Universal Access to Free School Meals and Student Achievement

US school districts, panel exploiting staggered Community Eligibility Provision adoption across and within states

US public-school students in high-poverty districts eligible for the Community Eligibility Provision. High-income country; food-insecure subpopulation.

Difference-in-differences using within- and across-state variation in the timing of CEP adoption. CEP lets sufficiently high-poverty schools offer free meals to every student regardless of individual income eligibility, which both extends free meals to some non-qualifying students and removes the stigma and paperwork of individual eligibility. The design identifies the effect of the POLICY, which bundles nutrition with household income relief (families no longer pay for meals), reduced administrative burden, and reduced stigma. It cannot isolate nutrition.

Universal free meals raise breakfast participation 38% and lunch participation 12%. Math performance improves in districts with LOW baseline free-meal eligibility - i.e. where CEP actually extends free meals to students who previously paid - and particularly among racial/ethnic groups with low income-based take-up. Where nearly everyone already qualified, there is little to move and little effect. The heterogeneity pattern is what an income-and-take-up channel predicts.

Low for selection - CEP eligibility is a statutory function of district poverty measures, not a family choice. But the treatment is a bundle, so a positive effect is not evidence about nutrition.