skip navigation

Economics Seminar Series – Robert McMillan, University of Toronto

Education Reform in General Equilibrium: Evidence from California’s Class Size Reduction
December 1, 2017, 3:30pm -5:00pm
444 Fronczak Hall

Abstract: Understanding general equilibrium responses to major education reforms is important for policy, given their potential to swamp any direct reform effects. This paper focuses on a general equilibrium sorting response that is likely to matter whenever a reform improves public school quality significantly, leading some households to re-optimize by switching out of private schools. It does so in the context of the unique roll-out, grade-by-grade, of California’s statewide class size reduction program of the late-1990s – a measure inspired by the STAR Experiment, and up to that time, the most costly state education reform ever implemented in the United States. Using a transparent differencing strategy that exploits the grade-specific timing of the reform, we first show significant reductions in local private school share for relevant elementary grades, marked changes in public school sociodemographics, and large house price increases in areas where the policy had been implemented. These reduced-form findings motivate an estimable structural framework that allows us to gauge the policy’s direct and indirect effects on a common scale for the first time, as well as their respective persistence rates. Identifying the structural parameters using a generalization of the differencing approach, our estimates reveal a significant pure class size effect of 0.11σ (in terms of mathematics scores), and an even larger indirect effect via induced changes in school demographics, of 0.17σ: ignoring this indirect response would thus understate the reform’s full impact by well over half. We find both indirect and direct effects of the reform persist positively, invalidating a simple difference-in-differences approach; further, a bounding exercise suggests that the indirect sorting effect involves positive spillovers for students already in public school. The analysis draws attention, more broadly, to conditions in which the indirect sorting effects of large-scale education reforms in other contexts are likely to be first-order.