Cooperative credit access and women's agricultural market participation in rural Tanzania: a fuzzy regression-discontinuity analysis of village savings-group eligibility thresholds

Village savings and loan associations (VSLAs) in rural Tanzania typically admit members through eligibility thresholds that are not uniformly enforced, creating a fuzzy boundary between households with and without cooperative credit access. The analysis draws on the DHS Program indicator 'Discontinuation due to access/availability' (FP_RDIS_W_ACC), comprising 74 country-survey observations across 28 African countries from 1992 to 2025, with a mean of 1.712 and a standard deviation of 1.37. The paper specifies the identification assumptions, the bandwidth selection rule, the local polynomial estimator and the compliance-weighting logic required for a credible fuzzy design at VSLA thresholds. It argues that the credibility of any such estimate depends less on the estimator than on three institutional features of Tanzanian VSLAs: the informality of eligibility rules, the endogeneity of savings behaviour to anticipated credit, and the gendered structure of market participation. The contribution is a design template and a set of falsification tests that future empirical work can apply to VSLA administrative data, not a completed causal estimate.