EdChoice Expansion Scholarship

The mechanism question is the one that determines whether districts lose money. Under the original EdChoice design, a scholarship was funded by deducting from the resident district’s foundation payment — the district lost the student and the money, and if the student had never attended the district, the district lost money for a student it had never educated. That last case is the core of the districts’ complaint. verified Under the Fair School Funding Plan it is not a deduction — established negatively and exhaustively against the FY2027 calculator, and pinned by a test. See mechanism below, which is where the confirmation and its remaining open edge live. verified

The program was held unconstitutional by a Franklin County trial court in June 2025, on Ohio constitutional grounds distinct from the federal question Zelman settled. That ruling is under appeal. verified

There are no income bands — from FY2025. The award is a continuous function of family income, and both sections that compute it have been in the committed statute extract since it was wired. For FY2024 there were eight, prescribed as a table by the same act that made the program universal, so a flat claim that the award has never been banded is false of the program’s first universal year. verified See the findings.

Participation counts by resident district are still not held. open Payments by fiscal period now have a bound rather than a blank, but not an appropriation, and that holds at both stages of the bill: LSC’s analyses of the proposal and of the enacted act each say the budget allocates no amount to any individual scholarship program, funding all five from one earmark. A figure for this program by fiscal year is therefore LSC’s estimate, twice over, and nothing states the appropriation. verified

What this repository computed # ↑ Contents

Not what Ohio publishes. Figures derived here from committed fixtures, each one citing the test that reproduces it.

Two advocacy scorecards rank Ohio on this program, and neither measures what this corpus measures. ALEC’s Index of State Education Freedom puts Ohio 10th of 50 at 63.79 points, grade B+, on programs in law as of June 1, 2024; the Heritage Foundation’s Education Freedom Report Card put Ohio 12th of 51 on education choice and 29th overall in its 2023 edition. verified Both are published by organizations that advocate for the channel they score, and neither rank is evidence about adequacy — ALEC’s index has no spending category at all, and Heritage’s return-on-investment category, where Ohio ranks 47th, treats lower per-pupil spending as better in terms. What the two establish jointly is that this channel is large and lightly conditioned by the standards of its own advocates. That is a fact about the channel, not about whether Ohio’s districts are adequately funded, and the ranks are two editions two years apart across the universal expansion — so they are not a series and differencing them would invent one. inference

The voucher ceiling is pegged to the public formula’s own base cost # ↑ Contents

R.C. 3317.022 does not fix $5,500 and $7,500. It directs both to “increase in future fiscal years by the same percentage that the statewide average base cost per pupil increases in future fiscal years”. verified crates/project Three amounts in the section carry that clause — these two and the Jon Peterson base — and the two $34,000 special-needs ceilings carry none.

So raising base cost raises every EdChoice and pilot-project ceiling with it, automatically and without a vote on the scholarship channel. The lever is a narrow one: it is the statewide average base cost per pupil of R.C. 3317.018(A), so an input-year refresh that lifts that average carries the voucher ceiling up with it, while the phase-in does not — the phase-in interpolates a district’s funding toward the computed amount and leaves base cost per pupil where it is. inference The two channels are not independent parameters of one budget; one is indexed to a term inside the other.

That is a different relationship from the one the deduction question asks about. A deduction would make the scholarship a transfer out of a district’s payment; this makes the scholarship’s ceiling a function of what districts are computed to need. The corpus established the first is absent under the plan; the second is present and was not recorded anywhere.

The first universal year was a table, and only that year # ↑ Contents

H.B. 33 did not switch the program to the income curve when it made it universal. For FY2024 it prescribed eight bands as a schedule, and the logarithmic function applies only “beginning in FY 2025”. verified crates/project

family income        K-8 award    9-12 award
at or below 450%     $6,165       $8,407
451% to 500%         $5,200       $7,050
501% to 550%         $3,650       $5,000
551% to 600%         $2,600       $3,550
601% to 650%         $1,850       $2,500
651% to 700%         $1,300       $1,750
701% to 750%           $900       $1,250
above 750%             $650         $950

The table is not a discretization of the curve — it is more generous than it at every band edge. Evaluated at each band’s upper bound against the FY2025 formula on the same base amounts, the prescribed award is higher in all fourteen comparisons. The floor differs too: the table’s $650 for K-8 sits $33.50 above the tenth-of-base the statute pays from FY2025. verified crates/project So the year the program’s participation tripled was paid on a schedule no reading of the current section reproduces, and any average award for FY2024 must be read against this table rather than against the curve.

LSC’s own arithmetic checks the crate’s reading of the statute twice, and turns a derivation into a reading. The greenbook works two points of the FY2025 formula independently — a family at 550% of poverty receiving 50% of the base amount, one at 650% receiving 25% — and both reproduce exactly against expansion_award. It then states the income at which the floor binds as 783%, where this node had derived 782.19% and could only mark it as inferred, because “the statute gives the floor as an amount and never says where it binds”. LSC says where it binds, and its figure is this one rounded up to the whole per cent. verified crates/project

And the published average is not evidence about the ceiling # ↑ Contents

Every award is the lesser of the ceiling and the school’s base tuition net of discounts, so the 2025 report’s $4,958.41 average for this program is bounded above by a mix of two indexed ceilings and by tuition, and cannot be read as saying which of the three binds. inference A reader wanting to know how much of the ceiling is used needs tuition by school, which the corpus does not hold.

But it does establish that at least one of them binds # ↑ Contents

R.C. 3310.08(A)(2) makes this program’s base amount the same schedule the traditional program pays — $5,500 and $7,500 by grade band — so the smaller of the two is a floor on any full-award average whatever the grade mix. In FY2025 those figures are the operative ones without indexing, because H.B. 33 froze the statewide average base cost per pupil the ceiling is indexed to at its FY2024 level, and a ceiling indexed to a frozen number does not move. verified crates/project

The published average is $541.59 below that floor, which no grade mix can account for — $54,667,553.01 across the program’s recipients that the schedule directs and the year did not pay. verified crates/project It holds on the report’s other average too, at $617.18, and it is made harder rather than easier by the students still paid the undecayed amount under R.C. 3317.022(A)(10)(a)(ii)(I): part of the program is on the flat schedule, so whatever reduces the rest reduces it further.

What the gap against the traditional program cannot settle # ↑ Contents

Traditional EdChoice averaged $6,808.75 against this program’s $4,958.41, and the 27.2% gap looks like a measurement of the income decay. It is not: three quantities move an average award — grade mix, the decay, and a tuition below the ceiling — and two published averages cannot separate three unknowns. The two populations also push opposite ways, since this program is universal and therefore richer, while the traditional program is triggered by a building’s performance and plausibly sends its recipients to cheaper schools. inference

Attributing the whole gap to income — the largest the decay can be — puts the mean factor at 0.72824 of the base amount, which the statute pays at 4.95751 times the federal poverty guidelines; because the decay is convex, that is a floor on the mean family income under that reading rather than an estimate of it. verified crates/project

Any one of three columns would close this and none is held: participation split by grade band, the report’s published maximum award per program, or the income distribution of recipients. The report links a per-district breakdown and the route returns 404. open

Payments by fiscal period are bounded, and the bound is an estimate rather than an appropriation. Table 5 of LSC’s FY2026-27 redbook estimates this program’s payments at $475.4 million in FY2025, $521.2 million in FY2026 and $563.5 million in FY2027 — the largest of the five programs in every year, and 45% of the channel’s estimated $1,251.1 million in FY2027. verified crates/project, the_points_that_bound_the_participation_hole Two things stop that from answering the appropriation question this node had open. The redbook says in terms that the proposal “does not allocate specific amounts to each scholarship program”, so the table is LSC’s arithmetic over a lump sum and not a schedule anybody enacted; and the table does not quite close on itself — its FY2025 column sums $0.1 million short of its own printed total, and its FY2025 estimate for the traditional program, $292.7 million, sits above the $283.1 million the annual report went on to report actually spent. verified

The enactment does not change that, and says so in the same words. LSC estimated the same table twice: Table 5 for the executive proposal, and Table 3 of the enacted greenbook for the act that passed. The enacted analysis repeats the proposal’s sentence verbatim — the budget “does not allocate specific amounts to each scholarship program”, funding all five within an earmark from GRF line item 200550 that also pays districts, community schools and STEM schools. So there is no appropriation by program and fiscal period at either stage of the bill, and that is now a finding about both editions rather than a gap in what this corpus has read. verified crates/project

What the second edition does say is what the enactment moved. Across five programs and three years, three do not move by a cent — the two EdChoice limbs and Cleveland — and the only two that do are Autism (+$7.7m in FY2026, +$8.2m in FY2027) and Jon Peterson (+$5.1m, +$5.5m), lifting the channel to $1,264.9 million in FY2027. LSC gives the reason outright: the maxima for most programs rise with the statewide average base cost per pupil, the act holds that quantity at its FY2024 value through the biennium, and so the EdChoice and Cleveland maxima are held with it while Autism’s and Jon Peterson’s are raised by hand. This program’s estimate is therefore frozen between the two editions for a reason external to it: it is indexed to a parameter the act chose not to move. verified crates/project

The channel’s participation is bounded the same way and by figures that do not share a denominator. The redbook puts about 87,000 FTE students on this program in FY2024 against the annual report’s 100,939 students in 2024-25; LSC’s FY2014 figure is 992 kindergarteners, the only grade the expansion then covered, and its FY2019 figure is over 10,800 in grades K-5. verified Four figures, four populations, and none of them is the count this node wants. The enacted table adds a fifth: a student column the redbook’s never carried, estimating 113,962 on this program in FY2027 against 100,011 in FY2025. It is forward-looking and it is LSC’s, so it bounds the channel’s expected size rather than measuring anyone; what makes it worth holding is that its column closes on its own printed total in all three years where the dollar column beside it does not. verified crates/project The whole census of such points — one hundred and two of them, from eight budget analyses and one issue of LSC’s Budget Footnotes, with the reasons not one can be joined to either committed series — is scholarship-bounds.tsv, asserted by crates/project. verified

    family income        K-8 award    9-12 award
    up to 450% FPL       $5,500       $7,500
    550%                 $2,750       $3,750
    650%                 $1,375       $1,875
    750%                   $687.50      $937.50
    782.19% and above      $550         $750