FY26 TRAD State Foundation Funding Calculator — recovered from the archive

Source. Ohio Department of Education and Workforce, Office of Budget and School Funding. FY26-TRAD-State-Foundation-Funding-Calculator-9-22-2025.xlsx, dated 22 September 2025. Type. Primary source — the department’s own working model for FY2026. Location. The Internet Archive, and nowhere else. Capture 20251008001131 of education.ohio.gov/getattachment/…/Traditional-School-Districts/. The department publishes one calculator and replaces it in place; when the FY2027 model went up this file stopped existing at any address the department serves.

What it contains. The same thirty-odd sheets as the FY2027 model, a year earlier. Only the Transportation sheet’s statewide factor block is extracted here — see What is not taken below.

Why it was worth recovering. The two transportation rates are the Fair School Funding Plan’s only endogenous parameter: R.C. 3317.0212(C) and (D) define both as trimmed means of what districts reported spending in the prior year, so they move without an act and nobody chooses them. Transportation Cost Rates carried “the rates before FY2027” as an open field from the day it was written. This file closes one year of it, and one year is the difference between a value and an interval.

What it shows. The two rates move in opposite directions:

	FY2026	FY2027	change
per weighted rider	$1,275.000	$1,337.175	+4.88%
per mile	$7.060	$6.867	−2.73%
SpEd transportation proration	1.0	0.917459740976215	first proration
minimum state share	45.83%	50%	+4.17pp

Between the two cost years those rates are drawn from, Midwest diesel rose 17.5% (Midwest No. 2 diesel retail prices — the EIA monthly series). The per-mile rate fell against it. Whatever the mile base responds to, it is not fuel in any direct way — and it is the base 350 districts and $308.8M of school bus payment are actually paid on.

Access constraints. Free. No live publisher URL exists; verify checks the archived bytes against the committed digest like any other source.

Caveats:

• The FY2025 calculator is not recoverable and this is what that looks like. Every archived copy — two capture dates, two filenames, four URLs — is exactly 1,048,576 bytes, truncated at one mebibyte mid-stream, and fails a zip integrity check while answering 200. The length is recorded here so a later reader does not spend an afternoon rediscovering it. FY2022 to FY2024 are not in the archive at all.
• A 200 from the archive is not provenance, so the extractor does not rely on one. The builder recomputes each district’s published per-rider base from the rate it read out of the sheet header and fails on any disagreement. That is the condition “An archived source is still a source” admits the address on, and it is a stronger claim than any live source in the registry makes about itself.
• The aggregate row has no name. As on the FY2027 Transportation sheet, the statewide total sits in the district table with a blank district name, so a name-based filter misses it and a naive column sum comes out exactly double. It satisfies the per-rider identity, so the builder’s check passes over it harmlessly; anything computing a statewide total from this sheet must exclude it.
• There is an “Inflation factor” of 0.05 beside the rates, in both years, and it is in no section. In FY2026 the adjacent cell carries $63.75, which is 5% of $1,275; in FY2027 that cell is zero. What it does is not established. open
• The Performance Supplement sheet is the FY2027 one’s, headers included. Its O1, O2 and O3 match the FY2027 workbook’s for all 612 rows, so they carry the same stale headers. O2, headed 2023-2024, is the 2024-25 progress rating, and O3, headed 2022-2023, is the 2023-24 one. No FY2026 panel is extracted, so only the FY2027 columns are pinned by a test. verified

What is taken, and what this entry got wrong first. It read:

What is not taken. The per-district tables. They are the FY2027 model’s a year earlier and the counts are very nearly identical — bus miles are the same 825,828 in both, public riders the same 651,702 — so committing them would be four megabytes to restate a panel already held.

Both facts are true and the conclusion does not follow from them. Bus miles and rider counts are a survey a year apart and they are stable. The columns two corpus nodes had recorded as unmeasurable “because the corpus holds one year of the calculator” are not: per-pupil local capacity moves +9.34% at the median and 2 districts of 609 are unchanged, and disadvantaged pupil impact aid moves −8.58% with none unchanged. Generalizing from the one sheet that was opened to the thirty that were not is the error, and it foreclosed the question the file was closest to answering.

Sixteen columns are taken now — the two sides of the state share, the local capacity build-up and the disadvantaged pupil blend — as crates/project/fixtures/fy26-department-model.csv, about a hundred kilobytes. Built by crates/connect/src/fixtures/fy26.rs and read by project::prior_model. The rest of the per-district tables genuinely do restate what FY2027 already holds, and are still not taken.

And the second year is what makes a vintage checkable at all. The CTE and EL sheets head their count columns -FY21, and four corpus nodes read a six-year freeze off that suffix. It is a stale label on a sheet built for the FY2022 report: the English learner column carrying it moves in 512 of 611 districts between these two workbooks, and the department’s own Directions table — rewritten for each year’s file, unlike the headers — puts the categorical counts at FY26 in both. One workbook could not have shown that. Every count the plan multiplies is committed for both years at crates/project/fixtures/calculator-counts.csv, and the two vintage tables at crates/project/fixtures/calculator-vintages.tsv, built by crates/connect/src/fixtures/counts.rs.

Feeds connector #

dew-foundation, built by crates/connect/src/fixtures/transport_rates.rs crates/connect/src/fixtures/fy26.rs and crates/connect/src/fixtures/counts.rs.