آیا پروژه های توسعه جامعه محور ، سرمایه اجتماعی را بهبود می بخشند ؟ مدارک و شواهد از فیلیپین

We explore the social capital impacts of a community-driven development project in the Philippines in which communities competed for block grants for infrastructure investment. The analysis uses a unique panel data set of about 2100 households, aggregated at the village-level, collected in 66 treatment and 69 comparison communities. We provide both difference-in-differences and propensity score matching estimates. We find that the project increased participation in village assemblies and the frequency with which local officials meet with residents and had a negative impact on collective action. There is also more limited evidence of a positive impact of the project on bridging (i.e., generalized) trust and of a negative impact on group membership.

International aid organizations, multilateral organizations and national governments are increasingly favoring bottom-up approaches that involve local communities in project design and implementation. In addition to purely normative benefits (Sen, 1999), participation is expected to lead to better outcomes through better poverty targeting, reduced project costs, improved maintenance and allocative efficiency. Moreover, often grounded in social theory (Chase and Woolcock, 2005), such programs are expected to enhance social capital. Community-Driven Development (CDD) projects are one way the World Bank has applied conventional wisdom about the value of participation. Such projects constitute 9% of World Bank lending annually (World Bank, 2007).2 In a typical CDD project, communities prepare subproject proposals with the support of facilitators, compete over block grants to finance investments for local public goods, and are then responsible for implementation and maintenance of those investments. The CDD approach promotes stakeholder involvement in all stages of the process from subproject identification to subproject maintenance. While the literature on some of the expected benefits associated with direct stakeholder participation in development projects is growing (see Arcand & Bassole, 2007, Prokopy, 2005, Chase et al., 2006, Khwaja, 2006 and Olken, 2007 among others), there is little empirical understanding of its institutional impacts. However, since those projects often have the objective of empowering communities, it seems important to address questions like: as a result of a CDD project, does participation in local governance activities actually increase? Do group membership rates increase? How do the relationships between local officials and citizens evolve? This paper explores the impacts on community-level social capital of the KALAHI-CIDSS, a large-scale CDD project in the Philippines. Under this project, all villages within a municipality can prepare proposals for small-scale infrastructure investment, while only a subset of villages (i.e., those with the most deserving proposals) actually implement those investments. The paper takes advantage of a unique and detailed panel dataset of 2092 households, aggregated at the village-level, in 135 villages in 16 municipalities of the Philippines. The sample includes eight municipalities in which the project is implemented and eight matched comparison municipalities.3 In this paper, we define social capital as ‘the ease with which community members act collectively’. While it is possible to analyze social capital at the household level, given the way CDD projects are implemented, we treat social capital as a community aggregate. Our indicators of social capital will include measures of participation in both formal and informal institutions as well as of the ties between citizens and the local and national governments. Our paper builds on existing research on participatory projects' impact to enhance social capital and foster collective action. First, we use household panel data, aggregated to the village-level, collected both before the project started and after one cycle of subproject implementation in both treatment and comparison municipalities. The survey instrument includes information on a broad range of household and village-level social capital measures. Thus, we can analyze whether changes in social capital indicators can be attributed to the project, controlling for initial village characteristics and time-constant unobserved characteristics. Further, the availability of household-level panel data allows us to check if our results could be driven by attrition bias. Second, communities followed very detailed, facilitated participatory processes throughout the project, tailored to Philippine conditions. Since opportunities for participation are clearly defined, we improve upon previous research that used data from CDD projects in which the decision rules within communities and with local authorities were not always clear. Note that while the subproject implementation cycle is repeated three times in each village, this research summarizes data currently available, collected after the first cycle. More data will be collected after the end of the third cycle which will enable us to analyze the impacts of the full projects (i.e., of the three cycles of subproject implementation). Our difference-in-differences and propensity score matching estimates highlight that the project impacts both participation in local governance activities and community social networks. As a result of the project, treatment villages experienced increases in participation in village assemblies and in the number of meetings with residents organized by elected village leaders. There is also limited evidence that generalized trust (i.e., toward strangers) also increases with the project. However, there is a decline in collective action and, to a lesser extent, group membership. This might indicate that different forms of social capital are substitutes but we can't test this formally with the data available. The paper is organized as follows. Section 2 defines social capital and discusses how CDD projects interact with the concept of social capital. Section 3 describes the CDD operation and data. The estimation strategy is presented in Section 4 while theresults are discussed in Section 5Section 6 concludes.

In this paper, we build on and expand existing research on the impacts of community driven development approaches. First, using a unique household panel dataset collected before the project started and after one cycle of subproject implementation in both treatment and matched control communities, we assess whether we can attribute to the project changes in social capital indicators. Second, we refine previous research by having data on a large number of social capital measures. Further, we provide both difference-in-differences and propensity score estimates. As a result, we are able to provide a more accurate picture of how CDD projects operate within Philippine villages. Overall, our results indicate the CDD operation led to changes in village-level social and institutional dynamics. Focusing on the outcomes for which estimates are statistically significant with both methods, we find that the project increased participation in village assemblies and the frequency with which local officials meet with residents. Further, the project had a negative impact on collective action. There is also more limited evidence of a positive impact of the project on bridging (i. e., generalized) trust and of a negative impact on group membership. Our results are only an initial step in understanding how CDD projects work with and build on pre-existing social dynamics in the communities in which they operate. We see at least three areas where further research could generate useful knowledge. First, efforts should be devoted to understanding the household-level impacts of participation in project activities. Given the potentially strong endogeneity of this decision, this would likely require an experimental design. For example, if in a particular setting sending invitations to households increases their likelihood of participation, randomly sending invitations could provide a credible instrument. Second, development practitioners should devote more effort to understanding the role facilitators play in ensuring broad representation within communities and significant benefits from the project. The outstanding operational question is whether and how increased facilitation can ensure that a broader proportion of community members participate than would be the case without such facilitation. Third, practitioners need to assess whether the results discussed in this paper hold or change after additional subproject implementation cycles. This could provide useful information to policymakers when deciding how many cycles should be funded. Another round of data collection is expected to take place after the third cycle of subproject implementation which will allow us to answer those questions. Fourth, it would be interesting to assess if, in light of the benefits associated with the project, municipalities adopt the CDD approach in allocating some of their resources. There is anecdotal evidence that some do but analyzing why could prove extremely fruitful.