SODECI provinces. The institutions involved are the for more favorable terms, with SODECI assum- Ministry of Public Works, for physical develop- ing more risk for proper operation in the urban ment of water sources; the Ministry of Health, for areas. In March 1988 the strategy for the rural water quality and health education; and the areas changed, with local contractors and artisans Ministry of Home Affairs, for institutional devel- becoming involved. Responsibility for mainte- opment and community participation. Decision- nance was gradually turned over to villages, and making powers for selecting communities and local people were trained in maintenance and allocating resources are delegated to either the repairs. The latest survey shows that, despite district or the subdistrict, depending on the com- intensive technical training in repairs and distri- plexity of the technology. bution of spare parts through local automotive These projects involve NGOs and the private dealers, one-third of all pumps remain broken. sector in a major way. Sri Lanka and Nepal con- Difficulty in coordination, and lack of demand duct large publicity campaigns, disseminating and incentives at the village level to keep pumps information about the project and the Fund, to functioning, were two problems identified as encourage any type of agency to participate in a causes for this outcome. project. C6te d'Ivoire involves the private sector in management and implementation that goes Funds and Financial Flow beyond just contracting for civil works, by stressing the importance of the regulatory and Institutional issues cannot be fully separated incentive structure as well as the different from financial and fund channeling issues. The approaches needed for rural and urban sectors key in managing demand-responsive approaches (Triche 1990). is availability of funds at the community level, Societ6 de Distribution d'Eau de C6te d'Ivoire funds that can be disbursed when needed, (SODECI), a joint foreign and local, private-stock through simple mechanisms, with transparency 70 The Contribution of People's Participation: Evidencefrom 121 Rural Water Suipply Projects and accountability. In a review of funding mech- ing example of the importance of having incen- anisms across sectors, Alexandre Marc (1992) tives at the agency level for making the receiving found five features to be especially conducive to mechanism work. beneficiary participation: SENASA, which was set up as a decentralized * Flexibility in the use of funds is important unit of the Ministry of Health in 1971, initially since it is impossible to predict at the outset of had responsibility for a range of activities in a five-year project what will need to be water and environmental sanitation. In mid-1991 funded, when it will be funded, and how SENASA was restructured to streamline opera- much will be needed. tions and to clarify responsibilities, and three * Mechanisms for accessing and managing autonomous directorates were created: one for funds at the local level need to be simple since water and sanitation, one for environmental many local communities (and sometimes health, and the other for administration. NGOs and local governments) are not yet Although the general approach to community accomplished at management. participation and the mechanism for delivering * Mechanisms need to permit small disburse- finances to the community were developed in ear- ments of funds, given the limited capacity of lier projects, the third Paraguay project focuses on small communities to absorb large dis- correcting the high level of arrears in debt service bursements. on loans that SENASA makes to communities. * Mechanisms must be transparent to foster Discussions revealed that SENASA has had little trust. incentive to press the juntas to service their debts, * Funding mechanisms need to be sustainable since these repayments were required to be to support long-term capacity building at the deposited at the Central Bank of Paraguay in a local level. non-earning account, and the approval process Changing the way finances are managed and for withdrawing even small amounts of money how money is allocated in the sector is a funda- for project use was cumbersome and lengthy. mental feature of institutional reform. Both the As a result of negotiations with the govern- fund delivery mechanism (from donor to gov- ment, SENASA is now permitted to manage its ernment to intermediary) and the receiving own financial resources, including recuperation mechanisms (at the community levels, and at the of past loans. The agency was also allowed to agency level in recovering costs from the com- establish an interest-bearing account to operate a munity) need to be addressed. revolving fund-under regulations and proce- All projects that depart from direct provision dures promulgated by SENASA itself-to expand of services by the public sector employ massive water services to other rural communities. publicity campaigns at the outset to encourage None of the projects lends directly to NGOs or wide-based participation and to prevent distor- to the private sector. In Bangladesh, however, the tion of the fund's use. Transparency is the best Grameen Bank makes loans to the poor for guarantee against abuse. In Sri Lanka all financial investment in wells and latrines. The approach investment decisions will be made by the has been successful; consequently, the United CWSSPU and its field offices, not by the drainage Nations Capital Development Fund turned over board. Using implementation agencies and inter- its large tubewell construction program to the mediaries will cut down on the total number of Grameen Bank. individual requests, but when the project reaches Channeling funds to communities for respon- its peak, 400 requests will most likely be sible use is the key to stimulating large-scale com- processed and implemented each year. munity initiatives, without controlling the form The third Paraguay rural water supply and and extent of direct participation at the commu- sanitation project is the only one that specifically nity level. Fiscal decentralization and establish- targets sector financing as its primary policy ment of funds for direct community access may objective. The project states that its first objective be new phenomena for the rural water supply is to organize "a sector financing system based on sector, but the Bank has considerable relevant greater reliance on water users self-financing experience with social funds and municipal capability." This project also provides an interest- investment funds. What is required for the water Translating Lessons into Design Features 71 sector is that the ex-ante control mechanisms cur- Indonesia; the government will allocate a fixed rently used for disbursement of funds, as well as amount for capital costs and communities will Bank procurement and auditing procedures, be make up the difference if higher service levels simplified and adapted to give control and power are desired. to communities themselves. For the Bolivia Social Investment Fund, the Approaches to Planning: Master Plan Bank approved the use of a unit-price costing sys- or Learning Process? tem supporting direct contracting instead of open bidding. In Indonesia, where NGOs are not rec- The conventional master plan approach which ognized by the central government and therefore focuses on "optimizing" technology for priority are not allowed to compete in local bidding, the investment areas, is most common in the water Bank encouraged management firms to enter into sector. Extensive village-by-village surveys, joint-venture arrangements with NGOs to com- focusing on population statistics, water availabil- pete in local bidding (Bhatnagar and Williams ity, health, and socioeconomic and hydrogeolog- 1992). ical conditions, typically are the basis for the A related issue concerns the legal status of master plan. When the surveys are completed, community groups. Unless these groups have preengineering designs are made; long before legal status, it is often difficult to channel implementation, "blueprints" are developed for resources or transfer ownership of assets to them. the communities, including detailed engineering Many water projects have already addressed this designs with established costs. Contracts have problem. Projects that depend on community often already been awarded by this stage. The groups to be key actors-for example, the Nepal several volumes of data and engineering draw- Community Hill Forestry Project, the India ings that exist for any one district or province Integrated Watershed Development Plains encourage development of how-to manuals for Project, and the Nepal Bhairawa-Lumbini project execution well before any implementa- Groundwater Irrigation III Project-include spe- tion begins. cific programs to legalize groups. The master plan (or blueprint) approach calls for very little consultation with local people; Technology and Choice of Service Level there is no need to establish local partnerships. A master plan approach is therefore appropriate for As long as technology and service level are prede- centrally managed and controlled systems and termined and inflexible, community demand can- for large construction projects. It is inherently not be met, even if such demand has been assessed antagonistic to a demand-based approach, how- and aggregated. Studies done during preparation ever, which depends on responding to what com- of Bank projects show that all clients-even the munities want. poor-want high service levels, and they are will- The alternative to the master plan is an ing to pay substantial amounts for reliable service. approach that emphasizes learning by doing; the Although many projects prohibit or limit the num- learning-process approach presumes that not ber of house or yard connections and insist instead everything can be known prior to implementa- on providing minimal service levels to the largest tion. NGOs have long used this approach, but it number, a few World Bank projects do offer service has only recently been applied to Bank projects. beyond the minimum, provided that users pay the In essence, a learning-process approach is a additional costs. systematic way of learning by doing and of build- Thus, in Sri Lanka, rather than drawing up ing in flexibility in the planning process (Korten detailed engineering designs for each village in 1980). It is a way to manage uncertainty when the a project area, the intermediary institutions doer does not know what will and what will not spend time-after a community has expressed work in a particular context. It is a way to manage interest-explaining and working out the short- risk, and to minimize the risk of failure by not and long-term costs of different options; the being forced to define one correct way of doing community then chooses the technology and business before the evidence is in. The emphasis is service level. A similar approach is planned in on trying different institutional and technological 72 The Contributtion of People's Participation: Evidencefrom 121 Ruiral Water Supply Projects options that appear feasible in a particular context. were being developed throughout the prepara- Because no two communities are the same, stan- tion process. Sixty-two villages (or starter areas) dardization of details of implementation strategies were selected in the six provinces to start imple- is not only irrelevant but counterproductive and mentation. Intensive and participatory data gath- wasteful of resources. ering was limited to these starter areas. Data were The process does not assume that there is one gathered by facilitators from the kecamatan and right answer; it seeks to identify, through moni- kabupaten levels through village self-assessment. toring and evaluation, the key principles that can Findings regarding willingness to pay and other guide implementation and inform policy. The data were used to develop the village action plan most important learning takes place at the local (VAP), which is a detailed document completed level; central to the concept, therefore, is involve- at the end of the village consultation process; the ment of all relevant stakeholders in the process. village and its water committee are signatories to That involvement is critical for building local the VAP. capacity, promoting local ownership, and induc- Because the government wanted to use a ing policy change. learning approach to make the project design When it is assumed that all the answers are responsive to community demand, and because not known but multimillion-dollar projects are the project was decentralized, the preappraisal still being planned, it is important that learning process could be participatory In the capital of be effective and efficient: hence, the special one of the provinces, the process included project importance of internal and external monitoring preparation teams from the provinces and senior and evaluation procedures to guide project deci- officials from the center and the provinces. The sionmaking. The acceptance of the learning first two days were spent developing a common approach to planning has profound implications understanding of a learning-process approach for the project-preparation process and for imple- through a series of hands-on activities involving mentation strategies. Success in this approach is sixty participants. The workshop was managed conditional on clarity of objectives and indicators by the UNDP-World Bank Water and Sanitation of success, openness to learning, experimentation Program. with different strategies, ceding of agency con- The workshop started with officials drawing trol, and tolerance of ambiguity. their vision of community management. Other The current water project in Indonesia is activities included card-sorting exercises that pri- based on a "structured-learning" approach to oritized the behavioral attributes needed at the community management. Even preparation for community and agency levels to make commu- the project differed from that for conventional nity management work and that identified the rural water supply projects. Since community critical decisions in the community-management management of rural water supply through gov- process and the appropriate government levels at ernment had never been successfully achieved which decisions should be made. thus, the work- in Indonesia, the World Bank team, the staff of shop clarified key concepts through an intensive the UNDP-World Bank water and sanitation participatory process and set the stage for evalu- program in Jakarta, and the Indonesian govern- ating provincial proposals based on common cri- ment became partners in collective problem teria identified by the participants. solving. Project preparation for the six provinces A limited number of engineering designs was decentralized to the provinces, assisted by were developed to establish cost parameters; pro- teams of government and local consultants from curement procedures were simplified to involve NGOs (CARE staff became consultants to the local contractors. Much time was spent in clarify- government). ing and simplifying the flow of funds to ensure As objectives, principles, and strategies that expenditures would be responsive to emerged, it became obvious to all that what was demand and that decisions for simple systems being proposed was a radical departure from the would be made as close as possible to the com- past and that extensive data gathering for the munity level. The project includes a detailed entire project area would not be useful. What was training strategy to orient and train staff to sup- needed instead was to try out the strategies that port community-management strategies; 20 per- Translating Lessons into Design Featuires 73 cent of the budget is earmarked for capacity effective, thus, minimizing risks during large- building at the agency and community levels. scale implementation. Given the importance of monitoring and An increasing number of Bank projects there- evaluation (M&E) for internal and external plan- fore begin pilot activities during project prepara- ning, one of the first project workshops will focus tion. This has been done in Indonesia, Karnataka on developing the M&E framework, that is, the (India), and Nepal. The Nepal rural water sup- indicators and processes to be monitored at dif- ply project (with involvement from the UNDP- ferent levels. The capacity for monitoring World Bank program) includes a pilot project process and outcome will be strengthened at the financed by the Japanese Grant Fund. The $1.5 kabupaten level. Thus, rather than extensive data million pilot phase will be implemented for gathering prior to implementation, data gather- twenty months before the effective date of the ing throughout the project cycle will become key loan. It will test different service delivery to adjusting and refining plans during imple- options, document results, and test the working mentation. Planning will be undertaken for one- of the Fund approach proposed for the main pro- year periods, with resource allocation for the ject (see box 6.2). following year dependent on the performance of the previous year. This "rolling conditionality" Monitoring and Evaluation ensures that the data gathering is put to use in decisionmaking. The budget is indicative only, Of course, managers worry about potential loss with the expectation that after the first couple of of control and accountability when decisions are years of implementation, better estimation will made by hundreds of communities rather than by be possible. a central authority. Nevertheless, flexibility and One constraint to the World Bank's involve- accountability can coexist, if there is also a moni- ment in the rural water supply sector is that toring and evaluation system that can simply and there are few large-scale successes in the coun- effectively indicate what is going on at different tries where assistance is needed. It is also clear levels. that successes depend on radical institutional In both the Sri Lankan and Indonesian rural reform, coupled with intensive community- water supply projects, monitoring and evalua- level outreach activities. hence, success depends tion activities are important. Both projects seek to on new learning. The interval between a pro- ensure that project agencies do not dominate the ject's identification and the effective data of a community decisionmaking process, and both credit or loan is often long, and that time could emphasize community self-assessment activities. be used to implement the new strategies and Selected information gathered at the community options proposed by the project. That experi- level flows upward through the system to keep ence, in turn, could help refine the principles and the process simple, yet still ensure responsive- strategies of the project once the loan became ness to communities. 74 The Contribution of People's Participation: Evidencefrom 121 Rutral Water Sutpply Projects Box 6.2. Nepal: a new funding mechanism In 1992 preparations began for the proposed US$18.1 ml- ronmental soundness). SO completes a pre feasibility lion national Rural Water Supply and Sanitation (RWSS) study of a sub-project which forms the basis of a contract Program in Nepal. The program is designed to include an with the fund for the development phase. autonomous RWSS-FUND, which will support demand- 2. Development phase. The water users' committee, led, community-based water and sanitation initiatives. In with SO assistance, prepares a feasibility study of its own March 1993, as part of project preparation, an innovative water supply and sanitation system. This feasibility twenty-month field-testing program was initiated to test study forms the basis for a contractual agreement among and refine proposed strategies, including the funding the Fund, the beneficiary community, and the SO for the mechanism. The pilot program is funded by a Bank- next two phases (for example, implementation and post- executed Japanese grant (approximately US$1.5 million) implementation). and managed by the UNDP/World Bank Water and 3. Implementation phase. SO provides hygiene and Sanitation program (TWUWU). sanitation education, and trains the water users' com- The RWSS-FUND will be managed by a board with mittee and village maintenance workers. The beneficia- representatives from both the govemment and the pri- ries construct the sub-project with support from the SO vate sector (NGOs). The board will be autonomous, and as needed. fully responsible for the Fund's management. Funding 4. Post-implementation phase. The continuation of will come from the Ministry of Finance through a sim- hygiene and sanitation education, and latrine promotion plified procedure consisting of the release of block grants by the SO: with operation and maintenance of the water once the Fund budgets have been approved by the Fund system by the beneficiaries. Board and Parliament. Preliminary findings indicate that district-based The Fund mechanism, which is being tested in the NGOs, localized NGOs, and CBOs are effective service pilot program, is largely similar to the final mechanism. providers, especially in many aspects of software. Large For example, the pilot Fund's advisory committee is sim- SOs have proven extremely useful as well, especially in ilar in composition to the proposed Board of the RWSS- their research and technical capabilities. Effective con- FUND. However, the pilot Fund is managed by project sortia agreements are being forged between the SOs. staff and funded through a grant. So far, the mechanism With the support of SOs, most communities have appears to be working well. successfully formed water users' committees, the major- Fourteen implementing agencies (support organiza- ity of which are representative in terms of gender and tions /SOs) from the private sector (mostly NGOs) have ethnicity. Water users' committees in almost all areas qualified for funds, and 84 communities representing have made advances in planning and construction. about 31,000 beneficiaries are participating. In addition, Although the development phase process (phase 2 a pipeline of sub-projects is being identified for imple- above) is sometimes seen as time-consuming, some SOs mentation when the main project is underway. The four have reported that the process results in much stronger phases of the sub-project cycle and the responsibilities of water users' committees, as well as fewer post-construc- each group of participants are outlined below: tion problems. 1. Pre-development phase. Selection of SO and sub-pro- ject according to a set of transparent eligibility criteria (felt need, sustainability, technical, economic, and envi- Soutrce: Legrain (1994); Pfohl (1993); World Bank. CHAPTER 7 Conclusions and Recommendations The lessons from this study have important appropriateness of technology are important, but implications for policy, project design, and the study found that without beneficiary partici- implementation strategies for rural water- pation (and hence beneficiary responsibility), supply programs. The overarching one is that water systems are unlikely to be sustainable even rural water projects must be fundamentally rede- when spare parts and repair technicians are avail- signed in order to reach the one billion rural poor able. Beneficiary participation is also a significant who lack a sustainable water supply Redesign contributor to other performance outcomes, must encompass a shift from supply-driven including the percentage of water systems in planning to demand-responsive, participatory good conditions, overall economic benefits, per- approaches to ensure beneficiary participation, centage of the population covered, equality of control, and ownership. Changes must be made at access, and environmental effects. the sectoral policy and community levels; ignor- Among the proximate determinants of out- ing either level will create problems for both. comes, beneficiary participation was found to be The conclusions and recommendations of the a significant determinant of both overall quality study are organized around three questions: What of implementation (instead of quality of overall do we know about participation? If we know so project design) and quality of operation and much, why hasn't participation occurred on a maintenance (instead of quality of construction). large scale? What can we do differently? The quality of design was significantly negatively affected by complexity of the design, as measured What Do We Know about Participation? in part by the number of different activities and organizations to be coordinated. Beneficiary participation-including participa- All relationships between proximate deter- tion by women-in the rural water subsector is minants and beneficiary participation were essential for project effectiveness as well as for weaker when measured separately for participa- local capacity and empowerment of people for tion at a particular stage of the project. Thus to sustainability. Multivariate analysis established maximize its benefits, participation must be that even after controlling for the effects of eigh- treated as a continuous process. Beneficiary par- teen other determinants, beneficiary participa- ticipation is important throughout the life of a tion is the single most important factor project; it is a cyclical, iterative process that can- contributing to project effectiveness. Availability not be broken into elements. In other words, par- of spare parts and repair technicians and the ticipation cannot be effective when it is limited 75 76 The Contributtion of People's Participation: Evidencefrom 121 Rtral Water Supply Projects to the later stages of a project, as it is in the formal, indirect forms through various repre- "handing over" syndrome. sentational strategies. As goals are met, mecha- Study findings prove that participation is not nisms set up to manage intense levels of a nebulous concept; it can be defined, measured, participation often are dissolved or adapted to and observed. Participation is dynamic, and its new activities. nature changes over a project cycle; mechanisms and forms of participation vary widely depend- Why Hasn't Participation Occurred ing on project objectives. on a Large Scale? The extent of beneficiary participation achieved is determined by characteristics of both Notions of people's participation have been the beneficiaries and the agencies. The two key around for a long time in one form or another. beneficiary characteristics are demand, or com- What has gone largely unrecognized is that mitment made before project implementation, participation is a process with important impli- and the degree to which beneficiaries are orga- cations for policies, organizations, budgets, nized for their role. The three key agency char- resources, and staff skills. In other words, partic- acteristics are the extent to which local ipation affects institutions, and the main reason knowledge is incorporated in a project, the estab- for lack of participation on a large scale has been lishment of participation as an agency goal that reluctance to frame the issue in institutional is monitored and evaluated, and the relative rather than technological terms. autonomy of the agency to manage its affairs. Viewing rural water supply primarily as a Participation is most likely to occur when both technology issue reduces participation to simply beneficiaries and agencies perceive the net bene- one more add-on task which gets done as long as fits to be high, as they do when beneficiary it is convenient, free, controllable, and available demand is high and agency commitment to the when commandeered. Such a narrow conception goal of participation is firm. of participation as a microlevel activity with no Markers on the road to realizing participation linkages to policy and institutional reform are user financial investment in the project, the hampers large-scale beneficiary participation. degree to which local groups achieve control and Moreover, several pervasive, damaging myths ownership, responsiveness of the agency to about the water sector bear directly on the ability clients, and the extent to which users listen to to institutionalize participation. field agents or extension workers. In turn, local ownership and control are more likely to be Myth: The poor are unwilling to pay and they can- achieved when beneficiaries are organized, not afford to pay; therefore, water should be when the organization is based on local tradi- providedfor them. tional collectives, when clients are skilled and Fact: The poor are already paying, often more than have leadership qualities, and when agencies the rich; the poor will pay if they get reliable themselves have autonomy to allow local control services. of subprojects. Achieving participation can be costly, consum- World Bank-sponsored studies demonstrate ing perhaps as much as 15-20 percent of the pro- that poor families sometimes pay ten to one hun- ject budget, and it is not necessarily simple. The dred times more than the rich for water, since rich, the elite, and the formal leaders can be effec- they have to buy it from vendors. The poor also tively coopted, however, to support poor people's travel the farthest to get water, often from pol- participation in decisionmaking and to bring addi- luted sources that adversely affect health. tional resources to their community. And although Experience in projects around the world estab- neither communities nor subgroups within com- lishes that when communities are in charge, they munities, such as women, are homogeneous, com- apply culturally appropriate ways of instilling monality of interest can bring people together to responsibility and ownership without excluding solve problems when the need is great. the poor. These strategies range from setting fees Over time there is movement from intense, on a sliding scale to having the poor and destitute informal, direct forms of participation to more contribute their labor to a project. Conclutsions and Recommendations 77 Myth: Poor people cannot solve problems or manage longer hours and have less free time; they have technology; they do not know what is goodfor less income; they are more isolated; they receive them. less information; they have poorer nutrition and Fact: People in poverty are creative. Every society health; they have less education; and they are employs elaborate organizational systems and often more illiterate than men. They are rarely rulesfor managing natural resources. community leaders, and they do not participate in community decisionmaking bodies. Although Experience demonstrates that where govern- women are the primary carriers of water, they ment largesse and population pressure have not have limited power, access, and control over destroyed traditional water-management systems, resources. In this context, how can women possi- these function effectively and maintain a balance bly be reached without special support and between human needs and environmental protec- investment in their development? tion. People in poverty can adapt traditional insti- The study found that of the twenty most par- tutional forms to meet new challenges; they have ticipatory projects, about half successfully done just that in many parts of the world when reached women. The study also found that the they have had assistance in doing so. Innumerable factors determining women's participation were examples of local organizations finding solutions different from those affecting overall beneficiary in the face of ecological, financial, and technology participation. Unless programs make women's constraints attest to human ingenuity. empowerment a specific goal, women's partici- pation will not move beyond tokenism. Myth: To achieve equity, the service level provided must be minimal so that limited resources Myth: Public sector agencies and engineering can be spread as far as possible. departments must be the service providers, Fact: If people are not given what they want, they since the main task is construction and the will not organize to undertake collective main indicator of success is construction action or pay requestedfees. completed. Fact: In rural areas where there arefew economies of Participation cannot correct for fundamental scale, engineering departments are successful inability to match what people want with what is when they monitor and provide technical sup- offered. While it may be important to spread lim- port to others (NGOs, the private sector, and ited resources to a large segment of the popula- other nonengineering government depart- tion, strategies have to be developed so that ments). The key task is creation of local man- subsidies do not detract from the community self- agement and institutional capacity for selection process. The service level should be sustainability after construction. based on the willingness of a community or household to match or add to partial subsidies to The technology issues in rural water-supply achieve the desired service level. projects are well known. Further technological Studies from India and Thailand prove that innovations may be important, but the key chal- people are willing to pay for a high-service level lenge is to create local ownership, organization, (namely, hard-piped water connections), although and responsibility for the management of assets. they are unwilling to pay lesser amounts for com- Typically, public engineering departments have munal facilities. neither the capacity nor the culture or tempera- ment to empower the poor. They should do Myth: If beneficiary participation is made a goal, what they can most efficiently do, which is to women will be reached. provide technical assistance and monitoring as Fact: In most cultures, unless women are specifi- needed. cally targeted and strategies are developedfor Of the twenty most effective projects in the their empowerment, they will not be reached. study, over 90 percent were executed by NGOs, parastatals, and government ministries other In most rural societies poor women are more than the public works or engineering depart- disadvantaged than poor men. Women work ments. The renowned Malawi self-help project 78 The Contributtion of People's Participation: Evidencefrom 121 Rutral Water Suipply Projects began in the country's community develop- of institutional options must be tried, monitored, ment department in the late 1960s; by the time and refined over time to match community it was taken over by the public works ministry, demand and local management capacity. But risk it had absorbed a strong community- can be reduced by offering options and by plan- development orientation. ning for shorter periods of perhaps a year, with later plans developed after evaluating the experi- Myth: Master plans based on extensive data gather- ence of the previous year. ing before implementation are needed so that a uniform approach can be implemented. Myth: User decisionmaking is important but should Fact: Blueprint, master plan documents stifle the be limited to 7vell-defined parameters; control growth and evolution of participatory pro- should remain with project managers. grams; data collection need not be extensive Fact: The whole participatonj process is about giving before implementation, but it does need to con- people a voice and a choice. Participation can- tinue throughout implementation. Too-early not be turned on and off by outsiders; partici- standardization of implementation procedures patory processes mean giving control to leads tofailure. communities. Extensive planning before implementation There is a tendency to underestimate what shortcircuits community decisionmaking. The people in rural communities can achieve, temptation is to manipulate community choices although most communities manage communal so that they fit what has already been decided, yet and private resources effectively. Yet giving peo- success depends on communities taking initiative ple choices does not mean that they must do and choosing what they want. Extensive, village- everything themselves and become experts in all by-village data gathering is unnecessary; rather, technical issues. Explaining new technology in data collection before implementation should aim terms of short- and long-term costs, and laying simply to set the parameters and develop the out the ultimate financial and organizational processes and strategies for managing a learning implications of the new technology, helps com- process. Detailed data collection is relevant only munities make informed choices. during the implementation process, as comm- Participation is a long-term process that can- unities, together with community facilitators, not be effective if it is restricted to certain stages of develop village inventories for analyzing their a project (for example, handing over water sys- particular situations. In the absence of blueprints, tems to communities after construction is com- two-way information flow becomes an important pleted). When communities choose and supervise tool for managing change within a short planning contractors who are required to satisfy them cycle. At the same time, the focus of project data before receiving payment, not only do costs collection before implementation shifts to institu- decrease but the shift in the power balance and tional analyses, which are accompanied by a accountability reinforces community empower- search for experience across sectors to guide the ment and ownership and responsibility for the flow of funds so that financial mechanisms can project's physical assets. support community self-determination. Managers who use the blueprint approach to Myth: Participatory approaches take a very long planning have difficulty tolerating the high lev- time and can only be done on a small scale. els of uncertainty and ambiguity inherent in Fact: When projects respond to demand, action is participatory projects. The blueprint design rapid and the community organizational approach provides a sense of security by creating process occurs quickly. the illusion that all is known and under control. Experts who manage large sums of money When agencies assume that the need for a pro- believe they can reduce risk by knowing all the ject is great in the absence of a community expres- answers. They are therefore reluctant to adopt a sion of need, eliciting participation becomes a structured-learning approach, which presup- hit-or-miss affair. Things go well if a match poses that all cannot be known and that a range between demand and response is in fact Concluisions and Recommendations 79 achieved, but eliciting participation becomes It is often assumed that participation and problematic if the community has different pri- other project "software" are too messy to be mea- orities. In this situation, educational activities surable. This is a fallacy. Most planners will not have to precede community organizational activ- fully consider participation unless it can be mea- ities. Experience shows that community organi- sured and its benefits proved, at least to some zational activities generally take between two extent. Making participation a goal at the agency and six months, depending on the degree of self- level, combined with monitoring and evaluating reliance sought. In sum, the study demonstrates participation and linking it to successful perfor- that participatory approaches can be used on a mance, was an important determinant of benefi- large scale. ciary participation in the projects covered in this study. But unless using participatory strategies Myth: Beneficiary participation is difficult to repli- and studying their impact are considered as indi- cate on a large scale because it requires the cators of success or evaluation criteria, there is presence of charismatic leaders, NGOs, and ultimately little incentive for project staff to sup- other special, gifted people. port this relatively unpredictable, uncontrollable Fact: Beneficiary participation is replicable. task. Fortunately, the study also shows that sim- Charismatic leaders are usefulfor starting the ple measures and benchmarks can be developed, process, but broad-based leadership keeps the and that community people can come up with the process going. NGOs are often more success- most situation-specific and meaningful indica- fiil in adopting empowerment strategies, and tors of their participation. they can be effective intermediaries. Like other technical specialties, skills in designing and What Can We Do Differently? implementing participatory programs are learned. Obviously, the main challenge at the country level is to change policies, institutional structures, pro- The presence of charismatic leaders at the ject design features, and implementation strate- community level may be needed to start and gies to support beneficiary participation in energize collective action, but the study estab- demand-responsive service delivery. It is impor- lishes that broad-based leadership qualities are tant to involve all stakeholders in this process, and more important than charismatic leaders for sus- to ensure that the process is participatory, so that taining the organizational process. ownership of institutional reform and project NGOs played important roles in many of the design clearly lie with the country. Six areas are most effective projects in this study (see appen- particularly important at the national or subna- dix 2). About half of these projects were executed tional levels for managing the necessary changes. by NGOs, and NGOs were involved as partners These concern the clarity of objectives, the fund- with government agencies in other projects as ing mechanism, the selection criteria, the plan- well. Other executing agencies for the most effec- ning process, the intermediaries, and the tive projects included parastatal organizations community organization. and government departments of health and com- munity development. Clarity of Objectives Myth: Participation is a nebulous concept that is Objectives must go beyond general statements of difficult to define and measure. The goal of "provision of improved water and health." The human development through participation most important specific objectives to address are in decisionmaking is important but imprac- sustainability and effective use of the water system tical. and capacity building of local people, especially Fact: The concept of participation can be put into the marginal-the poor and women-through operation and measured simply. Measuring, participation in decisionmaking. If a project aims monitoring, and evaluating participation for institutional reform to achieve these objectives, makes agencies more accountable to supporting this should be stated at the outset so reform can human development through participation. drive the project; if the objective is to set up 80 The Contribution of People's Participation: Evidencefrom 121 Rural Water Suipply Projects decentralized, community-based systems, that raising all (or a certain percentage) of capital costs needs to be stated; if the objective is to achieve before construction begins. Communities must community participation and women's empower- have a stake in the continued functioning of the ment, the goal should be stated as such. Only then water system. can links among objectives, planning strategies, Three tools for assessing demand are particu- and monitoring and evaluation criteria be clarified larly useful: participatory needs assessment, bene- and mutually reinforcing from the start. ficiary assessment, and willingness-to-pay studies. The latter are especially important in gauging the Ftnding Mechanism interest in different service levels and in guiding the early planning for the overall project. The funding mechanism for projects has to be set up to quickly respond to demand, if projects are Planning Process to truly support local-level decisionmaking. The study finds, in fact, that the funding mechanism The planning process should focus on learning. is the most important factor affecting the In project design, this implies clarity and agree- demand-responsiveness of projects. Funds that ment on objectives, clear definition of indicators can be disbursed easily, in small amounts, and of success, and a framework for monitoring and with minimum bureaucracy and with account- evaluation to guide management decisionmak- ability, enable local-level decisionmaking on a ing at all levels to allow for flexibility. When this large scale without creating entirely new organi- learning process is applied to projects, budget- zational structures. This is one of the main rea- ing becomes an effort to estimate expenditures in sons that NGOs have been able to work with different categories, with the express recognition communities more successfully than have large that pinpointing costs for narrowly disaggre- government agencies. gated categories of expenditures against a time- Decentralized funding with transparent crite- line is counterproductive. Long lists of village ria for accessing funds is critical for effective local inventories are not needed for the planning government involvement; it is also necessary for process; instead, data should focus on institu- getting the private sector and NGOs involved in tional capacities, creation of appropriate fund- service delivery as intermediaries or as providers ing mechanisms, strategies to interact with of specialized services needed by the community. communities, and pilot activities to try out the Decentralized funds are essential to close the gap strategies suggested while the project is still between where the construction takes place and under preparation. where the decision is made to commit and release program funds. Intermediaries Except where fees are collected to provide ser- vices, community funds for operation and main- When programs have to respond to the needs of tenance should be left with the community to hundreds of scattered communities, large, central- manage. This increases local ownership, respon- ized agencies in the public sector are neither effec- sibility, and accountability, and minimizes trans- tive nor efficient in service delivery. Greater action costs. attention must be paid to institutional analyses of other agencies that can serve as intermediaries. If Selection Criteria institutional reform and institutional capacity building are envisaged, these must be made spe- The selection criteria should support the primacy cific objectives; the goal must go beyond providing of user decisionmaking. Since the policy environ- hardware and infrastructure and setting up special ment is not the same in all developing countries, units for preparing and monitoring projects. the focus needs to be on the principle of commu- The choice of engineering departments as the nity self-selection, which includes user invest- lead agencies should not be a foregone conclu- ment in capital costs. It is important to institute a sion. Instead, other agencies (including NGOs) process whereby communities establish their with an established community orientation and commitment by undertaking particular tasks and presence should be drawn into new partnerships. Conclusions and Recommendations 81 A multiplicity of executing agencies fosters com- that are expected to manage the water systems on petition and encourages centralized agencies in their own than for groups that merely report the public sector to give up control of the details breakdowns to higher levels. of implementation. Given the record of limited government resources, scattered communities, and poor Community Organization maintenance, most rural water programs must invest substantial resources in building the The effective use, operation, and maintenance of capacity of community-level institutions to work a water system at the community level depends autonomously or in partnership with other sup- on users, individually and collectively. The port agencies. This investment in community degree of organization necessary to keep the organization recognizes that technology is technology working and effectively used dep- embedded in a particular social fabric and ends on the functions that the community orga- requires the interest, talent, and skills of individ- nization is to perform. Thus, investment in uals working in coordination to manage strengthening the capacity of community water resources to keep the system functioning into a groups must be greater for water-user groups changing future. t i I i i I I Appendixes Appendix 1A. Participation with Other Direct Inputs into Water Project Effectiveness Difficulty Adequacy in holding Availability Clarity Dependent variable Adjusted R2 Participation offacilities staff of parts of objectives Overall effectiveness 0.85 0.28 0.14 -0.05 0.57 0.22 (5.3) (1.8) (-1.1) (9.6) (2.9) Percentage of water system 0.69 0.30 0.10 0.17 0.71 -0.02 in good condition (3.1) (0.7) (-2.0) (6.6) (-0.17) Economic value of benefits 0.67 0.27 0.23 -0.11 0.38 -0.8 (4.1) (2.5) (-2.0) (5.3) (0.12) Percentage of target 0.22 0.17 -0.05 -0.12 0.13 0.07 population reached (1.9) (-0.4) (-1.6) (1.4) (0.5) Environmental effects 0.11 0.23 0.01 0.03 0.11 -0.11 (2.8) (0.07) (0.4) (-0.9) (-0.9) Equality of access 0.14 0.26 -0.13 -0.05 0.08 0.22 (2.3) (-0.77) (-0.5) (0.5) (1.44) Community empowerment 0.74 0.59d 0.24a -0.04 0.33d -0.14 (8.2) (2.3) (-0.7) (4.1) (-1.4) Water-system task capacity building 0.68 0.70d 0.16 -0.02 0.22a -0.19 (8.3) (1.3) (-0.3) (2.4) (-1.7) Extent local organizations 0.72 1.01d -0.02 -0.03 0.10 -0.03 strengthened (10.0) (-0.16) (-0.4) (1.0) (-0.3) Net effect on leaders 0.18 0.26b 0.04 0.03 -0.02 -0.03 (2.9) (0.3) (0.4) (-0.2) (-0.2) Note: Significance levels are indicated thus: a = significant at 0.05; b = significant at 0.01; c = significant at 0.001; d = significant at 0.0001. Figures in parentheses are t-statistics. 83 84 The Contribution of People's Participation: Evidencefrom 121 Rutral Water Supply Projects Appendix 1B. Participation with Eighteen Direct and Indirect Inputs (Model 3) into Project Outcomes Availability Objectives Appropriate Outcome' Adjusted R2 Participation of parts specified technology Overall project effectiveness 0.86 0.24 0.44 0.19 (3.8) (5.6) (2.3) Objective value of benefits 0.72 0.26 0.26 -0.30 (3.6) (2.8) (-2.6) Percentage of water 0.65 0.29 0.69 system in good condition (2.4) (4.4) Environmental effects 0.13 0.23 0.11 (2.3) (0.9) Percentage of target 0.26 0.25 0.07 population reached (2.4) (0.5) Equality of access 0.40 0.17 -0.14 (1.5) (-1.0) a. Except for participation and availability of parts, values are reported if statistic is significantly beyond the 0.05 level. Note: For overall project effectiveness, the t-statistics were significant at the following levels: participation = significant 0.0004; availability of parts = significant 0.0001; appropriate technology significant 0.02; quality of management = significant 0.10. Appendix lB. Participation with Eighteen Direct and Indirect Inputs (Model 3) into Project Otutcomes 85 Agency Geologicall Political understanding Management Skills of staff environmental context context Social context 0.22 (1.64) 0.3 (1.9) -0.35 0.31 (-2.10) (1.19) -0.29 (-2.4) 0.38 0.4 0.61 0.54 (1.9) (2.4) (2.6) (2.7) 86 The Contribution of People's Participation: Evidencefrom 121 Rutral Water Stipply Projects Appendix 2. Projects Rated High in Overall Effectiveness Country Cost (duration (population Technology of project) reached) Institutionalframework (choice) Objectives Swaziland $8.3 million . Ministry of Natural Resources, Buried pipe from * Provide piped water and (1979-89) (52,000) Land Utilization, and Energy protected source to associated sanitation (MNRLUE) standpipes and taps improvements to rural * Ministry of Health (MOH) (Yes, as guided communities * Rural Water Supply Board by RWSB) * Alleviate water-borne (RWSB) of the MNRLUE diseases * Health Education Unit and the * Develop GOS institutions, Health Inspectorate of the MOH especially Public Health * USAID/Government of Swazi- Engineering land (GOS) Ethiopia SEK 8,744,300 . Arsi Rural Development Unit Spring source, * Supply clean water within (1982-86) (56,000) (regionally based govemment gravity-fed, 2.5 km agency) community taps * Release 8,000 women from * Revolutionary Ethiopia's (No information) burden of drawing water Women's Association (REWA) * Supply enough water to (national women's association) improve hygiene * Swedish International Develop- * Provide training ment Authority * Women's participation in all phases * Empower women Panama $27.7 million * Department of Environmental Primarily piped- * Improve health and quality (1972-81) (approximately Health (DEH) water systems, of life 200,000) * Ministry of Health (MOH) household/yard * Reduce rural-to-urban * Water Task Force connections, some migration * Community Health Committee handpumps * Provide safe water for * USAID (No information) domestic use * Lessen women's burden * Community participation and responsibility for O&M Ecuador $825,527 * Catholic Relief Services (CRS) Gravity-fed pipe * Produce change in living (1984-86) (20 sites with * Ecuadorian Institute for systems standards populations Sanitary Works (No information) * Provide safe drinking water of 75-2,500) * Secretariat of Integrated Rural at low cost Development * Have active participation * National Health Council of community responsible * Ministry of Public for O&M * Ecuadorian Episcopal * Produce spin-off activities Conference (religious NGO) focused on health education * Local water committees/juntas * USAID/CRS Appendix 2. Projects Rated High in Overall Effectiveness 87 Community participation Demand and its assessment and specific strategies Success indicators * Community seeks out RWSB aid * Utilizes community meetings * Provides access to water systems to * Community fulfills RWSB * Provides training approximately 9% of rural Swaziland requirements, forms committee; * Involves community in assessing needs * Community develops new projects establishes maintenance fund; * Community in charge of O&M * Effective O&M community must have resettled * Consults women in assessing water and * Women form majority of most water according to rural development sanitation needs, identifying water sources, and sanitation committees agency guidelines; provides labor; and in siting standpipes * Women hold key WSC positions, selects water minders; constructs manage the money in operations fund latrines * Women recognize time savings and * Community completes survey convenience in project area showing all above requirements have been met * Dodota women recommend * Participation an objective * Time savings for women used for drinking water project * Work through regional REWA income generation; childcare; meeting * Women's Association members all * Recruit women for training through attendance; social activity; home agree is water most important need elections at Women's Association gardens; domestic duties * Community pays user fees * Provide focused training to recruited * Women satisfied with training * Community contributes labor women * Women successfully complete salaried * Pay salaries to tap attendants and other jobs trained personnel * O&M good, including preventive * Tap attendants financially accountable for maintenance unconfirmed gaps between water use and * Women feel freer (for example, from total collected fees reliance on husband or from agriculture . No blueprint plan to follow-allows time as primary income source to plan future and opportunities for participation * Community sees women as capable of * Public recognition (ceremonies) for training doing new jobs * Trained/salaried women seen as models for others * Communities submit WSS request * Participation is an objective * Community satisfaction with water to national representative * Work through Community Health quality and quantity * Communities contribute cash and Committee * Effective O&M management by labor 20% of construction costs * DEH informs community of obligations; community * Create community health DEH and community sign contract; MOH * Evidence of community problem committee contacts local leader; public ceremony to solving (for example, raised funds to hand over system; DEH trains community buy new pump; capped new spring; for O&M, community is responsible for elected new managers, created O&M graduated payment systems) * Installation postponed if criteria not met * Time savings for women; used for * Involved local leaders income generation, childcare * DEH provides technical assistant after * Women empowered (management role project transfer on water committees; emerged as leaders) * No systems abandoned * Community requests system * Community participation key criterion in * Community trained in record keeping, * CRS evaluates request screening request applies for bank loans, builds new * CRS selects request; implements * Community elects junta facilities in-depth community survey; * Junta collects fees and is responsible * New homes being constructed approves request for O&M * Health improved * Community signs agreement * CRS gives funds directly to junta * Women in junta and are involved in * Community expresses willingness - Junta purchases inputs and labor and construction and O&M to participate and builds family employs maintenance operators . Water quality/quantity good sanitation facility (Table continues on thefollowing page.) 88 The Contributtion of People's Participation: Evidencefrom 121 Rutral Water Supply Projects Appendix 2. (continued) Country Cost (duration (population Technology of project) reached) Institutionalframework (choice) Objectives India Rs3l.42 lakhs * Rayalaseema Seva Samithi Borewells, hand- * Improve water and (1988-91) (handpumps) (RASS) (an NGO) pumps, piped- sanitation (construct Rsl40,000 water schemes, handpumps, pumped water (pumped water latrines schemes, latrines) schemes) (No) * Provide water to region Rsl4 lakhs without potable water 7-8 (Latrines) months of the year (137 villages * Empowerment and with participation population of 300-2,000) Kenya Approximately * National Water Conservation Piped-water, gravity * Improve health and hygiene (Phase 1: $6.5 million and Pipeline Corporation scheme and water * Improve dairy farming and 1980-83) (68,000) (NWCPC) (parastatal) treatment vegetable production (Phase 2: * Ministry of Water Development (Yes) * Reduce burden of carrying 1984-88) * DANIDA water Malawi $9 million * Government of Malawi Gravity-fed systems, * Improve basic living (1980-88) (465,000) Department of Lands, communal conditions and health of Valuation, and Water (DLVW) standpipes rural population * Ministry of Health (No information) * Reduce water-related * Village Health and Tap diseases Committees * Increase disposable time for * USAID rural women and children * Provide health education Togo $16.7 million * Ministry of Public Health, Social Boreholes, springs, * Provide safe drinking water (1980-87) (600,000) Affairs, and Women's Welfare cisterns, latrines to rural villages in plateau * Ministry of Civil Works, Mines, (No information) and savannah regions Postal Services, and Tele- * Establish village communications development committees * Ministry of Planning (VDCs) in 350 villages * National level coordinator * Provide training to field * Regional level coordinator agents, VDC members, and * Sector/Prefecture coordinator village volunteers in project * Village development activities committees * Prevent water-related * Multiple donors diseases * Increase community self- sufficiency, especially in maintenance; community development Appendix 2. Projects Rated High in Overall Effectiveness 89 Community participation Demand and its assessment and specific strategies Success indicators Region has no potable water * Participation is a specific objective * Achieved good portion of construction 7-8 months of the year * Capacity building through education, * Women trained as caretakers leadership training, forming committees * Panchayats in water-scheme villages * Utilizes existing committees passed resolution to undertake * Community contributes funds and labor operation and maintenance * Community assists in site selection * Community participation in providing * Utilizes local "social workers" funds, site selection, road repair, land as community representatives donation * Conducts meetings and 2-day "awareness generation camps" * Lobbying by Aguthi village water * Work with Aguthi water committee * Provides water to 68,000 committee . Cooperative relationship with chiefs and *Metering reduces wasting of water * Labor input subchiefs * Tariffs collected successfully . Dug trenches for own connections . Water access denied if no labor contributed (approximately 91% of water sales * Consumers pay monthly charge * 98% of households metered; consumers revenues) billed monthly by NWCPC * Households have access to two to three times more water * Time savings for women * Women more mobile and able to take part in other activities * Milk yields appear to have risen * Health appears to have improved * Communities requested assistance . Work with preexisting network of village * Postplanning participation high, from government committees (project committee, tap especially among women * New projects requested by district committee, health committee) * Communities felt sense of ownership development committees, the * Provide institutional strengthening to and pride in maintenance Ministry of Agriculture and committees * The longer taps were in use, the more Natural Resources, or DLVW . Use project staff experienced in community- they became the exclusive water source * Communities contribute to based development techniques for villages construction * Involve village leaders * Communities provided one-fifth of * Formal and informal staff training, training capital costs, two-thirds of O&M costs of health committee members * Time savings * Work through volunteer community leaders * High participation of women in health for self-help activities (for example, and tap committees (fair in repair teams construction) and low in project committees) * Involve women, especially on health * Water available at community taps 90% committees of the time; average repair time five * Community (tap committees) responsible days for maintenance; DLVW monitors and provides backup; project committee oversees tap committee * Government agencies chose sites . Form or strengthen VDCs * Clean water more accessible; based on degree of water need, size * Increase community organizational communities using water for cooking of population, accessibility of development through participation and drinking drilling equipment, groundwater * Involve women * Time savings for women characteristics of geologic . Use participatory training methodology; * Easier access to water led to women's formations provide training to field agents, community, and youths' income-generating activities * Community contributes to VDC members/officers, and maintenance * VDCs successfully mobilized maintenance fund, provides and ORT volunteers communities for construction manual labor, purchases cement * Develop training throughout life of project and maintenance * Maintenance program a * VDC responsible for installation, O&M, * Latrine use high prerequisite of installation organizing community for future * Repairs carried out by community development activities repairmen (Table continues on thefollowing page.) 90 The Contribuition of People's Participation: Evidencefrom 121 Rural Water Suipply Projects Appendix 2. (continued) Country Cost (duration (population Technology of project) reached) Institutionalframework (choice) Objectives Mali $4.5 million * Direction Nationale de Boreholes with * Drill 150 successful (1984-89) (approximately I'Hydraulique et de l'Energie handpumps boreholes in the Sikasso and 30,000) (DNHE) (No) Kadiolo areas, which have * Ministry of Hydraulics and chronic water supply deficits Energy * Provide workshop for * DANIDA boreholes * Train local personnel to eventually replace expatriates on the project Haiti $6 million * CARE Mainly spring-fed, * Improve quality of life (1984-1989) (160,000) * National Potable Water Service gravity pipeline * Construct or rehabilitate . Local Water User's Association systems, rural potable water systems (CAEP) standpipes, some * Strengthen community * USAID reservoirs institutions through (Choice possible, if establishing and training able to pay) water associations for O&M * Educate users in water related sanitation Yemen Arab $20 million * Rural Water Supply Department Pumped water from * Improve access to potable Republic (70,000- (RWSD) wells, cisterns, water for Yemen's rural (1979-89) 109,0000) * Local Cooperative Councils for springs to public population Development (LCCDs) and individual taps * Improve health and quality * New Transcentury Foundation (Yes) of life (NTF) * USAID Ecuador $2.5 million * CARE-Ecuador Spring-source, * Provide improved access to (1982-89) (approximately *Ecuadorian Institute of Sanitary gravity-fed water, sanitation, and health 62,000) Works (IEOS) systems; power education in communities in * Ministry of Health (MOH) pumped systems two relatively poor areas in * Provincial Councils of Azuay (rare); metered Andean highlands and Loja Provinces household * CARE/Canada connections; "pour- flush" latrines (No information) Appendix 2. Projects Rated High in Overall Effectiveness 91 Community participation Demand and its assessment and specific strategies Success indicators * DNHE conducts survey; chooses * Community contributes implementation * 91% of villages paid initial contribution village for installation; informs funds * 10% pump failure rate village of conditions * Village chooses two "repairmen" who * Cooperation with local authorities * Village agrees to meet conditions; inform project office of breakdowns excellent signs contract. Conditions: Village * Community contributes to repair costs * Maintenance systems functioning and pays 250,000 F. CFA; forms water * Local blacksmiths to be recruited for sustainable point committee; ensures maintenance in phase II; will be paid by * Low maintenance costs cleanliness of installations; builds community * Contributed to the overall improvement protective enclosure; pays cost of of water supply pump repairs * System utilization satisfactory * Project in Sikasso 100% completed * CARE notifies communities of * Communities elect CAEPs * Quality of construction good project availability * CAEPs organize local labor groups * Performance of systems satisfactory; * Interested communities send letters * Food for work incentive not always water quality high; water use increased of request successful (for example, delays in food * Approximately one-third of CAEP * CARE reviews requests and payments cause work stoppage or discredit officers are women; two-thirds of COQ coordinates with other local water CAEP) members are women; almost entire staff projects * CAEP responsible for O&M; collect fees; of user education department are women * CAEP formed through local establish bank accounts * Time savings for women used for election * Comite de Quartier (COQs) responsible for childcare and education, agriculture, * Communities required to provide maintenance of individual standpipes; cottage industries volunteer labor (1 meter of trench project design called for mostly women * Democratization increased through per person); additional work members electoral system for CAEP officers reimbursed with food for work * Provide training to CAEP, local plumbers * CAEP removed ineffective COQ * Communities contribute local * Women participate as community leaders, members materials; O&M fees; land or right- staff, and beneficiaries * O&M satisfactory, repairs made of-way (although role of Haitian national water association not clearly understood by those involved) * Community and LCCD make * Utilize existing local organizational system, * Completed systems 2 1/2 years ahead of project request to RWSD, sheikhs (that is, LCCDs, sheikhs) schedule lobby RWSD for projects * LCCD signs agreement with donors, NTF, * Community satisfied with system * RWSD assigns project to NTF RWSD * 180 persons from 90 villages received * Community contributes funds, * Local leaders (slheikhs) assist in surveys, O&M training roughly one-third of costs avoiding disputes in pipeline placement * No examples of nonparticipation * Four weeks classroom and on-the-job by villagers training for selected community members * Community pride in system in O&M * System caretakers treated with respect * Villagers contribute labor, local materials, * Monthly fees collected from each portion of cost household used for O&M, paying * Community responsible for distribution caretaker's monthly stipend system, O&M * Repairs, major and minor, handled by community; hire private mechanics for major repairs * Community interest and * Communities contribute labor and local * Systems function reliably willingness to contribute materials * Quantity good; quality questionable * Each household must build latrine . Communities establish water committees * Designs are appropriate before new water system handed to coordinate construction and O&M * 37% of communities had women over to community * Use mingas (traditional communal work- on water committees gangs) * Decrease in diarrhea reported in * CARE meets with community children representatives, IEOS, to coordinate project * Fees collected, systems maintained * Form consortium of water committees * Pour-flush toilets well constructed * Team of operators must include one woman and maintained * Provide training for system operators * CARE employs women as project staff * Water committees responsible for O&M; IEOS monitoring (Table continues on thefollowing page.) 92 The Contribtition of People's Participation: Evidencefrom 121 Ruiral Water Suipply Projects Appendix 2. (continued) Country Cost (duration (population Technology of project) reached) Institutionalframework (choice) Objectives Rwanda No information * CARE/Rwanda Capped springs, * Construct reliable water (1984-89) (No information) * Communal Centers for gravity systems, systems Continuing Education and communal * Develop functional, self- Development standpipes managing water user * Community level management (No information) associations for sustained organizations operation and management * Ministry of Public Works and of new water systems Energy * CARE Malawi No information * Ministry of Community Gravity-fed, * Bring abundant and clean (1968-78) (668,000) Development community taps water from mountains to * Christian Service Committee (No information) communities at base and in surrounding plains Peru $766,114 * Catholic Relief Services (CRS) Communal * To provide water and (1983-86) (17 com- * Community Development standpipes, sanitation services to munities) Institute (Instituto de Desarrollo individual communities affected by Communal) connections floods in the north (N) and * Institute for Rural Promotion (Yes) drought in the south (S) (Instituto Animacion Campesina) * Development Corporation of Piura (a public service corporation) India $1.04 million * Social Work and Research Step wells with * Improve the physical and (No information (approximately Center (SWRC) (an NGO) Persian wheels and social conditions of -1986) 140,000) * People's Action for handpumps Bicchiwada inhabitants, Development, subagency of the (No information) especially "scheduled" tribes Ministry of Rural Development and castes, by increasing * Swedish International agricultural production, Development Authority educating groups in health and hygiene, and providing safe water and improved sanitation facilities Appendix 2. Projects Rated High in Overall Effectiveness 93 Community participation Demand and its assessment and specific strategies Success indicators * Government-initiated project * Involve local political leaders throughout * Community expressed clear sense of * Individuals donate part of small project, including project planning, financial ownership and responsibility for water landholdings for construction of management, and support for water system; felt confident in ability to standpipe and yard committees sustain project * Community contributes labor * Community selects criteria for committee * Facilities well maintained members; elects management, water point, * Committees successful in collecting user committees most fees * Use local government meetings to * Community identified problems with disseminate water system information payment of services; replaced ineffectual * Community pays flat fee; some individuals committee members refused to pay because of distance from * Women served on all committees water source * Some women held top positions, but * Specific strategies to involve women, (for majority were fee collectors example, quotas for committee membership) * Time savings, but resulted in added work for women * Projects initiated by government . Project activities planned to ensure * Community satisfied with water quality officials, and village headmen have community participation and quantity significant role . Community must accept project * Water easier to access * Community must accept project . Community participates in planning, * Community using piped water for and be willing to participate construction, O&M drinking, cooking, bathing * Community elects committees . Community elects main project, branchline, * Community labor accounted for about * Community must provide proper and tap committees 25 to 30% of total construction and aprons and soakaways before water . Committees responsible for project maintenance costs connected from branchline to tap construction and O&M * Women's contribution recognized * Community provides labor for construction by community and maintenance * Community feels project ownership * Community need for water and * Community and project holder signs * Community successfully organized sanitation improvement contract work brigades * Communities assume responsibility * Organizational development for * Water supply increased; quality during design and construction, communities relatively high, less expensive and for O&M * Community participates in service level, site * Community organization strengthened selection, management system * Home gardens resulted from increased * Community elects water committee water supply * Community contributes labor, local * Women gained confidence, assumed materials, cash O&M responsibilities * Community organizes O&M committees (N) * Water-borne diseases decreased * Women participate in construction, elect and * Projects catalyst for new health and train health committees (N) development activities * Involves community leaders * Water wastage low * Community develops separate water-use * Repairs completed (S) monitoring committee (N) * SWRC assessed villagers' * Most labor provided by project staff and * 1,371 improved or new safe water development priorities and any community labor paid for supplies formulated project plan * SWRC trains village mechanics (men) in * Two years after construction, no wells pump maintenance had gone dry * Village committees form to handle O&M * During training, mechanics repaired * Mechanics responsible to local village government-installed pumps; all committees functioning at time of evaluation * Mechanics paid per repair or maintenance * Time savings for women visit; responsible village committee * Guinea-worm prevalence reduced confirms pump surveyed and maintained * Out-migration declined in target areas * Mechanics salary linked to job performance * SWRC initiates multisectoral development, (for example, income-generating activities, health education, tree nurseries, smokeless oven program) * Women have active role in health, income- generation, and smokeless oven programs (Table continues on thefollowing page.) 94 The Contribiution of People's Participation: Evidencefrom 121 Ruiral Water Suipply Projects Appendix 2. (continued) Country Cost (duration (population Technology of project) reached) Institutionalframeu7ork (choice) Objectives Indonesia $4.23 million * CARE/Indonesia Gravity-flow, piped- X Reduce incidence of water- (1978-91) (No information) * Community Water Committees water systems, related diseases. * CARE/USAID rainwater * Provide adequate and catchment tanks, reliable water system and and handpumps sanitation facilities (No information) * Develop mechanisms for sustainable O&M * Improve community organization and management before construction * Maximize community inputs * Encourage communities to use credit to pay for systems and encourage banks to provide credit Indonesia $4.2 million * CARE field offices Gravity systems, * Provide proper (1978-90) (130,000+) * CARE International umbrella hydraulic ram water/sanitation system to agreement with Ministry of pumps, sand relieve water-related Home Affairs filtration, diseases and reduce the * CARE/Indonesia formal communal amount of time women and partnership with provincial standpipes children spend gathering governments (No information) water * Community water committees * Reduce CARE's direct *Canadian International village-level operational Development Agency involvement and increase its catalytic role Appendix 2. Projects Rated High in Overall Effectiveness 95 Community participation Demand and its assessment and specific strategies Success indicators * CARE selects potential community; * Community participation is a stated * Well-designed systems conducts needs survey objective * Communities participated actively * If site meets selection criteria, * Provide community training in organization * Majority of communities successfully CARE initiates training and management, resource mobilization, operate and maintain systems (some for * Water scarcity, accessibility, and system design, construction, O&M, financial up to 10 years); strength of local quality motivates community to management, sanitation and hygiene leadership most important factor in participate education success * Community contributes cash, labor, * Work through community organizations and * Women involved, but few in local materials leaders decisionmaking * Active community participation in all levels, from planning to O&M * Use community problem-solving approach * Community water committees responsible for organizing and collecting community contributions * Communities make request to . Community participates in all phases * Project adaptable and strengthened provincial planning ministry * Project involves community leaders (formal community participation over life of the * CARE screens communities based and informal) project on selection criteria * Project meets with community, leaders, * Very few design problems; simple, * Community contributes cash, local community organizations reliable design materials, labor * Community forms water committee * Households pay for and construct responsible for planning, O&M, latrines; latrine use high management of health component * Women involved in water committees * Project provides training for community and (13% of membership) water and health committees in * Women actively involved in health management, health, O&M committees (95% of membership) * Local health committees deliver health * Women satisfied with systems messages * CARE utilized women employees * Incorporates community-initiated income- * lime savings for women used for help generation activities in agriculture and income-generating activities (sewing, embroidery, handicrafts) * Community motivated and capable regarding O&M; repairs made successfully * Community pays water fees, some 6 months to a year in advance * Public taps became village meeting places (Table continues on thefolloiving page.) 96 The Contributtion of People's Participation: Evidencefrom 121 Rural Water Supply Projects Appendix 2. (continued) Country Cost (duration (population Technology of project) reached) Institutional framework (choice) Objectives Peru $2.9 million * Departmental Development Gravity flow, * To improve health of as (1984-88) (44,745) Corporations in La Libertad and protected spring many people as possible in Cajamarca departments source, yard rural mountainous areas of * Department-level health units standpipes northern Peru of MOH (No) * Construct and rehabilitate * CARE Canada/Peru water systems, educate * Canadian International community members in Development Agency maintenance and health issues, and form and support village water committees -Supply adequate amount of water to villages * Establish effective maintenance and repair system in each community * Introduce use of latrines in 50% of homes * Introduce personal hygiene and housecleaning * Use of oral rehydration therapy by 50% of mothers with children under 5 years * 50% of children receive vaccinations by age 1 Thailand $4.8 million * Sanitary Engineering Division Water treatment . Help Thai government win (1966-72) (600,000 to (SED), Thai Ministry of Public plant with storage loyalty of rural population 1 million) Health tower and piped in NE * U.S.-based engineering firm distribution system . Help develop government * USAID Public taps capacity to plan and eventually changed administer a National to metered private Potable Water Program connections. * Improve health in 600 (No information) "security sensitive" communities through provision of potable water Appendix 2. Projects Rated High in Overall Effectiveness 97 Community participation Demand and its assessment and specific strategies Success indicators * Communities request water system * Community provides labor and local * Increased use of potable water * Communities donate land for water materials * Water quality and quantity improved system * Community responsible for O&M * Communities exhibit pride in water * Communities provide local * Most communities elect water committee to systems materials and labor manage O&M * 98% of surveyed villages have water * Evidence of community * Counterpart provides training in O&M to committees; 91% collect water fees commitment is specific selection many community members * Overall quality of O&M high; water criteria * CARE provides financial and institutional systems functioning in all evaluated * Community members who do not development support to counterparts villages, some systems for over 5 years contribute labor excluded from * CARE provides tools for each water * Comiimunities have successfully made water system committee repairs * District officials often initiate - Community chooses community member as * Most systems functioning effectively project; officials discuss proposed plant operator more than 10 years after construction project with village chiefs and * SED provides 5 weeks' on-the-job training to * Most systems financially self-sufficient villagers operators; provides refresher training * Time savings and increased water use * Amount of community financial * SED provides monthly supervision yield economic benefits (increased contribution or self-help significant * Plant and distribution system turned over to gardening, more craft activities, and factor in selection local government for O&M animal raising) Community should: * Community develops rate structure Successful systems had: * Have existing but not potable water * Operator responsible for system operation * Initial community contribution of time, supply and fee collection labor, funds * Be readily accessible by road * Training and support for local operators * Have high interest in obtaining * Gradual evolution of viable rate system and contribute to structures construction Problems: * Be willing to develop rate structure * Some users not drinking water for O&M and future expansion * To cover O&M costs communities changed to metered, private taps; poorer villagers lost access to water system 98 The Contribution of People's Participation: Evidence from 121 Rutral Water Sutpply Projects Appendix 3. Summary of Differences between Blueprint and Learning Process Approaches Issue Blueprint approach Learning process approach Purpose For large-scale construction projects For institutional reform and demand-based approaches Role of government Provide services Create policy framework to facilitate and initiate beneficiary and stakeholder involvement Role of people/users Peripheral; pay utility charges Central; take initiative, learn; problem solve; process facilitated by "external" agents, as needed Project documents Detailed, accurate masterplans Broad guidelines; detailed plans produced at essential for success; technology choice community level; no master plans but clear and service levels predetermined working goals, strategies; monitoring and evaluation criteria essential Role of managers Manage construction activities Manage unpredictability by creating a problem- solving environment Personal evaluation criteria Construction completed, adherence Primarily promotion of local reliance to schedules, unit costs Role of data Extensive physical, economic data base Limited data collection before implementation, before implementation by experts including cultural and social data; continued throughout by community people and project staff Role of evaluation Primarily a terminal entity; conducted Ongoing evaluation by community people and by external experts project staff Indicators of success Quality and quantity of construction; Effective use of facilities, sustainability, unit costs empowerment Interagency collaboration Consultations needed; collaboration Working collaboration essential for achieving during implementation not essential indicators of success Souirce: Adapted from Korten (1980). Appendix 4. Indicatorsfor Sociological Monitoring System 99 Appendix 4. Centre for Social Research: Indicators for Sociological Monitoring System- Karonga Lakeshore Integrated Rural Groundwater Supply Project Indicator Purpose of indicator Specifications Data needs Timeframe Community institutions: To ensure that Numbers, composition * AAC reports on new Ongoing proportion of village-level when established, elections Reviewed every communities with structures are in meetings held * SME 91/92 2 months VWHC, PC, and PAs in place and operating * MA monthly records place and functioning * CDA visits * VWHC/PC records * Extension worker reports Degree to which To ensure that * Gender * SME 91/92 Ongoing committees are communities will * Traditional authority * MA monthly records representative of all support the * Other power/influence * CDA visits sectors of the established struc- * Elected/coopted * Training records community tures; to indicate * Method and frequency of where training, communication mobilization, and elections are required Proportion of committees To identify training Responsibilities of each * Training records and Ongoing that understand their needs and need for linkages training received follow up responsibilities in institution building * SME 91/92 relation to VLOM * MA monthly records * CDA visits * VWHC/PC records Proportion of To identify needs for Ability to pay; type of * SME 91/92 Ongoing communities that have training, community arrangements: * MA records established a viable mobilization and * Collection mechanisms * CDA visits O&M revolving fund support * Accounting procedures * Training records * Defaulters * VWHC/PC records * Number of users per pump * Time lapse for repair Degree to which To identify needs for * Communication with * VWHC/PC records Ongoing; communities training, support communities * User interviews especially understand and accept and mobilization * Community * MA records crucial in year their role in VLOM understanding of VLOM * CA visits one * Acceptance of level of * Extension worker reports financial contribution * SME 91/92 Note: Abbreviations used in this appendix are as foUows: CDA = community development assistant; FHH = female-headed household; MA = maint- enance assistant; O&M = operations and maintenance; PA = pump attendant; PC = pump committee; SME = sociological monitoring exercise; VLOM = village-level operation and maintenance; VWHC = village water and health committee. Souirce: Gaynor and Jespersen (1992). ,-- Notes 1. The gender issue was also addressed in the analy- 8. For cross-tabulation, the cardinal scale of 1 to 7 sis, but because of the importance and complexity of the was converted into low: 1-3; medium: 4-5; and high: findings, that subject will be examined in depth in a sep- 6-7. arate paper. 9. The chi-square test of independence of the two 2. All dollar amounts are U.S. dollars. variables based on the two-way classification in figure 3. Pearson-product moment correlation measures 4.2 gives a value of 64, which easily rejects the null the strength of linear association between two variables hypothesis of independence, since the critical value of and ranges from -1 to +1. A zero correlation between the test at the 1 percent significance level is 13.3. two variables means that they have no linear relation- 10. Although the coefficient from a linear regression ship. A higher, positive number means that one value is is not directly comparable to a familiar correlation coef- likely to have a high value when the other one does, ficient (rho), the results of a bivariate regression give because the two are positively associated. Association nearly the same results as simple correlations when the does not, of course, imply causation. dependent and independent variables have nearly the 4. Multivariate regression is a statistical method of same variance. In table 4.1 the regression coefficients are determining the percentage of variance in the depen- reported so that the bivariate and multivariate results dent variable that is explained by a combination of inde- are comparable. pendent variables, or by any one variable, after control- 11. The value of a t-statistic reveals the likelihood that, ling for the effects of other independent variables. given the precision of the estimate, an estimate of the value 5. The variable reduction mostly involved eliminating observed could have resulted by pure chance even if the those variables with few observations or unreliable esti- true value of the coefficient being estimated were zero. The mates, or those that were collinear. As seen in chapter 4, results of hypothesis testing are generally expressed in two inclusion of variables was, if anything, overly generous. ways. A significance level (most commonly 5 percent or 1 6. To reduce the number of performance variables to percent) and the computed value of the t-statistic can be be studied, a factor analysis was run for twenty perfor- compared to the "critical value" for that significance level; mance outcomes. Overall project effectiveness emerged for instance, for a large number of observations the critical as the principal factor, accounted for the greatest vari- value for the 5 (or 1) percent level of the t-test is 1.96 (2.57). ance (73 percent), and had the highest factor loading The other use of a t-statistic is to calculate the significance (0.98) of the twenty variables. The relative importance of level at which a given t-statistic would reject the hypothe- overall project effectiveness justified its use in this study sis of a zero coefficient (this is referred to as a p-level). For as the main indicator of performance. instance t-statistic values of 2, 3, or 5 are said by the first 7. The many measures of participation form a clus- method to "reject the null hypothesis at the 5 percent ter, as supported by the high correlations among the level"; the p-level, or the significance level at which the measures (0.7-0.85) as well as by a factor analysis per- three statistics (for 121 observations) could reject the formed on nine of the participation variables with fewer hypothesis, reveals very different levels of rejection: 0.047, than ten missing data. Overall beneficiary participation 0.003, and 0.000002. emerged as the principle factor, which explained 72 per- 12. A simple exposition in equations may be helpful. cent of the variance. It had a factor loading of 0.97. The basic regression is between an outcome, or perfor- 101 102 The Contributtion of People's Participation: Evidencefrom 121 Ruiral Water Suipply Projects mance, variable as the dependent variable (OPE), partic- 23. Since the average number of users was based on ipation (OBP), a set of direct participation determinants design criteria, large piped systems received high (call them Z), and a set of direct/indirect determinants scores. These may be multicommunity systems with (call them W). The regression is: pumping stations or gravity-fed systems. Just as com- plexity contributes negatively, a negative correlation OPE = a + * OBPi y * Z, + 6 * WI El. exists also between overall effectiveness and the largest water systems. The columns of table 4.1 report the partial correlation 24. The last finding, a significant negative relation- coefficient b on participation, with various sets of the Z ship between project complexity and quality of project and W variables. The major distinction between the Z design, is important. One of the main arguments justify- and W variables is that variables in W may also affect ing investment in improved water supply is health performance through participation, or participation may improvement. Studies done in the last decade establish affect these factors. that health impacts are maximized when water improve- 13. The basic result from multivariate econometrics ments are combined with sanitation facilities (toilets) is that the exclusion from a multivariate regression of a and health education. Hence most water projects try to variable that is both positively related to the outcome combine these three elements in integrated projects. (dependent) variable and an independent variable While some projects have been able to complete all three results in an upward bias on the estimated coefficient of activities successfully, the fact that in most countries dif- the included variable. Intuitively, the regression attribut- ferent government ministries have responsibility for es (falsely) to the included variable some of the outcome these activities makes successful implementation diffi- associated with the excluded variable. cult. Given the difficulty in changing the fundamental 14. The variables expressed as percentages are approach in the sector from a supply to a demand rescaled to 1 to 7 points so that the partial correlation approach, it is important to sequence activities carefully. coefficients are comparable across variables. Only after successful completion of one activity should 15. The correlation between project effectiveness and additional activities be undertaken. In most rural com- overall water-system sustainability was high at 0.88; munities, this would mean water first, sanitation later. accordingly, results for sustainability will not be report- 25. Based on her in-depth analysis of 480 hand- ed separately. pumps in Zimbabwe, Cleaver concluded that three-tier 16. The results of the other determinants are much maintenance was more myth than reality. She identified less strong than for project outcomes. Other vari- four prerequisites for effective community participation ables-namely, local social and cultural conduciveness, in the long-term maintenance of pumps in Zimbabwe: economic context, and understanding between agen- * A strong felt need for a protected water supply cies-were more important than the direct determi- * Knowledge that government will not provide signif- nants. icant maintenance support 17. The question was specifically related to partici- * A strong, well-motivated local leader pation; hence the correlation is very high and the vari- * The anticipation of some tangible reward. able is not independent. 26. The Swiss agency Helvetas undertook a twenty- 18. Some experimentation with using the coders' five-year retrospective analysis of its assistance to subjective assessments of the coding reliability on a pro- Cameroon, during which 114 piped-water systems had ject-by-project basis as weights revealed no significant been built. Findings revealed a very similar pattern. The differences in the results. study found a steady increase in total costs, steady 19. The basic result is that with random-measure- decline in village contributions, and a very poor mainte- ment error, a regression coefficient is attenuiated, that is, nance record, despite the presence of project-trained biased toward zero. In the simple bivariate case, the caretakers for 95 percent of the water systems. amount of the bias is the ratio of variance of the mea- Maintenance was supposed to be handled by care- surement error to the total observed variance of the takers chosen, employed, and paid by villagers. independent variable. However, out of 105 caretakers, only 34 were regularly 20. Of course, with hindsight it is clear that having compensated for their work. Fifty-eight received no one coder code the outputs without knowing the pur- compensation, and 19 received compensation rarely or pose of the data would have produced even stronger irregularly, sometimes in the form of a bottle of beer. protection from a halo effect. Not surprisingly, the lack of compensation and the lack 21. It should be noted that this result is also consis- of community interest in routine maintenance when tent with certain mixes of measurement error (with a water was still flowing made caretakers quickly lose downward bias) and halo effect (with an upward bias) interest in their work. The villagers, who did not feel that just offset each other. any responsibility for a system financed and maintained 22. The issue of establishing causation from data is by outside agents, had no interest in organizing to discussed more fully in Isham, Narayan, and Pritchett undertake maintenance. (1994), who use statistical estimation techniques appro- 27. Similarly, studies of nineteen small piped-water priate to identifying structural relationships. Those systems in Andhra Pradesh, India, illustrated the diffi- econometric results, which are beyond the scope of this culty in generating and collecting payment for public paper, fully support a causal relationship. standposts, but the willingness of communities to pay Notes 103 for higher service levels. Despite regulations against pri- ticipatory process, including the number of women who vate connections, the study found that 20 percent of the showed up, spoke up, or challenged what was said at people had connected illegally to the pipe system, meetings. Over one year, leadership of the water groups which severely hurt both revenue collection by the local shifted from village leaders to others who had not previ- councils and water flow at public standposts and at the ously held leadership positions. villages at the very end of the system. The same study 29. Greater autonomy is also, not surprisingly, postulated that if house connections were allowed, not strongly associated with greater local control (see table only would the revenue base of local councils be 5.3). sounder but the communities themselves would want to 30. A training exercise has been developed for use in take over the systems from the local councils, get workshops with task managers and sector and WID involved in management, and raise the revenue to (women in development) specialists in the World Bank. expand the distribution network of the pipe systems It applies a participation matrix to analyze the demand (PRED Study Team 1991; and Job and Shastry 1991). orientation of eight design features of a rural water sup- 28. In a community water project in Indonesia, a ply project supported by the Bank. The case study is project manager reported that, at the beginning, field hypothetical, but it is based on Bank experiences, partic- workers' diaries reflected no mention of women and ularly in Asia. For more detailed information on use of poor people. Field workers stayed with the chief and the material, see Workshop Facilitator Notes, "Participa- "hung out" with village officials. After being repeatedly tory Design of Demand-Driven Water Projects," Human queried and told that they were not doing their jobs, Resources and Social Development Division, Asia field workers gradually started spending more time Technical Department; and UNDP-World Bank Water with women and the poor and reported being less and Sanitation Division, Transportation, Water and embarrassed at doing so. From these early experiences Urban Development Department, World Bank, the project developed simple criteria to monitor the par- Washington, D.C., 1993. . Bibliography Altaf, Mir Anjum, Haroon Jamal, and Dale Whittington. Carroll, Thomas F. 1992. "Capacity Building for 1992. Willingness to Pay for Water in Rural Ptunjab, Participatory Organizations." Paper prepared for a Pakistan. UNDP-World Bank Water and Sanitation World Bank Workshop on Popular Participation, World Program, Washington, D.C. Bank, Washington, D.C. Bess, Mike. 1990. Mutnugi-Muigumango and Ngorika Piped CEFIGRE/CIEH. 1989. Drinking Water Projects in Western Water Schemes: Case Sttudies in Management Assistancefor Africa: Evaluiation Report. Report 1989-E6. Copenhagen: Suistainability. UNDP-World Bank Water and Sanitation DANIDA. Program, Regional Water and Sanitation Group Cernea, Michael M. 1993. "Culture and Organization: The (RWSG), East Africa, Nairobi, Kenya. Social Sustainability of Induced Development." In Bhatnagar, Bhuvan, and Aubrey C. Williams, eds. 1992. Sutstainable Development: The Ctultural Dimensions of Participatory Development and the World Bank: Potential Sustainability, ed. David Ip and Reg Little. MBC Directionsfor Change. World Bank Discussion Paper no. University Press, 1993. vol. 1, no. 2, 18-29; World Bank 183. Washington, D.C.: World Bank. Reprint Series no. 473, Washington, D.C. Boase, Robert, H. Subagyo Martodipuro, Sjahrul M. Christian Service Committee of the Churches in Malawi. Nasri, and Glenn Gibney. 1989. Evaluiation of the CARE 1981. An Evaluation Report of Christian Service Committee Sulawesi Rural Community Development Project. Fuinded Water Programmes, 1980. Blantyre, Malawi: CARE/Indonesia. Planning and Research Department. Boerma, Pauline. 1993. "Lessons Learnt from the CIDA/SIDA. 1993. Government of Uganda: The Soutth-west Implementation of the RUSAFIYA project in a Number Integrated Project. Report of a Project Monitoring of Selected Communities in Nigeria." Draft report. Mission, December 1992. Kampala. UNDP-World Bank Water and Sanitation Program, Cleaver, Frances. 1990. "Community Maintenance of World Bank, Washington, D.C. Handpumps." Harare: Department of Rural and Urban Bolt, Eveline, and Christine van Wijk-Sijbesma. 1992. Planning, University of Zimbabwe. "Women, Water, and Sanitation." Abstract Journal 2. Cohen, John M., and Norman T. Uphoff. 1977. Rutral The Hague, The Netherlands: IRC International Water Development Participation: Concepts and Measuires for and Sanitation Centre and PROWWESS/UNDP. Project Design, Implementation and Evaluation. Rural Briscoe, John, and David de Ferranti. 1988. Waterfor Rutral Development Monograph no. 2. Ithaca: Rural Communities: Helping People Help Themselves. Development Committee Center for International Washington, D.C.: World Bank. Studies, Cornell University. Bromley, Daniel W. 1986. "The Common Property . 1980. "Participation's Place in Rural Challenge." Proceedings of the Conference on Common Development: Seeking Clarity through Specificity." Property Resource Management, April 21-26, 1985. World Development 8: 213-25. Washington, D.C.: National Academy Press. Coreil, Jeannine, and Jean Beaudoin. 1989. An Evaluiation of Bromley, Daniel W., and Michael M. Cemea. 1989. The the Participatory Process in CAREiRwanda Byuimba South- Management of Common Property Natural Resources: Some east Water Systems Project. Water and Sanitation for Health Conceptual and Operational Fallacies. World Bank. (hereinafter WASH) Field Report no. 267. Washington, Discussion Paper no.57. Washington, D.C.: World Bank. D.C.: U.S. Agency for International Development. 105 106 The Contributtion of People's Participation: Evidencefrom 121 Rutral Water Supply Projects DANIDA. 1991. AgutthiRural WaterSuipplyProject, Kenya: Isham, Jon, Deepa Narayan, and Lant Pritchett. 1994. Evaluationi Report. Report by an evaluation mission vis- "Does Participation Improve Project Performance? iting Kenya, October 8-26, 1990. Evaluation no. Analysis of Subjective Measurement, 'Halo Effects,' and 1991/1. Causality." Background paper for World Development Donnelly-Roark, Paula. 1987. New Participatory Frame- Report 1994. Washington, D.C.: World Bank. worksfor the Design and Management of Suistainable Water Israel, Arturo. 1992. Issuiesfor Infrastructuire Management in Suipply and Sanitation Projects. WASH Technical Report the 1990s. World Bank Discussion Paper no. 171. no. 52 and PROWWESS Report no. 50. Washington, Washington, D.C.: World Bank. D.C.: U.S. Agency for International Development/ . 1987. Institutional Development: Incentives to UNDP-World Bank. Performance. Washington, D.C.: World Bank. Doyle, Brendan, A. 1990. "Community Participation Jackson, Justin R. 1988. "Community Self-Financing: A towards Greater Self-Reliance in Rural Water Supplies." Solution to Indonesia's Clean Water Needs." Paper pre- Ph.D. diss., Department of Social Science and pared for presentation at the International Resource Administration, London School of Economics, Association's Sixth World Congress, May 29-June 3, University of London. 1988, Ottawa, Canada. Dworkin, Daniel M., and Barbara Pillsbury. 1980. The Jensen, Erik S. 1987. Implementation of Water Master Plans Potable Water Project in Ruiral Thailand. A.I.D. Project for Iringa, Ruivuma and Mbeya Regions: Water Impact Evaluation Report no. 3. Washington, D.C.: U.S. Consuimption in Six Villages, 1987. United Republic of Agency for International Development. Tanzania: DANIDA. Esman, Milton J., and Norman T. Uphoff. 1984. Local Job, P. J., and D. S. Shastry. 1991. Village-Level Water Supply Organizations: Intermediaries in Rural Development. Management: Study on Nalgonda Mandal, Nalgonda Ithaca: Cornell University Press. District. Hyderabad, India: Panchayati Raj Engineering Espejo, Norah. 1989. Water Committees in Latin America: Department. Tasks and Training. The Hague, The Netherlands: IRC Jorstad, Mette. 1986. "The Bicchiwada Block Water Project International Water and Sanitation Centre. Dungarpur District, Rajasthan, India." Alternative Finsterbusch, Kurt, and Warren A. Van Wicklin III. 1987. Strategiesfor Involving Rural Women in the Water Decade. "The Contribution of Beneficiary Participation to New York: UNDP/PROWWESS. Development Project Effectiveness." Puiblic Admin- Judd, Mary. 1992. Personal communication and field istration and Development 7:1-23. notes. Regional Water Sanitation Group, World Bank, Gaynor, C., G. Banda, and C. B. Jespersen. 1992. Karonga Jakarta, Indonesia. Lakeshore Integrated Ruiral Groundwater Supply Project: Kammings, Evelien M. 1991. Economic Benefits from Village Level Operation and Maintenance. First Improved Rutral Water Suipply: A Review with a Focus on Sociological and Technical Monitoring Report. Centre Women. Occasional Paper 17. The Hague, The for Social Research, University of Malawi, Zomba, Netherlands: IRC International Water and Sanitation Malawi. Centre. Ghai, Dharam, and Cynthia Hewit de Alcantara. 1990. Khan, Akhter Hameed. 1992. Orangi Pilot Project "The Crisis of the 1980s in Sub-Saharan Africa, Latin Programs. Karachi, Pakistan: Orangi Pilot Project America and the Caribbean: Economic Impact, Social Research and Training Institute. Change and Political Implications." Development and Khan, M. Siddique. 1992. "Community Participation in Change 21:389-426. Social and Economic Development: An AJ&K Glennie, Colin. 1982. A Modelfor the Development of a Self- Example." Draft report. Regional Water Sanitation Help Water Sufpply Program. World Bank Technical Paper Group, World Bank, Pakistan. no. 2. Washington, D.C.: World Bank. Korten, David. 1980. "Community Organization and - . 1983. Village Water SLupply in the Decade: Lessons Rural Development: A Learning Process Approach." from Field Experience. Chichester, U.K.: John Wiley and Putblic Administration Review (September-October): Sons, Ltd. 480-511. HELVETAS. 1989. Evaluation of Water Suipply Systems, Korten, David C., and Felipe Alfonso, eds. 1983. Construtctedfrom 1964 to 1989. CD/HELVETAS Co-oper- Butreauicracy and the Poor: Closing the Gap. West Hartford, ation in Cameroon 1964-89. Zurich: HELVETAS. Conn.: Kumarian Press. HHC Team Consultants, Inc. 1989. An Evaluation of CARE Laredo, David, James Dawson, and Mouna Hashem. Canada's Water Projects in Four Latin American Countries: 1986. Small Ruiral Water Systems Project in the Yemen Arab Final Report. Prepared for CARE Canada, Ottawa, Repuiblic: A Midterm Evaluation. WASH Field Report no. Ontario. 197. Washington, D.C.: U.S. Agency for International Hino, Toshiko. 1993. "Community Participation in Development. 'Programme de restructuration de l'hydraulique villa- Larrea, Oscar R. 1988. Final Evaluiation of USAID/Catholic geoise' in C6te d'lvoire." Draft report. Africa 5 Relief Services Water and Sanitation Program in Ecuiador, Department, World Bank, Washington, D.C. Perui, Gutatemala, Honduras and the Dominican Republic. Institute for Development Anthropology. 1992. The WASH Field Report no. 201. Washington, D.C.: U.S. Central Tunisia Rutral Potable Water Institutions Project, Agency for International Development. 1986-91: Synthesis Report. Binghamton, N.Y.: Institute Marc, Alexandre. 1992. "Funding Mechanisms and for Development Anthropology. Participation: A Brief Review of World Bank Experience Bibliography 107 and Related Issues." Paper prepared for a World Bank Pfohl, Jake. 1993. "Progress Reports for the People's Water Workshop on Participatory Development, Washington, and Sanitation Program: Janta Ko Khane Pani ra Safai D.C. Karyakram UAKPAS)." August-November. UNDP- Mathew, Kochurani. 1990. "A Seminar of the Women by World Bank Water and Sanitation Program, Regional the Women." Jalasandesh 3 (1, 2): 9-10. Water Sanitation Group, World Bank, Kathmandu, McCommon, Carolyn, Dennis Warner, and David Nepal. Yohalem. 1990. Commutnity Management of Rural Water Picciotto, Robert. 1992. "Participatory Development: Supply and Sanitation Services. WASH Technical Report Myths and Dilemmas." WPS 930. DGO (Operations no. 67. UNDP/World Bank Water and Sanitation Evaluation Department), World Bank, Washington, Program, World Bank, Washington, D.C. D.C. McGowan, Rick, Rahardjo Soewandi, and Judi Aubel. Poluha, Eva, Goran Engstrand, Annika Idemalm, Johan 1991. Final Evaluation Report of CARE/International's Water Melchert, and Judith Narrowe. 1989. The Dodota Water and Sanitationfor a Healthier Environmental Setting Project. Suipply Project: Impact, Suistainability, and Replicability. Jakarta, Indonesia: CARE/International Indonesia and Project report. Stockholm: Swedish International U.S. Agency for International Development. Development Authority. Meehan, Robert, Anamaria Viveros-Long, Jose Manuel PRED Study Team. 1991. Village-level Water Suipply Hernandez, and Sonia de Heckadon. 1982. PANAMA: Management Study, Report 3: Pentapadut Mandal, West RLural Water. A.I.D. Project Impact Evaluation no. 32. Godavari District. Hyderabad, India: Panchayati Raj Washington, D.C.: U.S. Agency for International Engineering Department. Development. Roark, Philip, Judi Aubel, Kossi Hodin, Ikalie Kankarti, Mehra, Sudir. 1991. "Community Participation in the and Ayc Maria. 1988. Final Evaluiation of the USAID/Togo Water and Sanitation Schemes Implemented by the Ruiral Water Suipply and Sanitation Project. WASH Field Rayalaseema Seva Samithi (RASS), Tirupati." Draft Report no. 228. Washington, D.C.: U.S. Agency for report. Regional Water Sanitation Group, World Bank, International Development. New Delhi. Roark, Philip, and Jacqueline N. Smucker. 1987. Midterm Minnatullah, K. M. 1993. "Profile of a Village with Evaluation of the USAID/CARE Commutnity Water Community-based Rural Water System in AJK." Draft Systems Development Project in the Repuiblic of Haiti. report. Regional Water Sanitation Group, World Bank, WASH Field Report no. 205. Washington, D.C.: U.S. Pakistan. Agency for International Development. Narayan-Parker, Deepa. 1986. "Case Study Report: Roe, Emery, and Louise Fortmann. 1982. Season and Women, Water, and Sanitation: An Action-Study Strategy: The Changing Organization of the Rutral Water Programme by PKK, NTT, Indonesia." New York: Sector in Botswana. Ithaca: Rural Development PROWWESS/ UNDP. Committee, Center for International Studies, Cornell 1988. People, Puimps and Agencies: The South Coast University. Handppump Project. Technical Series on Women in Water Salmen, Lawrence F. 1987. Listen to the People: Participant- and Sanitation: Lessons, Strategies, Tools. New York: Observer Evalutation of Development Projects. New York: PROWWESS/UNDP. Oxford University Press. Narayan, Deepa. 1989. Indonesia: Evaluating Commutnity . 1992. "Beneficiary Assessment: An Approach Management. Technical Series on Women in Water and Described." Working Paper 1. Poverty and Social Policy Sanitation: Lessons, Strategies, Tools. New York: Division, World Bank, Washington, D.C. PROWWESS/UNDP. Sundararaman, Veena. 1986. The Social Feasibility Stuidy in - . 1993. Participatory Evalutation: Tools for Managing the Role of Women in Rutral Sanitation: Report of the Stutdy Change in Water and Sanitation. World Bank Technical in Four Villages in Maharashtra State. Bombay: Research Paper no. 27. Washington, D.C.: World Bank. Center for Women's Studies, SNDT Women's Oakerson, Ronald. 1986. "A Model for the Analysis of University. Common Property Problems." Proceedings of the Confer- Therkildsen, Ole. 1988. Watering White Elephants? Lessons ence on Common Property Resouirce Management, April from Donor-Fiunded Planning and Implementation of Rural 21-26,1985. Washington, D.C.: National Academy Press. Water Supplies in Tanzania. Publication 7. Uppsala, O'Brien, Dan. 1992. Internal memorandum, CARE, Inc. Sweden: Centre for Rural Development Research, Regional Water Sanitation Group, World Bank, Jakarta, Scandinavian Institute of African Studies. Indonesia. Triche, Thelma A. 1990. "Private Participation in the Ole Shani, Daniel, and Mitali Perkins. 1991. Participation Delivery of Guinea's Water Supply Services." WPS 477. in Development: Learningfrom the Maasai People's Program Water and Sanitation, Infrastructure, and Urban in Kenya. World Vision Staff Working Paper no. 12. Development Department, World Bank, Washington, Monrovia, Cal.: World Vision International. D.C. Paul, Samuel. 1987. Commuinity Participation in Development Tunyavanich, Nongluk, Prasit Leerapan, Thawatchai Projects: The World Bank Experience. World Bank Boonchote, Subarn Panvisavas. 1987. Research Report on Discussion Paper no. 6. Washington, D.C.: World Bank. the Evalutation of the Promotion and Suipport for Women's - . 1991. "Accountability in Public Services: Exit, Participation in a Village-Based Water and Sanitation Voice, and Capture." WPS 614. Country Economics Project. Bangkok: Faculty of Social Sciences and Department, World Bank, Washington, D.C. Humanities, Mahidol University 108 The Contribution of People's Participation: Evidencefrom 121 Rtural Water Supply Projects United Nations Centre for Human Settlements (Habitat). Warren, D. M. 1991. Using Indigenotus Knowledge in 1989. Malawi Gravity-Fed Rutral Piped Water Programme: Agricultutral Development. World Bank Discussion Paper A Case Stutdy. Nairobi: UNCHS. no. 127. Washington, D.C.: World Bank Uphoff, Norman T. 1986. Local Institutional Development: Water and SanitationforHealthStaff. 1990. Lessons Learned An Analytical Souircebook with Cases. West Hartford, from the WASH Project: Ten Years of Water and Sanitation Conn.: Kumarian Press for Cornell University. Experience in Developing Countries. Washington, D.C.: Uphoff, Norman T., John M. Cohen, and Arthur A. U.S. Agency for International Development. Goldsmith. 1979. Feasibility and Application of Rutral Water International, Inc. 1986. A Report on the Evalutation Development Participation: A State-of-the-Art Paper. of SANAA-CARE Community Program: Community Ithaca: Rural Development Committee, Center for Water Systems Construction and Watershed Conservation, International Studies, Cornell University. Hondutras. Ottawa: Water International, Inc.. VanSant, Jerry, James Sonnemann, and Rita Klees. 1991. Whittington, Dale, Donald T. Lauria, and Mu Xinmning. Impacts and Lessons Learned from the Rutral Water-Borne 1989. "Paying for Urban Services: A Study of Water Disease Control Project, Swaziland: A Ten-year Vending and Willingness to Pay for Water in Onitsha, Retrospective. Prepared under WASH Task no. 221. Nigeria." Infrastructure and Urban Development Washington, D.C.: U.S. Agency for International Department, World Bank, Washington, D.C. Development. World Bank. 1993a. "Community Management of Rural van Wijk-Sijbesma, Christine. 1985. Participation of Women Water Supply in Rwanda and Burundi: Comparative in Water Suipply and Sanitation: Roles and Realities. Case Study" Environmentally Sustainable Develop- Technical Paper no. Series 22. The Hague, The ment Division, Africa Technical Department, World Netherlands: International Reference Centre for Bank, Washington, D.C. Community Water Supply and Sanitation. . 1993b. "Janta Ko Khane Pani ra Safai Karyakram Wade, Robert. 1986. "Common Property Resource (AKPAS Project)" (The People's Water and Sanitation Management in South Indian Villages." Proceedings of Project). Brochure. Partnership Funds for Field-Testing the Conference on Common Property Resource Management, Service Delivery Options for Rural Water and April 21-26, 1985. Washington, D.C.: National Sanitation. UNDP-World Bank Water and Sanitation Academy Press. Program, World Bank, Washington, D.C. Warner, Dennis, Raymond B. Isely, Craig Hafner, and Wright, Albert. 1993. Personal communication. John Briscoe. 1988. Malawi Self-Help Rural Water Supply Yacoob, Mary, and Jane Walker. 1991. "Community Program: Final Evaluation. WASH Field Report no. 186. Management in Water Supply and Sanitation Projects: Washington, D.C.: U.S. Agency for International Costs and Implications." Aquia (United Kingdom) 40 (1) Development. (February 1991): 30-34. I i i i I 9 A 4 t!?1 ''4.' W r7 , AV jj t 44 - .4 B =1 1IA -A ;l| N, 'i ''. On.|w - Awl I 4 tsk'i't i I I ________