Association for Public Policy Analysis & Management

A substantial body of research has investigated the educational impact of TELs in Massachusetts (Bradbury et al 2001; Shadbegian and Jones 2005), Illinois (Dye and McGuire 1997; Dye et al. 2005), Oregon (Figlio 1998), and nationwide (Figlio 1997; Shadbegian 2003; Blankenau and Skidmore 2004; Downes 2007), but neither the state-specific nor national studies have produced a clear consensus on whether and how education inputs are significantly affected, as reviewed in Ballal and Rubenstein (2009), largely due to the lack of precise measure of TELs’ stringency and a series of methodological challenges.

This article adds to the literature by examining the most recent case of TELs in New York (NY). School districts in NY have been under a new property tax levy limit since 2013. Students on TELs mostly agree that relative to other TEL types, property tax levy limit is highly likely to be fiscally constraining local governments. Under this limit, school districts in NY are not allowed to have a budget that requires a property tax levy exceeding the prior year’s levy by more than 2 percent or the rate of inflation, whichever is less. This limit can only be overridden with a 60-percent supermajority vote in a public budget referendum.

In this paper, we seek to examine how this new levy limit affects education inputs, including total spending, different categories of expenditures (instructional v. administrative), teacher salaries, and student-teacher ratios. We hypothesize that the levy limit negatively affects education inputs. We also hypothesize that the limit has more severe effects on high-need school districts. In addition, compared with previous studies, we explicitly examine the role the override policy plays in the causal dynamics.

Using New York school district-level data between 2011 and 2017, we compare districts whose levies are close to their limits (limit-binding or treatment districts) and those whose levies are far from their limits (comparison districts). Based on event-time study and difference-in-difference methods, we find that the limit, especially for ‘at limit’ and ‘very near limit’ districts (highly constrained districts), significantly reduces total education operating expenditures. Also, consistent with the empirical findings in Figlio (1998), this reduction in total expenditures is driven mostly from a reduction in instructional spending. In other words, school districts are induced to lower their instructional spending while keeping administrative spending intact. Highly constrained districts also see no growth in teacher salaries and a slight (but insignificant) increase in student-teacher ratios. Finally, as expected, the levy limit more severely affects high-need school districts.