Universal private school choice programs began proliferating after the start of the Covid-19 pandemic in 2020. In total, 12 states have adopted programs that allow all families—regardless of income level or whether they were previously enrolled in public schools—to access public funds and use them for private school tuition or other educational services.

Opponents of private school choice, and even some who support targeted programs that only serve disadvantaged students, have been critical of universal policies. They see them as welfare for the rich since many of the families who now benefit could already afford private schools. ProPublica published a report this summer arguing that Arizona’s universal education savings account (ESA) program was primarily to blame for the state’s $1.4 billion budget deficit.

On the other hand, school choice proponents argue that the programs save money. Jason Bedrick and Corey DeAngelis make the case in the Wall Street Journal that Arizona’s ESA program benefits taxpayers because it educates kids for just over half the cost of what public schools spend per student.

Which is true: does universal school choice cost taxpayer money or save it?

Part of the problem is that there are multiple ways to evaluate the fiscal effects of school choice. Opponents of universal choice are correct in pointing out the total cost of these programs, but the sticker prices don’t include offsetting fiscal savings. Conversely, school choice proponents will rightly point out that choice scholarship amounts are often only 40 percent to 70 percent of per-student funding at traditional public schools and thus generate taxpayer savings. While this is again an important data point, it doesn’t consider students who wouldn’t have attended public schools without the choice program and wouldn’t have otherwise received taxpayer support.

Another empirical way to measure the fiscal effects of choice is to look at fixed and variable costs in public school expenditures and estimate fiscal impacts based on how many students using a program would have attended public schools otherwise. For instance, public schools can consolidate classrooms and decrease staffing when students leave, which saves taxpayer money. While this is a helpful lens, it focuses on the types of costs that public schools can save when enrollments decrease rather than dollars that are directly freed up in state budgets when students participate in a school choice program. Public schools are often able to cut costs when they lose students to choice programs, but those dollars don’t always get returned to state budgets.

However, the question that most concerns policymakers differs from those addressed above—namely, how does a school choice program impact the state budget? The answer isn’t as straightforward as one might think. While it’s likely that universal school choice programs cost states money over the near and medium term, they’re not nearly as expensive as opponents often claim—and this cost must be weighed against the benefits of a policy that fundamentally changes K–12 education.

Understanding the key variables

How school choice programs affect state budgets is primarily determined by four variables.

1. Scholarship amounts per student

2. Total school choice program enrollment

3. Portable public education funding per student: The amount of funding a school district loses when a student leaves the public education system to participate in the choice program. Note that this isn’t a total of what public schools spend per student since not all dollars (e.g., some local funds, federal funds, state categorical grants, etc.) get deducted from budgets when a student leaves. Usually, any funds that are explicitly allocated based on student counts, like state funding formula dollars or other enrollment-based state grants, can be recouped by the state as savings when students leave public schools and enroll in a choice program.

4. Switcher rates: A program participant is considered a “switcher” if they would have otherwise enrolled in a public school.

Generally, a choice program will only save state money if (1) portable public education funding per student is higher than the average choice scholarship amount per student and (2) switcher rates are sufficiently high. While exact figures depend on state context and program design, these two conditions must be met.

Case Study: Arizona

Arizona is one of only a few states with enough data to observe which students participate in a new universal school choice program with no caps on cost or participation. The Arizona Department of Education has published two years of data on the state’s universal ESA program. According to the latest quarterly report, about 75,000 students participate in the program. Of that number, about 56,000 students participate under the state’s new universal eligibility criteria, which began in the 2022–23 school year. The following analysis evaluates how much all participants in the ESA program cost the state on net based on the critical variables described earlier.

Determining average scholarship and portable funding amounts

Before examining student participation under Arizona’s ESA expansion, it’s first essential to estimate the funding variables that will impact the state budget. Traditional public and charter school students in Arizona are funded through a weighted student formula, which provides each student a base dollar amount and supplementary weights based on individual needs, grade levels, and other factors. Additionally, public schools receive some state funds outside the main formula that the state can recoup if they disenroll from public schools. Altogether, these sources averaged $8,725 per student in the 2022–23 school year.

The ESA scholarships are funded based on the same weighted formula used for public charter schools. This formula includes supplementary state funding that charters receive to account for the fact that they can’t access local property tax funding like school districts can. Charter formula calculations are then multiplied by a factor of 90 percent—a feature designed to guarantee some state savings for switcher students. Thus, multiplying the statewide average charter formula funding (which includes all formula weights) by 0.9 yields an estimated average ESA amount of $7,786 per student.

Incorporate state-reported participation and switcher data

Now that student-level funding variables have been estimated, actual student participation data reported by the Arizona Department of Education (ADE) can be incorporated. The most recent data published by ADE reports 48,618 new enrollees in grades 1 through 12 after the ESA program became universal and that 15,407 attended public schools in the immediate previous year (the best proxy available for ascertaining switchers). This results in a switcher rate for new ESA enrollees of 31.7 percent over the first two years of universal eligibility.

Additionally, the ADE reports that another 18,988 ESA students use the previous eligibility criteria for high-need students, such as adopted students or students with disabilities. Since most students in this latter group are required to have previously attended public schools, this analysis assumes 90 percent of them are switchers.

Finally, the department also reports another 6,600 kindergarten students for which switcher figures aren’t available since it was their first year of formal education. This analysis assumes that the switcher rate for those kindergarten students would be the same as that for grades 1 through 12 (31.7 percent).

After sorting all ESA participants into switcher and non-switcher categories based on the above assumptions, Arizona’s ESA program’s estimated overall switcher rate is 46.6 percent. Each switcher student generates a net state savings of $939 each, while each non-switcher costs the state $7,786. Multiplying each value by the corresponding number of students in each category yields the estimate that Arizona’s universal school choice program costs the state an estimated $276 million annually. The calculations are summarized in Table 1. Note that positive values connote state costs and negative values connote state savings.

As Table 1 illustrates, the two necessary conditions for generating state fiscal savings aren’t met in the case of Arizona’s universal ESA. While switchers yield savings of $939 per student, there aren’t enough to offset the non-switchers’ cost. Under these assumptions, Arizona would need a switcher rate of 88 percent or higher to be fiscally neutral or to generate state budget savings.

Caveats

There are limitations to these calculations. First and foremost, counting choice participants who were immediately previously enrolled in public schools as the only switchers in the ESA program likely understates the actual switcher rate. The Goldwater Institute, a conservative Arizona-based think tank, has pointed out that Arizona’s public school population is shrinking despite being one of the fastest-growing states in the country. This suggests that some students who would have otherwise attended public schools opt for the state school choice offerings when their families migrate to Arizona. Students in this cohort would be incorrectly identified as non-switchers and ascribed a net cost to the state since they didn’t previously attend public schools in Arizona.

Arizona’s immediate prior public school attendance measure also doesn’t identify students who were previously using one or more of Arizona’s four other tax-credit scholarship programs as switchers. However, many former participants in the tax-credit programs did switch from public schools before enrolling in those other programs, three of which are targeted toward higher-need students or students previously enrolled in public schools.

Data published by Arizona’s Department of Revenue corroborates the assumption that some reported ESA non-switchers came from these programs. While each tax-credit program had grown steadily over the last decade, the latest report for FY23 shows a notable decrease in the number of scholarships granted between FY22 and FY23—the first year of the state’s universal ESA. There was a decrease of 22,322 scholarships across all the tax-credit initiatives between the two fiscal years.

Importantly, these tax-credit scholarship programs can be “stacked,” meaning one student may receive multiple scholarships. Thus, calculating an unduplicated number of program participants isn’t possible with the available data. However, since three of the four programs are targeted toward higher-need students, it’s likely that some students identified as non-switchers in the state’s quarterly ESA reports came from one or more of the four tax-credit scholarship programs and had switched from public schools previously. Consequently, the actual switcher rate for Arizona’s universal ESA is likely higher.

Even so, a further adjusted switcher assumption would still fail to reach the necessary switcher threshold to avoid net state costs. The Common Sense Institute, a free market think tank in Arizona, estimates that on the low end, the average tax-credit scholarship student receives 2.2 scholarships. Suppose one assumes that all 22,322 of the reduced tax-credit scholarships between FY22 and FY23 resulted from the ESA expansion and that all those students were switchers. The result would be 10,146 more switchers and fewer non-switchers by the same amount, adjusting the switcher rate to 60.3 percent.

To be fair, scholars at other choice-supporting institutions like the Heritage Foundation and EdChoice have adopted different assumptions from this analysis and estimated switcher rates of 63 percent and 65 percent, respectively. Regardless, any of these switcher rates still fall short of the 88 percent threshold to achieve fiscal neutrality for the state and would still result in about $180 million in annual net state costs.

Altogether, it’s unlikely for a universal school choice program to have switcher rates high enough to save state money in the short run. The average state has about 9 percent of its students enrolled in private schools. Assuming most families already using private education would jump at the chance to get state support, substantial percentages of public school students would need to opt for the scholarships for the state to attain fiscal neutrality or savings. This shouldn’t necessarily be surprising, because adopting a universal school choice program with no enrollment or cost caps is tantamount to adding tens of thousands of new students who weren’t previously there onto a state’s budget.

Discussion

From the 2000s through the late 2010s, the private school choice programs states adopted were usually restricted to middle- and low-income families and students with disabilities. Under those limitations, they had a solid appeal to policymakers. They allowed higher-need students to get out of their struggling public schools, and they often saved taxpayer money because most participants were switching from public schools.

However, those two features—while real benefits—were never the sole advantage of advancing school choice. For decades, many school choice advocates have envisioned a system where all families have educational options. Limiting school choice initiatives to only the most disadvantaged students led to many states having small, stagnant programs that were merely an escape hatch for some families away from struggling public schools and amounted to only a marginal change from the K–12 status quo. The era of universal choice, which started only four years ago, has already done more to raise the reach and popularity of educational choice than decades of limited programs ever did.

Opponents’ complaints about universal choice programs are most animated against the idea that wealthier families who can already afford private education should get any public subsidy. But as 50CAN’s Derrell Bradford has pointed out, wealthy families already do get generous public subsidies when they attend the nicest, highest-funded public schools in exclusive school districts.

It’s also important to recognize that many children already enrolled in private schools are not from wealthy families. In Iowa, a state that’s in the process of phasing in its universal ESA program, one-third of early participants were already enrolled in private schools even though the program was only open to families below 300 percent of the federal poverty line that year or a household income of less than $90,000 a year for a family of four.

Finally, the likely costs of universal programs still shouldn’t be overblown. The net cost of Arizona’s universal ESA program is still equal to only 1.7 percent of the state’s budget, and it is much less than the $920 million funding increase the state’s public schools received between FY22 and FY23 alone. In that context, concerns that universal choice programs will bankrupt a state are hard to take seriously. If leaders are concerned about minimizing costs, scholarship amounts can be differentiated based on family income—an approach adopted in states like Ohio, Oklahoma, and more recently Louisiana.

Conclusion

Universal school choice programs will likely constitute a net cost to state budgets in the short run. That recognition can set state legislators up to budget wisely and implement universal programs successfully. Even after recognizing the potential state budget impact, the substantial and well-documented benefits of expanding options for all families make the tradeoff worth it. Universal programs allow for large-scale innovation—more educational providers, specialization, and pluralism. More importantly, they exert stronger competitive pressure on the public education monopoly than means-tested choice programs ever did.

Christian Barnard is assistant director of education reform at Reason Foundation.