Negotiating Impossibilities in Community-driven Development in Indonesia

PNPM

Community driven development has been an important part of the development landscape in Indonesia. In 1998, the World Bank Social Development unit in Jakarta designed and led the Kecamatan Development Program (KDP) with the objective to support development planned and implemented by rural communities (Guggenheim, 2004). Following its success, CDD programs followed in urban areas, most notably the Program Penangggulaangan Kemiskinan di Perkotaan (Urban Poverty Alleviation Program), or P2KP. These programs were based on the premise that the people know best how to solve their problems, at the same time that communities were undergoing a period of moral crisis (Effendy, 2015; Li, 2006). In order to overcome urban poverty, “communities” need to be morally strengthened, given the resources and capabilities to undertake their own development. The PNPM-Urban was a continuation of these programs, with the aims of “fostering community participation, improving local governance, and delivering basic needs at the community level” (Baker et al., 2013, p. vi).

An important feature of the PNPM-Urban is the establishment of a board of volunteers to implement the program at the local level. There is one BKM (Board for Community Self-reliance) in each Kelurahan: an urban administrative unit with a population between 10,000 and 30,000 residents. The BKM are responsible for developing a Community Development Plan and implementing projects in three areas: infrastructure (building of roads and drains etc.); economic (rotating funds and small enterprise development); and social (small gifts of welfare and training). Each BKM has between 7 and 12 members, roughly one per lingkungan (the lowest level of urban administration below the kelurahan). An election is held to select the members and a coordinator. The government continues to offer support and guide the BKM members through PNPM facilitators, and until recently, provided funding for activities in each of the three aforementioned areas. The BKM work with the Kelompok Swadaya Masyarakat (community self-help group) who develop proposals for their lingkungan to be funded by the PNPM.

The BKM members in Medan are socio-economically and ethnically diverse. In Medan, they include a former Lurah (head of the kelurahan), retired civil servants, wives of the kepala lingkungan (head of the lingkungan), and retired private sector workers. The vast majority are, however, from modest backgrounds. Some had small businesses such as a roadside warung (restaurant), while others worked in low-paying jobs such as carpark attendants or becak (similar to rickshaw) drivers. Many BKM members and a couple of coordinators were housewives. The BKM membership was diversified in terms of age and gender. In the ethnically diverse city of Medan, BKM members were Javanese, Batak-Toba, Melayu, and Mandaling among others, with seemingly no concentration of a single ethnic group in any one BKM. We did not, however, encounter, a single BKM member who identified as Chinese, (although a few had mixed heritage). ² During the research project, we interacted with BKM members from several different BKMs across the city of Medan, concentrating on four BKMs nominated among the best in Medan. The BKM members can be described as CBWs as follows:

voluntary or low-paid villagers encouraged to shape and assume tasks in service delivery…who live in the community that they serve, are selected by and are accountable to that community, receive a short, defined training, and may or may not be attached to a formal institution. (Boesten et al., 2011, p. 44)

I use the term volunteer to emphasize that they are not paid, and as the term they use to describe themselves. CBVs are critical to CDD, as both the operations and outcomes of the programs depend upon their agency (Boesten et al., 2011). The World Bank similarly note the important role of “community volunteers” in the PNPM program, in “initial planning, actual administration, implementation, and maintenance of projects” alongside “ensuring that knowledge about the program and information about its activities is widely disseminated to the public” (Baker et al., 2013, p. 10). With this role comes significant responsibility, which as I argue below, makes them a suitable target on which to devolve failure.

Individuals had diverse reasons for becoming CBVs. Self-representations of motivations most commonly included the following: the desire to become a more moral person or a better Muslim; wanting to give energy as a form of charity for volunteers too poor to give money; building social connections and gaining public recognition in the locality; and to realize what they considered a natural disposition toward social work (Yumasdaleni & Jakimow, 2017). Many researchers are cynical of such claims to be motivated by helping others, and self-representations do only tell partial truths (Butler, 2005). To disregard them completely, however, overlooks the value that BKM members ascribe to these opportunities. That is, BKM members do benefit, and their participation is, therefore, arguably in pursuit of self-interest, but the nature of this benefit is not found in the pursuit of political and economic gain. Rather CBVs are (also) engaged in and actively seeking benefits such as emotional satisfaction and self-actualization. These motivations need to be taken seriously in order to understand how and why CDD does, or does not work in practice and the consequences of this failure for BKM members.

Negotiating Impossibilities

Part of a consequence of being motivated to do good (or for the cynical among us, be perceived as doing good) is that BKM members are for the most part genuine in their desire to achieve the normative expectations of CDD. They have mostly subscribed to the need to include as many people as possible in decision-making processes, to target the needy, to be transparent and accountable, and achieve value for money. At the same time, they are acutely aware of the scrutiny over the PNPM, and by extension themselves, in which researchers, government officials, the media, and general public look for signs of failure. I examine the efforts to meet these normative expectations, and how they result in what I describe as negotiated impossibilities: that is, the practices arising as a compromise for expectations impossible to achieve. By revealing the consequences of scrutiny, I argue for evaluations that are both realistic in their expectations, and which take the accounts of BKM members as seriously as those of their detractors.

A central objective of CDD programs is to empower the poor and marginal by putting them in the driver’s seat of development. Critical to this objective in the PNPM is “community participation in planning, decision making activities, and project implementation activities” (Baker et al., 2013, p. 10). Not only does this provide groups with a voice, participation has additional benefits including “better designed projects, better targeted benefits, more cost-effective and timely delivery of project inputs, and more equitably distributed project benefits” (Mansuri & Rao, 2004, p. 6). The PNPM offers people with two opportunities to drive community development: first, become BKM members and second, attend community meetings including voting in BKM elections. The latter is intended to act as a counter-balance to the power invested in BKMs, as community consultation and democratic procedures reduce the risk that boards will dominate decision-making (Fritzen, 2007). To achieve “empowerment,” poor and marginalized individuals should participate in both these respects.

The failure to achieve these normative expectations is evidenced in studies that highlight the extent of elite capture in CDD. The evidence that local social hierarchies influence who can and does participate has become so overwhelming that Lund and Saito-Jensen (2013) suggest it is now assumed the elite capture is almost an unavoidable element of participation. Even in relatively egalitarian societies, elites dominate decisions and the processes of managing a public good (Mansuri & Rao, 2004). The “inevitability” of elites taking over the leadership of CDD has led some researchers to investigate whether elite control translates into elite capture (Dasgupta & Beard, 2007). Several studies demonstrate that control over decision-making does not necessarily divert resources away from the groups most in need, and furthermore, the ability of marginal groups to challenge elite domination of decision-making may change over time resulting in “empowerment” (Fritzen, 2007; Labonne & Chase, 2009; Lund & Saito-Jensen, 2009).

This literature leads to two questions for the PNPM. Do elites dominate the BKM boards and community meetings? And if so, does this result in the implementation of projects that disproportionately benefit those closest or politically useful to decision-makers. Studies suggest the evidence is mixed. McCarthy et al. (2014) find that participation in the PNPM-rural was uneven, with marginal groups most likely to be excluded due to transaction costs. When they did participate, their projects were least likely to gain approval, and there was a tendency for elites to take over leadership. Triwibowo (n.d., p. 16) found that even when there was widespread community participation, “Poor communities and women are only consigned to the role of passive participants, while the decision-making process remains under the control of the elite.” Triwibowo (n.d., p. 16) is adamant that within the PNPM Mandiri, community engagement is merely “procedural and mechanical in nature, and [as such] incapable of empowering poor communities.” Both studies found, however, that project benefits do reach the poor, albeit often distributed according to political expediency.

The elite capture of BKMs that is seemingly a feature of PNPM-rural, and in other CDD programs in urban areas (Marcus & Asmorawati, 2006), was not observed in Medan. Here, as noted above, members and coordinators are from modest socio-economic backgrounds. The relatively smaller budgets (compared to villages), the lack of support from local government officials, and the preoccupation in profit making activities perhaps dampening the appeal of elite involvement in urban CDD. People in financial hardship are, however, rarely BKM members due to the significant transaction costs (returned to below). Some BKM coordinators were tokoh (leaders) in the local community prior to being elected, while others became respected persons as a consequence of their new position. To classify the latter as entailing elite capture (Marcus & Asmorawati, 2006) seemingly undermines the intent to empower marginal individuals to become leaders. Enhanced social status does enable a degree of social mobility for BKM members. Some members, who previously thought of themselves as suffering hardship, have re-evaluated this assessment after interacting with the very poor, and now use language that suggests they are descending to the grassroots in their activities. Social hierarchies are therefore not irrelevant in the PNPM Medan, but the BKMs cannot be considered to be captured by elites. A bigger problem is finding sufficient people to become members and remain active.

People in cities are busy, preoccupied with daily livelihood activities (Marcus & Asmorawati, 2006). These same conditions also make it a challenge to convince people to attend community meetings. Nationally, the PNPM-urban has suffered poor participation rates and low attendance at meetings due to the time-costs involved (Baker et al., 2013). An overlooked challenge is the sheer scale of “community” at this level, with each Kelurahan in Medan having a population of between 10,000 and 20,000 adults. This creates challenges in informing people of the program (sensitization), and inviting them to specific meetings. Further, poor people’s networks are widely dispersed in cities, and there is therefore no pre-existing unit for social capital formation (Marcus & Asmorawati, 2006). Although units of social cooperation are found in urban areas across Indonesia (Guinness, 2009), these are usually at levels below the kelurahan. The design of CDD presumes a pre-existing “community” that can be restored for collective action: a task in this case assigned to BKM members. The literature and PNPM evaluations are silent to the impossibility of this task in cities such as Medan.

The second question is whether this patchy (at best) participation in community meetings results in projects being implemented that benefit the most needy. The allocation of PNPM funds available through the BKM entails individuals or groups such as the KSM submitting proposals (e.g., the building of a drain or a training program) that are then discussed at the aforementioned community meetings, prioritized according to “community” needs, and then selected. As the budget is limited, not all projects can be funded. Accusations circulate in Medan that BKM funds are prioritized in neighborhoods closest to BKM members, and in ways that benefit their family and friends. The difficulty of encouraging people to attend community meetings would partially explain a funding pattern concentrated in this way. The time cost of encouraging people to attend meetings is high, and BKM members often fall back on those closest to them (geographically and socially) to come. This has implications for the proposals that are discussed and accepted.

The accusations themselves must also, however, come under scrutiny. Coordinators complained that “clever people” will protest decisions to manipulate future resource allocations, also provoking other people to become cynical of the BKM. Other people fail to attend meetings, but still complain when their proposals are not funded. Accusing “developers” of favoritism, especially if done publically, can be a strategy to ensure resources flow their way next time. Complaints about the allocation and implementation of PNPM projects are also (mis)directed to BKM members. The expectations of the people are often higher than the funds available. For projects that are funded, the government is often slow to release funds, resulting in accusations that BKM members have pocketed the money. Infrastructure projects (paved roads or drains) may be budgeted for a particular distance, resulting in complaints when they fail to reach a locality not included in the original proposal. As BKM members are enmeshed in local relations (or are “heavy” to use Redfield’s (2012) terminology), they are more likely to bear the brunt of complaints and accusations, even when the cause of the problem is out of their control.

The emotional toll of these accusations and complaints is higher due to the degree of truth that underpins them. For the reasons outlined above, the dispersal of PNPM funds frequently fails to live up to the normative expectations of CDD, though not for reasons of the self-interest of BKM members. While accountability procedures, horizontal monitoring, and democratic processes can encourage responsibility of community leaders (Dasgupta & Beard, 2007), or increase the costs of non-compliance to program norms, these technical solutions will not help overcome the difficulty of mobilizing a large, non-existent community and satisfying their expectations with limited (and unreliable) funds. Indeed, the motivations of BKM members, many tied to ethical self-cultivation, creates more than enough incentive to allocate project funds as fairly as possible with the most impact. These same personal investments increase the weight of accusations and the emotional costs of failure.

In contrast to the distribution of projects that prioritize the needs of the “poor,” targeting refers to the identification of needy people ³ to receive specific development goods, such as gifts of welfare or loans. A normative expectation is that CDD achieves more precise targeting (Baker et al., 2013; Mansuri & Rao, 2004). As CBVs are from the locality, they can identify the people most in need and “worthy” of benefits, with such community involvement in selection increasing transparency and perceived fairness of the programs (Robertson et al., 2014). This expectation is not always achieved, however, and evidence shows that the poor often miss out on benefits due to social hierarchies, the nature of leadership, and information gaps (Mansuri & Rao, 2007; Robertson et al., 2014). While CDD has achieved good geographical targeting—reaching the poorest areas—it has proven less effective at targeting the very poor within diverse localities (Mansuri & Rao, 2007; Marcus & Asmorawati, 2006). Despite experimental evidence to the contrary (Schüring, 2014), the criticism persists that local people distribute program benefits to family members, friends, or for political gain.

Precise targeting, or tepat sasaran, is a key objective and basis for evaluation of the PNPM. To achieve “poverty targeting” (Baker et al., 2013, p. 10), BKMs conduct “self-mapping,” gathering information from each household to compose lists of potential recipients of benefits according to their own set of criteria. The BKM members claim to be more accurate than other government programs, as they go to the people directly, rather than relying on lists composed by the kepling (head of lingkungan) who may direct resources to family, friends, and supporters. Further, they not only have intimate knowledge of who is genuinely poor, but also what is required to eke out a minimum standard of living in the city: knowledge often gained through their own experience of poverty. For example, one BKM member stated that government programs exclude people who have a motorcycle, but the BKM realize that such transport is essential for many in order to make a livelihood. While we can be skeptical of self-representations that suggest benefits are always targeted appropriately, it is worth noting that selecting the correct people to receive charity is a duty under Islam (Benthall, 2003); to not do so, even unintentionally, is a sin. Several BKM members said they became anxious that they had failed to reach the most appropriate people due to this reason.

BKMs in Medan have nonetheless been criticized for incorrect targeting. In one example, academics from a local university presented an unpublished evaluation of the PNPM in Medan at a workshop for PNPM facilitators. Deploying an approach described as “net balance method,” they captured whether people had a positive or negative opinion about the PNPM according to certain criteria, including targeting. They concluded that the identification of beneficiaries was “subjective.” Recipients of micro-finance loans were, they claimed, geographically concentrated around the homes of BKM members, and included friends and family members. The micro-finance component of the PNPM in particular had failed to reach the very poor (McCarthy et al., 2014). Several facilitators present were incensed at the findings. They complained that the researchers had spent barely 10 minutes with each BKM, used aggressive tones to demand access to documents thereby creating fear, and made generalizations about the whole program from one example. One facilitator noted that many BKM members fulfill the criteria of “needy,” and therefore family members are legitimate recipients.

Facilitators have a stake in presenting program success (Mosse, 2005), and their accounts are therefore not disinterested. Our observations suggest, however, caution in assessing the validity of the evaluation. Many people we encountered made similar accusations about the BKM. On further prompting, however, the majority of these people had very little knowledge of the PNPM, and confused it with other programs—for example complaining that they did not receive rice through a government program (Raskin) over which the BKM has no jurisdiction. An atmosphere of cynicism and suspicion surrounds development activities in Medan, especially those funded by the state. The opinions of “random” citizens are therefore as likely to be based on rumor and suspicion as grounded in first-hand experience. The evaluation presented in Medan raises the question as to why accusations and criticisms by members of the public are most often taken at face value, while the accounts of CBVs are treated with skepticism, if not outright incredulousness. This is often the case even though BKMs have substantial documentation that supports their version of events.

That is not to say that there is no truth behind these criticisms. BKM projects do not always reach appropriate people, but this can only partially, if at all, be explained by favoritism. Instead, the difficulties of selecting and encouraging people to take part in projects, even as recipients, are grossly under-estimated. The first difficulty is the aforementioned issue of scale, and the impossibility of knowing all needy residents within the Kelurahan. Receiving so-called “benefits” also entails a costs in terms of time and money. The need to verify BKM activities requires ceremonies to mark activities such as the distribution of school uniforms to needy children, or sembako (essential daily goods) to widows. The poor must act as “grateful recipients” at such events, often waiting hours for government officials to turn up, while bearing the embarrassment of being identified as someone in “need.” Further, as many projects are designed on the basis of self-help, they require financial investment that poor people find difficult to make. For example, a housing scheme required recipients to take a loan to cover half the cost of construction. The BKM had trouble in identifying suitable recipients, as no eligible (that is poor) person wanted, or could take the risk of such a large loan. In such circumstances, it is often not a case of selecting the most appropriate recipients from a willing pool of people, but rather of convincing people to take part in the program.

An example of a “failed” BKM project highlights these difficulties, and the costs of an emphasis on precise targeting. The BKM members were tasked with finding potential participants for a 1-week training program for women entrepreneurs organized and funded by an NGO. The “beneficiaries” had to meet two criteria set by the NGO: (a) they earned less than the regional minimum wage, at the time about USD 170 a month; and (b) had their own business. Finding potential participants who met both criteria was, according to BKM members, impossible. The official minimum wage is insufficient to cover basic needs in Medan, let alone enable one to invest in one’s own business; the two criteria were therefore mutually opposed. Even after relaxing the wage requirement, they struggled to find entrepreneurs willing to invest a week of their time. With some compromises made on the criteria, and after considerable time locating and convincing women to take, the BKM and a worker from the NGO recruited approximately 30 participants. The labor involved in locating and convincing potential participants was not recognized by the NGO.

Nor was the relaxation in criteria acceptable. An NGO staff member lambasted the BKM coordinator, and censured the NGO staff member involved in recruitment, even though both had expressed the need to relax the criteria prior to the start of the training. They cancelled the remaining activities of the project, disappointing not only the BKM members but also the participants, many of whom had invested a week of their time in order to receive the later benefits of the project. ⁴ Ensuring that programs are fully subscribed requires compromises on criteria, and/or necessitates that CBVs draw on personal networks to encourage participation. Evaluations that characterize this as mis-targeting and nepotism obscure the difficulties involved in recruiting “beneficiaries,” while obscuring the costs imposed on CBVs to fulfill this task. The BKM members spend considerable time recruiting participants, while also acquiring future obligations to people who have given their time in order to make the program a success (Jakimow, 2013).

Transparency and downward accountability are two of the most important principles of CDD (Baker et al., 2013). Through an operational commitment to openness and democratic procedures, there is a normative expectation that CDD can reduce corruption. The PNPM-urban has several mechanisms to ensure “good governance”: “(i) randomized audits; (ii) severe penalties for misuse of funds; (iii) rewards for well performing BKMs and KSMs; (iv) an accountability matrix for all operations; and (v) a complaints handling service” (Baker et al., 2013, p. 14). These multiple layers of accountability have seemingly been successful in minimizing corruption across the PNPM and its predecessors in both rural and urban areas (Guggenheim, 2004; McCarthy et al., 2014). Suspicions remain, however, both among the general public (as outlined below), and among scholars. Hadiz (2010) highlights the dearth of evidence to demonstrate that decentralized governance reduces corruption, and instead argues that it has merely devolved opportunities to a greater number of local-level actors. CDD, based on many of these same principles of good governance as decentralized governance, is susceptible to the same misuse of government resources for personal enrichment.

The majority of BKM members stated that corruption was impossible due to the vigorous procedures and as evidenced by the people who quickly left the BKM once the inability to make money became apparent. Some claimed that the PNPM was designed to bypass the level “corrupt” local government, maintaining that money through the BKM reached the community in full. Only two BKM members admitted that corruption was possible in the PNPM, without, of course, admitting to such practices themselves. Conflicting accounts and the unlikelihood that BKM members would confess to corrupt practices means that I can only speculate as to their prevalence in the PNPM in Medan. My purpose in highlighting this issue is rather to examine how BKM members experience the criticisms that the PNPM is corrupt, while also trying to maintain (and take advantage of) accountability and transparency procedures.

Despite BKM members’ claims that the PNPM is designed to bypass the local state machinery, operationally they must still engage with officials in the Kelurahan and Kecamatan (sub-district) offices. Government officials must sign project budgets, bring prestige to ceremonies to mark project activities, and verify completed projects. The linkages between BKMs and these officials vary across Medan, and indeed Indonesia (Baker et al., 2013). While some BKMs have productive relationships, others are strained due to the struggles over power and capital (Asmorowati, 2012). Some BKM members claim that Lurahs see them as competitors for development money that would otherwise be channeled through them. These Lurahs attempted to frustrate BKM activities by demanding payments for signatures, not attending BKM events, and openly criticizing the BKM in public. It was common for PNPM facilitators and city-level officials to express that Lurahs were the number one problem for the PNPM in Medan. The BKM members have to deal with these tensions, are held accountable if projects fail, and must balance accounts when illegitimate fees are unavoidable.

Corruption does not have to take place to affect BKM members, as the very idea of it impinges on their everyday work. Anti-corruption discourses circulating in the Indonesian public sphere increase vigilance of development activities, at the same time that a morality of bagi-bagi (each taking their share) means that people involved in distributing money are often assumed guilty until proven otherwise. The media is increasingly playing a negative role in perpetuating suspicions. While originally important to enhance transparency and accountability by exposing misappropriation of funds in CDD (Guggenheim, 2004), today in Medan, several journalists concoct false stories and demand payment from victims to keep them quiet, ⁵ and several BKM coordinators had fallen victim to these demands. None we spoke to had given in to them, but rather used the documentation produced through the aforementioned mechanisms to prove their innocence. These incidents are deeply disquieting, however, and may cause social embarrassment if false accusations are published.

These accusations are part of what makes CDD “risky” for BKM members, and many coordinators in particular said that they suffered from anxiety. False accusations are less stress-inducing, however, then the possibility of inadvertently doing wrong. Few BKM coordinators are able to manage their books without assistance from PNPM facilitators, resulting in nagging doubts as to whether they may have been involved in wrong doing or a careless mistake without their knowledge. The fear was often palpable when we met BKM coordinators for the first time. Many assumed we had come to look over their books and were armed with various forms of evidence to convince us of their sincerity in running the program. It is of course critical to the credibility of the PNPM, and to the value of CDD in general that corrupt practices be identified, and reduced if not eradicated. At the same time, had we spent our time only pouring over the documentation, we would not have had the opportunity to speak to BKM coordinators about their experiences with the program, and the challenges that they face. Too frequently we witnessed evaluations, or read reports, where the preoccupation with transparency and accountability detracted from a more holistic understanding of how the program worked.

A final (and controversial) normative expectation is that CDD is a cost-effective way of delivering services to under-served populations (Boesten et al., 2011; Russell & Vidler, 2000; Triwibowo, n.d.). Rao (2005) argues that community participation explains the relative success of New Order Indonesia in improving health and education outcomes compared to India despite similar post-independence economic conditions. The mobilization of community through gotong royong, a (some would say invented) tradition of mutual assistance and volunteering for communal projects (Bowen, 1986), and the deployment of low-level state cadres (CBVs), enabled the delivery of much-needed services at a low cost (though Rao admits to the sometimes coercive nature of this volunteer labor). The PNPM, while rhetorically breaking with the New Order style of top-down development, mobilize these same institutions. Part of the rationale for community involvement through the PNPM-urban is that they can deliver infrastructure that responds to the community at a lower investment cost (Baker et al., 2013). While the evidence of whether this is achieved is inconclusive, the World Bank’s evaluation found that maintenance was better for PNPM infrastructure projects, and primarily undertaken by volunteer labor (Baker et al., 2013).

The controversy arises as some question whether the primary purpose of CDD is reducing the financial burden of the state, rather than its more loudly declared intention to empower local communities (Tribowo n.d.). Boesten et al. (2011, p. 42) argue that strategies to use CBWs “reside within a contemporary neo-liberal logic which seeks to keep state spending on social services low through decentralisation and privatisation” (Lacey & Ilcan, 2006). While the objectives of CDD are cost effectiveness and delivering services democratically, often the result is that tasks are devolved down to “insufficiently trained and compensated, community members” (Boesten, 2005, p. 13). Russell and Vidler (2000) found that the Sri Lankan government and NGOs alike used community development councils (similar to BKMs) in an extractive way to reduce the costs of implementation, while maintaining an asymmetric relationship with CBVs.

Our observations suggest that BKMs do help to reduce the costs of development. Several BKM members boasted that they are better able to mobilize volunteer labor (gotong royong) than government officials, and therefore can reduce the costs of infrastructure projects. A more significant saving, however, is ensuring high repayment rates of loans. The BKM members expand considerable effort collecting the small weekly or monthly payments of PNPM loans. This assists not only the government in maintaining a viable rotating savings and loans scheme, but also enables partner organizations including commercial banks and NGOs to recover debt. According to the NGO manager of the above-mentioned housing scheme, a primary reason for partnering with BKMs was their ability to select trustworthy recipients, and ensure they make regular repayments on the loan component. As BKMs undertake the time-consuming tasks of recruiting participants, sensitizing (explaining) programs to them, and undertaking implementation tasks, NGOs can reduce other costs, particularly salaries (Jakimow & Harahap, 2016).

The reduction of costs for the government and NGOs comes, however, at significant costs for BKM members. The opportunity cost, in terms of time lost to other activities including livelihood activities, is acknowledged in the program design and evaluations (e.g., Baker et al., 2013). Less recognized is the costs in terms of hard cash. One coordinator noted: “BKM members are very busy… as volunteers we must go here and there, and we must pay for our petrol, phone, we need to have food at meetings.” The small amount given to BKMs to cover “operational expenses” is insufficient, meaning that BKM members, and particularly coordinators, pay out of their own pocket. As many volunteers are from low socio-economic classes, the costs of participation are a significant burden, and many are now calling for a small honorarium. Russell and Vidler (2000) observed similar complaints and a call for financial compensation for CBWs in Sri Lanka, while Boesten et al. (2011, p. 54) argue that while CBVs may be motivated, “material reward through stipends, sitting fees, and access to resources is also important.” More important than maintaining BKM members’ involvement and motivation, I suggest that fairness dictates that the drive for cost effectiveness does not merely devolve costs to people less able to bear them.

Costs are not only financial, but also emotional. Many of these are mentioned above, and include the embarrassment and stress when confronted with complaints when projects do not run smoothly or are not directed to particular people, and the tension and anxiety when accused of corruption. Recovering loans is an additional activity that has a high emotional toll. In tears, a BKM member told us of the sadness she feels when she has to ask people to repay loans when she knows they are in difficult circumstances. The same BKM member has also been publicly humiliated, called a money lender in the streets. She, like the majority of BKM members, remains motivated and speaks more frequently of the satisfaction and public recognition that they receive from their work than of the emotional burdens. At the same time, anxiety is a recurring theme, and instances of complaints at times outweigh the instances of gratitude. The sentiments of a community volunteer in Sri Lanka echoes the concerns of BKM members: “it can be a thankless and dangerous job, often not worth the effort” (Russell & Vidler, 2000, p. 80).

Conclusion and Discussion

The potential benefits of CDD approaches to urban development make ongoing reflection and focused evaluations worthwhile. It is unfortunately rare, however, that the perspectives of people on the front line of CDD—the CBWs or volunteers—are taken seriously. Often CBVs are assumed to be benefiting materially and/or politically from their work: an assumption that becomes the starting point to interpreting failed program objectives. While I agree that self-representations of CBVs need to be read with a degree of caution, to disregard them completely results in an incomplete understanding of the factors shaping CDD practices and outcomes. As CBVs have a stake in the social production of success (Mosse, 2005), revealing the challenges they face and the compromises they make requires ethnographic observation of everyday practices, and most critical in an atmosphere of suspicion, building relationships of trust. This article contributes to understanding how CDD works in urban contexts through the insights afforded by such an approach.

The case study of the PNPM in Medan reveals that normative expectations of CDD of inclusive participation and the precise targeting of benefits are not always possible in their ideal form. The BKM members necessarily compromise on these ambitions in order to achieve results. Rather than applaud volunteers for negotiating the impossibilities of CDD to arrive at beneficial outcomes, evaluators, development partners and the general public frequently take them to task for their failure to achieve the impossible. The responsibility for failure is thereby devolved to CBVs, obscuring the need to reflect on the design principles of CDD, and the necessity to compromise on some of its more ambitious aspects in large cities. Rather than CDD programs designed in rural areas being replicated in urban ones, there is a need to design models suitable for each context (Marcus & Asmorawati, 2006). I suggest that the ways CBVs (and other implementers) negotiate the mechanisms of current programs are important indicators of what such models might look like. This is not to advocate a one-size-fits-all model for urban CDD, but to incorporate the knowledge of CBVs in the ongoing re-design on programs for each unique context.

In November 2016, 1 year after the PNPM-urban ended, I ran into a couple of BKM members on their way from the mayor’s office. Contrary to expectations (at least mine), despite the end of government funding, the BKM was going strong, partnering with city-level government, NGOs and others to conduct development activities in their kelurahan. While the lack of central funding has made their work more challenging, they remained committed to their work. Development is a mix of failures and successes, and poorly resourced volunteers are perhaps more prone to mistakes. But they are also an inspiration and have achieved real positive outcomes in their locality. They are a reminder that efforts to reflect on and improve CDD are worthwhile.