Introduction

Since universal eligibility in 2022, Arizona’s Empowerment Scholarship Account (ESA) program has seen rapid growth. A lack of reliable and consistent data about who is using the program and how is contributing to public uncertainty. The relatively inconsistent reporting by official sources, like the program’s administrators and marketplace vendor, has enabled misinformation and exaggeration to thrive.

Subject to the available data, this regular report is intended to provide ongoing and consistent answers to some recurring questions that policymakers and the public have been asking, like:

How many people are using the ESA program, how much bigger will it grow?
What is that going to cost?
What do these new ESA users look like, in terms of their demographic characteristics?
And is the program being well-run and managed in the public interest?

This is our second release of that report, with the latest data available from the U.S. Census Bureau, the Arizona Department of Education, and other sources.

Key Findings

CSI expects the ESA program to grow to 91,500 students next year and enroll between 86,500 and 89,500 this Fiscal Year (as of May 2025). The program has exhausted its rapid “catch-up” growth period and now depends on students switching from other programs for future growth. Because school-aged population growth in Arizona is now negative, total enrollment across all K-12 formula programs in Arizona may have peaked.
The cost of the ESA program is estimated to reach $882M this year, increasing to $939 million next year. For context, total public funding for K-12 in Arizona will be over $10.0 billion this year; ESA spending is about 8% of total K-12 funding. Universal-eligibility users now account for an estimated 52% ($460M) of total ESA costs.
CSI estimates that most ESA users continue to come from middle-income households (defined as those earning between $75,000 and $150,000 per year). 56% of ESA students live in ZIP codes where the median income of families with at least one child is between $75k and $150k; 35% of all Arizona families earn between $75k and $150k. While their expected income has drifted higher in the latest data, a disproportionate share of middle-income households use an ESA.
CSI estimates that over 18% of ESA users are nonwhite, and that nearly 13% live in rural areas.
For the entire period that data is available, CSI estimates that the ESA program has consistently approved far fewer expenditures than there is available funding. This implies a substantial and growing backlog of pending purchases, and growing ESA fund balances. In Q4 2024, ADE approved expenditures for 88% of expected available ESA dollars (a significant improvement from 71% in the prior period).
Sustaining an approval rate closer to 100% of awarded dollars will indicate ongoing improvement in the administration of the ESA program by ADE and its third-party vendor.

How Much Larger Will The Program Grow?

As of May 15, 2025, 89,783 students were enrolled in Arizona’s ESA program; given our best estimates of Arizona’s total school-aged population and enrollment in other programs, there are a total of 108,241 students eligible for an ESA or STO in Arizona today. Considering uncertainty about how many students are using an STO scholarship, it is likely all or nearly all ESA-eligible students in Arizona are now enrolled. Further growth in the future will have to come from students leaving other publicly-funded options. According to the Department of Education, over half of students newly-enrolling in the ESA program today previously attended public school.

Arizona today effectively provides K-12 formula funding on a fully per-student basis; any 5-17 year old Arizona resident is guaranteed at least one publicly-funded option to offset the costs of their K-12 education, regardless of where they attend. Of these options (public district, public charter, private school with an STO scholarship, or home- and private-school with an Empowerment Scholarship Account) there is a universe of students choosing each one.

Based on the 2023 American Community Survey, CSI estimates that there are 1,187,874 school-aged children in Arizona as of 2024. Of those, an estimated 108,241 school-aged children in Arizona do not attend District or Charter public schools. Most of these students are eligible for either an ESA or STO or both. For the last two years, that population has rapidly enrolled in the Universal-eligibility category of the ESA program, contributing to a temporary growth spike.

In FY 2024, STO’s gave 61,742 scholarships to students attending private schools. The average scholarship was reported to be $3,161. Because students can receive multiple scholarships and STO’s are not required to report recipient student counts (only the number of scholarships), CSI estimates that participating students receive an average of between 1.36 and 3.40 scholarships, or the equivalent of up to $10,700. Given this assumption and based on historic growth rates in the scholarships given, CSI estimates the STO population is between 16,850 and 41,972 students and are ESA ineligible. Given nearly 90,000 ESA recipients, the ESA program is likely fully- or nearly-fully subscribed; approximately all eligible Arizona students are publicly funded today.

In summary, CSI estimates that there were 108,241 school-aged children in Arizona who were not attending a public school full-time, and therefore eligible for an ESA. Of those, 89,783 were enrolled in the ESA program (as of May 15th, 2025). After accounting for uncertain STO utilization, the ESA program is now for all intents and purposes fully subscribed, subject to new growth and “switchers”.

During the catch-up period when the ESA program was expanded to universal eligibility, enrollment boomed. In 2024, enrollment began to slow as the universe of students eligible but not enrolled shrank. Enrollment growth has slowed even more as most students are accounted for by some system of education funding; future enrollment in the ESA program will be a function of demographics and choice.

Because of a declining school-aged population, ESA growth must now come from other publicly funded options, in an often net-savings or cost-neutral fashion. As noted in our previous report, STO scholarships have historically held steady at between 90,000 and 98,000. In the last 2 fiscal years, scholarships have fallen from 98,138 to 61,742 (this decline could generate up to an estimated $104.5 million in new income tax payments, given falling donations and the average scholarship amount). Most of these students likely switched to the ESA program. The portion of the STO students that decreased the least from FY 22 to FY 24 were the Low-Income and Disabled Corporate STO’s, suggesting that these students are receiving more in STO scholarships than they would with an ESA (potentially enough to cover a full private-school tuition); students receiving less in STO awards than they could from an ESA are most likely to switch. However, some students may remain on an STO because it is easier to use, administratively, or is perceived to be more stable, politically. There are fewer reporting requirements and costs placed on parents and schools (providers) in the STO program.

This trend is confirmed in the increasing share of students newly enrolling in the ESA program who previously attended public school. Intuitively, we would expect this ratio to rise over time as the number of students otherwise available to enroll goes down. According to ADE, in late-2022, fewer than 25% of newly enrolled ESA students had previously attended public school; today that share is 55%.

How Much Will the ESA Program Cost This Year?

The FY 2025 enacted budget appropriated $9.93 billion for K-12 “Basic State Aid” formula funding. This funding is allocated to all District and Charter schools and ESA students based on a per-pupil formula and makes up about two-thirds of all state and local public education funding in Arizona.

The FY25 Basic State Aid appropriation was expected to fund 82,400 ESA enrolled students, 852,000 District enrolled students, and 229,900 Charter enrolled students. These enrollment estimates were produced in August 2024. In recent years, enrollment has been particularly difficult to predict due to changes in educational preferences and demographic declines leading to falling enrollments – particularly in District public schools.

As of May 2025, CSI estimates actual Base State Aid formula costs will be approximately $97.3 million above the appropriation.

Empowerment Scholarship Accounts. $822 million was specifically appropriated for an estimated 82,400 ESA students. Funded enrollment (Fiscal Year 2025) as of May 12, 2025 was reported to be 86,862 students, and in the 2025 Quarter 2 Annual Report the average ESA award amount was $10,261. Based on these figures, CSI estimates the actual cost of the ESA program in FY 2025 will be $882 million.

By the end of this fiscal year (on June 30th, 2025), CSI projects no further growth or even slight declines from current levels. Historically, ESA enrollment has grown the fastest around September and then again at the beginning of the calendar year. Program enrollment historically declines in the summer.

From the period of February 26, 2024 to the FY 2024 peak enrollment on April 8, 2024, ESA enrollment grew 1.99%, while between the 2024 and 2025 school years the program grew about 10%. Between 2% and 10% more growth in the next school year would see the ESA program enroll about 91,500 students in FY 2026. This growth will not affect the FY 2025 cost however, as students enrolled after March 31st, 2025 will not receive funding this year. At current average award amounts, FY 2026 ESA program cost given this growth would be approximately $938.9 million.

The total expected supplemental need for the ESA program this year above its appropriation is $59.8 million. This is $9.2 million above what JLBC Baseline projected the supplemental need to be in January. The FY2025 adopted budget estimated the program would reach 89,700 students in FY 2026. Given slowing growth, actual enrollment next year may still be in this range, but the program will reach that level of enrollment earlier than forecast.

Additionally, while the total cost of the program has increased substantially since 2022 and creation of the Universal-eligibility enrollment category, other qualifying populations have continued to see growth. For school year 2022 (the last year prior to Universal-eligibility), there were 12,127 students in the ESA program receiving an average annual award of $15,566. For the school year ending in 2025, there were approximately 22,131 students enrolled in these same categories (+82%). Assuming this population would receive the same average annual award as the pre-Universal cohort (adjusted for formula changes since), this group is receiving over $421 million of approximately $882 million in total awards. Succinctly: Universal-eligibility contributes $461 million (52%) to the total costs of the ESA program, despite enrolling three-fourths of all students.

District Public Schools. Approximately $7.0 billion was specifically appropriated for an estimated 852,000 District students. This was the lowest projected District enrollment figure in at least the last 20 years. According to the most recent estimates available to CSI (ADE October 1 enrollment counts), actual enrollment in traditional District public schools for the 2024-25 school year was 848,700 students – less than the number of students that were enrolled in FY2003.

District public schools are awarded formula funding based on “100 day counts”, which should have been tabulated in October. However, there is both a significant lag and revision to these figures which can affect these estimates over a fiscal year.

This creates an expected General Fund savings of $28.2 million by the end of FY 2025, relative to the appropriation amount in last years enacted budget. CSI’s total expected cost for district public school students is $114 million below JLBC’s Baseline estimate released in January.

Charter Public Schools. Approximately $2.11 billion was specifically appropriated for an estimated 229,900 Charter school students. Actual enrollment (according to the latest published ADE October 1 enrollment counts) was 230,922 students, at an estimated cost of $2.17 billion. This creates $65.7 million in cost relative to the amount appropriated this year and $106.8 million in cost relative to JLBCs revised estimate. While Charter students on average receive less in total per pupil funding than their District peers, they (like ESA recipients) are relatively more dependent on the State General Fund than District students.

Overall, the total cost of Basic State Aid to District, Charter, and ESA students is likely to exceed $10.0 billion this year. ADE is facing a deficit of $97.3 million in formula costs for the Fiscal Year ending on June 30, 2025. Of that total amount, $881.9 million is attributable to the ESA program ($59 million more than budgeted and $8 million more than JLBC’s Baseline estimate).

Long-Term State K-12 Budget Implications of ESA

Since Universal-eligibility, ESA participation (and budgeted costs) have grown rapidly. This growth coincided with other rapid spending growth in the state budget, and an unrelated reversion of revenue growth to long-term averages, which necessitated a budget correction last year. Potentially lost in this discussion was larger consideration of an issue CSI has been calling attention to for some time now: demographic change and preference shifts among the state’s school-aged children driving long-term declines in traditional public school enrollment.

This trend began before Universal-eligibility ESA’s were available to families, and has persisted since. Today, there are 50,000 fewer students enrolled in Arizona public schools than there were prior to the pandemic; conversely, there are 75,000 more students enrolled in the ESA program than there were at that time. While considerable attention has been given to the costs of this growth in the ESA program, policymakers should also consider the offsetting savings from declining enrollment in other publicly-funded programs.

Arizona allocates funding to nearly all K-12 students based on a per-pupil formula; when enrollment in one of these formula-programs declines, funding declines commensurately. When enrollment grows, funding grows. The state also prepares long-term forecasts based on expectations about how these programs will grow. Prior to 2020, the non-partisan Joint Legislative Budget Committee consistently predicted long-term District and Charter enrollment growth of about 0.5%/year. Since 2020, actual growth in these programs has been negative, but total formula-funded growth has been much closer to these long-term predictions; it has simply come from the ESA program instead of traditional Public schools. Critically, taxpayers should realize that in practice they are paying for approximately the same number of students as was expected 5 years ago; the only change is what program those students are being funded through (District and Charter growth has been much slower than expected pre-pandemic, ESA growth much faster).

Going forward, because the ESA program is now saturated and demographic decline among the 5-17 year old population is likely to continue, excluding future policy changes total K-12 funding has peaked; annual baseline decreases due to enrollment declines are the probable norm. Policymakers may adjust the formula amounts to continue funding growth, however.

Who’s Using an ESA?

As of May 2025, 56.3% of all ESA recipients were living in Arizona ZIP codes with a median family income of between $75,000 and $150,000. This is down from 56.7% in January 2025. For context, 49.0% of all households with children under 18 reside in these ZIP codes; on the other hand, 35.8% of all Arizona families have household incomes between $75,000 and $150,000.

Because information about ESA recipients is released only in summary form by the Department of Education, program observers have been forced to draw inferences about the population, including their incomes and backgrounds. These inferences have been a prominent part of the political debate. The Department does not collect or report information about family income, race, or other characteristics; it does collect and report ZIP-code level geographic data, however.

The American Community Survey does report information about the incomes and demographics of Arizonans by ZIP code. However, the sample - especially the 1-year – is relatively small. The low response counts in some individual ZIP codes introduces significant risk of survey error; on the other hand, statewide averages – and in particular statewide averages from the 5-year surveys – produce much more reliable estimates about the true population’s characteristics.

This is especially true when trying to draw conclusions about households with school-aged children. In general, these households are concentrated in certain parts of the state, and those households (and ZIP codes) differ systematically from those without children. Families and households with children are higher income than households without, and most children in Arizona are found in wealthier ZIP codes.

CSI uses the best data available to us to try and draw the most likely conclusions, given these limitations and uncertainties. For ESA users, that is ZIP code data, but for Arizona families in general, that is ACS 5-year statewide averages. However, for transparency, we also now report the share of all Arizona households with school-aged children residing in ZIP codes by median income, using the latest ACS 1-year estimates; while an “apples-to-apples” comparison with our ESA methodology, we repeat our caution regarding the uncertainty and small sample size.

For further discussion of our technical assumptions and there justification, see our “2025 Quarter 1” Report from earlier this year.

CSI again notes how small the number of households involved are (and this occurs often in ZIP code-level data); any reporting of this data by ZIP that tries to “normalize” (for example, reporting x per thousand kids) conceals this scarcity issue and implies something that isn’t true (namely that there are potentially large numbers of ESA users or children in, say, Paradise Valley – a very high income ZIP code [85253] with a high ESA utilization ‘rate’ but with only 710 ESA users [0.8% of all ESA’s] and 2,728 kids total).

In general, given this uncertainty, we rely on median values for ACS 1-year and ZIP-code based analyses (which are subject to our largest possible sample skew and survey error), but also calculate and report arithmetic means when feasible and defensible (where robust sample sizes minimize potential error and skew and allow the theoretical advantage of the mean to come through).

ESA users continue to lean disproportionately middle-income (defined here to be households earning between $75,000-$150,000). While a quarter of ESA users may live in households earning less than $75,000 per year, up to 40% of all households with children earn this much. On the other hand, while we estimate less than 3% of ESA-using households earn over $200,000 per year, up to over 13% of all Arizona families earn this much.

CSI continues reporting other expected demographic characteristics of the ESA population here, using similar methodology as our approach to estimating incomes. This data relies on the same ZIP-code based methodology as our income estimates and is subject to the same measurement uncertainty. Again in general, what is striking is how similar ESA users to the overall K-12 population. We note three areas of marked difference (and will update these figures quarterly as data allows):

While over 18% of ESA users are estimated to be nonwhite, this is statistically less than for Arizona’s overall population (21.3%).
Over 12% of ESA users reside in rural ZIP Codes – higher than the 10.7% of Arizonans overall.
ESA recipients are probably more likely to reside in married-couple households (52.7%, versus 47.0% of Arizona households with children overall).

Measuring Program Efficiency

As of February 25th, 2025, the most recent Quarterly Report released by the Department of Education applied to ESA activity occurring between October and December 2024 (FY25 Q2). There has not been an update to how quickly reimbursements are being approved since March 2024, and at that time it took an average of 13 days to approve reimbursements. Marketplace transactions were approved at 88% of their expected rate in Quarter 2 of FY 2025 (an increase from 71.8% in the prior quarter).

Reporting about reimbursements and efficiency measures from the Department is inconsistent, so CSI uses a proxy measurement to estimate how efficiently the Department is processing transactions made by ESA users. Assuming ESA users were on average exhausting their scholarships, and the Department was timely approving all expenditure requests, on average and over time approved expenditures would average out to 100% of annual scholarship awards.

For the six quarters that data is available, ADE has never achieved a 100% of theoretical maximum expense approval rate. Approval rates are currently at their peak of 88.8%. On average, the Department reports approving expenses for only about 77% of awards. These trends imply growing and systemic backlogs and lengthy approval/reimbursement processes.

The decision to automatically approve reimbursement requests of less than $2,000 in December 2024 should have helped alleviate this issue; the most recent data suggests this relief may be occurring. But the most recent data covers the period of October 1, 2024 through December 31, 2024, so part of the increase in market transaction processing rates may have occurred after the decision made on December 9. Later data should confirm both that the improvement is ongoing and whether we are getting closer to the 100% target.

CSI continues to monitor administrative changes that may impact the efficiency of the ESA program. The latest draft of the 2025-2026 school year Parent Handbook included limits on certain supplemental items such as instruments, computers and technology, and physical education equipment. For these items, parents would only be reimbursed up to a certain amount by the ESA program. The Department claims “the coverage amounts help reduce delays in the approval process for allowable expenses and expedite the approval process for families”. In practice, CSI cannot identify clear statutory authority to impose these kinds of limits under A.R.S. Title 15 Chapter 19. As of this report, the draft has not been formally adopted by the State Board of Education.

As a reminder, under current Department of Education policy, staff must review and pre-approve each purchase, order, and reimbursement in the ESA program, before it can occur. In Quarter 4 of 2024 alone, there were 1,300 applications, 155,000 reimbursements, and 475,000 orders individually reviewed by Department staff. Each of these transactions are subject to the same process of review and pre-approval, regardless of amount or scope. For example, the Department requires its limited staff resources to individually review and approve 38,449 purchases of “curriculum”. The average purchase amount? Just $93.

To make the program more usable and to reduce the backlog of transactions needing approval each quarter, the Department of Education must move to a risk-based auditing approach that automatically deems most valid and complete purchase requests approved pending later review or audit, while identifying unusual, high-dollar, or high-risk transactions for added scrutiny. The Superintendent has indicated changes - consistent with this direction – intended to allow the Department staff to process all transactions in less than 30 days and reduce the 89,000 transaction backlog that Department Staff face; more recently, though, he has suggested this was a temporary measure and implied the Department may go backwards.

Bottom Line

The rapid “catch-up” growth in the ESA program has come to an end. Going forward, program growth will be much slower, and new enrollments (“switchers”) will reduce public costs elsewhere; because of universal ESA and demographic change, net-new public K-12 spending in Arizona due just to population growth is at an end.

The posture of the programs administrators since universal expansion has been defensive and driven by this rapid but temporary “catch-up” phase of growth. The result has been a program that’s slow and cumbersome for (disproportionately middle-income) participants to use. A lack of consistent and regular reporting to the public and policymakers contributes to these issues.

ESA’s in Arizona Q2 2025 Report