62007 WORLD DEVELOPMENT REPORT 2011 BACKGROUND PAPER CDD IN POST-CONFLICT AND CONFLICT-AFFECTED AREAS: EXPERIENCES FROM EAST ASIA Patrick Barron* University of Oxford July 16, 2010 The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the World Development Report 2011 team, the World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. * Thanks to Sean Bradley, Gillie Brown, Sarah Cliffe, Scott Guggenheim, Bert Hofman, Sana Jaffrey, Mary Judd, Markus Kostner, Julien Labonne, Dave McRae, Adrian Morel, Andrew Parker, Ben Powis, Susan Wong, and Rob Wrobel for providing comments on drafts of this paper. Comments should be directed to Patrick.Barron@nuffield.ox.ac.uk I. INTRODUCTION Community-driven development (or CDD) projects are now a major component of World Bank assistance to many developing countries. While varying greatly in size and form, such projects aim to ensure that communities have substantive control in deciding how project funds should be used.1 Giving beneficiaries the power to manage project resources is believed by its proponents to lead to more efficient and effective fund use. It is also claimed that project-initiated participatory processes can have wider ‘spillover’ impacts, building local institutions and leadership, enhancing civic capacity, improving social relations and boosting state legitimacy. Given these claims, it is unsurprising that CDD projects are increasingly utilized in areas affected by localized or escalated violent conflict and in post-conflict areas. Such places face multifold challenges. In areas with pervasive localized violence, there are risks that conflict may escalate. In higher conflict and post-conflict areas, poverty levels are usually high and welfare outcomes low. The stability and social cohesion necessary for development is frequently lacking. And often there are not strong and legitimate institutions to address poverty and manage conflict. Violent conflict is more likely to (re)emerge in such areas, leading to further impoverishment, undercutting social cohesion, and eroding institutions. The result can be a vicious cycle of deprivation and insecurity from which it is difficult to emerge. Doing development in such contexts is extremely hard. Even where violence has not escalated, localized conflict may provide the sparks for future fires, and development programs can become implicated in cycles of violence. CDD is viewed by its proponents as an appropriate vehicle for alleviating poverty and enhancing security in such places. Effective CDD projects can distribute resources quickly and to remote, rural areas. In devolving decision-making they can help ensure resource distribution is fair and popularly accepted and can operate in areas with security risks. Programs may also have peacebuilding impacts: by incentivizing forms of collective action that can work across conflict divides; by contributing to local institution building; and by strengthening vertical society-state linkages. CDD is increasingly viewed as a useful mechanism for preventing the emergence of violence and for continuing to channel development aid in its wake (World Bank 2006). The World Bank has developed an approach for conflict-affected and post-conflict areas that utilizes CDD along with support to supply-side reforms and outcomes that cannot be achieved at a community level.2 1 Projects usually involve the provision of block grants to localities; communities, sometimes competing with each other, propose any project within a more-or-less open menu and then have decision-making power over which projects should receive support. Local choice usually extends to procuring goods and project mechanisms aim to ensure that a wide section of the population (including marginalized groups) participates. 2 See Cliffe, Guggenheim, and Kostner (2003) for a summary. Ghani and Lockhart (2008) discuss the model with particular reference to Afghanistan. That country’s National Solidarity Programme (or NSP) is probably the largest CDD program working in areas of severe conflict. 1 This paper briefly reviews the World Bank’s experience of using CDD in conflict-affected and post-conflict areas of the East Asia and Pacific region.3 The region has been at the forefront of developing large-scale CDD programming including high profile ‘flagships’ such as the Kecamatan Development Program (KDP) in Indonesia and the Kapitbisig Laban Sa Kahirapan-Comprehensive and Integrated Delivery of Social Services (KALAHI-CIDSS) project in the Philippines. As of the end of 2007, CDD constituted 15 percent of the lending portfolio in East Asia compared with 10 percent globally (World Bank 2007a: 2).4 Many of East Asia’s CDD projects have operated—consciously or not— in areas affected by protracted violent conflict. CDD has also been used as an explicit mechanism for post-conflict recovery in Mindanao in the Philippines and in Timor- Leste, and for conflict victim reintegration in Aceh, Indonesia. How successful have such efforts been? Through what mechanisms have projects had impacts (or not)? And what factors—related to project design or to the context in which programs are operating—have affected performance? This paper provides a framework for assessing the impacts of CDD projects in post-conflict and conflict-affected areas.5 It tries to unpack the potential causal channels through which projects may have their desired, or other, impacts. It then looks at the evidence on whether and how projects have achieved these outcomes, focusing on a range of recent and current projects in Indonesia, the Philippines, Thailand, and Timor-Leste. The analysis summarizes results, draws on comparative evidence from other projects in the region and elsewhere, and seeks to identify factors that explain variation in outcomes and project performance. The paper concludes with a short summary of what we know, what we don’t, and potential future directions for research and programming. II. CONFLICT & CDD IN EAST ASIA 2.1 A Picture of Violent Conflict in East Asia and the Pacific Intra-state violent conflict of varying forms and levels impacts most countries in the region.6 Conflict-related fatalities are slight compared to many other regions in the world, in part because violent conflicts have often been localized.7 Over recent years, violence levels have also de-escalated in a number of the major conflicts in the region. 3 The paper refers to the countries in the EAP region as East Asia. Conflict-affected areas include both areas with pervasive localized violence and with larger-scale violent conflict. 4 Mansuri and Rao (2004: 2) argue that World Bank lending for CDD has been rising and estimate that globally by 2003, US$ 7 billion had been lent for such projects. 5 The paper builds upon previously published reviews and studies, including papers on CDD in East Asia (World Bank 2007a) and in Conflict-Affected Countries (World Bank 2006), project reports and documentation, evaluations, and the personal experiences of some of those who have worked on CDD programming in the region. It should be noted upfront that the quality of evidence on different projects varies massively. For some projects, notably KDP/PNPM in Indonesia, a large amount of research has been conducted. For others, findings cited in this paper are more anecdotal. 6 There have been no inter-state wars in the region in the past decade. Tensions exist between North and South Korea, and small disputes have taken place between a number of countries (e.g. between Thailand and Cambodia and between Malaysia and Indonesia over disputed territory). Yet, by and large, these have not escalated into larger-scale violence. 7 According to the Geneva Secretariat for Armed Violence (2008: 16), there were 5,410 direct conflict deaths in East and Southeast Asia between 2004 and 2007. This figure, which includes only reported deaths from conflicts that have killed at least 100 people, is 3 percent of the global total. 2 Nevertheless, a number of states are struggling with large-scale armed violence and/or to emerge from violent conflict, and localized forms of violence—destructive, too— affect large swathes of the region. Large-scale armed violent conflicts Current large-scale armed conflicts in the region are sub-national insurgencies. Areas currently experiencing large-scale organized violent conflict include Mindanao in the Philippines, Myanmar’s border areas, and southern Thailand. Most deadly is Mindanao where over 120,000 have been killed in fighting since the 1970s with two separatist groups (the MNLF and MILF) seeking independence (Schiavo-Campo and Judd 2005). While a peace deal with the MNLF, which resulted in the Autonomous Region in Muslim Mindanao (ARMM), has generally held, an agreement between the national government and the MILF has not been reached and fighting continues in parts of the island, including with a radical separatist Islamic group, Abu Sayyaf (ICG 2008). In Mindanao and elsewhere in the country, there have also been battles between the military and the armed wing of the Communist party, the New People’s Army (Ploughshares 2009). Myanmar’s post-independence history has been marked by numerous violent conflicts between different groups and the state, with every significant ethnic group in revolt at some point or other (ICG 2004: 12). Some of these groups have signed peace agreements, but violent conflict continues in many border areas. Violent conflict in southern Thailand is prevalent in the historically contested Pattani, Yala, Narathiwat and Songkala provinces. Around 4,000 lives have been lost since 2004 with violence intensifying in 2009 (ICG 2009a; Pathan 2009). Much of the violence is localized—local-level grievances meld with the broader narrative of ethno-national separatist contention, leading to varying patterns of unrest. ‘Post-conflict’ areas Many other areas in the region have experienced recent large-scale violent conflict. While such places are currently ‘post-conflict’, many are experiencing challenges in maintaining peace.8 Worldwide, risks of violent conflict resumption in such areas remain high.9 Even where large-scale violent conflict does not re-occur, new forms of violent conflict and crime often emerge (Geneva Declaration Secretariat 2008). In Timor-Leste, for example, widespread violence re-emerged in May 2006 leading to large-scale displacement and the return of international security personnel from Australia and New Zealand. Localized gang violence involving youth is also prevalent. In the words of one assessment, “Timor-Leste currently offers an enabling environment for violence and crime” (World Bank 2007b: 3). A number of regions of Indonesia are emerging from protracted violence. The fall of Suharto’s New Order government in 1998 was accompanied by inter-communal 8 I use the term ‘post-conflict’ to signify areas where large-scale violent unrest has reduced signficiantly. Of course, conflict, and indeed violence, is often still present in many of these areas. 9 Collier, Hoeffler and Soderbom (2008) find that countries emerging from civil war face a 40 percent chance of conflict resumption within ten years. 3 violence in the Malukus, Central Sulawesi and parts of Indonesian Borneo killing around 8,000 people (Varshney, Tadjoeddin and Panggabean 2008); these violent conflicts had largely de-escalated by 2002 through a combination of policing and military operations and informal domestic peace settlements. In Aceh, where a three-decade separatist conflict resulted in almost 30,000 deaths, a 2005 peace agreement has held (MSR 2010). Other ‘post-conflict’ areas in the region include Bougainville in Papua New Guinea, where a separatist uprising in the late 1980s and 1990s led to 10,000-15,000 deaths. In the Solomon Islands, ethnic violence broke out in the late 1990s displacing tens of thousands. In 2003 an Australian-led Regional Assistance Mission to Solomon Islands (RAMSI) arrived to try to help consolidate peace. If one goes further back, large-scale violence in Cambodia, which subsided somewhat in the early 1990s with the arrival of a UN mission, has had impacts that shape society and politics to the present day. Localized violence Across East Asia, persistent localized violence is common. Conflict is an inevitable outcome of processes of social, political and economic change (Polanyi 1944, Bates 2000); where effective institutions are not in place to manage these pressures, conflict can become violent. Such violent incidents tend to be much smaller in scale—involving fewer actors and with smaller impacts per incident. Yet, collectively, localized violence can have significant human security impacts and can causes challenges for the operation of CDD and other development projects. The spread and scope of such violence is much less documented than for the larger- scale conflicts discussed above. In the southern Philippines, rido, or family feuds, sometimes escalate into inter-community violence. A total of 1,266 rido cases occurring between the 1930s and 2005 have been documented, with around 5,000 killed and thousands displaced (Torres 2007). Many areas have a problem with localized political violence; in late November 2009, 57 men and women, including supporters of a rival to the incumbent mayor of Maguindanao province and journalists, were killed in a brutal massacre as they were on the way to register his candidacy for an upcoming election (ICG 2009b). In Indonesia, ‘routine’ violence also appears to be rife (Mansoob Murshed, Tadjoeddin and Chowdhury 2009; Barron and Sharpe 2008). Welsh (2008: 481), for example, records over 5,500 victims from vigilante lynchings in just four provinces between 1995 and 2004. In some areas, such as Papua, routine violence mixes with and accentuates other forms of contention, including separatism. Other countries, such as Papua New Guinea, experience high levels of localized tribal conflict. Xinjiang and Tibet in China has experienced long-running sporadic separatist uprisings (Dwyer 2005). Varying contexts There is thus immense diversity in the forms violence has taken in the region. In some areas, large-scale violence has focused on challenging the sovereignty of the central state. In others, violence has been inter-communal, although often driven by underlying political contestation. In some places, security responses or peace settlements have formally ended violence. In others, unrest continues. In some ‘post-conflict’ areas, forms 4 of localized violence have emerged, with the potential for large-scale violence to reoccur, while others are (relatively) peaceful. In many areas that have not been affected by large-scale armed violence, local violent conflicts are common and may provide the seeds for future larger-scale unrest. Variation not only exists between countries. Within states, some areas are affected by violence while others remain relatively peaceful. Significant violence, for example, is taking place in southern Thailand and political competition has turned violent in Bangkok and other urban areas. Yet other parts of the country remain largely unaffected. Even locally, conflict conditions and impacts can vary massively. And conflict conditions vary over time; ebbing and flowing violence patterns makes it hard to characterize the conflict environment in a particular locality. Table 1 provides an overview of conflict conditions in the four countries considered in this paper.10 Table 1. Examples of Current Conflict Contexts in East Asia Country Large-scale violence Large-scale violence Localized violence (ongoing) (ended) Indonesia - Aceh Common across most – separatist violence ended provinces. Forms by peace settlement, 2005 include land conflict and vigilante violence. Malukus, C Sulawesi, W and In some places (such as C Kalimantan Papua) this is linked to - inter-communal violence broader unrest greatly reduced by peace agreements and/or security responses, 2001-2002 Philippines Mindanao: Mindanao Politically-related - Government vs. MILF - Government vs. MNLF violence, and rido feuds Government vs .Abu particularly prominent Sayyaf Group and deadly National: - Government vs. NPA Thailand South - Bangkok/urban areas11 - Government vs. - Political unrest separatist group Timor - National - Isolated gang violence Leste - Post-referendum violence in parts of Dili and some (1999) districts - Isolated incidents Dili relating to development - Widespread riots and projects street violence (2006); - Domestic violence regular youth gang and thought to be 10 The lack of information on localized violence in most countries should be noted. By and large conflict studies, by scholars and practitioners alike, have tended to focus on large-scale escalated violence at the expense of examining local violence. Indonesia is the exception where a comprehensive dataset is currently being built recording incidents of both large-scale and smaller-scale violent incidents. 11 While the location of such incidents have been largely in urban areas, they reflect larger rural issues. 5 election-related violence widespread (2007), often connected to land disputes 2.2 CDD projects in violence-affected areas Most countries in the region have, or have had, World Bank-supported CDD programs. As of late 2006, there were 17 active CDD programs in East Asia.12 The current and pipeline portfolio includes CDD projects in Cambodia, China, Indonesia, Laos, Mongolia, the Philippines, PNG, the Solomon Islands, Thailand, Timor-Leste, and Vietnam. Many of these projects are operating in areas that are affected by violent conflict. This paper focuses on a selection of current and recent CDD operations in four countries— Indonesia, the Philippines, Thailand, and Timor-Leste (Table 2).13 12 Adapted from the project list in World Bank (2007: 2-3). That document lists 42 projects. However, in many of these CDD is just a small element of expenditures. Many of the projects listed are also different phases of the same project. The seventeen projects are distinct and at least 75 percent of expenditures are for CDD. 13 Note that all CDD operations in each of these countries are not reviewed. The paper picks a selection of operations that operate in different types of conflict contexts. 6 Table 2. Some CDD projects in violent conflict-affected East Asia Project Timeline Cost Conflict Context Objectives (US$ million) Indonesia KDP/PNPM- 1998 - 3,000.0 Varying conflict dynamics in different Villagers to benefit from improved socio-economic and local rural (by areas, and at different times. Has governance conditions through the provision of investment 2011) operated during periods of large-scale resources to support productive proposals developed by violence, in post-conflict areas, and in communities, using a participatory planning process. places where local violence is rife. BRA-KDP mid 2006 – 23.0 Implemented in post-conflict Aceh To assist conflict-affected communities to improve their living mid 2007 soon after signing of peace agreement. conditions through the provision of small projects that accord with Provided assistance to conflict victims their needs. Secondary: (a) to overcome mistrust in government that (but not ex-combatants) as part of the is a result of the conflict; (b) to improve relations between different Government’s reintegration program. conflict-affected groups. Philippines KALAHI-CIDDS February 182.4 Varies by area. Some areas (e.g. To support the GoP in strengthening local communities’ participation 2002 – Mindanao and NPA strongholds) in barangay governance, and developing their capacity to design, May 2010 experience large-scale violence. Other implement and manage development activities that reduce poverty. areas have localized violence. To avoid overlap with the ARMM Social Fund, the KALAHI does not operate in the ARMM. ARMM Social May 2003 40. 6 (21 In theory, a peace agreement ended To foster sustainable development in the ARMM through reducing Fund – May for CDD) violence in the ARMM area. In reality, poverty and supporting mechanisms for the promotion of a peaceful 2010 (with + further continuing fighting in some areas. and safe environment in the conflict-affected areas therein. plans for 3 30.0 year expected extension) Mindanao April 2006 50.014 Sporadic but deadly violent conflict Communities of violence-affected areas, IDPs and rebel returnees Trust Fund – benefit from the visible restoration of some basic services of their 14 The total value of the Mindanao MDTF is US$ 50 million, of which an estimated US$ 12 million had been committed as of May 2010. 7 Project Timeline Cost Conflict Context Objectives (US$ million) Reconstruction – Dec 2015 across many project areas choice and sustained by Local Government Units (LGUs). and Intermediate objectives: (a) Above groups have satisfactorily decided Development on, planned and implemented sub-projects of their choice; (b) LGU Program (MTF capacity for project planning, resolving conflicts, project –RDP) implementation and seeking resources for projects improved; (c) Local implementing partners, including the BDA, have achieved sufficient implementing capacity to work with LGUs and local groups. Timor-Leste Community March 18.55 Implemented as part of recovery To strengthen local-level social capital to build institutions that Empowerment 2000 – response after massive post- reduce poverty and support inclusive patterns of growth. Program (CEP) March referendum violence. 2004 Youth July 2008 – 2.1 Designed in the aftermath of the 2006 To promote youth empowerment and inclusion in development by Development December crisis. expanding the capacities of and opportunities for youth groups to Project 2011 initiate and participate in community and local development initiatives. Thailand Community October 2.6 Working in areas affected by on-going To develop alternative and effective community approaches to local Approaches in 2007 – insurgency in the south. development which create “space” an opportunity for increased Conflict June 2012 interaction within and among communities and between Situation communities and the state apparatus in the effort to promote trust (CACS) building in the conflict-affected areas of Yale, Narathiwas, and Pattani. 8 Indonesia The Kecamatan Development Program (KDP), recently re-titled the National Community Development Program (PNPM-Mandiri to use its Indonesian acronym), is the largest CDD project in the world. Starting in the immediate aftermath of the 1998 Asian Financial Crisis, the program has scaled-up to cover every rural village in Indonesia. By 2011, the cumulative budget for the program will be over US$ 3 billion. KDP/PNPM was not initially intended to be a ‘conflict project’. Yet its roll-out coincided with a wave of large-scale violence that followed the fall of the New Order government and KDP found itself utilized as a mechanism for reconstruction, development and sometimes peacebuilding in high conflict areas. Most project funds are used by communities for local infrastructure—roads, bridges, irrigation channels, etc. KDP ran in Aceh from 1998 during a period of intense separatist insurgency and counter-insurgency. After a 2005 peace agreement, a new local government agency (the BRA) was established to support the reintegration of former rebel combatants and conflict-affected civilians. One resulting program was BRA-KDP which adapted the KDP model to provide US$ 22.7 million of government money to conflict victims in one round of grants. With separate programs to support them, ex-combatants were excluded from benefitting. Despite having the same open menu as regular KDP, communities chose to use most funds for private goods. This was a result of communities viewing BRA-KDP as providing compensation for conflict hardships, of the widespread need for capital following the conflict’s end, and of messages from BRA’s leadership that individual assistance should be prioritized. An explicit aim of the project is to support social cohesion at the local level. A planned second round, covering the remaining villages in Aceh, did not happen after a change of leadership at BRA. Philippines A range of CDD projects operate concurrently in the Philippines. The largest is the US$ 182.4 million KALAHI-CIDDS. Launched in 2002, the KALAHI is active in the 42 poorest provinces (out of 80) in the country. Within each province, the poorest one-fourth of municipalities participate. Though it was not specifically designed for conflict-affected areas, given the widespread nature of violence in the Philippines the KALAHI operates in many areas affected by conflict in Mindanao and elsewhere. The KALAHI follows similar procedures to KDP/PNPM, with communities free to choose how to spend funds, and most resources are used for small-scale public goods. Two other CDD projects operate in Mindanao and have a more explicit peacebuilding focus. The ARMM Social Fund for Peace and Development (ASFP) works in the Autonomous Region in Muslim Mindanao, which was created in 1989 after the signing of a peace agreement with the MNLF. The project began in mid-2003 and around two- thirds of the project’s US$ 40.6 million is used for community development assistance.15 Additional financing of US$ 30 million for another three years is planned. The project is implemented by the ARMM government and a major focus of the project is on building that body’s capacity and legitimacy. 15 Other funds are used for strategic regional infrastructure and institutional strengthening. 9 The Mindanao Trust Fund Reconstruction and Development Program (MTF-RDP) provides reconstruction and rehabilitation assistance to conflict-affected areas in Mindanao. The program is starting with CDD support. In the absence of a peace agreement between the MILF and the Government, violent conflict is frequent in many project areas. The first phase of the project has focused primarily on building the local institutional capacity of the Bangsamoro Development Authority (BDA), the development arm of the MILF, and other local implementing partners. As of December 2009, the MTF had funded projects in 43 conflict-affected municipalities.16 Thailand World Bank-supported CDD programming has been more limited in Thailand. The Social Investment Fund ran from 1998 to 2001 but did not focus on conflict issues. In April 2009, the second phase of a State and Peacebuilding Fund (SPF) pilot project began, financing CDD activities in nine villages in the conflict-affected south, which will eventually increase to 36 villages in three provinces.17 Block grants to communities are accompanied by a ‘peacebuilding partnership fund’ that provides small grants to civil society organizations working on peace issues. Timor-Leste The Community Empowerment Program (CEP) was initiated in late 1999 in the immediate aftermath of the massive destruction that followed the referendum on independence. Building on the KDP model, which had operated before independence, CEP was a major component of the international post-conflict recovery response. Over three phases, US$ 18.55 million was provided, a large sum for a country of less than one million people. CEP-3 finished in early 2004. The Youth Development Project began in mid-2008 and is expected to run until the end of 2011. Its genesis and design stem from an assessment of the challenges facing Timorese youth in the aftermath of the 2006 crisis (World Bank 2007b). That report identified youth violence as a (potential) major barrier to development. The resulting project combines support to the Government’s national youth strategy with CDD block grants targeted at young people. In the first round, the most popular use of grants was for sporting facilities. 16 The second phase of the program, which focuses on providing assistance to communities, was meant to begin after the signing of a MILF-GoP peace agreement. In the absence of this, there have been discussions among participating donors to remove the phasing from the MTF (Social Impact Inc. 2009). 17 The first phase involved a conflict study to aid design of the operation. This initial phase was funded by the Post-Conflict Fund, which was the precursor of the SPF. 10 III. CHALLENGES, AIMS AND CAUSAL MECHANISMS The eight projects have a range of objectives, are of varying sizes, and are being implemented in very different contexts. The goals of CDD projects, the challenges they face, and the mechanisms through which they have impacts will differ depending on the context in which they are working. Yet underlying each project is a common logic of how aid can contribute to development and peacebuilding in the face of violence and contestation. This section seeks to flesh out the assumptions that underlie the design of such projects as a means of providing a framework to help understand the impacts they have. Development is more difficult in areas affected by, or emerging from, violent conflict.18 Violent conflict impacts on economic and social life and the institutions that govern it. This creates a complex set of challenges. Economic deprivation is likely to be higher in areas experiencing larger-scale violence.19 Yet channeling resources into areas with serious divisions can trigger new conflicts, as groups compete and incumbent elites resist challenges to their power, which may turn violent.20 And the social and institutional bases necessary for the management of development in non-violent ways are often eroded by violence. This creates particular development challenges that CDD projects aim to address. There are a number of mechanisms through which projects may have impacts. This section provides an outline of (post)conflict challenges, the aims of CDD projects, and potential causal mechanisms. The level of challenge will likely vary by the degree of violence. These are summarized in Figure 1. 18 There is a vast literature on development challenges in conflict-affected and post-conflict areas. Useful starting points include Collier (2009) and Call and Wyeth (2008). 19 One of the few econometric findings on the ‘causes’ of civil war that is robust across different studies is that poor countries are more at risk. See, Kalyvas (2008). 20 A long-standing strand of social theory has posited that conflict, sometimes violent, is likely during times of rising incomes (e.g. Polanyi 1944, Moore 1966, Skopcol 1979; Cramer 2006). Historically, as Bates (2000) has observed, while poverty is associated with violence, violence and prosperity also go hand in hand. 11 Figure 1. The theory of CDD in conflict-affected and post-conflict areas 3.1 (Post) Conflict Challenges and CDD Aims Economic deprivation Economic development needs are typically greater in areas that have been affected by large-scale violence. Civil wars, for example, on average lead to annual GDP reductions of 2-2.2 percent;21 incomes are reduced by 15 percent and the proportion of people living in poverty increases by one-third (Moser 2006). More localized violence can also have significant economic impacts. The cost of lost productivity from criminal violence outside of warzones, for example, is estimated at US$ 95-163 billion (Geneva Declaration Secretariat 2008). There are multiple sources of economic impact including reductions in human capital (through impacts on the education system, forced migration and a brain drain, etc), weak government institutions, and distortions in the tax system, all of which impact on private sector investment and growth (World Bank 2009). CDD projects cannot impact these macro-level determinants of growth. Rather, economic gains are likely to be through the provision of public infrastructure, the replacement of basic assets, and the capital necessary for private asset replacement and accumulation. Infrastructure investments are often needed to replace assets destroyed 21 Estimates differ in scale. See: Collier (1999); Hoeffler and Reynal-Querol (2003); Restrepo et al. (2008). 12 in conflicts. In Timor, for example, violence around the 1999 referendum destroyed 70 percent of public buildings (Rohland and Cliffe 2002: 1). The war in Aceh damaged over 4,000 schools, two-thirds of those in the province, 7,700 km of road and almost 2,200 bridges (MSR 2010). Infrastructure that is damaged or deteriorating due to lack of maintenance is less likely to be repaired in conflict-affected areas, in part because of security risks, in part because conflict often erodes the capacity of local government institutions. Violent conflict can also lead to a reduction in the private assets of ordinary civilians. Most directly, private homes are destroyed, livestock and crops die, people leave possessions behind when they are displaced, and economic retraction limits job opportunities and savings. People may also sell productive assets to cover short-term needs. Accessing credit also often becomes more difficult as banks and other credit providers become less likely to lend or only lend at higher rates (Nagarajan and McNulty 2004). In many CDD projects, funds are used for saving and loans schemes or one-off cash transfers to households are provided. This may help individuals’ and families’ recovery. Instability Lack of security is a barrier to development, both directly—through limiting basic freedoms—and through its instrumental impacts on welfare. Development projects in areas affected by violent conflict typically aim not only to do development despite the conflict but to affect the social and economic bases that make an area conducive to violence. One way to do this is by attempting to reconfigure social and society-state relations. Violent conflict has social impacts.22 Where different groups previously co-existed side by side, war tends to accentuate markers of social difference such as ethnicity, religion, class or political affiliation.23 Large-scale displacement, all too often a result of conflict, can change the composition of villages, leading to new tensions and negatively impacting on local problem solving capacity.24 Trust in others is often diminished. This can result in barriers to collective action with people unwilling to participate for fear that others will shirk. While these processes are likely to be most marked for large-scale violence, and particularly where conflicting parties live side-by-side, localized violent conflict can lead to serious social divisions at the local level. During and after violence, particular groups may also be disruptive of social life. Programs to ‘reintegrate’ combatants are based on the premise that long-term social cohesion requires a particular focus on those who participated in war.25 22 Colletta and Cullen (2000) map the ways in which violent conflict transforms social capital and how this affects prospects for peacebuilding and development. 23 See, for example, Kalyvas (2006) who argues that group affiliation and differences are in large part a product of violent conflict rather than a cause of it. 24 Cernea and Guggenheim (1993) highlight how large-scale involuntary resettlement as a result of dam construction can create new needs and break down local institutional structures that were used to manage problems. 25 On difficulties ‘reintegrating’ ex-combatants, see Muggah (2009) and Colletta, Kostner, and Wiederhofer (2003). 13 Many CDD projects thus have a specific goal of enhancing the relations between different groups. In areas affected by conflict, it is often claimed that CDD holds potential to encourage new forms of collaboration across conflict divides, which can improve trust and make communities less prone to fresh violence.26 In post-conflict Aceh and Rwanda, CDD has also been used to target particular conflict-affected groups. However, CDD project may also inadvertently accentuate violence, in particular in areas marked by horizontal group divisions, if competition over project resources cannot be contained. Institutional weaknesses A third set of challenges relates to the impact of violent conflict on the capacity and legitimacy of institutions at multiple levels. Violent conflict tends to occur where institutions are weak; it can also further impact on institutional functioning leading to a vicious cycle of economic retrenchment and new violence (Bates 2008). Indigenous civil society institutions, which may substitute for state functions, can also be negatively affected by violence as local informal leaders become co-opted or their legitimacy is eroded. Rebuilding institutions in violence-affected or post-conflict areas can be difficult. Institutions may not be deemed legitimate; communities may have lost confidence in the ability and willingness of leaders and structures to act in their collective interests. Institutional capacity can also be eroded with population flight, the war economy distorting skills accumulation, and reduced fiscal resources impacting on functioning. If institution building is particularly important in such contexts, the aid effort can undermine it. Where aid agencies deliver services that were previously provided by the state, they can undercut state legitimacy or cause the state to move its resources and energies to fighting wars (Pendergast 1996). Skilled locals may leave their jobs in the civil service or universities to work at much higher wages for aid agencies (Ghani and Lockhart 2008). CDD projects typically aim to improve local institutional performance by creating a common platform for coordination and planning for development investments at the local level and by increasing demand for ‘good governance’, by improving citizen-state relations. At the same time, however, in establishing or supporting service delivery systems that run in parallel to government line ministries, they have the potential to undermine the legitimacy of the latter. Aid approaches and CDD goals In post-conflict and violence-affected areas, donors have traditionally developed separate projects for dealing with security and development needs: capacity building programs to improve institutions; peacebuilding programs to improve social cohesion; and economic recovery programs to repair conflict-affected infrastructure. CDD approaches differ in that they seek to use one intervention to address all three at the local level. Successfully addressing one goal is seen as indirectly contributing to the others: potential economic gains from participation in projects, for example, incentivize 26This claim is reflected in the project appraisal documents for most of the projects examined in this paper. For more general claims, see World Bank (2006). 14 collaborative decision-making that may lead to improvements in inter-group relations; more participation in the local public sphere may increase the demand for responsive institutions; and improved institutions may lead to quicker and better economic recovery and better management of tensions. As such, CDD programs aim to provide one mechanism through which to address a range of local needs. 3.2 Mechanisms Implicit in the ‘theory’ of CDD are a number of claims about how the program has impacts and how performance is superior to other ways of addressing goals. These can be expressed as a set of hypotheses. These will be evaluated in the next section; they are outlined just briefly here.27 Improved local public infrastructure and more private assets H1. CDD results in a better matching of project resources with local needs. Aid ineffectiveness is often a result of the wrong things being supplied. Project staff may not understand the needs of those they are trying to assist. Giving recipients control over project resources may reduce information asymmetries leading to an allocation of project resources that fit with local needs.28 H2. CDD is more efficient because it reduces unit costs. Because communities are in control of project funds, they have greater incentives to ensure they are used efficiently. This may create bottom-up accountability of those involved in project procurement or financial management, limiting corruption. This may mean that CDD resources go further than those spent through other approaches. H3. CDD projects are more likely to be maintained than other projects. Community members may value the projects completed through CDD funding because of perceived greater ownership. This may make it more likely for communities to contribute towards maintaining public goods than for other projects where they have less say. In areas affected by violent conflict, communities may take steps to ensure CDD-financed buildings are not destroyed. H4. CDD can more effectively work in high violence areas than other projects. CDD projects may rely less on ‘outsiders’ than other development approaches. Procurement can be done locally, and planning (including facilitation) can be delegated to locals. This may make it possible for projects to continue operating in high conflict areas. CDD projects may also be more likely to reach people in rural violent conflict-affected areas than other post-conflict approaches that are often capital-centric. Improved social relations and cohesion H5. CDD increases participation in civic life beyond the project leading to improvements in trust between different (conflicting) groups. Well-facilitated CDD projects involve participation from across the community. Involvement at various stages in programs may improve the inter-group relationships, through the promotion of collective action 27 These are generated from a number of World Bank publications, especially Mansuri and Rao (2004), Barron et al. (2004), Chase and Woolcock (2005), World Bank (2006), World Bank (2007a) and Barron, Diprose and Woolcock (2006, 2007). 28 This draws on work on the importance of local knowledge in decision-making, such as Scott (1998) and Chambers (1999). 15 that operates across groups. This in turn may lead to improvements in trust that mitigate against violent conflict.29 H6. CDD results in wider acceptance of project resource distributions, limiting project- related conflict. Development resources can exacerbate destructive conflict.30 Aid, of all types, creates ‘winners’ and ‘losers’, and in the process can lead to conflict that (at times) turns violent.31 Where state systems are weak, the most vulnerable may not receive aid. CDD projects typically aim to create consensus within communities on how resources are allocated. Even when people do not get their proposals funded, they may be more accepting if they have been involved in the process of decision-making. Community-based targeting can also ensure that those most in need receive benefits. H7. CDD provides mechanisms for defusing project-related conflicts. CDD projects typically have mechanisms for dealing with project-related problems such as complaints systems and locally stationed facilitators. This may defuse any latent tensions, limiting escalation and associated social impacts. More effective and responsive institutions H8. CDD creates demand for more responsive institutions. CDD projects typically aim to provide an alternative template for how state project-related functions can be performed. If beneficiaries are happy with CDD process, they may demand that other projects and services are delivered in similar ways or following similar principles. H9. CDD increases interactions between state officials and villagers improving trust between them. If CDD incentivizes more responsive actions from government, it may lead to increases in the perceived legitimacy of the state, in the process eroding a driver of conflict. H10. CDD creates a common platform for planning and resource allocation improving coordination. Post-conflict areas often receive large amounts of aid from numerous agencies. Coordination can be weak, between aid agencies and with government. CDD projects offer the potential for common cross-sectoral planning at the local level, in the process improving coordination between actors. IV. THE IMPACTS OF CDD IN POST-CONFLICT & CONFLICT-AFFECTED AREAS 4.1 Dealing with Economic Deprivation Analysis of the different CDD operations shows that, by and large, they are effective at addressing sources of economic deprivation that lie within communities, and that this can impact on poverty levels. Projects can have such impacts across the range of conflict 29 This is implicit in the work of Varshney (2002) on riots in India. He argues that the presence of institutionalized associational interaction across ethnic groups limits the potential for violence incidence and escalation. 30 See, for example, Anderson (1999) and Uvin (2002). 31 In some conflict areas, aid has inadvertently supported one side to the conflict, such as assistance to refuges in Congo and Tanzania, some of which was used to arm rebel groups (see MSF 1994 and Kaldor 1999). Such cases are extreme. Yet elsewhere, some people will benefit from aid and others will not, unleashing processes of contention that have the potential for escalation. 16 contexts, including in areas where large-scale violence is occurring and in places with deep social divisions. The evidence of economic impacts is strongest for KDP/PNPM. Alatas (2005) shows that the first phase of KDP had positive impacts on communities’ income, compared with matched control areas, and that these increases are greater over time. KDP-2 resulted in per capita consumption gains among poor households that were 11 percent higher in project areas than matched control locations; the proportion of households moving out of poverty was 9.2 percent higher; and vulnerable households were 4.5 percent less likely to fall into poverty (Voss 2008).32 Evaluations of the KALAHI also suggest positive economic impacts, although the quality of evidence is somewhat lower.33 PNPM has been scaled-up to a nationwide program, and the KALAHI covers more than half of the provinces in the Philippines; they operate in some conflict-affected and post-conflict areas, but also in others where violence is not a significant problem. Unfortunately, data is not disaggregated by region, making it difficult to assess whether positive impacts also hold in higher conflict areas.34 Evidence is more limited for high violence and post-conflict areas, but does suggest that CDD projects can have economic impacts in these environments. Assessments of the Community Empowerment Program in Timor note that it built infrastructure that matched with local needs and that projects, in particular those addressing water supply and irrigation issues, were likely to address poverty. However, no data was collected that allows for the measurement of poverty impacts (Conway et al. 2003).35 BRA-KDP, which operated in post-conflict Aceh, resulted in a strong set of welfare gains: poverty declined by 11 percent more in villages that received the program compared with similar villages that did not (Barron et al. 2009). Given that almost 90 percent of funds were used for purchasing private goods, this suggests that CDD projects can be an effective mechanism for distributing one-off cash transfers in emergency situations (Morel, Watanabe and Wrobel 2009). For almost all of the projects, components that focused on providing credit have been less successful. Evidence from the two Mindanao projects is far more limited. The project performance report for Phase 1 of the Mindanao Trust Fund contains no information on economic 32 Both studies utilized rigorous ‘difference of difference’ assessments, which match treatment and control sub-districts, with data collected before and after implementation. Unfortunately such evaluations are rare for other projects in the region. 33 As the KALAHI is only completing its first phase of implementation a full impact evaluation has not yet been completed; however, a mid-term assessment showed average economic rates of return of 20 percent for the project (Araral 2006). (For KDP, average ERR is 53 percent—Torrens 2005). The project’s mid-term evaluation report also found that in treatment locations, 60 percent of households reported their households were accessible all year round after one round of the program, compared to 44 percent at baseline. However, there is little evidence of significant differences in welfare outcome indicators between treatment and control locations (Asia-Pacific Policy Center 2007: 4). The midterm was field after just one cycle (out of three) of the project, and only one-third of the treatment villages actually received a project in this round. 34 There is some evidence of impacts in high conflict areas. Barron et al. (2010), for example, note the effects of KDP in North Maluku where large-scale ethno-religious violence occurred. See also National Management Consultants (2002). 35 The otherwise critical report of the Independent Evaluation Group (2006) acknowledged the success of the project in building a large amount of infrastructure and creating short-term employment rapidly, although it also questioned the long-term appropriateness of the infrastructure that was built (IEG 2006: 12-13). 17 impacts (Social Impact Inc. 2009).36 The ARMM Social Fund post-project assessment provides only anecdotal evidence. It finds, for example, a 6.3 percent greater decline in the proportion of households in project area Lanao del Sur experiencing reduced food consumption than in control locations. But the evaluation does not provide data on this indicator for other project areas (ARMM Social Fund 2009). Through which mechanisms are projects having economic impacts? Is it a result of better matching of project resources with local needs? This is a difficult hypothesis to test given the inherent subjectivity of needs. Yet a few claims can be made and empirically supported. First, for all the projects for which there is evidence, community members tend to be happy with fund use and usually feel that funds are spent more effectively than for other development projects.37 Second, given consensus on the need for local infrastructure such as village-to-market roads in post-conflict areas (e.g. World Bank 2009), it is clear that the types of projects financed tend to be appropriate.38 The experience in Aceh shows strongly how giving communities choice over project money can lead to effective use of funds. Despite pressures from some facilitators to use funds for public goods, communities most often chose to distribute funds in cash, which was often used to help clear land that could not be farmed during the conflict. The result was a doubling of usable land for conflict victims in areas that received the BRA-KDP program (Morel, Watanabe and Wrobel 2009). That said, it is important to recognize the limitations of CDD projects in spurring growth in violence-affected and post-conflict areas. The types of local infrastructure typically financed by CDD projects are unlikely alone to lead to large-scale sustainable poverty reduction; CDD alone is not able to undertake the vast infrastructure reconstruction often needed in post-conflict areas. CDD projects, however effective, can be only one component of an economic response in high conflict and post-conflict areas.39 The evidence is also reasonably strong that per unit costs are typically substantially lower for CDD projects than for others. Table 3 provides data from the KALAHI and KDP/PNPM, the only two of the projects considered that collected systematic data on input costs. 36 This may be because poverty reduction is not part of the project’s development objective. However, other project documentation does note that the Fund’s impact will be in part through its impact on growth and development (World Bank 2007d: vi). 37 In Aceh, 88 percent of recipients felt that both BRA-KDP and KDP funds were spent on the most important needs (Barron et al. 2009). The ARMM SF assessment (2009: 27) notes that “all FGD participants and key informants commend the process undertaken by the ASFP in terms of subproject selection and implementation … They said that this is the first time they encountered a project that seek [sic.] meaningful involvement of the community members at all phases of the project.” The MTF-RDP report notes initial community skepticism followed by strong community commitment to the project (Social Impact Inc. 2009: 5). Supervision reports and aide memoires from CEP paint a similar picture. 38 One critique of CEP is that the infrastructure it financed was often not directly productive (IEG 2006). The same critique has been made of BRA-KDP, where communities often chose to rebuild community centers and prayer halls, and could be leveled against the Youth Development Program in Timor, which is financing mainly sports facilities. However, financing such things may help to build community cohesion. 39 This seems obvious yet it is sometimes forgotten by advocates of community-driven approaches. It is also a point neglected by the Independent Evaluation Group in its criticisms of CEP for not being able to rebuild large amounts of Timor’s infrastructure after the referendum violence. 18 Table 3. Estimated unit cost savings of CDD approach KALAHI-CIDDS KDP (%) (%) Roads and bridges 8 – 59 3240 Water supply 71-76 36 Health centers 44 - Irrigation - 24 Source: World Bank (2007c) These savings are consistent with those reported for CDD projects in other conflict- affected contexts. A project in Kosovo, for example, was able to build and rehabilitate schools 19 percent cheaper than other projects that did not have significant community participation (World Bank 2006: 8). Data on the maintenance of CDD-financed local infrastructure is limited. None of the projects considered collects time series data on infrastructure maintenance.41 A number of the supervision mission reports for CEP and the projects in the Philippines note concerns about maintenance. There is limited evidence from elsewhere that projects with community participation and ownership are maintained more effectively.42 It has been claimed that communities in Afghanistan are more likely to actively protect buildings financed by the National Solidarity Programme than other buildings during conflict.43 But it is hard to find conclusive evidence. Finally, can CDD projects work more effectively in high conflict areas than other projects or move into recent post-conflict areas more quickly? Again, there is little formal comparative evidence on the degrees to which different types of programs can work during periods of large-scale violence. The evidence does show that CDD programs can operate even in areas of high conflict and in places with serious social divisions, even within communities. KDP kept going in most of the high conflict areas when violence was at its highest (National Management Consultants 2002). In Aceh, the program operated throughout martial law when most other projects pulled out. In Central Kalimantan, Poso, Maluku and North Maluku, where large-scale ethnic or religious violence took place, there were temporary suspensions but the program was able to re-start after a short period. The three Philippines projects have managed to keep going in challenging environments in Mindanao and elsewhere, although the ‘quick response’ component of ARMM was very slow in part because of security conditions (ARMM Social Fund 2006). The Thai pilot is able to work despite suspicions and 40 Other estimates put savings higher. Cliffe, Guggenheim and Kostner (2003: 4) report that KDP roads cost US$ 4,000 per kilometer compared to US$ 11,000 when built by the public roads agency, controlling for the types of technology used. School repairs showed a similar difference costing US$ 2,000 compared to US$ 5,600 when using local contractors. 41 World Bank (2007: 26-27) reports “mixed evidence that CDD approaches produce sustainable O&M [operations and maintenance] systems” based on a review of projects in Cambodia, Vietnam, Philippines, and Indonesia. 42 See Khwaya’s (2001) study of projects in Pakistan, cited in Mansuri and Rao (2004). 43 From discussions in Kabul, September 2007. The same has been claimed for infrastructure financed through the Poverty Alleviation Fund in Nepal. 19 tensions locally. In East Timor, the independent evaluation of CEP concluded that it was able to work quicker than other projects.44 Supervision reports from all projects note the challenges but also the resilience of projects (and their facilitators). In many cases, other development projects have had to pull out. 4.2 Promoting stability To what extent are CDD projects effective as a mechanism for limiting violent conflict and instability? Experience shows that CDD projects alone do not affect the likelihood that areas will experience violent conflict, at least in the short run. However, there is some evidence that over time they can improve levels of trust and interaction, which may lead to resilience against local level violence. Few studies have been undertaken that have sought to directly assess the impacts of CDD programs on levels of violence.45 Barron, Diprose and Woolcock (2006) use reports of violent conflicts in local newspapers in two “medium level” conflict provinces to assess the impact of KDP. They find little evidence of a project effect.46 The ARMM Post- Project Review also found little evidence of positive impacts drawing on survey evidence: local disputes had declined in both treatment and control areas, but greater decline in control areas (ARMM Social Fund 2009: 53-54). An interim evaluation of Afghanistan’s National Solidarity Programme finds no evidence that NSP affects the prevalence of village disputes or tribal feuds, or the probability of a village suffering an attack (Beath et al. 2010). A 2004 USAID report cited in World Bank (2006: 35) argues that there is no evidence from anywhere of CDD impacting on broader political violent conflicts. The lack of such observed impacts should not be surprising. It is difficult to envision through what mechanisms projects could affect sources of tension that are exogenous to the communities in which the projects work.47 Indeed, claims that CDD projects are a “solution” to violent conflict should be viewed with suspicion (Barron, Diprose and Woolcock 2007). Yet projects may indirectly contribute to stability in the long-run through their effect on aspects of social relations such as trust, inter-group collective action, and local participation. This may make local level violent conflict less likely and communities more robust to outside violence. Here, there is some evidence of impact although, unfortunately, it is all-too-often anecdotal. Project reports from the Youth Development 44 “The overall evaluation of the poverty impacts of the CEP program is that it reached all villages of Timor with funds for projects that addresses [sic.] basic poverty issues, and that it did so within a shorter time than any other efforts being made following the emergency situation in early 2000” (Conroy et al. 2003: 13). 45 A new study is assessing the impact of the KALAHI-CIDDS program on conflict but results are not yet available. 46 Violent conflicts were more likely in areas that had received KDP in one province (NTT) and slightly less likely in another (East Java). Deaths were greater in KDP areas for both. The authors argue that this does not show that the project is making areas more prone to violence, as there are systematic differences between areas that have received the project and those that have not—i.e. the former areas are in general poorer—that may be correlated with violent conflict risk (Barron, Diprose and Woolcock 2006, chapter four). 47 This is one of the bases of Li’s (2007) critique of KDP. She argues that in focusing on reshaping communities, it ignores broader power relations that keep people subservient and that are the real drivers of revolt and violence in the post-Suharto era. 20 Project in Timor, ARMM Social Fund, Mindanao Trust Fund and Community Empowerment Project all cite examples of communities working together across conflict divides, but it is unclear how common this is and whether there are any systematic differences with non-project areas.48 Three studies on the social impacts of KDP, the KALAHI and BRA-KDP provide more concrete evidence. In two provinces of Indonesia, Barron, Diprose and Woolcock (2006) find that across a range of different identity cleavages, KDP significantly contributed to improvements in inter-group relations. Where villages had the program for longer, these impacts were greater. There was also great variation in changes based on local conditions and (more importantly) on the performance of local facilitators. The KALAHI study found that the project changes social dynamics and practices in villages, but that these impacts were not uniformly positive or negative. Labonne and Chase (2008) report that the project leads to general increases in trust, but that it results in declining trust in neighbors and a decline in collective action. They argue that trust is a function of repeated interaction. Building roads (a major component of KALAHI expenditures) improves trust with those in other villages; but such projects do not increase trust within communities as villagers already interact with each other (Labonne and Chase 2010). BRA-KDP in Aceh is a particularly interesting example given its focus on supporting post-war village-level reintegration. The evidence for improvements in social cohesion is weak (Barron et al. 2009). Levels of social acceptance of returning groups, reported social tensions, divisions and conflict and community efficacy are similar between villages that received the program and villages that did not. There is evidence that BRA- KDP resulted in lower levels of acceptance of ex-combatants by conflict victims. The authors hypothesize this may be a result of the project ‘empowerment’ effect, whereby communities felt more able to stand up to ex-combatants who tried to steal money from the program. It could also be the case that relations with ex-combatants were diminished when they tried to capture funds from the program.49 The limited positive impacts on social relations may be a result of the program only running for one year (the KDP study found that such impacts were much stronger in years three and four of a 48 The assessment of the Mindanao Trust Fund (Social Impacts Inc. 2009: iv) is illustrative: “The process of starting and implementing a CDD project appears to provide one such opportunity for community members to work and solve common problems together. For example, focus group participants in a case study reported that the project at least partially addressed the root cause of the conflict because it promoted better understanding, cooperation and a sense of closeness among citizens. In another case study, focus group participants reported, nine months after project completion, that the Muslims, Christians and Indigenous People were united and sustained good relationships. And in another case study, focus group participants also noted that internal community relations had improved, including between Christians and Muslims.” There is no comparison with control areas, or account of the frequency with which such changes were observed across project locations. 49 BRA-KDP funds were not for ex-combatants who were to be targeted by other programs. However, problems with other reintegration programs meant that community funds often arrived before assistance to ex-combatants. This led to some ex-combatants trying to capture funds from the program. This may have increased resentment from communities. Interestingly, where communities chose to provide some assistance to former fighters, improvements in relations were greater (Morel, Watanabe and Wrobel 2009). 21 program) and because the program provided mainly private goods, which involved less cooperative work.50 Projects of course may be as likely to cause conflict as resolve it. This is particularly true for CDD projects that involve explicit competition over finite resources. Project reports shows that this is rarely a problem in areas where large-scale violent conflict is not active. Even in ‘fragile’ villages, in post-conflict areas such as Aceh and Maluku, project- related processes rarely lead to violence.51 Where escalated violence is ongoing the ability of projects to operate in ways that generate consensus over resource prioritization is more limited. In Thailand, two block grants were provided to different ‘communities’ within one village (Local Development Initiative 2010). The KDP study found that in areas of localized violent conflict, villagers broadly accepted the outcomes of project processes, even when they ‘lost’, because they had opportunities to participate in decision-making. Only where corruption or malfeasance occurred did project-related conflicts escalate (Barron, Diprose and Woolcock 2006). The study also showed that KDP-related conflicts were far less likely to escalate into violence than were conflicts related to other development projects (Table 4). The reasons for this included the presence of effective complaints channels to defuse tensions before they escalated. Table 4. Development Conflicts and Their Impacts in Indonesia Type of # Conflicts # Violent Fatalities Injuries Properties Program Conflicts (% Damaged Violent) KDP 42 1 (2%) 0 2 0 Other 713 36 (5%) 4 38 11 Government Other 19 1 (5%) 0 1 0 Program Source: Barron, Diprose and Woolcock (2006: 57) Yet while grievance systems are part of all the CDD projects considered in this paper, in most projects they are acknowledged as not working effectively. One of the most commonly cited problems in supervision mission reports is that complaints handling systems are not operating as intended (if at all). Clearly this is an area that needs attention. 4.3 Strengthening Institutions 50 Evidence from elsewhere is also limited. The most rigorous study to date involved a randomized assessment of a small CDD project in Liberia, which found that program processes can enhance collective action. Using a field experiment, Fearon, Humphreys and Weinstein (2009) find communities that received the program contribute 9 percent more of their own resources for community public goods. The NSP interim evaluation found increased social interaction on some measures (such as inter-village connections among women) but found no evidence of improved trust between villagers (Beath et al. 2010). 51 Evidence from post-genocide Rwanda, for example, shows that divided communities can work together to decide on the allocation of public goods (Cliffe, Guggenheim and Kostner 2003). On how this process worked in North Maluku, Indonesia, see Barron et al. (2010). 22 The CDD projects considered in this paper typically aimed at promoting institutional change at two levels. At the community or inter-community level, projects have the objective of creating new avenues for citizen-state interaction, and generating bottom- up demand for better governance from local officials. At a higher level, many of the projects aim to consolidate (or reform) national or regional state institutions and to help facilitate coordination amongst the different actors involved in providing development and reconstruction assistance. Many of the CDD projects considered appear to be leading to institutional changes at the local level. Macro impacts are less clear, in part because in complex transitional environments they are hard to measure. The goals at the community level are largely similar for each of the projects considered but approaches differ. CDD projects typically provide a model for accountable decision- making in the hope that community satisfaction with ‘doing things differently’ results in demand for changes in the way other non-project related decision-making occurs. In some cases, this involves working closely with existing formal power structures; in others, more explicit attempts are made to reengineer local institutional structures. For example, in Timor, CEP excluded village chiefs from decision-making roles in projects;52 other projects such as the ARMM Social Fund have a formal role for local government staff. The evidence shows mixed impacts from different programs.53 The KALAHI mid-term evaluation finds that in places that received the project, there is now greater attendance in (non-project) village assemblies and participation in development planning compared with control locations (Asia-Pacific Policy Center 2007). Labonne and Chase (2008: 19-20) find that participation in non-project meetings increased by 20 percent more in treatment areas than control locations after one cycle of the program. Importantly, this also appeared to lead to some changes in the behavior of officials: elected village officials met more often with villagers in KALAHI areas than in the control locations. In villages that received funds from the project, trust towards local officials increased by 10.7 percent. Decisions that affect the village were also increasingly made during village assemblies as a result of the project, controlling for baseline values. Community and local leader preferences were equally represented in community proposals (Labonne and Chase 2009). Barron, Diprose and Woolcock (2006) report similar results for KDP. They find that there has been a “positive reconfiguration of citizen-state relations at the local level”, helping to democratize village life. Fifty percent more villagers reported that more marginalized groups were coming to (non-project) village meetings than in the past, compared to matched control areas. The Philippines and Indonesia findings fit with 52 Those involved in the project at the time point out that when CEP was first implemented there was a lack of clarity on who exactly village leadership was. Often within villages, there were several people who felt they were in charge, each of whom was trying to establish (or maintain) a power base at the local level. Formal leaders tended to be different people to traditional leaders, who held much de facto power. In a rapidly transitioning environment, the project chose not to support village heads in order to avoid playing in to power struggles taking place, and to avoid consolidating any one model, when these struggles were still playing out. Comments from Gillie Brown, former task team leader of CEP. 53 One reason for this may be that institutional impacts are more difficult to measure than economic impacts. Methodological approaches, such as household surveys, suitable for understanding economic and poverty effects may be less suitable for measuring cognitive and behavioral effects in the non- economic realm. 23 results from the interim evaluation of NSP in Afghanistan that finds that the project leads to an increase in the frequency of village meetings and villagers attendance.54 In contrast, other projects had greater difficulties in affecting such change. The mid- term review for the ARMM Social Fund notes: “the ASFP investments were apparently unable yet to instill in the minds of most people the local government’s capability in confronting rising poverty, in delivering basic services like those in education and health, or in establishing meaningful linkages with capitalists/industrialists” (ARMM Social Fund 2006: 38). An evaluation of BRA-KDP showed that the project had no impact on citizen-state relations (Barron et al. 2009). The impacts of CEP on the practices of local elites were limited by the exclusion of village chiefs from project decision-making councils (Chopra and Hohe 2004). This led to resistance from elites and undermined the legitimacy of the project fora. Which factors caused variation in impacts? Three spring out. First, is the length of time projects have worked in an area. Barron, Diprose and Woolcock (2006), for example, show that the governance impacts associated with KDP increase over time, with villages that have had the project for three or more years far more likely to see changes in local political behavior. Projects like BRA-KDP involved only one cycle of assistance; it is unsurprising that this does not change governance practices, which have evolved over decades or longer. Second, and related, the quality of project facilitation is key. Within projects, there appears to be massive variation in local institutional impacts. In part this is a result of varying local contexts; where an environment is “conducive” to change, projects are more likely to act as a catalyst (Barron, Diprose and Woolcock 2007). Yet equally important is the ways in which facilitators engage with local leaders and communities. This is not just a function of their capacity (e.g. for community mobilization and conflict mediation). The “positionality” of facilitators is also important; in Afghanistan, for example, facilitators prove to be more effective when they are from the same tribe as the communities they work in.55 Supervision reports for a number of the projects show successes against the odds where facilitators have worked hard with local communities and leaders. Third, project design is important. CDD projects often face a difficult dilemma. They seek to support behavioral and cognitive changes at the local level—the empowerment aim. Yet they are only likely to be successful if they involve existing local elites who may be resistant to such changes (Powis 2007). There are inherent trade-offs and getting them right will, in part, dictate a program’s impact.56 Maintaining support at higher levels is important: both BRA-KDP and CEP were closed before their planned duration 54 NSP is found to improve male villagers’ perceptions of government and to increase interactions between villagers and officials, but it does not make villagers more accepting of state taxation or jurisdiction over local crimes. NSP also increases the number of functioning local councils and increased attendance at village meetings. But it does not affect villagers’ perceptions of the quality of local governance (Beath et al. 2010). 55 Thanks to Rob Wrobel for this observation. 56 See Chopra and Hohe (2004: 299-303) who identify different options for participatory intervention by the degree of “social engineering” they undertake. Sometimes choices are dictated by political imperatives. CEP, for example, was under pressure from Timorese leaders to exclude those who had been village chiefs during the Indonesian occupation from decision-making roles (Cliffe, Guggenheim and Kostner 2003: 9). 24 because of a lack of support from government agencies in Banda Aceh and Dili. What works in one place will not in another. Perhaps the most important lessons are (a) to ensure that there is flexibility to alter a project’s institutional arrangements over time, as project experience grows, and (b) to undertake continuous consensus building, making it clear how a well functioning project can serve the interests of local politicians, civil servants and other elites. Both of these objectives require extensive and continuous supervision and management attention. This is particularly key in areas prone to, or affected by, violent conflict. Unfortunately, supervision tends to be seriously under- funded. Task team leaders are often in charge of multiple projects or based at a distance from where they operate. This limits the ability to address emerging problems, seize opportunities, and learn from experiences. Where local flexibility and widespread consensus-building has been achieved, such as with PNPM which has expanded across all of Indonesia, there has been scope for projects to influence broader political dynamics. Community-based approaches, which channel resources down and preferences on developmental priorities up, now characterize the broader suite of Indonesia’s anti-poverty programs. While causality is difficult to prove, KDP/PNPM may have shifted the way that the government thinks about poverty and development issues, and have played a role in changing the ways in which citizen-state accountability is structured. Ultimately such shifts should translate into improved vertical relations between citizens and higher-level institutions, key in areas that are, or have been, affected by violent unrest. In areas of separatist violence (such as Aceh, Mindanao, and southern Thailand) conflict cleavages are not just between different social groups but also between citizens and the state. Yet such changes are slow and hence often hard to observe. And in areas where national governments have not embraced the widespread use of CDD approaches, or where other government actions undermine community trust, these impacts are not likely to eventuate. Documentation for projects such as the southern Thailand pilot, BRA-KDP and the ARMM Social Fund is explicit about the goal of improving state-society relations. But in the absence of broader processes of reform, these goals are perhaps over-ambitious. It is hard to see how CDD projects could alone affect these dynamics, at least in the short-term. Community perceptions of the state are likely to be a function of state performance in a range of areas, including service delivery, impartial provision of security and justice and, in many poor areas, the equitable provision of public sector jobs (Barron 2010). Recognition of local cultural and religious practices is also often important. Altering perceptions thus requires more extensive changes in the way the state functions. Government commitment to reform is more important than the design and functioning of any one program. CDD projects may play a role in the longer-run: they have the potential to shape attitudes to reform. This has been the case in Indonesia, where more than ten years into KDP/PNPM, decision on other government resources are now made based on bottom-up processes. As government officials realize CDD-type mechanisms provide a useful means to distribute resources in ways that consolidate their positions, they are often increasingly supportive of them, and this may help shape norms that can have powerful effects. But such virtuous cycles are not entirely under the control of the project and its designers and supporters; in the absence of broader buy-in from national governments to the principles of CDD-type programming, project effects in this area are likely to be limited. 25 Finally, CDD projects offer the potential of providing a mechanism for coordinating the approaches and programs of different donors and government at the local level. This is extremely important in post-conflict areas, where often hundreds of agencies are providing assistance. CDD programs offer a common channel through which different agencies can provide funds; alternatively, CDD programs can be used for developing a common local needs assessment, which different donors (and governments) can use to coordinate their assistance. There is little evidence, however, of this happening systematically. Other donors provide funds through KDP in Indonesia and Aceh and the two Mindanao projects. Yet more projects worked through separate delivery systems. A rare example of effective collaboration amongst donors and government was the delivery of public information materials related to the peace process through the KDP network in Aceh. Yet when it came to providing reintegration assistance, other development agencies preferred to use their own mechanisms and the needs identification generated by the program did not impact the programming of other agencies (Barron and Burke 2008). Ownership of programs by national and local governments is key if better coordination is to happen. At a national level, PNPM has been fairly successful at harmonizing donor assistance under one umbrella. This only happened because PNPM became the cornerstone of the government’s stated policy framework, and because the government was strong enough to insist donor’s played game. Where the government was weaker, as in Timor, it was impossible to harmonize donor approaches and hence programs. V. CONCLUSIONS CDD projects can be an effective mechanism for dealing with local drivers of poverty in areas affected by violent conflict. The projects reviewed in this paper have had positive impacts on welfare, although more rigorous evidence would be desirable. They tend to be cost effective. There is little evidence on improved maintenance, but the trend towards community contributions would lead us to expect that projects may be more likely to be maintained than others. Further, the projects considered have been able to have such impacts in a range of conflict environments, including areas experiencing large-scale armed violence. Yet it is also important to note the limitations of CDD projects: alone, they cannot transform the economies of conflict-affected and post- conflict areas. This requires investment in larger-scale infrastructure and policy responses to promote growth. CDD is only part of a strategy for development in (post)conflict areas. There is little evidence that CDD affects aggregate levels and impacts of violent conflict, whether it is localized violence or larger-scale violent unrest. However, when projects work well, and are in areas for a number of years, they can have indirect conflict impacts, affecting social relations and behavior in ways that may make communities more robust to dealing with local problems. This may prevent future conflict escalation. This has relevance both in areas affected by localized violence and in those experiencing larger-scale violent conflict, where master conflict narratives often belie the local nature of tensions (Kalyvas 2006). There is also evidence that projects may be less likely than other development approaches to lead to violent conflict over project resources. Key here is the performance of complaints systems and local facilitators. Unfortunately, a 26 quick assessment of aide memoires and supervision reports show consistently that training, complaint units and monitoring and supervision are the elements of CDD operations that tend to receive the least attention. In areas experiencing larger-scale violence, capacity tends to be more limited, only increasing the importance of having these things in place. In higher violence areas, larger supervision and monitoring budgets will be necessary to ensure projects function well. CDD projects, when designed well, can also have institutional effects at the local level. Impacts on vertical linkages between citizens and the government take much longer and are only likely if accompanied by broader processes of reform. In some cases, such as PNPM in Indonesia, CDD approaches have affected to some extent politics at a national level and, as a result, have potential of contributing to a strengthening of the social compact in the longer-run. But where (potential) improvements from CDD projects are not linked to larger changes in accountability between society and the state, such impacts will be less. As such, it is unclear how small-scale projects (such as the Thailand pilot) will seriously impact upon the main drivers of conflict. Efforts to utilize CDD projects as a mechanism for coordination of development aid have also been less than successful in most cases. CDD is thus no ‘magic bullet’ response to conflict. Most of the CDD programs discussed do not have an explicit conflict resolution aim. But the evidence does suggest that CDD projects can support poverty reduction, help build social cohesion at the local level and strengthen local institutions. Importantly, CDD projects have been able to do this in places where violence rages, in communities that are divided, and in weak institutional environments. Broader impacts are contingent on higher level political support and processes of change. Maintaining, or generating, wide-ranging political support is thus extremely important. Who to get on-board will differ from area to area; understanding this is particularly important in conflict-affected areas. CDD projects need to be conceived as part of broader governance reform strategies aimed at improving institutional performance and accountability relations. This requires understanding deeply the (often local) political economies that shape patterns of conflict and violence and development outcomes. Both project implementation and context are important. Context can be conceptualized at multiple levels: villages within an area affected by one conflict share a common context, but more localized factors—such as the quality of leadership, historical memories of conflict and local identity cleavages—also shape the context. Projects can have local level effects in places where the context is not “conducive” (e.g. in divided communities). But they will only do so if project design is built on a clear understanding of local realities and constraints. (Projects have had mixed success here, largely a function of the level of investment in analytic work to understand local conditions and of levels of supervision and monitoring). These will change over time, so the ability to adapt approaches is important. Again, this requires ongoing analytic work and a large investment in supervision. One-off conflict or social assessments will rarely provide information of the types necessary for projects that operate in dynamic environments. What a well-designed CDD operation looks like will vary between areas; and understanding of local conditions is essential for designing effective programs. Yet as important as up-front analysis is ensuring that there are continuous flows of information—on challenges faced, on successes achieved—which 27 can allow for programs to be re-designed as they go along. This is particularly important given that CDD operations such as the KALAHI, CEP and KDP/PNPM cover(ed) large territories with very different conditions in different areas, and that conditions change frequently in conflict-affected areas. A final conclusion relates to the quality of existing evidence on CDD impacts, how these are achieved, and sources of variation in performance. Most of the projects considered in this paper are not collecting basic information that allows for an assessment of impacts.57 Strategies for generating ongoing information on both project performance and on the contexts in which they are working is lacking for most projects. CDD projects have an amazing resource in the networks of facilitators stationed in villages. Projects need to develop strategies to utilize these for data collection to provide real-time information on local conditions and to allow for responses to be adapted as local conditions change. 57Mansuri and Rao (2004) provide a comprehensive review of independent and peer reviewed evaluations of CDD operations. But very few of these are of projects in East Asia, or of projects operating in conflict zones. 28 REFERENCES Alatas, Vivi (2005). “An Evaluation of Kecamatan Development Program.” Mimeo. Jakarta. Anderson, Mary B. (1999). Do No Harm: How Aid Can Support Peace – Or War. Boulder, CO: Lynne Rienner. Areral, E. (2006). “Mid-Term Project Review Economic Analysis.” Manila: World Bank. ARMM Social Fund (2006). “ARMM Social Fund Project: Mid-Term Review Mission, January 23-February 3, 2006. Aide-Memoire.” World Bank project report. ARMM Social Fund (2009). “ARMM Social Fund Post-Project Assessment.” Draft final report. 23rd June. Asia-Pacific Policy Center (2007). “Tracking Progress Towards Community Empowerment & Welfare: KALAHI-CIDDS Midterm Evaluation Report.” Manilla: Asia-Pacific Policy Center. Barron, Patrick, with Paul Arthur, Peter Bartu, and Teresa Whitfield (2010). “Sub- National Conflicts in Middle and Higher Income Countries: The Cases of Aceh, Basque Country, Mindanao, and Northern Ireland.” Background paper prepared for the 2011 World Development Report. Mimeo. Barron, Patrick and Adam Burke (2008). Supporting Peace in Aceh: Development Agencies and International Involvement. Policy Studies 47. Washington, DC: East- West Center. Barron, Patrick, Rachael Diprose, David Madden, Claire Q. Smith, and Michael Woolcock (2004). “Do Participatory Development Projects Help Villagers Manage Local Level Conflicts? A Mixed Methods Approach to Assessing the Kecamatan Development Program, Indonesia.” Conflict Prevention and Reconstruction Paper No. 9 (revised edition). Washington, DC: World Bank. Barron, Patrick, Rachael Diprose, and Michael Woolcock (2006). “Local Conflict and Community Development in Indonesia: Assessing the Impact of the Kecamatan Development Program.” Indonesian Social Development Paper No. 10. Jakarta: World Bank. Barron, Patrick, Rachael Diprose and Michael Woolcock (2007) Local Conflict and Development Projects in Indonesia: Part of the Problem or Part of a Solution? Policy Research Working Paper No. 4212. Washington, DC: World Bank. Barron, Patrick, Macartan Humphreys, Laura Paler, and Jeremy Weinstein (2009). “Community-Based Reintegration in Aceh: Assessing the Impacts of BRA-KDP.” Indonesian Social Development Paper No. 12. Jakarta: World Bank. Barron, Patrick, Sri Kusumastuti Rahayu, Sunita Varada, and Vita Febriany (2010). “Disturbing the Equilibrium: Movements Out of Poverty in Conflict-Affected Areas of Indonesia,” in Deepa Narayan and Patti Petesch (eds.), Moving Out of Poverty: Rising from the Ashes of Conflict. Washington, DC: World Bank/Oxford University Press. 29 Barron, Patrick and Joanne Sharpe (2008). “Local Conflict in Post-Suharto Indonesia: Understanding Variations in Violence Levels and Forms Through Local Newspapers.” Journal of East Asian Studies 8(3): 395-424. Bates, Robert H. (2000). Prosperity and Violence: The Political Economy of Development. New York: W.W. Norton & Co. Bates, Robert H. (2008). “State Failure.” Annual Review of Political Science 11:1-14. Beath, Andrew, Fotini Christia, Ruben Enikolopov, and Shahim Ahmad Kabuli (2010). Randomized Impact Evaluation of Phase II of Afghanistan’s National Solidarity Programme (NSP): Estimates of Interim Program Impact from First Follow-up Survey. Kabul: NSP. Call, Charles T. with Vanessa Wyeth (eds.) (2008). Building States to Build Peace. Boulder, CO.: Lynne Rienner. Cernea, Michael M. and Scott E. Guggenheim (1993). Anthropological Approaches to Resettlement: Policy, Practice and Theory. Boulder, CO: Westview Press. Chambers, Robert (1999). Rural Development: Putting the First Last. London, Lagos and New York: Longman. Chase, Robert and Michael Woolcock (2005). “Social Capital and the Micro-Institutional Foundations of CDD Approaches in East Asia: Evidence, Theory, and Policy Implications.” Paper given at the Arusha conference ‘New Frontiers of Social Policy’, December 12-15. Chopra, Jarat and Tanja Hohe (2004). “Participatory Intervention.” Global Governance 10: 289-305. Cliffe, Sarah, Scott Guggenheim, and Markus Kostner (2003). “Community-Driven Reconstruction as an Instrument in War to Peace Transitions.” Conflict Prevention and Reconstruction Paper No. 7. Washington, DC: World Bank. Colletta, Nat J., Markus Kostner, and Ingo Wiederhofer (2003). “Disarmament, Demobilization, and Reintegration: Lessons and Liabilities in Reconstruction,” in Robert I. Rotberg, ed. When States Fail: Causes and Consequences. Princeton, NJ: Princeton University Press. Colletta, Nat J. and Michelle L. Cullen (2000). Violent Conflict and the Transformation of Social Capital: Lessons from Cambodia, Rwanda, Guatemala and Somalia. Washington, DC: World Bank. Collier, Paul (1999). “On the Economic Consequences of Civil War.” Oxford Economic Papers 50(4): 168-83. Collier, Paul, Lani Elliot, Havard Hegre, Anke Hoeffler, Marta Reynal-Qeurol, and Nicholas Sambanis (2003). Breaking the Conflict Trap: Civil War and Development Policy. New York: Oxford University Press. Collier, Paul, Anke Hoeffler, and Mans Soderbom (2008). “Post-Conflict Risks.” Journal of Peace Research 45(4): 461-478. 30 Collier, Paul (2009). Wars, Guns, and Votes: Democracy in Dangerous Places. New York: Harper. Conroy, John, Jakob Krikemann Hansen, Lene Ostergaard, and Bob Scouller (2003). “East Timor—Community Empowerment and Local Governance Project Independent Review.” Draft mimeo. November. Cramer, Christopher (2006). Civil War is Not a Stupid Thing: Accounting for Violence in Developing Countries. Hurst & Co.: London. Dwyer, Arienne M. (2005). The Xinjiang Conflict: Uyghur Identity, Language Policy, and Political Discourse. Policy Study 15. Washington, DC: East-West Center. Fearon, James D., Macartan Humphreys, and Jeremy M. Weinstein (2009). “Can Development Aid Contribute to Social Cohesion after Civil War? Evidence from a Field Experiment in Post-Conflict Liberia.” American Economic Review: Papers & Proceedings 99(2): 287-291. Geneva Declaration Secretariat (2008). Global Burden of Armed Violence. Geneva: Geneva Declaration Secretariat. Ghani, Ashraf and Clare Lockhart (2008). Fixing Failed States: A Framework for Rebuilding a Fractured World. New York: Oxford University Press. Guggenheim, Scott (2006). “Crises and Contradictions: Understanding the Origins of a Community Development Project in Indonesia” in Anthony Bebbington, Michael Woolcock, Scott Guggenheim, and Elizabeth Olsen (eds.) The Search for Empowerment: Social Capital as Idea and Practice at the World Bank. Bloomfield, CT.: Kumarian. Independent Evaluation Group (IEG) (2006). “Timor Leste Project Performance Assessment Report: Community and Local Governance Project, Second Community Empowerment Project, Third Community Empowerment Project, Agricultural Rehabilitation Project, Second Agicultural Rehabilitation Project.” Report No. 36590. June 27th. Washington, DC: World Bank. International Crisis Group (ICG) (2004). “Myanmar: Sanctions, Engagement or Another Way Forward?” ICG Asia Report No. 78. Yangon/Brussels: ICG. International Crisis Group (ICG) (2008). “The Philippines: The Collapse of Peace in Mindanao.” ICG Briefing No. 83. Manila/Brussels: ICG. International Crisis Group (ICG) (2009a). “Southern Thailand: Moving towards Political Solutions?” ICG Asia Report No. 181. Bangkok/Brussels: ICG. International Crisis Group (ICG) (2009b). “The Philippines: After the Maguindanao Masscare” ICG Asia Briefing No. 98. Manila/Brussels: ICG. Hoeffler, Anke and Marta Reynal-Querol (2003). Measuring the Costs of Conflict. Washington, DC: World Bank. Kaldor, Mary (1999). New and Old Wars: Organized Violence in a Global Era. Cambridge: Cambridge University Press. 31 Kalyvas, Stathis N. (2006). The Logic of Violence in Civil War. New York: Cambridge University Press. Kalyvas, Stathis N. (2008). “Promises and Pitfalls of an Emerging Research Program: The Microdyanmics of Civil War” in Stathis Kalyvas, Ian Shapiro, and Tarek Masoud (eds.) Order, Conflict and Violence. New York: Cambridge University Press. Khwaja, Asim Ijaz (2001). “Can Good Projects Succeed in Bad Communities? Collective Action in the Himalayas.” Mimeo. Harvard University, Department of Economics, Cambridge, MA. Labonne, Julien and Robert S. Chase (2008). “Do Community-Driven Development Projects Enhance Social Capital?” Policy Research Working Paper No. 4678. Washington, DC: World Bank. Labonne, Julien and Robert S. Chase (2009). “Who is at the Wheel When Communities Drive Development? Evidence from the Philippines.” World Development 37(1): 219-231. Labonne, Julien and Robert S. Chase (2010). “A Road to Trust.” Mimeo. World Bank. Li, Tania Murray (2007). The Will to Improve: Governmentality, Development, and the Practice of Politics. Durham, NC.: Duke University Press. Local Development Initiative (2010). “Progress Report: Piloting Community Approaches in Conflict Situation in Three Southernmost Provinces in Thailand – May 1 to October 31, 2009.” Mansoob Murshed, Syed, Mohammad Zulfan Tadjoeddin, and Anis Chowdhury (2009). “Is Fiscal Decentralization Conflict Abating? Routine Violence and District Level Government in Java, Indonesia.” Oxford Development Studies 37(4): 397-421. Mansuri, Ghazala and Vijayendra Rao (2004). “Community-based and driven development: a critical review.” World Bank Research Observer 19(1): 1-39. Medicin Sans Frontiers [MSF] (1994). “Breaking the Cycle: Calls for Action in the Rwandese Refugee Camps”. November 10. Available at www.doctorswithoutborders.org/publications/reports/before1999/breaking_1 994.cfm Moore, Barrington (1966). Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making of the Modern World. Boston: Beacon Press. Morel, Adrian, Makiko Watanabe, and Robert Wrobel (2009). “Delivering Assistance to Conflict-Affected Communities: The BRA-KDP Program in Aceh.” Indonesian Social Development Paper No. 13. Jakarta: World Bank. Moser, Caroline O. N. (2006). “Reducing Urban Violence in Developing Countries” Policy Brief 2006-1. Washington, DC: Brookings Institute. 32 MSR (2010). Multi-Stakeholder Review of Post-Conflict Programming in Aceh: Identifying the Foundations for Sustainable Peace and Development in Aceh. Banda Aceh/Jakarta: MSR. Muggah, Robert, ed. (2009). Security and Post-Conflict Reconstruction: Dealing with Fighters in the Aftermath of War. London and New York: Routledge. Nagarajan, Geetha, and Michael McNulty (2004). “Microfinance Amid Conflict: Taking Stock of Available Literature.” Washington, DC: United States Agency for International Development. National Management Consultants (2002). “Kecamatan Development Program: KDP in Conflict Areas.” Jakarta: Ministry of Home Affairs Community Development Agency and National Management Consultants. Pathan, Don (2009). “Conflict and Development in Thailand’s Southernmost Provinces.” Mimeo. World Bank. September. Ploughshares (2009). Armed Conflicts Report: Philippines—CPP/NPA. Plougshares. < http://www.ploughshares.ca/libraries/ACRText/ACR-PhilippinesN.html> Accessed 5th January 2010. Polanyi, Karl (1944). The Great Transformation. Boston: Beacon Press. Powis, Benjamin (2007). “Systems of Capture: Reassessing the Threat of Local Elites,” Social Development Working Paper no. 109. Washington, DC: World Bank. Restrepo, Jorge, Brodie Ferguson, Jukliana M. Zúñiga, and Adriana Villamarin (2008). “Estimating Lost Product Due to Violent Deaths in 2004”. Unpublished background paper for the Small Arms Survey. Geneva/Bogota: Small Arms Survey/CERAC. Rohland, Klaus and Sarah Cliffe (2002). “The East Timor Reconstruction Program: Successes, Problems and Tradeoffs.” Conflict Prevention and Reconstruction Paper No. 2. Washington, DC: World Bank. Schiavo-Campo, Salvatore and Mary Judd (2005). “The Mindanao Conflict in the Philippines: Roots, Costs, and Potential Peace Dividend.” Conflict Prevention and Reconstruction Paper No. 24. Washington, DC: World Bank. Scott, James C. (1998). Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed. New Haven: Yale University Press. Skocpol, Theda (1979). States and Social Revolutions: A Comparative Analysis of France, Russia and China. New York: Cambridge University Press. Social Impact Inc. (2009). “Mindanao Trust Fund – Reconstruction and Development Program: Project Performance Report, Phase 1.” Submitted to the MTF Secretariat, World Bank. October. Torrens, Anthony (2005). “Economic Impact Analysis of the Kecamatan Development Program Infrastructure Projects.” Jakarta: World Bank. 33 Torres, Wilfredo Magno (2007). Rido: Clan Feuding and Conflict Management in Mindanao. Manilla: The Asia Foundation. Uvin, Peter (2002). “The Development/Peacebuilding Nexus: A Typology and History of Changing Paradigms” Journal of Peacebuilding and Development 1(1): 5-24. Varshney, Ashutosh (2002). Ethnic Conflict and Civic Life: Hindus and Muslims in India. New Haven: Yale University Press. Varshney, Ashutosh, Mohammad Zulfan Tadjoeddin, and Rizal Panggabean (2008). “Creating Datasets in Information-Poor Environments: Patterns of Collective Violence in Indonesia, 1990-2003.” Journal of East Asian Studies 8(3): 361-394. Voss, John (2008). Impact Evaluation of the Second Phase of the Kecamatan Development Program in Indonesia. Jakarta: World Bank. Welsh, Bridget (2008). “Local and National: Keroyakan Mobbing in Indonesia.” Journal of East Asian Studies 8(3): 473-504. World Bank (2006). “Community-Driven Development in the Context of Conflict- Affected Countries: Challenges and Opportunities.” Social Development Department Report No. 36425. Washington, DC: World Bank. World Bank (2007a). “Enabling East Asian Communities to Drive Local Development: East Asia Region CDD Flagship Report.” Unpublished mimeo. World Bank (2007b). Timor-Leste’s Youth in Crisis: Situational Analysis and Policy Options. Washington, DC: World Bank. World Bank (2007c). “Economic Results of CDD Programs: Evidence from Burkina Faso, Indonesia and the Philippines.” Social Development Notes No. 109. Washington, DC: World Bank. World Bank (2007d). “Philippines: Mindanao Trust Fund Reconstruction and Development Program.” Washington, DC: World Bank Social Development Unit. World Bank (2009). Aceh Growth Diagnostic: Identifying the binding constraints to growth in a post-conflict and post-disaster context. Jakarta: World Bank. 34