Every calculator claims to be accurate. Almost none will tell you what they had wrong last month, which leaves you no way to tell a site that checks its figures from one that has never looked.
So this is the list. It covers changes to figures you could have acted on — a rate, a deduction, a threshold. Where the fix was to our plumbing rather than to a number, it is in the commit history instead, because padding this page would bury the entries that matter.
Each entry names the file that still carries the explanation. A test fails the build if any of those files stops existing, so this page cannot quietly become a list of unverifiable claims.
Want to know about the next one without checking back? Subscribe to the RSS feed — no email address, nothing to sign up for. Or jump straight to a state: ohio · illinois · georgia · pennsylvania · new york · maine · arizona
Local income tax coverage (all states) (this state only)
September 10, 2026
- What was wrong
- The methodology page said local income tax was modeled for 3,521 jurisdictions. That number was the register count plus the individually-checked count, added together — but the two sets overlap. Ohio's register already contains Columbus and Cleveland, which are also among the individually-checked localities, so they were counted twice. The published datasets page, reading the registers directly, said 3,508 at the same time.
- What it says now
- The two are stated separately and never summed: 3,508 jurisdictions from the state registers, and 13 checked one at a time against each locality's own schedule, some of which also appear in a register. No combined total is published, because no source supports one. The register figure is the one used everywhere a single number is needed.
- How it was found
- A routine check while adding a local-tax link to the salary pages. The helper used for the link reported 666 municipalities for Ohio against a register of 663, which is the same overlap showing up somewhere it could be seen.
- Documented in
src/data/tax/localities-2026.ts
Illinois — flagged, not corrected (this state only)
September 3, 2026
- What was wrong
- The wage-garnishment floor may be overstated. This site says protection starts at 45 times the greater of the federal or Illinois minimum wage — $675 a week. The only reachable text of 735 ILCS 5/12-803 says 45 times the FEDERAL minimum wage, which would be $326.25.
- What it says now
- Unchanged, and flagged in the code as unverified. The only copy of the statute reachable from here is an Internet Archive capture from June 2025; the Illinois General Assembly's site does not respond. Rewriting a published figure on the strength of a fourteen-month-old snapshot would be a worse error than leaving a marked one.
- How it was found
- Adding a source quote for the statute, and reading it. Nobody finds this by looking at the file.
- Documented in
src/data/tax/state-garnishment-2026.ts
Georgia (this state only)
August 15, 2026
- What was wrong
- Every figure was a year stale — the rate, the standard deduction and the dependent deduction.
- What it says now
- HB 463, signed 2026-05-11 and retroactive to 2026-01-01: rate 5.19% → 4.99%, standard deduction $15,000 single / $30,000 joint, dependent deduction $4,000 → $5,000.
- How it was found
- A re-read of the cited source. The URL was already correct — the Department publishes two revisions of the same document at the same address, and the older one had been read.
- Documented in
src/data/tax/states/south.ts
Philadelphia, Pennsylvania (this state only)
August 13, 2026
- What was wrong
- The city wage tax was 3.75%, which had been the rate two cuts earlier.
- What it says now
- 3.735% for residents, from the City of Philadelphia's own effective-date table. The rate had fallen twice since the figure was first entered — to 3.74% in 2025 and 3.735% in 2026.
- How it was found
- A locality-by-locality re-check against each city's own publication, rather than a spot check.
- Documented in
src/data/tax/localities-2026.ts
Yonkers, New York (this state only)
August 13, 2026
- What was wrong
- The resident surcharge was 16.775% of New York State income tax.
- What it says now
- 16.75%, per New York State's IT-2105-I (2026) line 22a. The extra digit had no source behind it.
- How it was found
- The same locality re-check.
- Documented in
src/data/tax/localities-2026.ts
Maine (this state only)
August 13, 2026
- What was wrong
- The standard deduction was $8,350 single / $16,700 joint — figures that were not Maine's at all.
- What it says now
- Maine's own 2026 amounts. The wrong pair appeared in more than one state file, which is how a single unsourced number spreads.
- How it was found
- Checking each figure against the state's own withholding tables instead of trusting the value already in the file. It was corrected twice: the first replacement turned out to be Maine's 2025 figures.
- Documented in
src/data/tax/states/northeast.ts
Arizona (this state only)
August 13, 2026
- What was wrong
- The standard deduction was wrong.
- What it says now
- $15,750 single / $23,625 head of household / $31,500 joint, per HB 4168, signed 2026-06-13 and retroactive to tax years after 2024-12-31.
- How it was found
- Tracing the figure to the statute. Arizona ties its deduction to the federal amount under A.R.S. 43-1041(H), so the number moves when a bill rewrites the base.
- Documented in
src/data/tax/states/west.ts
Why one of these says “flagged, not corrected”
Because it is not fixed. The Illinois garnishment floor may be overstated on this site, and the only copy of the statute reachable from here is an Internet Archive capture from June 2025 while the Illinois General Assembly’s own site does not respond. Rewriting a published figure on the strength of a fourteen-month-old snapshot would be a worse error than leaving a marked one, so it is marked.
If you can read 735 ILCS 5/12-803 from a network that reaches ilga.gov, that closes it. The question is whether the floor is 45 times the federal minimum wage or 45 times the greater of the federal and Illinois figures — $326.25 a week against $675.