FSFP Targeted Assistance

Two tiers that measure different things and add # ↑ Contents

[G] = [C] + [F], and the addends do not answer the same question:

• [C] the capacity amount is 0.8% of however far the district’s total weighted wealth falls below the statewide median district’s. It is about the size of the tax base, full stop — so a small district with a small base scores well on it regardless of how few pupils that base serves.
• [F] the wealth amount is a rate against weighted wealth per resident pupil, cutting off where that reaches 1.25 times the state median.

The capacity tier has a size cliff # ↑ Contents

Because that tier measures total wealth rather than wealth per pupil, a very small district would otherwise draw equalization designed for one twenty times its size. The department’s guard is a step function rather than a per-pupil measure: a district below 200 ADM receives none of the tier, one from 200 to 400 receives 5% of it, and the fraction ramps linearly to 100% only between 400 and 600. verified

A supplemental tier that qualifies districts and pays them nothing # ↑ Contents

[I] Supplemental Targeted Assistance is zero for every district in the FY2027 model. Its eligibility test still runs and still qualifies 36 districts: [H] is YES where a district’s FY2019 wealth index exceeded 1.6 and its FY2019 enrolled ADM was below 88% of its total ADM — a district that was poor and had already lost more than an eighth of its pupils to open enrollment and community schools. verified

The zero is a repeal. H.B. 96 eliminated the payment outright — item 5 of the act’s own list of changes to the school financing system. verified In FY2025 it paid $52.5 million to 36 districts, the same 36 the eligibility column still qualifies. What survives is the column, not the program.

What this repository computed # ↑ Contents

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

The tier was repealed while the condition it answered remained # ↑ Contents

The greenbook says who the 36 were and why the tier existed: districts that “might appear relatively wealthier on a per-pupil basis because of changes in how the formula in place since FY 2022 counts students” — mostly lower-wealth urban districts with large shares of resident pupils in another district, a community school, or a nonpublic school on a scholarship. verified

So the supplemental tier was compensating for an artifact of the plan’s own pupil count, and H.B. 96 removed the compensation and left the artifact. The eligibility test still runs each year and still names the districts, which means a future biennium could fund the line without changing anything else — a statement about the spreadsheet rather than about the law.

It behaves like an equalization and not like a categorical: it falls to exactly zero for 135 of 609 districts, which no payment for pupils ever does. Columbus receives none of it. verified

The rate the repealed tier paid at is recoverable, exactly # ↑ Contents

“A future biennium could fund the line without changing anything else” was true of the spreadsheet and not of this corpus, which could name the districts and not the amount. R.C. 3317.0218 is repealed, so there is no statutory text to read, and the department’s own description gives a range and no formula — “the amount of per-pupil funding is scaled to increase based as the level of wealth decreases … Districts who qualify for this supplemental funding receive between an additional $85 and $750 per student educated.” The rate was the one input the panel did not carry.

It is in the last payment. Against the FY2025 report’s 36 districts the rate is linear in the FY2019 wealth index, running from $75.0000 at the eligibility threshold of 1.6 to $750.00 at the state’s highest index — Youngstown City, 2.81996153 — with a worst relative residual of 6.4e-08 across all 36. That is not a fit; it is the schedule. verified crates/project/tests/the_rate_a_repealed_tier_was_paid_at.rs

Two details had to be right together. The denominator is current-year enrolled ADM — “per student educated” is the department’s own phrase — and not either of the FY2019 counts the eligibility test uses; fitted against FY2019 enrolled ADM the relation is only approximately linear and East Cleveland misses by 35%. And the $85 is not the floor: it is what the lowest qualifying district receives, Middletown City at index 1.6194, $85.76. The floor is $75, at a threshold no district sits on. verified

The top of the scale is a rank. $750 is paid to whichever district holds the highest FY2019 wealth index, and everything below is interpolated toward $75. Nobody sets the number the scale tops out at; one district’s position does — the same shape as R.C. 3317.017(A)(4)‘s benchmark, which is the other parameter in this formula that reads as discretionary and is a statistic of a distribution. See FSFP Local Capacity Measure. verified

So the tier is a lever now, and restoring it is priceable. Run forward against the FY2027 panel at the schedule it was last paid on it costs less than the $52.3m of FY2025, not more: the eligibility test is frozen on FY2019 data so the population does not move, and the districts in it have fewer pupils than they had. verified

Statewide the split is 24% capacity and 76% wealth, and per district it can be either, both, or neither: 299 draw the first, 438 the second, 135 nothing. Which tier a district is on is a statement about it that the total conceals. verified

It binds for five districts today, and that is not why it matters. A threshold written into the formula is a standing property of the formula; the number of districts on the wrong side of it is a fact about this year’s enrollment. Sixty-eight districts sit between 200 and 600 pupils, on the shelf or the ramp, and a shrinking district crosses these boundaries in the direction that costs it money. inference

The wealth tier’s 0.8 cutoff is not a policy threshold # ↑ Contents

The tier is (median_per_pupil x 0.014 - district_per_pupil x 0.0112) x ADM, gated by IF(index < 0.8, 0, ...). And 0.0112 is exactly 0.8 x 0.014, so the bracket reaches zero at precisely the index the gate cuts off at. The IF is not choosing where to stop paying; it is stopping the formula from charging wealthy districts. verified

That is worth knowing before anyone treats 0.8 as a dial. Moving it without moving the coefficient pair either opens a discontinuity or starts paying negative aid. inference

Two pupil counts, one line apart # ↑ Contents

[D] divides weighted wealth by resident ADM — enrolled, less those open-enrolling in, plus those open-enrolling out. [F] then multiplies the resulting rate by enrolled ADM. Both are defensible alone: wealth per pupil should arguably be measured against the pupils a district’s own taxpayers are responsible for, and aid should arguably be paid on the pupils it teaches. Together they mean a district with heavy inbound open enrollment is measured as wealthier per pupil than it is and then paid on the larger count. The measurement and the payment disagree about who the district’s pupils are. verified

Where it belongs in this ontology # ↑ Contents

Beside local capacity, not beside the weighted categoricals. Both blend assessed valuation with federal adjusted gross income — 60/40 here against local capacity’s 60/20/20 — and both are about the tax base rather than the pupil. The two blends being different weights on overlapping inputs, inside one formula, for two purposes, is exactly the kind of thing a single “categoricals” node would have hidden. verified

What it replaces, which is one of the two candidates and not the other # ↑ Contents

The charge-off era ran two components here, and the corpus had them as one question. They are not alike. verified crates/project

Gap aid could not have had a successor. LSC files it under the heading “Charge-off Supplement (Gap Aid)” and defines it as the payment made “to districts whose actual operating revenue is lower than the local share assumed by the formula”. It supplemented an assumption — the 23-mill charge-off — and when the charge-off went there was nothing left to supplement. Its $73.5m across 145 districts in FY2008 is the last biennium any analysis carries it.

Parity aid is the ancestor, and three of this component’s features are its. Am. Sub. H.B. 94 created it to equalize “an additional 9.5 mills (above the adequacy level) to the 80th percentile district’s wealth level”, on “a weighted average property wealth (2/3) and income wealth (1/3)”, reaching about 492 districts. A millage instrument, a blend of valuation with income, and a rank threshold — which is what this component is.

FY2002  9.5 mills  2/3 : 1/3   the 80th percentile district   492 districts
FY2008  8.0 mills  —           the 80th percentile district   410 districts
FY2009  8.5 mills  75 : 25     the 80th percentile district   367 districts
FY2014  6 mills    1/2 : 1/2   the 490th lowest of 612        489 districts
FY2022  14 / 11.2  60 : 40     the median district

The threshold is what actually moved, and no analysis says so # ↑ Contents

The 490th lowest district of 612 sits 80.1% of the way up that panel, which is parity aid’s threshold to within a district or two. So H.B. 59 renamed the component, changed the blend and cut the rate, and left parity aid’s threshold exactly where it was — 489 districts eligible against parity aid’s 492 a decade earlier. verified

The Fair School Funding Plan is the act that moved it, from the wealth of a fairly rich district down to the wealth of the middle one, and raised the rate from 6 mills to 14 to pay for the shorter distance. That is a larger change than any of the rate changes in the series above, and H.B. 110’s analysis presents it as a new formula rather than as a moved threshold: the word “percentile” does not appear in its account at all. verified

Whether the two are equivalent in effect — a longer distance charged softly against a shorter one charged hard — is a distributional question this corpus can pose and has not run. inference the framing open the comparison is unmade

The scale is recomputed every year, and it moves further than the cost side does # ↑ Contents

Both medians move between the department’s two published models. Median weighted wealth goes from $362,326,436.80 to $392,151,306.63, +8.23%; the per-pupil median from $253,337.38 to $276,708.97, +9.23%. Every weight and every base cost in the plan is the same number in both years, so the scale this component equalizes to moves nine per cent in a year while the scale the rest of the formula prices against does not move at all. verified crates/project

And the exclusion was computed once, made no difference, and was left behind # ↑ Contents

The second pair — excluding North Bass and Middle Bass — equals the main pair to the last published digit in the FY2026 model. Two districts out of 611 cannot move a median, which is what a median is for. In the FY2027 model those cells still hold the FY2026 numbers, beside a main pair that has moved eight and nine per cent. So it is not an alternative calculation maintained beside the live one; it is last year’s value nothing recomputes. verified crates/project

That is the third stale cell found in this workbook, after the -FY21 suffix on the career-technical and English learner counts and the FY21 heading on transportation’s guarantee base. A value in these files can be stale in three distinct ways: a heading that no longer describes its column, a table of vintages two biennia old, and a number nothing recomputes. inference

The capacity tier’s index is a measurement of enrollment # ↑ Contents

The description above records that the tier “is about the size of the tax base, full stop”, and reads the department’s 200/400/600 ramp as the guard against that. Both are right and neither is the measurement. The tier compares two whole-district totals — the median district’s weighted wealth against this district’s — so what its index varies with is how many pupils the base is under.

Across all 609 districts, ln(total weighted wealth) regressed on ln(base cost enrolled ADM) has a slope of 0.9855 and an r-squared of 0.8126. The same districts’ weighted wealth per resident pupil on the same predictor has a slope of -0.0429 and an r-squared of 0.0085. verified crates/project A slope of one and four fifths of the variance on the quantity the tier reads; no gradient at all on the quantity it names. Ohio’s small districts are not poorer per pupil than its large ones, and this tier pays them as though they were.

So the tier’s $334,036,712.23 reaches 299 districts at a median 876 pupils and no district above 3,259.83, which is the first district of the largest ADM sextile. Per pupil, by sextile, it is $2,349, $1,616, $433, and then nothing. verified crates/project

It is the half of the guarantee’s size curve nothing had accounted for # ↑ Contents

The statewide guarantee rate by ADM sextile is non-monotone — lowest among the smallest districts, highest in the third — and base cost’s staffing floors are half of it. The residual was recorded as “not size”, on the reasoning that base cost has no other size-dependent term. The reasoning holds; the conclusion does not, because this tier measures size outside base cost.

ordered on base cost enrolled ADM    1   2   3   4   5   6   gap
as the plan stands                  18  39  71  60  49  57    53
staffing floors struck out          48  51  74  61  49  57    26
capacity tier struck out            52  72  88  63  49  57    36
both struck out                     89  82  89  64  49  57     0

Strike both out and the first-to-third gap closes to zero. verified crates/project The tier alone keeps 87 districts off the guarantee, against the floors’ 46, and unlike that counterfactual this one needs no bound: [C] is paid in full rather than at a state share, it feeds no statewide average, and FY2027’s phase-in is 100%, so a dollar off it is a dollar off foundation funding exactly. verified crates/project

The size cliff the corpus recorded is the part of the tier that pushes the other way # ↑ Contents

It is natural to read the 200/400/600 ramp as the mechanism. It is its opposite. The ramp withholds $91,859,433 from the 73 districts it touches; pay them in full and the smallest sextile’s guarantee count falls from 48 back to 20 — 28 of the 30 districts the floors’ removal added to that band go straight back off the guarantee. verified crates/project The drafters bounded the tier’s size dependence with the one provision visibly about size, and left the rest of it in the index, where nothing names it. inference

And it is the denominator rather than the rate # ↑ Contents

Take the same eight mills against the shortfall of weighted wealth per resident pupil below $276,708.97 — the published median the tier’s own sibling [F] already indexes against — and multiply by enrolled ADM. Same section, same rate, same published median, and $24.2 million less in aggregate. The sextile row reads 86 / 74 / 86 / 56 / 45 / 56, first-to-third gap zero. verified crates/project

That is an attribution and not a proposal: it holds the rate and very nearly the money fixed and changes only what the shortfall is measured against, which locates the whole residual in the comparison division (B) makes. Whether the total is the wrong measure is a different question — the per-pupil restatement is [F]’s own form at a third rate, and two measurements of the same tax base at different rates may be what the General Assembly wanted.

The total is inherited, and the biennium that chose it says what it is for # ↑ Contents

Why division (B) compares two whole-district totals was recorded as unestablished, on the reasoning that the section carries no recital and H.B. 110’s analysis derives nothing. Both of those hold. The question was put to the wrong act.

Capacity aid’s own comparison was already total against total. The component H.B. 110 absorbed was charged on “the amount a district can raise with one mill” — District capacity amount = Three-year average valuation x 0.001 — and its ratio divided the statewide median of that quantity by the district’s. A one-mill yield is a whole-district dollar amount by construction, which is what a property tax raises. So (B) did not choose a total; it kept one. verified

And the act that created it states the purpose the current section omits: capacity aid “targets funding to smaller districts with relatively low total property valuation” — LSC, in H.B. 64’s analysis, restated verbatim in H.B. 49’s, and in no other biennium. The size dependence recorded above as sitting “in the index, where nothing names it” is named, two regimes back, by the document arguing for the instrument the index came from. verified crates/project

The pupil count follows the shortfall, so it was never dropped # ↑ Contents

The premise that made this look like a drafting choice — that the predecessor was per-pupil and the plan removed its ADM multiplier — is true of capacity aid’s payment and false of its comparison. And the removal is not a decision about size.

Both predecessors computed an index and multiplied the payment by it: H.B. 59’s tier one by its wealth index, capacity aid by its capacity ratio. R.C. 3317.0217 computes both indices, at (B)(3) and (C)(3), and multiplies by neither — each survives only as a gate, at 1.0 and at 0.8. One transformation, applied to both tiers: index demoted to a threshold, payment restated as a flat millage on a shortfall. verified

Once that is the form the pupil count follows from the shortfall. [F] multiplies by enrolled ADM because a per-pupil shortfall is not money until it does; [C] does not, because a difference of two whole-district totals already is. Capacity aid’s ADM multiplier was what turned a dimensionless ratio into dollars, and a difference needs no such conversion. verified the uniformity inference the drafters’ reasoning is not recorded anywhere

What it costs, and what stands where the cap stood # ↑ Contents

The restatement is not free, because the two forms differ in the limit. Capacity aid was x Formula ADM, so it vanished with the pupil count, and its generosity per pupil was bounded absolutely — per-pupil amount x multiplier x 2.5, one statewide ceiling for every district. [C] has neither factor, so as weighted wealth falls it approaches $3,137,210.45 — eight mills of the median district’s — with no pupil count dividing it.

Kelleys Island Local has 3.51 enrolled pupils and (B)(4)(b)(i) computes $2,461,241.79 for it: $701,208.49 a pupil. All 17 districts the cliff and the 5% shelf reach would draw more per pupil than $3,773.82, the most any district the tier pays in full draws. verified

So the 200/400/600 ramp stands where capacity aid’s 2.5 cap stood, not where its ADM multiplier stood. Both bound the same divergence, and only the older one is written as a bound on generosity rather than as a bracket on a pupil count. inference

And nothing committed records the change # ↑ Contents

H.B. 110’s analysis names capacity aid once in its whole length, in the sentence replacing it, and never states what it did — no multiplier, no cap, no mill. The third instance of a shape this corpus keeps finding: a change presented as a new formula rather than as a moved part, after the threshold above and the benchmark rank on local capacity. verified

Nor was there a budget choice to explain. H.B. 1 of the 134th repeals R.C. 3317.0218 — capacity aid’s section — and enacts a new R.C. 3317.0218 under the same number for the supplemental tier, in one document, with (B) already reading as it reads in force. The plan arrived drafted; a greenbook explains what a budget changed. The remaining question belongs to the plan workgroup’s own papers and not to any act, which is where Local Capacity Percentage left the benchmark rank. verified