Per-Pupil Operating Expenditure
What a district actually spends per student on day-to-day operations — salaries, benefits, purchased services, supplies — excluding capital outlay and non-elementary-secondary programs. It is the outcome side of every adequacy and equity claim about Ohio’s districts, and the measure against which base cost per pupil should be read: base cost is what the state says an education costs, this is what districts spend.
metric/per-pupil-operating-expenditure · 5 nodes point here · 1 correction
This metric is not the state’s business and that is the point. A district’s operating expenditure reflects local revenue, levy history, and local choice as much as state aid. Regressing it against valuation per pupil is the wealth-neutrality test; comparing it against base cost per pupil is the adequacy test. It answers neither question by itself. inference
The denominator here is a headcount, and its sibling’s is not. DEW’s report card publishes expenditure per equivalent pupil, which divides by an ADM count weighted upward for disadvantage, English learners, and disability. That measure runs about 17.8% lower at the median — $12,856 for 2024-2025 against $15,646 here for FY2024 — and it is the one sitting next to the Performance Index on the same report card, so it is the one an Ohio spending-versus-results argument will reach for. The two are not interchangeable in either direction, and the difference is not cosmetic: a need-weighted denominator is a need adjustment that suppresses exactly the gradient this metric exposes. verified
The report card splits this figure by where the money came from Contents
What this repository computed Contents
Not what Ohio publishes. Figures derived here from committed fixtures, each one citing the test that reproduces it.
The statewide series demonstrates why deflation is not optional. From FY2000 to FY2022,
Ohio public school operating expenditure per pupil rose from $7,065 to $15,314 — 116.8% in
nominal terms. Adjusted for inflation, the increase is 26.1%. verified crates/deflator
The same fact, stated two ways, supports two entirely different arguments about whether Ohio
has increased its investment in schools, and both numbers are correct. This is the concrete
case for the deflator calculator being a prerequisite rather than a refinement.
A third framing is also correct and lower still: excluding federal COVID relief funds, FY2022 operating EPP was $14,493, making real growth 19.4%. verified crates/deflator The $821 per-pupil gap between the two FY2022 figures is larger than most policy changes this corpus models.
Total operating expenditure roughly doubled over the period, $13.0 billion to $25.8 billion, while enrollment fell 8.3% from about 1.84 million to about 1.68 million. verified Per-pupil growth is therefore partly a numerator effect and partly a denominator one, and attributing it wholly to spending decisions is wrong.
Growth is broadly uniform across functions, which is itself informative — it rules out the common claim that administrative bloat drove the increase. In real terms FY2000 to FY2022: instruction +25.7%, support +28.2%, other +11.4%. verified Support grew fastest, but only slightly, and from a smaller base.
Every dollar of the rise since FY2010 is the denominator Contents
The federal district panel now runs FY2009 to FY2024 and carries the numerator beside the
quotient, which the statewide series does not. Over FY2010 to FY2024, on Ohio’s comparable
districts: verified crates/dispersion
real spending per pupil +8.86%
real total spending -1.11% FY2024 is 0.9889 of FY2010
fall membership -9.16%
Those are one statement, not three. A quotient’s growth is its numerator’s over its denominator’s, and the identity closes to the fourth decimal. Ohio’s traditional districts spent slightly less in FY2024, in constant dollars, than they spent in FY2010, and the per-pupil rise is 152,889 fewer children.
This node already warns that per-pupil growth “is partly a numerator effect and partly a denominator one, and attributing it wholly to spending decisions is wrong”. On this population over these years it is not partly. It is all of it.
Part of the enrollment fall is children moving rather than children not existing: 88,974 pupils sat outside the comparable set in FY2010 against 134,885 in FY2022, about a third of the district loss over the same span, and the rest is demographic. That changes what the arithmetic is about and not the arithmetic. verified
So the rule has a third clause Contents
The node’s existing rules are name the index and name the window. The panel adds name the denominator — and it is the same class of error as the weighted-versus-headcount confusion this metric was built around, one level up. Per pupil and in total are two answers to “did Ohio increase spending”, they disagree in sign over a fifteen-year window, and neither is wrong. inference
The pandemic year is the same error in miniature Contents
FY2020 to FY2021 the panel’s real spending per pupil is flat — up a seventh of one per cent — while its real total spending falls 3.3%. Fall membership dropped 54,777 pupils in a single year and was still 43,000 short of FY2020 two years later. A reader with the per-pupil series alone sees a year in which nothing happened. verified
The disagreement about this series is periodic, not only definitional Contents
OCG White Paper 015 makes the best available case that Ohio’s funding argument is a quarrel about definitions: it names seven standards for “defunding” and answers each. Its per-pupil and real-reduction rows both read No — up ~25-26% real, which is correct endpoint to endpoint and is this metric’s own headline figure.
It is still not the whole answer, because every standard in that table is applied FY2000 to FY2022. Deflated year by year this series falls about 7% between FY2010 and FY2014 and does not regain its FY2010 level until FY2018. A speaker who says Ohio defunded its schools and is thinking of those years is describing something that happened, under the paper’s own real and per-pupil standards, and no choice among its seven definitions surfaces it. inference
So the rule this metric adds to the deflation rule: name the window as well as the index.
An endpoint pair cannot see an interior reversal, and this series contains three. Two
speakers can select different windows on one record and both be correct, which is a second
and independent reason the public argument does not resolve. See
skills/deflator.
That gradient, computed over the same FY2024 cross-section behind the equity findings above: operating expenditure per pupil correlates with the economically disadvantaged share at +0.383 (Spearman +0.348), and mean spending by quintile of disadvantage runs $15,526, $15,148, $15,787, $16,103, $18,990. verified Ohio’s highest-poverty fifth of districts spends about $3,400 more per headcount pupil than its lowest. Any analysis reporting that spending and need are unrelated should be checked for which denominator it used first.
The check has now been run, on one numerator against both divisors. Correlated against the Performance Index, FY2025 operating expenditures give −0.337 per unweighted pupil and −0.015 per weighted pupil; this metric’s own FY2024 series gives −0.388. verified Two headcount measures from two publications agree; the weighted one is the outlier. The denominator is not a presentation choice.
A second, independent series also lands here. The FY2025 Expanded List’s headcount median of $16,289 sits 4.1% above this metric’s FY2024 Cupp Report median of $15,646 — two publishers, consecutive years, consistent level. verified It is the closest thing the corpus has to an external validation of this series, and the first time this metric has been checked against anything but itself.
exp_per_equivalent_pupil_federal and its state-and-local twin have been in the committed
fixture since the report card was first extracted and reached no consumer. They are complete
for all 607 districts and add to the whole to within a rounding — the largest disagreement in
the fixture is $1. verified
Federal money is a median 4.20% of a district’s operating spending, and 29.04% at the
top. 49 districts are above a tenth; Alexander Local in Athens County is the most exposed,
at $3,280 of $11,294 per equivalent pupil. The floor is 0.7%. verified 2024-25 report
card, computed in crates/bundle
This is the only quantity in the corpus that measures exposure to a decision Ohio does not make. Every other figure here describes a choice by the General Assembly, a county auditor or a local electorate; this one describes what happens to a district if a federal appropriation changes, and the answer differs by a factor of forty across the state. inference
The share, not the dollars, is the portable number. Both parts are published per need-weighted pupil, the denominator this corpus has repeatedly warned about and this project has twice been caught mixing. A ratio taken inside one denominator carries none; a comparison across two does. Any use of the dollar figures has to name the basis, and they are not additive with the function breakdown, which divides by an unweighted headcount. verified
And the correlation that must not be reported alone. Federal share against the Performance Index is -0.546 raw and -0.115 holding poverty constant. verified crates/bundle Federal education money is allocated substantially by poverty, so the raw figure is very largely the poverty relationship read backwards; quoting it without the control would state a confound as a finding, which is the error this corpus most often finds in secondary reporting. Neither figure identifies an effect in either direction. verified
Properties Contents
| Name | Per-pupil operating expenditure |
|---|---|
| Unit | dollars-per-pupil |
| Definition | operating_expenditure_per_pupil = total operating expenditure / enrollment Operating expenditure covers day-to-day spending on running schools and educating students: salaries, benefits, purchased services, supplies. It excludes capital outlay (buildings, buses) and programs not directly related to elementary and secondary education (adult education, community services). Must be reported with (a) whether federal relief funds are included, (b) nominal or constant dollars and which deflator, and (c) which entity population is counted. Omitting any of the three produces a figure that cannot be compared to any other figure. |
| Inputs | Total operating expenditure and enrollment by entity and fiscal year — NCES National Public Education Financial Survey and Common Core of Data, via census-f33 and nces-ccd. Deflator — bls-cpi, national CPI-U all items, June of each year, aligned to fiscal year end. |
| Values over time | |
Statewide, all Ohio public school entities, nominal: FY2000 $7,065 verified FY2002 ~$8,100 verified chart label, approximately $100 precision FY2004 ~$9,000 FY2006 ~$9,700 FY2008 ~$10,300 FY2010 ~$11,200 FY2012 ~$11,300 FY2014 ~$11,400 FY2016 ~$11,900 FY2018 ~$12,900 FY2020 ~$13,700 FY2022 $15,314 including COVID relief; $14,493 excluding verifiedReal growth FY2000-FY2022: +26.1% including relief funds, +19.4% excluding. verified crates/deflator CPI growth over the same span: +71.9%. verified crates/deflator Constant FY2022 dollars, computed by the deflator calculator (cargo run --example epp_real_series -p deflate):FY2000 $12,143 baseline verified crates/deflator FY2002 $13,341 +9.9% inference FY2004 $14,058 +15.8% inference FY2006 $14,166 +16.7% inference FY2008 $13,948 +14.9% inference FY2010 $15,226 +25.4% inference FY2012 $14,591 +20.2% inference FY2014 $14,173 +16.7% inference FY2016 $14,630 +20.5% inference FY2018 $15,169 +24.9% inference FY2020 $15,747 +29.7% inference FY2022 $15,314 +26.1% verified FY2022 $14,493 +19.4% excluding federal relief verifiedIntermediate rows are inference on the NOMINAL side only: the figures are chart-label approximations to about $100. The deflator no longer contributes uncertainty — every CPI observation from FY2000 to FY2022 has been checked against the Bureau of Labor Statistics' own published series by the bls-cpi connector, and crates/connect/tests/deflator_matches_bls.rs re-checks all twenty-three on every test run. verifiedThat check found one error. FY2016 had been transcribed as 241.038 where the published index is 241.018, which moved this row from $14,629 to $14,630 — a difference too small to change any claim here, and exactly the size of error that a plausible-looking series hides. Twenty-two of twenty-three points were typed correctly. The one that was not could only be found by going back to the source, which is the argument for the connector rather than for more careful typing. TWO FINDINGS THE NOMINAL SERIES CONCEALS. inference Both are now pinned in crates/deflator/tests/ohio_epp_real_series.rs rather than left as prose, including a check that perturbs both trough endpoints by the full chart-label uncertainty in the direction that would erase the decline and finds it intact.Real per-pupil spending did not grow steadily — it fell. From FY2010 to FY2014 it dropped from $15,226 to $14,173, a real decline of 6.9%, leaving FY2014 statistically indistinguishable from FY2006's $14,166. Eight years of no real gain. That trough sits inside the Bridge formula decade and coincides with the withdrawal of federal stimulus. Causation is still not established, but the shape is no longer merely "near-flat" — it is a decline and a recovery, and the recovery did not reach the FY2010 level again until FY2018. Real spending peaked in FY2020, not FY2022. At $15,747 it stands about $433 above the FY2022 figure of $15,314. verified crates/deflator Federal COVID relief arrived while prices were still low and was then eroded by the FY2021-22 inflation, so the nominal high and the real high are two years apart. The federal district panel does not reproduce that peak. On exact figures rather than chart labels, FY2020 and FY2021 sit within $21 of each other, FY2022 is $112 below FY2020, and FY2023 and FY2024 are both above everything before them. The dip after FY2020 is there and it is a tenth the size; the summit is not. verified crates/dispersionThe two are different entity sets, so this does not settle the Auditor's series — its wider population includes the community schools and service centers the panel excludes, and those were where a large part of the relief money landed. What it does settle is that "real spending peaked in FY2020" cannot be carried as a general fact about Ohio: on the one exact per-district record the corpus holds, it is false. Whether the Auditor's own population peaks there is still unresolved and now has a specific shape. open THE TROUGH HAS NOW BEEN LOCATED DISTRICT BY DISTRICT, not from the Auditor's dashboard but from the federal survey underneath it: the Census F-33 district panel, extended back to FY2009 by three NCES archives published under a _txt suffix the FY2012-FY2022 files do not use. The panel had opened at FY2012, three years inside the decline. verified see crates/dispersion/tests/f33_ohio_panel_trough.rsIT WAS BROAD. Of the 610 comparable districts the panel carries in both endpoint years, 490 — 80.3% — spent less per pupil in real terms in FY2013 than in FY2010. Median change -5.6%, first quartile -9.1%. verified crates/dispersion This was not a handful of districts cut hard enough to move a statewide average; four in five actually spent less. The denominator is the inner join and not a district count. It was written here as 607 — the report card's population, a different panel for a different year — and 490 of that is 80.7%, which is where the 81% came from. A district the survey gained or lost between FY2010 and FY2013 cannot carry a change, so the population that can is smaller than either year alone. AND NOTHING ABOUT A DISTRICT PREDICTS HOW FAR IT FELL. The obvious hypothesis — a contraction coinciding with the withdrawal of federal stimulus, inside the Bridge formula decade, should land hardest on districts most dependent on state aid — is not supported. Against the depth of the fall: FY2010 state share of revenue +0.020 verified crates/dispersion FY2010 federal share of revenue +0.026 verified crates/dispersion FY2010 local share of revenue -0.024 verified crates/dispersion economically disadvantaged share -0.033 [later vintage, directional] assessed valuation per pupil +0.033 [later vintage, directional] log enrollment -0.197 verified crates/dispersion FY2010 spending per pupil -0.168 verified crates/dispersionOnly size and prior level register at all, both weakly and the second partly mechanical. The deepest decile is real — 61 districts averaging -17.5% — but carries 12.0% of pupils, so it is a tail rather than the story. verified crates/dispersion Every figure in that table moved when it was bound. The test behind them asserts a band — r.abs() < 0.06 for the three revenue shares, close(r, -0.186, 0.02) for the two that register — and the corpus published the band's center as though it were the measurement. The bands still hold and every one of the values inside them had drifted, which is the eighth occurrence of that pattern in this issue and the reason convention four exists.A SECOND SOURCE CONFIRMS THE SHAPE. Enrollment-weighted, the F-33 panel falls 6.8% from FY2010 to FY2013 verified crates/dispersion against the Auditor's 6.9% from FY2010 to FY2014 — different survey, different entity set, levels about $1,500 apart, same decline. verified The two series are not interchangeable and should never be quoted as one; that they agree on the shape is what licenses reading them together. FY2014 is absent from the NCES archive under every naming the surrounding years use. That used to mean the district measurement stopped at FY2013, at the open end of the series. It no longer does: the panel now runs to FY2024, FY2013 is the minimum of all fifteen years it holds, and FY2015 — which the trough reading did not have — is 1.7% above it. The missing year sits between two observed ones rather than beyond the last. It is still unpriced, and it is a materially smaller unknown than it was. open Per-district series also exist on the Auditor's own dashboard, on the same entity basis as the statewide series here, and remain unextracted. That is now the only route left to the FY2020 question above, because the entity basis is what the two records disagree about; it is no longer the route to a per-district series at all. open | |
| Caveats | Entity scope is wider than the 611 districts receiving foundation funding: it includes JVSDs, educational service centers, community schools, STEM schools, and state- and federal-run schools, plus tuition those entities pay to private and out-of-state schools. Not interchangeable with a district-only figure. Excludes capital entirely, so a district rebuilding its schools under facilities assistance shows no increase here. NCES definitions differ from Ohio's; this is not the operating cost the Ohio formula computes, and comparing it directly to base cost per pupil requires stating the gap. Enrollment denominators differ between NCES counts and Ohio ADM. verified |
| Calculator | dispersion (for cross-agency distribution); deflate (required) |
Where this appears on the site Contents
The pages outside the corpus that link here, and the section of each the link sits in.
| Compare | In “Side by side”. |
|---|---|
| Cleveland Municipal | In “Position among Ohio's 609 districts”. The same link is on all 609 pages like this one. |
Links Contents
| Derived from | Enrolled ADM |
|---|---|
| Contrasts with | General Fund Cash Balance |
| Instance of | Metric |
| Derived from | Assessed Valuation Per Pupil |
| Compares against | Expenditure Per Equivalent Pupil |
| Makes measurable | Adequacy |
| Makes measurable | Equity |
| Computed for | Fiscal Year 2022 |
| Compares against | Base Cost Per Pupil |
| Sourced from | Longitudinal School Finance Study — Ohio Auditor of State |
| Sourced from | OCG White Paper No. 015 — Has Ohio Been Defunding Public Education? |
Also mentions
Pointed at by
| Bridge Formula | mentions |
|---|---|
| Enrolled ADM | mentions |
| Expenditure Per Equivalent Pupil | Compares against |
| General Fund Cash Balance | mentions |
| Title I, Part A | mentions |
What this node used to say Contents
The corpus is not rewritten to have always been right. Each entry is a claim this node carried, what replaced it, and the thing that settled it.
Correction 1 of 1
It said
The figure was carried as a single per-pupil total, with no source split.
It says
exp_per_equivalent_pupil_federal and its state-and-local twin are complete for all 607
districts and add to the whole to within a rounding. verified
Settled by
The committed report-card fixture, which has held both columns since the report card was first extracted.
What else it touched
Nothing was wrong; something was unread. The split matters most for FY2020-FY2024, where the federal column is ESSER and a total read as ordinary operating spending is not one.