Estimating the Benefit Incidence of an Antipoverty Program by Propensity Score Matching

Income gains from participation in economic programs are estimated as the difference between income with the program and that without it. The "with" data can be collected without much difficulty. But the "without" data are fundamentally unobserved, since an individual cannot be both a participant and a non-participant of the same program. It is common practice to estimate the unobserved income without the program as income with the program minus wages received. However, there are opportunity costs of participating in the program. Ignoring these foregone incomes of participation will result in over-estimation of the gains from the program. We apply recent advances in propensity-score matching methods (PSM) to the problem of estimating the distribution of net income gains from an Argentinean workfare program. PSM allows us to draw a statistical comparison group to workfare participants from a larger contemporaneous and comparable survey of non-participants. The average incomes of the comparison group are compared with the average incomes of the participants to assess the direct income gains from the program. The average gain is found to be about half the gross wage. Over half of the beneficiaries are in the poorest decile nationally, and 80% are in the poorest quintile. Our PSM estimator is reasonably robust to a number of changes in methodology, including a instrumental variables test for selection bias after matching.