Research

Abstract: This paper examines experimentally the impact of classroom inputs on student achievement from kindergarten through 6th grade. We use data from a large cohort of elementary school students in Ecuador, who were randomly assigned to different classrooms at the start of each academic year. We estimate reduced form and structural models of the process of skill accumulation to show that learning at the end of a grade is close to an additive function of classroom quality experienced in all previous grades. There is no evidence of dynamic complementarity between classroom quality across different grades.

Abstract: In many low- and middle-income countries, most children walk to school through dense and often unsafe neighborhoods. These commutes impose daily coordination burdens and safety risks on families. Yet despite its consequences for attendance, learning, and well-being, school commuting remains an overlooked dimension of education policy. This paper evaluates walking-to-school caravans (WSC), in which trained monitors accompany groups of children along safe routes to and from school. We study Bogotá's Ciempiés program, launched in 2019 and rolled out across public primary schools. Using administrative data from 2016–2024 and a differences-in-differences design that exploits the staggered adoption of the program, we estimate causal effects on educational outcomes and road safety. Overall, WSC reduce school switching by 6.8 percentage points and branch switching by 10.9 points—over 40% declines from pre-treatment rates in the control group. We also find suggestive reductions in grade repetition and dropout, though no effects on traffic accidents. At USD 285 per child annually (between 24% and 51% cheaper than busing), WSC are a cost-effective, scalable intervention that eases family coordination burdens and improves school permanence and progression, making them a promising model for other dense urban settings worldwide.

 [AEA RCT Registry]