Program > Papers by author > Silva Joana

Tuesday 13
C1 - Education, development programs III
Chair: Joana Silva
› 10:15 - 10:40 (25min)
› Room 003 - G. Dandelot
Improving Learning Outcomes Through Social Assistance: Regression-Discontinuity Evidence from Brazil
Joana Silva  1@  , Flavio Cireno  2@  , Rafael Proenca  3@  
1 : The World Bank
2 : Ministry of Social Development
3 : Harvard University

Social assistance benefits are often conditional on school attendance. However, they will only lead to higher human-capital accumulation and increased labor market earnings through learning. We use a regression-discontinuity design to examine whether conditional cash transfers impact test scores in national exams. Identification of causal effects exploits an unexpected change in the eligibility cutoff of Brazil's Bolsa Famlia that gives rise to exogenous variation in program participation. The analysis draws on detailed administrative data linking individual records from the programs payment sheets, the nation's single registry of the poor and vulnerable, and educational outcomes from student censuses and test scores. The results reveal that the program leads to significant improvements in learning of Mathematics and Portuguese language. Investigating the mechanisms through which these effects take place, we find that the program causes parents to monitor student effort more frequently, and students to increase effort in school and decrease engagement in work.



  • Other
Online user: 1