Assessing the impact of public transfers on private risk sharing arrangements : evidence from a randomized experiment in Mexico

We adopt a structural approach to studying the effects of public transfers on consumption smoothing, risk sharing and welfare in small village economies. We calibrate the key parameters of a dynamic limited commitment model using data gathered as part of the Mexican Progresa program, and take advantage of the randomized experimental design of the data to validate the model using the treatment sample. The limited commitment model enriched to allow for unobserved heterogeneity in preferences can reasonably well explain consumption dynamics and cross-sectional distributions. The calibrated model correctly predicts the increase in consumption smoothing of transfers’ recipients, and the decrease in risk sharing between beneficiaries and non beneficiaries of the program. Progresa transfers are found to crowd-out between 3% and 10% of the pre-existing private transfers, but the overall direct effect of the subsidy on consumption is welfare improving for all households. Last, we use our structural model to evaluate a counterfactual, fully funded, insurance scheme.