Essays on Development Economics and Education

Mismatch of the language of instruction in schools and locally spoken vernacular language can adversely affect foundational learning. Thus, bilingual education can improve the schooling outcomes of young children. On the other hand, large-scale bilingual programs in school instruction can be resource-scarce, resulting in more demands on already thinly spread teacher effort. In chapters I and II of this dissertation, I empirically study a large-scale bilingual program to resolve this ambiguity: a nationwide bilingual education reform in Peru aimed at correcting instructional mismatch. I exploit the timing of the policy's rollout in a difference-in-differences framework that harnesses both publicly available data and confidential administrative records. I find that the program results in a 40 percent increase in the self-reported probability of offering BE by school principals in treated schools. I also detect large effects on access to inputs for BE. In schools targeted by the program, the percentage of teachers trained in BE and the probability of receiving textbooks in the mother tongue increased by 60 percent and more than 1,000 percent, respectively. Next, I document the effects on students' outcomes. While the failure rate for students from grades 2-4 decreased by 10 percent, I find evidence of adverse short-term effects on math standardized test scores. This effect is possibly driven by a change in the recruitment process for teachers in the treated school, which requires that teachers be proficient in the local language while inducing negative selection in academic ability.

In the third chapter, in collaboration with Professor Sarah Turner, we study how adverse local labor market shocks such as those generated by exposure of local industries to international trade impact on high school enrollment and completion. The effects are theoretically ambiguous. Incentives to stay in school increase when employment prospects are weak; yet, public resources for local schools may also shrink. We demonstrate that, while high school enrollment rates increase significantly, high school degree attainment does not show commensurate growth. Diploma counts relative to the population indicate only a modest increase, while the share of young adults with a high school degree in a community does not change. The correspondence between high school enrollment and diploma receipt of young adults reflects important measurement issues, as ‘outmigration’ of young adults, changes in the timing of degree receipt and shifts from GED attainment to diplomas may complicate measurement. Moreover, the negative impact of trade exposure on secondary school resources per student operates in the opposite direction of enrollment demand to attenuate gains in high school degree attainment.