Evidence of the Impact of Children's Domestic and Market Labor on Learning from School Census Data in Brazil

Abstract

In this study we analyze the impact of domestic and market child labor on learning using Prova Brazil census data from 2007 and 2011. To do that, we created a large panel dataset with students in 5th and 9th Grades. To measure the impact of children working in the labor market and/or in their households on Portuguese and Mathematics test scores, we estimated fixed effect models. An instrumental variable approach, proposed by Lewbel, was applied to the models to control for the endogeneity of child labor. Possible attrition bias was taken into account through inverse probability weights. Results show that the work performed by children either in the household, or in the labor market was detrimental to their academic performance, with working only in the labor market showing the largest impact for both girls and boys, followed by working both in the market and inside the house, and by working in the house alone. The largest impact was for 5th Grade girls who suffered a reduction of 19% in their Portuguese test scores when they worked only in the labor market. The largest impact for boys was a reduction of 14% in their 5th Grade Portuguese test scores when they worked only in the labor market.