School Attendance, Child Labor and Cash Transfers. An Impact Evaluation of PANES

In this paper we analyze the impact an emergency social assistance program, PANES, on school attendance and child labour. The program was carried out in Uruguay from April 2005 to December 2007. Specifically, we analyze the effects of the cash transfer component of the plan (Ingreso Ciudadano), and explore potential explanatory channels such as labour market outcomes, income and awareness of conditionalities. This research is based on a panel of successful and unsuccessful applicants to PANES. The first wave uses the administrative records of the program and the second wave is a follow-up survey that was gathered two months after the program ended and was specifically designed to carry out the impact evaluation of the program. In order to check the robustness of our results, we provide evidence based on two different identification strategies: a regression discontinuity approach using data from the second wave of the panel, and a difference-in-difference approach that exploits the longitudinal nature of the collected data. Our results indicate that the program did not affect school attendance or child labour, whether children are considered as one group or are disaggregated by age or sex. We also do not find any impact on household income, which suggests that income substitution does not explain the lack of results in terms of schooling. It therefore appears that either the size of the transfer was not generous enough to promote school attendance or that the determinants of child school attendance are more complex and require complementary interventions. Our results are particularly relevant for understanding of the role of cash transfers in middle-income countries where attendance rates at primary school are already high, and where the main challenge is to keep students in school at the secondary level. The data also allows us to explore the role of conditionalities. Only a small share of households was aware of the school enrolment condition (20%). Conditionalities were announced and are present in other social security programs in Uruguay, but were ultimately not monitored in this case. We did not find the conditionality to have any robust impact (as perceived by the household) on children’s school enrolment.