A Distributional Difference-in-Difference Evaluation of the Response of School Expenditures to Reforms and Tax Limits

Abstract

Prior work uses a parametric approach to study the distributional effects of school finance reform and finds evidence that reform yields greater equality of school expenditures by lowering spending in high-spending districts (leveling down) or increasing spending in low-spending districts (leveling up). We develop a kernel density difference-in-difference approach to isolate how tax limits and/or education finance reform affect the full distribution of education expenditures. Simulations of the difference in distributional differences across school finance regimes over time suggest that parametric approaches offer an incomplete description of the distributional impacts of policy changes. Using data for the population of U.S. school districts in 1972, 1982, and 1992, we find that educational reforms and tax limits yield greater equality of expenditures by reducing the number of districts in the tails of the distribution, particularly when both are adopted during the same decade. Our results also suggest that the incompleteness of the parametric descriptions used in prior work can suggest greater variation in the distributional consequences of reform than is actually present.