A Truly Universal Education Savings Account Proposal, Including Fiscal Implications


Byron Schlomach, Vance H. Fried


This paper proposes a $4,500 ESA for every school-age child in Oklahoma that parents can supplement. It recommends specifics of what purchases are allowable under an ESA. It discusses how the program would be funded, how it would be administered, and academic accountability.

Full Text HTML

A Truly Universal Education Savings Account Proposal, Including Fiscal Implications

Byron Schlomach and Vance H. Fried

Education Savings Accounts (ESAs) are state-funded and administered accounts to fund educational services for school-age children under the direction of the child’s parents or legal guardians. The funds can be used for a variety of educational purposes such as private school tuition, tutoring, books, and online materials. Students can participate only if they contractually agree to not attend a traditional public school (i.e., participants cannot double-dip on taxpayers’ funds). An incentive to economize is provided by allowing funds remaining in an ESA account upon graduation to be used for post-secondary (college and career) education.

The 1889 Institute has published Designing an Education Savings Account 2016 as well as Education Savings Accounts and Improving Oklahoma Student Achievement, which explain the ESA concept in some detail. Both are available on the 1889 Institute’s website.[1]

Here, we propose specific parameters for an ESA law for the state of Oklahoma, and discuss its financial ramifications for the State and for local school districts.

Basics of an ESA Law


Any school-age child, kindergarten and above, eligible to attend an Oklahoma public school is eligible for an ESA. [OPTION: Make children from very small districts ineligible. More on this below.] Only children whose parents sign a contract promising not to enroll a child receiving an ESA in full-time public school may receive an ESA. A child who accepts an ESA and who originates from an Oklahoma public school shall be immediately removed from the weighted average daily attendance (WADA) count of the school district of origin.

Most other states’ ESA programs have been limited to benefit only special student populations. Only Nevada has what is considered a universal ESA bill, but even Nevada’s program is not truly universal. A child must have attended a Nevada public school or charter school for at least 100 days immediately prior to receiving an ESA.[2] The proposal presented here envisions true universality wherein all of Oklahoma’s public school-eligible children can qualify.

Fairness is the reason for including children currently attending private schools. The State and its school districts are legally obligated to serve all comers. So, if every privately schooled child in Oklahoma suddenly showed up at the doors of the state’s public schools, the school district must provide every one of them the same education they are providing their existing students. However, the parents of these children have relieved the state of the significant financial burden of educating their children. They are willing to bear the burden for the State because they want their children to get a better education than they feel their district school provides. These parents should not be financially punished under an ESA because of their willingness to make financial sacrifices for their children. From a fairness standpoint, students currently attending private schools should be included. However, for short-term financial reasons, these students may need to be phased into the system. (See Fiscal Note discussion below.)

The logic for the optional language on small districts is explained in Fiscal Impact to a District School.

Due to a lack of evidence that early childhood education has lasting effects, and the partial-day nature of early childhood education, the program should be limited to children kindergarten and above.[3]

The contract language is a standard provision in ESA laws. It helps to ensure financial stability of the system and simplifies funding by avoiding refunds and constant transfers.

The provision on reducing WADA count is so that the state will not reimburse a school district for phantom students.


Individual ESAs will be entirely state funded at $4,500 per year from the state’s general appropriation for formula funding of common schools. Parents may supplement this amount with their own money. The program administrator has authority to determine when to distribute funds. Deadlines will be set by rule for election by parents to receive ESAs so that public education formula funding can be calculated in a timely manner.

At $4,500 the state will save a significant amount without reducing per student funding to district schools since $4,500 is much less than the state spends on its district schools. (See Fiscal Note and Fiscal Impact on a District School.) As Table 1 shows, Oklahoma public school districts spent an average of $9,600 per student during the 2014-2015 school year. Most of this money comes from either the State or local district property taxes. Some was spent on debt service and other fixed costs that change little with enrollment. However, the bulk of education spending is variable. That is, as enrollment fluctuates, so does the amount of money needed by schools. Adding instruction, support services, transportation, and 60 percent of physical plant per-student expenses as variable costs, amounts to $7,500, about 78 percent of total per-student spending.[4] Over time, even debt service is variable, but it and “miscellaneous” are both ignored, as is Child Nutrition (largely federally funded). Thus, $7,500 is the approximate variable cost per student in a district school.


Funding ESAs at $4,500 per student results in a net savings of approximately $3,000 per student ($7,500 – $4,500) when a child leaves public schools for the ESA. There are additional savings from reduced teacher retirement and health care costs that are not reflected in Table 1 because some expenses are paid directly by the state rather than following through the school district. Health insurance paid directly by the state is about $560 per student and state retirement contributions amount to almost $460.[6] However, this additional $1,020 in annual savings does not occur immediately in the year that the student leaves a district school for an ESA.

While few private schools price as low as $4,500, this amount is enough that even parents with modest means can afford to send their children to most private schools. Further, virtual school tuition is well-below $4,500. The state might provide a higher ESA amount for students from low income families. This would complicate some of the financial aspects discussed below, resulting in less savings, but is achievable with minimal additional operational complexity.

Parents would have to decide whether to participate in advance of a school year so that the appropriate amount of ESA funding could be deducted from the state’s general appropriation for common schools. Once that deduction is made (by multiplying the number of ESA students by $4,500) formula funding can be calculated with the remainder. This likely early summer decision, with the deadline determined by rule, also gives the district school sufficient time to match staffing to enrollment.

A Note on Federal Funding

President-elect Trump has proposed using $20 billion for a federal school choice program.[7] The federal government already spends over $60 billion on public education each year. Whether the $20 billion is in addition to this amount or is to be taken from current funding (e.g. school lunch funding follows the low- income student to a private school) is unknown. Assuming Oklahoma got a one percent share of that $20 billion, the state would get $200 million for use in education choice programs. As it stands, too little is known about what Trump proposes to speculate how this funding might work with an ESA program. Nevertheless, it appears that some form of federal support is likely.


Accounts and account activity shall be handled by the State Treasurer. The Treasurer shall contract with a private vendor to create a real-time closed vendor payment system that automatically rejects ineligible purchases. The Oklahoma Department of Education (ODE) shall cooperate with the Treasurer and ESA Board in determining student status and enrollment in order to enforce the provisions of the ESA law.

A five-person ESA Board to administer the program shall be appointed as follows: one member by the President of the Senate, one member by the Speaker of the House, one member by the State Treasurer, one member by the Lieutenant Governor, and one member by the Governor. The members appointed by the Speaker of the House and President of the Senate shall be parents of ESA students. No member of the board may be: 1) an education material and/or service vendor serving ESA students, 2) an employee of the State, or 3) an employee of a public school.

The Board shall:

1) Coordinate with the State Treasurer to determine rules for the proper distribution of funds,

2) Settle any disputes that might arise concerning a child’s eligibility,

3) Determine what constitutes adequate progress as measured by norm-referenced tests,

4) Determine deadline dates for ESA application, in coordination with the Oklahoma Department of Education,

5) Approve items for inclusion as allowable expenses,

6) Maintain a vendors’ registry, and

7) Hear appeals regarding allowable ESA purchases.

Board staff requirements shall be filled by the Board.

As envisioned here, the Treasurer would administer the ESA program, the ESA Board would provide policy guidance, and the ODE would provide information for which it has developed expertise in collecting.

The Treasurer has expertise in objectively controlling and disbursing funds with proper accounting controls. While the Treasurer is well-suited to handle financial administration, there also needs to be a body designed to make decisions surrounding implementation and accountability.

This body should be independent of the State Board of Education and State Superintendent. Experience in other states has demonstrated that state departments of education are not organizationally or temperamentally suited to running an ESA program. Their focus on the public education system and bureaucratic processes prevent the streamlining of systems and there is often hostility toward school choice programs. Nevertheless, the ODE has established protocols for identifying students with unique identifiers and has the ability to match students with student-specific data. Recreating a similar system would be an unnecessary cost, so ODE should retain some purely administrative functions when it comes to ESA students.

Allowable ESA Expenses

1) Private school tuition,

2) Tutoring,

3) Educational materials (both print and online),

4) Educational and extra-curricular services voluntarily offered by public schools,

5) Traditional co-curricular and extra-curricular activities such as art, athletics, drama, music, student clubs, and other academic-related activities,

6) Educational therapies (often used by special needs students),

7) College and career tuition for concurrent enrollment students,

8) Testing,

9) Post-secondary education (up to an age to be determined, at which time any remaining funds revert to the state).

Parents shall be allowed to allocate funds among these eligible expenses in any way they see fit and buy from registered vendors. Any funds left over at the end of a school year shall carry over into the next year. Funds can be used for post-secondary education (college or technical), including AFTER graduation when the student is no longer eligible to attend public school.

The list here is fairly standard for ESA programs already in place in other states. Part of the administrative and implementation cost will be creating systems that disallow purchases of items and services not on the list while allowing for timeliness in directing purchases by parents. Creating an approved vendors list saves time and administrative cost in addition to making financial accountability considerably simpler.[8]

Academic Accountability

A student’s parents must have the student tested yearly. The test shall be chosen by the parents from a list of nationally-recognized norm-referenced tests selected by the ESA Board. Once chosen, the same test must be used for a minimum of three consecutive years. Results must be reported to the Oklahoma Department of Education (ODE) which will compile the information and share fully with the ESA Board.

Failure on the part of a child to make adequate academic progress, as determined by the ESA Board, failure to submit to testing, or failure to report test results shall result in revocation of a child’s ESA and any unspent account balance.

There should be no need to completely re-create the resources needed to keep account of test results; thusly, the use of ODE resources where expertise and required computerized tools already exist. ODE’s role would be purely administrative record-keeping and parents could independently receive test scores for their ESA-participating children. Academic independence is maintained through the use of a menu of existing norm-referenced tests with a specific instrument chosen by parents.

Opponents of school choice often complain that there is too little accountability in school choice programs while proponents argue that restrictive testing requirements limit flexibility and the ability to find the best fit for a child. This proposal gives parents adequate flexibility and simultaneously insulates the taxpayer from risk by placing the onus of responsibility for a child’s education where it should be – on the parents.

Financial Accountability

The Board shall provide a registry for bonded vendors, including specifying the level of bonding required. The Board shall also promulgate a list of allowable expense items and services based on the basic list in law.

Parents shall not be allowed to resell items except after they have been used and then only through approved channels. Proceeds from such sale shall accrue to a child’s ESA account. Proceeds from the return of unused items shall accrue to the child’s ESA account.

Parents and vendors shall be prohibited from kickback arrangements that, if discovered, shall be prosecuted as theft.

Relatives of first and second degree cannot be paid out of a student’s ESA for their services (e.g., tutoring). Exceptions can be made by appeal to the ESA Board.

The child of a parent who is convicted of fraud or theft arising from this program shall retain any account balance with allowed expenditures to be directed by the recipient upon attaining the age of 18.

Most of this is self-explanatory. The ESA board should provide resources to parents so that they can be sure to comply with the specifics and spirit of the law without having to interpret the law themselves. The bottom line is to make sure expenditures from an ESA are for legitimate education expenses of the child an ESA is intended to benefit.

Administrative Cost

The Treasurer may deduct a maximum of 3 percent from each ESA in order to cover ongoing administrative costs of state agencies. Further, the Treasurer may borrow up to $350,000 from the state in order to finance administrative startup costs within Treasury and ODE.

This solution is based on Nevada’s approach. Nevada chose a higher-cost disbursement model than Arizona, the first ESA state, in order to better prevent fraud. Nevada’s experience with Benefit Wallet (a Xerox subsidiary), online parent portal and vendor-provided payment system (system created by a vendor as contracted by the Treasurer to make immediate transactions that are compliant with the law) make it likely that startup costs will be minimal, assuming Oklahoma chooses a system similar to Nevada’s. Over time, the administrative cost per student will drop with an increase in the ESA population and increased administrative expertise.[9]

Fiscal Note

Administrative costs do not need to be addressed if they are handled in a manner as recommended above.

Scenario 1: Every private school child receives an ESA, 5,000 home schooled children participate, and no public school children participate.

Approximately 30,000 children attend private schools in Oklahoma. It is likely most would participate in the ESA program. It is estimated that nearly as many children are home schooled in the state. However, relatively few would participate, at least initially, given the reluctance of home school parents to participate in choice programs. If all current private school children and 5,000 home schooled children participate while no current public school children participate, net general appropriations to the public education system would decline by:

<$157,500,000> = 35,000 x $4,500.

This is 8.4 percent of the state’s $1.87 billion general appropriation to public schools in 2016, 5.6 percent of $2.8 billion in total state funding to public schools, and 3 percent of the all-funds revenues for public schools.[10]

Alternatively, the state could appropriate an additional $157.5 million in order to cover the additional cost and not rely on existing formula funding to cover the cost.

Scenario 2: 35,000 non-public-school children receive an ESA and 180,000 public school children (just over a quarter of all current public school children) participate.

As noted above, each child in Oklahoma’s common schools represents an average variable cost of about $7,500. While these funds technically come from both state and local sources, as explained in the 1889 Institute’s A Primer for Understanding Oklahoma’s School Funding System, it can be accurately argued that the entire funding system, controlled by the state, funds each student. Thus, every child who opts out of the public education system in favor of receiving an ESA represents a net savings of $3,000, a savings that can be accurately said to accrue to the state.

Net “cost” calculation under Scenario 2:

Cost of 35,000 non-public-school children participating:

$157,500,000 = 35,000 x $4,500.

Cost of 180,000 public school children participating:

$810,000,000 = 180,000 x $4,500.

Savings to common schools from 180,000 public school children participating:

$1,350,000,000 = 180,000 x $7,500.

Net savings of Scenario 2:

$382,500,000 = $1,350,000,000 – $157,500,000 – $810,000,000.

This does not include savings for state costs that do not run through the district. Assuming $1,020 per student in teacher retirement and health insurance costs, there would be an additional $183,600,000 savings, resulting in a total net savings of $566,100,000, or a little less than 8 percent of the state’s general fund budget.

No Net-Cost ESA

Many argue the state cannot afford additional expenses at the present time, so it is imperative to keep any new programs from costing the state. As illustrated by Scenario 2, an ESA program potentially means a great deal of savings for the state. However, in the short term, the greatest demand for ESAs is likely to be from children already in private schools for three reasons. First, parents of private school children have an obvious pecuniary interest in participating. Second, it will take time for knowledge about ESAs to become wide-spread among public school parents. Finally, in the short-term private schools may not have enough excess capacity to absorb every new student wanting to enroll. We do not have numbers on the amount of private school excess capacity, but assuming private schools are currently at 80 percent of their capacity, that means there are only 7,500 slots available state-wide for new students.

Over time, private school capacity will expand in response to increased demand caused by ESAs. Further, virtual schools and home schooling do not face capacity constraints. Nevertheless, on balance, for the first few years a lack of private school capacity will likely be a significant problem.

Since ESA participants who leave the common school system create a net savings, an ESA program would have no net costs if the ratio of ESA participants from private schools (PSP) to ESA participants from common schools (CSP) is low enough. As it happens, three ESA participants from the common schools, results in a net savings of $9,000. This savings is sufficient to fund two ESA participants from private schools.

A truly universal ESA law could include a provision that restricts the number of ESAs for children who originate from private schools, determining a PSP/CSP ratio that results in no net cost. This would require setting the deadline date for parents of public school children to elect to participate early enough so that net savings could be calculated and a lottery held to determine which private school children can participate. A lottery will not be required in the event that the number of applicants from private schools compared to those from public schools is sufficiently low.

A provision restricting the PSP/CSP ratio should have a definite expiration date. This will encourage private schools to build additional capacity, and for new private schools to emerge in anticipation of increased demand. A date-certain allows for planning and greater confidence. It also encourages public schools to respond in a healthy way, by improving their educational product, rather than putting their hope in lobbying to continue to restrict competition.

No-Net-Cost ESA Provision

For a period of five years after the effective date of an ESA law, the ratio of Education Savings Account recipients who attended private school before receiving an ESA to ESA recipients who attended public school before receiving an ESA shall be determined by the ESA Board.

The Board, in consultation with the Oklahoma Department of Education, shall:

  1. Set a deadline for parents of children attending common schools to elect to participate for the upcoming school year,
  2. Based on the number of common school students who will receive an ESA and the dollar amounts of their respective ESAs, determine the maximum number of students from private schools who can receive an ESA for the school year,
  3. Conduct a lottery to determine which students from private schools will receive an ESA if the number of applicants exceeds the number determined under 2) above.

Revised Fiscal Note

Scenario 3: The PSP/CSP ratio does not exceed 2/3.

Scenario 3A:

Suppose 4,000 students from Oklahoma’s common schools opt to receive an ESA. Then, up to 2,666 students of Oklahoma’s non-public-school students can participate at no net cost.

“Cost” calculation under Scenario 3 with 4,000 common school participants:

Cost of 2,666 private school children participating:

$11,997,000 = 2,666 x $4,500.

Cost of 4,000 public school children participating:

$18,000,000 = 4,000 x $4,500.

Savings to common schools from 4,000 public school children participating:

$30,000,000 = 4,000 x $7,500.

Net savings of Scenario 3 with 4,000 common school participants:

$3,000 = $30,000,000 – $11,997,000 – $18,000,000.

Additional savings from reduced retirement and health insurance contributions of $1,020 per student would result in total savings of $4,083,000.

Scenario 3B:

Suppose 45,000 common school students participated, allowing all 30,000 private school students to participate.

Net calculation under Scenario 3 with 45,000 common school participants:

Cost of 30,000 private school children participating:

$135,000,000 = 30,000 x $4,500.

Cost of 45,000 public school children participating:

$202,500,000 = 45,000 x $4,500.

Savings to common schools from 45,000 public school children participating:

$337,500,000 = 45,000 x $7,500.

Net balance under Scenario 3 with 15,000 common school participants:

$0 = $337,500,000 – $202,500,000 – $135,000,000.

Additional savings from reduced retirement and health insurance contributions of $1,020 per student would result in total savings of $45,900,000.

Fiscal Impact to a District School

Revenue Impact

School district revenue can be broken into four basic pieces. First are revenues used to derive formula funding – dedicated state revenues, state appropriations, and some local property taxes. Second are some property tax revenues that go to the school district and are not used in formula funding calculations. Third is property tax revenue dedicated to servicing any bonds issued by the district to fund capital expenditures. Finally, districts receive various federal grants and federal funding for the free/reduced price lunch program.

Assuming a district is not self-funded and does receive state money for formula funding (which is the case for the vast majority of children attending Oklahoma’s common schools), the following happens when a child leaves a district:

  1. The district loses the full amount of formula funding for that child,
  2. The district retains all local property taxes not included in the funding formula calculation, and
  3. The district retains all revenues dedicated to bond service, with no loss to formula funding.

The net result is that the total district funding decreases on an absolute basis, but will likely increase on a per student basis. Savings are also likely to accrue to the state as a whole from students leaving the public system sufficient to slightly increase per-student funding for those students remaining in the public system once the state recalculates formula funding factors. If the number of private school children who opt for an ESA fails to meet the maximum allowed under a no-net-cost provision, the amount of funding per-student in the public system will likely increase.

If a large number of children from self-funded districts opt for an ESA, the amount of formula funding per student in the public schools will be decreased since these students would represent a pure cost to the state, similar to previously privately-schooled children. The share of funding to public schools represented by local revenues will increase as children opt for ESAs, but the total funding per student, especially taking account of bond revenues and local revenues not included in the formula, will likely increase due to the small number of children in self-funded districts, which also tend to be small districts.

About 21,000 students attend self-funded and partially self-funded school districts. While it is theoretically possible that 5,000 of these students would participate, it is highly unlikely. Nevertheless, if 5,000 did participate, the $22,500,000 in cost represents less than $33 per public school student.

Expense Impact

When only one student leaves a district, the district loses all of the formula funding for that student, but experiences little in the way of operating savings since a single student leaving is unlikely to result in a reduction of any personnel or facilities, so profits (revenues over cost) decline by the amount of lost revenue. Conversely, the revenue from adding a single student is almost all profit. However, subtracting several students means costs can be reduced. Losing students is rarely a problem for most districts since the absolute number of students enrolled is high. As a result, medium and large schools cost per student can be adjusted proportionally downward when students are lost, but this may not be true of a small school.

For example, a medium-size school district might have 300 first grade students in 10 classes, but a small school has 30 first graders in one class. If both lose 10 percent of their students to an ESA, the medium district has lost 30 students, but can easily deal with this problem by going to 9 classes of 30. They save money because they need one less teacher.

Indeed, medium and small districts are constantly changing staffing to match increases and decreases in enrollment. However, the small school may not have this option. When our hypothetical small school loses 3 students, it will have one class of 27 students and cannot save money by hiring fewer teachers. As a result, losing students can have a negative financial impact on very small schools – revenues decline but costs do not. One possible solution to this problem is to exempt small districts from an ESA program. Of course, doing so would make the program no longer universal.


  1. Schlomach, Byron and Vance H. Fried, Designing an Education Savings Account 2016 (Oklahoma City, OK: 1889 Institute, January 2016), http://nebula.wsimg.com/3e65befa20f16820263db580741ff62f?AccessKeyId=CB55D82B5028ABD8BF94&disposition=0&alloworigin=1.Butcher, Jonathan, Education Savings Accounts and Improving Oklahoma Student Achievement (Oklahoma City, OK: 1889 Institute, February 2015), http://nebula.wsimg.com/83e2e0f3e69541338ffa7e47d7371773?AccessKeyId=CB55D82B5028ABD8BF94&disposition=0&alloworigin=1.
  2. “Education Savings Account,” website, Nevada State Treasurer, http://www.nevadatreasurer.gov/SchoolChoice/Home/.
  3. Richwine, Jason, “When Social Programs Hurt Kids,” National Review, November 21, 2016, http://www.nationalreview.com/corner/442357/preschool-study-government-social-programs-hurting-kids.
  4. In the very short or immediate run, there are no variable or marginal costs. Everything is fixed. If one student leaves a school, in the vast majority of circumstances, there are no personnel or operations/maintenance savings. On the other hand, if a new student attends it is arguable there are no added costs. In both cases, however, formula funding changes significantly. Some might say that there are no operations/maintenance savings when a child leaves a school. However, the likelihood is that significant numbers of children will leave a given public school under a choice program and, in fact, there will be significant savings in operations/maintenance on an average per-student basis statewide. The 60 percent threshold chosen for operations/maintenance savings is somewhat arbitrary, but the bulk of all school costs in all areas of expense are in labor and as public schools depopulate, fewer personnel will be needed in all areas of school operations.
  5. Author calculations based on ” State Expenditure Details Based On All Funds Submitted By Districts,” website, Oklahoma State Department of Education, https://sdeweb01.sde.ok.gov/OCAS_Reporting/StateExpenditureDetails.aspx.
  6. Figures derived from the total amount appropriated for the Flexible Benefit Allowance and the total amount of state/federal contributions to Oklahoma’s Teacher Retirement system as reported, respectively, by:Financial Services Division, Technical Assistance Document (Oklahoma City, OK: Oklahoma State Department of Education, February 2016), 1, http://sde.ok.gov/sde/sites/ok.gov.sde/files/documents/files/FY%202016%20TAD.pdf, andFinance Department of the Oklahoma Teachers Retirement System, Comprehensive Annual Financial ReportFor the Fiscal Year Ended June 30, 2016 (Oklahoma City, OK: Oklahoma Teachers Retirement System, December 16, 2016), 125, https://www.ok.gov/TRS/documents/2016%20CAFR.pdf.
  7. Carey, Kevin, “Why Trump’s Education Pick Won’t Be Able to Privatize U.S. Schools, New York Times, November 23, 2016, http://www.nytimes.com/2016/11/23/upshot/why-donald-trumps-education-pick-would-face-barriers-for-vouchers.html.
  8. Schlomach and Fried, Designing an Education Savings Account 2016.
  9. Peshek, Adam, Director of Education Choice for the Foundation for Excellence in Education, telephone interview, December 5, 2017.
  10. Author calculations using Financial Services Division, Technical Assistance Document (Oklahoma City, OK: Oklahoma State Department of Education, February 2016), 1, http://sde.ok.gov/sde/sites/ok.gov.sde/files/documents/files/FY%202016%20TAD.pdf.