31631 Scaling Up Community Driven Development: A Synthesis of Experience Stuart Gillespie Submitted to the Social Development Family of the World Bank International Food Policy Research Institute 2033 K Street, NW Washington, D.C. 2003 December 31, 2003 Acknowledgements I am very grateful to the following people who have contributed significantly to this project and ultimately thereby to this synthesis: James Garrett, Suneetha Kadiyala, Muge Kokten Finkel, Ayse Kudat, Nurbek Omuraliev and Manohar Sharma for their excellent case studies which illuminate and ground this paper; Rob Chase, Keith McLean, Dan Owen and Jeff Ramin at the World Bank for their guidance and comments throughout; Lawrence Haddad of IFPRI and Steen Jorgensen of the World Bank for assistance in initiating this project; Jay Willis and Lynette Aspillera of IFPRI for help with document finalization and project administration, and finally all the participants of a series of brown bag seminars held at the World Bank in October/November 2003 for their feedback and suggestions. I would also like to thank the World Bank for funding this work. 2 Abstract While many Community-Driven Development (CDD) initiatives may be successful, their impact is often limited by their small scale. Building on past and ongoing work on CDD, this study addresses the fundamental question: how can CDD initiatives motivate and empower the greatest number of communities to take control of their own development? What are the key contextual factors, institutional arrangements, capacity elements and processes related to successful scaling up of CDD, and, conversely, what are the main constraints or limiting factors, in different contexts? Drawing upon recent literature and the findings from five case studies, key lessons on how best to stimulate, facilitate and support the scaling up of CDD in different situations, along with some major challenges, are highlighted. Lessons include the need for donors and supporters of CDD including governments to think of the process beyond the project, and of transformation or transition rather than exit. Donor push and community pull factors need to be balanced to prevent "supply- driven demand-driven development". Overall, capacity is pivotal to successful CDD and its successful scaling up over time. Capacity is more than simply resources -- it includes motivation and commitment which in turn requires appropriate incentives at all levels. Capacity development takes time and resources but is an essential upfront and ongoing investment, with the capacity and commitment of facilitators and local leaders being particularly important. A "learning by doing" culture ­ one that values adaptation, flexibility and openness to change -- needs to be fostered at all levels, with time horizons adjusted accordingly. The building of a library of well-documented context-specific experiences through good monitoring, evaluation and operational research will be useful in advocating for improvements in the contextual environment. Ultimately, for CDD to be sustained, it should be anchored within existing contextual systems (government), frameworks (e.g. PRSP) and processes (decentralization) even where these may be imperfect. 3 1. Introduction Community Driven Development (CDD) is a process that recognizes that poor people are prime actors in the development process, not targets of externally-designed poverty reduction efforts. In CDD, control of decisions and resources rests with community groups who may often work in partnership with demand-responsive support organizations and service providers, including elected local governments, the private sector, NGOs, and central government agencies. Experience has shown that, given clear rules of the game, access to information, and appropriate support, poor men and women can effectively organize to provide goods and services that meet their immediate priorities. Not only do poor communities have greater capacity than generally recognized, they also have the most to gain from making good use of resources targeted at poverty reduction (Alkire et al 2001) According to the World Bank's Voices of the Poor study based on interviews with 60,000 poor people in 60 countries, poor people demand a development process driven by their communities. When the poor were asked to indicate what might make the greatest difference in their lives, they responded: (a) organizations of their own so they can negotiate with government, traders, and NGOs; (b) direct assistance through community- driven programs so they can shape their own destinies; and (c) local ownership of funds, so they can end corruption. They want NGOs and governments to be accountable to them (Narayan and Petesch 2002). CDD's potential is increasingly recognized. Individual studies have shown that CDD can increase the effectiveness, efficiency and sustainability of projects or programs, making them more pro-poor and responsive to local priorities. Other objectives include developing capacity, building social and human capital, facilitating community and individual empowerment, deepening democracy, improving governance and strengthening human rights. A recent paper by the World Bank's Development Research Group finds the evidence for these claims to be mixed (Mansuri and Rao 2003). There are indeed examples of success, but there are few rigorous evaluations ­ even of small-scale CDD -- that clearly delineate what works, where and why. The Issue of Scale CDD is promoted by governments, multilateral agencies, and NGOs alike, but with some exceptions, initiatives are usually small-scale. Its potential for wide-reaching impact, through adoption by a large number of communities, has not yet been demonstrably realized. Though it is no longer true to say that CDD operates at the margins, it has yet to fully infiltrate the mainstream when large-scale programs are at stake. No matter how effective a scattering of small-scale initiatives are, the achievement of national and international goals, including the Millennium Development Goals, will not be possible without effective large scale action. 4 The following statement from the World Bank website encapsulates the rationale for scaling up CDD and is the starting point for this study: "The ultimate goal of CDD projects is to move from being "islands of excellence" that serve and empower a handful of communities, to operating at a national scale, where all regions of the client country can benefit from the project's approach. By "scaling up," the aim is to reach the greatest possible number of poor people, and to motivate and empower the greatest number of communities to take control of their own development" There are many other definitions of scaling up -- or what, two decades ago, used to be referred to as "going to scale" ­ for example, bringing "more quality benefits to more people over a wider geographical area, more equitably, more quickly, and more lastingly." (IIRR 2001). The focus in this paper is ultimately on achieving large-scale, sustainable impact of CDD. Such a focus necessarily includes the environment/context within which such initiatives emerge and grow. Past work has examined the relationship between the scaling up of community-driven initiatives and the `scaling down' of governmental power, responsibility and resources (e.g. Uvin and Miller 1994). But what exactly is being scaled up to achieve this large-scale impact -- technologies, processes, projects, innovations, methodologies? The four core features of CDD which form part of the World Bank Africa Region's Vision for CDD (Binswanger and Aiyar 2003) provide some guidance here. They are: · Real participation and empowerment including devolution of authority and resources, community co-financing, local participation in all stage of project design, and availability of technical assistance and facilitation from higher administrative levels; · Accountability: horizontal and downward to peers, not just upward to central levels; · Technical soundness: context-specificity, designing in phases, ensuring full participation, administrative and fiscal, as well as political, decentralization, fully utilizing local capacities, use of clear field-tested manuals and tools for each actor, provision of technical advisory services; · Sustainability: fiscal, asset, environmental and social In the last two years there has been an upsurge in interest in how to translate small-scale success ­ the numerous `boutiques' dotted around the developing world -- into large- scale impact. Driven in large part by the World Bank, this has led to recent several initiatives, reports and papers on the issue of scaling up. This is one of them. Others include the Mansuri/Rao review of the evidence for CDD effectiveness, the Binswanger/Aiyar paper on theoretical underpinnings and design implications of CDD upscaling, and Jim Hancock's paper on scaling up `good practice' in support of the 5 Bank's Rural Development Strategy (Mansuri and Rao 2003; Binswanger and Aiyar 2003; Hancock 2003). 2. Objectives The aim of this synthesis is to compare and contrast the case study findings with those of earlier work in order to elicit core features, tendencies, success/failure factors for given contexts, and to highlight challenges which need to be confronted to maximize the potential of CDD for large-scale impact. Comparison will necessarily be indirect as these parallel studies where not originally designed to be directly comparable. These case studies differ from many others in that they are studies of community-based initiatives that are not directly supported by the World Bank, carried out by a multidisciplinary team from outside the Bank. The ultimate goal was the same ­ essentially to synthesize experience, add to the growing body of knowledge of what works, and thus respond to the fundamental question: "How can CDD initiatives motivate and empower the greatest number of communities to take control of their own development?" In order to operationalize the findings, our study will be structured to answer the question: what are the key conditions, elements and processes relating to successful scaling up of CDD, and, conversely, what are the main constraints or limiting factors, in different contexts? Through examining the conditions and processes depicted in Figure 2 in the five case studies and past literature, we anticipate that it will be possible to derive certain widely applicable principles and lessons as to how best to stimulate, facilitate and support the scaling up of CDD in different contexts, and also what not to do. A caveat is required here. Not all of the case study examples conform entirely to the strict definition of CDD, though they are all community-based development initiatives that seek to shift control over decisions and resources toward communities. This broadening of the scope allows for inclusion of more experiences from which to distil findings and generate lessons. 3. Conceptual Framework This study is premised on the notion that, while CDD initiatives that succeed in a given context are not transferable per se, there are certain basic operational principles related to success that may be effectively applied more widely. These principles broadly relate to the particular set of institutional arrangements (see Figure 1), capacities and processes that underpin and facilitate successful CDD initiatives in given contexts. These conditions and processes ­ including those relating to central and local government -- also affect the potential for CDD initiatives to effectively scale 6 up. One other important dynamic element comprises the triggers, catalysts or "sparks" for scaling up. Figure 1: Examples of Common Institutional Arrangements for CDD A B C Partnerships between CBOs Partnerships between CBOs Direct partnerships between and local governments and NGOs/private firms CBOs and central government or central fund Community- Community- Community- based based based organizations organizations organizations Local or Local or NGOs and municipal municipal private firms government government Elected local Central NGOs and or municipal government or private firms government central fund Central Central government or government or central fund central fund Source: Alkire et al 2001 Figure 2 below is an initial conceptual framework portraying the relationships between the following integral conditions and processes. · Context: This includes the macro-economic, social, cultural, political and historical conditions and trends in which the CDD initiatives have taken root and are being scaled up. The literature suggests that even where there is spontaneous diffusion, there may be a limit or "context roof", dictated by such contextual factors. In order to maximize impact it will be important to anticipate such boundaries and develop a specific plan for surmounting them. · Institutional arrangements relate to the particular configuration of community- based organizations (CBOs), civil society organizations or NGOs (CSOs), local and central government (LG and CG) as per the typology developed by the World Bank (Figure 1). Where there is a provincial government involvement (e.g. India), this would also need to be included (between central and local). Another level, not depicted here, is that of donor involvement. · Capacity is here defined as "the ability of a person, community or organization to take control of its own destiny and to manage and direct its development process through an iterative process of assessment, analysis and action" (Gillespie 2001). 7 Capacity includes the quantity and quality of the following types of resources: human resources (skills, time, knowledge etc), social, organizational, financial, natural and material. But it goes beyond resources to include such factors as authority, power, responsibility, incentives, leadership, and systems of problem- solving, decision-making and communication. · Triggers, catalysts or `sparks' for scaling up ­ the motivating or driving forces e.g. a successful pilot project, a charismatic individual leader or a change in the local or global context. There may be a combination of factors/events coming together. Such triggers/catalysts may come unexpectedly or they may be deliberately planned, emerging from the felt need among certain stakeholders to show large-scale impact. It will be important to recognize the type and source of such factors/events. Figure 2: A schematic overview of processes of scaling up community-driven development CONTEXT; CONTEXT; VISION/GOAL --local local Impacts (outcomes, benefits) --wider wider FUN P CTIO AL N QUANTITATIVE NAL ITILO C LA IZATIO AN G R O Sparks CDD ACTIVITIES CAPACITY ELEMENTS INSTITUTIONAL ARRANGEMENTS Scaling Up Processes How does scaling up occur? Through what types of process? Scaling up processes may be considered to be a sub-set of CDD processes that relate to all aspects of decision- making from problem assessment, causal analysis to choice and implementation of relevant actions. This would include the community's decision on initial priority activities. 8 To help structure our analysis of the five case studies, and other experiences, a scaling up taxonomy adapted from that used by Uvin and Miller (1994) and Korten (1980) is developed here. Essentially four different types of scaling up processes are differentiated: · Quantitative where a program expands in size, geographical base or budget (also referred to as `scaling out'); · Functional, involving increases in the types of activities and integration with other programs; · Political, involving increases in political power and engagement with wider political processes; and · Organizational, involving increases in organizational strength. At any time during scaling up, one or more of these processes may be underway (see Table 1) Table 1: Taxonomy of Scaling Up Quantitative scaling up (or scaling out) Spread increasing numbers of people spontaneously adhere to the organization and its programs, perceiving them to serve their interest/preferences. Replication a successful program (methodology and organizational mode) is repeated elsewhere. Nurture a well-staffed and well-funded outside agency, using a specific incentive-based methodology, nurtures local initiatives on an increasingly large scale. Integration a program is integrated into existing structures and systems and in particular government structures after it has demonstrated its potential Functional scaling up Horizontal unrelated new activities are added to existing programs, or new programs are undertaken by the same organization Vertical other activities related to the same chain of activities as the original one are added to an existing program (i.e upward or downward linkages are made) Political scaling up First generation essentially service delivery Second community capacity development for self-reliant action. Through better information generation and mobilization, an organization's members or local communities are stimulated to participate in the body politic. Third beyond the community, influence policy reform to foster an enabling environment. This may involve networking and aggregation of organizations into federative structures designed to influence policy. Fourth beyond specific policies, catalyze social movements, and/or direct entry of grassroots organizations, or their leaders, into politics (either through creating a political party or joining one). Organizational scaling up Internal Increasing organizational capacity and improved management processes (links to management effectiveness and efficiency) Financial Increasing financial viability/autonomy, including self-financing, through viability subcontracting, consultancy or fees-for-service. Institutional both internally and externally (including diversification of donors) and linkages with diversification other actors/organizations 9 There are several other classifications of scaling up processes. Hancock (2003) for example makes the differentiation between organizational growth and institutional and policy change -- both of which encompass direct and indirect routes. Organizational growth could be direct through replication or spread, or indirect through the formation of partnerships and integration. Institutional and policy change could be direct, via capacity development and mainstreaming, or indirect through advocacy and lobbying. Another system is employed in the Kecamatan (KDP) program in Indonesia which differentiates three main types of scaling up process: physical, social and conceptual: · Physical scale-up depends on simple systems -- systems for disbursement, for decision-making, and for implementation, monitoring and evaluation, which are fully described in manuals. · Social scale-up arises from the experience of success by stages, and is manifested in the development of community capacity to work synergistically towards common goals in relation with local government. · Conceptual scale-up implies that the meaning and dimensions of CDD will continue to evolve, that CDD principles will be progressively incorporated into policy goals, institutions and new poverty programs, and that there will be an increasing understanding of the larger issues that affect community well-being. Myers (1984) refers to scaling up through three different approaches ­ expansion, explosion and association. Another classification, used in the past by IIRR (2001), suggests that scaling up needs to be viewed from the following perspectives: · Institutional: the need to look at the processes and mechanisms involved in the scaling up process; that scaling up involves wider stakeholder participation, involving as many development actors as possible and that it has to promote participation. · Geographical/spatial: that scaling project/activities involves expanding coverage to other communities/ municipalities. · Technological: that scaling up may also mean broadening or implementing appropriate technologies or implementing complementary or additional activities/technologies to increase productivity or to better manage ecosystems more sustainably. · Temporal: the timing and duration of scaling up. · Economic: cost of scaling up, economic viability. It is useful to keep the polar opposite universalist and contextualist positions in mind too, as described in Hancock (2003) and shown in Figure 3. Universalists view scaling up essentially as expansion of a pre-designed project, while contextualists see the environment as all-determining. 10 Figure 3: Universalist and Contextualist Perceptions of Scaling Up Universalist vs. Contextualist (Hancock 2003) C ontext Policy chang e Inspiration Expansion R eplication Source Source A daptation Dynamics of Scaling Up Sequencing and timing ­ the temporal dimension -- is clearly important. Binswanger and Aiyar (2003) refer to three sequential stages of moving towards large-scale CDD: · Initiation ­ may include enhancing participation, engaging in dialogue on de- centralization, and/or piloting CDD · Scaling up ­ after a successful pilot. Requires planning for training and logistics, development and field-testing of manuals etc. · Consolidation -- may include going for national coverage, moving from participation to full empowerment, capacity development, expanding and deepening CDD functionally to address issues that may not have been first priorities e.g. chronic malnutrition or HIV/AIDS, and/or forming networks or federations of stakeholders. Twenty three years ago, David Korten also differentiated the dynamics of project effectiveness, efficiency and expansion (quantitative scaling up) over time, highlighting some of the trade-offs that may be made between these three goals and emphasizing the process of scaling up as one of learning (see Figure 4). 11 Figure 4: Program Learning Curves The learning process approach to program development proceeds through three stages with each stage involving a different learning task ­ effectiveness, efficiency and expansion · In stage 1 ­ learning to be effective ­ the focus is on participatory assessment, knowledge generation (e.g. through action research and piloting) and capacity development. · In stage 2 ­ learning to be efficient ­ the focus shifts to reducing input requirements per unit of output. At this stage, there may well be some loss in effectiveness in exchange for efficiency, as more serious attention is now given to achieving fit between program requirements and realistically attainable organizational capacities. · Stage 3 ­ learning to expand ­ is the scaling up stage where emphasis is on development of organizational capacity along with continued refinements in the program. By the end of Stage 3 the program will have become relatively stable and large-scale in operation (quantitative scaling up). The organization can then turn to finding solutions to new problems (functional scaling up). Or, if by this time the community has prospered as to upset the fit previously attained, there may be need to repeat the learning cycle to redefine the program and realign organizational capabilities accordingly (possibly moving towards the political or the organizational type of scaling up). 12 Clearly there is much overlap among these many classifications. While accepting there is no one universally recommended system, we aim to proceed with Table 1 as a starting point in our enquiries, whilst keeping in mind the contributions offered by the other classifications. 4. Methods and Cases As its starting point, each case study addressed a series of cross-cutting questions, within the limits of the case context, available data and the given time-frame of the project. These questions related to different parts of the conceptual framework (see Box 1): Box 1: Adaptable Checklist of Cross-Cutting Questions 1. Scale, impact, vision Evidence of impact (small and large scale), degree, sustainability What initial vision (scaled up project or scaled up impact)? Whose vision? Is it shared by all stakeholders? 2. Context, institutional and financial arrangements Socio-economic, cultural, political, policy context and trends Linkages between central, local govt., NGOs, community orgs etc Financing levels, flows, systems, sustainability 3. Capacity Capacity strengths and weaknesses/gaps Resources (human, economic and organizational) Authority, responsibility, motivation Systems for monitoring, communicating and learning 4. Triggers, catalysts, `sparks' Internal or external? Accidental, spontaneous or planned? 5. Scaling up processes Scaling up what (what is replicable/standardizable?) When to scale up (phases and sequencing)? How to scale up? Different types of scaling up (see taxonomy overleaf) 6. Appropriate forms of external support for CDD 13 In general the study used a mix of qualitative and quantitative methodologies for addressing the key questions. The study drew upon secondary data sources (project documents, evaluations etc), supplemented, where possible, with site visits and focused key informant interviews. The following briefly describes the five case studies that were undertaken. Detailed descriptions, prepared by the authors, are provided in the Annexes. · Zambia: PROSPECT (Program of Support for Poverty Elimination and Community Empowerment) evolved from PUSH (1992-97) and aims to reduce poverty reduction in peri-urban areas of Lusaka and Livingstone through representative community organization. The main foci are social "empowerment" (institution building), personal "empowerment" (microfinance) and the development and maintenance of infrastructure. Supported by CARE International (Author: James Garrett) · Malawi: Scaling-up HIV/AIDS Interventions Through Expanded Partnerships (STEPS) is a multi-sectoral response to HIV/AIDS comprising community mobilization/planning by village AIDS committees aimed at prevention, care, support, mitigation, and supported by Save The Children (US). (Author: Suneetha Kadiyala) · Kyrgyz Republic. Comparison of two programs: the UNDP-supported Decentralization Program which is CDD working through strengthening CBO-LG links with emphasis on local governance. And the DFID-supported Sustainable Livelihoods Program among livestock producing communities, which is driven by social mobilization through "village working groups" and catalyzed by a local NGO, bypassing local government. (Authors: Ayse Kudat, Muge Kokten Finkel and Norbek Omuraliev) · Nepal and India. Comparison of two approaches to extending microfinance to the poor. Nirdhan Utthan Bank Ltd. (NUBL) in Nepal which extends group-based loans to targeted clients (self-proclaimed `Grameen replicator'), and the Self Help Group Linkage Banking Program of the National Bank for Agriculture and Rural Development (NABARD) India which acts as an apex organization linking informal "self-help groups" with decentralized commercial banks. (Author: Manohar Sharma) · Kerala, India. Kudumbashree (KDB) is a multi-sectoral state-wide poverty alleviation program, based on village-level micro planning, the formation of `thrift and credit societies', microenterprise, and women's empowerment through social mobilization and group formation. (Author: Suneetha Kadiyala) The criteria for choice of case studies were mixed ­ we looked for geographic balance, though oversampling regions with a higher incidence of poverty (Sub-Saharan Africa and South Asia). We also purposively selected cases from geographical and sectoral areas that IFPRI researchers had been, or were still, active ­ to build on their knowledge of the context. As mentioned some of these cases do not fully meet the strict definition of fully- 14 fledged Community-Driven Development but they nonetheless do represent attempts to scale up initiatives that seek to mobilize and empower poor communities. Institutional arrangements differed ­ the Kyrgyz UNDP-supported project aligned with Example A in Figure 1, while Kudumbashree aligned with example C. The most common arrangement was B ­ partnership between CBOs and NGOs, with central or local government playing a supportive role. 5. Synthesis In this section, the main highlights of the case study work, supplemented with those of the literature review, are described as they apply to the different parts of the framework and the type of questions listed in Box 1 above. These include both facilitating and constraining factors. `Success factors' or strengths are outlined, while failures or weaknesses are viewed as challenges, with recommendations on how they may be addressed or prevented. In the concluding section, key lessons are highlighted. 5.1 Scale, Impact, Vision Patchy Data A major problem throughout all these case studies (and indeed elsewhere as described by Mansuri and Rao 2003) is that good evaluations are thin on the ground. It is crucial, for all stakeholders, to establish basic effectiveness and impact -- firstly of the initial CDD initiative and secondly of the CDD scaling up process and outcomes. Crucial because learning and change, both within and beyond, requires an understanding of what worked where, and ideally, why. Donors and governments also need to be convinced of the benefits of CDD. From the cases, there is some evidence of increased income in Kyrgyz, near universal on- time loan repayment by self-help groups in the Indian NABARD program, and increased social capital in Malawi (less stigma, more commitment re: AIDS, more examples of collective action) where STEPS is active. In general there is a significant bias toward data on coverage and outputs, with few data on outcomes (e.g. reductions in the rate of HIV spread or in the severity of downstream impacts of AIDS in the case of STEPS, or reductions in poverty or malnutrition in the case of KDB). Where outcome data are available they are usually not comparable with a baseline ­ either because there is no baseline, or the study design precludes valid comparison ­ a basic failing of many projects. A straightforward before/after with/without design would provide firmer understanding of changes and how much was attributable to the project. It would also be interesting to make more direct comparisons -- say, between a CDD approach and a more service delivery-oriented system. Quality is important -- key questions with regards to community-based structures might include how representative they are, how successful they are at increasing community participation and trust in governance, how capable they are of leading development, how much community residents interact with them, how they are incorporated into municipal 15 planning, how financially legally and politically stable they are, and what needs to be done to address weaknesses. Participation is both an end and a means, and needs to be viewed as such in process and outcome evaluations (see below) 5.2 Context, Institutional and Financial Arrangements Several qualitative studies indicate that the sustainability of CDD initiatives depends crucially on the existence of an enabling institutional environment (Mansuri and Rao 2003), one in which policy and institutional reforms are oriented toward increased control of decisions and resources by community groups and/or participatory elected local governments. In this section, some of the more important aspects of such an environment are discussed. Political and Financial Commitment This is manifested in the proactive opening up of space for CDD by political leaders i.e. the creation and maintenance of an enabling environment. Commitment should not be measured by policy statements alone. Evidence of budgetary re-allocations, real policy change and the willingness to make reforms ­ for example, the decentralization of authority and resources -- are key. Pro-CDD legal mandates in the form of new decrees are common features. In Kerala, the system of Community Development Societies, on which KDB is based, emerged following the passing of a unique set of bylaws which empowered such groups to receive funds from governments, banks and other donors. The State Government of Kerala (GOK) also stipulated a statutory provision of 2% municipality revenue for poverty eradication through the CDS structures. Women's participation has increased following the GOK reservation of local government (LGB) seats for women. Kerala, of course, is quite different to most other Indian states regarding its history of social and labor movements, and the relative absence of socio-economic, caste, educational and gender inequalities. As Kadiyala (2003) puts it: "politicians in Kerala do not fear empowered groups". On the other hand, the fact that GOK is a bureaucracy like any other state government, and one in which departmentalism is rife from top to bottom, represents a constraint to multi-sectoral programming. In Kyrgyz, political stability is enhanced by continuous efforts toward decentralization. Support for local self-governance is visible in many spheres. Municipalities (AOs) can be further encouraged to build capacity through upgrading qualifications of their staff, improving governance practices, establishing service departments, and upgrading the overall profile of the local congress in their communities. Donor support for decentralization is strong and translates itself into willingness to fund CDD activities at the AO level. The Kyrgyz example also highlights the need to minimize bureaucratization, paperwork, and over-regulation of CBOs in order to truly enable, and not micromanage, the CDD process. 16 In Nepal, credit-based NGOs such as NUBL would not have been permitted prior to the shift to democracy in 1990. Stability too was important, with the scaling up of NUBL subsequently hampered by Maoist insurgency. Looking at the Asian microfinance programs, banking reform has been key to their expansion. India's Central Bank classified loans to Self Help Groups as `priority sector' loans, providing major incentive for commercial banks to lend to SHGs. In Nepal, NUBL also received concessional loans from commercial banks following government regulation of at least 12% of the loan portfolio to be channeled into `priority sectors'. In Malawi, the establishment of several frameworks contributed importantly to the successful scaling up of STEPS. The president declared HIV/AIDS a national emergency in 1999. This was followed by the adoption of a comprehensive national HIV/AIDS strategic framework based on the premise that effective action to combat the epidemic requires a multisectoral approach, implemented by a wide spectrum of partners within public and private sectors, civil society and faith communities. In 2000, the National AIDS Commission was established to coordinate the national response, and was later instrumental in Malawi's being awarded a major grant from the Global Fund. During this time (1998-2001), according to a study by the Policy Project, the policy environment in Malawi improved greatly across seven indicators. Yet more mainstreaming is required -- AIDS still does not feature prominently in the current PRSP plan. Decentralization of Power and Resources The historical experience of decentralization is mixed (Binswanger and Aiyar 2003). Too often local governments have been created without administrative authority or fiscal resources. Decentralization has three key dimensions--- political, administrative and fiscal--and all three need to be harmonized to work well. One precondition for effective decentralization is a high degree of central state capacity (Platteau and Abraham 2002). If weak states cannot successfully pursue decentralization, it is because when a weak state devolves power it is more often than not simply making accommodations with local strongmen rather than expanding democratic spaces (Heller 2001). A process of "recentralization" ­ essentially elite capture of decentralized resource flows ­ has been documented and is discussed below. But there are clear benefits. The system of decentralized AIDS committees in Malawi is a major factor in STEPS's success. In Kyrgyz, political decentralization has increased trust in local leadership. The election of local governments has substantially increased communities' trust in their elected officials, and their belief that they can hold officials accountable for their work. The process of decentralization and local elections has also created an unprecedented opportunity for the newly elected officials to build new coalitions. Instead of aligning themselves with the regional governors, who used to appoint them, local leaders now have broader networks of supporters in their own communities. Do governments need to `scale down' to provide the space for CDD to scale up? `Scaling down' may not be a wholly accurate term for the changes that may be required. 17 Conventionally `scaling down' refers to a decentralization of authority, resources, and capacity. A broader definition would include the enabling, supporting, or facilitating function whereby the government adopts modes of functioning that allow local communities and organizations to build conceptual, operational, and institutional capacities. It thus refers to a shift in the balance between top-down supply and bottom-up demand. In finding this balance, governments and communities need to agree on where their respective comparative advantages lie ­ CDD, for example, may work well with microfinance, while child immunizations require a more government-controlled approach. `Scaling down' thus does not in any way imply the need for government to disengage from community-driven development. But it does require it to be more flexible and more responsive to locally-generated demand -- to ensure the terrain is fertile enough for community organizations to emerge, learn and grow. Considered in this light, for CDD to be sustained, it will be necessary for some direct interaction with government systems, and for this to happen, governments too will need to change their ways of operating. Community Cohesiveness The degree of community cohesiveness is in a sense a measure of the degree to which the microenvironment is enabling. Conversely, the fractionalization of communities acts as a bottleneck or even a roadblock to CDD. Recent work by Das Gupta et al 2003 has shown that with greater intra-community equality, there is less of a trade-off between using communities' advantages in local information and elite capture. In the Bangladesh food for education program Galasso et al (2000) also found that power in community decision- making clearly mattered in pro-poor intra-village targeting, with greater intra-community inequality associated with greater marginalization of the poor in village level targeting. In Malawi a significant difference was found with regard to the success of STEPS -- between rural and urban areas and between populations dependent upon agrarian livelihoods and those more dependent upon more individuated livelihoods such as tourism. Communities with indigenous traditions of collective action and experimentation have a clear advantage. In Kyrgyz, pre-soviet and Soviet traditions, including the practice of ashar in which all community members contribute to community projects, were important. Use of such pre-existing traditions of mutual help should be maximized in scaling up, argue Kudat et al 2003. Interestingly, the Zambian experience was different ­ here the one-party state and socialist history was said to have reduced community initiative -- "communities expected everything to come from the state" (Garrett 2003) 18 Harmony and Responsiveness of Institutional Arrangements to Evolving Context Comparisons between the scaling up of microfinance in Nepal and India show the importance of institutional responsiveness to context. In Nepal, where commercial banks were thinly dispersed, NUBL adopted an approach of catalyzing group formation. In India, where there exists a dense network of commercial banks throughout the country, an apex role of facilitating the linkage of groups to banks was more appropriate for NABARD. As activities change over time, institutional arrangements evolve and the configuration of actors and their inter-relationships may change. New groups may emerge (e.g. water user groups in Kyrgyz) and become increasingly instrumental in prioritizing local needs. Das Gupta et al 2003 describe how higher levels of government can form alliances with communities, putting pressure on local authorities from above and below to improve development outcomes at the local level. These alliances can also be very effective in catalyzing collective action at community level, and reducing local capture by vested interests Working with Government and Other Existing Structures There are few examples of scaled up community development approaches which have bypassed government structures and been sustained. CDD is fundamentally aligned with a rights-based approach to development, which holds elected governments ­ as duty bearers ­ ultimately accountable for the facilitation and fulfillment of the rights of its citizens. Bypassing government may lead to the government bypassing this fundamental obligation. Summers (2001) has also highlighted the risk of parallel structures undermining democratically elected national governments. The dangers of the initial approach of PROSPECT in Zambia of avoiding governmental involvement was highlighted by Zambian government (GOZ) suspending new elections among the area-based organizations (ABOs) effectively blocking the scaling up process for three years. GOZ was increasingly being eclipsed in influence and access to donors by PROSPECT and GOZ-elected local councilors felt threatened by a parallel structure with resources, albeit with no legal authority. The hiatus ended when an agreement was reached and the PROSPECT ABOs were institutionalized as the lowest tier of government, providing stability, continuity and authority. Integration with government structures may bring the greater legitimacy and provide a mechanism for scaling up, but it may also bring a new dimension of vulnerability -- for example, if a "new crop of councilors come into power and try to push us out". Group federations were considered to be one mechanism for insuring against such practices. In KDB in Kerala, involvement of state and local government in designing the pilots was considered one of the most important determinants of scaling up of the CDS approach. The current director of KDB used to be the district collector of Malappuram, where one 19 of the two successful pre-KDB pilots was initiated. The embedding of KDB in a permanent institutional framework ­ the local government bodies (LGBs) which have substantial funds due to fiscal devolution and a solid tax revenue ­ ensures sustainability. So government should not be bypassed, but nor should government remain static ­ it needs to become more conducive and responsive to different types of community demand. Communities should not be artificially `sectoralized' to conform to long- standing bureaucratic incentives that prioritize specialization. Such incentives and systems need to change. Inter-Donor Coordination Donor compartmentalization or `balkanization' ­ whether by sector or by geographical area ­ may constrain CDD scaling. So too will donor duplication where there is over- emphasis on a certain problem or a certain region, potentially sucking funds from elsewhere. Part of the space required for the scaling up of CDD relates to the ability of external agencies to provide appropriate resources when they are demanded by communities (see below). This applies primarily to local and central governments, but it also applies to donors. Effective donor coordination ­ in terms of who, when approached, will provide what type of support, where -- is itself a form of functional scaling up of external support. 5.3 Capacity What are the essential capacity pre-requisites of scaling up? Conversely, what are the critical capacity gaps/weaknesses that prevent or limit scaling up? A full definition of capacity development, as mentioned, is the strengthening, enhancing and nurturing a community's abilities to take control of its own destiny and to manage and direct its development process through an iterative process of assessment, analysis and action (see Figure 5, adapted from UNICEF 1990). Figure 5: The Triple A Cycle ASSESSMENT Of the problem and situation ACTION ANALYSIS Based on the analysis Of the causes of the and available problem, roles and resources patterns and capacity 20 The process of assessing, analyzing and continuously developing different elements of capacity is absolutely pivotal to effective CDD. Capacity assessment is essential, though not so frequently undertaken. A critical stage of the PROSPECT scaling process, for example, involved a PANA ­ participatory appraisal and needs assessment ­ to keep track of problems, capacity strengths and weaknesses as well as opportunities. Box 2: Key Elements of Capacity RESOURCES Human -- total number, skills, time, motivation, knowledge, beliefs Financial ­ savings, income, flow, diversity of sources, financial services Social/organisational -- formal and informal institutions e.g. family, CBOs, networks Physical ­ infrastructure such as transport, shelter and buildings, water supply and sanitation, affordable energy; and access to information (communications). Natural resources ­ land, water, common property resources AUTHORITY Legal status of individual/organisation in wider society (including traditional law) Social/political legitimacy or credibility of the individual or organisation Mandate established within organisation or in wider fora Relationship to constituency ­ perceived relationship by individual/organisation and by constituency, mechanisms of representation/feedback RESPONSIBILITY/MOTIVATION/LEADERSHIP Clear responsibilities corresponding to key functions in relation to goal Incentives -- salaries, indirect economic benefits, professional and social status or recognition, training or learning opportunities, personal encouragement Attitudes and values -- related to institutional norms and practices at different levels in open society and within organisations Leadership/direction Strategic vision related to relative niche Credibility SYSTEMS (OF ASSESSMENT, ANALYSIS AND ACTION) For detecting & analysing changes in the context and relevant response (i.e. problem detection and solving). For determining particular niche given mission and mandate (i.e. positioning individual or organisation in context vis-ą-vis other stakeholders) For planning action which corresponds to context and mission/mandate For analysing, documenting and improving own performance (or fora where analysis could take place) including drawing in lessons from own experience and that of others; Existence of functioning feedback loop from action to re-assessment (which permits learning) Access and ownership of communication systems Degree of participation in communication; connectivity of all relevant actors Source: Gillespie 2001 `Capacity' is much more than resources per se. Attitudes, values, motivation ­ particularly on the part of community members and government personnel -- are all key factors in generating and sustaining community driven development at all levels. Capacity assessment needs to focus on the type of capacity elements outlined in Box 2. 21 Capacity Development Capacity development is a process, not a one-time undertaking. Once existing capacity strengths and gaps have been assessed, strategies for developing or strengthening capacity can be designed. These strategies need to be grounded in a theory of action that links capacity development to expected results. The preparation of manuals or sourcebooks, which provide method menus and guidance, will be important here. These manuals should be field-tested, adaptable `living documents'. For example, in Kudumbashree, about 24 manuals were prepared to train staff at all levels (state, district, and local government body (LGB) on various topics. A cascade approach to training is followed, with KDB staff training state and district level resource persons in urban and rural LGBs, who then trigger Neighborhood Group (NHG) formation and train NHG members on various topics including human resource development, micro- entrepreneurship development, technical trainings in health and education, thrift and credit operations, infrastructure, convergence, and the preparation of micro -and mini- plans. In NUBL, once groups have been formed, all group members undergo a Compulsory Group Training (CGT) in order to fully understand the modus operandi of group activities. Completion of the training and certification by NUBL staff is required before lending can begin. The skill sets required of different stakeholders will change, just as the activities within CDD initiatives may also change, during the scaling up process. In the Zambian case, for example, in the transition from PUSH I to II, the proportion of social development workers increased as that of engineers declined. NGO sustainability and capacity to facilitate, keep staff trained and motivated, is important. In Zambia, CARE came to learn over time that it was important to build capacity by providing opportunities for communities to do something new, meaningful and concrete ­ to galvanize and harness the latent energy that existed. It also practised what it preached, and sought to empower its staff to think on its feet and remain flexible and adaptive in its decisionmaking. Several cases highlight the importance of nodal capacity support. Save the Children, for example, is planning to scale-up through various innovations including the establishment of Community-Based Learning Centers (CBLCs), based on the principle of " seeing is learning". The idea is to establish a library with all the STEPS-relevant documentation and a training facility in the SC district office in Lilongwe. In addition, in each of the districts, two active community AIDS committees (CACs) and four active village committees in each of these CACs will be chosen to serve as "living universities" ­ to train other CBOs/NGOs and AIDS committees at different levels from other areas. In Kyrgyz, Rural Advisory Services has been created and supported by donors to ensure delivery of capacity building elements to CBOs. These services have been streamlined, and are expanding their scope and geographical coverage. 22 Well-Defined Stakeholder Incentives, Roles and Responsibilities Capacity is translated into action through motivation that in turn requires well-defined incentives. When considering appropriate incentives, subsidiarity is key ­ the devolution of power and responsibility for certain functions to the lowest level able to manage them, or as close as possible to those who will ultimately benefit. The particular roles and responsibilities of different stakeholders at different levels are often spelt out, without being set in stone, in user-friendly operational manuals. For the sake of clarity, transparency and efficiency, such roles may need to be ­ as Sharma suggests ­ "officialized". Incentives are required at all levels. Examples include the availability of discounted refinancing rates for loans to self-help groups which encourage local commercial banks in India to lend to self-help groups. In PROSPECT, a major incentive for ABO membership is the chance of becoming a community facilitator. One incentive at the community level may be the availability of matching grants from the local government for good CDD proposals. Likewise, local government bodies that are more accountable to communities may be rewarded through increased funding from central government. The type of community-central government alliances described by Das Gupta et al (2003) is another example of incentive ­ in this case an incentive via a double squeeze, for local government bodies to improve services to communities. Where commitment needs to be built in external organizations, measures to reward increased cooperation may include providing additional funding, entering into performance contracts, or developing improved monitoring systems (Heaver 2002). Incentives may or may not be financial. Peer recognition, status and the opportunity to learn are valuable incentives in many social programs. Other incentives include performance-linked rewards such as certificates, plaques, stars, designing systems for recognition by the local community, initiating competition between service areas, based on comparison of performance statistics, and rewarding performance through promotion. As well as positive incentives, there is a need to guard against disincentives -- for example the pervasive culture of handouts in Malawi (encouraged by the President himself who frequently gives away cash on his travels within the country.) Committed Leaders and Facilitators In Malawi, STEPS staff and committee members at various levels consistently identified local leadership as the most critical factor in generating collective action. Weak leadership tended to correlate with low community participation. In Kyrgyz, the high level of reciprocal trust that exists between local leaders and community members greatly facilitated CDD, while national level leadership was instrumental in the creation of an enabling environment for CDD scaling up. 23 In order for communities to drive their own development, it is usually necessary for a cadre of competent facilitators to be in place -- certainly at the beginning when communities are being mobilized. While social capital may be the bedrock for sustained community action, it often needs to be unlocked and activated by an external mediator or facilitator. Such an activation of latent capabilities may also trigger scaling up. Qualitative evidence suggests that the role of external agents, such as project facilitators, is a major contributor to successful CDD (Mansuri and Rao 2003). Platteau and Abraham (2002) argue that endogenous community imperfections are being downplayed in the donor-fuelled rush to decentralized or community-driven development. They conclude that overlooking the crucial role of external facilitators and capacity development in such approaches could backfire seriously, and result in the appropriation of resources by local elites (elite capture) operating within a logic of patronage. The Kecamatan (KDP) program in Indonesia involved both technical and social facilitators with clearly defined, multiple roles ­ as information provider, teacher, mediator, motivator and challenger. KDP facilitators are trained in facilitation methods including structured experience and visionary observation. Structured experience consists of five stages ­ do and experience, express and discuss, process and analyze, conclude, then apply (Wong 2003). Capacity to Participate Participation is fundamental to CDD and numerous papers have been written on its meaning, value and how to promote it. Narayan (1997) states that participation in community-based development depends on reversing control and accountability from central authorities to community organizations. Successful design requires tapping into local needs, understanding and building on the strengths of existing institutions and defining the changes needed in intermediary implementing agencies to support community action. With regards to CDD, participation is both a means and an end. But participation of whom, in what activities, and in what way? Case study authors where asked to consider the different types of involvement of key actors within their projects, with regard to the distinctions shown in Table 2 below. Several questioned the distinctions made in this table, and the implication that "collective action" as defined here, was the universally optimal approach. Many of the "C"s above could be happening in the same project at the same time. Or a project could start in consultative mode and evolve and grow towards collective action. The balance in most of the cases lay within consultation (e.g. the South Asian microfinance initiatives), cooperation and co-learning. One near-universal response was that the implied evolution in this table ­ in suggesting the phasing out of facilitators ­ was problematic. Community groups will always need to work with outsiders in some manner (see "capacity to demand"). 24 Table 2: A Participatory Continuum Relationship of Research or Mode of Participation Involvement of Local People Action to Local People Cooption Token representatives are chosen but with no ON local people real input or power. Compliance Tasks are assigned with incentives; outsiders FOR local people decide the agenda and direct the process. Consultation Local opinions are asked; outsiders analyze and FOR/WITH local people decide on the course of action. Cooperation Local people work together with outsiders to WITH local people determine local priorities; responsibility remains with outsiders for directing the process. Co-learning Local people and outsiders share their WITH/BY local people knowledge and understanding to create new understanding and work together to form action plans, with outsider facilitation. Collective action Local people set their own agenda and mobilize BY local people to carry it out in the absence of outsider initiators and facilitators. Source: Cornwall (1996). As Mansuri and Rao (2003) found, key concepts that underlie the "theory" of CDD such as "participation," "community" and "social capital" are sometimes naļvely conceptualized which can lead to poor project design. "Empowerment" too is an over- used term whose meaning is often blurred. Empowerment does not automatically equate with self-sufficiency, as the Uphoff quote below illustrates. True empowerment implies control over decisions and resources, the capability to participate and the capability to make demands1. "One reason for breaking away from thinking in terms of top-down versus bottom-up development is to avoid two fallacies about assisting the poor. The first is the paternalistic fallacy: the belief that planners, technicians, and experts possess all the knowledge, wisdom, and virtue needed to achieve development and that the poor should be responsible and grateful beneficiaries. Similarly mistaken is the populist fallacy that the poor themselves possess all that is needed for their own advancement, that they can do entirely without bureaucrats and technocrats. While there are some impressive self-help examples and enclaves, those regional and national programmes that benefit the poor on a significant scale have been concurrent mutual endeavours from above and below." Source: Uphoff 1988 In a recent IFPRI study of community participation in South Africa's public works programs, while there was a generally shared view among all stakeholders that the 1The World Bank differentiates four main types of participation along a continuum ranging from sharing information (one-way flow) to consultation (two-way flow) to collaboration (increased control over decision-making) and finally empowerment (World Bank 1998) 25 community should participate in some way, there were widely divergent views as to what participation actually meant and what form it should take (Adato et al 2003). De facto participation (who is actually responsible) was quite different to de jure participation (whose name was on the contract), with the former being far more important for the achievement of project outcomes. There was in fact no evidence that de jure participation affected these outcomes. When considering participation as a process, a means to an end, we need again to consider what type of capacities and incentives are required for people to effectively participate in furthering the goals of CDD. Clearly, not everyone needs to, desires to, or is able to participate in all or even most activities. Cooke and Kothari (2001) make an interesting and timely critique of the `tyranny of participation', while others have referred to decisions being impeded or critically delayed by `participation paralysis' (e.g. Edwards and Hulme 2002). What is the balance between the participation of poorest groups and elite capture in scaling up? Even in the most egalitarian societies involving the community in conducting development initiatives is a process that will almost always be dominated by elites because they tend to be better educated and have fewer opportunity costs on their time (Mansuri and Rao 2003). Running parallel with decentralization reform, a process of "recentralization" at the local and provincial levels has been documented with the reemergence of elites (Schonwalder, 1997; Vernon 2001 in Kerala). Kumar and Corbridge (2002) provide one example in India, where "the Project was less successful in targeting the poor, than richer families were in targeting the Project". It is not clear, however, that this always represents "capture" in the sense that elites appropriate all the benefits from the public good. In PROSPECT, Garrett states, so long as elections are free and fair, and information on candidates' backgrounds and positions widely available, then ­ although those chosen for the RDC probably are "elites" -- they are also leaders with certain skills, education, or business acumen. Mansuri and Rao (2003) likewise differentiate benevolent from malevolent capture. Some elite domination, they suggest, is inevitable ­ particularly where the elite are among the few who can read project documents and communicate with outsiders. The risk of malevolent capture may be reduced in relatively homogeneous communities, but collective action is still possible in more heterogeneous communities with strong leadership. Is there a fundamental tension between participation and scaling up? How can effective participation be maintained or stimulated as initiatives grow in size? Earlier reviews are inconclusive on this, in that there are examples of increasing elite capture with increasing scale, and examples where participation is not sacrificed overall, but may take on different forms for different people at different times (IIRR 2001). The perceived tension between participation often stems from a political context ­ in particular, the frequently adversarial relationship between NGOs and the state/government where the former may reluctantly deal with the latter to mainstream their successful small-scale innovations. This stereotypical characterization may becoming less common ­ the emerging "hybrid" paradigm seeks to combine the best of state-led, NGO-led and community-led approaches 26 where the state enables and nurtures the growth and contribution of NGOs and community groups. Our cases differ with regard to the degree that this is evident­ in some, it is words, or policy statements, rather than action. In others e.g. PROSPECT, there are roadblocks as increasing community resource mobilization threatens the power of local politicians until roles, responsibilities and mutual benefits are visualized and activities resumed. Capacity to Demand Communities need not only capacity to act but also capacity to make demands. CDD is thus not about self-sufficiency but more about self-reliance. Local capacity to demand also needs to be matched by external capacity and commitment to respond. One example here from the Kyrgyz Republic is the need for credit institutions to use existing and newly generated social capital as collateral (Kudat et al 2003) In a review undertaken for the World Bank, Heaver (2002) defines commitment as an integral component of capacity, which aligns with the broad definition used here. In the medium and long run, sustained bottom-up demand from organized community groups is probably the best way to maintain the commitment of both policymakers and service providers that was discussed earlier. Empowering households and communities to know their rights, understand their needs, and participate in program monitoring and management is an important tool for building and sustaining commitment. In this way, communities can become directly involved in shaping the enabling environment within which they can flourish. Community-based systems: for financing, planning, management, and learning Regarding financial arrangements, cost data were in general hard to come by. There was simply not enough readily available information to assess to what extent funding constrains scaling up, the degree and type of involvement of communities in financing CDD and its expansion? Community self-financing is not only vital for sustainability, it has also been shown to be important for increasing cost-effectiveness of actions. When resources are locally generated they are usually well managed (Binswanger and Aiyar 2003). In Nepal, NUBL has seen an upward trend in its measure of financial independence or operational self- sufficiency (the ratio between total financial income and total expenses). In NABARD, a Microfinance Development Fund was established to finance scaling up. The financial sustainability of PROSPECT in Zambia on the other hand is on shaky ground (Garrett 2003). The plan to channel 5% of the revenues from the water scheme toward supporting the ABOs is poorly implemented, and it is considered unlikely that the municipal government will allocate funds to support ABOs, given that one of the motives for decentralizing was to reduce city budget commitments. Who decides what issues get addressed in the CDD planning process? In the KDP program in Indonesia, there was an open menu -- essentially a blank sheet of paper ­ 27 though there did exist a "negative list" of activities that were effectively banned (including bribery, politics, use of toxic chemicals, military activities, religious buildings, child labor, and destruction of the environment). Community plans need to be exactly that ­ not an agglomeration of individual wish lists. What happens to issues that fall off the radar screen, or never make it on? As systems of planning scaled up in Kerala, KDB shifted its focus from nutrition and child health to microenterprise which raises the question: is there ever a case for earmarking external support for certain activities? Binswanger and Aiyar (2003) suggests there is ­ at least, with regard to problems such as HIV/AIDS which are both serious and less likely -- through stigma and social exclusion -- to be articulated as community priorities. Acceptable criteria and mechanisms for such `earmarking' at different levels need to be agreed. A well functioning management information system (MIS) should highlight bottlenecks or roadblocks for corrective action or operational research, and it should be based as far as possible on the two main principles for the use of information for action -- first, to collect the minimum, feasible amount of data required to inform and improve decisions leading to action, and second, maximize the use of data at the level they are collected (the subsidiarity principle). Systems need to reflect the reversal of accountabilities, so that information is not routinely collected for immediate transmission to more central levels, but is put to direct use by communities. In response to the recommendations of earlier evaluations, STEPS established such a MIS system in 2001, with a database of 53 monitoring indicators collected at three main levels ­ district, health catchment area and village. Data are collected and discussed monthly in local AIDS committees, then used to inform community-level action prior to the feeding of key information upwards to higher levels. A simple, transparent, streamlined MIS linking options for action with different types of information collected by different actors at different levels needs to be drawn up to support CDD upscaling. This might include the specification of "trigger points" or critical thresholds for action with respect to certain indicators and certain responsible actors. Just as blueprint projects usually do not make sense in the context of scaling up, nor do blueprint evaluations -- those that use close ended questionnaires with preset questions, and statistical analyses based on linear assumptions. A shift toward participatory research and evaluation has run parallel with the move toward CDD approaches. There needs to be a balance between involvement of internal and external evaluators, and between use of qualitative and quantitative methods. Early work on community-based development (Korten 1980, Paul 1992) clearly showed that in successful programs, planning, evaluation and implementation where not divorced from each other, and that the same individuals were involved in all to some degree. Figure 5 illustrates this well. Evaluations are only useful if they are used to inform and improve action. The generation and sharing of knowledge between different actors at different levels, including communities, NGOs and local government, is a fundamental contributor to the building 28 of trust and transparency. The link between information gathering for project evaluation and systems of learning and communication needs to be explicit so that success can be communicated, internalized, and ultimately used to generate more success. Mechanisms, processes and indicators for assessing, sharing, learning from and adapting "good practice" both within a project (e.g. monitoring, MIS, meetings) and externally (e.g. networking) need to be thought through (see "Communicating Past Success" below). 5.4 Triggers, Catalysts, "Sparks" The studies showed that the triggers or catalysts were both endogenous (e.g. a charismatic local leader) and exogenous (e.g. shift in donor priority), both planned and spontaneous. The following were seen to be important in facilitating the move from small to larger scale modes of operation. Leaders and Champions Leaders are not only a crucial element of human capacity, as described above, they are also often highly influential in scaling up. This may be through recognizing and/or capitalizing on the opportunities for increasing impact, or through opening up such opportunities via their own endeavors. For example, NUBL was initiated by the ex- deputy governor of Nepal's Central Bank who quit his job to pioneer NGO banking services in Nepal (Sharma 2003). One early study of community-based development showed that stability, commitment and continuity of program leadership were more important than charisma (Paul 1992) Communicating Past Success Success breeds success. Or at least it inspires. Such demonstration effects may be internal and/or external. Internally, the successful implementation and impact of a small-scale pilot may be the trigger e.g. the urban Alleppey and rural Malappuram programs in Kerala, or the successful piloting of a community mobilization approach with STEPS in Namwera. High repayment rates in the South Asian microfinance pilots paved the way for scaling up, while the success of microcredit programs in Kyrgyz constituted an important trigger for emergence and scaling up of community driven initiatives. Key bridges built between associations of micro-credit groups and other CBOs facilitate a poverty-focused collaboration between CBOs and local governments for sustainable village level investments, as illustrated by the UNDP Decentralization pilots. Externally, for example, the success of the Grameen Bank in Bangladesh was a major influence, spawning a plethora of `Grameen replicators' in Asia ­ similar attempts to extend microfinance to the poorest groups, particularly women. NUBL is one example. Looking forward, there is evidence that these study initiatives are having such an effect themselves--for example, there are STEPS-like models of multisectoral response to AIDS being adopted and adapted in Malawi, and the Mahavita program in Madagascar is based to some degree on PROSPECT in Zambia, also facilitated by CARE International. 29 But past success of its own is not enough. It needs to be communicated (see "community based systems" above). The media has a key role here, as do good internal communication strategies. STEPS facilitates community-to-community interactions through its program of "Open Days" wherein mobilized STEPS communities welcome visits from other communities -- another self-fuelling mechanism for triggering scaling up. Some of the pilots of these CDD initiatives have won awards, which has helped communicate their success more widely. In 1995, the Alleppey municipality won the "We, the Peoples: 50 Communities" Award initiated in commemoration of the 50th anniversary of the UN. Malappuram also won a prestigous Commonwealth Association Award in 2000. As part of its `enabling' function, governments may facilitate communication e.g. dissemination of the success of the Kyrgyz program by the Congress of Local Communities (the decentralization body in the Prime Minister's Office) Institutional Innovation Scaling up may be initiated when a new institutional innovation becomes widely available e.g. the Local Development Fund or the proposed community-based loan committees in Kyrgyz. Support for the establishment of such independent loan committees, which would provide information about loans, grants, and development activities, has been growing since local elections in Kyrgyz. Communities would like to see this committee also as a means to hold local officials accountable if they fail to observe fair distribution of resources (Kudat et al 2003). From the donors' point of view, such a committee could also help reduce initial project costs and ensure sustainability of CDD initiatives after donors leave the country. New Information A light may be shone on critical problems or new opportunities through the availability of new information ­ whether information on problems, gaps, weaknesses or on strengths or potential. A new survey may reveal gaps e.g. the 1981 NABARD debt/investment survey that showed significant gaps in access to credit among the poorest. Another study in 1987 demonstrated the feasibility of linking self-help groups (SHGs) directly to pre-existing financial institutions and was the trigger for this mode of scaling up ­ that is, NABARD as apex organization focusing on catalyzing formation and strengthening of SHGs and linking them to the highly accessible network of commercial banking outlets in India. New information from outside may catalyze a drive toward scaling up (see "communicating past success" above) 30 Change in Context As discussed above, shifts in the enabling contextual environment for CDD have triggered/facilitated scaling up. Examples from the case studies include the Nepalese government stipulation of priority lending to most deprived sections, and a new government decree in Kyrgyz. Contextual changes may of course be negative too -- simultaneously increasing the urgency for scaling up while decreasing the capacity to achieve it. One prominent example here is the generalized HIV pandemic in southern Africa which has rendered it imperative to generate responses that match the breadth, depth and scale of the impacts of AIDS. The main district-level criteria for upscaling of the STEPS program were high HIV prevalence and absence of other NGOs aiming to combat it. Other contextual or institutional changes may be important. A donor review mission, a shift in donor philosophy, or a new donor coordination mechanism may all act as triggers or catalysts. In 1993 a donor mission to Zambia triggered the shift from a top-down food-for-work program (PUSH I) to a more process-oriented community-driven approach ­ in which actual activities were not specified -- that became PUSH II (the precursor to PROSPECT). PUSH I was seen as fairly ineffective as a safety net and non-empowering. The combination of exogenous and endogenous is exemplified by the case of the Kyrgyz Republic. Here, the CDD concept was originally initiated by donors, adapted by the government, then further improvised by community-based organizations, which themselves had formed around indigenous traditions of mutual self and collective action. 5.5 Scaling Up Processes Key approaches to scaling up are summarized for each case study in the Annex II tables. Multi-Dimensional Some or all of the different dimensions of upscaling described in the taxonomy in Table 1 above may apply simultaneously or sequentially. Indeed, CDD projects that only scale up in one or even two dimensions are rare. The Zambian case is interesting. In the second phase of PUSH II (1994-1997), CARE sought to move from direct implementation to facilitation of implementation and to building the capacity of the community to identify development needs and act to address them ­ that is, from first to second generation in terms of political scale (see Table 1). PROSPECT then represented a multidimensional scaling up of PUSH II quantitatively, functionally, organizationally, and politically. Most importantly for this study, PROSPECT is an instance where a donor (DFID) and an operational organization (CARE) scaled up CDD in ways that it became integrated into national and municipal structures. This has then allowed for widespread "replication" of CDD (Garrett 2003). 31 In Malawi, STEPS provides an interesting example of multi-dimensional scaling, moving through various phases with evolving roles, activities and donor relationships (Kadiyala 2003). The role of the donor itself (Save the Children) has changed from service delivery implementation to facilitating community mobilization. More recently STEPS has sought to form partnerships with other networks, including other NGOs as well as the government, to facilitate scaling. Specific partnerships may be developed to improve the coverage or quality of specific activities e.g. home-based care, where it partners with the Umoyo NGO network. Such partnerships were not envisioned in the first phase of STEPS. Again, the environment provided by the National Strategic Framework is conducive to such partnerships. Niches, comparative advantages and mutual benefits become clearer when organizations align with such national frameworks. The STEPS example also shows how scaling down ­ in this case of Save the Children's role and district-level presence -- may be required to create the space, and to devolve the power and resources, to enable CDD scaling (see "decentralization" above). Project or process ­ what is scalable? Indonesia's Kecamatan program sought to scale up principles of transparency, accountability and participation. Binswanger and Aiyar (2003) speak of scaling up the principles of empowerment, accountability and sustainability. The fact that, in these cases, functional scaling runs parallel to other forms of scaling up, suggests that it is principles, processes and simple rules that are `scalable' -- not activities per se or specific procedures which necessarily are governed by the local context from which they emerge. For example, in Kudumbashree, it is the principles and process that govern local planning that are scaled up through the CDD architecture, with the selected activities spinning off as they are prioritized. Scaling up is a learning process. Though Korten in 1980 referred to "learning to expand", perhaps `learning to grow' is more apposite, reflecting the development of capabilities and opportunities, as well as size. The link to learning is through the systems of monitoring, evaluation, communications and information dissemination, described above. Variable Sequencing and Pace Binswanger and Aiyar (2003) differentiate three consecutive phases of upscaling ­ initiation, scaling up and consolidation. ­ which are commonly seen, but do not always apply. For example, the NABARD program went through a piloting phase followed by a mainstreaming phase before entering into an expansion phase. Sequencing varies, as does the speed and the determinants of this pace. In Kerala, the pace was set by the government, while in Malawi communities themselves set the pace for scaling up. Certain components take longer than others ­ in PROSPECT, for example, the institution building component, which was inherently political, relying more on developing community relations, took a lot longer than the roll out of the water scheme. Upfront capacity development will, if done properly, take time, but this is an investment, as progress will accelerate later when these new capabilities are unleashed. 32 Anticipating and Resolving Trade-Offs For effective CDD to be effectively scaled up, there's a need for mechanisms for addressing some fundamental trade-offs. A primary trade-off is between the conventional need to demonstrate short-term impacts -- through pre-defined inputs achieving expected outputs -- and the need for impacts to be sustained, which requires a slower and more costly start-up as capacities are developed. This to some extent parallels the distinction between quantity and quality e.g. universalization of project coverage (scaling out) vs. consolidation and deepening (functional, organizational scaling). And how to weight the increased cost of outreach to very remote areas (which may also be the poorest) with that of reaching more accessible communities? There may be three-way trade-offs between, for example, quality, commitment/capacity and speed/scale (see Figure 6). It will be important to be aware of these and to understand that, at least initially, pushing on one or two of these objectives may affect achievement of the others. Figure 6: Possible Trade-Offs Quality Speed Spe Commitment Scale Capacity aci 5.6 Appropriate Forms of External Support to CDD How should external support change as scaling up proceeds? How can true CDD align with planning structures and budget cycles and output orientation of governments and donors? Should donors think in terms of exit or of transition? There is a real need to consider the process, not just the project, in scaling up CDD. The following quote of a Neighborhood Group member in Kerala sums it up well: "We know how to do thrift and credit. We know who to access for trainings and micro-credit and where. So even if Kudumbashree is not there we can manage." Part of the process comprises the development of capacity to demand resources when these are not available at community level (see above). 33 Uvin et al (2000) suggest returning to the notion of projects as policy experiments rather than solutions to problems in and of themselves. Donors need to balance "push" and "pull" factors. They need to guard against falling into the "supply-driven demand-driven development" trap of pushing communities too hard or too fast (Mansuri and Rao 2003) ­ the drive for quick time-bound results which compromises quality, capacity development and is anathema to a true "learning by doing" approach. Platteau and Abraham (2003) argue there is no escape from a gradual approach that refrains from disbursing quickly vast amounts of money over as many communities as possible. Too much "push" increases the risk that local leaders will seek to gain legitimacy from the outside world rather from their own people, possibly increasing the risk of malevolent elite capture. Outside money may "corrupt the process of local institutional development by allowing leaders to eschew negotiation with members for support and material contributions, thereby preventing autonomous organization-building based on total accountability of leadership vis-a-vis the members" (Platteau and Gaspart 2003). Kadiyala (2003) comes to a similar conclusion with regard to STEPS in Malawi: "how finances are channeled can make or break community ownership". Investing in thorough institutional and human capacity strengthening is a vital and often time-consuming first step. External resources could then be released according to observable milestones in the development of absorptive capacity. Garrett (2003) suggests support might be divided between capacity development, including participatory community-driven planning, and funding support for the implementation of community proposals. Flexibility, adaptability and openness to change applies to all stakeholders in CDD processes. The "learning by doing" culture of CDD involves a lot of "learning by making mistakes" and is costly. How much funding and time are important concerns for governments and donors who support CDD. Learning-by-doing is also antithetical to a traditional project cycle with its linear progression from problem identification to project preparation, appraisal, implementation, supervision and evaluation. Such a project cycle assumes that solutions to known problems can be fully determined and costed in advance and implemented to a fixed timetable. The logic of CDD runs counter to this. But there are positive signs of these considerations being taken on board in the re-aligning of support systems and structures. Ultimately, more and better evaluations and better communications will foster learning and help reduce these constraints. When considering appropriate roles for donors, the notion of `phase-out' is dangerous terminology. In PROSPECT, for example, at perhaps the most crucial time in transition -- when the need for supporting and training local structures to be sustainable was greatest, was just when many of the most capable staff left, fearing future redundancy as CARE pulled out (Garrett 2003). In an external evaluation of STEPS in Malawi, Williamson and Donahue (2001) recommended "phase down, not phase out". Norman Uphoff has stated that the concept of withdrawal or phasing out in development programs is almost a violation of the very principles and values that make them successful in the first place (IIRR 2001). Mathie and Cunningham (2003) suggest the need for donors to be present with a light touch -- both "hands on" and "hands off". The notion of transition -- not 34 withdrawing but including people in a larger movement -- is more useful than that of exit or phase out. 6. Conclusions As an earlier World Bank review concluded: since the success of CDD is crucially conditioned by local cultural and social systems, it is best done not with a wholesale application of best practices applied from projects that were successful in other contexts, but by careful learning-by-doing. "The best practice here may be the absence of a best practice" (Mansuri and Rao 2003). While this synthesis has described some of the more important features of successful CDD scaling up, it is not possible to rank these in any way. Some are more important in certain contexts, some less so. Many interact synergistically. With such interactive systems it is not possible to make useful generalizations about the relative importance of one factor versus another. Rather than ranking, it is more useful to view these features in toto as guidance ­ in terms of what to strive for (with new CDD initiatives) or what to look for (in evaluations). Box 3 provides a summary of key lessons. Box 3: Summary of Key Lessons · Overall, keep in mind context, institutional arrangements, capacity, and the triggers and different processes of scaling up · Donors and governments need to think of the process beyond the project, of transformation or transition, not exit. Need to balance push and pull factors and avoid "supply-driven demand-driven development". Community-driven is citizen-driven, not client-driven. · Capacity is pivotal and is more than simply resources; it includes motivation and commitment which necessitates appropriate incentives at all levels. Capacity development takes time and resources but is an essential investment. The capacity and commitment of facilitators and local leaders is particularly important. · Learn by doing -- and by communicating, monitoring, evaluating and changing. Learn from failure, but learn faster from success. Start with the positive (what's working) not the problem (what isn't) and build on that. Be adaptive, flexible and open to change. Anticipate and address trade-offs. Apply realistic time horizons (10-15 year, not 5 year cycles) · Build library of well-documented context-specific experiences through good monitoring, evaluation and operational research. Use these to advocate for improvements in the contextual environment. · To sustain CDD, anchor it within existing contextual systems (government), frameworks (e.g. PRSP) and processes (decentralization) even where these may be imperfect. Ultimate aim is to weave and embed sustainable CDD in national social, political, cultural and institutional fabric. The CDD paradigm is about adaptation and learning rather than mere replication. It values diversity over standardization, and promotes synergy rather than substitution or competition. Ultimately it is about processes and principles, not projects per se -- processes that need to become anchored in national policy frameworks and embedded within the social, cultural and institutional fabric of the country. 35 References Abraham, A and J. P. Platteau. 2000. Participatory development in the presence of endogenous community imperfections. Belgium: University of Namur-Department of Economics. Adato, M., Hoddinott, J and Haddad, L 2003 Power, Politics, Efficiency and Effectiveness: benefits and Challenges of Community Participation in South Africa's Public Works program. IFPRI Discussion Paper, forthcoming. Alkire, S et al. 2001. Community- driven development. In PRSP Source Book. Washington DC: World Bank. Binswanger, H. and Aiyar, S. (2003) Scaling Up Community Driven Development: Theoretical Underpinnings and Program Design Implications, February 2003 draft, World Bank. Cooke, B. and Kothari, U. (eds) 2001 Participation: The New Tyranny? London: Zed Press. Cornwall, A. 1996. Toward participatory practice: Participatory rural appraisal and the participatory process. In Participatory research in health: Issues and experiences, ed. K. de Konig and M. Martin. London: Zed Books. Appearing in B. de Negri, E. Thomas, A. Illinigumugabo, I. Muvandi. Empowering communities: Participatory techniques for community-based programme development, Volume I: Trainers Manual. June 1998. Das Gupta, M. Grandvoinett, H. and Romani, M (2003) Fostering Community-Driven Development: What Role for the State? World Bank Policy research Working paper 2969, World Bank. Edwards, M. and Hulme, D. 2002 Making a Difference: NGOs and Development in a Changing World., Earthscan Publications, London. Galasso, E and M Ravallion. 2000. Local knowledge vs local accountability. Washington DC: World Bank. Garrett, J 2003 Community Empowerment and Scaling up in Urban Areas: The Evolution of PUSH / PROSPECT in Zambia, IFPRI, Washington DC. Gillespie, S.R. 2001 Strengthening Capacity to Improve Nutrition. IFPRI Discussion paper 106 Gupta, D. B. and A Gumber. 1999. Decentralization: Some initiatives in health sector. Economic and Poltical Weekly February 6. 36 Hancock, J. (2003) Scaling Up the Impact of Good Practices in Rural Development: A working paper to support implementation of the World Bank's Rural Development Strategy. Report No. 2031, 8 June 2003, World Bank. Heaver, R. 2002 Improving Nutrition: Issues in Management and Capacity Development, HNP Working Paper, World Bank. Heller, P. 2001 Moving the State: The Politics of Democtaric Decentralization in Kerala, South Africa and Porto Alegre. Politics and Society 29 (1), 131-63. Iannnotti, L and S Gillespie. 2002. Successful community nutrition programming: Lessons from Kenya, Tanzania and Uganda. Washington DC: AED, USAID, UNICEF and Regional Center for Quality of Health Care, Makarere University, Uganda. International Institute of Rural Reconstruction IIRR, 2001. Going to Scale: Can We Bring More Benefits to More People More Quickly? Workshop Highlights Presented By The CGIAR ­ NGO Committee and The Global Forum for Agricultural Research with BMZ / MISEREOR / Rockefeller Foundation / International Rice Research Institute / International Institute of Rural Reconstruction www.ngoc- cgiar.clades.org/index0.html Kadiyala, S. 2003 Scaling-up HIV/AIDS Interventions Through Expanded Partnerships (STEPs) in Malawi, IFPRI, Washington DC. Kadiyala, S. 2003 Kudumbashree: Collective Action for Poverty Alleviation and Women's Empowerment, IFPRI, Washington DC. Korten, David. 1980. "Community Organization and Rural Development: A Learning Process Approach" A Ford Foundation Reprint from Public Administration Review. Kudat, A., Kokten, M and Omuraliev, N 2003 Community Driven Development in the Kyrgyz Republic: Evaluating the Potential for Scaling Up, World Bank, draft. Kumar, S and Corbridge, S (2002) Programmed to Fail? Development Projects and the Politics of Participation. Journal of Development Studies 39 (2), p73-103. Mansuri, G. and Rao, B. 2003 Evaluating Community Driven Development: A Review of the Evidence. February 2003 draft, World Bank. Mathie, A. and Cunningham, G. 2003 Who is Driving Development? Reflections on the Transformative Potential of Asset-Based Community Development, Coady International Institute, Canada. (www.stfx.ca/institutes/coady) Myer, R. 1984 Going to Scale. A Paper prepared for UNICEF for the Second Inter- Agency Meeting on Community-based Child Development, New York, October 29-31, 1984. 37 Narayan, D. 1997. Designing community based development. Social Development Working paper No. 7, World Bank. Narayan, D. and Petesch, P. (eds) 2002 Voices of the Poor: From Many Lands. World Bank and Oxford University Press. Paul, S. 1992. Accountability in public services: Exit, voice and control. World Development 20(7): 1047-1060. Platteau, J.P and Abraham, A. 2003. Participatory development in the presence of endogenous community imperfections. Journal of Development Studies 39 (2), p104-36. Platteau, J.P. and Gaspart, F. 2003 The Risk of Resource Misappropriation in Community-Driven Development. World Development 31 (10), 1687-1703. Schonwalder, G. 1997. New democratic spaces at the grassroots? Popular participation in Latin American Governments. Development and Change 28: Sharma, M. 2003 Scaling Up Microfinance in South Asia, IFPRI, Washington DC. Summers, Lawrence (2001) Speech at World Bank Country Director's Retreat, June 2001. As cited in Rao and Mansuri (2003) UNAIDS. 2001. Reaching out, scaling up. UNAIDS Best Practice Collection . Geneva: UNAIDS. UNICEF 1990. Strategy for improved nutrition of children and women in developing countries. UNICEF, New York. Uphoff, N. 1988. Assisted self-reliance: Working with, rather than for, the poor. In Strengthening the Poor: What have we Learned?, ed. J. P. Lewis. Overseas Development Council, Washington D.C. Uvin, P. 1999. Scaling up, scaling down: NGO paths to overcoming hunger. In Scaling up, scaling down: Overcoming malnutrition in developing countries, ed. T. J. Marchione. Gordon and Breach. Uvin, P, P. S. Jain, and L. D. Brown. 2000. Think large and act small: Toward a new paradigm for NGO scaling up. World Development 28(8): 1409-1419. Uvin, P. and Miller, D. 1994. Scaling Up: Thinking Through the Issues. The World Hunger Program www.globalpolicy.org/ngos/role/intro/imp/2000/1204.htm Veron, R. 2002. The ' new" Kerala model: Lessons for sustainable developement. World Development 29(4): 601-617. 38 Walt G., M. Perera, and K. Heggenhougen. 1989. Are large-scale volunteer community health worker programmes feasible? The case of Sri Lanka. Social Sciences and Medicine, 29 (5). In Jennings et al. (1991) Managing successful nutrition programmes. ACC/SCN State-of-the-Art Series, Nutrition Policy Discussion Paper No. 8. Williamson, J and J Donahue. 1998. Community mobilization to address the impacts of AIDS: A review of the STEPS II program in Malawi. Washington D C: Displaced Children and Orphans Fund. Wong, S 2003 Indonesia Kecamatan Development Program: Building a Monitoring and Evaluation System for a Large-Scale Community-Driven Development Program, Environment and Social Development East Asia and Pacific Region Discussion Paper, World Bank. World Bank (1998) Participation and Social Assessment: Tools and Techniques, World Bank, Washington DC World Bank. 2000. Scaling up community driven development for poverty reduction World Bank. 2000. Lessons on community-driven development. Lessons and Practices No. 12. Washington DC: World Bank-Operations Evaluation department. World Bank. 2001. Decentralization and governance:Does decentralization improve public service delivery? PREM Notes No. 55. Washington DC: The World Bank. The World Bank - Participation Sourcebook http://www.worldbank.org/wbi/sourcebook/sbhome.htm World Bank 2000 Community-Driven Development - Basic Concepts & Principles. (and case studies) World Bank http://wbln0018.worldbank.org/essd/CDDWk2000.nsf/GWeb/cdd_concepts 39 ANNEX I: CASE STUDY ABSTRACTS 1 Community Empowerment and Scaling up in Urban Areas: The Evolution of PUSH / PROSPECT in Zambia (James Garrett) CARE began PROSPECT (Program of Support for Poverty Elimination and Community Transformation) in 1998. PROSPECT aims to reduce poverty in peri-urban areas of Lusaka. It employs a community-based approach to carry out three types of activities: social empowerment (institution building at the local level), personal empowerment (microfinance), and infrastructure improvement (mostly water supply schemes). PROSPECT has attempted to carry out these activities largely through its support of area- based organizations (ABOs) that now form part of city government. The Zone Development Committees (ZDCs) and Residents' Development Committees (RDCs) are the basic components of the ABO structure. These are community-level representations of municipal government; they are the community's mechanisms for expressing its voice and driving development. PROSPECT is itself an extension of an earlier project, PUSH II (Peri-Urban Self-Help Project). PUSH II and PROSPECT are fundamentally about developing community- based and community-driven development (CDD) mechanisms and strengthening community capacities to identify and respond to community needs. The paper examines the scaling-up experience of PUSH II and PROSPECT, looking especially at the mechanisms of CDD, the ABOs. Evolution of PUSH and PROSPECT PUSH I (1991-1994) was a food-for-work (FFW) program sponsored by the Government of Zambia, the World Food Program (WFP), and the Canadian International Development Agency (CIDA). The objective of PUSH was to alleviate the negative effects of structural adjustment and stabilization and the 1991-92 drought on well-being and food security. In the second phase of PUSH, PUSH II (1994-1997), CARE sought to move from direct implementation to facilitation of implementation and to building the capacity of the community to identify development needs and act to address them. A new donor, the U.K.'s ODA (Overseas Development Administration) and CARE's senior staff encouraged this more community-driven approach. PUSH II initiated work in 3 compounds in Lusaka and was a learning and "process" project. That is, communities determined project interventions only after initial participatory assessments. PROSPECT scaled up from the 3 compounds of PUSH II to 11. After the initial "process" orientation of PUSH, while maintaining a community-based focus, PROSPECT began to revert to a more traditional project arrangement. PROSPECT's proposal, for instance, specifically identified water provision and microfinance as interventions. 40 Facilitating Factors and Challenges Although supposedly water supply was to be simply an entry point for more in-depth work to build up the ABOs and encourage CDD, CARE's efforts to follow up on this have been less than optimal. The two principal factors that militated against this seem to have been 1) political conflict between the ABOs and the traditional hierarchical power structures and 2) the difficulty of structuring organizational incentives to encourage consistent long-term actions to support CDD. Political tension is almost inevitable as CDD scales up. The spreading of the RDCs threatened the authority of the City Councilors, who traditionally had been in charge of development in the compounds (although not necessarily very effective). With PROSPECT, residents saw the RDCs managing substantial resources and promoting investment beneficial to the community. The RDCs were an emerging threat to the councilors' status. In 1999 the Council suspended further RDC elections. In response, CARE worked with stakeholders to develop a new legal framework for RDC and Council relations. Elections only resumed in mid-2002, so for almost 3 years the ABO structure was not fully operational. (Some project activities continued as the suspension only affected new elections, but the conflict cast a pall over ABO, CARE and Council relations.) The result of the suspension has been a more robust legal framework for the ABOs and, as a significant step forward for scaling up of CDD, integration of the ABOs into the municipal government structure and its spread to all peri-urban compounds in Lusaka (whether CARE works in them or not). Still, the crisis showed how significant an obstacle politics can be to CDD and scaling up. The political conflict between the ABOs and the City Councils illustrates the necessity of involving government structures, rather than ignoring them, in promoting CDD, especially in urban areas. To resolve the crisis, PROSPECT worked with the City Council and ABOs so that everyone understood their roles. CARE's position as important international NGO with resources to fund the dialogue and to meet some community needs possibly allowed them to play that mediating role in a way that a less prestigious or less well-funded organization could not. Scaling up CDD may thus require the prestige and resources of an outside catalytic change agent and almost certainly requires that stakeholders' and participants' roles and responsibilities are clarified from the beginning and quite possibly "officialized" in some way. The history of PUSH II and PROSPECT also illustrates the organizational difficulty of supporting CDD. Despite PUSH II's initial emphasis on CDD and empowerment, PROSPECT over time came to focus more on promoting specific interventions (water provision, microfinance) than on energizing CDD in a sustainable way. This is not entirely surprising given the genuine community need for water, external evaluations that encouraged water schemes, and donors that stressed outputs over process. As a result, the focus shifted from process to specific activities of previously determined project components. PROSPECT's own monitoring and evaluation system also focused more on outputs or their impacts, rather than on indicators of effectiveness and sustainability of 41 ABOs. Efforts to support CDD thus will require changes in project design and performance indicators, and in donor perspectives and expectations. Other factors were also important in shaping CDD. Democratization and a national policy of decentralization, along with encouragement from key CARE staff, encouraged a community-based approach to development. Project management was careful to train staff in participatory methods. This was important for them to understand how to promote CDD. Project and mission management's own insistence on decentralizing authority and empowering staff further enhanced this thrust. All this created a national and institutional environment important to promoting CDD. Funding continues to be a severe challenge to sustainability. Neither Council nor ABOs have adequate resources to build capacity or implement projects to meet identified needs. Without resources to meet community demands, CDD becomes a hollow exercise and structures will wither. The evolution of PROSPECT suggests that the nature of support to CDD will change over time. The promoting organization therefore must be a learning and empowering organization. Initial promotion of CDD may involve establishing local structures and working to create an enabling environment, but supporting organization must move to build their capacity on a number of fronts, including needs identification, project design, fundraising, and project management. As community, municipal, and national structures become stronger, the promoting organization will need to move from the center to the side. It may still be involved in strengthening organizations and shaping the enabling environment, but the skills and activities required to do this are very different from those required at the beginning, when the institutional landscape is more barren. The promoting organization itself must be ready for this, and think in terms of transforming its support (its activities, funding focus, and staff numbers and skills), rather than simply "shutting down" or "exiting." The study also finds that scaling up occurs rather naturally along a number of different dimensions over time. In PROSPECT's case, not only were the ABOs ultimately integrated into the social and institutional fabric of urban government in Lusaka, PROSPECT's approach had impact beyond the project itself. Seeing a need for project learning and documentation, CARE instituted a new project, Urban INSAKA. CARE's C3 is a relatively new project that, like the Zambian Social Investment Fund, funds community initiatives, thereby supporting CDD with responsive resources. CULP has repeated PROSPECT's approach in 3 more cities in Zambia, while Mahavita has transferred the experience to Antananarivo in Madagascar. These last examples of scaling up depended to a great extent on the international reach of CARE and key advisors who took the experiences, structures, approach, and lessons from PROSPECT and transferred them to other locations. 42 2 Scaling-up HIV/AIDS Interventions Through Expanded Partnerships (STEPs) in Malawi (Suneetha Kadiyala) This paper discusses the factors that are enabling and constraining the scaling-up of Scaling-up HIV/AIDS Interventions Through Expanded Partnerships (STEPs)2 initiative, supported by Save the Children U.S (SC), to combat HIV/AIDS in Malawi. It also discusses some potential threats to and contextual factors limiting scaling-up of STEPs. This report draws primarily upon the available literature and qualitative data collected during a five-day visit to SC Malawi in December 2002. STEPs started in 1995 (then called Community -based Options for Protection and Empowerment, STEPS) as a service-delivery program in one district in Malawi, Mangochi, to assist children affected by HIV/AIDS. Through evaluations, SC realized such an approach was unsustainable, cost-ineffective and not scalable. Based on the recommendations of the evaluations and the field experience, the program revitalized the dormant decentralized AIDS committees (at district, community and village levels) and their technical sub-committees, under the National AIDS Commission (NAC), in the Namwera community in Mangochi to mobilize collective action to combat the epidemic. Based on the positive experience in Namwera, the program changed its initial strategy to that of an external change agent, assisting communities with community mobilization and capacity building so that communities become empowered to act collectively to address their own problems. Village AIDS Committees (VACs) identify the vulnerable; plan responses on the basis of the nature and magnitude of vulnerability within the villages, needs of the vulnerable and capacity within the villages to respond. They also monitor the activities and mobilize resources. As the needs of the most affected communities are crosscutting, the program evolved into a truly multi-sectoral program with activities along the continuum of prevention, care, support and mitigation. STEPs has also been influencing national policies related to HIV/AIDS and children. STEPs now functions in four districts and aims to expand to two more, covering a total of 15% of Malawi's population by 2005. Through partnerships and training other NGOs/ CBOs in the program approach of community mobilization and facilitating collective action, STEPs and similar models aim to cover 75% of Malawi's population. Contextual factors critical for scaling-up include an enabling policy environment with a strong commitment of the current government, especially NAC, to a multisectoral approach of combating HIV/AIDS. Organizational factors enabling scaling-up include: a well trained and motivated staff, adoption of a community mobilization model through capacity building of the district, community and village AIDS committees, its commitment to document and disseminate lessons learnt; and reaching more affected 2 2Until June 2003, STEPs was known as Community -based Options for Protection and Empowerment (STEPS). However for the sake of legal issues involved in the use of the acronym STEPS due to copyrights, we use STEPs throughout the paper. However, the readers should interpret STEPs as STEPS before June 2003 43 populations through partnerships. Factors specific to communities include, leadership within the community, whether they are urban or rural (with rural communities more easy to mobilize than urban), the nature of livelihoods, history and culture of the communities with respect to collective action. Planning along with the communities for a phasing down of SC's presence and scaling-up of the role and responsibilities of the AIDS committees and funding mechanisms that do not threaten community ownership of the problem and response have also been identified as critical in enabling and scaling-up sustainable collective action. Important factors that threaten or limit the scaling-up of STEPs include: the magnitude of the epidemic eroding community resources; the current food crisis that is diverting resources to sheer survival; the gap between the resources that communities need to make a visible change and what they have access to consequently, undermining the spirit of volunteerism; weak commitment of donors to a truly community driven multisectoral response; and the overall context of poverty and under development in which it takes longer to mobilize communities and build their capacities. 3 Kudumbashree: Collective Action for Poverty Alleviation and Women's Empowerment (Suneetha Kadiyala) This paper discusses the factors that enabled and constrained the scaling up of a multi- sectoral poverty alleviation program called Kudumbashree, initiated by the government of Kerala (GOK) in 1998 to eradicate poverty by 2008. It also discusses some potential threats to, and trade-offs, of scaling-up Kudumbashree. This report draws primarily upon the available literature and qualitative data collected during a five-day visit to Kudumbashree in March 2003. In 1991, the GOK, along with UNICEF, initiated the Community-based Nutrition Program (CBNP) in Alleppey town to improve the health and nutritional status of children and women. CBNP facilitated collective action, by forming and developing the capacity of three-tiered community development societies (CDS), the members of which are exclusively women. Women from families identified as poor, using a nine ­point non-income based index, were organized into neighborhood groups (NHGs) comprising 20-45 families. Each NHG member elected a five-member committee called the NHG committee (NHGC) to coordinate and facilitate action at the NHG level. The NHGs were federated at ward level as an Area Development Society (ADS). The ADSs were then federated at the municipal level as a CDS. Based on the positive experiences in urban Alleppey and subsequently in rural Malappuram, the GOK scaled up the CDS strategy to the entire state in 1998 under the name Kudumbashree. The State Poverty Eradication Mission implements Kudumbashree through the department of Local Self Governments (LSG), formed and empowered in 1992 by the 73rd and 74th Constitutional Amendments. Convergence of various government programs and resources at the CDS level, participatory anti-poverty planning and implementation, formation of thrift and credit societies and development and nurture of micro-enterprises are the key Kudumbashree strategies. 44 Our findings show that an enabling environment, especially decentralization and the concurrent devolution of finances to the local government bodies (LGBs), was critical in scaling-up Kudumbashree. The CDS structures are now considered as a further step to decentralization. As the CDS structures are affiliated to the LGBs, their financial sustainability is ensured through various modalities, e.g. through convergence, Women's Component Plan and ear marked assistance to women's self-help groups (SHG). The unique context of Kerala coupled with leadership of a few motivated and innovative officials was key in both the decentralization process as well as the scaling-up of Kudumbashree. Kudumbashree itself is an inter- departmental initiative making it conducive for a multi- sectoral response to poverty. Other institutional arrangements such as partnerships with the central government and the National Bank for Agriculture and Rural Development (NABARD) continue to play a role in an expanded response to poverty alleviation through CDS structures. Training and capacity building of the LGBs and the CDS/ADS/NHGs was critical in building ownership. Two factors initially constrained the scaling-up process. The government initially wanted to include all women, not just those below the poverty line in the CDS structures. The negotiations between various stakeholders delayed the process by a year and half. Many LGBs resisted strengthening of the CDS/ADS/NHGs, as they perceived them to be a threat to their authority. This still continues to be a problem in many gram panchayats. The potential threats and trade-offs include a shrinking focus on maternal child health and nutrition issues with an increasing preoccupation with micro enterprise initiatives. The quality of collective action needs to further strengthened. The current plans tend to be a catalogue of individual needs with group needs often not addressed, and there are signs that the NHG/ADS/CDSs are becoming bureaucratic. 4 Scaling Up Microfinance in South Asia (Manohar Sharma) This case study examines the scaling up experiences of two microfinance institutions: the Nirdhan Utthan Bank Limited (NUBL) in Nepal and the Self-Help Group (SHG) ­ Bank linkage program of the National Agricultural Bank for Agriculture and Rural Development (NABARD) in India. Both NUBL- and NABARD-groups use self-regulation (peer selection, peer monitoring and peer enforcement of contracts) as key to access services otherwise not accessible to them. There are two community-based drivers. Loan products are heavily driven by clients' preferences as evidenced by strong demand to join the program, high repayment rates, and very low drop out rates. Second, the process of organizing clients into groups itself has a significant empowering effect, providing voice ­ and the attendant bargaining power - to an impoverished class, when there was none before. Standardization of rules of conduct and basic service delivery mechanisms (and also financial products in the case of NUBL) was key to swift replication in both India and 45 Nepal. In Nepal where the density of banking services was low, NUBL chose to be the financial service provider itself. In India, where this density was already very high, NABARD recognized the core advantages of group-based finance, but decided to facilitate linkage banking instead. The NABARD experience is government-led. NUBL, on the hand, was established as an alternative to government action. However, in both cases, government policy in the form of mandatory "priority sector" credit played ­ and continues to play - a critical role in facilitating expansion. The subsidy content (explicit and implicit) of both NUBL and the NABARD program is quite high, and continued expansion of both programs is highly conditional on whether the policy regime of directed credit continues in the future. Any change in this policy will deal a severe blow to both these institutions. Provisioning group-based credit is costly in that it is highly staff time-intensive. In the case of NUBL, staffing requirements are so high that it has so far not been possible to scale-up services in remoter and more sparsely populated mountainous areas of Nepal. In India, expansion of services in the remoter northeastern states has been hindered by the high cost of setting up and operating SHG promoting institutions. One option in both countries is to induce the development of group federations that eventually become self- financing and self-regulating. Instances of well-functioning group federations are emerging in parts of India and federations may well be the key to consolidating the gains made so far, in ensuring that the programs are primarily driven by the interests of clients, and to an eventual weaning out of subsidies. Finally, quality of the broader national environment is very important in facilitating growth of institutions. NUBL's growth leveled off exactly when expansion of SHGs accelerated in India. This was not a coincidence. The Maoist insurgency in Nepal severally restricted the development of microfinance sector while the supporting environment in India facilitated its unparalleled expansion. 5 Community Driven Development in the Kyrgyz Republic: Evaluating the Potential for Scaling Up (Ayse Kudat,Muge Kokten Finkel and Nurbek Omuraliev) Thos case study discusses issues for scaling up in light of the lessons learned from rural community driven initiatives in the Kyrgyz Republic. In order to compare the processes, and to analyze the system wide and household level impacts of alternative CDD projects, the study focuses on two programs: UNDP's Decentralization Program and DFID's Sustainable Livelihoods Program. The study demonstrates that there are important similarities between the processes imbedded in the pilots of the two programs. Both programs adopt a multi-sectoral approach to development, supporting social mobilization for micro-credit, providing advisory services for agricultural and livestock production, and supporting individual or group based enterprise activities. Both projects try to strengthen CBOs and their capacities for business planning and project implementation. However, they differ in their core principles and their institutional arrangements, which in turn, affect not only the steps of their impacts but also their potential for scaling up nation-wide. The UNDP 46 program aims to bolster transparent and accountable local governments that are able to work with their citizens in a sustainable manner with locally owned and generated resources. Hence it works with CBOs, local governments and their representation at the national level, the Congress of Local Communities, and installs incentives for cooperation between all three parties. It also has a higher level of commitment with respect to both financial and human resources. As a result, the program's procedures are fully incorporated into the national institutional structures and policies, providing a solid chance for national scaling up. In comparison, the DFID program bypasses local governments and works directly with villagers through elected `working groups'. It experiments with innovative techniques for farming, livestock breeding, and collective resource management to improve the livelihoods of the communities, hence it emphasizes capacity building and shared systems of information, making these experiences significant for efforts to scale up CDD. The Kyrgyz experience concludes that several facilitative national and contextual factors are crucial for scaling up CDD successfully, including the political commitment to decentralization, a positive agrarian economic environment supporting micro-credit and village based development initiatives, a relatively educated labor force, and most importantly indigenous self-help mechanisms and previous exposure to community based projects. Programs that utilized existing mechanisms of self-help and social mobilization have been more successful in expanding the reach of their projects and keeping their costs down. Donor coordination and experience sharing have also had a positive impact for scaling up, as projects can build on each others experiences. Institutional innovations, such as the Local Development Fund in the Kyrgyz Republic, have also been very effective when there exists a serious commitment from donors and institutional support from the national policymaking processes and authorities. At the same time, the Kyrgyz experience also highlights continuing challenges to successful CDD scaling up in the future. Among them are incomplete efforts of fiscal decentralization, prevailing high rural poverty rates, excessive regulations concerning the activities of CBOs and bureaucratization of micro-credit applications, and the absence of crediting procedures that reward and utilize existing and generated social capital through collective projects. 47 ANNEX II: CASE STUDY SUMMARY TABLES PROSPECT in Zambia INSTITUTIONAL Funder: PROSPECT: ODA (now DFID); PUSH II: also WFP ARRANGEMENTS ABOs (RDCs and ZDCs), the focus of this study, are responsible for overall community development and liaison with Lusaka City Council (LCC) CAPACITY ELEMENTS Training of ABOs, councilors and council staff (leadership, democracy, conflict resolution, gender, etc.) Working with councilors and council staff to develop policies and programs TRIGGERS ODA review mission and focus on process projects CARE's desire to continue work in periurban areas once PUSH I ended CARE senior staff commitment / background in participatory development FACILITATING FACTORS National policy of decentralization, need for community representation Need to bring peri-urban areas under municipal authority Recognition that national / municipal government did not have resources to supply services Donor (funding, philosophy) and country mission support Capacity building among staff in participatory methods / learning organization partnered with government Had resources available Prestige and "weight" of CARE Began smaller and grew over time LIMITING FACTORS / Government disinterest CHALLENGES Political obstacles Donor frameworks and requirements Focus on outputs, budget setups and cycles, M&E indicators Resources to respond and build capacity Staff capacity and skills Sustainability of focus on community over time Transformation and transition Staff, funding lines, stakeholders, activities Weaving sustainable CDD into social and institutional fabric Spreading to other SCALING UP PROCESSES Quantitative Expanded operations from 3 to 11 compounds Scaling Up Facilitated integration of ABO into municipal governance structure for all compounds Assisted in finding funding for ABO initiatives Functional Added Environmental Health, Gender, and HIV/ AIDS components Scaling Up Linked with other projects like KAR (urban waste management program) ABOs constructed clinics and police posts and community centers Expanded down to create and support ZDCs Supported creation of ABO federation Expanded up to build capacity among council staff and councilors, and sit on national boards Political Service delivery: FFW built infrastructure (PUSH I) Scaling Up Community capacity development: redirected ABOs Policy reform: integration of ABOs into local governance structure; member of NDCC; promotion of federation of ABOs Social movements / leadership entry into politics: some RDC members have become councilors Organizational Cross-training and cross-sectoral coordination of CARE staff Scaling Up Training of ABOs and LCC members, council staff Connecting ABOs to water schemes for revenue Assisting with finding donors, writing proposals 48 Spinoffs and incorporation of RDCs in other structures STEPS in Malawi INSTITUTIONAL Partnership with DACC/CAC/VAC is central to the STEPS strategy ARRANGEMENTS Partnership with other NGOs/CBOs to build their capacity in initiating STEPS like models in other districts; to leverage their presence within the same district Partnership with NAC and capacity building of CBOs is the primary mode of scaling-up in the next phase of STEPS CAPACITY Community mobilization has been critical in building capacity ELEMENTS Training at various levels (DACC/CAC/VACs) critical Funding issues continue to constrain capacity building of the AIDS committees TRIGGERS Few visionaries in SC The success community mobilization model in Namwera was a trigger to expand the strategy to other districts Within districts communities themselves have been catalysts FACILITATING Decentralization of AIDS committees to district, community and village levels FACTORS Working through the existing structures An enabling policy environment Institutional arrangements SC's ability attract, retain high quality staff and keep them motivated through trainings, encouraging innovation and competitive salaries LIMITING Inadequate training of the DACC/CAC/VAC members FACTORS Lack of adequate funding Other factors such as the magnitude of the epidemic, the ongoing food crisis, weak commitment of donors, lack of a full time DACC coordinator and the over all context of poverty and underdevelopment SCALING UP PROCESSES Quantitative The community mobilization model, revitalizing DACC/CAC/VAC structures Scaling Up replicated in the four districts Other NGOs are also adopting a similar strategy in other districts of Malawi and elsewhere Functional From primarily a psychosocial orphans support project, STEPS evolved into a multi- Scaling Up sectoral initiative addressing issues along the continuum of prevention, care, support and mitigation Political Through STEPS, SC is influencing national policies: It was a member of the National Scaling Up Orphan Task Force of Orphan and played a key role in drafting guidelines for orphan care in Malawi; leading member of Children and Violence Task force; formation of Wills and Inheritance Core group Organizational As STEPS scaled-up it recruited a HBC officer and an M& E officer. Plans for Scaling Up recruiting additional staff to scale-up. STEPS is scaling up partnering with the organizations, including the government. 49 Kudumbashree in Kerala, India INSTITUTIONAL An inter-departmental government initiative ARRANGEMENTS Partnership with UNICEF was critical in the initial stages Strong partnerships with LGBs, various line departments, the Central government, NABARD and other commercial banks Uneasy relationship with the NGOs CAPACITY The nature of the initiative (cutting across departmental lines) enables multi-sectoral ELEMENTS action Training at various levels (Kudumbashree staff, LGB leaders, CDS/ADS/NHG) critical Financial sustainability through convergence, WCP TRIGGERS Alleppey and Malappuram CNBP were triggers FACILITATING Decentralization and fiscal devolution was the catalyst in setting pace FACTORS Leadership of a few motivated government officials TCS and microeneterprises LIMITING Inadequate training at the LGB and CDS/ADS/NHG levels to build ownership due to FACTORS the rapid and mandatory pace of scaling- up Political interference and bureaucracy of the CBOs themselves SCALING UP PROCESSES Quantitative CDS/ADS/NHG structures replicated in the entire state Scaling Up Functional A multi-sectoral initiative with an increasing emphasis on reaching the destitute and Scaling Up children. From primarily a nutrition initiative, it scaled up to include micro-enterprises etc. But scaling-up to other activities has reduced the focus on maternal and child health and nutrition Political CDS/ADS/NHG structures are considered as a further step to decentralization Scaling Up CDS are considered to be the "wings" of the LGB Organizational Kudumbashree was set up in 1998 with the explicit function of undertaking the program Scaling Up to alleviate poverty in Kerala 50 Microfinance in Nepal and India Self Help Group ­Bank Linkage Program of the National Bank for Nirdhan Utthan Bank Limited (NUBL) Rural Development (NABARD), India INSTITUTIONAL Incorporated as Development Bank. Subject to Self-help groups (SHG) are informal, but banks are authorized to ARRANGEMENTS prudential banking laws and supervision by conduct formal financial transactions with them. Nepal's central bank. Governed by a Board of NGOs or commercial banks help form these groups. Directors selected from among shareholders. NGOs receive financial assistance from NABARD, and Expected to issues shares to clients in near future commercial banks may refinance loans to SHGs from NABARD and representation of clients on Board expected at a special rate. after that. Currently capitalized at Nepalese Rs NABARD created in 1982 by a Special Act of parliament and 10 million. jointly owned and funded by the government of India and India's Provides group-based loans to targeted clients central bank. NABARD is an apex institution, accredited with all matters concerning policy, planning and operations in the field of credit for agriculture and other economic activities in rural areas in India. It is capitalized at Indian Rs20 billion. CAPACITY Professional senior management staff. Recently set up Microfinance development fund to scale up ELEMENTS Field workers hired locally and trained. linkage program. Staff capacity in training and promotional Currently developing Management Information activities adequate. System critical for further expansion. TRIGGERS Personal commitment of founder and CEO who Debt-investment survey of 1981 showing inadequate access to quit position as deputy governor in Nepal's banking services by the rural poor triggered experimentation and central bank to start NGO banking services. piloting. Establishment of credit-based NGO only Demonstration of success by Grameen Bank in neighboring possible after democratic rule started in 1990 Bangladesh led to adoption of group-based approach. Used Grameen methodology that was successful High density of pre-existing branches of commercial banks led to in neighboring Bangladesh adoption of the linkage approach. Donor support based on demonstration of sound management and poverty outreach. FACILITATING Banking law requiring commercial banks to Nationalized banking system heavily regulated by central bank FACTORS invest in "deprived sector" has resulted in easy Central bank directive making it possible to open group accounts access to loanable funds Pre-existing NGOs that could be tapped to promote SHGs High repayment rates Subsidy grant from central bank and government India Banking laws on "priority sector lending" High repayment rates LIMITING Low capitalization Remoter regions lack pre-existing NGOs to promote groups FACTORS Inadequate management information system Remoter regions lack bank branches Group based method highly staff-intensive ­ cost Excessive dependence on subsidy and financial sector regulation of service delivery high in remoter areas Excessive dependence on donor funds and priority sector credit Maoist insurgency SCALING UP Expansion rate in tandem with management Intensive promotion of SHG banking among NGOs and bank PROCESSES capacity, experience management through outreach, education, and training Intensification of services within established Assured refinancing SHG loans from NABARD made it attractive areas of operations rather than scaling up across for participating banks. districts. Distinct periods of piloting, mainstreaming, and scaling up Quantitative From 125 clients in 1993 to 35,268 clients in From 225 SHGs in 1992 to 566,826 SHGs in 2003. Unevenness of Scaling Up 2001. Number of clients leveled off after growth regionally, however. 2001due to Maoist insurgency. Functional Some diversification in loan products achieved: Three flexible linkage models developed Scaling Up group loans supplemented by voluntary savings Flexibility in loan size, purpose, and loan terms introduced in to accounts, house loans, individual-based loans, match clients' demand and livestock insurance. Political Firmed position by incorporating as development Scaled up a national-level program. Position firmed by strong Scaling Up bank. Was successful attracting investment from financial backing from central bank and central government. private sector banks. Well-known banking personalities sit on governing board. Organizational Area based management structure with each area Scaling Up office reporting to headquarter. New MIS being developed currently. 51 Kyrgyz Republic UNDP-DE DFID-SLLPC UNDP-PAC INSTITUTIONAL Partnerships between SHGs, Direct Partnerships between Partnerships between elected ARRANGEMENTS SHGAs and LGs, CBOs, LGs and central village working groups and government through CLC NGOs/private firms; CAPACITY KAFC and RAS KAFC and RAS KAFC and RAS ELEMENTS Local UNDP activists Local UNDP activists International experts until a Meaningful micro-credit Meaningful micro-capital grant core team is trained opportunities and Local Development Fund Training seminars and small opportunities credit opportunities TRIGGERS Institutional commitment to Close cooperation with national Encouraging results from poverty reduction decentralization strategy and CDD projects in other CAS, previous with the GoKR, and the WB FACILITATING Indigenous traditions of self PAC experience; Cooperation Institutional support from FACTORS help; trained local activists; with CLC; trained local KAFC and RAS; Pilot activists; substantial budget; communities' previous National commitment to exposure to CDD through decentralization PAC; LIMITING SHGs are not formalized except Registration fees and taxes for No incentives for `working FACTORS for credit applications, don't CBOs and CBO Associations; groups' to cooperate with have permanent roles at LGs; excessive paper work and LGs, possible competition; not depressed housing markets corruption; lack of integration enough funding for the minimizing chance for of social capital into collective projects themselves collateral; high interest rates collateral processes SCALING UP PROCESSES Quantitative Already active in 7 oblasts, and Easy with low initial costs due Emphasis has been on nurture; Scaling Up in more than 150 of the 487 to collaboration with CLC, and spread and replication has LGs already significant: 531 CBOs been sporadic; integration has are active in 15 pilot AOs, 543 not been pursued CBOs active in non-pilot AOs Functional Both horizontal and vertical Ongoing; vertical scaling up is Horizontal scaling up is Scaling Up scaling up has been achieved more pronounced than integral to the project; and has horizontal scaling up also been successful in vertical scaling up Political First and second generation are Achieved full political scaling First and second generation Scaling Up achieved, third may be up scaling up is achieved, third is achieved naturally planned, fourth is not pursued actively Organizational Achieved internal management Achieved internal management, Achieved internal Scaling Up and institutional diversification, financial viability and is most management and financial financial viability may require successful at internal viability, not pursued more time institutional diversification institutional diversification 52