Group-participation and Women Empowerment: Matching as an Evaluation Estimator A District-level Study in West Bengal, India

The paper aims at isolating the effect of group-participation on women empowerment using primary data on 1500 individual women collected during 2007-08 from two districts of West Bengal, India, namely Hooghly and South 24-Parganas. Since the impact evaluation exercise typically suffers from the problem of counterfactual, in the absence of biologically identical observations proxy has been constructed in terms of pairing of statistically identical observations by applying matching techniques based on propensity-scores. It is observed that mere inclusion in a SHG is not sufficient for any woman to enjoy the benefits of better connectivity with the social capital. Here both the duration of membership and the quality of participation matter. Moreover, in terms of a comparison of the probability of inclusion in the program with the proportion of actual inclusion for subjects with same p-scores, the extent of program mismatch has been assessed. This indicates a bias from the supply side where the more likely agents are being included in the absence of special effort to reach out the relatively more vulnerable ones.