83022 Designing Community-Driven Development Operations in Fragile and Conflict-Affected Situations Lessons from a Stocktaking S O C I A L D E V E L O P M E N T D E PA R T M E N T Designing Community-Driven Development Operations in Fragile and Conflict-Affected Situations Lessons from a Stocktaking Jacomina de Regt, Shruti Majumdar, and Janmejay Singh S O C I A L D E V E L O P M E N T D E PA R T M E N T Copyright © 2013 The International Bank for Reconstruction and Development / The World Bank Group 1818 H Street, NW Washington, DC 20433, USA All rights reserved First printing: November 2013 Manufactured in the United States of America Please cite this report as follows: World Bank. 2013. “Designing Community-Driven Development Operations inFragile and Conflict-Affected Situations: Lessons from a Stocktaking.” Social Development Department, World Bank, Washington, DC. Photo credit: Curt Carnemark, World Bank, 1993. The findings, interpretations, and conclusions expressed in this report are entirely those of the authors and should not be attributed in any manner to the World Bank, or its affiliated organizations, or to members of its board of executive directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. ii Contents Acknowledgments.............................................................................................................................................. v ......................................................................................................................... vi Abbreviations and Acronyms. Executive Summary.........................................................................................................................................viii I. Introduction...................................................................................................................................................1 II. Methodology and Sample for the Stocktaking.................................................................................. 5 III. Summary of the Main Findings from the Stocktaking...................................................................... 8 IV. Choices and Emerging Guidance in 10 Design Areas.........................................................................15 4.1 . Preparatory Diagnostics................................................................................................................................................................. 15 4.2 . Overall Objectives. ........................................................................................................................................................................... 17 4.3. Institutional Arrangements at the National Level.............................................................................................................18 4.4. Institutional Arrangements at the Sub-national Level.................................................................................................... 21 4.5. Learning by Doing. ........................................................................................................................................................................... 23 4.6. Targeting............................................................................................................................................................................................... 24 4.7. Menu Choices, Grant Sizes, and Grant Cycles................................................................................................................... 27 4.8. Community Mobilization and Facilitation...........................................................................................................................29 4.9. Governance (including Fiduciary) and Monitoring and Evaluation. .......................................................................... 32 4.10. Peace Building and Social Cohesion Elements................................................................................................................... 34 V. Some Overall Guiding Principles for Design...................................................................................... 36 Looking to the Future: Areas for Further Research and Action.................................................... 39 VI.  ............................................................................................................................... 42 VII. Concluding Thoughts. References.......................................................................................................................................................... 45 ............................................................................................................................................................... 47 Annexes. 1 Definitions of Fragility and Conflict.............................................................................................................................................48 2 Tables Related to the Sample.........................................................................................................................................................49 3 Select Summary Tables from the Stocktaking Study.......................................................................................................... 55 4 Summary Tables and Figures to Understand and Inform Design Choices.................................................................68 Boxes 1 The Challenge of Community-Driven Development in Fragile and Conflict-Affected Situations...................3 2 Hawalla Dealers: Adapting to the Local Banking System in Afghanistan.....................................................................33 3 Summarizing the Core Recommendations Emerging from the Stocktaking on CDD-FCS Design. .................43 4 Simplified Procedures and Flexibility in the New OP 10.....................................................................................................44 iii Figures 1 Sample CDD-FCS Operations Studied, 1997–2013................................................................................................................... 6 2 Fragile Versus Post-conflict Countries in the Sample.............................................................................................................7 3 Typical Assessment Types for Preparatory Diagnostics....................................................................................................... 15 4  Methods of Approaching Preparatory Diagnostics and Associated Constraints.................................................... 16 5 Different Objectives and Mechanisms that CDD Caters to in FCS. ...............................................................................18 6 Coordinating and Implementing Aagency Choices at the National Level. ................................................................. 19 7  Design Choices and Actors that Define Sub-national Institutional Arrangements in CDD. .............................. 22 8 Four Choices for Learning by Doing............................................................................................................................................ 23 9 Three Levels of Targeting Decisions and Modalities............................................................................................................ 25 10 Decision Choices around Eligible Activities and Grants..................................................................................................... 27 11 Hiring Facilitators: Typology, Advantages, and Disadvantages........................................................................................30 12 Options for establishing community committees................................................................................................................. 31 13 Examples of Peace Building and Social Cohesion Elements............................................................................................. 34 A4-1 Steps in the Community Cycle....................................................................................................................................................... 74 A4-2 Options for Flow of Funds .............................................................................................................................................................. 75 A4-3 Social Accountability Tools by Primary Objectives in CDD Projects........................................................................... 76 A4-4 Elements of a Good Monitoring and Evaluation System for a CDD Project. ........................................................... 77 Tables A2-1 List of CDD Operations in FCS Studied.....................................................................................................................................49 A2-2 Number of Projects (167) in FCS by Region..............................................................................................................................50 A2-3 Comparison of Countries and Projects Covered in CDD-FCS Review Studies........................................................ 51 A2-4 Contextual Variation within Sample............................................................................................................................................ 53 A3-1 Overall Objectives in Projects and their Continuations..................................................................................................... 55 A3-2 Language for Project Development Objectives..................................................................................................................... 57 A3-3 Specific Components or Program Activities to Address Social Cohesion................................................................58 A3-4 Level of Decentralization Versus the Project Design...........................................................................................................60 A3-5 Size of Grant by Community........................................................................................................................................................... 61 A3-6 National Implementing Institutions and Institutions with Final Authority............................................................... 62 A3-7 Locus of “validation” of community plans............................................................................................................................... 63 A3-8 Authorization Versus Implementation of Sub-projects..................................................................................................... 63 A3-9 Design Choices for Flow of Funds by Project.........................................................................................................................64 A3-10 Choices in Menus and Public/Private Goods Offered in the CDD FCS Sample. ....................................................65 A3-11 Variations of the flow of funds to communities and local government................................................................... 67 A3-12 Community contribution requirements (cash or in-kind).................................................................................................. 67 A4-1 Peace Building Versus State Building: How Designs May Change..................................................................................68 A4-2 Conceptualizing the Design Choice on Sub-national Institutional Arrangements................................................69 A4-3 Examples of Goods and Services Delivered in CDD Projects by Core Outcome................................................... 71 A4-4 Conceptualizing the Choice between Open and Limited Menus................................................................................. 72 A4-5 Conceptualizing the Design Choice on Number of Grant Cycles................................................................................. 72 A4-6 Stages in the Community-Level Sub-project Cycle..............................................................................................................73 iv Acknowledgments This paper provides a summary of a broader stocktaking study that was part of an overall effort to develop an operational toolkit for designing community-driven development (CDD) opera- tions in fragile and conflict-affected situations (FCS). The work was jointly led by the World Bank’s Social Development Department (SDV) and the Fragile States unit in the Operations Policy and Country Services (OPCFC) network. It received funding from the PKNOW trust fund that was set up to finance knowledge and research on development issues in FCS contexts. The study could not have been completed without the valuable contributions of many col- leagues across the World Bank, who provided insights and advice on the historical, political, and operational issues that surround the design and implementation of CDD programs in FCS. The authors would like to thank the following who, through frank discussions, provided valuable insight and material that contributed to the report as a whole: Sarah Adam, Naila Ahmed, Ian Bannon, Laura E. Bailey, Sean Bradley, Richard Calkins, John Elder, Scott Guggenheim, Bernard Harbonne, Elisabeth Huybens, Zishan Karim, Markus Kostner, Nicholas Kraft, Balakrishnan Menon, Adrian Morel, Meena Munshi, Daniel Owen, Michelle Rebosio, Caroline Rusten, Colin Scott, Rad- hika Srinivasan, Matt Stephens, Yasmeen Tabba, and Varalakshmi Vemuru. The comments of the peer reviewers—John Elder, Marcelo Jorge Fabre, Scott Guggenheim, and Susanne Holste—greatly improved the study. Additional comments were also received from Laura E. Bailey, Eric Bergthold (USAID), Kenneth Green, Francoise Gopal, and Richard Hogg; also the staff of the South Asia Rural Development and Livelihoods Unit (SASDL) contributed exten- sive comments. Special thanks go to Atif Ansar for his initial research work that informed the analysis in this study, as well as to Kaori Oshima, who provided excellent inputs, comments, and visuals to help make the report more user-friendly for readers. Helene Grandvoinnet (cluster leader, Social Accountability Team) and Susan Wong (sector man- ager, SDV) provided overall guidance for the study. Finally, we are grateful to Kristin Hunter and Laura Johnson for excellent editorial support, design, and formatting assistance. All errors and omissions are the sole responsibility of the authors. v Abbreviations and Acronyms AGETIP Agence d’Execution des Travaux d’Interet Public contre le sous-emploi [Public works and employment agency] ARMM Autonomous Region of Muslim Mindanao (Philippines) AusAID Australian Agency for International Development BP Bank Policy BRA Badan Reintegrasi Aceh [Aceh reintegration agency] CAR Central African Republic CBO community-based organization CDC community development committee CDD community driven development CDFP Community Development Fund Project (Kosovo) CDP Community Development Project (Bosnia & Herzegovina) CDRDP Community-Driven Recovery and Development Project (Somalia) CoP community of practice CPIA Country Policy and Institutional Assessment CRDP Community Reintegration and Development Project (Rwanda) CRP community resource person CSDP Consultative Service Delivery Program (Iraq) DFID Department for International Development (UK) EAP East Asia Pacific (region) FAS Fundo de Apoio Social [Social development fund-Angola] FCS fragile and conflict-affected situations FY fiscal year GAC governance and anti corruption GET global expert team GRM grievance redress mechanism HQ headquarters IBRD International Bank for Reconstruction and Development ICM Implementation Completion Memorandum ICR Implementation Completion Report IDA International Development Association IDP internally displaced person IEC information, education, and communication IGA income-generating activity ISR Implementation Status and Results Report J-PAL Jameel Poverty Action Lab KDP Kecamatan Development Program vi Abbreviations • vii LGU local government unit LICUS low-income country under stress LIL learning and innovation loans (LILs) M&E monitoring and evaluation MDG Millennium Development Goal MMRDP Magdalena Medio Regional Development Project (Colombia) MRRD Ministry of Rural Rehabilitation and Development NELSIP North East Local Services Improvement Project (Sri Lanka) NGO non-governmental organization NSP National Solidarity Program (Afghanistan) NUSAF Northern Uganda Social Action Fund OED Operations Evaluation Department O&M operations and maintenance OP Operational Policy PAD Project Appraisal document PCAP Post-Conflict Assistance Project (Cote d’Ivoire) PDO project development objective PIU project implementation unit PMO prime minister’s office PMU project management unit PNPM National Program for Community Empowerment (Indonesia) PPF project preparation facility PRA Participatory Rural Appraisal RALDP Reconstruction and Local Development Project (Guatemala) SVGCDP Support to Vulnerable Groups Community Development Project (Central African Republic) TTL task team leaders SDF Social Development Fund (Yemen) SDV Social Development Department SIL specific investment loan SPF State and Peace Building Fund UN United Nations VNDP Village and Neighborhood Development Project (West Bank & Gaza) WB World Bank WB&G West Bank & Gaza WDR World Development Report Note: All dollar amounts are U.S. dollars unless otherwise indicated. Executive Summary This paper reviewed the designs of a representative sample of 17 community-driven develop- ment (CDD) programs implemented in fragile and conflict-affected situations (FCS) over the past decade and a half to distill lessons that can be informative for task teams working on CDD-FCS. With “reverse engineering,” project documentation, and task team interviews and focus group discussions with task team leaders (TTLs), this paper summarizes the main findings of the stock- taking, provides design guidance in 10 areas, and outlines four overarching principles that are fundamental to CDD-FCS projects. On overall design, the stocktaking reaches four conclusions: 1. There is a lot of variation in the menu of design choices across the CDD projects in the sample. 2. The menu of design choices seems similar to what teams would likely face in non-FCS settings. 3. Within the sample, there is little differentiation in design choices selected between “(post) conflict” and “fragile” situations. 4. Driven by several FCS specific criteria that influenced decision-making by project teams, we observed patterns in the choices made from the overall menu. In terms of preparatory diagnostics, most projects relied on informal analysis and very few under- took formal social, political or conflict studies. Of the overall objectives, the dominant ones were improved service delivery, restoration, and governance, while other socio-economic goals, such as reducing conflict, peace building, increasing social cohesion, or creating livelihoods, were less common, but often mentioned as additional objectives. When it comes to institutional arrangements, the study sample from past CDD-FCS projects displays a much higher use of outsourcing to (semi) independent agencies, although such arrangements decline in subsequent projects. At the sub-national level, most projects opted for a hybrid or mix of roles assigned to local governments units (LGUs) and communities. Moreover, almost all projects in the sample incorporated “learning by doing,” gathered either from earlier pilots, project preparation, or phased implementation. Perhaps most surprisingly, all but two projects limited themselves to one grant cycle or block grant per community per project, even for follow-on projects. Finally, while all task teams attested to the importance of facilitators in FCS, project documentation shows little about their precise role. viii Executive Summary • ix Based on the stocktaking, this paper provides design guidance in 10 areas: 1. In terms of preparatory diagnostics, understanding the context through political economy and conflict analysis is crucial. This may well take the form of ongoing informal stakeholder consultations. Whether or not formal studies are feasible is determined by urgency of rapid progress, budget, availability of data, and need for political buy-in. 2. In terms of overall objectives, limiting the first phase to one objective and then setting in place building blocks for the development objectives in subsequent phases is crucial. 3. When choosing the agency to take charge of implementation at the national level, it is important to assess whether the agency has the capacity and power to protect and deliver the program—particularly if it is in the hands of a government agency—and assess whether the agency is neutral if it is outsourced. In the case of the latter, an exit strategy needs to be planned and considered as well. 4. For sub-national-level institutional arrangements, a hybrid of community and local govern- ment units should be considered, with adequate capacity building and empowerment built into the design. 5. In FCS, past experience suggests that investing in some form of learning by doing is well worth while, but it should be a continuous endeavor—not a one-off exercise—through- out the life of the project as it goes to scale. 6. The sixth recommendation concerns decisions about where to target benefits, which can be highly sensitive in FCS. Different factors should be weighed at different levels: the macro-level should be determined by political economy analysis that assesses whether proj- ect benefits should be spread across the whole area or zoom in on particular regions. At the inter-community level, transparently communicating the criteria used and the rules of the game is of utmost importance. And at the intra-community level, targeting should be used sparingly, especially in the first phase. 7. Using a wide-ranging closed menu—an “in-between” arrangement of sorts—is likely the best choice for the first phase, which can move toward a full open menu over time. More- over, when it comes to the number of grant cycles, task teams should plan to repeat (block) grants to cement participation and accountability. 8. The importance of gaining trust and obtaining buy-in from critical stakeholders at local and higher levels, when setting up the facilitation and community leadership arrangements, cannot be emphasized enough. 9. Both monitoring and evaluation (M&E) and governance require special attention in project design and they need to be much more robust in FCS, since governance and anti-corruption (GAC) risks are much higher in these contexts. 10. And finally, in order to influence social cohesion and peace building in FCS, task teams must have a much broader strategy of aligning with and influencing different stakeholder interests and incentives. x • Designing Community-Driven Development Operations in Fragile and Conflict Situations Apart from overall design guidance, the interviews undertaken as part of the stocktaking study also revealed some cross-cutting principles that task teams repeatedly mentioned: 1. Be responsive to context and, within this context, treat CDD as a tactic or a part of a toolkit of things that countries can do, rather than a strategy for development. 2. Thinking long-term is key, even if responding quickly is necessary in the short run. The theory of change should be to plan for the evolution of a CDD project into the govern- ment’s participatory local-planning and/or service-delivery mechanism. 3. It is important to keep things simple across all design elements, and certainly when it comes to donor coordination. 4. Task teams recommended that designers “think scale.” In other words, the design team has to ensure that the design demonstrates the benefits of the approach at a scale that can attract attention and be showcased. This design stocktaking study should be seen as the first step in a deeper and longstanding effort to improve our understanding and guidance of designing CDD programs in FCS. Looking forward, much more work is clearly needed to help better guide our work on CDD in FCS and build upon its “science of delivery,” in the words of World Bank President Kim. There are four areas that deserve special attention: 1. We need to better understand fragility and CDD design. 2. We must gather much more evidence on how design choices play out in real life. 3. We also need to better develop our operational toolkit on peace building, social cohesion, and conflict resolution. 4. Finally, there is a clear case for greater “south-south” exchange, donor harmonization, inter- nal World Bank harmonization and quality control, and external/internal alliance building around CDD in FCS. I. Introduction Fragile and conflict/post-conflict situations (FCS)1 present the most profound challenges to development in the world today. In both fragile and conflict-affected situations, poverty lev- els are usually high and welfare outcomes low. The stability and social cohesion necessary for development is frequently lacking. And often there are no strong and legitimate institutions to address poverty and manage conflict. Violent conflict is more likely to reemerge in such areas, leading to further impoverishment, undercutting social cohesion, and eroding institutions. The result can be a vicious cycle of deprivation and insecurity which is difficult to dislodge or change. Based on analysis of data in Global Monitoring Report 2013 (World Bank 2013b), only 20 countries with fragile and post-conflict-affected situations have met one or more Millennium Develop- ment Goal (MDG) targets and only six are on track to do so by the 2015 deadline.2 Development work in such contexts remains extremely hard. Even where violence has not escalated, localized conflict may provide the spark for future fires. Community-driven development (CDD) programs have long been viewed as particularly relevant development interventions in FCS. Defined as an approach to local development in which con- trol of decision-making and resources for local infrastructure and service delivery is wholly or partly transferred to community groups, the critical advantage of CDD lies in its flexibility to respond to the scale and nature of community needs in diverse local conditions. As World Development Report 2011 (World Bank 2011) argues, the “range of purposes [that CDD operations] can serve and the relative ease of adapting their design to different needs” makes CDD useful in FCS. Effective CDD projects can distribute resources quickly and to remote, rural areas. In devolving decision-making, they can help ensure that resource distribution is fair and popularly accepted, and can operate in areas with security risks. Programs may also support peace-building by providing incentives for collective action that can work across conflict divides, by contribut- ing to local institution building, and by strengthening vertical society-state linkages. Based on the context, CDD operations can also be designed to help target specific vulnerable groups, increase livelihood security of communities, or help improve local shared services. An additional feature of CDD in post-conflict settings is that it helps establish a direct link with the central government structure, which in many cases is lacking during conflict/post-conflict years. Creat- ing grassroots institutions of the poor (i.e., participatory and inclusive community groups) allows a weak and/or new government to gain legitimacy and trust at the ground level as it works towards reconstruction and rehabilitation of the country. Further, while the emerging evaluation 1. See annex 1 for definitions of fragility and conflict used in this paper. Note that all fragile situations do not necessarily experience or move into conflict, but in conflict situations there is, per definition, fragility. Moreover, these situations are “fluid” in both space and time. 2. Reported in World Bank (2013c) available at http://www.worldbank.org/content/dam/Worldbank/Feature%20Story/ Stop_Conflict_Reduce_Fragility_End_Poverty.pdf. A few countries included in our report are among those that have reached some MDGs. 1 2 • Designing Community-Driven Development Operations in Fragile and Conflict Situations evidence shows mixed results for social capital and conflict outcomes, it does suggest some success when it comes to service delivery and economic outcomes (Barron 2011; Wong 2012).3 While the practice of deploying CDD in FCS has spread rapidly,4 guidance on designing programs in these settings has been lacking. Undertaking CDD programs in FCS is particularly challenging (see box 1). Neutral information is hard to get, project delays can reignite conflict, and weak governance and capacity is pervasive. The work on assessing these types of challenges and the potential for CDD operations in FCS as a consequence started more than a decade ago.5 While several of these past studies already have reviewed design issues, what has been missing is an in-depth and comprehensive understanding of what design choices are available and how to select between them in such situations. After all, the flip side of the immense flexibility in CDD is that it can end up covering a very broad remit of designs; thus CDD projects come in all shapes and sizes. There are several initial, primary questions to consider: What institutional models have been used and why? How should funds flows be arranged? How should facilitation methods be set up? What menu of sub-project activities should be available for financing? What levels of cost sharing, if any, should be involved? Although questions like these are relevant for a CDD task team in any context, the need to find careful answers to these issues is even more pertinent in FCS. Guidance on how to approach these design questions and an analysis of what has been done in the past is currently missing. This paper summarizes a stocktaking of the designs of a representative sample of CDD in FCS projects in an effort to fill this knowledge gap. By looking at the designs of a representative sample of CDD programs that have operated in FCS over the past decade and a half, the stock- taking study aims to provide in-depth and user-friendly operational guidance to CDD task teams working in these difficult contexts. The stocktaking adopts a “reverse engineering” strategy to help unpack the design choices that CDD teams face and uses a combination of literature review and interviews to provide some preliminary guidance and insights in design choices. The 10 spe- cific areas of design covered in this report6 are: (1) preparatory diagnostics; 2) overall objectives; 3) national institutional arrangements; 4) sub-national institutional arrangements; 5) learning by doing arrangements (phasing and piloting); 6) targeting; 7) menu of activities, grant sizes, and cycles; 8) community level sub-project cycle and facilitation arrangements; 9) governance and monitoring and evaluation (M&E) measures; and 10) peace building and social cohesion. Through task team interviews and focus group discussions with groups of task team leaders (TTLs), we delved deeper into some of these choices. The guiding research questions behind the 3. It is important to emphasize, however, that CDD should not be seen as a “magic bullet,” but as one tactic among many that should be linked to a broader strategy of interventions in FCS. In this way, CDD should always be thought of as a means to an end, which like any other will face limits. In fact, as the team task leaders (TTLs) interviewed for this stocktaking pointed out repeatedly, “we often load too much on CDD projects…it is important to keep in mind what we are trying to achieve in these projects and not oversell this.” 4. World Bank CDD database. As noted later, analysis over the fiscal year (FY) 2000–2010 period, as part of this stocktak- ing, showed that the Bank had financed at least 168 operations using CDD approaches in FCS contexts. 5. See, for instance, Cliffe et al. (2003), Strand et al. (2003), and Owen et al. (2006). 6. The broader stocktaking report covers the further design areas of safeguards and fiduciary arrangements. Section I: Introduction • 3  he Challenge of Community-Driven Development in Box 1. T Fragile and Conflict-Affected Situations Many characteristics of FCS countries pose particular challenges for the design and implementation of CDD projects: • The nature of the risks: Risks to project funds and to development objectives are not simply higher than in more normal countries, they are fundamentally different: – Interim/coalition governments consisting of former rebel leaders, for example, may be ill-prepared for the administrative challenges of running a government, while the staffing of key ministries may consist mainly of former combatants who need to be given jobs, regardless of merit or qualifications. – Sector-level implementing agencies may be weak, totally dysfunctional, or altogether absent. – Poor governance, weak internal controls, and ineffective oversight agencies may be accompanied by entrenched, conflict-related corruption, including criminal activities, such as drug trafficking, while the opportunities to tackle crime and corruption may be constrained by the peace accord and the need for political stability. • The consequences of failure: These risks are complicated by the interdependence between the political, security, justice, and develop- ment spheres, where failure in any one area could lead to the failure of all. Failures in the security area, for example, may prevent the restoration of basic services, the reconstruction of infrastructure, the creation of employment, and the improvement of livelihoods, and thus the achievement of the peace dividend. The consequences of failure include the possibility of a resumption of conflict. • The need for a quick response: In some cases, delays in the restoration of critical services—from health care to trade and commerce— may have catastrophic consequences for civilian populations who survive the conflict, but remain vulnerable to disease and death from social and economic impoverishment. This is true for the survivors of natural or man-made disasters (such as earthquakes, floods, and droughts, or financial, food, and fuel crises), to which FCS countries are especially vulnerable. In such cases, the risks to donor funds and development objectives must be weighed against the consequences of failing to act, or failing to act in a timely manner. • The importance of not making things worse: Channeling new resources into targeted areas, or to targeted groups within a given area, may have implications beyond those intended. Such programs create “haves” and “have-nots,” winners and losers, and “new elites” with potential risks to political and social stability. The basic principle is to “do no harm,” but that is easier said than done. • The need for simplicity and flexibility: The profound lack of capacity at all levels in many FCS countries requires simplicity in project design, especially in the administrative burden that the project may impose on government systems (central and local) and on communities. Source: World Bank (forthcoming, 2013). document review and interviews were: (1) what were some of the choices considered along each of these 10 areas of design; (2) how did the particular context and nature of fragility and conflict influence these choices (that is, how are some of these choices different from what we are likely to see in a non-FCS sample?); (3) what were some key adaptation features; and (4) what finally worked and what didn’t? In other words, what are some things that the TTLs would recommend or discourage future teams from doing? Implicit was the hypothesis that by answering these questions there would be sufficient information to provide guidance on the adaptive features of CDD design in FCS. This guidance is what we summarized in this synthesis paper. The hope is that this guidance will be helpful not only for ongoing projects and current TTLs but equally 4 • Designing Community-Driven Development Operations in Fragile and Conflict Situations informative for new projects and TTLs working in these settings.7 It represents a first step toward filling this critical knowledge gap in our work in FCS.8 The paper is structured as follows. In section II, we give a brief description of the methodology of the study and describe the sample used for analysis. We proceed with a summary of main findings from the stocktaking, analyzing some of the trends we found in section III. Section IV then looks at the results of the stocktaking and the reverse engineering to elaborate on the choices in the ten areas of design covered here with suggestions on how to select between them. In section V, we illustrate some key principles that task teams should bear in mind when designing these projects. In section VI, we provide some ideas for future research work and action on CDD in FCS. Finally, section VII summarizes the design guidance revealed from this stocktaking exercise, and offers some concluding thoughts. 7. For example, South Sudan and Myanmar are the two latest additions to the CDD-FCS portfolio. 8. The full stocktaking study is meant to be more of a reference document or sourcebook that can be accessed by CDD- FCS teams as and when needed for more detail on design questions in specific operations. Methodology and Sample II.  for the Stocktaking The stocktaking was informed by a combination of a desk review of project documents and interviews with task teams. The project documentation included project appraisal documents (PADs), project implementation completion reports (ICRs), operations manuals, project imple- mentation plans, and aide memoires of the sample of CDD projects. These were used to list the design choices deployed and get information on how they related to the countries’ contexts and desired outcomes. Data from project documentation were combined with inputs from TTLs (via individual interviews, as well as through three “market test” sessions of focus groups that brought together over 50 CDD experts and TTLs) to get direct feedback and insight into the parameters and judgments that lay behind decisions on design of these operations. In addition to questions about design choices, we asked the teams to reflect on the success and effective- ness of these decisions during implementation and draw the lessons for other task teams facing similar contexts. Finally, we also reviewed the secondary literature on CDD design, concepts, and applications to FCS.9 A sample of 17 projects was selected from the universe of Bank-wide CDD operations in FCS. We then obtained a data sample of the Bank’s CDD projects in FCS from the larger number of CDD operations maintained by the SDV anchor since fiscal year (FY) 2000. It contained close to 800 CDD operations for the period FY 2000–2010 in over 90 countries, which we narrowed down for the FCS universe using the following criteria: • Projects in countries on the official FCS list for 2010, that is, with a country policy and institutional assessment (CPIA) rating of 3.2 or less, or the presence of a United Nations or regional peace-keeping (or peace-building) mission (e.g., African Union, European Union, NATO), excluding border-monitoring operations, during the past three years • Projects in countries that are otherwise stable with a CPIA greater than 3.2, but contain specific regions of conflict (e.g., Mindanao in the Philippines) • Projects in countries with a history of conflict or post-conflict reconstruction and develop- ment in the past 10–-15 years (e.g., such as Rwanda) were also added. This broader FCS universe covered 167 CDD projects in 29 countries across all regions where the World Bank operates (see table A2-2 in annex 2 for a list of the 167 projects). After assembling this CDD-FCS projects universe, we purposefully chose a sample of roughly 10 percent (about 17 projects) by applying the following rules: 9. See references for a complete list of the studies reviewed. 5 6 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Figure 1. Sample CDD-FCS Operations Studied, 1997–2013 ‘97 ‘98 ‘99 ‘00 ‘01 ‘02 ‘03 ‘04 ‘05 ‘06 ‘07 ‘08 ‘09 ‘10 ‘11 ‘12 ‘13- Afghanistan: NSP Phase I–III Angola: FAS II Bosnia & Herzegovina: CDP Colombia: MMRDP Phase I–II Central African Republic: SVGCDP Cote d’Ivoire: PCAP Guatemala: RALDP Indonesia (Aceh): BRA-KDP Kosovo: CDFP Phase I–II Iraq: CSDP Philippines: ARMM Social Fund Rwanda: CRDP Somalia: CDRD Sri Lanka: NELSIP Uganda: NUSAF West Bank & Gaza: VNDP Yemen: SFD Phase I–IV • Each project in the sample must belong to a unique country. • Each region in which the Bank operates must be represented with at least two projects.10 • Moreover, we also built on prior studies and other stocktaking initiatives, given the ease of availability of documentation (see table A2-3 in annex 2). The resulting sample of 17 programs is provided in annex 2 and the years covered by them can be seen in figure 1. As shown, many of the projects had additional financing or second and third phases, leading to an analysis (to some extent) of some 32 projects. The sample includes mostly post-conflict settings (figure 2), although there was significant diversity in the context each proj- ect faced (see table A2-4 in annex 2). The processing arrangements for the projects differed significantly.11 While three programs in Afghanistan, Central African Republic (CAR), and Cote d’Ivoire were explicitly set up as emergency projects, most projects were specific investment 10. This criterion added the Guatemala Reconstruction and Local Development Project to ensure an additional experi- ence from Latin America and Caribbean region. Given that many of the FCS countries are in Africa, this regional quota naturally skewed the sample against that region. 11. The choice of lending instrument can have profound implications for the design of an FCS project. In the Central African Republic, for example, the team switched from a specific investment loan to an emergency loan and many pre- paratory steps were folded into the first 18 months of implementation. As a result, more (most) funds went to the quick disbursing component (the Priority Response Fund), which financed construction of necessary infrastructure selected by the government in consultation with the communities through an external agency (AGETIP, Agence d’Execution des Travaux d’Interet Public contre le sous-emploi [Public works and employment agency]). The CDD component, especially the grant disbursement, did not get off the ground till almost the end of the project. Section II: Methodology and Sample for the Stocktaking • 7 Figure 2. Fragile Versus Post-conflict Countries in the Sample On-going conflict Immediately post conflict Fragile Colombia Afghanistan I-III Guatemala Northern Uganda Bosnia & Herzegovina Aceh, Indonesia Somalia Central African Republic Yemen Angola I–II Cote d’Ivoire West Bank & Gaza Afghanistan I-III Iraq (Kurdistan) Mindanao, Philippines Rwanda Sri Lanka Angola III Kosovo I and II Note: The projects are labeled as such, not based on the present day situation, but based instead on their state of fragility during project design. Some countries fall into more than one category as there is fluidity, as well as regional diversity in fragility, within FCS countries. Table A2-4 in annex 2 describes the conflict context in each country in the sample. loans (SILs), and in two countries—Rwanda and Colombia—there were learning and innova- tion loans (LILs). West Bank & Gaza (WB&G), Somalia, and Iraq were financed through specific trust funds. The sample is representative cross-regionally and reviews of both implementation completion reports (ICR) and implementation status and results (ISR) ratings suggest that it was well performing with almost every project rated moderately satisfactory or above. It is important to be cognizant of the limitations of this report. First, given that this is largely a desk review backed by task team interviews, and not impact evaluations or field surveys, we have little evidence on the impact of the design choices made or of their benefits vis-à-vis other programs in FCS (although we do know from the ICRs that the completed projects have deliv- ered the expected outputs with high beneficiary assessment ratings). Our goal is thus to present the overall menu of available design instruments and illustrate patterns of choices, as well as reflections (based on the opinions of the concerned task teams) on why these choices were made. Second, as a stocktaking, this study naturally looks at the past, which means that the study is subject to a recall problem. Typically task teams adjust design to context, as deemed most appropriate at the time, so some of these design choices may not have been active choices, but rather were tacit decisions based on informal knowledge that was never formally documented. Thus, in many instances, it is difficult to ascertain why certain choices or changes were made since official documents often do not capture the thought processes and alternatives that may have been considered. Also, many additional design choices, which may have been introduced during implementation as adaptations to FCS reality, were not necessarily brought into the discussion by the TTLs. Finally, since this study is sample based, with projects covering largely post-conflict (rather than fragile) situations, and since some projects are no more than small pilots, readers should be aware that some of the results are influenced by this distribution of the sample. Given these caveats, the following section presents some of the key findings from the stocktaking. Summary of the III.  Main Findings from the Stocktaking While the broader stocktaking study holds all of our findings and reflections, separated into a number of specific design elements—development objectives, project preparation choices, tar- geting, menu of choices, national institutional arrangements, sub-national institutional arrange- ments, community-level sub-project cycle and facilitation arrangements, fiduciary arrangements, governance and social accountability measures, safeguards, and monitoring and evaluation—it is possible to distill eight main conclusions and findings. First, from an overall design perspective, the stocktaking reaches four conclusions: 1. The menu of design choices across the CDD projects in the sample showed a lot of variation. 2. The menu of design choices seemed similar to what teams would likely face in non-FCS settings. 3. Within the sample, there was little differentiation in design choices selected between (post) conflict and fragile situations. 4. There were, however, some observed patterns of choices from the overall menu, driven by several FCS-specific criteria that influenced decision making of project teams. The projects in the sample displayed wide variations in the menu of instruments and choices made in almost all areas of design, be they institutional arrangements, funds flows, targeting, or overall objectives. This in and of itself was not surprising because it truly showed the kind of flexibility that CDD is known for. Moreover, it made it possible to apply the reverse engineering methodology effectively to develop typologies for design choices (as discussed in the next sec- tion). Arguably, the variations in the choices made from this menu reflect the FCS context-driven realities that governments and task teams faced when conceiving the operations and establish- ing their decision-making criteria. Upon further analysis, reflection, and discussions with task teams, what also became clear is that this menu of choices and instruments did not seem that distinct from non-FCS settings. 8 Section III: Summary of the Main Findings from the Stocktaking • 9 (Notable exceptions are the use of “on/off” switches12 and conflict filters.)13 This finding can be interpreted in different ways. On one level, given that the CDD approach was originally estab- lished as a crisis-response mechanism, it is not surprising that the menu of design elements and models, now used much more extensively, is not so different when comparing FCS to non-FCS. (This points to the built-in malleability of CDD projects, which can indeed be applied to difficult contexts in much the same way as non-FCS.) Moreover, that the menu of design choices can be the same in FCS and non-FCS may be true for many other development sectors. For instance, the menu of choices involved in designing a road, a school, or a health center will largely be the same in FCS and non-FCS. However, on another level, this finding does raise questions about whether the actual choices or selections made from this menu are different. That is, was the adaptation of design to the FCS context enough to reveal distinct patterns of design? In this regard, we did indeed observe that certain patterns of choices emerged as responses to FCS and this is reflected in the findings of our stocktaking below. But, at the same time, as the example of the on/off switch from Indo- nesia shows (see footnote 12), the reality of implementation in FCS is sometimes only revealed in informal conversation: it may not be documented and maybe not even be written up in the operational manual. Therefore, a detailed answer to the question of adaptation (beyond what is captured in this study) is not possible without deeper, cross-sectional, and longitudinal empirical evidence. Related to this, we observed little or no difference between design choices made in (post) con- flict versus fragile contexts in our sample. Arguably, one would imagine that the two contexts might demand different strategies in terms of institutional arrangements (e.g., weak governance versus conflict-ridden situations), types of capacity building (e.g., a fragile situation could occur in a non-fragile state), delivery mechanisms (e.g., working with international agencies may be required after a conflict), and so on. However, in our sample, no clear differences like this were seen. This again poses the questions of whether and how CDD project designs should be altered when applied in a fragile (and not post-conflict) setting—which we return to later in the paper. Second, in terms of preparatory diagnostics, most projects relied on informal avenues of analy- sis14 and very few projects undertook formal social, political, or conflict studies. This was a 12. In Indonesia, during its years of conflict, the task team designed and negotiated a special “on/off” switch with the finance ministry and auditors that allowed them to stop action when conflict flared up without losing an entire budget year. Otherwise, as happened in the National Solidarity Program [NSP] in Afghanistan, they could wind up paying a huge amount of overhead to idled facilitators. Such context-driven adaptations (which clearly one would not find in a non-FCS country) are further illustrated in the longer stocktaking report. 13. It is important to note that, when we refer to the difference between FCS and non-FCS CDD projects, our knowledge of the latter does not draw on any existing stocktakings of the same but rather on the experience of the authors with non-FCS CDD projects. Moreover, in our interviews and focus group discussions, we urged task teams to reflect on what in the context (the fragility and post-conflict nature of it) set some of these decisions apart from non-FCS projects—and we draw on that accumulated knowledge as well. 14. This, of course, does not include mandatory social assessments that many projects did as part of their safeguards requirements. 10 • Designing Community-Driven Development Operations in Fragile and Conflict Situations somewhat unexpected finding since one would imagine that the need for a preparatory diagnos- tic in FCS would be enhanced. However, as task team interviews revealed, the urgency of project preparation and pressure to start implementation and show results quickly made carrying out formal studies an impractical luxury. Instead, task teams and projects (e.g., in Iraq, Afghanistan, and West Bank & Gaza) relied on existing post-conflict studies and informal means of understanding local socio-political and conflict contexts. They also looked at institutional capacity by consult- ing key stakeholders15 or hiring expert consultants or partners, who had grassroots knowledge and experience, as well as those who had done prior work with similar programs in that country or elsewhere. This being said, there were some reported cases of conflict and political economy analysis. For instance, during appraisal of the Kosovo Community Development Program (CDP) II, a special social and conflict analysis was carried out to deal with (presumably ethnically mixed) communities. Likewise, the Sri Lanka North East Local Services Improvement Project (NELSIP) used a “reconciliation and conflict filter” to analyze design choices in the local context when designing the project. (See annex 13 in the NELSIP project appraisal document.16) Third, in terms of overall development objectives, the two dominant objectives have been improved service delivery and restoration, and improved governance.17 Tables A3-1 and A3-2 in annex 3 provide a snapshot of project development objectives (PDOs) across all projects, and examples of language used in PDOs, respectively. Other socio-economic goals, such as reducing conflict, building peace, increasing social cohesion, and creating livelihoods, were less common, although often mentioned as primary objectives.18 At first instance, this tilt against the latter themes may seem inconsistent with FCS—after all, shouldn’t themes like cohesion, state build- ing, reduction of conflict, or restoration of livelihoods be at least as important as infrastructure and service delivery in FCS situations? However, further analysis of project documents and dis- cussions with task teams suggest that there are several reasons for this trend: • Context and actors matter. Most of the CDD projects in our sample were operating in post- conflict countries, where realizing a peace dividend quickly was the top priority.19 Therefore, the focus on services and infrastructure restoration was primary. Moreover, the actors (both in the governments and in the Bank) may have influenced the core focus, given the implicit theory of change they adhered to and their understanding of the philosophy behind CDD. If this philosophy or theory of change centers on local governance and uses CDD as mecha- nism for instilling democratic traditions, then the objectives and corresponding design will be quite distinct from a vision that sees CDD largely as a delivery mechanism for infrastruc- ture and services. For instance, as one TTL put it, “The key role of CDD in such contexts is 15. For instance, CAR used the stakeholder meeting as a design tool in light of the processing under its emergency procedures. 16. World Bank (2010a). 17. Improved governance was seen in many project approval documents (PADs) to be equally important as an objective at both the community and local government level through participatory planning and implementation mechanisms (creating capacity), with the rationale that, when communities learn how to plan and implement in a participatory way, this leads to better (local and central) governance and citizen empowerment. This, in a way, described and justified the use of a CDD approach as a means to an end. 18. For instance, 6 of the 17 projects explicitly adopted social cohesion as a primary or secondary objective, but only half of them have measured it or have a separate outcome indicator for it. 19. There are several tradeoffs between the two approaches—quick peace dividends versus longer state-building proj- ects. These are discussed in greater detail in the sourcebook. (For a summary, see table A4-1 in annex 4.) Section III: Summary of the Main Findings from the Stocktaking • 11 not as much social cohesion, as it is to deliver services and benefits quickly to a population that has voice on what the development menu should be.” The pattern of PDOs seen in our sample seems to match this overall thinking around the theory of change of CDD. • Phasing and sequencing matters. TTLs also emphasized the fact that considering broader socio-economic outcomes beyond service delivery is rarely an either/or decision. They commented how complicated and challenging livelihoods and social cohesion are in this environment, given the high risk of elite capture. And even though these are high priority objectives, in the short run they may need to be secondary goals. In this instance, then, the theory of change is to emphasize service delivery in the short run, with the expectation that it will play an instrumental role over time in building trust and cohesion by improving accountability, bringing the state and its citizens closer, and bringing about better gover- nance.20 And one finds in the sample that, in cases where there were several successor proj- ects (e.g., Angola, Yemen, and Somalia), an evolution of objectives occurred, moving slowly towards broader social and governance outcomes.21 • Keeping things simple is critical. Many projects (e.g., Afghanistan NSP-I) explicitly men- tioned that they did not include livelihoods-type components as these need different implementation arrangements and expertise, which would add complexity to the design. In other cases (e.g., Angola II and Sri Lanka), there were complimentary projects focusing on livelihood dimensions. As one TTL remarked, “It is in a TTL’s best interest to keep the project’s objectives narrow and focused on service delivery; otherwise we would be setting ourselves up for failure.” • Finally, it is also important to note that even though not many projects put social cohe- sion, peace building, or livelihoods as explicit objectives; nonetheless, several included components or activities to promote cohesion and livelihoods, which were often for tar- geted population groups, such as youth or conflict victims. (Table A3-3 in annex 3 provides examples from each project of the specific components that were social cohesion-related.) Fourth, when it comes to institutional arrangements, the sample of past CDD-FCS projects displays a much higher use of, or outsourcing to, (semi) independent agencies, although such arrangements decline in subsequent projects. In fact, only in only 3 of 17 countries (Sri Lanka, Rwanda, and Indonesia) were all or part of the implementation not assigned, or contracted out, to some independent agency. These agencies had varying responsibilities, ranging from just car- rying out implementation (e.g., the use of AGETIPs in CAR and Cote d’Ivoire) to having final authority at the central government level (e.g., the social funds in Angola, Guatemala, Uganda, Philippines, and Yemen). Choices are well argued in PADs: Afghanistan NSP I considered using an agency model versus a line ministry and opted for the line ministry with community implemen- tation. In the end, though, it outsourced fiduciary oversight to an independent agency because the objective was to get better and new community-level governance. Angola FAS I–II opted for 20. Of course, even in this theory of change that puts service delivery objectives at the forefront, adequate attention to cohesion and trust building still needs to be provided through instilling and enforcing the transparency, accountability, and participation elements of CDD in the project design. 21. In Angola, for example, service delivery and local community empowerment for secure areas ranked high in FAS I (Fundo de Apoio Social [Social development fund]), while economic infrastructure was emphasized more in FAS II to kick start livelihood improvements. In FAS III, the integration of the local governments nationwide became more prominent. Somalia, on a much smaller scale and with shorter funding cycles by various donors, also moved from a purely community driven approach in phase I to an approach linked to local government and government planning by phase III. 12 • Designing Community-Driven Development Operations in Fragile and Conflict Situations the autonomous Social Fund, given the general state of conflict and weak government presence overall in the country; Angola FAS III, on the other hand, considered moving toward implementa- tion through local governments, but considered their delivery capacity too weak and opted for capacity building instead. In Colombia (with the Magdalena Medio Regional Development Proj- ect, MMRDP I and II), the Bank joined an existing program with an independent agency and the PAD endorsed the reasons for rejecting the alternatives of central government implementation (weak capacity and no legitimacy) or local municipality implementation (weak and controlled by interest groups) because the objective was to improve governance. Interviews with TTLs corroborate that, indeed, FCS justifies the high use of independent agen- cies: the speed with which national governments have to show a peace dividend, the weak state apparatus, and usually weak intermediate or local government institutions all justify the use of high capacity (presumably more neutral and autonomous) agencies in implementation, if not overall national coordination of CDD operations in FCS, even if they result in higher costs. How- ever, although going through independent agencies is common, further analysis reveals that over time their use has steadily declined, especially in the longer-running programs.22 Generally, as the projects evolve and as situations in FCS become more stable and decentralization emerges, the trend is for local and intermediate levels of government to start playing a greater role. Fifth, most projects opted for a hybrid, or mix, of roles assigned to local government units (LGUs) and communities. These sub-national institutional arrangements reflect the level of decentralization in the country. Traditional CDD literature has often differentiated between the community, the LGU, and the agency models of CDD. However, as our design stocktaking revealed, the reality of sub-national institutional arrangements in CDD is not so simple and very few projects followed these ideal types—only three countries (Afghanistan NSP I–III, Somalia CDRD I–II, and Yemen SFD I–IV)23 chose to use the pure community model, not involving LGUs. Only Sri Lanka NELSIP and Bosnia & Herzegovina CDP adopted a pure LGU model, where all funds and implementation was done via local government. The rest of the projects set up their sub-national institutional arrangements so that communities and LGUs shared responsibilities. As interviews suggested, this implies that the thinking on CDD has gradually shifted away from ideal types to finding optimal co-production between LGUs and communities, based on the local context, while recognizing that no one model will be perfect everywhere. The implicit theory of change behind this that CDD design teams increasingly embrace is that longer term sustainability of the program will require stronger LGUs. 22. For example, in Iraq CSDP, the level of engagement of community action groups with local government has increased over the life of the program, and the role of the International NGO (ACDI/VOCA) is increasingly one of facilitation rather than of primary implementation. Similarly, Afghanistan NSP II–III rely far less heavily on internationally recruited oversight consultants and much more on provincial management units. In addition, the Ministry of Rural Rehabilitation and Devel- opment (MRRD) had an expanded role in NSP III. 23. In these three countries, the local governments were either non-existent or non-functional, often due to the active conflict. Demonstrating from the bottom up that there is community capacity and that development can take place was important in each of these situations. Section III: Summary of the Main Findings from the Stocktaking • 13 Moreover, an analysis of the contexts within which the projects operated suggests that these institutional arrangements at the local level mirrored or matched24 the state of decentraliza- tion of the country, and that these decisions were based on an assessment of policy risk and performance risk associated with the local government system. The institutional arrangements matched the level of decentralization, not only at the beginning of the projects, but also as they evolved over time to match shifting levels of decentralization. This implies that design teams clearly paid significant attention to the future role of LGUs and indeed adapted the design according to context.25 There were also cases, such as Uganda and Colombia, where the CDD projects had a direct impact on the decentralization process and helped change local govern- ment attitudes and practices.26 Overall, there was a pattern over time and across countries of increasingly providing the LGUs with more roles, more responsibilities, and more authority—of course, depending on how decentralization progressed. Sixth, almost all projects in the sample incorporated “learning by doing” from earlier pilots, project preparation, or phased implementation. This learning by doing took four forms: (1) a pilot project prior to the project, (2) proof-of-concept testing during project preparation (financed through preparation advances or trust funds), (3) a phased implementation approach (testing the whole approach in stages), and (4) pilot projects during implementation (usually for specialized interventions). Project teams clearly used this learning-by-doing approach to main- stream practical lessons into the design and implementation of the larger program. The longer stocktaking report extensively documents all the pilots undertaken for each project and the kinds of lessons learned and incorporated into the project design from it. One example is the FAS Phase I in Angola that had a pilot to support income-generation activities, but which was discontinued in Phase II after an evaluation of the pilot showed its limited contribution to sus- tainable commercial income-generating or micro-enterprise activity. Aside from this, some of the longer programs instituted pilots during later phases to experiment with new activities and interventions. Seventh, almost all projects limited themselves to one grant cycle or block grant per com- munity per project, even for follow-on projects. Only the Somalia Community-Driven Recovery and Development Project (CDRD) and the West Bank & Gaza Village and Neighborhood Devel- opment Project (VNDP) were designed to provide two rounds of grants to each community to enable learning by doing and to have skills and governance processes become more institution- alized. That more projects explicitly did not choose to offer multiple rounds of financing or block grants to the same community in weak FCS is a curious finding because, intuitively, one would imagine that empowerment, building trust, social cohesion, and poverty reduction goals 24. Serrano-Berthet et al. (2010) distinguished three different strategies for deciding on decentralization: lagging, match- ing, and leading strategies. A lagging strategy does not take full advantage of the opportunities provided by an existing policy environment. Such a strategy reflects a reluctance to entrust LGUs with responsibilities for setting priorities and implementing projects. A leading strategy is employed to accelerate change in an existing institutional environment, reflecting confidence in the path of reforms and often providing capacity building and “proof of concept” for instituting more functions and resources at the local government level. A matching strategy straddles the other two strategies. Typically the institutional arrangement in this case mirrors the on-the-ground capacity of the LGUs as stipulated in the ex- isting decentralization policy and legislative structures. This last approach seems to be the most common in our sample. 25. Table A3-4 in annex 3 provides an analysis of how different projects compared to the level of decentralization in the countries they were operating in, while table A3-7 shows where decisions on sub-projects were made. 26. See chapter 8 in the longer stocktaking report for a fuller discussion of the linkages of the projects in the sample with their respective decentralization contexts. 14 • Designing Community-Driven Development Operations in Fragile and Conflict Situations would only occur through multiple rounds of grant financing.27 So, why do task teams not opt for multiple rounds? Interviews with task teams suggested that this is largely related to political economy factors and budget constraints. In most post-conflict situations studied, there was an overwhelming need for infrastructure rehabilitation and, given limited resources, the desire to spread the peace dividend as widely as possible meant that there was a trade-off between repeating grants versus covering more communities; project teams seemed to have opted for the latter. But, as they admitted, this did diminish the likelihood that participatory governance and cohesion objectives would be achieved.28 In short, the number of grant cycles is not an isolated discussion and needs to be woven into a larger discussion of what the project hopes to achieve. Related to this was also a huge variance in the size of grants, which ranged from very small cash grants in the Indonesia BRA-KDP project29 to grants up to $150,000 in the Guatemala Reconstruction and Local Development Project (RALDP). Table A3-5 in annex 3 provides a sum- mary of the sizes of grants used in different projects. Finally, while all task teams attested to the central importance of facilitators in FCS, proj- ect documentation reveals little about their precise role. Communities have a great need for facilitation after conflict and in crisis times, sometimes higher than anticipated at design. However, the task teams varied in what they saw the role of the facilitators to be. While some TTLs believed they were meant to solve micro- or household-level problems and bolster local capacity building, others envisaged a much bigger role for facilitators in mitigating conflict and building trust. Task teams pointed out that this is often contingent on whether the conflicts are characterized by inter-community (and sometimes inter-regional) conflict—where facilitators are seen to play a larger role—or whether the conflict is between state and non-state actors. It is important to clarify this distinction in project design, not just for its own sake, but also because the kind of skills and training expected in frontline workers are contingent on the role envisaged for them in these situations. However, in FCS it is often tough to find good facilitators, especially women, and different proj- ects used different methods to fill this gap, such as piggybacking on other projects in the coun- try.30 For instance, the Yemen Social Development Fund (SDF) dealt with a shortage of facilitator skills in general, and in specific areas of the country, by financing formal training at universities for selected students from those regions to prepare them for future work with SDF. In a similar vein, the West Bank & Gaza VNDP recognized the invaluable role youth was playing in facilitation and at the mid-term review started offering them training and stipends. Overall, countries used a combination of outsourcing and in-sourcing, sometimes switching from the former to the latter as capacity developed in government agencies (e.g., in LGUs). What did not emerge from project documentation, however, was whether the facilitators got special training, selection procedures, and mentoring—things which task teams highlighted in interviews as important. 27. In fact this has been a longstanding critique of CDD projects in general. OED (2005) had a similar conclusion, namely that several rounds of financing were needed in CDD projects to ensure that such processes would be institutionalized. 28. “One grant in 11 years won’t get you much,” like in NSP, was one of the comments from task teams, for instance. 29. BRA-KDP: BRA = Badan Reintegrasi Aceh [Aceh Reintegration Agency]; KDP = Kecamatan Development Program. 30. In Indonesia, for instance, through the broader National Program for Community Empowerment (PNPM) program, the non-profit Community Facilitator Certification Institute has been set up to establish competencies and certification standards, offer refresher training, and develop curricula for universities. This will bring about sustainability and help professionalize the facilitators that the program in Aceh, for example, will benefit from.  hoices and Emerging IV. C Guidance in 10 Design Areas This section offers some guidance on specific areas of design that emerged from our analysis. While we have already made clear that the purpose of the stocktaking was not to make prescrip- tive statements on what to do when, the reverse engineering methodology that was deployed has allowed us to unpack the menu of choices available to task teams in different design areas. And, based on the insights and feedback from TTLs in interviews, we inferred broad parameters or factors that come into play when choosing a specific design. We present the guidance on choices and the criteria by which to select between them, across 10 areas of design. This is, of course, a summary and readers who want more detailed examples on design areas should refer to the broader stocktaking study. 4.1 . Preparatory Diagnostics Understanding the local context through some kind of analysis of political economy, fragility, and conflict is crucial and need not be a formal study. Designing a CDD operation that specifi- cally responds to the fragility or conflict of a country requires diagnostic analysis of that context and more specifically of the dynamics of the conflict—both at the micro and macro levels. The stocktaking shows that most teams use informal means to carry out these diagnoses. Formally, however, there are several types of assessments, and methods of approaching them, to choose from, as shown in figure 3. Some assessments come as requirements, such as the social and environmental assessments linked to the safeguards OP 4.01. Figure 3. Typical Assessment Types for Preparatory Diagnostics Social and Environmental Political Economy Conflict Analysis Beneficiary Assessments Assessments Assessments/ Stakeholder Analysis • Useful to identify • Useful for problem • Used to map potential drivers of conflict diagnosis and needs social and environmental • Useful for mapping and • Useful not only during assessment by the impacts of project and assessing capacity of preparation but also target population safeguard against existing institutions implementation since • Useful for social and adverse consequences and actors context changes economic assessment • Allows broad • Useful for bringing rapidly of the beneficiaries consultations with together stakeholders (e.g., IPs) affected populations to collect FCS-specific experience 15 16 • Designing Community-Driven Development Operations in Fragile and Conflict Situations  ethods of Approaching Preparatory Diagnostics and Figure 4. M Associated Constraints Formal Data Collection • Qualitative studies/focus groups • Quantitative surveys Required if there is no existing data, • Expert reports but time would be needed and security • Analysis of poverty and would have to allow for data collection. socio-economic data • Review of past studies, projects, and relevant literature Informal Mechanisms Under pressure to move rapidly and • Consultations generate buy-in, this is a better approach, • Multi-stakeholder workshops especially in the absence of reliable data. • Inputs from other CDD experts Others, such as a beneficiary or (participatory) poverty assessment, may well be undertaken as a preliminary step in the community-level social mobilization process that CDD programs often begin with. But other assessments—like macro or local political economy analysis or drivers of conflict and fragility analysis—are also options and their role may be more important in FCS, given a fragile and complex context. Methods for undertaking these analyses vary, ranging from direct formal studies directly or literature reviews, to more informal mechanisms, such as local consultations and stakeholder opinions from round tables and workshops, and expert inputs (see Figure 4). Regardless of the timeline, task teams strongly advised taking advantage of the social assess- ments, required under the safeguards OP, as well as mechanisms to undertake fragility and local political economy analysis. The decision to include “conflict filters,” as in Sri Lanka’s NELSIP, in order to review design choices with the goal of minimizing the risk of fueling greater conflict and tensions was also suggested as an option. Using as much existing data available from past projects and engaging or partnering with experts with local understanding was recommended in order to have “eyes and ears on the ground.” Task teams also stressed that this type of political economy and/or fragility analysis should be ongoing throughout the life of the project at all levels, not just in preparatory stages because FCS is, by definition, dynamic and volatile. Much of the value of the political economy analysis is to provide guidance on how to interact with rapidly evolving government and social contexts. This argues for continuity of decentralized teams that operate at the grassroots level. At the same time, several task teams noted the importance of Section IV: Choices and Emerging Guidance in 10 Design Areas • 17 south-south learning across CDD-FCS projects to benefit from the experiences of others.31 Finally, at a more strategic level, task teams advised that the purpose of the analysis and prepara- tory diagnostics should also be to figure out where and when a CDD approach is needed within a broader development strategy in the FCS country. In this way, the analysis not only informs the design of the CDD program but also where, why, and for how long it should be applied. 4.2 . Overall Objectives CDD projects in FCS can cater to multiple objectives, but the primary goal will be likely deter- mined by the specific nature of the conflict or fragility—in other words, by the context. As noted in the introduction and depicted in figure 5, CDD projects can deliver various outcomes. The four key outcomes are improving public infrastructure and service delivery, improving social cohesion (and preventing conflict), promoting better governance, and enhancing livelihoods (including common property resource management). The primary focus determines which design choices make sense or not; task teams advised that this should be based on the political economy and context analysis. In particular, whether the overall strategy of the project is to produce a short-term peace dividend or develop a longer-term mechanism for state building will fundamentally change the sequencing of objectives. While the former prioritizes service delivery and infrastructure, the latter typically means that governance, social cohesion, capacity building, and possibly livelihoods become primary.32 Among factors that influence the choice on overall objectives are issues of context, simplic- ity, capacity, sequencing, and mandate. Based on the stocktaking interviews, experienced TTLs argued that design teams should ideally limit themselves to one primary development objec- tive, especially for the first phase, and set in place building blocks for the development objec- tives for subsequent phases.33 TTLs equally strongly suggested that governments and the Bank have a long-term perspective in mind when introducing a CDD operation in FCS. In many cases, the primary objective of the first phase will be to provide a peace dividend and provide rural infrastructure and services, moving to governance in the second phase—as was the case for many projects covered in the stocktaking. Specifically for more intangible outcomes, such as empowerment, social cohesion, or enhancing trust, they advised limiting objectives to what is realistically attainable, as most CDD projects have little to no influence on the drivers of conflict at a macro-level.34 31. The Iraq team for instance used the global expert team (GET) for FCS to provide insights on international experiences with CDD in FCS settings to inform the strategy around a proposed scaling-up of the CDD program there. 32. Table A4-1 in annex 4 attempts to show the ways in which design choices may change based on whether the primary goal is a quick peace dividend (i.e., restoring services and infrastructure rapidly) versus longer-term state building (i.e., developing sustainable participatory local governance systems). This distinction is often temporal or determined by the state of the political settlement in the country, but it also reflects divergent philosophical differences on how to approach development in FCS. 33. Task teams, however, noted that having one primary objective should not be equated with having just one sector of focus, given that the multi-sectoral nature of CDD is one of its strengths. (All the projects covered in the sample were multi-sectoral.) In fact, a multi-sectoral approach requires simplicity in terms of processes and institutions since it can be wasteful and confusing for communities to have multiple community decision-making bodies for different sectors, instead of a single community development committee or LGU covering a range of local development programs. 34. In general, current knowledge of CDD operations has limited understanding of the impact of these projects on such intangibles and it is a potential area for future research. 18 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Figure 5. Different Objectives and Mechanisms that CDD Caters to in FCS Economic Deprivation Challenges in Collapsed Fragile and Conflict Settings Instability Services Institutional Weaknesses Improved Public Effective Project Aims Improved Social Infrastructure and Institutions and Livelihoods Cohesion Services Good Governance • Better matching of • Greater participation • Demand for more • Improve physical resources with needs in civic life and responsive institutions assets at the Mechanisms • Improved efficiency improved relations • Improve citizen-state community level • Better maintenance • More accepted relations • Sustained income • Ability to work in resource distribution • Common platform growth high conflict areas • Mechanisms for for programming • Promotion of diffusing problems economic growth Source: Adapted from Barron (2011). 4.3. Institutional Arrangements at the National Level Choosing the appropriate national implementing agency can be one of the biggest deter- minants of success for a CDD program in FCS. In terms of choices, as seen in figure 6, the first fundamental choice is which agency will have final authority: one within existing government ministries or a (semi) independent non-state agency. Within government, one can opt for a cross-cutting ministry or agency (e.g., finance or planning), the prime minister’s office (PMO), or a line ministry (rural development, social welfare, etc.) as the lead coordinating or implementing agency. Independent or non-government agencies can include international agencies, such as the United Nations or special purpose bodies (e.g., social development funds). Then there is an additional choice on what actual implementing modality is used. Projects can vest final author- ity in a particular ministry, but implementation may be outsourced to external agencies (e.g., non-governmental organizations, NGOs) or kept with an existing department or special project implementation/management unit in the ministry. Section IV: Choices and Emerging Guidance in 10 Design Areas • 19 Figure 6. Coordinating and Implementing Aagency Choices at the National Level Coordinating Agency (with authority) Semi-Government/ Independent Type Government Semi-Autonomous Non-Government Coordinating Ministry Line Ministry Special Purpose International Agency Agency (e.g., MoF, MoI, (e.g., Rural Devt., Agency (e.g., UN body, MoPlanning, PMO) Social Welfare, etc.) (e.g., Social Fund) Intl. NGO) Project Outsourced to Implementing Implementation/ Existing Unit or external agents Modality Management Unit Department (e.g., NGO, fund) (PIU/PMU) There are advantages and disadvantages to each of these choices.35 A cross-cutting ministry typically has the advantage of having more power and influence, which it can use to change procedures and processes. It can also work and convene actors across sectors more easily. Also, in scaling up, CDD programs working with a cross-cutting ministry have fared much better than the other two options. The disadvantage is that, unless the cross-cutting ministry actually runs or monitors LGUs, its sub-national presence can be limited (this is particularly the case in FCS) and it will need to hire all the project implementation staff in some form or another. Line minis- tries, on the other hand, usually have a strong sub-national presence, as well as sectoral technical expertise, although both can be severely limited in FCS. Therefore, to the extent that the project is delivering certain core services (such as health, education, and water), working through a line agency will ensure that the necessary supply-side technical inputs—required to supplement the CDD grants—are available.36 The downside is that some of these line ministries will not be able to convene the players in other sectors, turf battles may ensue (especially when scaling up), and in FCS line ministries may not have the clout and power to streamline government procedures. 35. These are more extensively discussed in the longer stocktaking report. 36. In fact, evidence from evaluations shows that sector CDD programs have stronger service delivery outcomes, likely as a result of this supply-demand linkage. See Wong (2012). 20 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Finally, the speed with which national governments have to show a peace dividend, the weak state apparatus, and often weak intermediate institutions means many projects (as seen in our sample) will find it useful to use independent agencies or social funds, which have the advan- tage of being able to circumvent the government bureaucracy, have more flexible rules and procedures, have the ability to influence the culture of a new organization, and have the ability to hire all staff from the market. The downside of this, as task teams pointed out, is that such an agency in FCS creates parallel systems, will have less clout (unless it reports directly to the prime minister or president), raises questions about sustainability, and invites turf battles, especially with line ministries and sometimes with local government when scaling up.37 Such an approach is also inconsistent with the “New Deal for Engagement in Fragile States” that emphasizes the use of country systems.38 Given these pros and cons, it is not surprising that a variety of models are seen across CDD-FCS projects.39 Factors influencing the choice of national level institutional arrangement include context, mandate, capacity, neutrality, and power. Design teams in post-conflict contexts are likely to be under pressure to deliver quick results in countries or regions where administrative capacity is weak, non-existent, and/or not very well connected to the local population because of the recent conflict. Trust in either central or local government may also be very weak. Moreover, the capacity of communities to organize and prioritize may also be constrained at this stage. In post-conflict settings, there is also a coordination problem: many humanitarian organizations may still be operating; donors may be entering the country with similar or different approaches, intent on showing support as a peace dividend and wanting quick results; and central govern- ment agencies may have their own different approaches. While it may be tempting to set up a new agency just for CDD implementation, this must be tempered by the need for sustainability and institutionalization. Given all of this, task team advice was to make sure that the final agency in charge of imple- mentation has the capacity to deliver, has the neutrality and “power” to protect the program, as well as the ability to scale-up and transition the program. There will be trade-offs. For fast delivery, a parallel structure may be needed. Yet, for the longer term, the authority and imple- mentation capacity may need to move from an independent agency to a government agency (or agencies), and from central government to LGUs as decentralization progresses. This same trade-off will also be tempered by the extent to which the state-building imperative trumps short-term delivery of infrastructure and services—likely to be an issue in fragile contexts. Per- ceptions of state legitimacy and the level of trust in public institutions in general will also matter; if they are widely associated with the root causes of conflict, then it may be worth going with an independent agency even when there is capacity within the state. Another important aspect within this area is the choice of the executive director, who should be perceived as a fair arbiter 37. However, fierce turf battles can ensue in any of these arrangements if and when the government does not have a very strong strategic goal of using the CDD project approach as a comprehensive planning and implementation approach for all its development activities at the local level. 38. The “New Deal for Engagement in Fragile States” was announced by the members of the International Dialogue on Peacebuilding and Statebuilding, comprised of the G7 (the finance ministers of seven industrialized nations: the United States, the United Kingdom, France, Germany, Italy, Canada, and Japan) plus a group of 19 fragile and conflict-affected countries, development partners, and international organizations as part of the “Fourth High Level Forum on Aid Ef- fectiveness” in Busan, South Korea, in 2011. For more information on the New Deal, visit www.newdeal4peace.org. 39. Table A2-6 in annex 2 shows how our sample is distributed across these categories. Section IV: Choices and Emerging Guidance in 10 Design Areas • 21 between the two (or more) sides of a conflict, who should be “clean” and either so powerful or so well protected that they can manage with minimal political interference from outside. This is especially important in the first phase and when or if a program is implemented by a more independent agency.40 With an independent or transitional model for project delivery, an exit strategy needs to be considered. As noted, this is also in accordance with the “new deal” principles, which favor government managed systems. Assuming that the exit strategy is a decentralized system, the program designers must ask themselves whether the policy environment adequately supports decentralization. The ministry chosen to coordinate or implement the program needs to see CDD as an instrument that supports and promotes its own agenda and/or decentralization; likewise, the ministry in charge of promoting decentralization should see CDD as a proper and integrated part of the system that will provide a platform for multi-sectoral interventions. In sev- eral countries in the stocktaking are strong and positive examples of linkages with the ministry of local government (Rwanda) or less successful ones where the CDD project did not have an impact (West Bank & Gaza). In post-conflict situations, where decentralization is often still incipi- ent and where many donors may be pushing their version of decentralization, integrating the CDD approach requires that the CDD project have sufficient scale to demonstrate results, that a long-term time frame has been considered, and that consistent linkages are being established. This enables the CDD project to become a country-led strategy of decentralized participatory planning and implementation. The stocktaking shows one additional possibility of an exit strat- egy, where the independent agency remains, but it transitions to take on other government and donor programs. The Foundations in Kosovo and Bosnia evolved into such a role, as did the Social Fund in Angola. 4.4. Institutional Arrangements at the Sub-national Level At the sub-national level, the design choice relates essentially to the division of roles between LGUs and community organizations. As mentioned earlier, traditional CDD literature has often differentiated between the community, the LGU, and the agency models. However, as this design stocktaking revealed, the reality of sub-national institutional arrangements in CDD is not that simple. In fact, as suggested in figure 7, there are at least six design questions regarding the decisions about who is in charge. And for each of these, the options include going through a community organization, the LGU, an external agency like an NGO, or some combination of joint decision-making and management by any or all of these actors. Thus, the universe of potential design combinations at the sub-national level is quite large. In terms of guidance, the clear message from the stocktaking was to think about the appropriate “hybrid” of community and LGU arrangements. Parameters that influence the decisions include: • where (or how high) the LGU is situated, • capacity and responsiveness of LGUs, 40. The role of the executive director was mentioned as especially critical in Kosovo and in Colombia. 22 • Designing Community-Driven Development Operations in Fragile and Conflict Situations  esign Choices and Actors that Define Sub-national Institutional Figure 7. D Arrangements in CDD Design choices Possible actors Who plans? Community Organization Who decides? NGO/Agency Who controls/manages funds? Who implements? LGU Who maintains projects? Who monitors? Combination • degree to which the LGU is associated with the conflict (e.g., if there were local officials who perpetuated or failed to control violence), • fiscal and administrative decentralization structures, and • the amount of specialized targeting needed.41 Other contextual factors, such as the homogeneity or heterogeneity of communities under an LGU or the existence of strong polarization along ethnic or class lines, may also matter in terms of deciding how to distribute roles. As task teams noted, however, this is not a static decision and arrangements do and should evolve over time, based on learning from successes as well as failures (e.g., Mindanao and Guatemala). Most cases should see a growing role for LGUs. It is key to ensure that adequate capacity building is built into design, as well as adequate monitoring 41. Table A4-2 in annex 4 outlines several of the key factors and parameters that influence the choice between these models based on past literature. Tables A3-7, A3-8, and A3-9 in annex 3 summarize how the stocktaking CDD-FCS sample was distributed in terms of some of these parameters (e.g., who decides versus who implements, etc.). Section IV: Choices and Emerging Guidance in 10 Design Areas • 23 of results to build trust. Planning for such a transition takes deliberate effort and a government strategy to empower both LGUs and communities.42 4.5. Learning by Doing Past experience suggests that investing in some form of learning by doing is well worth it in FCS. As mentioned earlier, the stocktaking shows the importance that design teams placed on learning by doing. There are essentially four choices that task teams have (figure 8). The first is to invest in a stand-alone pilot project, often financed through a trust fund, such as the State and Peace Building Fund (SPF) or Japanese Social Development Fund (JSDF),43 but it may also be financed and implemented by another donor or aid agency. The second is to undertake some kind of “dry run” during preparation as a “proof-of-concept” test (e.g., through a project prepa- ration advance).44 And finally, one can integrate the learning by doing during implementation in two ways, as part of a phased roll-out of the program and/or a pilot test of some specific intervention.45 Figure 8. Four Choices for Learning by Doing 1 2 3 4 “Dry Run”/ “Proof of Concept” Phased Stand-Alone Piloting during Testing during Implementation Pilot Project Implementation Preparation (e.g., Uganda NUSAF, (e.g., Afghanistan NSP) (e.g., Angola FAS) (e.g., Rwanda, CRDP, Sri Lanka NELSIP) Colombia MMRDP) 42. In Afghanistan, for example, after 10 years of implementation through project implementation units staffed by con- tractors and consultants, there is now a concerted effort to streamline salaries, benefits, and allowances of contract staff, and to strengthen civil service and government agencies so they can take over the role of program implementation. In some agencies, capacity building is focused on such core competencies as procurement, human resources, financial management, security, and management. 43. The Post Conflict Fund (PCF) and State and Peace Building Fund (which replaced the PCF) were used in three cases in the sample, and none used JSDF funding. Six of the 17 projects based their “learning” on prior WB financed projects. NSP I was one of those projects that were able to learn from prior smaller pilot projects: the ICR notes the Afghanistan Emergency Community Empowerment and Public Works project (June 2002–December 2004) was de facto “phase zero” of the program. 44. In Colombia, where the implementing Foundation was already operating the program, a Japanese grant, as well as project preparation facility (PPF) financing, was available prior to the first LIL, which assisted with project preparation, such as the participatory social assessment. Similarly, Rwanda conducted four Participatory Rural Appraisals (PRAs) to learn which priorities communities would pick, especially given the massive return of refugees in 1996–1997; to identify project interventions to address these needs; and to identify specific local institutional arrangements. This resulted in the possibility of financing the construction of houses for returnees and homeless families if and when a community develop- ment committee (CDC) decided on such a project. 45. For instance, in mid-2004, the Afghanistan NSP initiated a participatory district governance pilot in six districts—the District School Construction and Rehabilitation Window (DSCRW)—to test the feasibility of linking CDCs with the lowest level of state administration in joint-planning exercises with a defined budget. It was accessible on the condition that a minimum of 70% of the CDCs participated and a majority endorsed the plan agreed upon with the district administration. 24 • Designing Community-Driven Development Operations in Fragile and Conflict Situations The key is not to think of learning by doing as a one-off exercise, but as a continuous endeavor throughout the life of the project as it goes to scale. Given the constant in-flux nature of fragile contexts, trial runs and pilots must be planned for their pay-off in terms of learning. Task teams must weigh the trade-offs between a more in-depth pilot versus a rapidly deployed project. There is also a decision on scale—pilots that are too small can sometimes just become “islands of excellence” that do not provide the level of demonstration effect required to go mainstream and influence policy dialogue for scaling up the CDD program. Therefore, even if the decision is made to conduct a pilot, the task teams recommended that it should be within the context of a national scale program, perhaps as a “phase zero” during the preparation or roll out of the program. TTLs in fact suggested that one should conceive of the CDD program in FCS as a multi- tiered, interlocking set of phases or pilots, through which to build in flexibility and adapt the design using the regular review processes of the operations manual. Pilots can even be useful in restructuring the whole project. This runs counter to traditional project culture46 and requires strong backing from management, but the uncertainty of the FCS may require such an approach. 4.6 Targeting Decisions on where and who to target through the CDD program can be highly sensitive in FCS. While poverty is often the primary targeted variable in a normal context, in FCS, design teams need to go beyond poverty to look at levels of destruction and conflict threats, and to consider which groups have been particularly affected by the conflict (e.g., child soldiers, ex-combatants, internally displaced persons [IDPs], refugees, widows, orphans, and those handi- capped by war injuries). Moreover, in FCS, typically these choices need to be made in light of limited or unreliable data regarding who and where to target. Some sources of data will not be neutral; the question is whether accuracy and verity can be checked through triangulation. And finally, the fact that targeting takes place, by whatever criteria, can become an additional source of tension in FCS, more so than in any other context. Thus, establishing fair and transparent methods for allocating resources and communicating these widely and frequently is essential. There are three levels—each requiring different data and different methods—at which tar- geting decisions need to be made (see figure 9): 1. The first target is at the macro-level, deciding where regionally or geographically the CDD program will operate. Determining this, in many ways, is a question of scale and less a find- ing from pilot programs. Should the program only target heavily conflict-affected areas? Or should it avoid them in order to show proof of concept in terms of the CDD model and move to the tougher areas once capacity is built? Going small also allows a program to undertake significant learning by doing, but it runs the risk of being seen only as an isolated island of excellence that is not scalable. A larger overall coverage area allows for more politi- cal appeal and ability to scale-up, but it puts a lot of implementation pressure and runs a greater risk of political interference as the program gets influenced by multiple regional interests.47 46. It may be simpler in the future under the new OP 10, as discussed further in the concluding section. 47. During the stocktaking, we found a lot of variation on this overall macro targeting question. Section IV: Choices and Emerging Guidance in 10 Design Areas • 25 Figure 9. Three Levels of Targeting Decisions and Modalities • Political economy considerations Macro will determine whether project will Level cover whole country or regions versus a pilot • Use available simple indicators, such as poverty, conflict damage access to services from many Inter- sources, and triangulate community • Requires complex formulae that pull indicators together • Include as component the improvement of data systems • Community itself targets, using tools such as participatory wealth ranking Intra- • Separate component or funding community for pre-identified groups • Ear-marked funding within community grant for pre-identified group 2. Once the overall geographic coverage of the CDD program is decided, the second level of targeting is inter-community—that is, deciding how resources and program activities will be distributed across communities within that geography. This can now bring in more tech- nical criteria, such as poverty, conflict damage, access to services, and so on. However, more sensitive decisions about ethnicity (if based on geography), for example, may also come into the picture. The criteria can range from simple indicators to fairly complex formulae.48 3. The third level of targeting is intra-community, which refers to targeting specific groups within communities. Broadly, there are three ways of doing this: participatory methods, such as wealth ranking applied by communities themselves to identify the families most in need (e.g., Philippines, Indonesia, Sri Lanka);49 funding pre-identified groups, such as ex-com- batants, widows, and orphans, and their project proposals (separate components, possibly with parallel implementation arrangements that address the needs of vulnerable groups, as occurred in Cote d’Ivoire’s Post-Conflict Assistance Project, PCAP; and the Northern Uganda Social Action Fund, NUSAF); or ear-marked funding for specific target groups within overall 48. The Northern Uganda Social Action Fund (NUSAF), for instance, used nine target indicators, using data from various sources, while Kosovo created a composite population/poverty index. See the larger stocktaking report for details. 49. If participatory community processes are the goal in general, then the mode of targeting involving the community in the identification of the targeted groups and their problems or solutions could in fact lead to greater inclusion. 26 • Designing Community-Driven Development Operations in Fragile and Conflict Situations community financing (e.g., Somalia’s Community-Driven Recovery and Development Proj- ect, CDRDP).50 Different factors come in to play when making targeting decisions at the three levels, macro, inter-community, and intra-community (figure 9). At the macro level, task teams specifically cau- tioned that the targeting requires governments, through their political economy analysis, to weigh the trade-off of their optimal choice for that moment: peace dividend for all or only for certain regions, areas with direct conflict, or areas that hosted refugees and IDPs, or just a pilot project. This in turn, requires design team teams to know the art of negotiation and achieving acceptable compromises. For inter-community targeting, ideally, the design team should have data and time, neither of which is easily available to design teams in FCS. Ideally, task team would have access to clear diagnostics to assist in establishing criteria for selection, adequate time to reach an agreement with stakeholders (government and the warring parties, donors, humanitarian NGOs, etc.) about these diagnostics and criteria, and enough time to develop transparent communication about the selection criteria and actual selection (Owen et al. 2006). The design of targeting mecha- nisms should be simple, explicit, easy to monitor, based on objective criteria, protected against elite capture, and impervious to leakage of benefits—a tall order under any circumstances. TTLs advised using all available data sources first, imperfect though they may be (NGOs, military, peace keeping forces, UN World Food Program, and other humanitarian agencies), to target the meso level and to triangulate across sources. But above all, TTLs stressed that the communica- tion around the rules of the game—the criteria of selection, who will benefit first, and who will benefit later, etc.—must be clear, persistent, and transparent because, despite best intentions and the use of “objective” data for targeting, the targeting process is inherently political in nature as one is choosing “winners and losers.” Knowing the conflict landscape at each of the three lev- els is critical in this regard so that design teams are prepared for the outcomes of their targeting strategy. Likewise, investments in information campaigns will be key. Finally, be cautious with intra-community targeting, per the feedback from TTLs. Providing a grant to a community has the advantage of bringing the whole community together in an inclusive, participatory, and transparent manner, the benefits of which may outweigh those from targeting specific groups for special benefits. Providing grant funds to targeted population groups, especially for private goods (e.g., income-generation grants for war widows), creates project management complexity, as separate skills are required to facilitate such groups to suc- cess. Beyond beneficiary satisfaction, there is no clear evaluation of success for these small components. Although it is tempting from a political economy perspective, the TTL’s advice was that such targeting should be used sparingly as it can be highly sensitive in FCS. Especially in the first phase, channeling resources to specific groups via separate, parallel projects is preferable, but again only if the administrative capacity to do so exists as there is also the risk of fuelling 50. This option is also chosen when another donor is willing to fund just one particular priority, for example, an NGO that addresses the needs of handicapped war victims. This has the distinct advantage for the community, in that the participatory process only has to be undertaken once, and for the donor in that it can benefit from a pooled approach which saves administrative costs. Section IV: Choices and Emerging Guidance in 10 Design Areas • 27 Figure 10. Decision Choices around Eligible Activities and Grants Menu Approach Type of Good Size of Grants Number of Cycles Grant Selection Process • Limited • Public • Micro (<5k) • One • Competition for • Open • Private • Small (<50k) • Multiple grants • Medium/Large • Fixed allocations (e.g., 50–150K) per community pressures if things go wrong.51 There will be cases in FCS, however, when the conflicts overlap with ethnic and indigenous populations, and in those cases the Indigenous People’s Policy (OP 4.10) will apply and require specific targeting to these groups within the project.52 4.7. Menu Choices, Grant Sizes, and Grant Cycles The overall objectives of the CDD program influence the design choices for grants and eligible activities. As the stocktaking analysis revealed there are a lot of choices to be made here (figure 10). The menu of eligible activities may be open (usually with a small negative list) or closed (restricted to certain types of benefits), and the latter can include both public and private (individual benefit) goods and services. The decision between these options depends largely on the overall objectives of the program, as well as the implementation capacity and coverage, since a more closed menu makes it easier to provide supporting technical expertise. In annex 4, tables A4-3 and A4-4 provide some sample public and private goods and services provided in CDD programs, as well as the generic factors that would influence the decision between open and closed menus. Table A3-10 in annex-3 summarizes how the projects in our sample chose the goods and services to provide. There is finally also a choice of providing grants to each commu- nity that submits a plan based on a pre-determined fixed allocation, or of having communities within a LGU compete with each other for sub-project grants out of a larger allocation. Guidance from TTL interviews suggests that using a wide-ranging closed menu—an in-between arrangement of sorts—may be the best choice for the first phase, which can then move toward a full, open menu over time. The recommendation to establish a wide-ranging closed menu in the initial phase caters to a restricted set of common needs and allows time to make technical assistance and relevant supply side inputs available in advance. The menu should, however, be based on good participatory research in communities,53 so that it reflects community priorities 51. An interesting example of the dangers of targeting was revealed in the Afghanistan NSP program, where free wheat was distributed to targeted groups, but proved to be very difficult. The case is discussed in further detail in the longer stocktaking report. 52. Both CAR and the Philippines, for instance, prepared indigenous people’s plans. 53. The stocktaking suggests this was a useful methodology used in the Autonomous Region of Muslim Mindanao (ARMM) in the Philippines. 28 • Designing Community-Driven Development Operations in Fragile and Conflict Situations and retains the demand-driven nature of CDD. In practice, most projects—even with fully open menus—end up with similar sub-project typologies because communities often address basic needs first.54 But given capacity and availability of technical assistance and of supply side inputs (crucial to early and rapid success), a more restricted menu may still be beneficial. As the country context and technical capacity changes, the menu can become fully open. Teams also suggested including an “other” category in order to escalate and approve innova- tive community priorities. If peace building or cohesion is an objective, communities can make several prioritized lists: one of things they can do or finance themselves, one for cohesion, and one for infrastructure and other needs. Having different finance windows, such as two grants per community for different purposes, was also an option suggested here (e.g., Cote d’Ivoire). TTLs cautioned against mixing public and private goods, however, because this increases management complexity and adds to governance risks. That being said, having some private good facility for vulnerable groups (e.g., conflict victims, widows, etc.) may be unavoidable from both a social and political perspective in order to build legitimacy and trust in the program. For grant sizes, the choices are micro (under $5,000), small (usually under $50,000), and medium/larger grants (usually under $150,000); the choices for grant cycles are whether to have one cycle per community or many during the life of the project. These choices are largely based on overall objectives, as well as the overall availability of funds and the method chosen (sub-project grant or block grant for a community development plan). For example, when a project looks to achieve a short-term peace dividend over a broad coverage area, having one cycle of mostly small or micro grants is acceptable, especially if it avoids the perception of inequitable distribution of funds and the risk of (re)igniting conflict.55 On the other hand, setting up a whole infrastructure (CDC, facilitation, etc.) for one grant cycle seems relatively inefficient. In particular, if the objective is to go beyond a fast peace dividend, there should be more than one grant cycle that moves beyond meeting basic needs, cements learning about participation and community empowerment, progresses to a community-level socio-economic development plan, links communities to LGU structures, and improves accountability of LGUs.56 Similarly, if one targets very specific groups, then using private micro transfers makes sense; but if one is attempting to reconstruct larger, cross-community infrastructure, then larger grants are required. The emerging guidance in this regard suggests that investing in more than one cycle of grants, with amounts sufficient to make a difference, is going to be key for enhancing the impacts of CDD in FCS, if this is fiscally affordable. Of course, local contextual factors will need to influence this decision, but broadly the retrospective feedback of many task teams was that one cycle of grants is simply not enough. Generally, one cycle makes sense only when there is a significant budget limit and when it is felt that more than one grant cycle per community may be perceived 54. The other danger of a closed menu is that communities will feel forced to select one of the available sub-projects, even if it is not their priority, just to secure funds. But recommending a “wide-ranging” menu based on good PRAs, as suggested above, would guard against these distortions. 55. Table A3-5 in annex 3 provides some generic factors influencing the decision on number of grant cycles. 56. As mentioned earlier, most projects in the stocktaking sample provided only one cycle of (block) grants, and the grant sizes varied greatly (see table A2-5 in annex 2). Section IV: Choices and Emerging Guidance in 10 Design Areas • 29 as unfair and spur conflict.57 Otherwise, the local governance and empowerment effects that CDD programs can offer will simply not occur with just one round of grants. This also appears to be the evidence emerging from recent impact evaluation literature.58 Therefore, over time and definitely over subsequent phases of the CDD program, the goal should be to have multiple grant cycles per community. Similarly, the size of the grant has to be sufficient to cover costs of basic infrastructure and large enough on a per capita basis to improve welfare outcomes. It is simplest to calculate allocations or size on a per-capita basis and then extrapolate from there. The medium-term fiscal envelope is also an important criterion when deciding the size of grants and the number of cycles in the project or in the project series because, in some cases, the amount of donor funding and development aid available to an FCS country immediately after a conflict may be much larger than what it can afford on a sustainable basis over the longer term.59 Finally, regarding the selection process, guidance was that a non-competitive selection model is less risky and contentious in FCS. Competition in the grant selection process allows better proposals to be selected and can be perceived as more merit-based. Competition is the norm in the East Asia Pacific (EAP) region and some African countries. However, in FCS there is a greater risk that this could be perceived as unfair or biased against weaker communities (which actually may have more need) and hence can ignite conflict. Therefore, a process that provides a fixed allocation to communities without competition may be a safer and more equitable approach. Within the sample, discussions with task teams revealed that, when it comes to awarding grants to communities, a non-competitive model was chosen, even in EAP. In Aceh (BRA-KDP), this model ensured that grants reached everyone who was affected by conflict; and competition was suspended in Mindanao, Philippines, in order to avoid igniting conflict. 4.8. Community Mobilization and Facilitation Decisions around community mobilization and facilitation need to be carefully made in FCS. As the stocktaking revealed, there was not much discussion around this in project documenta- tion. However, the analysis of designs suggests that many choices must be made. When it comes to the community mobilization and sub-project cycle, this could be separated into four phases: 1) preparation and planning, 2) decision-making, 3) implementation, and 4) operations and main- tenance (see table A4-6, annex 4). There are at least 14 decision points for design across these four phases. While the broader stocktaking report discusses all of them, we zero in on three that are particularly relevant for FCS: ensuring buy-in among stakeholders, ensuring a good pool of facilitators, and selecting community development committees (CDCs). First, the stocktaking, as well as TTL interviews, point to the importance of trust and of obtain- ing buy-in from critical stakeholders at local and higher levels when setting up the facilitation 57. As some TTLs noted, this “perceptions risk” with multiple grants per community is not insignificant because the unfair distribution of benefits from development programs could well have been one of the sources of the original conflict or tension in FCS countries. Starting with one grant per community in such a situation may be a safer option, even if it is less optimal from the local governance and capacity-building perspectives, at least for the first phase. 58. See, for instance, Barron (2011) and Wong (2012). 59. Afghanistan is a case in point here, where the NSP program at one point covered close to 20% of the government’s entire development budget. But with the upcoming U.S. military withdrawal and scaling back of international donor funding to the country, this kind of funding is likely to be untenable. 30 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Figure 11. Hiring Facilitators: Typology, Advantages, and Disadvantages Hiring Facilitators Outsource Insource Private Sector NGOs Government Staff Independent Community Consultants Agency Resource People Examples Afghanistan, Afghanistan, Angola, Bosnia Colombia, Uganda, Afghanistan III CAR, Guatemala, CAR, Somalia, and Herzegovina, Yemen Rwanda, Uganda, Cote d’Ivoire, Cote d’Ivoire, WB&G Iraq, Uganda Indonesia, Rwanda, Sri Lanka, Uganda Advantage • Experienced • Experience and • Sustainable • Sustainable • Cost effective facilitators knowledge of capacity capacity • Creates • Existing capacity political economy • Neutral party community jobs • Existing capacity • Trust Disadvantage • Cost • May be difficult • Cost • Not trained in • Low sustainability to fire when early years so • Limited availability underperforming may require • May lack support • Distrust heavy training, network and • Drawn away from mentoring, often political economy other tasks by NGOs knowledge • Distorts wages • Elite capture • Low capacity and community leadership arrangements. Neutrality and transparency in selecting the “mes- sengers” and “messages” become key factors here, especially in volatile post-conflict situations, in order to avoid creating new tensions and/or conflicts. While some buy-in is obtained through the joint government-World Bank design decisions (e.g., the national institutional implemen- tation arrangements), the advice is to pay much more attention to deliberately planning the communication, based on stakeholder analysis at all levels. Testing the procedures and messages, either in pilots before full implementation or during implementation with annual adjustment of the operating manuals, is even more important than in non-conflict situations, as weak capacity or resumption of conflict may create unexpected bottlenecks. A second key decision point relevant in an FCS setting is choosing the facilitation arrange- ments. Here, one can essentially choose to in-source capacity within the government by using existing government staff, hiring consultants, or directly recruiting community professionals. Alternatively, one can outsource the facilitation function to NGOs or private consultants (see figure 11). Because facilitators are the key front-line workers who implement the sub-project cycle and manage tensions at the community level, their pivotal role cannot be overemphasized. Section IV: Choices and Emerging Guidance in 10 Design Areas • 31 Figure 12. Options for establishing community committees Elect Select Appoint (via gov’t) Combination of these (via community options plus quotas • Democratic option • Capable leaders or inclusion criteria • If by secret ballot, all • Capable community would be selected by community members leaders would be government • Allows for normally would have equal selected and would • But in FCS, there is underrepresented voice/power have trust of community danger of lack of trust groups to participate • But, if secret ballot, • But high likelihood by community in in decision making might introduce a that selection would government appointed (e.g., women in totally new process be of existing elites, leaders Afghanistan) • And if in open meeting, who may not be • But, they may not have social pressure for some development oriented “real voice and agency” candidates and so don’t influence consequent elite decisions capture possible • Will need a lot of capacity building Communities’ need for facilitation after conflict and during crises is high, sometimes greater than anticipated at design. Further, depending on the social dimensions of the conflict (e.g., if it was ethnicity based), the choice of who facilitates can be very sensitive. In FCS, therefore, facilitators need to be selected carefully and trained in social and conflict analysis, conflict pre- vention, and mediation, and special attention should be given to gender balance, skill sets, safety and security, supervision, and monitoring. Implementation reports often show that attention to all these aspects takes up a lot of design energy. However, most FCS countries rarely have the capacity to do facilitation in-house—they often lack a ready pool of skilled public professionals and there are safety concerns as well. Therefore, they rely instead on NGOs, which may already have trained facilitators in pilot or humanitarian programs. In fact, only two projects, Sri Lanka and Uganda, carried out facilitation in-house. How- ever, there is competition for these skilled facilitators, and key questions arise in FCS about what level of qualification one is willing to accept and how to compensate in training and supervision. These factors often mean that outsourcing facilitation (at least in the first instance) is the most efficient or practical approach to obtain experienced social and technical facilitators. However, in the longer term, the costs of outsourcing as the program scales up can be prohibitive, in which case a gradual shift to either a government in-sourced model or the use of community resources was advised, provided the government had sufficient capacity as well as the human resource systems to manage these in-house facilitators.60 The final design point is the selection of the CDCs that manage the sub-project decision- making. As the ultimate face of the project at the community level, these CDCs need to be seen with credibility, trust, and legitimacy for the program to work. Choices here are either to set up a democratic election process or allow selection or appointment either by the community or 60. The latter is cheaper because of the ability to contract individuals directly (instead of firms or organizations), lower remuneration rates, and little or no travel costs. All of these can reduce overhead costs associated with facilitation. 32 • Designing Community-Driven Development Operations in Fragile and Conflict Situations through government. Any of these processes can also be combined with some inclusion criteria or quotas (see figure 12). Separate sub-committees can also be set up under the overall CDC. The choice on how to go about it depends on several factors, including culture, objectives, existing heterogeneity of communities, capacity, and cost. For CDCs, governance arrangements—includ- ing appointment, duration of tenure, rules of conduct, and so on—have to be established and made public so that communities can trust their authorities. Governance (including Fiduciary) and . 4.9.  Monitoring and Evaluation While design choices around governance and monitoring and evaluation (M&E) are similar in CDD, design teams in FCS need to adjust to the weaker capacity, risk, and trust contexts. As noted earlier, the stocktaking shows little variance in the FCS sample in terms of the menu and choice of governance (including fiduciary and social accountability elements), safeguards, and M&E approaches and instruments, compared to what one normally finds in a non-FCS CDD cohort. (Figures A4-2, A4-3, and A4-4 in annex 4, as well as tables A3-11 and A3-12 in annex 3, provide some of the generic elements and options available for CDD design in these areas, as well as summaries of how our sample of 17 projects was distributed on some of them.) However, the choices made are FCS-driven; in particular, fiduciary arrangements for the flow of funds are crucial to ensure that resources reach the ground as quickly as possible to respond to the urgent needs faced by FCS communities. While use of country systems is advised, using independent agencies may well be required in this regard for both efficiency and transparency because quick disbursements are a peace dividend tool and delays in disbursements may (re)ignite conflict. Similarly, one needs to take advantage of the built-in flexibility around community procurement and disbursement to allow for community-based procurement mechanisms, for direct sourcing from the usually limited number of suppliers in FCS contexts,61 and for direct disbursements based on progress in implementing the sub-project rather than receipts. The lack of a banking system may well require the use of innovative alternative mechanisms for making payments, such as the use of hawalla dealers62 in Afghanistan (see box 2). However, all task teams acknowledged that M&E and governance require special attention in CDD design and should be much more robust in FCS. M&E, in particular, is often treated as a stepchild in project design Bank-wide, and CDD in FCS is no exception. Developing a culture of learning by doing by including better impact studies, documenting experiences with pilots, and feeding this into future design is a critical need for CDD projects and is further enhanced in FCS contexts.63 In interviews, task teams stressed the need for stronger use of participatory commu- nity-based governance and monitoring systems because they also build local capacity and are a promising avenue in more insecure areas. Technological advances may also ease the data analysis burden on M&E, and some projects indeed have made good use of this—for instance, use of 61. Examples of special procedures for community participation in procurement include local shopping, local bidding, direct contracting, off-the-shelf purchases, and community force accounts. See Guidance Note for Design and Management of Procurement Responsibilities in Community-Driven Development Projects (World Bank 2012). 62. Hawalla refers to traditional Afghan money lenders, who carry and transfer cash from one place to another. They operate as a parallel banking system and rely on a network of money brokers. 63. This was also a key recommendation from the recent Policy Research Report, Localizing Development: Does Participation Work? (Mansuri and Rao 2012). Section IV: Choices and Emerging Guidance in 10 Design Areas • 33 Box 2. Hawalla Dealers: Adapting to the Local Banking System in Afghanistan The NSP program in Afghanistan provides a good example of how one needs to adapt fiduciary arrangements to the local context. Overall, the flow of funds in NSP is from the International Development Association (IDA) to the Central Bank (DAB) to the independent financial management agent (FMA), which is in charge of managing funds for the project. The FMA maintains a float account from which to make block grant payments into CDC accounts that are maintained in provincial DAB branches or commercial banks. Considering that large parts of Afghanistan did not have any recognized banks or banking systems during the early years of the NSP, it used the hawalla system to transfer block grants to the communities in certain provinces. There have been two distinct uses of the hawalla within the NSP: 1. Former NSP oversight consultants (OC) and then the former NSP management support consultant (MSC) used contracted hawalla dealers to transfer block grant funds from the NSP headquarters in Kabul directly to the communities in select areas. 2. Certain CDCs used a hawalla dealer when they withdrew block grant funds from their DAB bank accounts and entrusted the funds to the hawalla dealer for their safe transfer from the provincial center (where the provincial DAB bank branches are located) to the com- munities, project sites, or other locations (such as where the goods for the project need to be purchased). Since the opening of DAB provincial branches, however, the use of the hawalla dealer for transfers, described in bullet 1, was reduced to only one province. Further, following a significant case of fraud involving a hawalla dealer in this province, the rules regarding the hawalla dealers were tightened. Oversight increased and communities are encouraged to open accounts with private banks in districts, so that the commute for cash transactions from the community to the bank is not as far. Source: NSP project documentation. GIS (geographic information systems) in Yemen SDF, GPS-enabled mobile phones in Afghanistan NSP, and geo-tagging in Mindanao. Regarding evaluation, TTLs strongly endorsed that it is time to ensure that meaningful evaluations (rather than those that use a wrong yardstick) are car- ried out for national programs that are already delivering the goods. In turn, such evaluations will provide a basis for policy support and the parallel development of the necessary policy framework. All CDD projects need to be designed so that they can be replicated and scaled up, but simultaneously M&E systems need to be designed to ensure that data are available to allow replication and scale-up with adjustments as needed. However, one needs to be cognizant that insecurity (as well as lack of financing) poses great risk to obtaining data, especially for a base line for start up, when insecurity may be the largest. Moreover, randomization may be politically impossible in FCS, which makes random control trials unfeasible. In such situations, one has to “think outside the box,” as one TTL put it, and develop alternative methods for evaluation.64 When it comes to governance and social accountability, FCS is assumed to have much higher governance and anti-corruption (GAC) risks than other countries. Thus, it is the task teams’ responsibility to identify, as clearly as possible, the nature and extent of those risks at coun- try, sector, and project levels. However, as gleaned from project documentation, even if risk 64. OPCS has issued a guidance note on results frameworks and M&E, which has a special section for FCS that is useful for CDD teams to look at in this regard (World Bank 2013d). More generic guidance on impact evaluations can be found in Gertler et al. (2011). More generic guidance on impact evaluations can be found in Gertler et al (2011), and on monitor- ing by non-state actors in World Bank (2013a). 34 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Figure 13. Examples of Peace Building and Social Cohesion Elements Cultural/sports/ Human security plans or Targeting/special weightage solidarity building conflict filters to conflict victims activities (e.g., CAR, Sri Lanka) (e.g., Indonesia) (e.g., Cote d’Ivoire) Housing rehabilitation for Training on peace building Social cohesion grants IDPs, returnees and reconciliation (e.g., Philippines) (e.g., Kosovo, Rwanda) (e.g., Uganda) assessments are carried out, they tend to focus on standard financial management and pro- curement, and may not carefully assess or take into account GAC agendas beyond fiduciary issues. During interviews, task teams emphasized that some form of grievance redress mecha- nism (GRM) should be incorporated in the project, however simple, so as to have an avenue by which to provide feedback, make complaints, and resolve disputes. Failure to be transparent can possibly weaken already fragile trust among community members and can even cause conflicts. A World Bank forthcoming (2013) paper that proposes the creation of a new paradigm for fidu- ciary management of CDD further argues that these arrangements in FCS need to be simple and flexible, build on community institutions, use technology where possible, and strengthen the engagement with CSOs (community support organizations).65 4.10. Peace Building and Social Cohesion Elements Last but not least, one design area that emerges as an important separate category in FCS is the inclusion of some element of peace building or social cohesion. As noted in the stocktaking summary, even though most of the CDD-FCS projects in our sample did not explicitly include peace building or social cohesion in their objectives, many had some components or activities addressing these goals (see figure 13 and table A3-3 in annex-3 for a summary). Given that the defining feature of FCS is the urgent need for restoring peace and social cohesion, it is important that future task teams attempt to cater to these goals in some form. However, this first requires that peace-building work is embedded by mapping a more complete picture of different stake holder interests and incentives, looking for possible alignment, and formulating a more holistic theory of change or alignment of these incentive systems before interventions are designed. This is particularly key as countries move out of immediate post-conflict environments to more stable state-rebuilding phases or when faced with a fragile (as opposed to a post-conflict) context. 65. See World Bank (forthcoming, 2013). Section IV: Choices and Emerging Guidance in 10 Design Areas • 35 However, there is little evidence on which to base guidance for peace building and social cohe- sion elements. The evidence for recommending peace-building elements is scant. The projects in the stocktaking sample did not carry out such evaluations, with the exception of BRA-KDP in Aceh, which was a one-year targeting program with a positive impact. In general, beneficiary satisfaction is high for the targeted grants, but little is known about the sustainability of such programs. For other components, projects relied on NGOs to deliver them, and these NGOs did not have good impact evaluations. Undoubtedly, this is an extremely important area for future research—and some of the newer projects are indeed pushing the envelope in this regard. For instance, in Senegal, a peace-building project will precede a CDD program to provide small grants to communities to work on peace-building initiatives. In the Philippines, a conflict window will be included in the national CDD program (NCDDP) to provide additional grants to communities to increase social cohesion in conflict zones, building on the ARMM Social Fund experience in Mindanao. Neither of these programs has, however, designed their interventions with a docu- mented theory of change based on an analysis of stakeholder incentives, so expectations for the durable impact of these interventions should be tempered. What is more, as one TTL cautioned, task teams need to realize that a CDD program is, at the end of the day, just a project—in other words, a bureaucratic instrument used by the state to transfer resources—where its ability to have a transformative impact on peace building and social cohesion will always remain limited. Therefore, one must have a much broader strategy of aligning with, and influencing, different stakeholder interests and incentives in order to truly influence social cohesion and peace build- ing in FCS. Some Overall Guiding V.  Principles for Design While the previous section attempted to lay out the choices available and some decision-making criteria for different design areas, the interviews undertaken as part of the stocktaking study also revealed some cross-cutting principles that task teams repeatedly mentioned. Four key design principles are discussed below. One, always be responsive to context. The statement, context matters, has become somewhat of a cliché in development in general and CDD in particular. But the enormity of the influence of context in FCS merits repeating it here. As noted in each of the design areas above, almost every design choice was somehow influenced by contextual factors. In FCS, where context is often fluid, knowing the background and designing for a constantly changing environment is even more important. Within context, understanding the most important drivers of the underlying conflict or fragility, especially the social dimensions, at national and local levels was mentioned by several task teams as particularly important and often overlooked. Analysis of break-downs and their inter-relationships will determine the long-term strategy for government and donor partners, namely, the timing and the sequencing of interventions and programs to be put in place. In this regard, as one respondent noted, it is important to realize that “CDD is not a strategy but a tactic.” That is, CDD is an approach to change and gets implemented as a program with an operational focus. But it is not per se a strategy for development; it is part of a toolkit of options that countries can undertake. In order to ensure that the long-term strategic objectives are achieved, CDD needs to be seen in context of a nexus of programs that government and donors may put in place to reach those objectives. The CDD approach needs to be embedded in the overall strategy and suite of programs. Sensitivity to context, long and short term, will in turn argue for flexibility and contingencies in the design.66 Two, think long term. Rarely do design teams in FCS start by having the time to design for the long term because they need to deliver results fast. Bank project horizons are normally five years and staff rotation in FCS is high. Yet, the meta-analysis of the impact evaluations of Bank- financed CDD projects67 revealed that projects in that sample had an average duration of about 11 years and this trend of extensions, additional financing, and repeater projects is seen across the CDD portfolio. So in practice, design teams have a longer-term planning horizon, even if initially they may lack the incentives to formulate longer-term scenarios for national scaled-up programs. Given this, the lesson for task teams is to manage the trade-off of a short-term versus a long-term planning horizon innovatively. They can respond quickly, but always keep the long 66. New CDD projects in Myanmar and Indonesia have, for instance, included zero components to allow for a quick redeployment of resources should circumstances change—in their cases, this refers to the onset of unpredictable, but regular, disasters. 67. See Wong (2012). 36 Section V: Some Overall Guiding Principles for Design • 37 term in mind (e.g., start with one project per community, but leave a second cycle of grants as a possibility if more funding arrives; start with an independent agency, but bring in LGUs as much as possible, etc.). The theory of change should be to plan for the evolution of a CDD project into the government’s participatory local planning and/or service delivery mechanism. Where possible, therefore, teams in FCS should also try and get management and government com- mitment upfront that the CDD operation is going to be the first phase of a long term program68 because CDD, at its core, is about behavioral, structural, and social change, all of which takes time and requires strong continued political support and policy changes. In FCS, this planned change process overlays a period of constant and uncertain change, often not desired by the participants. To rebuild trust that this can be an orderly change process, in which the participants have voice and agency, demands time. In addition, operationally, the CDD approach will require the establishment and development of many logistical systems, which can be quite difficult. The example of the use of the hawalla system of money lenders (used in Afghanistan for transport- ing cash to villages because there was no functioning banking system) is one of the more telling examples. Insecurity, lack of infrastructure at all levels, movement of large population groups, and healing of traumas all take time and slow down the speed of the development of a CDD approach. Hence, pacing and evolving the program, based on learning from success as much as failure, over a longer period is important in FCS. Three, keep things simple. CDD programs often suffer from the danger of what one respondent called the “coral reef effect”—that is, the risk that layers and layers of themes and activities (often pushed by different donor demands) get added on to program designs. This is often a recipe for failure in FCS contexts since they, in general, have much lower human, economic, and social capital. The institutional capacity is also usually low. This all requires innovative and creative design work, especially at the level of procedures which meet the spirit of the law, but may not necessarily be cut and pasted from other countries. As one respondent noted, “A 250- page community-level financial management manual in English is not practical and will create bottlenecks.69 Similarly, community-level procurement reports that barely literate community leaders must fill out shifts the burden to the facilitators and removes accountability from where it belongs. The simplicity principle would likewise apply to all manner of design elements— safeguards, flow of funds, and sub-project selection to name just a few.70 Governments often express their desire to limit the number of donor reporting and fiduciary requirements they need to deal with, and they demand that donors pool funding and use a single reporting system in order to make things simpler for the few staff who are conversant in donor lingo. Such coor- dination should also extend to competing World Bank interventions at the community level: one community platform would keep it simple(r) for the communities. Four, always think scale. For the CDD approach to become the standard acceptable and desired approach for service delivery and local governance, the design team has to ensure that it 68. This would also be consistent with the principles of the new deal and current change management discussions relat- ing to FCS within the World Bank. 69. And this is probably true in any CDD setting, be it FCS or not. See, again, World Bank (forthcoming, 2013) for new thinking in this regard, particularly around fiduciary arrangements. 70. Once the core systems are in place and prove to be functioning, then one can “add on the bling,” as needed, to quote one respondent. 38 • Designing Community-Driven Development Operations in Fragile and Conflict Situations demonstrates the benefits of the approach at a scale that attracts attention and can be show- cased. Only when operating at a sufficiently large scale can different aspects of the approach and the delivery be tested in different environments in the country and sufficient learning take place to plan for larger scale (horizontal scale-up) or deeper establishment of the approach (vertical scale-up). For instance, CDD programs can ensure that all sectors working in rural space use its local development plan and the community-level management structures in their plan- ning and delivery. This visioning of the CDD program going to scale is related to thinking long term and it has implications on the overall piloting, evaluating, sequencing, and evolution of objectives of the program. These four principles are both simple and well known,71 but often are tough to apply. The dif- ficulty of applying such principles lies, above all, in the incentive structure for both governments and the Bank’s design teams. When faced with difficult problems and the opportunity to apply resources to solutions, it is difficult to limit a program design to a simple one, given the complex- ity of the underlying FCS situation and the demands on such teams to find solutions. Yet, this temptation should be tempered by analyzing the complexity in depth, by isolating the most important drivers of the underlying conflict or fragility, and by formulating a theory of change that in the long term will have the most likely effect of influencing these drivers of conflict and of fragility through the CDD approach. In no way will the CDD approach be the only answer. Complementary policy and program solutions will have to be put in place. But defining the theory of change and the contribution that the CDD program can make in the immediate and longer term will enable design teams to be most effective.72 Similarly, a wider effort is needed within the Bank to address the structural incentives at work that cause final designs of CDD-FCS projects to violate the four design principles outlined above (context responsive, longer-term oriented, simple, and scalable).73 71. They are, in fact, very similar to the core principles outlined in the new deal for FCS mentioned earlier. 72. Within the cohort of FCS projects reviewed, the Afghanistan program comes closest to articulating its theory of change from the start, even though it faces its own transition and sustainability issues as the geo-political context there will change in 2014. 73. Reviewers of this paper suggested that there is a lack of quality control within the Bank and that establishing such a mechanism would go a long way to making the portfolio be more reflective of the lessons learned and less at risk. Looking to the Future: VI.  Areas for Further Research and Action This design stocktaking study should be seen as a first step in a deeper and longstanding effort to improve our understanding and guidance on designing CDD programs in FCS. As we men- tioned in the introduction, this study is a modest effort to fill a large gap in our overall design guidance, which started by looking back at what past projects have done. Our conclusions are based on a desk review of project documents describing design choices made in the beginning of a project cycle. They do not draw from CDD impact evaluations from those projects, which could provide evidence of one design choice being better than another, or from studies and reviews, which compare CDD to other development responses in Bank-funded projects. What evidence we have on how things fared in real life are drawn from select interviews with TTLs, task teams, and ICRs. It is, by definition, mostly anecdotal and subjective. Therefore, in looking forward, much more work clearly is needed to better guide our work on CDD in FCS and build upon its “science of delivery” to quote World Bank President Kim.74 Below we suggest some areas that deserve special attention. First, we need to better understand fragility and CDD design. As noted, our sample, although representative, ended up with only a few projects in fragile contexts. The majority was in post- conflict situations, which is true of the larger CDD-FCS universe as well. Therefore, if we are to throw more light on the specific situation of fragility, we need a more targeted study focusing only on those situations and with a much larger sample. We need to understand how design should change in cases that do not produce the immediate peace dividend and infrastructure reconstruction needs (which arise directly after a war or prolonged conflict), but are instead frag- ile due to lack of governance, institutional instability, or transition. The countries in the Middle East and North Africa (MENA) region in the aftermath of the Arab Spring are natural places where such understanding would be extremely valuable, as many of them are considering decentral- ization and CDD approaches in an otherwise fragile socio-political environment. Reviewers for this paper called for strong (and resourced) mechanisms to filter operational lessons on CDD in fragility projects to identify whether there are distinct or different issues, or other approaches that do not appear in CDD-FCS projects. This would be an important area of future focus for both the CDD and FCS communities of practice. 74. As the OPCS (2012) review concluded: “Designing a CDD program in a FCS context is not for amateurs or begin- ners.” Strong, experienced teams with frequent mentoring from senior experienced experts, adequate resources, and strong management support is the ideal combination. While this happens in flagship countries like Afghanistan, it should become the norm if the WB is serious about its contribution to FCS. 39 40 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Second, we must gather much more evidence on how design choices play out in real life. As noted, this study was just a desk review supplemented with task-team interviews. While this allowed us to reverse engineer the design choices made by design teams and, to some extent, validate why choices were made, this exercise was clearly not exhaustive. What is needed mov- ing forward are cross-sectional and longitudinal studies on specific design choices that really matter (e.g., targeting or facilitation arrangements) to see how these played out across countries and over time. Additionally, there should be sufficient resources to carry out more impact evalu- ations of CDD programs in FCS contexts. As a recent meta-analysis study (Wong 2012) of World Bank-financed CDD operations revealed, there are only six impact research evaluations of CDD in FCS. These can barely provide the guidance and knowledge needed for future CDD-FCS work, mandating a concerted effort to increase the evidence base of CDD-FCS program impacts and results. This can only come about through a partnership between the Bank, other donor agen- cies (e.g., AusAID, DFID) working on CDD-FCS, as well as academic and research organizations (e.g., J-PAL, 3iE, etc.). Third, we need to better develop our operational toolkit, specifically on peace building, social cohesion, and conflict resolution. The design guidance section above already noted the impor- tance of including these elements in the design of CDD programs in FCS contexts. And our stocktaking revealed some examples of what projects in the past have done. But if one looks at aggregate evidence of CDD in this regard it is not very encouraging.75 There is a lack of a holistic theory on how peace and cohesion building fits into the CDD project modality with far reaching impact, beyond the disbursement of grant activities, and there is a dearth of proven instruments, tools, and approaches to promoting peace building, social cohesion, and conflict resolution at all levels. Many more pilots need to be funded and these need to be evaluated rigorously to learn what might work under which circumstances.76 Finally, there is a clear case for greater south-south exchange, donor harmonization, inter- nal World Bank harmonization and quality control, and alliance building around CDD in FCS. While it was not a clear focus of our stocktaking, the analysis revealed a strong desire across task teams to have more south-south exchange, specifically around CDD projects in FCS contexts. Much of the existing learning of CDD has been driven through lessons from flagship programs, such as Indonesia, Brazil, Philippines, and India. Learning from FCS is much less common, and it was clear from our interviews that, barring some exceptions (e.g., Afghanistan), very few CDD- FCS projects had been in contact to share experiences and lessons with each other. A related issue, mentioned above, is that teams designing a CDD operation should also outline a strat- egy for scale-up, institutionalization, and sustainability of the CDD approach. This exit strategy requires start-up efforts that include donor harmonization, pooled donor funding mechanisms, alliance building in and outside of government, and intensive work with all the sectors. It also requires strategic investments of time and resources by the World Bank teams and, above all, management to ensure such coordination in house, in country, and with development partners. In particular, the significance of having a strategic consensus and framework that embeds 75. See for instance Wong (2012) and Mansuri and Rao (2012). 76. The increasing application of behavioral science and evaluation in development could be made use of in this regard as trust and inter-group behavioral dynamics have been recognized as key societal dimensions of fragility. See Marc et al. (2012). Section IV: Looking to the Future: Areas for Further Research and Action • 41 the CDD project with the country assistance strategies (CAS) and the key role of country directors and others in framing CDD operation within their post-conflict recovery programs cannot be emphasized enough. Moreover, to ensure that CDD projects in FCS conform with key design principles and lessons learned to date, a stronger quality assurance and review process for CDD in FCS, coordinated through the CDD community of practice and with the backing of both sustainable development and human development management, is something that several reviewers of this study recommended. VII. Concluding Thoughts This design stocktaking has revealed some concrete areas of design guidance for future CDD- FCS task teams. While we began this stocktaking by noting that our objective was not to pro- vide prescriptive guidance, as the previous section shows, this study has indeed provided a set of core recommendations, based both on the study findings and the professional judgments of the large cohort of CDD practitioners that were interviewed during the review. (Box 3 sum- marizes these core recommendations.) Given all the caveats about methodology, sample, lack of impact data, and so on, these should not be treated as ideal types or best practices that each task team needs to live up to, but rather a first stab at core elements that should prevail. The introduction of the new OP 10.0077 on investment lending provides added flexibility and guid- ance to put these recommendations into effect (see box 4). In conclusion, it is important to reiterate the relevance of CDD in FCS. This stocktaking was not meant to be evaluative in nature nor was it meant to be a how-to guide. Yet, we feel that there is enough evidence in this desk review to claim that the CDD model is a good fit for the FCS contexts in the sample because, indeed, the projects delivered on what they set out to do. This is because the CDD approach emphasizes, above all, that it is a not a single “right- off-the-shelf” approach, but one which is requires adjustment to context—in this case, the specific requirements and complexities of fragile, conflict and post-conflict countries. And the tremendous variety in design arrangements we found testifies to its adaptability. At any point in time, the communities and states with FCS contexts lie on various points of the crisis-devel- opment spectrum. In some, the entire population may have fled the fighting, while in others the population may not have been exposed to violence or an influx of IDPs. Given the dynamic and unique nature of these contexts, our best bet perhaps is to empower communities to drive the reconstruction process in strong collaboration with a capacitated local state, in this way ensuring that actual needs are met. This is precisely the philosophy behind a CDD approach. We hope therefore that as the Bank and client governments move forward in thinking through their development strategies in FCS contexts, they recognize the valuable role CDD can play. And if so, they are able to benefit from this effort to learn from past experiences that inform how to better design CDD programs in these FCS contexts in the future. 77. OP/BP 10, Preparation of Investment Project Financing: Situations of Urgent Need of Assistance or Capacity Con- straints. 42 Section VII: Concluding Thoughts • 43  ummarizing the Core Recommendations Emerging from the Stocktaking Box 3. S on CDD-FCS Design In terms of preparatory diagnostics, understanding the context through political economy and conflict analysis is crucial. This can well take the form of ongoing informal stakeholder consultations. Whether or not formal studies are feasible is determined by the urgency of rapid progress, budget, availability of data and the need for political buy-in. In terms of overall objectives, limiting to one objective in the first phase and then setting building blocks in place for the development objectives in subsequent phases is crucial. Which objectives become primary are determined by the specific nature of the conflict or fragility, the need to keep things simple, capacity constraints, and expectations (or mandate). When determining the final agency in charge of implementation at the national level, it is important to assess whether the national agency has the capacity and power to protect and deliver the program (particularly if it is in the hands of a government agency), and whether the agency is neutral (particularly if it is outsourced). In the latter case, an exit strategy needs to be planned and considered as well. For sub-national institutional arrangements, consider a hybrid between community and LGU with adequate capacity building and empowerment built into the design. Past experience suggests that investing in some form of learning by doing is well worth it in FCS, and yet it should not be considered a one-off exercise, but rather a continuous endeavor throughout the life of the project as it goes to scale. Targeting decisions can be highly sensitive and different factors should be weighed in at different levels. The macro level should be determined by political economy analysis that assesses whether teams should spread across or zoom in on particular regions. At the intra- community targeting level, communicating the rules of the game transparently is of utmost importance. Finally at the intra-community level, targeting should be used sparingly especially in the first phase. Use a wide-ranging closed menu—a hybrid arrangement of sorts—for the first phase, which can then move to a full open menu over time. Moreover, when it comes to the number of grant cycles, task teams should plan to repeat block grants to cement participation and accountability. The importance of trust and obtaining buy-in from critical stakeholders cannot be emphasized enough when setting up the facilitation and community leadership arrangements at local and higher levels. M&E and governance require special attention in project design and they need to be much more robust in FCS, since GAC risks are much higher in these contexts In order to influence social cohesion and peace building in FCS, task teams must have a much broader strategy of aligning with and influencing different stakeholder interests and incentives. 44 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Box 4. Simplified Procedures and Flexibility in the New OP 10 These are some of the flexibilities defined in paragraph 11 of the OP: • Condensed process: The normally sequential stages of identification, preparation, and appraisal may be consolidated; and the decision to authorize negotiation may be taken after a single consolidated review of a complete negotiations package. • Accelerated business standards: In addition to a consolidated decision point, there are accelerated business standards with shorter turnaround times in the case of a new operation following these instructions or in the case of additional financing: – Decision review: Advisors are given 3 business days to provide comments instead of the normal 5 business days – Board submission: Packages are submitted 10 business days before board presentation instead of the normal 18 business days. • Safeguards: If environmental and social requirements are deferred, based on Para 11 (a) and the RVP’s decision, the Bank prepares a project-level safeguards planning document referred to as an “action plan.” The safeguards action plan is a planning document that provides a time-bound plan setting forth the steps and the sequential planning and coordination for project activities and the prepara- tion by the Borrower of the relevant safeguards instruments to ensure compliance with these safeguards requirements. • Procurement methods and contracting arrangements: One new procurement method was added—Framework Agreements, in- troduced as an alternative to shopping or NCB methods—and application of several existing ones has been expanded to facilitate procurement under emergency situations (direct contracting/single-source selection and force account). Force accounts have always existed, but their application has been expanded to urgent repairs that require attention to prevent further damages, or works to be carried out in conflict-affected areas, where private firms may not be interested. In addition, similar facilitating-type revisions were made in two specific cases of use of procurement agents and construction managers. Source: OP/BP10.00 Instructions. References Adam, S. 2009 (Unpublished draft). “Rationalizing Design Choices for Community-Driven Development in the Philippines.” East Asia Social Development Unit. World Bank, Washington, DC. Afghanistan Reconstruction Trust Fund (ARTF). 2012. Afghanistan Reconstruction Trust Fund at a Cross-Roads: History and Future. Oslo, Norway: Scanteam. Barron, P. 2011. “CDD in Post-Conflict and Conflict-Affected Areas: Experiences from East Asia.” Background paper, World Development Report 2011. World Bank, Washington, DC: World Bank. Beath, A., C. Fotini, and R. Enikolopov. 2012. “Empowering Women: Evidence from a Field Experiment in Afghanistan.” Policy Research Working Paper 6269, Impact Evaluation Series 76. World Bank, Washington, DC. Binswanger-Mkhize, H., J.P. de Regt, and S. Spector. 2010. Scaling up Local and Community Driven Development (LCDD): A Real World Guide to Its Theory and Practice. Washington DC: World Bank. Bryld, E., and C. Kamau. 2011. “Evaluation of the Community-Driven Recovery and Development Project.” Prepared for the Danish Refugee Council. Copenhagen, Denmark: Tana Copenhagen. Green, K. 2012. “Using Frameworks for Addressing Environmental and Social Safeguards in CDD Projects.” PowerPoint presentation. World Bank, Washington, DC. Jennings, R.S. 2010. Promoting Stability and Peace at the Local Level in the MENA Region: Operational Findings on Local Service Delivery Initiatives in Conflict and Post-Conflict Areas. Washington, DC: World Bank. Jorgenson, S. 2005. (Unpublished draft). “Framing the Logic of CDD Operations: Towards a Project Theory for CDD Design and Evaluation.” Social Development Department. World Bank, Washington, DC. Krogh, E. 2012. “Mapping of Community-Based Approaches in South Sudan: Background Study for the Design of the Local Governance and Service Delivery Programme (LGSDP) of the Government of the Republic of South Sudan. Produced in cooperation with the World Bank, Washington, DC. Mansuri, G., and V. Rao. 2012. “Localizing Development: Does Participation Work?” Development Economics Research Department Policy Research Report. World Bank, Washington, DC. Marc, A., A. Willman, G. Aslam, M. Rebosio, and K. Balasuriya. 2012. “Societal Dynamics and Fragility: Engaging Societies in Responding to Fragile Situations.”In New Frontiers in Social Policy. Social Development Department. Washington, DC: World Bank. Operations Evaluation Department (OED). 2005. “The Effectiveness of World Bank Support for Community-Based and -Driven Development: An OED Evaluation.” World Bank, Washington, DC.. Owen, D. 2006. “Community-Driven Development in the Context of Conflict-Affected Countries: Challenges and Opportunities.” Social Development Department Report 36425-GLB. World Bank, Washington, DC. http:// siteresources.worldbank.org/INTCDD/Resources/CDD_and_Conflict.pdf. Serrano-Berthet, R., L. Helling, J. Van Domelen, W. Van Wicklin, D. Owen, M. Poli, and R. Cherukupalli. 2010. “Lessons from Africa.” In Local and Community Driven Development by H.P. Binswanger-Mkhize, J.P. de Regt, and S. Spector. 2010. Washington, DC: World Bank. Strand, A., H. Toje, A.M. Jerve, and I. Samset. 2003. (unpublished). “Community Development in Contexts of Conflict. Concept paper commissioned by Environmentally and Social Sustainable Development (ESSD), World Bank, and Chr. Michelsen Institute Development Studies and Human Rights, Washington, DC. Wicklin, V. 2012 (unpublished). “Guidance Note: Enhancing Governance and Accountability in Community Driven Development Operations.” Draft paper. World Bank, Social Development and Agriculture and Environmental Services Departments, Washington, DC, March 2013. Wong, S. 2012. “What Have Been the Impacts of World Bank Community-Driven Development Programs? CDD Impact Evaluation Review and Operational and Research Implications.” Social Development Department Report. World Bank, Washington, DC. 45 46 • References World Bank. 2008. “The Political Economy of Policy Reform: Issues and Implications for Policy Dialogue and Development Operations.” Social Development Department Report, no. 44288-GLB. Washington, DC: World Bank, SDD. ———. 2010a. “Sri Lanka: North East Local Services Improvement Project (NELSIP).” World Bank, Washington, DC. http://documents.worldbank.org/curated/en/2010/04/12309679/sri-lanka-north-east-local-services-improvement- project-nelsip. ———. 2010b. “How, When and Why to Use Demand-Side Governance Approaches in Projects.” Social Development Department How-To Notes. World Bank, Washington, DC. ———. 2011. World Development Report 2011: Conflict, Security and Development. Washington, DC: World Bank. ———. 2012. “Community-Driven Development Projects: A Review of Governance Issues.” Operational Policy and Country Services (OPCS) Network. World Bank, Washington, DC. ———. 2013a. “Monitoring through Beneficiary Feedback: A Stocktaking of Monitoring by Non-state Actors in World Bank-Funded Projects.” Social Development Department. World Bank, Washington, DC. ———. 2013b. Global Monitoring Report: Rural-Urban Dynamics and the Millennium Development Goals. Washington, DC: World Bank. https://publications.worldbank.org/index.php?main_page=product_info&products_id=24503. ———. 2013c. “Stop Conflict, Reduce Fragility, and End Poverty: Doing Things Differently in Fragile and Conflict- Affected Situations.” Feature Story (online). http://www.worldbank.org/content/dam/Worldbank/Feature%20Story/ Stop_Conflict_Reduce_Fragility_End_Poverty.pdf. ———. 2013d. “Results Framework and M&E Guidance Note” World Bank, Washington, DC. http://intresources. worldbank.org/INTOPCS/Resources/380831-1360104418611/Guidance_Note_Results_and_M&E.pdf (internal link). ———. 2013 (forthcoming). “Dealing with Governance and Fiduciary Issues in CDD Projects in Fragile and Conflict- Affected States.” How-To Notes, Social Development Department. World Bank, Washington, DC. Annexes 47 Annex-1. Definitions of Fragility and Conflict According to World Development Report 2011: Conflict, Security, and Development (WDR 2011), the term conflict denotes “organized violence” that covers the use or threat of physical force by groups, including state action against other states or against civilians, civil wars, electoral violence, and communal conflicts. WDR in turn refers to “fragility or fragile situations” as “peri- ods when states and/or institutions lack the capacity, accountability, or legitimacy to mediate relations between citizen groups and between citizens and the state.” From an operational standpoint, as per its new Operational Policy on Investment Lending (OP/ BP 10.0) the World Bank defines conflict-affected countries as countries that 1) have experi- enced violent conflict in the past 5–10 years, 2) are currently experiencing violent conflict, and 3) are perceived as being at risk of violence. The general term “conflict-affected countries” is not identical to the specific designation of “post- conflict countries,” which the Bank uses to define eligibility for exceptional International Develop- ment Association (IDA) allocations. As per the IDA Exceptional Allocations, a country is considered to be in, or in transition from, conflict by reference to 1) the extent of human casualties, 2) the proportion of the population that is internally displaced, and 3) the extent of physical destruction. A country is defined as a “fragile state” under the same policy if it: 1) currently has a harmonized average Country Policy and Institutional Assessment (CPIA) rating of 3.2 or less, or 2) has the pres- ence of a United Nations and/or regional peace-keeping or peace-building mission (e.g., African Union, European Union, Organization of American States) and has had such presence for at least the past three years (but excluding such a presence whose sole purpose is for border-monitoring operations). The policy also recognizes “fragility within a non-fragile country,” which refers to areas where development actions are limited, due to very low capacity, unstable and rapidly changing condi- tions, and limited or no functional state presence. These localized conditions negatively affect the borrower’s (or Bank’s) operational ability to implement its responsibilities, including—but not limited to—conducting adequate physical supervision of project activities, fully implementing all fiduciary and safeguards responsibilities, or freely interacting with target beneficiary populations. All the above categories of countries and situations (conflict-affected, post-conflict, fragile, and fragile in non-fragile country), as well as “small states” and post-disaster contexts, are subject to simplified processing and additional flexibility in terms of project preparation and implementation. Source: WDR (2011) and OP/BP 10.0 Instructions. 48 Annex-2. Tables Related to the Sample Table A2-1. List of CDD Operations in FCS Studied No. Country Region Project name Fiscal year Product line ICR/ISR rating* 1 Afghanistan SAR Afghanistan: Emergency National Solidarity 2004–present IBRD/IDA S Project (NSP) I–III 2 Angola AFR Second Social Fund Project (FAS) 2000 IBRD/IDA S 3 Bosnia & ECA Community Development Project (CDP) 2001 IBRD/IDA S Herzegovina 4 Colombia LAC Magdalena Medio Regional Development 1998 IBRD/IDA S Project (MMRDP) I–II 5 Central African AFR Support to Vulnerable Groups Community 2009 IBRD/IDA MS Republic (CAR) Development Project SVGCDP) 6 Cote d’Ivoire AFR Post-Conflict Assistance Project (PCAP) 2008 IBRD/IDA MS 7 Guatemala LAC Reconstruction and Local Development 1998 IBRD/IDA MU Project (RLDP) 8 Indonesia EAP Community-Based Reintegration in Aceh 2006 Multi-donor trust S (BRA-KDP) fund 9 Iraq MENA Iraq Consultative Service Delivery Program 2010 Recipient n/a (CSDP) executed (SBPF) 10 Kosovo ECA Community Development Fund Project (I–II) 2000 IBRD/IDA S 11 Philippines EAP ARMM Social Fund Project 2003 IBRD/IDA S 12 Rwanda AFR Community Reintegration and Development 1998 IBRD/IDA HS Project (CRDP) 13 Somalia AFR Community-Driven Recovery and 2008 Recipient MS Development (CDRD) executed (SBPF) 14 Sri Lanka SAR North East Local Services Improvement Project 2010 IBRD/IDA S (NELSIP) 15 Uganda AFR Northern Uganda Social Action Fund (NUSAF) 2002 IBRD/IDA S Project 16 West Bank & MENA Village and Neighborhood Development 2008 Special financing MS Gaza Project (VNDP) 17 Yemen MENA RY-Social Fund for Development (I–IV) 2004 IBRD/IDA S * Rating provided is for Achievement of Development Objectives (where available). S = satisfactory; MS moderately satisfactory; MU = moderately unsatisfactory; HS = highly satisfactory. Notes: World Bank regions are SAR = South Asia; AFR = Africa; ECA = Europe and Central Asia; LAC = Latin America and Caribbean; EAP = East Asia Pacific; MENA = Middle East and North Africa. Source: ICRs from latest phase (e.g., NSP II, Yemen SDF II, etc.) and latest ISR for ongoing projects (only till 2012). 49 50 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A2-2. Number of Projects (167) in FCS by Region South Asia Africa ECA LAC EAP MENA Afghanistan (15) Angola (5) Bosnia & Colombia (1) Indonesia (5) Iraq (1) Herzegovina (6) Nepal (13) Burundi (8) Georgia (5) Haiti (6) Philippines (6) Yemen (23) Chad (4) Kosovo (1) Timor-Leste (4) West Bank & Gaza (5) Comoros (2) Tajikistan (8) DRC (22) Cote d’Ivoire (2) Guinea (5) Guinea-Bissau (4) Liberia (2) Rwanda (2) Sao Tome and Principe (2) Sierra Leone (5) Sudan (1) Togo (1) Uganda (1) Zimbabwe (2) 28 68 20 7 15 29 DRC = Democratic Republic of Congo. Source: World Bank CDD Database. Annex 2. Tables Related to the Sample • 51 Table A2-3. Comparison of Countries and Projects Covered in CDD-FCS Review Studies World Strand Owen Barron Bank Wong Iraq GET This Countries (2003) (2006) (2011) (2012) (2012)* report paper Afghanistan: The P.E.A.C.E. Program (1997–2001) x Afghanistan: National Community Empowerment Program x Afghanistan: National Solidarity Program x II x x Angola: Social Action Fund (FAS) II III IV I–IV Bosnia & Herzegovina: Building Local Institutions and Social x Capital: The Tuzla Model (2002) Bosnia & Herzegovina: Community Development Project x Burundi Social Action Project I–II x Burundi Community Rehabilitation Project x CAR: Support to Vulnerable Groups Community x Development Project Colombia: Peace and Development Project x x Colombia: Magadalena Medio Regional Development Project I–II DRC: Emergency Social Action Fund x Ethiopia: Productive Safety Net III x Georgia: Self reliance for IDPs and host families in Samagrclo x Region Guatemala: Reconstruction and Local Development Project x Guinea-Bissau: Rural CDD Project x Haiti: Rural CDD project x India: Tamil Nadu Health x x Indonesia (Aceh): BRA x x x x Indonesia: KDP I–III x x III II x Indonesia: PNPM (Rural) x x Iraq: Consultative Services Delivery Program x x Ivory Coast: Post-Conflict Assistance Project (PAPC) x Kosovo: Social Development Fund I II II Kyrgz Republic: Village Investment Project II x Nepal: Poverty Alleviation Project x x Philippines: Promoting transition from conflict to peace and x development at community level in Mindanao (2001–2003) Philippines: ARMM x x x x x (continued) 52 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A2-3. Continued World Strand Owen Barron Bank Wong Iraq GET This Countries (2003) (2006) (2011) (2012) (2012)* report paper Philippines: KALAHI-CIDDS x x Philippines: Mindanao Trust Fund x Philippines: Reconstruction and Development Program x Rwanda: Community Reintegration and Development Project x Rwanda: Decentralization and Community development x x Project Scribia: Southern Serbia Municipal Improvement and x Recovery Program (2001) Sierra Leone: Decentralized Service Delivery x Sierra Leone: GoBiFo x Somalia: Health Service Recovery Project x Somalia: CDRD Project I–III x Sri Lanka: NELSIP x x Sri Lanka: Community Development/Livelihood II x Sudan: CDF Project x Tajikistan: Rural Infrastructure Rehab Project x Timor Leste: CEP I–III I x Timor Leste: Youth Development Thailand: Community Approaches in Conflict Situations x Northern Uganda: NUSAF x x Vietnam: Northern Mountains Poverty Reduction II x West Bank & Gaza: Integrated Community Development x ? Project West Bank & Gaza: Municipal Infrastructure Development x Project Yemen: Social Fund IV I–IV I–IV * Wong (2012) is a meta-analysis of impact evaluations of CDD projects in general, which included four CDD-FCS projects as well. Notes: GET = Global Expert Team. DRC = Democratic Republic of Congo. Annex 2. Tables Related to the Sample • 53 Table A2-4. Contextual Variation within Sample Country Project name Time Context Afghanistan NSP I–III 2004–2012 Weak project implementation capacity at the central and local levels. Continued violence and conflict in parts of the country, increasing in geographic scope and frequency during NSP implementation. Angola FAS II 2000 Protracted civil war for 3 decades, still ongoing; severe war damage. Some overall improvement in the security conditions in 2000. Almost no capacity at community or municipal levels. Bosnia & CDP 2001–2009 Post-war country after a devastating war. Ambitious recovery program with decentralization Herzegovina at the core. Destroyed infrastructure, but some capacity at the municipal level. Colombia MMRDP I–II 1998–2008 Conflict since 1960s increased in severity in the 1990s. Violence, insecurity, and high homicide rates limited opportunities. Rising unemployment affected youth significantly. Estimated 2 million IDPs due to ongoing violence. Central SVGCDP 2009– Has one of the lowest rankings of HDI largely due to its landlocked position, weak African institutions, a poor transportation system, an unskilled work force, and politico-military Republic crises. This situation has been compounded by the international financial crisis. Decades of conflict and poor governance have led to economic and social collapse. Cote d’Ivoire PCAP 2008– During conflict, economic and political power concentrated in central government. Need for poverty reduction, transition from war to peace, and decentralization. Limited local and central government capacity. Poor trust of communities towards government institutions. Guatemala RLDP 1998–2006 Chronic poverty, plus 36 years of civil war and high degree of income inequality. Needs decentralization and greater recognition of indigenous rights. Weak governmental capacity, particularly in reaching out to indigenous populations. Indonesia BRA-KDP 2005–2009 Oil-rich country with a history of violence and conflict rooted in regional inequality and ethnic differences. High levels of corruption and ineffective local government, weak judicial systems, and poor ministerial capacity. Since 1998, radical decentralization changed political, economic, and social contexts. More recently, end to a 30-year conflict between the separatist GAM movement and the government of Indonesia. Iraq CSDP 2010– Since 2003, the Kurdish region has enjoyed relatively good security (a “security permissive” operating environment) and better service delivery than the rest of Iraq, enabling development activity. Kosovo CDF Project 2000–2006 Ethnic conflict ended in June 1999. Country received generous donor support. Transitioning I–II to self-rule. Reconstruction efforts on track, although ethnic divides persist. Protection of disempowered minorities was key concern. Philippines ARMM Social 2003– In middle-income Philippines, this marginalized region (Mindanao) gained autonomy in 1989 Fund Project after peace agreement signed between the government and the separatist group MNLF. Despite the peace agreement, violence continued in some areas. A preliminary peace treaty was announced in Oct 2012 with the Moro rebel group, which would create an autonomous Muslim region in the five provinces by 2015. The regional government capacity was initially severely lacking, but this improved with time. (continued) 54 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A2-4. Continued Country Project name Time Context Rwanda CRDP 1998–2003 Poor governance at root of genocide. Strong government commitment to decentralization and public sector reform. Lack of institutional capacity throughout local government. Poor donor and NGO coordination added stress to weak institutions. Somalia CDRDP 2008–2010 Repeated cycles of violence. Has one of the lowest rankings HDI. Fragile transition with low capacity at central and local levels, threat of further violence, and corruption. Sri Lanka NELSIP 2010– War between the government and the Tamil rebels ended 2009 (northern province) and 2007 (eastern province). Weak local capacity and relatively better technical and administrative central capacity, but lacking legitimacy at the local level. Uganda NUSAF 2002–2010 Northern Uganda has highest incidence of poverty. Politically motivated conflict tied to Project livelihood insecurity and internal and cross-border conflicts. West Bank & VNDP 2008–2012 Newly established, but weak institutional capacity at the central level. Central municipality Gaza tensions. Lack of progress in the peace process resulted in renewed conflict. Stringent Israeli closure of WB&G reduced mobility and access to essential services. Substantial damaged infrastructure. Economic growth halted after intifada in 2000 and border closures. Political uncertainty and insecurity discourage growth and investments. High unemployment, poor infrastructure, and social services. Municipalities dominant in service delivery. Yemen SFD Phases 1997–present Long-standing political tensions and security concerns. Political crisis in 2011 made donors I–IV suspend their funding. Weak institutional capacity at the central and local levels. Notes: HDI = Human Development Index. GAM = Free Aceh Movement [Gerakan Aceh Merdeka]. MNLF = Moro National Liberation Front. Source: Various PADs. Annex-3. Select Summary Tables from the Stocktaking Study Table A3-1. Overall Objectives in Projects and their Continuations Improving, restoring, or constructing Enhancing service delivery Strengthening livelihoods and Building cohesion and local local government increasing incomes and preventing/ Countries and projects infrastructure and institutions of the poor mitigating conflict Afghanistan: Emergency National Solidarity 3 3 Project I Afghanistan: National Solidarity 3 3 Project II Afghanistan: National Solidarity 3 Project III Angola: Social Action Fund Project I 3 3 Angola Social Action Fund Project 3 (FAS II) Angola Social Section Fund (FAS III) 3 3 Angola IV: Local Development Project 3 3 3 Bosnia & Herzegovina: Community 3 Development Project Colombia: Magdalena Medio Regional 3 3 Development Project I Colombia: Magdalena Medio Regional 3 3 3 Development Project II CAR: Support to Vulnerable Groups Community 3 3 3 Development Project Cote d’Ivoire: Post-Conflict Assistance Project 3 3 3 Guatemala: Reconstruction and Local 3 3 Development Project Indonesia: Community-Based Reintegration In 3 3 Aceh (BRA-KDP) Iraq: Consultative Service Delivery Program I 3 Iraq: Consultative Service Delivery Program II 3 3 Iraq: Consultative Service Delivery Program III (continued) 55 56 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A3-1. Continued Improving, restoring, or constructing Enhancing service delivery Strengthening livelihoods and Building cohesion and local local government increasing incomes and preventing/ Countries and projects infrastructure and institutions of the poor mitigating conflict Kosovo: Community Development Fund Project 3 3 I Kosovo: Community Development Fund Project 3 3 II Philippines: ARMM Social Fund Project 3 3 3 3 Rwanda: Community Reintegration and 3 3 Development Project Somalia: Community-Driven Recovery and 3 3 3 Development I Somalia: Community-Driven Recovery and 3 3 Development II Somalia: Community-Driven Recovery and Development III Sri Lanka: North East Local Services 3 Improvement Project (NELSIP) Northern Uganda Social Action Fund Project I 3 3 Northern Uganda Social Action Fund Project II 3 3 West Bank & Gaza: Village and Neighborhood 3 Development Project Yemen: Social Fund for Development, Phase I 3 3 3 Yemen-Social fund for Development, Phase II 3 3 3 Yemen-Social fund for Development, Phase III 3 3 3 Yemen-Social fund for Development, Phase IV 3 3 Source: Various PADs. Annex 3. Select Summary Tables from the Stocktaking Study • 57 Table A3-2. Language for Project Development Objectives Objectives Language from project documents Strengthening local government and “…lay the foundations for strengthening of community-level governance” institutions (Afghanistan NSP I). “…focus on capacity building at the level of communities, partners, local government, and FAS” (Angola FAS II). “…rehabilitate social infrastructure and improve the capacity of local stakeholders to plan and manage local recovery in targeted areas” (Central African Republic CDP). “…build the capacity of local government to be able to take over the function and responsibility of helping communities help themselves (Somalia III). Improving, restoring, or constructing “…improve access of rural communities to social and productive infrastructure and services” service delivery and local (Afghanistan NSP I). infrastructure “…improvement of basic services and facilities through investments in non-revenue generating socially oriented projects and programs” (Bosnia & Herzegovina). Enhancing livelihoods and increasing “…improve conflict-affected communities’ and individuals’ opportunities for economic reintegration” incomes of the poor (Cote d’Ivoire Emergency Post Conflict). “…provide capacity building for women, youth, and other community groups for improving food security, employment opportunities, and household incomes (Philippines). Building cohesion and preventing or “…strengthen social cohesion and partnerships between and within communities in mitigating conflict the ARMM region” (Philippines ARRM). “…contribute to peace building and particularly to building the social contract between citizens and their government (CAR CDP). Source: Various PADs. 58 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A3-3. Specific Components or Program Activities to Address Social Cohesion Country and project Conflict mitigation/prevention strategy and tool Effect Afghanistan Democratically elected (through secret ballot), gender-balanced community Evaluation studies show success on NSP I–III development councils helped trans-fer authority away from traditional power many measures structures. Bosnia & Municipal participatory process for sub-project selection and management. This created trust, which facilitated Herzegovina the re-integration of returnees and created the conditions to keep them in place, rather than have them migrate to better conditions. Central African Each community development plan includes the design of a human security Republic element. This allows communities to monitor the human security status of their locality, as well as real and perceived threats to their personal safety, and the security of their homes and assets. This information is then transferred to the local authorities, along with the local development plans, for discussion on how security and development issues need to be managed so that these threats can be addressed and mitigated. Colombia II Pilot creation of humanitarian zones, which are commun-ities dedicated to Communities successfully took a building a peaceful society based on human rights and human dignity, and where stand and withstood threats from residents build alliances to create a secure and sustainable community that insurgents. rejects violence. In these communities, which were seriously affected by violence, the results demonstrate a new ability for inter-institutional collaboration. Cote d’Ivoire Restoration of social capital and social cohesion at the community level within a Peace committees were established spirit of stabilization and reconciliation. It includes: in post-conflict communities; radio programs re-knit the social fabric. 1) the systematic development and strengthening of community-based mechanisms to mediate and resolve conflicts; 2) a wide range of activities to promote solidarity and reconciliation, including public events and group activities (e.g., village theatre, traditional dancing, sport competitions, inter- communal meetings and radio programs); and 3) a special component addressing the needs for employment of those who perpetrated violence or are a threat: ex-combatants, youth, female soldiers. Indonesia-Aceh This project addresses the perceived and real perception of unjust treatment by Created more welfare. It did not the state. The project provides grants to communities with a high percentage increase acceptance of returning of conflict victims to support village level re-integration of returned groups, groups; in fact, it increased tensions including ex-combatants. Communities were to prioritize conflict victims. with ex-combatants who tried to bully their way into use of funds for conflict victims when their own funds were delayed. Overall, conflict victims received higher grant allocations than non-conflict victims. (continued) Annex 3. Select Summary Tables from the Stocktaking Study • 59 Table A3-3. Continued Country and project Conflict mitigation/prevention strategy and tool Effect Kosovo I–II Funds set aside for mixed and minority communities to create an effective CDF I demonstrated how a mechanism for ensuring equitable access to resources. Carried out special social/ community-based project can serve conflict analysis for such communities. as a catalyst for a process in which targeted commun-ities, especially Provided management support to provide housing for IDPs funded by other mixed and minority communities, donors. Specifically targeted inter-ethnic youth activities for support, based on gradually come together and special review of CDF’s capacity to strengthen inclusion of youth at mid-term. participate in the design, implemen- tation and monitoring of projects for the common good of the community. A full-time Serbian engineer hired as CDF staff member to facilitate activities in Serbian communities; hired Serbian sub-contractors. Local multi-ethnic NGO contracted to carry out programs in mixed and minority (including Serb) communities. Successfully carried out youth activities Philippines Grants component used to build social cohesion, with special attention to Peace watch report and monitoring vulnerable groups (women, IPs) and to youth prone to rebel recruitment. for early warning of violence/ Transparency, GRM, and LGU reforms to be “positive” re-enforcers of peace and conflict.25% of communities reported development. Parallel CIDA funded project for NGOs/CSOs on peace advocacy, increased perception of safety by reconciliation, and conflict management. Development of peace curriculum for 2010. schools. Provide housing for IDPs as part of sub-project menu. Rwanda Construction of houses for returnees or homeless families included in the open menu for grants. Uganda, NUSAF I Provide specific support for reconciliation in communities, training peace makers Successful reintegration of 458 (1,309), using 652 structured events (such as games, sports, drama), but also using displaced persons and ex-combatants. traditional institutions and authorities, and healing and cleansing rituals for Some 845 communities benefited communities and child soldiers. from reconcilia-tion efforts. Qualitative Impact Assessment: community reconciliation and conflict management support was credited with contributing to the resettlement of abductees and displaced persons, the resolution of latent conflicts among communities, and the capacity building of cultural institutions, NGOs, and CBOs. 60 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A3-4. Level of Decentralization Versus the Project Design Level of . CDD project design vs. decentralization status Country decentralization* Lagging Matching Leading Afghanistan Dysfunctional 3 Angola I–III Dysfunctional 3 Angola IV Dysfunctional 3 Bosnia & Herzegovina Incipient 3 CAR Dysfunctional 3 Cote d’Ivoire De-concentrated 3 Colombia Incipient 3 Guatemala Incipient 3 Indonesia Consolidating 3 Iraq Dysfunctional 3 Kosovo I Dysfunctional 3 Kosovo II Incipient 3 Philippines De-concentrated 3 Somalia I–II Dysfunctional 3 Somalia III Dysfunctional 3 Rwanda Incipient 3 Sri Lanka Consolidating Uganda Consolidating 3 WB&G Incipient 3 Yemen I–III Centralized government 3 Yemen IV Incipient 3 * Definitions of level of decentralization are drawn from Berthet et al. (2008). Source: PADs, ICRs, and secondary literature. Annex 3. Select Summary Tables from the Stocktaking Study • 61 Table A3-5. Size of Grant by Community Project name Country Size of grants to sub-projects Emergency National Solidarity Project (NSP) Afghanistan ≤ $60,000; average $26,000 Appraisal: $50,000/subproject; actual: $75,582/ Community Development Project (CDP) Bosnia & Herzegovina subproject Magdalena Medio Regional Development Project I Colombia Less than $50,000 * Productive sub-projects (component 3): Magdalena Medio Regional Development Project II Colombia ≤$100,000; social sub-projects (component 4): $5,000–$20,000 Support to Vulnerable Groups Community Development Project Communes up to $50,000; villages up to CAR (SVGCD) $12,500 (pooled up to $20,000) Post-Conflict Assistance Project (PCAP) Cote d’Ivoire Below $20,000 Reconstruction and Local Development Project (RALDP) Guatemala ≤ $150,000 Very small (mostly cash disbursements to Community-Based Reintegration in Aceh (BRA-KDP) Indonesia individuals) Iraq Consultative Service Delivery Program (CSDP) Iraq  No specifics On average, $71,780 for works projects; $7,584 Community Development Fund Project I (CDFP) Kosovo for social service projects Community Development Fund Project II (CDFP) Kosovo Average $112,599 No specifics, but allocation per barangay ARMM Social Fund Project Philippines (village or district) was calculated at PhP 2 million–PhP 2.5 million ($50,000–$60,000) Community Reintegration and Development Project (CRDP) Rwanda $215,000 divided among 3 types of sub-projects Community-Driven Recovery and Development (CDRD) Somalia Maximum $15,000 North East Local Services Improvement Project (NELSIP) Sri Lanka Not available ≤ $20,000, except when those sub-projects are approved by the LAs and endorsed by the NSC Northern Uganda Social Action Fund Project (NUSAF) Uganda at the recommendation of the NUMU; in that case ≤ $50,000 Village and Neighborhood Development Project (VNDP) West Bank & Gaza Not specified Social Fund for Development (SFD), Phases I–IV Yemen Not specified * Colombia II: The size of the productive sub-projects does not correspond to the size of the grant, as the grant was limited to technical assistance and the productive sub-projects obtained financing from the banking system. 62 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A3-6: National Implementing Institutions and Institutions with Final Authority Implementing institutions In-sourced Outsourced Project manage- Outsourced to ment team in Autonomous social foundation or . Institution with final authority ministry fund (public) int’l NGO Central or Prime Minister’s Office Cote d’Ivoire cross-cutting Ministry of Finance Bosnia & Herzegovina Ministry of Interior or Home Affairs Rwanda, Indonesia Ministry of Planning Colombia, Iraq In-sourced Ministry of Economic Development Sri Lanka Line ministry Ministry of Local Government West Bank & Gaza, Rwanda Ministry of Rural Reconstruction and Afghanistan Development Ministry of Social Affairs CAR International Somalia, Kosovo Outsourced organization Agency Independent agency Angola, Guatemala, Philippines, Uganda, Yemen Annex 3. Select Summary Tables from the Stocktaking Study • 63 Table A3-7. Locus of “validation” of community plans Community development committees or Communities Formal government Parallel project structure Local government Uganda, Indonesia Philippines (later) or inter-community forums Sub-national or Afghanistan Angola, Philippines (initially) regional National, project West Bank & Gaza, Rwanda, Bosnia & Herzegovina, implementation unit CAR, Cote d’Ivoire, Sri Lanka, Colombia, Guatemala, Iraq, (PIU), or agency Kosovo, Somalia, Yemen Table A3-8. Authorization Versus Implementation of Sub-projects Who implements? Community or Agency or project community-based Local government . implementation unit Who authorizes? organization (CBO) or line department (PIU) Outsourced Agency or PIU Cote d’Ivoire, Kosovo, Kosovo Guatemala (later), CAR (to NGOs) Somalia, Uganda, Angola, Yemen Yemen, West Bank & Gaza Local gov’t or line Bosnia & Herzegovina, Bosnia & Herzegovina, department Rwanda, Sri Lanka, Rwanda, Sri Lanka Uganda, Community or Afghanistan, Cote CAR (to NGOs) CBO d’Ivoire, Colombia, Guatemala (earlier), Indonesia, Philippines 64 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A3-9. Design Choices for Flow of Funds by Project S.No. Country Region Project Name 1 2 3 4 5 1 Angola AFR Social Action Fund (II and III) 3 2 Afghanistan SAR Afghanistan: Emergency National Solidary Project (I-IV) 3 3 Bosnia and Herzegovina ECA Community Development Project 3 4 Colombia LAC Magdalena Medio Regional Development Project (I and II) 3 5 Central African Republic AFR Support to Vulnerable Groups Community Development 3 Project 6 Cote d'Ivoire AFR Emergency Post-Conflict Assistance Project 3 7 Guatemala LAC Reconstruction and Local Development Project 3 3 8 Indonesia EAP Community Based Reintegration in Aceh (BRA-KDP) 3 9 Iraq MENA IQ-Iraq Consultative Service Delivery Program 10 Kosovo ECA Community Development Fund Project (I and II) 11 Philippines EAP ARMM Social Fund Project 3 3 12 Rwanda AFR Community Reintegration and Development Project 3 13 Somalia AFR Community Driven Recovery and Development 3 3 14 Sri Lanka SAR North East Local Services Improvement Project (NELSIP) 3 15 Uganda AFR Norther Uganda Social Action Fund Project 3 3 16 West Bank and Gaza MENA GZ-Village and Neighbourhood Development Project 3 17 Yemen, Republic of MENA RY-Social Fund for Development (I-IV) 3 Notes: 1. Flow of funds comes directly from a central account (held with a central-level government ministry (such as the Ministry of Finance), to communities, which bypasses all middle layers of government. 2. Flow of funds comes from a central account to communities, with the intermediation of provincial or local government. 3. Flow of funds comes from a central account to communities, with the intermediation of a non-governmental entity 4. Flow of funds comes from a central account to an intermediate layer of government (LGU control). 5. Flow of funds comes from a central account to an intermediate contracted entity (such as an NGO) or to Social Funds and Foundations 3 The primary approach adopted in the operation. Grey check means the secondary approach also used in the operation. Source: Various PADs, ICRs, operational manuals, and TTL interviews. Annex 3. Select Summary Tables from the Stocktaking Study • 65 Table A3-10. Choices in Menus and Public/Private Goods Offered in the CDD FCS Sample Support to vulnerable Productive groups and Social infra- Economic project, Productive other infra and Country structure* infrastructure** public project, private services Cohesion Afghanistan Human capital X Savings and credit Asset transfer development for women and for women and handicapped handicapped people people Angola X X X (in FAS I) Environmental protection, reforestation Bosnia & X X Services: Herzegovina maintenance, waste collection, community recreation, parks, lighting, day care CAR X X Human security plan for each community Colombia X X X Cote d’Ivoire X (to refurbish X X (for vulnerable Youth, ex- Peace programs, e.g., damaged infra- groups) combatants; plays, community radio; structure) employment ID cards creation Guatemala X X Cultural heritage Indonesia, BRA X X X X Communities Rehabilitation of chose private mosques goods Iraq X X Kosovo X X Vulnerable groups Inter-ethnic youth projects; public goods sub-projects requiring mixed ethnic commun- ities to work together; or >1 community to work together (continued) 66 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A3-10. Continued Support to vulnerable Productive groups and Social infra- Economic project, Productive other infra and Country structure* infrastructure** public project, private services Cohesion Philippines X X Train women, Indigenous indigenous ancestral domain people, and mapping; youth in food IDP housing processing rehabilitation Rwanda X X X X HIV/AIDS support; refugee housing Somalia X X Vulnerable people through service access grants Sri Lanka X X Uganda X X Youth, vulnerable Peace-building activities people West Bank & X X Gaza Yemen X X Micro credit Vulnerable people, cultural heritage * Social infrastructure and services: education, health facilities. ** Economic infrastructure and services: Power, roads, bridges, irrigation, markets. Annex 3. Select Summary Tables from the Stocktaking Study • 67 Table A3-11. Variations of the flow of funds to communities and local government  Funds channeled Funds passed on Funds only for Funds for Funds for Two windows to local govern- to communities projects going municipal plans municipal plans of funding: ment for their plans beyond one (from local (local govern- municipal plans and implementa- community government and ment versus and communities tion and/ or for community) commun-ity; each get direct urban seat they compete) funding Bosnia & 3 3 Herzegovina CAR 3 3 “first come, first served” Guatemala 3 Philippines 3 Rwanda 3 3 Sri Lanka 3 N. Uganda 3 Source: Various PADs, ICRs, operational manuals, and TTL interviews. Table A3-12. Community contribution requirements (cash or in-kind) Country Total contribution (%) Cote d’Ivoire 0 cash (in-kind; detail unclear) Sri Lanka 2% CAR 5%–10%, to be negotiated for each local development project (LDP) separately Guatemala 7% Uganda 5%–20% Afghanistan 10% minimum of the cost of infrastructure West Bank & Gaza 10% (excellent recording systems) Angola 10% or 15% (but in Angola II it was 8.6%, according to ICR) Sri Lanka 10%–20% from communities Kosovo 15%, from community, local government, or combination of both; 5% for poor communities Philippines 15% for local government Somalia 20% Source: Various PADs, ICRs, operational manuals, and TTL interviews Annex 4. Summary Tables and Figures to Understand and Inform Design Choices Table A4-1. Peace Building Versus State Building: How Designs May Change Design elements Peace-dividend focus State-building focus Project development Service delivery and infrastructure rehabili-tation will likely Governance, social cohesion, building capacity, and objective become the main focus. possibly livelihoods would become more important or primary. Participation Horizontal participation at community level is sufficient. Vertical participation between community and local government, and between other ministries becomes important. Scope and scale: All conflict-affected areas should ideally get covered to Could be phased with different stakeholders and cover macro-targeting maximize the peace dividend. only certain regions first. Targeting All communities should get covered so they “see visible Can use many different criteria for targeting as it will impact”; conflict damage can be used as criterion to help depend on the kind of state-building process to be hard-hit areas first. developed. Intra-community, IDPs, demobilized soldiers, and other conflict victims Entire conflict affected area population groups would population groups would be focus. be target. Menu Wide ranging closed menu, with infrastructure emphasis Open menu with peace building, cohesion, and likely to be best. capacity building as possible activities. Facilitator skills Horizontal participatory processes; bonding social capital Horizontal and vertical participatory processes; linking emphasize and conflict resolution. social capital. Grant cycle (Depending on overall budget) one grant cycle may be Multiple rounds become key here to learn process of default to be able to spread the dividend as widely as dealing with self-governance and relationship with possible. LGUs. National Capable, independent management agency may deliver Community and LGU capacity would be important. implementation the fastest results and parallel structure may be warranted. Also the national agency that can deliver this capacity arrangements building best should be selected. Fiduciary arrangements May favor independent agency over community Would favor community and LGU management to management. build long-term capacity. 68 Annex 4. Summary Tables and Figures to Understand and Inform Design Choices • 69 Table A4-2. Conceptualizing the Design Choice on Sub-national Institutional Arrangements Criteria/factors Preferred institutional set-up influencing Agency/ Community Local gov’t decision Type NGO model model model Comments Decentralization Advanced 3 Clear case for using local government model. policy framework Exceptions may be high degree of exclusion and targeting requirements, post-conflict/disaster context, and livelihoods delivery. Nascent 3 3 Community model can be justified as interim option with view to transitioning to local government model. Requires explicit capacity-building objective/ component. Local governance Strong 3 Local government model is most appropriate, with capacity possible exceptions for private goods financing and social exclusion. Weak 3 3 Agency and community models may be appropriate. As governance improves, seek transition to local government model. Choice between them will depend on other factors, primarily on whether deliver private goods and/or if exclusion (i.e., social protection) focused. Conflict/Disaster Yes 3 3 3 Community model is effective if goal is to build social context cohesion. If source of conflict is community/ethnicity- based, then agency model may be justified. Local government model would be medium-term strategy when state-building is explicit objective. No 3 3 Choice between community and local government models will depend on other factors, primarily the decentralization context. Type of good Public 3 3 In most contexts, local government model most financed appropriate. Community model may be justified in nascent decentralization and conflict/disaster contexts. Private 3 3 Agency model (partnership with NGOs, private organizations) preferred (e.g., as used in South Asia). Local government model would require partnering with non-state organizations (e.g., private sector). Targeting High 3 3 3 Agency model with specialized targeting technology requirements preferred. Community model with community identification of beneficiaries possible. Local government model okay if high capacity and no social exclusion risk. Low 3 3 Choice between community and local government models will depend on other factors, primarily the decentralization context. (continued) 70 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A4-2. Continued Criteria/factors Preferred institutional set-up influencing Agency/ Community Local gov’t decision Type NGO model model model Comments Low 3 3 Choice between community and local government models will depend on other factors, primarily the decentralization context. Vision of funding Long-term 3 Agency and community models should integrate local duration government involvement and seek to transition to local government model. Short-term 3 3 Short-term intervention may justify use of agency and community models. Choice will depend on other factors, mainly whether delivers private goods or exclusion (social protection) focused. Source: Adam (2009). Annex 4. Summary Tables and Figures to Understand and Inform Design Choices • 71 Table A4-3. Examples of Goods and Services Delivered in CDD Projects by Core Outcome Type of good or Increasing incomes Improving services Enhancing cohesion Strengthening governance service • Economic infra-structure • Small-scale social • Integrated community • Community spaces, local (access roads, bridges, infra-structure (primary spaces, sports facilities government buildings irrigation systems, schools, health posts, • Equipment and Goods markets) potable water supply) furnishings for local • Tool shops, ware-houses, • School furniture, text- governments and local storage facilities books, learning materials, bodies drug supplies • Vocational training • Training in health and • Information, Education • IEC campaign on local • Agricultural exten- sanitation practices and Communication (IEC) planning and budgeting Public sion services, animal • Capacity building campaign on social or • Training/facilitation on husbandry for teachers, nurses, inclusion issues participatory planning • Marketing and business extension workers • Social accountability • Training in commun- Services linkages support initiatives (e.g., social ity procurement and • Organizational audits) contracting development for • Group formation • Community involve- producer groups ment in school-based • Training in financial mgmt, management book keeping • Public advocacy • Social monitoring • Livestock • Provision of textbooks, • Cash or in kind transfers • ID cards • Household-level computers, health care to poor and excluded • Information and Goods productive assets supplies groups news materials for (generators, tube wells, • Seed capital communities equipment) • Food aid Private • Microfinance (sav-ings, • Conditional cash transfers • Targeted training, plus • Grievance redress credit, insur-ance and • Scholarships, meal and literacy to excluded or mechanisms payment services) drug vouchers vulnerable groups (ombudsman) Services • Land reform; • Legal services • Legal services • Training in produc-tion, • Therapy processing, marketing) • Work fare Source: Adam (2009). 72 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A4-4. Conceptualizing the Choice between Open and Limited Menus Criteria If… …Then Open Limited Coverage Multi-sectoral 3 Single sector 3 Hierarchy of objectives Empowerment 3 Sectoral priorities 3 Concerns regarding complexity High 3 Low 3 Emphasis on technical standards High 3 Low 3 Source: Adam (2009). Table A4-5. Conceptualizing the Design Choice on Number of Grant Cycles …Then Criteria If… One round of financing Multiple rounds of financing Technical capacity (of High 3 communities and/or local governments) Low 3 Relative emphasis on High 3 empowerment Low 3 Complexity of subproject design High 3 Low 3 Sequencing Desirable 3 Unnecessary 3 Desired coverage Wide 3 Targeted 3 Available resources Ample 3 Limited 3 Vision of funding mechanism Long-term 3 Short-term 3 Source: Adam (2009). Annex 4. Summary Tables and Figures to Understand and Inform Design Choices • 73 Table A4-6. Stages in the Community-Level Sub-project Cycle Stages Community level Necessary conditions 1 Ensure local buy-in from all stakeholders through meetings, IEC, stakeholder workshops. Select communities according Preparation and planning to targeting criteria. 2 Facilitate a participatory planning exercise (e.g., using tools from participatory rural appraisal Ensure that trained facili- toolkit) within communities. Ensure that the toolkits are widely available and that the tators are in place and simple choice of tools and the local implementation of the “how-to” is very context dependent. operational and procedural This choice requires skilled facilitators. manuals have been tested. 3 Community selects or elects a community development committee and various (sub) committees’ functions. 4 Participatory decision-making process ensures inclusiveness, leading to the prioritized and agreed-upon local development plan. Decision-making 5 Review of the proposed local development project (LDP) by another institutional Training of “review board” arrangement, “the validation committee.” Comments and adjustments recommended validating the LDP to community (e.g., in line with sectoral norms and guidelines, such as a school map or information that rebuilding a bridge is already in the budget and scheduled by the local government). 6 Community or local government adjusts its LDP, after which it is considered “approved” 7 In the case of community management, the community opens an account, which may Adjust fiduciary arrange- require that the community development committee or its finance committee is first ments as needed, so recognized as a legal entity, allowed to open accounts, and allowed to carry out financial communities can open an transactions. Co-financing in cash contribution is deposited in the community account. account. In the case of local government management, such accounting procedures are already in place, although some donors may require separate accounts. With agency management, this step will be unnecessary. 8 Facilitator and/or project management institution trains all the committees. Training materials prepared and tested for procurement, financial management, M&E, project management. 9a Community or local government contracts (or ensures contracting of) technical specialists Project management ensures who design, cost out, etc., the sub-project. At times, this is done with funds from the that there is a list of vetted Implementation overall project, prior to the submission of the LDP (e.g., for a technical assistance grant to contractors, both for the community or local government), to ensure that adequate costing, and analyses are technical assistance and for carried out and that the institution. To validate or approve the LDP, they will have the implementation minimum data requirements in front of them before they approve it (e.g., environ-mental analysis, economic and financial rate of return for an economic sub-project). 9b With design in hand (as well as relevant approvals from sectors and/or local government or other higher authority), community or local government calls for bids. Community procurement guidelines for CDD apply. 9c Community or local government evaluates bids and awards contract and signs contract. 10 Community or local government pays contractor to implement sub-project. Community provides “in-kind” co-financing in coordination with contractor (e.g., materials such as sand, labor). 11 Community or local government ensures that monitoring of contract is done by a committee, which is independent of the project management committee. 12 Community or local government receives sub-project formally and pays all obligations and contracts. (continued) 74 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Table A4-6. Continued Stages Community level Necessary conditions 13 Community or local government has structure or process in place to maintain and operate sub-project, such as a memorandum of understanding with local government or relevant Operations and ministry to place and pay for staff and materials. The community has established and maintenance trained an O&M committee (e.g., water user association with assistance of facilitator or sectoral ministry). 14 Community or local government evaluates both process and outcome, and continually monitors impact indicators it has set up (e.g., do more children finish primary school now that schools have all grade levels?). Figure A4-1. Steps in the Community Cycle Target communities Idenfiy, hire, and Governance Arrangements train facilitators O&M Arrangements 14. M&E 1. Buy in, IEC 13. O&M 2. Participatory 12. Receive and IEs Planning 3. Select 11. M&E Prep Committees O&M Plan Social 10. Implementation Accountability tools 4. Prioritized Imple- Decision Plan 9. Procurement mentation making Train 5. Validated Validation Board 8. Train Plan Committees 7. Open accounts 6. Adjustments, Final Plan, Training materials IEC ready and tested Fiduciary Arrangements Annex 4. Summary Tables and Figures to Understand and Inform Design Choices • 75 Figure A4-2. Options for Flow of Funds Central Agency Government 4 3 2 1 5 LGU NGO Community Notes: 1 = Flow of funds comes directly from a central account (held with a central-level government ministry (such as the Ministry of Finance), to communities, which bypasses all middle layers of government. 2 = Flow of funds comes from a central account to communities, with the intermediation of provincial or local government. 3 = Flow of funds comes from a central account to communities, with the intermediation of a non-governmental entity 4 = Flow of funds comes from a central account to an intermediate layer of government (LGU control). 5 = Flow of funds comes from a central account to an intermediate contracted entity in the community (such as an NGO). 76 • Designing Community-Driven Development Operations in Fragile and Conflict Situations Figure A4-3. Social Accountability Tools by Primary Objectives in CDD Projects Transparency Accountability Participation Information Dissemination Community/Participatory Participatory Decision Making and Demystification Monitoring • Participatory Planning • Information Campaigns • Community Scorecards, • Participatory Budgeting • Community wide meetings Community Report Cards Participatory Management to explain the project and • Community Social Audit • Community Management start the participatory Financial Management • Community Contracting planning process • Procurement Monitoring • User Management • Public displays of information • Participatory Physical Audit Committees on the subprojects being • Public Expenditure Tracking • Citizen/User Membership in carried out (e.g., painted on • Input Tracking of both Decision-Making Bodies the wall) community co-financing and • Public Reporting of expendi- Local Governments (LG) of inputs being provided by • The future CDD projects will tures by community and by others the agency providing outside focus on LG planning and • Timely and reliable informa- budget preparation based on (grant) funds (in community tion sharing when subprojects meetings and painted on participatory decision making. are procured centrally by • Community Management a wall) agency other than community • Community Radio and Community Contracting • The grant/implementation could also be implemented agency needs to sign a MOU with funds from LG. with the community members • MOUs between user for community to monitor the management committees progress and quality of work and LG or deconcentrated • Community Oversight ministry to provide required Grievance Redress Mechanisms budget and staffing. Source: Adapted from World Bank (2010), “How-to Notes: How, When and Why to Use Demand-Side Governance Approaches in Projects” (Washington, DC: World Bank). Annex 4. Summary Tables and Figures to Understand and Inform Design Choices • 77 Figure A4-4. Elements of a Good Monitoring and Evaluation System for a CDD Project Management Information Participatory or Community Grievance Redress System (MIS)— Monitoring— Mechanism (GRM) and including project, HR, financial, using different Beneficiary Feedback and capacity building data Social Accountability Tools Special Studies— Third Party Monitoring— Impact Evaluations— e.g., on governance, to triangulate information both qualitative and quality of infrastructure, from MIS and do quantitative and institutional maturity technical evaluations Source: Adapted from World Bank (2010), “How-to Notes: How, When and Why to Use Demand-Side Governance Approaches in Projects” (Washington, DC: World Bank). The World Bank The World Bank Group Social Development Department 1818 H Street, NW Washington, DC 20433 USA www.worldbank.org/cdd