Unintended consequence of centralized public school funding in Michigan education.

• Author: Zimmer, Ron

1. Introduction

   In an effort to create greater equity among school districts, a number of states have shifted the responsibility for school funding from local school districts to the state. Although many researchers have found that centralizing funding for public schools has been successful in creating greater equity (Murray, Evans, and Schwab 1998: Moser and Rubenstein 2002), other researchers have pointed out that centralizing funding can lead to unintended consequences. For example, Fischel (1989, 1992) suggests that families in wealthier districts, which have a stronger demand for education, are no longer able to match their preference for education under a system of equalized funding. In some cases, this may drive parents to seek private school alternatives (Dowries and Schoeman 1998). (1) In other cases, families meet their preferences by making private contributions to public schools (Sonstelie and Brunner 1997). Furthermore, Theobald and Picus (1991) argue that centralizing funding forces education, as a state service, to compete with other state services for funds, and as a result, the growth rate of spending per pupil decreases over time. (2) These articles illustrate that policies designed to centralize control of public school funding can produce unintended consequences.

   In this article, we introduce into the literature an additional consequence caused by centralizing control of school funding. We argue that centralized control of public school funding with the objective to equalize operating expenditures (i.e., salaries, supplies, and other ongoing expenses) across school districts while leaving capital acquisitions (i.e., purchases of school buildings and other facilities) under local control results in resource-rich school districts becoming more reliant on debt financing. (3) We find empirical support for this theory using data in the state of Michigan during its transition to a centralized school financing system in 1994.

   In the next section, we present an overview of Michigan's move to a centralized funding policy. In the third section, we present the regression model used to test the theory, followed by the description of the data and the estimated results. The final section of the study summarizes our findings.

2. Overview of Michigan's Educational System

   In 1993, a legislative anomaly lead to an abrupt change in the method of financing public education in Michigan. Prior to 1993, Michigan had one of the most locally controlled school systems in the United States. Over 65% of school operating expenditures came from local revenues, whereas the remaining 35% were derived from state and federal sources (Courant, Gramlich, and Loeb 1995). Property owners argued that local reliance on property taxes created an excessive burden and voiced their complaints to state representatives. Attempts to lower these taxes were defeated annually until Debbie Stabenow, a Democrat senator and candidate for governor, frustrated with the tax debate, proposed an absurdity to eliminate all property taxes without suggesting an alternative funding mechanism (Courant and Loeb 1997). Astonishingly, the legislature passed her bill, leaving the public school system with no means of funding. Seizing an opportunity to create greater equity among school districts, Michigan officials quickly devised a centralized system of financing public school operations, known as Proposal A, that reduces districts' reliance on local property taxes by increasing the state's sales tax. Integrated into Proposal A is a distribution formula that reduces the disparity in operating expenditures across high- and low-spending school districts, in part, by constraining the allocation of operating funds for high-spending districts.

   However, Proposal A does not affect local control of funding for capital projects (Watkins 2002). Michigan's state operating budget typically provides less than 5% of the total revenue needed to service school debt and no revenue for capital expenditures. (4) As a result, high-spending school districts may issue more bonds to offset the mandated expenditure reductions as a mechanism of maintaining their educational preference, which undermines the policy's objective to reduce spending disparities.

   Table 1 outlines the changes in tax rates created by the proposal. The two most prominent changes are the increase in state sales tax from 4% to 6% and a reduction in general property taxes from an average of 33 mill to only a 6 mill statewide property tax and an 18 mill property tax on nonhomestead property (business and rental property). (5) The reduction in property taxes and increased state sales tax centralizes the sources of funding and reduces local voters' ability to increase local revenue.

   These changes significantly altered the relationship between per pupil expenditures and property values, which were highly correlated prior to Proposal A. Because of the reliance on local property taxes prior to Proposal A, school districts with high property values had high per-pupil spending, whereas school districts with low property values had low per-pupil expenditures. As a means to bring greater equity to spending, all districts spending less than $3950 per pupil (low-spending districts) prior to Proposal A are guaranteed an inflation-adjusting foundation grant of no less than $3950, which essentially increases their spending levels. School districts spending more than $6500 per pupil (high-spending districts) prior to Proposal A find their spending levels limited under Proposal A. These districts are allowed an inflation-adjusting foundation grant starting at no more than $6500 per pupil, although the state permits them to assess an additional 3 mill enhancement tax on homestead property for up to three years. (6)

   Despite the allowed enhancement, the objective of the policy is to reduce the disparity in per pupil spending with the additional 3 mills acting as a pressure value for districts suffering too harshly from the funding cuts. As evidence that the cuts were meaningful, nearly all of the school districts that spent more than $6500 in 1994 voted for the 3 mill enhancement, which produced a weighted-mean fixed local rate (weighted by the percent homestead and nonhomestead property value) of 13.84 mills, which is considerably less than the mean rate of 30.85 mills before Proposal A. Although it may have been politically necessary to allow a 3-mill increase, the enhancement falls short of increasing funding for high-spending districts to locally desired levels.

   Because operating expenditures are regulated, high-spending districts now have an incentive to circumvent the policy by increasing the use of capital markets and debt financing. Cullen and Loeb (1999) suggest that when external constraints are placed on districts, they may have an incentive to carry out capital expenditures under the guise of operating expenditures (due to the fungible nature of budgets). Applying Cullen and Loeb's (1999) argument, one could imagine high-spending districts issuing bonds for expenses that could either be defined as capital or operating expenditures. For example, before Proposal A many districts may have paid for the refurbishing of a classroom, painting of walls, patching a roof, putting in a new intercom system, etc., out of operating expenditures. Much of these items could now be paid through the issuance of a bond and free up money to maintain the level of operating expenditures. However, it is not entirely clear that this is legal under Michigan law. Michigan school districts are given the authority to issue bonds as long as the bond is not issued longer than the lifespan of the asset (Michigan Department of Treasury 2001). Whether for capital accumulation or fiscal substitution, Proposal A should induce voters to approve more bond proposals, and thus we should expect the probability of issuing debt to increase for high-spending school districts.

   Proposal A provides a natural experiment for estimating the probability of issuing debt by high-spending districts after centralized funding since it was passed and was enacted within a matter of months with little time for preemptive action by district administrators. Thus, Proposal A represents a nearly clean policy "shock." The implementation date, school year 1994-1995, is a distinct point in time to examine the effect. Second, Proposal A targets school districts by expenditure levels and thus provides a clear...