TD NEREUS 09-2020
Gabriel Lyrio de Oliveira and André Luis Squarize Chagas
In this paper, we present some long-term effects of the largest Conditional Cash Transfers program in the world, and one of the pioneers, the Bolsa Fam´ılia Program (BFP). We focus on the effects on Schooling attained in early adulthood and Labor Market outcomes of individuals more or less exposed during their childhood. The estimates were enabled by linking identified data from Formal Labor Market, BFP Payment Records, and the Single Registry (SR). In this Natural Experiment, the main identification strategy relies on a rich set of control variables, and on the fact that the release of BFP resources for registered families is automatized and based on municipality poverty parameters estimated by the government. In an alternative identification strategy, we consider an instrumental variable, the observed proxy for the municipality effort to register vulnerable families. These strategies help to solve the potential selection bias of families to the SR, and consequently to the treatment. Nonetheless, since the program selects the most vulnerable families, the threats to the identification suggest that the estimates are lower bounds. Our main results show positive long-term effects on Schooling, and on the Formal Labor Market participation, while mixed results are observed for Earnings. Heterogeneity tests suggest that the effects are stronger for boys, for smaller cities, and for families with never formally employed parents.