How to read this, and where it comes from

How we counted levy elections. The Missouri State Auditor’s annual Property Tax Rates report prints a flag beside every tax rate it reviews: A for a new voter-approved rate (or one replacing an expired levy), B for a voter-approved increase, decrease or extension of an existing levy. We parsed every school-district operating row — in both Appendix VII and Appendix VIII — for 2018 and 2020–2025, seven tax years. The complete scan found 140 A/B-flagged district-year rows: 139 voter-approved operating-levy increases across 121 districts, and one excluded row (Strain-Japan R-XVI in 2021) where the full operating ceiling decreased. The yearly increase counts are 19, 24, 16, 14, 21, 20 and 25. Twenty-two winning districts were still below $3.43 immediately after at least one increase. The 2019 report has not been obtained, so that year is not included.

How we matched report names to today’s district names. We do not use fuzzy text matching. Each Auditor name and tax year is joined to the Auditor-derived statutory levy history, which carries the district’s stable DESE/Census identifier; that identifier then selects the current district record. All 139 events map one way only. This resolves the two formerly uncertain historical labels: Cass County R-V School District is Archie R-V (the Auditor begins using the Archie name in 2023), and Scott County R-V School District is Scott Co. Central. An automated check now refuses to regenerate the district-profile history file unless every voter-flagged row is either mapped to an included increase or listed as an explicit exclusion. That check also caught three increases still missing from the first correction: Smithville R-II (2020), Mehlville R-IX (2023), and Maplewood-Richmond Heights (2025). The Corrections page preserves both the 20 July correction and this later reconciliation.

What this count can and cannot see. It sees elections that passed. The Auditor records tax rates, not ballots, so a district that never appears may never have gone to its voters, or may have gone and lost. We cannot distinguish those, and we do not claim to: the site says a district has not succeeded, never that it has not asked. We also do not yet hold the 2019 report, so an increase approved in that one year would be missed. Separately, we tested and rejected a tempting shortcut — inferring elections from rises in the tax rate ceiling. Ceilings float in both directions with assessed value under the Hancock Amendment. Counting any ceiling rise above five cents would flag 130 districts as having “won an election” in 2024–25, against 45 voter-approved operating-levy increases actually identified by the Auditor’s A/B flags. A ceiling increase is therefore not evidence of a public vote, and that shortcut is not used anywhere on this site.

“At their ceiling.” Of the 211 districts below $3.43, 192 are within a half-cent of their certified ceiling. We use the half-cent rule because the public-facing rates are displayed to the cent; it does not claim that all 192 unrounded rates are exactly equal.

Where $3.43 came from. The performance levy was created by SB 287 (2005). It was the average operating levy of the original “performance districts” — the model districts used to set the state adequacy target — measured against 2004 assessed values. It was descriptive: a statement about what a particular group of districts happened to charge twenty-one years ago. It has never been indexed or recalculated, although the state adequacy target is recalculated every two years by statute. In June 2025, DESE’s deputy commissioner told the Governor’s funding task force that the formula’s local-aid assumptions are “a figure frozen in time.” SB 727 has now turned that frozen 2004 average into a compliance standard with a presumption of fault attached.

The finding that matters most. Within the 513-district nonvirtual wealth-analysis universe, 211 districts fall below the performance levy; 192 — 91% — levy within a half-cent of their certified tax rate ceiling, the maximum they may charge without a public vote. For 209 of the 211, that ceiling is itself below $3.43. Exactly two districts in the group (Bell City R-II and Brentwood) could reach the standard by board action; every other one would have to hold an election. Most districts in this group are already at or effectively at the no-vote ceiling and still fall short of the state’s own number — and the statute calls the result insufficient local effort.

The rate is not a measure of wealth. Across Missouri’s districts the correlation between the operating levy and property wealth per student is −0.001 — not weak, but nothing whatsoever. Knowing a community’s tax rate tells you nothing at all about whether it is rich or poor. The cleanest demonstration holds the rate perfectly still: 64 districts levy the identical operating rate of $2.75. Among them, a dime raises $57 per student in Thayer R-II and $702 in Pettis Co. R-XII — twelve times the money for exactly the same tax. Both fall below the performance levy; both must file notice with the state.

The poorest are the most likely to be caught. Sorted into quartiles by property wealth, 54% of the poorest quarter fall below the performance levy, against 35% of the richest. And 64 districts levy exactly $2.75 — the state-aid cliff described above. Fifty-six of those are rural; their median 2024 child-poverty estimate is 20.0% against a state district median of 15.4%. They are holding the only line the law truly forces them to hold.

And complying would reward the wealthy. If every district below the line raised its levy to exactly $3.43, the poorest quartile would gain a median of $437 per student and the richest quartile $813. Obeying the same law would leave the wealthy district 2.4 times better off than the poor one.

The performance levy measures a tax rate. It does not measure effort, it does not measure wealth, and it does not measure result. Whether a community already taxing itself to its legal limit should be deemed to have made insufficient local effort is not a question this site will settle for you. We present the statute, the arithmetic and the maps; the conclusion is yours. It is, however, a live question: the Governor’s School Funding Modernization Task Force must report by 1 December 2026, and Missouri already has the lowest share of state aid in the country (30% of school revenue) and is 7th in reliance on local property tax.

A note on sourcing. The levy figures on this site are the rates actually levied in 2025, taken from the Missouri State Auditor’s 2025 Property Tax Rates report (No. 2026-006), which reviews every taxing authority in the state and reports operating and debt-service rates separately. We previously used a DESE tax-rate field, and it was wrong in a way we first misdiagnosed. We described it as “the ceiling rather than the levied rate.” On re-examination the dominant fault is vintage: DESE labels its files by school year, so its “2025” figure is tax year 2024. It matches the Auditor’s tax-year-2024 ceiling for 99% of districts, and differs from the 2025 levied rate for 76% of them — because 2025 was a reassessment year and Hancock rollbacks moved almost every rate. It is not true that most districts decline to levy their full ceiling: 89% levy it to the cent. All levy figures here are now the Auditor’s levied rates for tax year 2025. Twenty-one St. Louis County districts levy a separate rate for each class of property (residential, agricultural, commercial, personal). The district’s own operating levy — the figure Missouri law uses — is those rates weighted by the assessed value in each class; the district lookup also shows the residential rate, which is what a homeowner pays.

Why Beating the Odds still uses the free/reduced-lunch rate — a test we ran, and lost. We have said plainly on this site that the free/reduced-lunch rate is no longer a usable poverty measure, so the obvious next step was to rebuild the Beating the Odds regression on the Census child-poverty figure instead. We tried it, and it made the map worse. We pulled SAIPE child poverty for every one of the eleven years the achievement data covers (2009–2019), so the two measures describe the same era, and refit the model. Free/reduced lunch explains 42% of the variation in district achievement level; child poverty explains 21% — half as much. And the weaker control does real damage: with child poverty alone, the median suburban district drifts to +0.06 standard deviations above prediction while the median city district falls to −0.09, and Ladue and Clayton climb into the state’s top over-performers — not because they are doing more with less, but because a model that under-controls for advantage hands the advantaged the credit. Fitting both measures together adds almost nothing (R² 0.434) and the two are so collinear (r = 0.82) that the child-poverty coefficient flips to a nonsensical positive sign.

The explanation is that free/reduced lunch, whatever its faults as a poverty statistic, is a better statistical control: it counts the actual children in the school at 185% of the poverty line, while SAIPE counts every child resident in the district, including those in private schools, at 100%. So the map keeps the free/reduced-lunch model — and, importantly, it uses the 2009–2019 average, the window before community eligibility spread widely enough to break the measure. In that matched file, average FRL correlates −0.648 with achievement level across 516 districts (R² = 0.420) but only −0.171 with learning rate across the 511 districts reporting it (R² = 0.029). Those are the source figures for the Achievement page’s “about 42% versus about 3%” comparison. We are documenting the failed test rather than burying it, because the intuition behind it was right and someone else will have it too.

Debt, building, and cash and investments. The “Who is building” map comes from the U.S. Census Bureau’s Annual Survey of School System Finances (the F-33), fiscal year 2024 — the same survey NCES uses for federal school-finance reporting. Long-term debt is the principal still outstanding at the close of the fiscal year; capital spending is what the district actually laid out on buildings, land and equipment that year; cash and investments are the F-33’s cash and security holdings across sinking, bond and all other funds. That total is not the same as unrestricted operating reserves: sinking funds are set aside for debt service, bond funds can include unspent construction proceeds, and the public file does not divide “other” cash into restricted and unrestricted portions. The displayed cash-to-expenditure ratio is ending cash and investments divided by the F-33’s total expenditure for the same fiscal year — not operating spending and not payroll. Per-student figures use the F-33’s own fall membership count, so numerator and denominator come from the same source. We verified the file’s internal debt identity (debt at the start of the year, plus what was issued, minus what was retired, equals debt at the end) — it holds for 503 of Missouri’s 518 records, the rest differing by a thousand dollars of rounding. We then cross-checked it against an entirely separate source, the State Auditor’s debt-service levies: of the 327 districts that levy for debt, 325 show long-term debt in the F-33. Two special-purpose districts (St. Louis County Special and Pemiscot County Special) are not K-12 districts and are excluded, leaving all 516.

What vote does it actually take? A distinction that matters, and that almost everyone gets wrong. Two different thresholds govern Missouri school ballot questions, and they are routinely confused — including, at first, by us.

The operating levy — the rate that pays teachers — needs only a SIMPLE MAJORITY. Mo. Const. art. X, § 11(c) imposes a two-thirds requirement on municipalities and counties, then carves schools out: in school districts the rate may be raised for school purposes, up to a total levy of $6.00, when “a majority of the qualified electors voting thereon shall vote therefor.” Every one of the 211 districts below the performance levy is far beneath that $6.00 ceiling. A bare majority is all any of them needs. The same section adds that a board whose voters say no “shall be free to resubmit any higher tax rate at any time” — there is no cooling-off period.

A general obligation bond — which builds buildings — needs a supermajority, and how big a supermajority depends on the day it is held. Mo. Const. art. VI, § 26(b): “four-sevenths at the general municipal election day, primary or general elections and two-thirds at all other elections.” So a bond needs 57.14% in April or November — but 66.67% at a February, March, June or August special. The same bond, with the same 60% of the vote, passes in April and fails in June. This is the supermajority most people have in mind when they think about school elections, and it is why a measure can win a clear majority and still lose. It is not the rule for the operating levy.

Why we labour the point: the natural defence of the 211 districts is that Missouri sets an impossible voting bar. It does not. The bar is 50% plus one, and it can be attempted every single year. Across the seven reported tax years, districts cleared it at least 139 times — roughly 20 times a year. Yet 211 districts remain below $3.43, 192 of them are within half a cent of their certified ceiling, and 22 successful districts were still below $3.43 immediately after at least one increase. The corrected record does not show that levy elections almost never work. It shows that winning an election is not the same as having enough property wealth to reach the state’s benchmark — and that Missouri keeps no statewide record of the districts that asked and were refused.

The failed levy elections: a search that came up empty, and what the emptiness means. Our levy-election count sees only the elections that passed, because a win changes the tax rate and tax rates are recorded. A district that asked its voters and was refused leaves no trace. On 13 July 2026 we went looking for the losses, and established — by exhaustion — that no public record of them exists in Missouri at any level above the individual county clerk.

The Secretary of State certifies returns for state offices and statewide measures; local ballot issues are certified by county election authorities and are not compiled by the state. DESE publishes nothing. RSMo 162.201 looks promising — county clerks must certify school election results to the state board of education — but it governs the formation of new districts, not tax levies. Ballotpedia, the only national aggregator, publishes its own scope: local measures within “the 100 largest cities in the United States, and state capitals.” In Missouri that is Kansas City, St. Louis and Jefferson City — 13 of the 211 districts, and blind to the other 198. Oregon County, whose district won a levy election in 2023, has an empty Ballotpedia page. The records exist only in 76 separate county election authorities.

The one compilation we can find is private. The bond underwriters who run these campaigns keep score, because it is their business to know which communities say yes. L.J. Hart & Company advertises “Election Results” as a standing service. We have requested it. If we obtain it we will publish it with its provenance stated plainly — an industry tally, not an official record — and only after reconciling their wins against our Auditor-derived series. If their wins do not reconcile, we will not publish their losses.

The emptiness is itself the finding. From 1 July 2026 the state requires 211 districts to answer for a levy most of them cannot raise, and points them at the ballot box. It records, meticulously, every district that succeeded. It keeps no record at all of who tried. The single fact that would show whether “insufficient local effort” is a fair description of these communities is the one fact Missouri has never thought to collect.

We put the dime test through an independent check — and it held. The dime test is the most quoted thing on this site, and until 13 July 2026 it rested on a single source for assessed valuation. After a run of errors that all shared that same root cause, we went back and tested it against two independent records.

First: every district’s assessed valuation, as used by the dime test, was checked against the State Auditor’s own published figure. All 403 districts we could match agree to within 0.1%. No transcription error.

Second, and this is the real test: we recomputed the entire dime test from DESE’s assessed-valuation file — a completely separate record, certified to DESE by county clerks rather than compiled by the Auditor. The structure survives intact. Naylor R-II is still the property-poorest K–12 district in Missouri ($35 a dime on the DESE figures against $39 on ours), and the gap from poorest to wealthiest is still roughly eighteen to nineteen-fold. Every figure sits about 8% lower on the DESE file for a reason we can name: it is tax year 2024, and 2025 was a reassessment year in which values rose. The two sources tell the same story about Missouri, one year apart.

What the check did turn up. Clayton and Brentwood are effectively tied for the wealthiest K-12 tax base — $716 and $713 a dime. Where this site names Clayton, read it as “Clayton, with Brentwood a few dollars behind,” not as a clear first place. And the Census F-33’s enrollment count disagrees sharply with DESE’s for a handful of very small districts (it reports 158 students in Appleton City R-II where DESE reports 321). The site therefore keeps each numerator with its matching denominator: the dime test and other Auditor/DESE per-student figures use DESE attendance, while the F-33 debt, capital and cash-per-student figures use F-33 fall membership. For DESE current expenditure per pupil, the downloaded MCDS field is reproduced as published because it does not identify which of DESE’s published enrollment-base variants it contains. No denominator is silently substituted for another.

Which rate counts as “the operating levy” — a correction we made to ourselves on 13 July 2026. Twenty-one St. Louis County districts levy a different rate on each class of property: one rate on homes, another on farmland, another on business property, another on cars and equipment. Until 13 July 2026 this site published the residential rate as those districts’ operating levy. For a map of what a homeowner pays, that is the right number. For the $3.43 performance-levy test, it is the wrong one — RSMo 163.011(12) defines a district’s operating levy for school purposes as the levy of the district, not the rate borne by any one class of property.

The district’s levy is its subclass rates weighted by the assessed value sitting in each subclass. We now compute it that way for every district and every year. The method was validated against DESE’s own published operating rate for tax year 2024 and reproduces it exactly — to four decimal places — for every district we could match (406 of 406), including two things our earlier parse had missed: voter-approved temporary operating levies (33 districts; East Buchanan Co. C-1 levies $5.23 plus a $0.64 temporary levy running to 2042) and Kansas City 33’s Article X, Section 11(g) levy, without which Kansas City reads $2.07 instead of $4.86.

What moved: 495 of 516 districts were unaffected — the only ones that change are those 21. Five leave the below-$3.43 group (Ferguson-Florissant, Hancock Place, Mehlville, Pattonville, University City). The count falls from 216 to 211; the share within half a cent of the certified ceiling is now 91%. The finding is stronger than it was, which is not why we changed it. The district lookup shows both numbers for those 21 districts — the levy of the district, and the rate a homeowner actually pays.

A dead end worth recording. RSMo 163.011(15) says “school purposes” pertains to the teachers’ and incidental funds — which reads as though the capital projects fund should be excluded from the operating levy, and excluding it would push the count to 239. It should not be excluded. Five districts land on exactly $2.75 — the state-aid floor set by the same statute — only when capital projects is counted, and on that measure no regular district falls below the floor at all. The rate a district sets is a single operating levy that it then allocates across those funds. We tested the narrower reading and rejected it.

What the tax base is made of — and the one number we will not give you. The four-way map of residential, agricultural, commercial and personal property comes from a records request to Missouri DESE School Finance, answered 13 July 2026. Missouri does not publish district-level assessed valuation by property class anywhere: the State Tax Commission reports county totals only, and the Auditor’s reports carry district total AV. The underlying numbers are certified to DESE by each county clerk. Districts that cross county lines are certified once per county, so we sum — the opposite of the Auditor’s tax-rate exports, where a multi-county district repeats and must be de-duplicated. All 516 districts matched on GEOID with none unaccounted for, and the four classes sum to each district’s certified total. Note the vintage trap that runs through every DESE file: DESE labels by school year, so its “2025” sheet is tax year 2024, which is what this map shows.

Here is what we will not do with it. Missouri assesses residential property at 19% of market value, commercial at 32%, and personal property at 33⅓% — so for those three you can, roughly, work backwards to a market value. You cannot do that for farmland. Actively farmed land is assessed at 12% of its agricultural productive value — a figure set by soil grade under RSMo 137.017, not by what the ground would fetch at auction. Dividing the agricultural column by 0.12 therefore does not recover the market value of Missouri farmland; it recovers a productivity estimate. We built that reconstruction, saw that it would overstate precision while understating the real gap, and threw it away. The correct statement is the one on the funding page: the shortfall between what Missouri farmland is worth and what a school district may tax is larger than these figures can show, and we decline to put a dollar figure on it.

We also tested, and rejected, the intuitive claim that farm-heavy districts are property-poor. The correlation between a district’s farmland share and its assessed value per student is −0.10, and the median AV per student is statistically identical either side of a 10% farmland threshold ($121,800 vs $122,350). Farm districts have thin bases and few students, and the two roughly cancel. The component that does track property wealth is commercial (r = +0.33). We are reporting the failed hypothesis because it is the one most readers will arrive with.

The levy-change map. The “Levy Change · Pick Your Baseline” map is built from the Auditor’s machine-readable tax-rate exports for every year from 2013 to 2025 — 516 districts × 13 years — and shades the rate each district actually levied, never the ceiling. Where a district levies by property class we again take the residential rate. Multi-county districts appear once per county in the Auditor’s file; we de-duplicate rather than sum. (An earlier build of this map used DESE’s tax-rate ceiling, and said so in its caption; it has been rebuilt.)