by John Garrett Clawson, Sarah Pollock, Katie Floyd, Rey De La Garza, & Cassie Davis. (Faculty supervisor/editor: Paul von Hippel.)

Table of Contents

Introduction

Education savings accounts (ESAs) are a school choice mechanism that withdraws some state education funds from the public school system and puts them in the hands of parents. Participating parents receive the funds in a secure account, which they can use for a variety of education services, including tuition, tutoring, and other services such as therapy for children with special needs.

The impact on the state budget depends on how much money the state withdraws from the public schools vs. how much the state spends to fund parents' education accounts. These amounts can vary according to the number of students that participate, their characteristics, and the characteristics of their school districts. 

On this Wiki page we estimate the impact on the state budget of an ESA bill proposed by Texas Legislators in the 85th Texas Legislative Session in 2017 (Senate Bill [SB] 3). We estimate that the state would expect to lose approximately $2,600 per student who enrolled in the ESA program. For example, if 20,000 students enter the program each year, it would cost the state $52 million in the first year, and double that in the first biennium.

Senate Bill 3

Since 1995, Texas Republicans have repeatedly proposed private school choice bills. In the 84^th Legislature (2015), a private school choice bill passed the Senate for the first time, though it did not receive a vote in the House. In the 85th Legislature (2017), new Lt. Governor Dan Patrick advocated for an ambitious private school choice proposal, which was carried by Senator Larry Taylor, the Senate Education Committee Chairman, as Senate Bill 3 (SB 3). SB 3 proposed two programs for private school choice: 1) ESAs for parents to use state dollars to fund their children’s education and 2) Tax Credit Scholarships to help pay for the tuition costs and other educational expenses for students wanting to attend private schools. This Wiki page focuses on the ESA component.

Senate Bill 3 received a hearing on March 21, 2017, in the Senate Education Committee. The hearing took 7 hours and 150 advocates and opponents signed up to give public testimony. Among the issues discussed was the cost of ESAs to the state treasury. As introduced, SB 3 would have required annual reports on the "net savings" due to "reduced" "state spending on public education," yet the day before the Committee meeting the Legislative Budget Board (LBB) published a fiscal note projecting that the ESA portion of SB 3 would cost the state $90-330 million over its first two years.

SB 3 passed out of Committee, then was engrossed (passed) by the full Senate on March 30. The bill went through significant changes between the version that was introduced by Sen. Taylor and the final version that was engrossed by the Senate. Several changes would reduce SB 3's cost to the state.:

• Eligible Grades: As introduced, SB 3 extended eligibility not just to children leaving public for private schools but to kindergarteners who were too young to have attended public school previously (born on or after September 1st, 2012). As the fiscal note pointed out, this meant that ESAs would be available to over 20,000 kindergarteners who never had any intention of enrolling in public school, and the cost of ESAs for those kindergarteners would not be offset by any reductions in public school spending. The cost of such "free riders" would increase over time, since in year 2 a new crop of kindergarteners would be eligible and previously eligible kindergarteners would remain eligible in first grade. To close this loophole and reduce expenses, the engrossed version of SB 3 limited eligibility to those children who attended a public school during the entire preceding academic year.

• Eligible Counties. As introduced, SB 3 extended eligibility to children throughout the state. As engrossed, SB 3 limited eligibility to children who lived in a county with a population of at least 285,000. It should be noted here that the limitation to large counties is less restrictive than it may seem at first; Texas' 18 largest counties have 82.5 percent of the state’s private schools and enroll 83 percent of the state’s private students. The main purpose of this restriction was not to reduce expenses, but to sway rural Senators who would have opposed the bill if it affected small-county constituents.

• Eligible Income Levels. As introduced, SB 3 extended eligibility to families of all incomes, although higher-income families were eligible for smaller amounts. As engrossed, SB 3 limited eligibility to families whose income was less than 175 percent of the threshold for the federal free-or-reduced-price lunch program. It is not clear how many interested families this restriction affects. A rough estimate, based on U.S. Census data from 2015 and assuming a household size of three for the calculation of income eligibility based on free-and-reduced-price lunch guidelines, shows that about half of Texas schoolchildren have family incomes above the threshold, but high-income children are more likely to attend private schools.

As engrossed, SB 3 offered different children different ESA amounts. Children with disabilities would receive "90 percent of the state average maintenance and operations (M&O) expenditures per student for the preceding state fiscal year," or $8,159 in year 1. Other children whose families were below the income limit would receive "90 percent of the state average maintenance and operations expenditures per student for the preceding state fiscal year," or $6,799 in year 1.

Table 1: ESA Amounts by Income Level 

Estimating Program Costs

Because the House has not taken up SB 3, no fiscal note has been published on the version that was engrossed by the Senate. Here we try to estimate the cost of the engrossed version to the state. 

Definitions

Because the amount disbursed to an ESA card is less than current M&O costs, it might appear as if the program would have no net cost, and might even have a benefit. But the program can have a net cost to the state, because ESAs are funded entirely by state dollars, whereas the state is only responsible for part of district spending.

To model the program's fiscal impact, define the Total Annual Net Cost as the total annual gross cost–the amount disbursed to ESAs–minus the amount that the state would have spent on the ESA participants in public schools. In districts where the state recaptures excess local property taxes, we should also subtract any increase in recapture funds that the state receives because of reductions in district enrollment.

• Net Cost = Gross Cost - (Reduction in State Support + Increase in Recaptured Funds)

Each district with wealth greater than is necessary to fund the education of its students at the state-assessed level is subject to recapture. Districts subject to recapture send money back to the state in the form of "attendance credits," whose costs vary (a little) depending on how "deep" a district is into recapture -- i.e. how much a district's wealth exceeds the threshold for the state to begin recapturing funds. To prepare for our analysis, we submitted a public information request for the attendance credit values of the districts in recapture. We also used state support levels that were publicly available from TEA. 

Program Modeling

To model an ESA program's budget impact, we must assume something about the characteristics of participating districts and students. A number of assumptions are possible. We illustrate the calculations using the assumptions below.

• According to the engrossed version of SB 3, eligibility is restricted to the 17 counties that had more than 285,000 residents in 2010. In addition, ESAs are only usable in districts that have private schools. Only 112 districts (out of 1,247 statewide) have private schools and are in one of the 17 largest counties.

• We use the assumptions made by LBB about characteristics of the students most likely to participate in the ESA program under the introduced version of SB 3. While the engrossed version was somewhat different, we determined that 90 percent of the students participating are likely to have an income at or below 175 percent of the FRPL program level and 10 percent are likely to have an eligible disability. Based on these figures, we expect the average ESA award total to equal $7,440, which is a weighted average of the two award levels (=90%*$8,159+10%*$6,799). We assume this weighted average is the same for every district. It isn't, but since the two award levels are not very different, small changes to the weights would have little effect on the weighted average.

• We assume that the districts in recapture, and their attendance credit costs, will remain the same as today. This is a reasonable approximation if relatively few students use ESAs. If many students use ESAs, it is possible that additional districts will be driven into recapture, or that districts will avoid recapture by, e.g., selling off property to reduce total property wealth. These are potential long-term consequences, which we neglect in our short-term analysis.

In each district, the program's average net cost is determined by the following formula. 

• Cost/Savings to State = (State revenue sent to district) + (value of attendance credit) - (ESA award amount)

The net cost in each of the 112 districts is tabled at the bottom of this page. The state would save money in just 8 of these districts, and lose money in the other 104. On average, the state would lose about $2,500 per participating student. There are several ways to calculate the average. We can calculate an unweighted average, in which a small district and a large student get equal weight, but this is unrealistic since clearly more students will participate from larger districts. As an alternative, we can calculate a weighted average, giving more weight to districts with higher public or private enrollments. In the end, we get about the same result no matter how the average is weighted.

Projecting the number of students that will ultimately utilize the program proves difficult. The Texas Education Agency believes that approximately 25,000 additional students would attend private schools because of the program every year (approximately 10 percent of current private school capacity in Texas). Using these assumptions, 25,000 students would utilize the program in the first year of the program. If this projection is accurate, the total number of participants would grow to approximately 100,000 by 2022 (assuming the program was implemented in 2019), though this seems high since only 250,000 Texas students are enrolled in private schools today. Using per student cost estimates produced by our analysis, if ESA participation grew by 25,000 students per year, the state could expect to spend roughly $65 million in 2019 and $262 million in 2022.  

Below is a table summarizing relevant information for eligible districts with private schools.