81336 enGender Impact: The World Bank’s Gender Impact Evaluation Database Group Diversity and the Impacts on Female Participants: A Quasi-Experimental Study of Income Generating Networks in India Author(s) Gayatri Koolwal Contact gkoolwal@worldbank.org Country India Organizing Theme Voice and Agency, Economic Opportunities and Access to Assets Status The Project is ongoing but the evaluation is complete Intervention Category Community Development Sector Social Development Community networks, such as credit associations and income generating groups, help pool resources and information among the rural poor. Often, women are targeted through all- female, socially homogeneous groups to foster greater communication and confidence among members. Using a new data set collected by the author from a quasi-experimental survey of women participating in World Bank income generating networks in India, this paper finds that women in mixed-gender groups have actually achieved improved influence and participation in household economic decisions as compared to participants overall, targeted nonparticipants, and control respondents. These outcomes include greater individual savings relative to household savings, economic reach within the village in terms of increased net loans to other village members, and greater participation over a range of household decisions since program implantation. In particular, I show that these impacts Abstract occur through non-monetary mechanisms (greater potential for learning through exposure to a wider skills set, for example). Heterogeneity in sub-castes or communities within the group, however, has negative or insignificant impacts on different respondent outcomes. Endogeneity of program targeting is resolved by surveying exogenously nontargeted villages. To address potential endogeneity in the composition of participants’ networks, overall participation is instrumented by whether a program official randomly visited the respondent at the onset of the program. Share of men in the group is instrumented by the agricultural season in which the program entered the village, impacting men’s availability to participate. Finally, since women’s household behavior in this region is largely determined by their own caste, village size is small, and permanent migration is rare, the share of different sub-castes in the network is instrumented by the number of households in the respondent’s ten-household neighborhood that are from different sub-castes. Gender Connection Gender Focused Intervention Gender Outcomes Savings, access to credit, intrahousehold decision making IE Design Quasi-Experimental Design and Instrumental Variables The main component of the Development Poverty Initiative is the organization of “community interest groups” (CIGs). CIGs are networks of roughly 5-6 members, each Intervention member from a different family, who participate in a specific income generating activity such as raising livestock, growing new crops, or forming a small enterprise. Targeted Last updated: 14 August 2013 1 enGender Impact: The World Bank’s Gender Impact Evaluation Database individuals self-select into the program and a particular network, often with the guidance of a program official. For each CIG, DPIP allocates 20,000 rupees (about US $400) per member in a joint bank account from which they can withdraw for the project Intervention Period 2000- Present The author administered a random sample to 274 women across 240 households in rural Sample population Madhya Pradesh. For participant households, all women who were members of CIGs were interview, while for non-participant households, all women 18 and older were surveyed. There was a block of villages that were considered eligible for DPID, but due to resource Comparison conditions constraints, were randomly excluded from targeting at the first stage. Unit of analysis Individual Level Evaluation Period November 2003-February 2004 Participants with more men in GIC achieved significantly greater savings. Individual loans to other members also increased in the mixed gender group. The greatest changes in household decision are from mixed gender groups. There is no benefit of a mixed gender Results group to expanding one's network beyond their caste. As compared to the share of men in the group, however, participants in CIGs with a greater share of members from different communities experienced insignificant or, in the case of participation in household decisions, worse outcomes as compared to more socially homogeneous groups. Primary study limitations Funding Source Cornell University Graduate School Koolwal, G. (2007) "Group Diversity and the Impacts on Female Participants: A Qausi- Reference(s) Experimental Study of Income Generating Networks in India" Link to Studies http://www1.worldbank.org/prem/poverty/ie/dime_papers/421.pdf Microdata Last updated: 14 August 2013 2