The Effects of Locally Financed Class-Size Reduction Initiatives

This research identifies the causal impact of locally approved bonds and parcel taxes on average class sizes and subsequent schooling outcomes in California k-12 public school districts. We build upon recent work by Cellini, Ferreira, and Rothstein (2010) who develop a dynamic regression discontinuity approach to examine the effect of school district expenditures by exploiting the natural discontinuity induced by voting thresholds in local bond referenda. Our analysis extends their work by introducing a regression discontinuity gains model and by comparing the differential effects of approving parcel taxes versus general obligation bonds. More importantly, our analysis focuses on those bonds and parcel taxes specifically targeted to fund class size reduction efforts. Across all of our specifications, bond passage leads to modest reductions in class size several years after approval while parcel taxes appear to minimally affect class size. We find that neither of the two types of locally funded ballot initiatives aimed at reducing class size have any measurable effect on average district scores on 3rd and 7th grade standardized tests. These findings suggest that while efforts to fund class-size reductions through general obligation bonds do lower class size, these reductions are not large enough to result in any meaningful improvements in academic achievement.