Twenty-Mill Floor

It is a small number that does something large. For a district above the floor, reduction factors neutralize valuation growth and revenue from existing levies stays roughly flat in nominal terms. For a district at the floor, reduction factors stop operating and revenue rises directly with assessed valuation. verified The two kinds of district are on fundamentally different revenue trajectories, and which one a district is on is determined by its levy history rather than by any current decision.

This makes the floor the highest-leverage single parameter in the system for questions about local revenue, and a frequent target of reform proposals in both directions. Raising it would move more districts onto valuation-linked growth; lowering it would do the opposite. inference Any credible simulation of a property tax change has to determine floor status per district first, because the same change produces opposite effects on the two groups.

The remaining candidate, confirmed in the statute itself # ↑ Contents

R.C. 319.301(A)(1): the reductions “do not apply to … taxes levied at whatever rate is required to produce a specified amount of tax money, including a tax levied under section 5705.199 or 5748.09 of the Revised Code, or an amount to pay debt charges.” R.C. 5705.199 is the emergency levy section. Fixed-sum levies are exempt from reduction as a matter of law. verified ohio-laws, crates/project/tests/the_statute_behind_the_weights.rs

An absence of an effect in one group is not an absence of the mechanism, and the correction is recorded rather than quietly replaced because the wrong reading was the more natural one.

What this repository computed # ↑ Contents

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

How much the reduction factors have actually taken: the median district has lost 42.4% of the millage its voters approved. The median voted current operating rate is 42.31 mills against a median effective Class 1 rate of 23.40, and the largest reduction in the state is 74.7%. verified FY2024 District Profile Report, current_operating_millage_ty23 against effective_class1_millage_ty23 This is the magnitude of the mechanism the floor bounds, and it went unstated here for as long as it did because the voted-millage column was present in the fixture and never parsed. Note the figure is the median of the per-district ratio, not one minus the ratio of the two medians, which is a different and meaningless quantity.

A defect in the corpus’s own calculator, found by running it against the tables. crates/millage implemented the floor as a clamp: any district whose reduced rate fell below 20 mills was raised to 20 only if its voted rate was at least 20, and otherwise took the reduced rate unchanged. That reduced the six sub-twenty-mill districts as if the floor did not concern them — a district voting 19 mills came out at 9.5 on a doubling of valuation. R.C. 319.301(D) states the floor as a guarantee of what twenty mills would raise, so a district already below twenty is under the guarantee before any reduction and none applies. The published data settles it: all six report voted and effective rates equal to four decimals. verified Fixed, with crates/xcheck/tests/against_published_tables.rs holding the crate to 1,218 published district-years.

How many districts sit at the floor: 170 of 606, or 28.1%, in TY2023. verified computed from the FY2024 District Profile Report; see crates/dispersion A further 63 sit within half a mill of it, so 233 districts — more than a third of the state — are at or effectively against the floor. The statewide median effective Class 1 rate is 23.40 mills and the maximum is 84.29. verified

That distribution is the single most important fact about Ohio local school revenue, and it is not a tail phenomenon. More than a quarter of districts are on a revenue trajectory where reappraisal raises income directly, and the rest are on one where it does not. Any statewide property tax analysis that treats districts as homogeneous is averaging across a regime boundary. inference

An anomaly the corpus’s model does not explain. 20 of the 606 report an effective Class 1 rate below the floor. 6 of them are straightforward: they never voted twenty mills — voted rates of 18.7 to 19.71 — and the floor cannot lift a rate above what voters approved. All six are in Appalachian and southern Ohio counties (Vinton, Morrow, Lawrence, Jackson, Pike, Scioto), which is itself worth noting: the districts levying least sit in the poorest part of the state. verified

The other 14 voted between 23 and 48 mills and still report effective Class 1 rates between 19.72 and 19.994 — just below the floor. verified Nothing in the corpus’s current account of R.C. 319.301 explains this. Two candidate explanations were offered when this was first recorded; Table SD-1 now lets one of them be tested, and it fails.

The combined-base hypothesis is disconfirmed. If the floor were applied to a combined Class 1 and Class 2 base rather than to Class 1 alone, the value-weighted real property rate for these districts would pin at 20 mills. It does not: on TY2024 it runs from 19.719 to 21.562 across the fourteen, and three of them land within a hundredth of 20. Statewide the same measure puts 62 of 611 districts at 20.00 against 155 on the Class 1 rate, so the Class 1 figure is the one the floor governs. verified TY2024 Table SD-1, all four counts from the same year

The TY2023 column of the same table says 76 of 611 on the combined measure against 172 on Class 1, with one of the fourteen within a hundredth of 20. Both years disconfirm the hypothesis and neither is the other, which is the reason for naming the year on every count: this paragraph previously carried the combined-base count from TY2023 beside the Class 1 count from TY2024 under a single TY2024 citation, over a denominator of 609 that belongs to the FY2027 funding model rather than to Table SD-1’s 611 districts.

R.C. 319.301 never says “twenty mills”. It says two per cent of the taxable value of the class, and two-tenths of one per cent for joint vocational districts. Those are the 20 and 2 mills this node carries, so the figures are right — but the name is a colloquialism for a percentage, and a reader searching the statute for the phrase will not find it. verified

Twelve of the fourteen sit within 0.12 mills of the floor and eleven within 0.05, which is more consistent with rounding inside the reduction-factor computation than with a different base. inference Bradford Exempted Village, at 19.7247, is the one that a rounding account would not cover.

And there is a term in the section the corpus’s model does not carry, which looked as though it would. (E)(2) does not require a district’s own current-expense taxes to reach two per cent. It requires them to reach it “when combined with the pre-1982 joint vocational taxes against that class” — defined at (E)(1) as the amount by which a joint vocational district’s TY1981 current-expense taxes exceeded two-tenths of one per cent of value, floored at zero. verified crates/project/tests/the_term_the_floor_model_omits.rs

This is not the combined-base hypothesis above wearing a different hat. That one changed the denominator and was tested and failed; this adds to the numerator, and a district carrying a positive amount of it may sit below twenty by exactly that much. Read backwards from the fourteen shortfalls, the implied 1981 joint vocational rates run 2.006 to 2.278 mills — a hair above the two-mill figure the same section names, which is where a joint vocational district’s current-expense rate sits. Bradford implies 2.278, an ordinary rate rather than an anomaly. inference

What makes it worth naming rather than guessing is the bound. No district that voted twenty sits more than three tenths of a mill below the floor, and every district more than half a mill below is one whose voters never approved twenty — the two populations do not overlap. A floor being violated has no bound; an additive term left out of a model has a bound the size of the term. verified computed

Confirming it looked as though it needed TY1981 joint vocational current-expense rates by district, which no source the corpus holds carries. It did not. The next section disconfirms the term without them.

The term is disconfirmed, by a column that was in the fixture the whole time # ↑ Contents

The membership lists the paragraph above called unpopulated were never missing. Table SD-1 publishes taxes charged twice — once excluding joint vocational operating levies and once including them — and the difference is the joint vocational levy on that district’s own parcels. 501 of 611 districts carry one in TY2024 and 110 carry none, and the 110 concentrate where career-technical education is not delivered by a joint vocational district: Cuyahoga, Stark, Hancock, Summit and Franklin hold 57 of them between them. verified crates/dispersion

The rosters at Eastland-Fairfield are still unpopulated and none of this populates them. The abstract answers whether a district is in a joint vocational district, and — below — which districts are in the same one, without ever saying which one that is. That is all the argument needs.

Membership is recoverable more precisely than yes-or-no. A joint vocational district levies one rate per class across the whole of its territory, so every member’s levy is the same class1 mills x class1 value + class2 mills x class2 value. Shelby County’s eight districts fit one such pair in every one of the four tax years — 2.67222 and 4.69303 mills in TY2024, with no district off by more than twenty-five millionths of its own levy — and those rates sit far enough above the two-mill floor of R.C. 319.301(E)(3) that the fit is not the trivial one any two floored joint vocational districts would also pass. Eight districts explained by two numbers, four years running, is co-membership. verified

And co-membership kills the term. The pre-1982 amount belongs to the joint vocational district rather than to the member, so it is the same for every district in the group. Anna Local’s Class I charges come to twenty mills within four millionths in each of the four years. Botkins, in the same joint vocational district, sits 0.0705 mills below. The term cannot be positive for Botkins and zero for Anna. Jefferson and Muskingum repeat the shape — Buckeye and Franklin below the floor beside three and two districts at exactly twenty — on the weaker evidence a fit at the two-mill floor gives. verified

And six districts below the floor are in no joint vocational district at all, so (E)(1), which defines the term by reference to “the joint vocational school district of which the school district is a part”, has no referent for them: Middle Bass, North Bass, McComb, Fort Recovery, Osnaburg and Morgan. McComb’s 0.1076 mills is the second largest shortfall among the fourteen. verified

And the other candidate does not survive either # ↑ Contents

Rounding inside the reduction-factor computation was the second explanation, and it was given eleven of the fourteen. R.C. 319.301(D)(1) recomputes the percentage every year against that year’s valuation, so a rounding residual is a fresh draw each year. It is not one. The published Class I rate is Class I taxes charged over Class I value to within half a unit of its own rounding in every tax year, so a rate below twenty is dollars that were not charged rather than a rounding of the column. Computed that way, 17 districts sit below the floor in all four tax years, and Southern Local of Columbiana County reports 19.99959, 19.99961, 19.99961 and 19.99960 across them — a shortfall reproducing in the fifth decimal place through four annual recomputations. verified

So both named candidates fail, and what is left is a different anomaly from the one recorded above: stable per district, in the dollars rather than the presentation, and not a property of a district’s joint vocational district. What survives untouched is the bound. open

The floor did not move in 2026 and its base did. Division (E) now excludes from “taxes charged and payable” any tax levied under R.C. 5705.194 or 5705.199 — emergency and substitute levies — approved by electors before 1 January 2026, until the first tax year from 2026 in which R.C. 5715.24 applies in a county holding part of the district. verified A district whose twenty mills included a pre-2026 emergency or substitute levy has that levy taken out of the floor computation until its county’s next reappraisal or update year, and may cross the floor without any rate changing.

This is a second and distinct thing from (A)(1)‘s exemption of fixed-sum levies from reduction, which this node records above. One says such a levy is not reduced; the other says it does not count toward the floor. A node recording the floor as “unmoved since 1976” is exactly the node most likely to miss it. inference

The count is 21, not 20, on the later table. Table SD-1’s TY2024 figures add Kelleys Island Local, which the District Profile Report omits entirely: 10.15 effective Class 1 mills against $21.2m of taxable value per pupil, the highest in the state by a factor of seven. It is the clearest instance of the six-district pattern — one mill raises $21,025 per pupil there, so twenty would be an absurdity — and it is invisible to the profile report the corpus had been counting from. verified

What two tax years establish that one could not. Of the 416 districts that began TY2023 above the floor, 301 saw their effective Class 1 rate fall by TY2024; of the 193 that began at it, 9 did. 97.1% of every effective-rate reduction in Ohio happens above the floor. verified computed in crates/bundle from TY2023 and TY2024 Table SD-1, over the 609 districts of the FY2027 funding model, at RATE_FALL_TOLERANCE

The split must be taken at the start of the interval. Classifying by the end year lets a district that fell to the floor be counted as a floor district with a falling rate, which puts the outcome being measured inside the category doing the measuring; it moves the at-floor fall rate from 4.7% to 21.6% and the headline from 97.1% to 87.7%. verified

And a caution against the regime language above. 54 districts sit above the floor by less than a twentieth of a mill, and 82 crossed 20.0000 in one direction or the other between the two tax years — 61 of those by under five hundredths of a mill. For roughly an eighth of the state the binary is settled in the fourth decimal place and is likely to read the other way next year. verified The regime switch is real and the boundary is sharp in the statute; the assignment of districts to sides is not stable, and a simulation that treats floor status as a fixed district attribute will be wrong about that eighth. inference

Which districts are on the far side of the boundary, measured rather than supposed # ↑ Contents

The passage above calls the floor “the single most important fact about Ohio local school revenue” and says a quarter of districts are on a trajectory where reappraisal raises income directly. It never said which quarter. Asked of the wealth distribution the equalization measure uses — Ohio’s comparable districts cut into quartiles on local revenue per pupil — the answer is monotone and steep: 46.1% of the poorest quartile sits at the floor against 12.5% of the richest, with 39.5% and 28.3% in between. verified crates/regime-diff, reappraisal_incidence::by_wealth, pinned by test

This is the general form of the observation already recorded above about the six districts that never voted twenty mills, all of them in Appalachian and southern counties. It is not six districts; it is the shape of the whole distribution.

What it does to a reappraisal. H.B. 920 holds a district’s yield flat as values rise, so a revaluation is a revenue event only on the far side of this boundary — which is to say, mostly for poor districts. Across FY2022 to FY2024 local revenue per pupil grew 22.5% in the poorest quartile and 11.3% in the richest, and removing the reappraisals from the data — putting every district on its own quiet-year rate — makes the FY2024 gap between the two $212 wider, not narrower. verified crates/regime-diff, reappraisal_incidence::gap, pinned by test

The identification is Ohio’s own staggered cycle rather than an assumption: the 88 counties revalue in different years, Table SD-1 carries four of them, and a cohort’s event year moves its class-I value 24 to 55 times as much as its quiet years. verified

This was recorded backwards elsewhere in the corpus. doctrine/equity held that Ohio’s FY2023 reappraisals “would raise local revenue in the richest quartile mechanically”, and offered that as one of two candidate causes for a break in the equalization series. The mechanism runs the other way, and the break was an artifact of a five-pupil district entering the survey population. See that node’s revisions.

The formula charges for the value on both sides of this boundary, and only one side collects it # ↑ Contents

Everything above is about revenue. Set the state’s side of the same reappraisal beside it. R.C. 3317.017 charges a district for assessed value with no reference to millage at all, so the value a revaluation adds raises local capacity and lowers the state share of base cost on both sides of this floor — while H.B. 920 decides, on only one of them, whether the district collected anything for it. verified crates/project

at the floor       87 districts    about 1.3 cents charged per dollar    2.0 cents collected
above the floor   181 districts    about 1.2 cents charged per dollar    nothing collected

At the floor the two roughly net. The state takes back a median 63.26% of what a dollar of new Class I value raises at the district’s own effective rate, rising to 74.2% and never reaching the whole of it. verified crates/project Above the floor there is nothing to net against: the levy yields the dollars it always yielded, the state share falls anyway, and a reappraisal is therefore a pure loss of state aid for the districts on that side. inference

This does not reverse the incidence measured above; it compounds it from the other direction. The districts at the floor are disproportionately the poor ones, so they are both the districts whose reappraisal produces revenue and the districts five eighths of whose new revenue the state recaptures. The wealthier districts above the floor produce no new revenue and are charged for the value regardless. Neither side is held whole, and the two are held short in opposite ways. inference