Bilingual education and child labor: Lessons from Peru

Abstract

This paper investigates the effect of bilingual education reform in Peru on child labor. The reform aimed to improve access to schooling for indigenous children in rural areas with greater integration of education providers and the delivery of lessons and material in vernacular indigenous languages. The core specifications use child-level panel data from the Young Lives Study and a two-way fixed effects difference-in-difference approach to test the causal effect of the program on child labor. The results show evidence to suggest that access to bilingual education has a negative and statistically significant effect on the extensive margin of child labor, with some specifications showing consistent results for its intensive margin counterpart. The preferred estimates are sizeable, with some specifications showing a decrease in the likelihood that an indigenous child enters child labor by between 12 and 18 percentage points. The findings are robust to several specifications including randomized inference techniques, synthetic difference-in-difference, and placebo tests. The paper also shows that concurrent policies, including Peru's conditional cash transfer program, are unlikely drivers of the core results. The paper's findings also suggest that the program resulted in greater inputs (i.e., time) into schooling and improved outcomes of education for indigenous children.