The market-level effects of charter schools on student outcomes: A national analysis of school districts

We study the total, market-level effects of charter schools, and their mechanisms, on a national level and across multiple outcomes. Using a generalized difference-in-differences method, we find that increasing the charter market share by 10 percentage points increases math and ELA elementary/middle test scores of the entire geographic district in which they locate by 0.01 standard deviations and increases the high school graduation rate by 1-2 percentage points. The effects are concentrated in larger, urban districts. The main mechanism appears to be the participant effect, though competitive effects are increasing in the participant effect and driven partly by the closure of low-performing traditional public schools. Causal interpretation of these findings is reinforced by analysis of potential endogeneity of charter school location and timing using placebo analysis and other methods. This study improves understanding of the array of charter mechanisms and their effects on multiple outcomes, on a national basis.