This paper uses data on Ohio school districts to estimate the short and long term impact of different types of school expenditures on student outcomes. Our identification strategy employs a dynamic regression discontinuity design that relies upon the exogenous variation in public school funding created by marginally approved or failed local referenda to fund Ohio schools. We find that additional school expenditures on operating, minor capital, and major capital expenditure categories do not have a statistically significant effect on the student test scores of the average public school. Importantly, however, operating expenditures have a large and statistically significant impact on student performance in higher poverty school districts. We also examine possible channels (e.g., class size, attendance, discipline, and teachers' compensation) through which each type of expenditure may affect outcomes, and we find that teachers' compensation is the only channel that is affected by additional operating and minor capital expenditures.