Test-Mex: Estimating the effects of school year length on student performance in Mexico

Estimating the impact of changing school inputs on student performance is often difficult because these inputs are endogenously determined. We investigate a quasi-experiment that altered the number of instructional days prior to a nationwide test in Mexico. Our exogenous source of variation comes from across states and over time changes in the date when the school year started and the date when the test was administered. We find that having more days of instruction prior to examination slightly improves student performance but exhibits diminishing marginal returns. The effects vary along the distribution of resources as determined by a poverty index, with lower improvements in poorer schools. These findings imply a weaker net benefit of policies expanding the length of the school year as they could widen the achievement gap by socioeconomic status.