The public education establishment routinely argues that school choice programs, where the money follows the child, harm students who remain in public schools. They suggest that students who remain in public schools are worse off because there will be fewer resources available for their education once some children depart public school districts via school choice. That is, there will be fewer students and, consequently, fewer taxpayer dollars to cover the substantial fixed costs of running a school.

Research shows that all forms of school choice tried in the United States have led either to improvement in academic outcomes for students who remain in public schools, or to no effect on academic outcomes for students who remain in public schools. Thus, the evidence on academic outcomes is one-sided. Greater school choice does not harm academic outcomes for students who remain in public schools.

But what about money? The evidence on the fiscal effects of school choice on public school districts is not readily available. Robert M. Costrell has shown that it is straightforward to design a school choice program that saves taxpayers money. He also suggests that the fiscal effect of a given school choice program on local school district budgets is more complicated. Specifically, school choice programs that allow school districts to retain funding for any fixed costs would not harm the fiscal health of public schools or decrease resources available to students who remain in public schools.

In a new report, “The Fiscal Effects of School Choice Programs on Public School Districts,” published in March by the Friedman Foundation for Educational Choice, I constructed the first-ever estimates for each state and the District of Columbia of the short-run fixed costs of educating children in public schools. I endeavored to make cautious overestimates of these short-run fixed costs.

In the interest of creating an overestimate of fixed costs, the report treated the following as fixed costs in the short run: expenditures on capital, interest, general administration, school administration, operations and maintenance, transportation, and “other” support services. Of course, if a significant number of students left a school district from one year to the next, some of these costs could be reduced immediately. For example, a school losing a large number of students could reduce the number of assistant principals from two to one; there could be fewer bus routes; two schools could be merged into one; etc. However, the goal of the report is to create an overestimate of fixed costs. A cautious overestimate allows us to be comfortable that school choice programs where “the money follows the child” can be designed in such a manner as to improve the fiscal situation of public school districts.

While I treat expenditures on capital, interest, general administration, school administration, operations and maintenance, transportation, and “other” support services as fixed costs in the short run, all of these costs are variable in the long run.

The definition of fixed costs from the leading introductory textbook in economics, the sixth edition of Principles of Economics by N. Gregory Mankiw, is: “Some costs, called fixed costs, do not vary with the quantity of output produced.” So fixed costs do not change one penny when the output decreases or increases. In the long run, all costs are variable. Mankiw continues:

Over a period of only a few months, Ford cannot adjust the number or size of its car factories. The only way it can produce additional cars is to hire more workers at the factories it already has. The cost of these factories is, therefore, a fixed cost in the short run. By contrast, over a period of several years, Ford can expand the size of its factories, build new factories, or close old ones. Thus, the cost of its factories is a variable cost in the long run.

Any microeconomics or accounting textbook would have similar phraseology. The implication of this is that fixed costs are only fixed for a given period of time. In the long run, all costs are variable.

Public schools can make new strategic decisions in response to permanent changes in their student counts. Thus, after a few years of a new school choice program, when enrollment trends become apparent, all taxpayer funds devoted to K-12 education can “follow the child” to the schools their parents deem better.

The proper way to think about this issue is not whether public school districts have in the past reduced costs when students in large numbers left the district for any reason. The issue is whether they are able to do so. Evidence that school districts increased expenditures when the number of students they served significantly decreased does not necessarily mean that they cannot decrease expenditures when students leave. Perhaps they did not have to reduce expenditures when students left because one or more levels of government chose not to reduce taxpayer funding, so districts did not reduce expenditures.

The key question for this analysis is the following:

If a significant number of students left a public school district for any reason from one year to the next, then is it feasible for the district to reduce some of its expenditures commensurate with the decrease in its student population?

The answer that comes from analyzing the finances of large and small school districts that lost students is “yes.” Both the large school districts and the small ones were able to reduce the combination of instructional and support expenses at a higher rate than the losses in students. Thus, these costs were variable, even in the short run.

Oklahoma’s average spending per student was $8,715 in 2008-09. I estimate that 33 percent of these costs can be considered fixed costs in the short run. The remaining 67 percent, or $5,860 per student, are found to be variable costs, or costs that change with student enrollment. The implication of this finding is that a school choice program where less than $5,860 per student is redirected from a child’s former public school to another school of his or her parents’ choosing would actually improve the fiscal health of the average public school district. And, it would provide more resources for students who remain in public schools.

If, however, students with above-average costs to educate leave a school district, such as students with special needs, then the short-run variable costs would be significantly greater than $5,860 per student. Thus, scholarship amounts for special-needs students well above $5,860 per student would not harm the fiscal health of public school districts in Oklahoma. Special-needs students typically receive much smaller pupil-teacher ratios than other students, which is why they tend to receive much higher instructional expenditures than other students. Instructional expenditures are a variable cost of educating students, even in the short run. In the report I show that school districts have been able to reduce teachers and other support personnel in a manner more than commensurate with decreases in students—and these decreases in students were not due to school choice programs. Thus, these costs are variable, even year to year.

I have also done analogous calculations for the Jenks and Union school districts. Jenks spends more per student than the Oklahoma average—$9,265 per student for Jenks versus $8,715 for the state average. However, Jenks spends fewer dollars on instruction and other variable costs than the state average. For example, Jenks spends only $3,667 per student on instruction versus an average of $4,508 per student in the state, about $841 less per student. Jenks spends significantly more money per student on capital and overhead than the state average. Consequently, the estimate of short-run variable costs for Jenks is $5,225 per student, or 56.4 percent of total expenditures per student.

Union spends a bit less than the state average, $8,567 per student. Like Jenks, Union spends more on overhead than the state average and less on instruction. Also like Jenks, this leads the estimate of short-run variable costs for Union to be less than the state average. For Union, short-run variable costs are estimated to be $5,041 per student.

The rationale as to why a loss of students and the funding associated with those students could increase the performance of traditional public schools is twofold. First, a large number of empirical studies have found very large differences in teaching effectiveness across public school teachers. If public schools lose students and funding, they could choose to lay off the least effective teachers. The remaining students would be reassigned to more effective teachers, which would lead to a significant improvement in their academic achievement. Second, Chakrabarti (2007) has shown theoretically and empirically that when more money follows the child, the incentives are stronger for public school leaders to improve their schools. In Milwaukee, they did improve the public schools when there was an increase in the amount of money that followed voucher students to private schools.

In short, public school leaders and their allies routinely argue, that “when one student leaves, we still have to pay for that student’s teacher. We still have to pay the electricity bill. We still have to pay …” The implication of their argument is that all costs of running public schools are fixed. Interestingly, I have never heard that argument made when there is an increase in the number of students. If a public school adds only one student, do the lobbyists for public school leaders suggest that the district should not receive any extra funding? I suspect that has never happened.

Fixed costs are only fixed for a given period of time. In the long run, all costs are variable.