EXPANDING AND DIVERSIFYING INDONESIAÕS PROGRAM FOR COMMUNITY EMPOWERMENT, 2007 Ð 2012 SYNOPSIS In 2007, Indonesia embarked on a multiyear effort to expand an innovative community-driven development program, first started in 1998, into the largest program of its kind in the world. For nearly a decade, the Kecamatan Development Program had empowered communities to determine how they wanted to use funds for their own development, whether for small infrastructure projects, health and education, or microcredit opportunities. Communities planned, implemented, and maintained projects on their own through village and intervillage committees. The program experienced very low levels of corruption, and in some communities it was the only government program to provide direct benefits that actually reached citizens. It was also successful in raising the incomes of Indonesians in poor parts of the country. From 2007 to 2012, the central government significantly expanded the program and launched pilot projects to extend participation to geographic areas and activities beyond the scope of the original program. Although the scale-up strained management, creating occasional delays and gaps in implementation, the program continued to raise the incomes of the poorest Indonesians. This case study offers several lessons about scaling up community-driven development. Jonathan Friedman drafted this case study based on interviews conducted in Jakarta, Indonesia in May 2013. The research benefited from additional interviews conducted by Rushda Majeed in October 2013. The case was prepared by ISS in partnership with the World Bank as part of the BankÕs Science of Delivery initiative. This case study is the second in a two-part series; see ÒServices for the People, by the People: IndonesiaÕs Program for Community Empowerment, 1998 Ð 2006.Ó Case published February 2014. INTRODUCTION At a cabinet meeting in 2005, Indonesian President Susilo Bambang Yudhoyono asked how his government could halve the national poverty rate from 16.7% to 8%, by 2009, the end of his first term.1 ÒAt the time, we had 52 poverty [reduction] programs, driven by 27 central ministries, but there was no linkage between them,Ó said Sujana Royat, deputy minister of poverty alleviation in the Coordinating Ministry for PeopleÕs Welfare, the ministry that coordinated efforts between government ministries that planned, monitored, and analyzed social welfare programs. In response, Aburizal Bakrie, the minister, presented two graphics. The first showed five of the governmentÕs largest programs, ranked by effectiveness. The second ISS is a joint program of the Woodrow Wilson School of Public and International Affairs and the Bobst Center for Peace and Justice: www.princeton.edu/successfulsocieties. ISS invites readers to share feedback and information on how these cases are being used: iss@princeton.edu. showed five estimates of how the independent expansion of each program would affect poverty. Two related programsÑthe Kecamatan Development Program (KDP) and the Urban Poverty ProgramÑoutperformed the others in the analysis. The KDP, based in the countryÕs subdistricts, or kecamatan, was a community- driven development program run by the Ministry of Home Affairs in cooperation with the national development planning agency, Bappenas. It provided block grants for poor rural communities, which were free to use the money, with some restrictions, for local projects of their choosing. The sizes of grants depended on communitiesÕ population densities, poverty incidence, and remoteness. Villagers determined communal needs through open discussion and developed their own proposals for projects like the construction of roads, bridges, and irrigation and drinking- water systems. Then they managed the implementation of projects that intervillage, subdistrict forums selected for feasibility and likely impact. The program tried to ensure that resources reached intended beneficiaries by depositing funds directly into community bank accounts. Project and budget details were posted on public information boards in the villages to promote transparency, and a village verification team monitored results. The central government initially financed the program through a series of World Bank loans and grants, but by 2002 it had increased its own contribution to nearly 22%. Households in KDP areas moved out of poverty in greater proportions than those in non-KDP areas. Studies found that most infrastructure projects were high quality, had high economic rates of return, and cost less than if the work had been handled through traditional government channels.2 The roads that connected villages to markets and the irrigation systems that opened up new areas to cultivation had some of the most significant economic impacts. The KDPÕs rapid expansion in support of areas affected by natural disastersÑsuch as the region of Aceh following a devastating tsunami in December 2004 and the region of Nias following an earthquake in March 2005Ñ added to the programÕs visibility and popularity. In the mid-2000s, as IndonesiaÕs economic growth buoyed federal tax revenues, the government had additional funds available to address poverty on a broader scale. The KDP was already one of the largest community-driven development programs in the world, active in roughly half of rural Indonesian villages. BakrieÕs presentation persuaded the president that the program had achieved significant gains. In August 2006, Bakrie unveiled an ambitious plan to combine the KDP and other community-based poverty reduction programs into a National Program for Community Empowerment (Program Nasional Pemberdayaan Masyarakat Mandiri), or PNPM Mandiri. (In Bahasa, IndonesiaÕs official language, the term mandiri means self- reliant and independent.) Outside the cities, the program operated under the name PNPM Rural. Under the presidentÕs plan, the KDP (renamed PNPM) would expand into every rural village across the nationÕs sprawling archipelago by 2009. Now the question was how to scale up the community-driven development program even further, expand the range of projects to include social programs as well as public works, and reach stigmatized or disenfranchised segments of the population. THE CHALLENGE The KDP started in 1998, at a turning point in Indonesian history. Nearly overnight, the Asian financial crisis of 1997Ð98 had reversed decades of economic growth and poverty reduction. The crisis reduced incomes, bred public demand for more effective and accountable government, and spawned an era of Reformasi, or Reform. Public anger over corruption, financial collapse, and rising poverty forced President Suharto from power after 32 years of authoritarian rule. Following decades of repression, media and civil society groups could for the first time speak more openly about corruption and transparency in government programs. The reforms were far- reaching, including a 1999 Big Bang decentralization law that devolved many services to district governments. During that period, it became possible to assemble high- level political support for a program that paid little heed to long-standing patronage systems. From 1998 to 2006, 34,233 villagesÑ about half of IndonesiaÕs total of 70,000Ñ participated in the KDP initiative, though not all in the same project cycle. The program covered 1,983 subdistricts, out of 4,290. Evaluations showed that large numbers of villagers attended KDP forums in their communities, participated in village-based KDP committees, and received leadership and technical training.3 In addition, the studies found that communities maintained the new infrastructure for years after completion. However, not all poor households benefited from KDP projects, and the benefits were mixed in areas that had higher average incomes, in part because basic economic infrastructure was already in place.4 To respond to the presidentÕs decision to scale up the program, the ministry had to address four broad challenges. It had to (1) reach populations that had not adequately benefited from the program; (2) extend its network of functional and oversight staff to thousands of additional subdistricts, many in remote areas; (3) maintain the programÕs sound record on preventing corruption; and (4) deepen engagement with elected local governments, helping build the PNPM principles of participation, transparency, and accountability in local government administration. At the same time, it was also important to maintain past gains in community empowerment and poverty reduction. Responsibility for meeting those challenges fell mainly to the Ministry of Home Affairs and the management structure it had created in the later years of the KDP. The ministry administered the program through a secretariat, by means of parallel operational and oversight systems. The secretariat consisted of a small number of civil servants from the ministry and roughly 40 Indonesian consultants in charge of overseeing the project. (Civil service rules prohibited the ministry from adding significant numbers of employees in the Community Development Agency, which was the office at Home Affairs in charge of PNPM Rural. To expand the agency beyond a certain size required either an exception to civil service rules or the creation of a larger directorate just for PNPM.) Daily operational responsibility rested with another team of private consultants who reported to the secretariat. The Ministry of Home Affairs hired Indonesian firms to implement PNPM Rural on the national and regional levels. The top level, called the National Management Consultants, was based in Jakarta. Seven Regional Management Consultants units, each responsible for several of the countryÕs 34 provinces, had offices both in Jakarta and in the provinces. Both groups had specialists in financial management, complaints handling, information management, infrastructure, training, and communication. The National Management Consultants managed teams of facilitators, who worked closely with the villages. Initially, the facilitators came from private firms, but later, they were hired individually and were usually from the communities they assisted.5 At first, there were two types. Empowerment facilitators encouraged participation in PNPM projects and built villagersÕ capacity through training and regular consultation. Their partners, technical facilitators, had backgrounds in civil engineering; they reviewed project design and monitored progress on public works projects. Later, districts added a third specialistÑin financial managementÑin part to better advise and monitor womenÕs revolving-loan fund activities funded by PNPM grants. At every level, facilitators monitored and supported their counterparts one level below and reported to facilitators one level up. Many of the procedural problems and other issues that arose could be dealt with at local levels through these layers of support and oversight without the need for direct intervention by the central ministry. To meet the demands of the scale-up, Home Affairs had to hire more consultants at the national and regional levels to provide support and oversight. The ministry also had to hire and train thousands of additional facilitators at the district and subdistrict levels. Because PNPMÕs continued success hinged on the facilitatorsÕ performance, the ministry had to find good communicators who could handle the sometimes sensitive job of working with ordinary citizens who had to learn how to identify communal needs and lead and manage projects. Further, the ministry had to extend and improve its financial-control processes to handle the large sums of money it had to distribute to thousands of subdistricts through PNPM. Tracking and accountability were more important than ever to ward off corruption. FRAMING A RESPONSE For assistance both in financing the expansion and in deciding what adjustments to make, the Ministry of Home Affairs turned to familiar associates. The World Bank and other development partners had supported the KDP, the PNPMÕs predecessor, through a series of grants and loans. Home Affairs, together with other government ministries, urged the partners to organize a dedicated trust fund for the support of PNPM. The new partnership was called the PNPM Support Facility (PSF or support facility). A Joint Management Committee chaired by the State Ministry of National Development Planning, or Bappenas, and co- chaired by the World Bank, decided on the use of support facility funds. A technical secretariat did much of the legwork in support of the committee. Other major development partners and participating ministries, such as the Coordinating Ministry for PeopleÕs Welfare, served as members. The World Bank managed the fund and employed staff to carry out associated activities. The World Bank, Australia, Denmark, and the Netherlands were early donors; the European Union, Canada, the United States, and the United Kingdom contributed resources later on. The PSF assisted Home Affairs in the latterÕs main tasks by providing monitoring and training services and conducting its own financial management reviews. Additionally, the PSF expanded the learning and innovation functions that had been strong parts of the KDP program. PSF-funded program evaluations would help Home Affairs determine how PNPM performed and where there were needs and opportunities. The PSF also provided a mechanism to pilot new variants of PNPM that targeted marginalized groups or to focus on development sectors that villagers generally did not elect to fund through PNPM block grants. Royat, the deputy minister at the Coordinating Ministry for PeopleÕs Welfare, said PSF supervision of financial matters was especially important in buttressing the Ministry of Home AffairsÕ internal supervisory mechanisms. ÒI wanted accountability to be first-class, so I requested of our development partners . . . that each penny, each rupiah, can be traced,Ó he said. PNPM Rural based its community-driven development process directly on the KDP, empowering communities to determine their own development needs (see Textbox 1). GETTING DOWN TO WORK In April 2007, President Yudhoyono officially launched PNPM Rural in the city of Palu in Sulawesi Central Province, more than 1,600 kilometers northeast of the capital, Jakarta. The choice of location was apt. Royat described the mayor of Palu as Òhungry about being effectiveÓ and eager to try out Òthe great innovations in combating poverty.Ó Throughout 2007 and 2008, Home Affairs worked to double project coverage. With support from the PNPM Support Facility (PSF), Home Affairs had to build a national network, hire and train thousands of new facilitators, and prevent corruption from taking root. Building Ties with Local Government The first step was to reach out to district and subdistrict governments that had not yet participated in PNPM RuralÑto explain how the program worked as well as the funding and oversight roles of local governments. Support from local government was crucial under PNPM Rural. The central government and development partners provided about 80% of the money for each block grant, but districts had to contribute about 20% from their own budgets depending on their financial resources. Though KDP had similar matching-fund arrangements, under PNPM Rural, national and subnational governments had to sign agreements in which both sides committed to specific funding levels. (In neither instance did the local governments have a say in how to spend the money.) Ayip Muflich, director general of community empowerment at Home Affairs, said local governments generally responded well to Home AffairsÕ overtures but added that it was easier to win acceptance in some districts than in others. He attributed the variation partly to different levels of local- government commitment to the principles of community empowerment. He said some local governments resisted PNPM Rural because of a perception that there were limited opportunities to benefit from corruption due to heavy monitoring of project funds and the programÕs ban on the use of contractors. Some local officials also appeared to play more active roles in organizing project stakeholders and using their local influence to support PNPM, such as in the district of Purwakarta, whereas others merely met the minimum funding and administrative requirements. Expanding the Network of Facilitators Citizen participation in PNPM was largely a function of facilitatorsÕ communication abilities, coaching skills, and commitment to community empowerment principles. ÒIf facilitators are not good enough, then it influences the implementation process,Ó said Home AffairsÕ Bito Wikantosa. From 2006 to 2009, Home Affairs tripled the number of facilitators to 1,249 at the district level and to 9,810 at the subdistrict level.6 It advertised facilitator positions through local media on a province-by-province basis. Wikantosa said it was fairly easy to Textbox 1: Four Stages of Community-Driven Development (1) At the beginning of each project cycle, the Ministry of Home Affairs organized provincial and district workshops in order to announce a new round of block grants and to make villagers aware of what had to be done to qualify for the grants. This phase usually took one or two months. The ministry invited local government officials, community leaders, and representatives of the local media, universities, and civil society groups to attend. The PNPMÕs subdistrict facilitators convened an intervillage forum that included government officials, respected community members, and three to six representatives from each village to agree to a schedule for moving ahead. This sensitization or socialization phase then extended to the villages, which each convened large, open meetings. Villagers elected two volunteer facilitators: one man and one woman. Fifty to 100 people typically attended village meetings. The PNPMÕs district and subdistrict facilitators trained village and intervillage teams and provided support at every stage. (2) Planning and prioritization came next and lasted two or three months. Subdistrict and village facilitators led discussions about communal problems and needsÑroad access or school materials, for exampleÑand collected project ideas. WomenÕs groups met separately to develop their own proposals. At a second open village meeting, community members debated proposals and collectively decided on up to three priority needs, two of which had to come from womenÕs groups. At the same time, the villagers selected representatives to write down the proposals by using simple templates that included each projectÕs location, the expected number of beneficiaries from each project, and a rough cost estimate for each. A subdistrict verification team comprising village leaders as well as PNPM subdistrict facilitators and technical staff reviewed the proposals for feasibility, technical quality, and likely impact. (3) During the selection phase, the intervillage forum reconvened to discuss the merits of village proposals, based on the verification teamÕs findings. The forum ranked the proposals based on feasibility, impact, how well villages had maintained previous projects, contributions from villagers, and other factors. Facilitators then helped communities whose proposals received high marks prepare detailed design plans and cost estimates. The PNPMÕs technical facilitators evaluated proposals, and a district engineer inspected the designs as well. The forum then met again and selected winning proposals, seeking consensus. The forum also elected an intervillage financial management unit to oversee project funds. (4) The head of the financial management unit, the subdistrict social facilitator, and elected village representatives next opened a joint bank account. The subdistrict-level officials from the Ministry of Home Affairs certified the selected projects and submitted a letter to the provincial branch of the Treasury to release the block grant to the communal account. The funds were released to the subdistricts in installments of 40%-40%-20%, based on the status of the work, and on to the villages as needed. To withdraw funds, the social facilitator, financial management head, and village representatives had to sign off. After receiving the first installment of the grant, a community elected a three-person village implementation team. Elected village implementation teams contracted with suppliers, paid poor villagers to supply labor, and tracked progress. Procurement of supplies exceeding 15 million rupiah (about US$1,500) followed simplified competitive-bidding procedures. The teams reported on project progress at village accountability meetings multiple times throughout the project cycle. District engineers had to certify completion of the projects before the Treasury released the final 20% installment. Following completion of the projects, village implementation teams presented results and accounted for funds to the villagers, and villages selected operation and maintenance teams to sustain the projects. recruit social facilitators because they could come from a variety of backgrounds. However, technical facilitators had to have knowledge of civil engineering, a requirement that posed a challenge in less-developed parts of Indonesia. Even in the countryÕs most populous island of Java, where civil engineers were in relatively large supply, it could be difficult to find people willing to serve in rural areas on the island. After reviewing applications, Home Affairs invited top applicantsÑusually 10% more than it would ultimately hireÑto a preservice training, so as to observe recruitsÕ communication abilities and performance on tests. ÒWe can see whoÕs good at communication and facilitation, so we can see which ones are good facilitators and which ones donÕt match the requirements,Ó Wikantosa said. The Ministry of Home Affairs added the names of members of the remaining 10% to a reserve list and hired from that pool if active facilitators resigned their positions or could not perform adequately. The predeployment curriculum covered the principles of community empowerment, the fundamentals of coaching and conflict resolution, and the specifics of the PNPM Rural project cycle. Because social facilitators came from a wide variety of backgrounds, said Wikantosa, training in how to encourage communities to organizeÑto support collective actionÑwas crucial. Support and training continued once facilitators were in the field. In addition to annual refresher courses for all facilitators, district facilitators met twice a month with their subdistrict counterparts to discuss progress and problems. They also provided on- the-job training for subdistrict facilitators as problems arose in their subdistricts. At the provincial level, regional consultants organized bimonthly conferences for all district facilitators. There, facilitators communicated their experiences and ideas about solving common problems in facilitation or project procedures. The events had other benefits, too. For example, district facilitators from Purwakarta in western Java said that story sharing at these meetings helped inspire facilitators and cultivated a friendly sense of competition (see Textbox 2). Beyond the regional variations in the pool of technically trained candidates, the ministry faced challenges in recruiting and retaining facilitators,. Working conditions often were poor in remote areas, and travel between villages within subdistricts was difficult. In addition, facilitators did not receive annual pay raises, and their contracts covered only the period of the project cycle. Budget delays at the central level left facilitators working for months at a time without being paid. And as a result, the positions offered little job stability and few clear opportunities for professional advancement. In response, Home Affairs developed a certification program for facilitatorsÑwith support from the PNPM Support Facility and the National Board for Professional Certification, a government institution that certified all professions in Indonesia. An association of PNPM facilitators (Ikatan Pelaku Pemberdayaan Masyarakat Indonesia) had pushed for such a program. MuflichÑwho led the certification program beginning in 2012, following his departure from Home AffairsÑexplained that the goal of the certification process was to develop the position of facilitator as a profession in Indonesia, thereby making the work more attractive to potential recruits. Still, the Ministry of Home Affairs continued to struggle to fill facilitator positions. In 2009, the ministry reported vacancies in 5% of positions, with significant variation between provinces; and the problem persisted.7 Textbox 2: A View from the Ground The district of Purwakwarta is located between the Indonesian cities of Jakarta and Bandung on the island of Java. Purwakarta contained 17 subdistricts, including 13 in rural areas. Before the scale-up, four subdistricts had participated in the Kecamatan Development Program. By 2009, however, all 13 participated. District facilitators in Purwakarta described the positive and negative impacts of the scale-up in their district. Most importantly, they said, more people were able to participate in the program and benefit from new roads, bridges, and health facilities. Additionally, the district government took a greater interest in supporting the program because it then had district-wide coverage and did not favor some areas over others. The district government enacted regulations to promote village maintenance of PNPM Rural projects. It also began offering small cash rewards for the top performing intervillage financial management teams, and provided additional funds for training of intervillage committees not covered by the central projectÕs training budget. On the negative side, the district facilitators said the original four subdistricts became less enthusiastic about PNPM Rural following the scale-up, as some of the excitement over the project had come from its exclusivity. Additionally, they said they had to oversee more subdistrict facilitators and projects than before the scale-up, but did not receive additional resources to do so. Richard Gnagey, head of a group of advisers to the Ministry of Home Affairs, said the burden on facilitators to oversee large numbers of committees and projects was too big in some cases. ÒWhen you have too many villages to be in charge of, that makes it difficult, because you really need to spend a lot of time in villages,Ó he said. ÒThere a limit to how many people you can supervise or coach effectively.Ó Combating Corruption In combating corruption, PNPM Rural was able to continue the solid work of its predecessor program, the KDP. One of the astonishing findings about the KDP was its very low level of corruption despite the huge sums disbursed. Audits by PriceWaterhouse and Moores Rowland found financial inconsistencies in less than 1% of projects, though even some of the programÕs strongest champions suspected that corruption was underreported and that the record, though strongly positive, was not quite that good.8 Home Affairs was determined to extend the programÕs record on corruption in new project locations and sustain it among existing project areas. Scott Guggenheim, a World Bank official who had helped craft the KDP and had served as initial head of the PSF, said, ÒThe model has always been that the first best measures on dealing with fraud are design measures.Ó The first and most important design principle, Guggenheim said, was to keep the design simple by limiting discretion and reducing the number of transactions. For example, the project limited the number of financial transactions by disbursing funds directly from provincial branches of the National Treasury Office to communal bank accounts at the subdistrict level. Transparency was key. An example was the use of multiple signatures to set up and withdraw funds from the account, including those from the head of the intervillage financial management unit, the subdistrict social facilitator, and elected village representatives. To take advantage of villagersÕ social capital, the design also created checks and balances by separating responsibilities for project selection, budgeting, implementation, oversight, and maintenance across committees and then by using their reports in the villageÕs overall discussions of how well the program was performing. The projectÕs modular design further facilitated the application of appropriate, nondistorting sanctions. Locating oversight at the subdistrict level ensured that if corruption was discovered, the grant to that subdistrict could be suspended without penalizing neighboring subdistricts. Limiting central government liability to the successful delivery of block grants similarly made it possible to enlist ministry support for addressing local level problems without antagonizing the entire government apparatus. The Ministry of Home Affairs conducted regular financial and compliance audits of subdistricts, which were also increasingly examined annually by the Audit Board of Indonesia. An earlier, randomized controlled trial conducted by the US-based Poverty Action Lab showed that the risk of audit had a greater impact in deterring corruption than some forms of community oversight did. Based on that research, in 2011 the Audit Board began to audit 20% of subdistricts each cycle.9 Communities could also take the initiative by reporting fraud to a dedicated PNPM Complaints-Handling Unit via phone, SMS, mail, or local drop boxes. Monitoring and Oversight The PNPM Support Facility had dedicated units to monitor budgets and activities in support of the Ministry of Home Affairs in this regard. Monitoring from Jakarta was possible because of regular reporting up the ranks from villages to the central level, but field visits and financial audits and compliance audits were also parts of the oversight strategy. The support facility created risk profiles to serve as an early warning system for a variety of problems that could arise. An implementation support team consisting of a financial management unit, a field unit, and other teams collected data from regional consultants to assess performance by using indicators such as the number of block grants and the amount of money disbursed in an area, turnover rate of facilitators and empty positions, incidence of fraud over previous quarters, number of field visits by consultants and facilitators, strength of a revolving loan fund that communities funded with PNPM grants including percentage of nonperforming and idle loans, and percentage of funds already provided that were used for planned expenditures as well as the unexpended balance. The support facility field team produced color-coded maps that reflected the prior performance and level of risk for each indicator for each of the 34 provinces. Low-risk provinces appeared in green, moderate-risk provinces in yellow, and high-risk provinces in red. Kun Wildan, director of institutions and community training in the Directorate General for Community Empowerment (the responsible directorate at Home Affairs), said the maps represented an easy way to digest a lot of information. He said the PNPM secretariat and the National Management Consultants wanted to develop their own color-coded mapping system going forward Òto help monitor broad coverage . . . to differentiate good ones and worse ones.Ó He said the maps helped people visualize the data and improved communication. Representatives of the support facility met quarterly with the national consultants and Bappenas to discuss which regions were at risk in subsequent months. Support facility teams conducted field inspections based on an assessment of risk factors. They visited an increasing number of subdistrictsÑat least 160 each quarter (out of approximately 5,000 participating in PNPM)Ñbeginning in 2011. On-site, they met with local government officials, facilitators, and village and intervillage oversight bodies to determine how well they understood the PNPM project cycle and how closely the participants followed procedures. When the support teams discovered problems, they were instructed to bring them for discussion before the Joint Management Committee of the PSF so that the Ministry of Home Affairs could take action. If the financial team found evidence of fraud, it would report it to a dedicated complaints-handling facility at the national consultants unit. Adapting to new demands Important new goals were to encourage use of the block grants for a wider range of activitiesÑnot just public worksÑand to expand access for the disenfranchised. And beginning in 2007, the Joint Management Committee of PSF developed several pilot projects to test ways to attain those goals. One of the new programs was PNPM Generasi Sehat dan Cerdas, or PNPM Healthy and Smart Generation, which began in 2007. Indonesia as a whole was making uneven progress toward its 2015 Millennium Development Goals in health and education, with certain poor rural areas lagging significantly behind. The question was how to encourage communities to use KDP funds for health and education instead of infrastructure. Generasi aimed to reduce maternal and infant mortality and increase primary school enrollment in underperforming areas. It adopted the PNPM Rural model, but villages could use its block grants only for health and education projects. Villages used the Generasi grants for a range of activities. The most common education projects were the provision of school materials, such as textbooks, and the provision of financial assistance such as scholarships and transport fees. Communities pursued projects in health that provided financial assistance for pregnant women for their transportation to clinics or to pay midwives and for supplementary feeding activities.10 The allocation formula departed from the PNPM model. In some experimental subdistricts, the sizes of block grants for subsequent project cycles depended on prior performance. In those areas, 80% of block grants were distributed based on the standard PNPM formula for population and poverty incidence; the other 20% was a bonus based on performance in the previous cycle, as measured by 12 health and education indicators. That design enabled observers to test whether such types of performance incentives would lead villagers to focus on health and education and to devise their own ways to solve problems in those areas. Generasi showed mixed results overall, with largely positive outcomes in areas with the lowest baseline performance in health and education.11 Generasi helped reduce childhood malnutrition rates and increase primary school enrollment in some areas, though baseline attendance was already quite high.12 In 2010, the government decided to expand the program from 1,800 pilot villages to almost 4,000 in 2012, with help from the support facility.13 Going forward, Robert Wrobel, head of the Generasi team at the support facility, said Generasi intended to focus on early childhood education and on increasing the supply of health services in rural areas. The expansion of the PNPM program included not only new funding to promote underprioritized types of village activities but also new methods to empower marginalized populations. Evaluations from the first years of the KDP revealed that certain groups were largely excluded from participating in the program at the village level or did not benefit from projects because of social stigmas and other causes of marginalization. Those groups included street children, sex workers, people with HIV/AIDS, indigenous groups, and female-headed households.14 In 2011, in response, the Coordinating Ministry for PeopleÕs Welfare launched PNPM Peduli. The Peduli program differed significantly from the PNPM model. Civil society organizationsÑrather than the governmentÑled implementation. Those civil society organizations had established networks and hard-earned trust among relevant communities. ÒWe felt that those most qualified to design and implement these programs were the community-based civil society organizations that have extensive experience working with [marginalized groups],Ó said Royat.15 The support facility directed grants to three large Indonesian organizations, which helped organize a network of nongovernment groups.16 In addition to funds, the support facility provided the organizations with training in financial management, report writing, and applying methods for working effectively with marginalized groups. The Peduli pilot phase lasted until the end of 2012. In the Peduli programÕs first year, 72 civil society organizations supported a variety of activities that affected more than 15,000 people in 24 provinces.17 Activities included providing birth certificates for street children so they could attend school, assisting indigenous groups with land rights, helping sex workers with professional training and funds to start small businesses, and other projects. Following the pilot, peopleÕs welfare and the support facility decided to extend Peduli for a second phase. OVERCOMING OBSTACLES The Peduli program encountered two obstacles that were not totally unanticipated but were not amenable to easy solution: regional variation and stresses at the top. Regional variation Extending the PNPM model was not easy in all parts of Indonesia. The eastern provinces of Papua and West Papua presented particular challenges for the Ministry of Home Affairs. Located in the western half of New Guinea and surrounding islands, Papua was among the poorest and most underdeveloped regions of Indonesia. Several indigenous Papuan separatist groups had sporadically clashed with security forces since 1963, and at the time of the PNPM scale-up, the area was the home of IndonesiaÕs deadliest conflict. In 2007, the president had announced a New Deal for the area, aimed at increasing development,18 but weak transportation infrastructure, long distances between villages in certain areas, and cultural distinctiveness presented challenges for extending PNPM. The job of the facilitator was especially difficult. In Papua, languages varied between villages even within the same subdistricts. It was usually hard to find a facilitator with the right language skills. Under the KDP and in order to expand the pool of recruits, the planners had gradually reduced the levels of credentials and experience required of people who applied to work as facilitators, and they paired each facilitator with a local resident, called an assistant facilitator, to help translate. Caste differences, too, could make the facilitatorÕs job difficult. Sentot Satria, a former consultant team leader, said, ÒIf you are a subdistrict facilitator from the lowest caste, it is difficult to monitor projects.Ó In 2008, Papuan provincial governments and Home Affairs launched a version of PNPM tailored to Papua, called PNPM Respek (PNPM Respect). Local governments provided funds for block grants, and Home Affairs provided facilitators. Unlike PNPM Rural, though, under the Respek program all villages received funds without an intervillage competitive-proposal process. Additionally, because facilitators had difficulty traversing long distances between villages, which resulted in a slower pace of implementation, PNPM Respek project cycles were permitted to extend beyond the normal 12- to 16-month period. And the local provision of funds freed Respek from central government disbursement deadlines. However, there was another challenge in PapuaÑnamely, the lack of civil engineers who could work as technical facilitators. Wikantosa of Home Affairs said the reason was that initial wage levels offered to facilitators were low, and Papuan civil engineers could easily find other jobs with better pay. Home Affairs raised facilitator salaries in Papua, Wikantosa said, but the increase did not solve the problem. Home Affairs then tried transferring facilitators to Papua from other parts of Indonesia, but many non-Papuan facilitators struggled to navigate the geography and the cultural map of Papua, Wikantosa explained. To correct the problem, the Ministry of Home Affairs revived a training program first introduced in 2003 called Barefoot Engineers. Richard Gnagey, an adviser to the ministry, recalled that at the time, Home Affairs needed to recruit technical engineers for 90 subdistricts in Papua, but only 19 people passed the preservice training, and 10 of those left in the first two weeks. In 2009, Home Affairs launched a second phase, with funding and assistance from the support facility. Home Affairs partnered with a local university in PapuaÑUniversitas CenderawasihÑto recruit and train high school graduates in a six-month course that taught the skills needed to perform as technical facilitators. Upon completion of the course, graduates were deployed as facilitators. In the 2009 cycle, 106 were trained.19 Based on the programÕs success, a third cycle was launched in late 2012 to train an additional 300 technical facilitators out of the total of 560 required for the region.20 Stresses at the top A second obstacle arose at the top levels of management. The program was thinly staffed at its headquarters in Jakarta. Although PNPM Rural accounted for 80% of the Ministry of Home Affairs budget, the implementation of PNPM Rural was the responsibility of just one directorate of one directorate general and consisted of about 10 civil servants. Muflich, the former director general of community empowerment, said the ministry did not elevate the directorate to the level of a directorate general because the project was originally temporary and not one of the ministryÕs core functions. As a result of its small size compared with the magnitude of the task it faced, the directorate sometimes struggled to respond quickly to issues or problems with PNPM, according to the 2012 governance review. It labored to address facilitator vacancies at both the local and national levels and to maintain the frequency and quality of training it offered to support facilitators. The ministry also struggled to disburse funds on time (though the ministry was not always the source of the problem), which delayed project implementation and left facilitators working for months without being paid. Data systems and the capacity to respond to complaints were also strained. When an aspect of the program lagged or Home Affairs did not have enough capacity to correct a problem, the Project Support Facility staff stepped in. Although the aim was to help out the ministry, the practice sometimes confused roles, created ambiguity, and generated additional stress. ASSESSING RESULTS From 2007 to 2009, Home Affairs almost doubled the scope of PNPM Rural from 33,300 villages in 1,971 subdistricts to 57,266 villages in 4,371 subdistricts across rural Indonesia.21 From the launch of PNPM through 2011, communities used project funds to build or repair nearly 69,000 kilometers of roads, 6,500 irrigation systems, 30,000 clean- water systems, 22,000 schools, and 11,000 health facilities.22 Government and development partners spent a combined US$4.4 billion on the program, including for block grants and administrative expenses, such as the wages of more than 9,000 facilitators in 2009 and even more in subsequent years.23 The scale-up of community-driven development influenced poverty levels and participation, but some of the same limitations apparent in the earlier KDP program persisted. Livelihoods A PNPM Rural impact evaluation conducted in 2010 and published in 2012 compared households in communities that started to participate in the program in 2007 with a control group that began to participate only in late 2009 or early 2010, just before data collection started (the closest possible equivalent to a no-treatment counterfactual).24 The study used consumption as an approximation of income. It found that the expanded program sustained the positive results of previous KDP cycles on poverty and household consumption, with some variations, although its impact also had similar limitations. From 2007 to 2010, consumption among poor households that participated in PNPM was, on average, 9.1 percentage points higher than consumption among poor households in the control group. The difference was equivalent to 39,000 rupiah (about US$4) per person per month. The increases were more pronounced for the poorest fifth of households, but the evaluation also found positive effects for households in the second and third consumption quintilesÑthe near poor. In 2009, the average increase in household consumption was 3.3 to 5.7 times the average cost of the program per household. Poor households were 2.1% more likely to move out of poverty in PNPM areas than were poor households in villages that were part of the control group. The gains were uneven, however. Households led by women and households led by men without a primary school education did not experience the same benefits as others did and were no more likely to move out of poverty than were their counterparts in non-PNPM areas.25 (The Peduli pilot program began later on in 2011, after the evaluationÕs study period.) Additionally, PNPM did not substantially Table 1: PNPM Project Scope 2006 2007 2008 2009 Coverage area 25,962 villages in 1,958 subdistricts 33,300 villages in 1,971 subdistricts 34,405 villages in 2,447 subdistricts 57,266 villages in 4,371 subdistricts Funds disbursed (millions of US dollars) 185 190 447 700 Number of program activities 27,849 25,835 49,836 70,006 Number of facilitators 1,633 5,061 6,471 11,059 help nonpoor households or households in nonpoor areas.26 A 2010 PNPM qualitative study provided possible insight into some of the causal mechanisms that linked the program to antipoverty outcomes.27 The study found that in poor areas without basic infrastructures, the needs of the poor and of the general community aligned. Everyone benefited from public works. As a result, communities invested in infrastructure projects that benefited the poor, such as clean- water systems or roads linking villages to markets. By contrast, in nonpoor areas, where basic economic infrastructure was already in place, communities often continued to invest their block grants in additional infrastructure projects. However, in those areas, the primary needs of the poor were not for additional infrastructure but for capacity development and access to capital. As a result, PNPM projects had less impact on incomes and consumption than they did in the poorer areas. A 2008 PNPM evaluation had encouraged PNPM organizers to consider adapting the PNPM model for nonpoor subdistricts, given the KDPÕs lesser effectiveness in those areas.28 But the government did not alter the basic project model, preferring to scale-up first. Susanne Holste, head of the operations team at the PSF, said given the immense task of the scale-up, Home Affairs had taken the right approach in holding to the basic model of PNPM. ÒFirst you achieve a scale-up, and then, when your systems are in place, you can think about offshoots . . . and regional adaptation,Ó she said. But the model had to remain responsive to learning, especially because the need for rural public works projects had diminished. Guggenheim suggested that the program would have to adapt to changing conditions and sustain its impact. Community Empowerment The KDP was originally conceived as a way to empower communities. The theory of change was to build engagement, generate social capital, and gradually win adoption of the principles of participation, transparency, and accountability in the broader social and political contexts. The actual results were mixed, though in some areas, the program was so new that it might have been premature to expect significant changes. According to Home Affairs annual reports under PNPM Rural, participation rates among women and poor villagers remained high. Of the almost 19 million villagers who attended PNPM meetings or participated in project committees in 2009, 53% were women and 49% were poor villagers. Of participants in 2007, 44% were women and 63% were poor villagers. A 2010 report produced by the support facility also found that the positive effect on citizen participation rates in PNPM Rural was sustained from earlier years, although the quality of participation varied.29 And although participation rates remained high, elites enjoyed outsize influence in guiding debates and in performing decision making. Village elites typically felt they represented their entire communities, and nonelite villagers passively agreed to elitesÕ decisions. And though womenÕs participation rates in village meetings (excluding women- only planning meetings) increased, men still dominated the decision-making process.30 A 2012 governance review found similar dynamics.31 The 2010 study also found that villagers did not demand similar rights of participation in local government planning. Instead, they adhered to whatever rules applied to a project, and if non-PNPM programs did not require public participation, villagers would not demand a voice.32 Elite dominance at the village level was one possible reason that the three values at the core of PNPMÑ participation, transparency, accountabilityÑ did not become part of the Òcommon senseÓ with which villagers approached their other relationships, especially their interactions with government or politics. Another was the lack of incentive to apply PNPM principles in local government administration. There was no guarantee that if villagers applied the program principles, they would be rewarded with a project, as in PNPM. Several observers said the participatory aspects of the PNPM project cycle were too mechanistic to become embedded in everyday interactionÑthat at times, facilitators and villagers followed procedures for the sake of fulfilling requirements and filling out reports rather than for animating communal deliberations. The comprehensive manual for facilitators and village committees helped regularize PNPM processes but also tended to overwhelm. Gnagey, the adviser to the Ministry of Home Affairs, also worried that facilitators were overburdened by the number of villages they had to oversee. ÒWhen you have too many villages to be in charge of, that makes it difficult, because you really need to spend a lot of time in villages.Ó Suahasil Nazara, coordinator of a policy working group at the Secretariat of the National Team for the Acceleration of Poverty Reduction in the Office of the Vice President, also observed that by incorporating paid labor and by providing grants, PNPM contradicted some elements of the reciprocity and community commitment that were part of Indonesian tradition. Although he remained committed to the PNPM model, saying it was largely consistent with gotong royong, or joint bearing of burdens, he said, ÒNow there are also studies that show that the PNPM block grant jeopardizes the gotong royong feeling. In the past, people were willing to help without money, but now, if they want to participate in PNPM projects, theyÕre asking for wages.Ó He said he personally did not think working for wages ÒjeopardizesÓ willingness to volunteer and to come together with others for the common good, while noting that the observation perhaps deserves more study. REFLECTIONS Built-in learning was a key component of PNPMÕs success. From the early days of the KDP, the World Bank emphasized rigorous baseline studies and impact evaluations to determine the effectiveness of the program. It financed several randomized, controlled trials, building on new applications of these methods to policy analysis.33 Under the PNPM support facility, development partners contributed even more than before to support research and learning. The support facilityÕs Joint Management Committee determined the schedule and the topics of studies. Natasha Hayward, head of an analytics team, emphasized that the research had to be useful for policy makers. Some studies were conducted to respond rapidly to requests from government agencies, she said. Others sought to inform policy making at more-strategic levels. ÒDemand from government is critical and is the major driving force,Ó Hayward said. ÒBecause of the unique role of PSF as interface, they [PSF] also generate ideas and theories to be tested, but weÕd never pursue those in isolation. It always must be followed up with, ÔIs this a priority for government, consistent with the challenges theyÕre facing?ÕÓ The rigorous evaluations and thematic reports identified gaps in PNPM that sparked the creation of various pilot projects. ÒWhen we know from evaluations, for example, that there are community needs other than infrastructure, we should then explore other interventions that may increase peopleÕs welfare,Ó said Vivi Yulaswati, director of social protection and welfare at Bappenas. ÒWe have good support from PSF not only in evaluations but also in thematic reports, so we learn a lot about what works and what doesnÕt work in communities. PNPM Generasi and Green and Peduli are examples of our effort to improve the effectiveness of PNPM.Ó Studies also created opportunities for government agencies and other stakeholders such as development partners, universities, and civil society organizations to discuss broader policy issues. In that way, the support facility served as a platform for government and nongovernment actors to share ideas. ÒWe serve three or four key government agencies here. They are not always necessarily aligned with one another,Ó Hayward said. ÒItÕs our challenge in some ways to try to steer a path that is responsive to what they want to see and that pushes them to see some of our ideas and thinkingÑto push the envelope a little bit and also to provide a framework around which they can coalesce and agree through us.Ó The support facilityÕs studies went beyond PNPM to look at core issues of poverty, local governance, inclusion, and social accountability throughout Indonesia. Suahasil Nazara, coordinator of a policy working group at the Secretariat of the National Team for the Acceleration of Poverty Reduction in the Office of the Vice President, said these broader studies helped policy makers think strategically about the future of poverty alleviation programs, local governance, and community empowerment in Indonesia. ÒPSF is generating the knowledge. ItÕs the most important function of PSF: producing the knowledge and evidence,Ó he said. Nazara stressed the need to draw on rigorous analysis in trying to explain to other countries why community-driven development has worked: ÒThis is where PSF is important,Ó he said. At the same time, because of the number of factors that influence national poverty rates, it was impossible to measure the direct impact of PNPM Rural on IndonesiaÕs overall poverty rate. However, the nationÕs official poverty rate did decline to 12% in 2012 from 16.6% in 2007. ÒI cannot say the poverty rate is declining just because of PNPM,Ó said Rudy Prawiradinata, director of poverty reduction at Bappenas, Òbut definitely, PNPM is a main vehicle to reduce poverty.Ó EPILOGUE In 2012, the Vice PresidentÕs National Team for the Acceleration of Poverty Reduction and the Coordinating Ministry for PeopleÕs Welfare agreed on a 12-point strategy to strengthen PNPM and thereby ensure the sustainability of the programÕs principles of participation, transparency, and accountability in local government administration. Called the PNPM Road Map, the strategy emerged from years of studies and consultations with government ministries, development partners, and civil society organizations. The Road Map proposed a number of changes. To strengthen PNPM, the government agreed to find solutions to the problems of facilitator vacancies, such as cultivating the facilitator certification program, raising wages systematically, and reducing the administrative burdens on facilitators. It would set up a regular, direct transfer mechanism to get funds from the central government to communities. It promoted good governance through increasing levels of accountability in PNPM programs and expanding and coordinating legal empowerment programs for villagers. And the plan sought to strengthen the village and intervillage committees formed through PNPM by creating a more robust legal basis for their activities. The Road Map also sought to extend PNPM project principles to local government administration. To date, a potential vulnerability under both KDP and PNPM Rural had been the directness of the linkage between the national level and the communities, effectively bypassing local government administration. Accordingly, the plan called for integrating PNPM mechanisms and support structuresÑincluding facilitatorsÑwith existing local planning procedures to form a single local participatory planning system with a common budget. In December 2013, the legislature passed a bill that created a single planning system and built KDP/PNPM into national law. Nazara was optimistic that the billÕs passage would normalize the application of community empowerment principles. ÒIntegration is successful if we can change the law,Ó he said. ÒA great number of principles of PNPM are to be included in the next village law. If thatÕs successful, then village operations will be close to the PNPM model. That is how it should be integrated.Ó Nazara said that despite all of the PNPM studies that looked at the quality of infrastructure produced and the impact on incomes, it was important to remember that PNPMÕs ultimate goal was community empowerment. ÒAnyone measuring PNPM as an antipoverty measure and who looks only at income, I think thatÕs wrong,Ó he said. ÒPNPM comes to say, ÔLook, your voice matters, what do you want to do, what would you like to happen in this village, to see in your community?Õ and thatÕs what matters.Ó Table 2: PNPM Timeline 2006 President Susilo Bambang Yudhoyono announces plans to expand the Kecamatan Development Program and combine it with other projects under the umbrella of PNPM Mandiri. 2007 The president launches PNPM Mandiri in Palu, Central Sulawesi. 2007 The Indonesian government, the World Bank, and other partners establish the PNPM Support Facility (PSF) to provide operational and financial support for PNPM Mandiri. 2007 PNPM Rural reaches 33,300 villages, 1,971 subdistricts, and 343 districts. 2007 The Indonesian government pilots PNPM Generasi in 1,605 villages and 5 provinces. 2008 PNPM Rural reaches 34,405 villages, 2,447 subdistricts, and 336 districts. 2008 PNPM Respek launches in Papua and West Papua. 2008 PNPM Green is established to adapt the PNPM Rural model to address environmental sustainability. 2009 PNPM Rural reaches 57,266 villages, 4,371 subdistricts, and 379 districts. 2009 106 high school graduates are trained as technical facilitators in the region of Papua in the second phase of Barefoot Engineers. 2010 PNPM Rural reaches 61,000 villages, 4,791 subdistricts, and 385 districts. 2011 PNPM Peduli pilot project is initiated and grant agreements between the PSF and three national civil society organizations are signed. 2012 PNPM Rural impact evaluation finds that from 2007 to 2010, the program raised household consumption by poor families by almost US$4 per person per month. References 1 ÒPoverty Headcount Ratio at National Poverty Line (% of Population),Ó World Bank Global Poverty Working Group. http://data.worldbank.org/indicator/SI.POV.NAHC. 2 Anthony Torrens, ÒEconomic Impact Analysis of Kecamatan Development Program Infrastructure Projects,Ó January 2005. 3 Karrie McLaughlin, Adam Satu, and Michael Hoppe, ÒKecamatan Development Program Qualitative Impact Evaluation,Ó World Bank, April 2007. http://psflibrary.org/catalog/repository/Kecamatan%20development%20program%20qualitative%20impact%20 evaluation.pdf. 4 John Voss, ÒImpact Evaluation of the Second Phase of the Kecamatan Development Program in Indonesia,Ó EASIS World Bank Indonesia, June 2008. http://www- wds.worldbank.org/external/default/WDSContentServer/WDSP/IB/2008/09/24/000334955_200809240519 44/Rendered/PDF/455900WP0P10701DP1Impact0Eval1Final.pdf. 5 Susan Wong, ÒIndonesia Kecamatan Development Program: Building a Monitoring and Evaluation System for a Large-Scale Community-Driven Development Program,Ó 2003, 3. 6 ÒAnnual Report PNPM Mandiri Perdesaan: Ministry of Home Affairs Republik of Indonesia,Ó 2009. http://www.pnpm-perdesaan.or.id/downloads/Annual%20Report%202009%20Eng.pdf. 7 Ibid. 8 John Voss, ÒImpact Evaluation.Ó 9Benjamin A. Olken. ÒCombating Corruption in Community Development in Indonesia,Ó Abdul Latif Jameel Poverty Action Lab. http://www.povertyactionlab.org/evaluation/combating-corruption-community- development-indonesia. See also Benjamin A. Olken, ÒMonitoring Corruption: Evidence from a Field Experiment in Indonesia,Ó Journal of Political Economy 115 (2007):200Ð248. 10 Benjamin A. Olken, ÒGenerasi Impact Evaluation,Ó 2011. 11 Ibid. 12 Ibid. 13 ÒPSF Progress Report 2011,Ó PNPM Support Facility (PSF), 2012. http://psflibrary.org/web/?q=catalog&recid=5929. 14 ÒMarginalized Groups in PNPM-Rural.Ó Bandung, Indonesia: Akatiga Center for Social Analysis, June 2010. http://psflibrary.org/catalog/repository/PNPM%20Marginalized_english.pdf. 15 ÒPSF Progress Report 2011,Ó PNPM Support Facility, 2012. http://psflibrary.org/web/?q=catalog&recid=5929. 16 The three organizations were Kemitraan, the Association for Community Empowerment, and Lakpesdam, part of the Nahdlatul Ulama. The first two were grant-making organizations with dozens of partnerships with local civil society organizations. Lakpesdam had a national network of branches to which it distributed funds. 17 ÒPSF Progress Report 2011.Ó PNPM Support Facility (PSF), 2012. http://psflibrary.org/web/?q=catalog&recid=5929. 18 ÒIndonesia: Dynamics of Violence in Papua,Ó International Crisis Group, 9 August 2012. http://www.crisisgroup.org/~/media/Files/asia/south-east-asia/indonesia/232-indonesia-dynamics-of- violence-in-papua.pdf. 19 ÒAnnual Report 2009 PNPM Mandiri Perdesaan: Ministry of Home Affairs Republik of Indonesia.Ó http://www.pnpm-perdesaan.or.id/downloads/Annual%20Report%202009%20Eng.pdf. 20 Ibid. 21 ÒPNPM Rural,Ó PNPM Support Facility (PSF). http://pnpm-support.org/pnpm-rural. 22 ÒPSF Progress Report 2011,Ó PNPM Support Facility (PSF), 2012. http://psflibrary.org/web/?q=catalog&recid=5929. 23 Ibid. 24 ÒPNPM Rural Impact Evaluation Report 2012.Ó Jakarta: 2012. http://psflibrary.org/web/?q=catalog&recid=6299. 23 25 John Voss, ÒPNPM-Rural Baseline Report.Ó EASIS, World Bank Indonesia, 2008. http://pnpm- support.org/data/pdf/pnpm-evaluations/2008/PNPMRuralBaselineReport.pdf. 26 ÒPNPM Rural Impact Evaluation Report 2012.Ó Jakarta: 2012. http://psflibrary.org/web/?q=catalog&recid=6299. 27 ÒThe Impact of PNPM-Rural 2010: A Qualitative Study.Ó SMERU Research Institute, December 2010. http://psflibrary.org/catalog/repository/Final%20PNPM%20Semeru%20-%20Edited.pdf. 28 John Voss, ÒImpact Evaluation.Ó 29 ÒThe Impact of PNPM-Rural 2010.Ó 30 Ibid. 31 Andrea Woodhouse, ÒGovernance Review of PNPM Rural,Ó January 2012. http://psflibrary.org/catalog/repository/Governance%20Review%20of%20PNPM%20Rural%202012.pdf. 32 ÒThe Impact of PNPM-Rural 2010.Ó 33 Benjamin A. Olken, ÒMonitoring Corruption: Evidence from a Field Experiment in Indonesia,Ó Journal of Political Economy 115 (2007):200Ð248. See also Benjamin A. Olken, Junko Onishi, and Susan Wong. ÒShould Aid Reward Performance? Evidence from a Field Experiment on Health and Education in Indonesia,Ó March 2013. http://economics.mit.edu/files/6923. Innovations for Successful Societies makes its case studies and other publications available to all at no cost, under the guidelines of the Terms of Use listed below. The ISS Web repository is intended to serve as an idea bank, enabling practitioners and scholars to evaluate the pros and cons of different reform strategies and weigh the effects of context. ISS welcomes readersÕ feedback, including suggestions of additional topics and questions to be considered, corrections, and how case studies are being used: iss@princeton.edu. Terms of Use Before using any materials downloaded from the Innovations for Successful Societies website, users must read and accept the terms on which we make these items available. The terms constitute a legal agreement between any person who seeks to use information available at www.princeton.edu/successfulsocieties and from Princeton University. In downloading or otherwise employing this information, users indicate that a. They understand that the materials downloaded from the website are protected under United States Copyright Law (Title 17, United States Code). b. They will use the material only for educational, scholarly, and other noncommercial purposes. c. They will not sell, transfer, assign, license, lease, or otherwise convey any portion of this information to any third party. (Republication or display on a third partyÕs website requires the express written permission of the Princeton University Innovations for Successful Societies program or the Princeton University Library.) d. In all publications, presentations or other communications that incorporate or otherwise rely on information from this archive, they will acknowledge that such information was obtained through the Innovations for Successful Societies website. A suggested citation format is as follows: [Document author if listed], [Document title], Innovations for Successful Societies, Princeton University, accessed at http://www.princeton.edu/successfulsocieties on [date accessed on Web] e. They understand that the quotes used in the case study reflect the intervieweesÕ personal points of view. Although all efforts have been made to ensure the accuracy of the information collected, Princeton University does not warrant the accuracy, completeness, timeliness, or other characteristics of any material available online. f. They acknowledge that the content and/or format of the archive and the site may be revised, updated or otherwise modified from time to time. g. They accept that access to and use of the archive are at their own risk. They shall not hold Princeton University liable for any loss or damages resulting from the use of information in the archive. Princeton University assumes no liability for any errors or omissions with respect to the functioning of the archive. Innovations for Successful Societies (ISS) is a joint program of Princeton UniversityÕs Woodrow Wilson School of Public & International Affairs and the Bobst Center for Peace & Justice. The Woodrow Wilson School prepares students for careers in public service and supports scholarly research on policy and governance. The mission of the Bobst Center for Peace & Justice is to advance the cause of peace and justice through mutual understanding and respect for all ethnic traditions and religious faiths, both within countries and across national borders. ISS is a joint program of the Woodrow Wilson School of Public and International Affairs and the Bobst Center for Peace and Justice: www.princeton.edu/successfulsocieties