Skip to content

Correct Utah taxpayer credit income-tax subtraction by year - #10843

Merged
MaxGhenis merged 10 commits into
mainfrom
fix-ut-taxpayer-credit-income-tax
Nov 16, 2026
Merged

MaxGhenis merged 10 commits into
mainfrom
fix-ut-taxpayer-credit-income-tax

Conversation

@MaxGhenis

@MaxGhenis MaxGhenis commented Nov 12, 2026 •

Copy link
Copy Markdown
Contributor

Correct Utah's taxpayer credit income-tax subtraction by tax year and retain the federal section 68 reduction.

Before 2018, claimants subtract income tax allowed in their federal itemized deductions under Utah Code 59-10-1018(2)(a)(ii)(A).

For 2018 through 2025, claimants subtract paid and reported income tax, capped at $10,000 for every filing status. Utah's Legislature adopted that definition in H.B. 293 section 39 and applied it from January 1, 2018 in section 46. See the May 8, 2018 edition, (1)(g) and (1)(h) and the 2023 edition, (1)(e) and (1)(f).

For example, a 2024 joint return with $10,000 of real-estate tax, $8,000 of income tax and $1,000 of sales tax deducts $10,000 of SALT under 2024 Schedule A, line 5e. Utah's deduction base falls from $10,000 to $2,000: $10,000 minus $8,000.

From 2026, claimants subtract the lesser of paid and reported income tax and the income-tax share allowed federally under the 2026 edition, (1)(e). Utah's Legislature applied the amendment from January 1, 2026 in H.B. 77 section 29.

Invariants

The YAML cases and test_ut_taxpayer_credit_income_tax_properties.py check these properties:

  • Sales-tax electors retain the federal itemized total, or the existing reconstruction when that total equals zero. Standard-deduction filers retain the federal standard deduction.
  • Income-tax electors subtract paid and reported income tax up to $10,000 in 2018 through 2025. The difference can remain negative before personal exemptions enter the credit calculation.
  • Before 2018 and from 2026, an income-tax elector's base stays between zero and the effective federal total.
  • Raising sales tax above income tax never lowers the base, including pairs with section 68 reductions.
  • Before 2018, filers exclude protected deductions when allocating section 68 reductions. From 2026, filers allocate the reduction across all itemized categories.

Methodology choices

  • Follow the statute over the TC-40 line 14 instructions for 2018 through 2025. Utah's Tax Commission tells filers to enter Schedule A line 5a in the 2019 instructions, up to $10,000 in the 2023 instructions, and line 5e less lines 5b and 5c in the 2025 instructions. For example, a joint filer with $40,000 of real-estate tax, $15,000 of income tax and $6,000 of other deductions in 2025 keeps $36,000 under the statute and $46,000 under the 2025 instructions. source_conflicts/candidates.yaml lists this question as the open candidate us-ut:taxpayer-credit-income-tax-paid-and-reported, and a maintainer must confirm the reading before registration.
  • Allocate allowed SALT to real-estate tax first, then income tax. Utah's Tax Commission retains that allocation in its October 20 draft of the 2026 instructions, line 12.
  • Before 2018, prorate income tax after section 68 using the 2017 line 14 worksheet, excluding protected deductions from the allocation pool.
  • From 2026, prorate across all itemized deductions because Congress replaced section 68 without category exclusions in P.L. 119-21 section 70111. Utah's Tax Commission does not specify this allocation in its 2026 draft; confirm it when final instructions appear.
  • Keep the signed 2018–2025 difference before adding personal exemptions. A separate filer with $5,000 of federal deductions, $10,000 of income tax and a supplied $6,000 exemption receives an initial credit of 6% × ($5,000 − $10,000 + $6,000) = $60.
  • Retain the existing election inference, tie handling, supplied combined-tax amounts and reconstruction when the federal total equals zero. Under that convention, the model can't tell a true zero from a missing total.
  • Keep the existing simulation cap on SALT at AGI minus exemptions. A forced low-income itemizer can therefore retain a negative Utah deduction base; the final taxpayer credit floors at zero.

Validation

Pull request run 43920478615 at c209c50ae0 passed all 37 jobs on GitHub's merge with main 8e1837790b.

ut_federal_deductions_for_taxpayer_credit.yaml holds 35 cases, and this PR adds cases 16 to 35. A comment on each new case carries its arithmetic.

  • Cases 16 to 18 cover 2017: income tax allowed with no SALT cap, proration after a section 68 reduction, and protected medical, investment-interest and casualty deductions.
  • Cases 19 to 25, 33 and 34 cover 2018 through 2025. Real-estate tax uses up the SALT cap in 2018 and 2024, income tax sits below the cap, and a separate filer exceeds the $5,000 federal cap. Three returns straddle the $10,000 limit, and the rest cover a binding $40,000 cap, the SALT phase-down, a negative base and a reconstructed federal total.
  • Cases 26 to 32 and 35 cover 2026: income tax within allowed SALT, real-estate tax that uses up allowed SALT, the phase-down for single and separate filers, the 2/37 reduction and proration across every category.

The property test runs up to six generated examples and five explicit ones over 2017, 2018, 2024, 2025 and 2026, and pairs every return with a copy that elects sales tax. Its 2017 and 2026 examples each hold a return with a $20,000 total, $4,000 of medical deductions and a $4,000 reduction. Excluding medical deductions in 2017 leaves a Utah base of $8,500, and prorating across every deduction in 2026 leaves $8,000.

This PR's ten captures under source_conflicts/sources/us-ut reproduce complete official documents: four editions of section 59-10-1018, H.B. 293 of 2018, and the TC-40 instructions for 2017, 2019, 2023 and 2025 and the October 20 draft for 2026.

Microsimulation impact

We estimate that this branch raises Utah income tax by $899,308 in 2025 on the default dataset, for 5 sampled tax units (4,411 weighted): tax rises for 4 and falls for 1. Household net income falls by $899,306 across 5 changed households (4,411 weighted), and federal income tax and household benefits don't change. The taxpayer credit after its income phase-out, ut_taxpayer_credit, falls by $899,310 across 6 changed tax units (4,411 weighted): it falls for 5 and rises for 1. The federal deductions that Utah counts toward the credit before that phase-out, ut_federal_deductions_for_taxpayer_credit, fall by $202.5m across 59 changed tax units (64,278 weighted): they rise for 35 and fall for 24. The initial credit, ut_taxpayer_credit_max, falls by $12.1m across the same 59. We compared main at aead96b524 with this branch at c209c50ae0.

Output Main total Change Records changed (weighted)
ut_federal_deductions_for_taxpayer_credit $43.63bn -$202.5m 59 (64,278)
ut_taxpayer_credit_max $2.77bn -$12.1m 59 (64,278)
ut_taxpayer_credit $1.58bn -$899,310 6 (4,411)
ut_income_tax $6.64bn +$899,308 5 (4,411)
state_income_tax $548.1bn +$899,308 5 (4,411)
income_tax $2,422.0bn $0 0 (0)
household_benefits $1,954.9bn $0 0 (0)
household_net_income $16,440.5bn -$899,306 5 (4,411)

In 2026, we estimate that this branch changes no tax unit's Utah income tax and no household's net income on the default dataset. The federal deductions that Utah counts toward the credit fall by $275.5m, for 18 sampled tax units (23,623 weighted), and the initial credit falls by $16.5m, for 16 (23,623 weighted). The credit after its phase-out changes for 1 sampled tax unit, whose weight rounds to 0. We compared the same two commits.

Output Main total Change Records changed (weighted)
ut_federal_deductions_for_taxpayer_credit $44.84bn -$275.5m 18 (23,623)
ut_taxpayer_credit_max $2.84bn -$16.5m 16 (23,623)
ut_taxpayer_credit $1.61bn -$4 1 (0)
ut_income_tax $6.92bn $0 0 (0)
state_income_tax $576.6bn $0 0 (0)
income_tax $2,594.9bn $0 0 (0)
household_benefits $2,060.1bn $0 0 (0)
household_net_income $17,246.4bn $0 0 (0)

axiom: TheAxiomFoundation/rulespec-us#2287 queued (Utah Code 59-10-1018(1)(e) and (1)(f) by tax year; us-ut/statutes/59-10-1018.yaml applies the 2026 text to every year, and the pinned corpus release lacks the earlier editions, so the signed encoder can't date them yet)

🤖 Generated with Claude Code

MaxGhenis and others added 8 commits November 12, 2026 01:09
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@MaxGhenis
MaxGhenis marked this pull request as ready for review November 15, 2026 21:30
@codecov

codecov Bot commented Nov 15, 2026 •

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (8999568) to head (c209c50).
⚠️ Report is 139 commits behind head on main.

Additional details and impacted files
@@            Coverage Diff            @@
##              main    #10843   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            1         1           
  Lines           24        40   +16     
  Branches         0         2    +2     
=========================================
+ Hits            24        40   +16     
Flag Coverage Δ
unittests 100.00% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

MaxGhenis and others added 2 commits November 15, 2026 23:28
The example judged a 2024 return by the 2025 TC-40 instructions. In its
2024 instructions, Utah's Tax Commission tells filers to enter Schedule A
line 5a up to $10,000, so that filer keeps $8,000 under the statute and
under that year's instructions.

Give the 2025 joint return of YAML case 24 instead. With $40,000 of
real-estate tax, $15,000 of income tax and $6,000 of charitable gifts,
the filer keeps $36,000 under the statute and $46,000 under the 2025
instructions.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
@MaxGhenis MaxGhenis closed this Nov 16, 2026
@MaxGhenis MaxGhenis reopened this Nov 16, 2026
@MaxGhenis MaxGhenis closed this Nov 16, 2026
@MaxGhenis MaxGhenis reopened this Nov 16, 2026
@MaxGhenis
MaxGhenis merged commit 99508bd into main Nov 16, 2026
80 of 84 checks passed
@MaxGhenis

Copy link
Copy Markdown
Contributor Author

US + core hub merge audit, us#10843 at c209c50ae0d233b28cfaa301d5c92e86c25c04c6 (squash): Correct Utah taxpayer credit income-tax subtraction by year.

  • Reviews:
    • The PR has no GitHub reviews. Three independent review rounds ran outside GitHub, with fixes after the first and the second.
    • The first round, at 22cf5ffeca, read Utah Code 59-10-1018 on le.utah.gov as it stood before 2018 and in 2018, 2021, 2023 and 2026. It also read H.B. 293 of 2018, H.B. 77 of 2026, P.L. 119-21 section 70111 and the TC-40 instructions for 2017, 2019, 2023 and 2025.
    • It ran nine households on that head and on its merge base 935b24ff4f, and every result matched its hand figures. For example, a 2024 joint return with $10,000 of real estate tax and $8,000 of income tax kept $2,000 at that head and $10,000 at the base.
    • In that round, all 35 YAML cases, the property test and the 103 source-conflict guard tests passed at that head, and 13 of the 35 cases failed against the base's formula. A mutant that protected medical, investment-interest and casualty deductions in 2026 failed one YAML case and survived the property test.
    • It found no defect in the formula or the parameters. It asked for a population estimate, a valid axiom: line, 2018 references on both parameters and medical draws in the property test. It also found that the source-conflict candidate marked its own reading as ruled, and asked for a maintainer's confirmation before registration.
    • The fix round, in four commits ending at fcfead2f0d, changed no formula, parameter value or YAML expectation. It cited H.B. 293 and the May 8, 2018 edition on both parameters and set the candidate to open. It renamed the property file, added medical draws and gave the 2017 and 2026 examples hand-computed expectations of $8,500 and $8,000. It replaced the capture excerpts with ten complete official documents.
    • The second round, at fcfead2f0d, read those four commits, the formula, both parameters, all 35 cases and both versions of the property test. It confirmed the three statutory periods and both effective dates from government sources, recomputed every YAML expectation and both new examples, and compared the ten captures with the official documents. It found no code defect.
    • It ran no test and couldn't read the fix round's logs, so it left those results unaudited. Its one open finding asked for CI at that head on current main, since every job had skipped on the draft. Run 43764047770 later passed at that head, and the run in the gates below passed at c209c50ae0.
    • It also asked for two description edits. I rewrote the Validation section to say what the tests cover and to take pass results from CI. It asked to keep axiom: needed. The description names a queued parity issue instead, since CLAUDE.md reserves needed for external contributors and a maintainer account opened this PR. The description also states the open source-conflict question in full.
    • A check before landing then found that the candidate's example judged a 2024 return by the 2025 instructions. Commit 3bc4bfb7c5 replaced the example with the 2025 joint return of YAML case 24 and changed nothing else.
    • Main had added 53 candidates to source_conflicts/candidates.yaml since the branch's base 935b24ff4f, so commit c209c50ae0 merged main at aead96b524. The merge had no conflict. Against main, the PR changes the same 18 files as before, and only the candidate's example differs from fcfead2f0d.
    • Two checks of those commits compared the merged candidates.yaml with both parents: it holds main's 482 entries byte for byte and this PR's one, with no duplicate id. Both recomputed the example by hand from the 2025 TC-40 instructions, the 2025 Schedule A and the statute's 2023 edition. The first found two stale sentences in the description, which I then corrected, and the second found nothing to fix.
    • The third round, at c209c50ae0, covered those two commits. It recomputed the example from the same three sources and confirmed that the old 2024 return kept $8,000 under the statute and under the 2024 instructions. It counted 430 and 482 candidate ids in the merge's parents and 483 in its result, and it found no other sentence in the PR that judges a return by another year's sources. It checked each of the 38 jobs in run 43804617238 at the head, and all 38 had passed. Its trial merge of the head with main at fd0a748698 had no conflict and held 484 candidate ids, each once. It reported no finding.
  • Methodology choices for maintainers:
    • For 2018 through 2025, the PR follows the statute over the TC-40 line 14 instructions. source_conflicts/candidates.yaml holds that question as the open candidate us-ut:taxpayer-credit-income-tax-paid-and-reported, and no maintainer has confirmed the reading yet. For example, a joint filer with $40,000 of real-estate tax, $15,000 of income tax and $6,000 of other deductions in 2025 keeps $36,000 under the statute and $46,000 under the 2025 instructions.
    • The 2018 through 2025 difference stays signed. No edition of 59-10-1018 floors the Utah itemized deduction, and the TC-40 subtracts line 14 from the sum of the exemption and deduction lines. A separate filer with $5,000 of federal deductions, $10,000 of income tax and a $6,000 exemption gets an initial credit of $60, where a zero floor would give $360.
    • From 2026, the PR prorates the allowed income tax across every itemized deduction after the section 68 reduction. Congress named no excluded category in P.L. 119-21 section 70111. Utah's October 20 draft of the 2026 instructions tells filers in the federal phase-out for state and local taxes to leave out "state income tax that was phased out" and gives no allocation method. A maintainer who keeps the pre-2026 protected categories would expect $73,333.33 in case 35, where the PR expects $72,000.
    • Before 2018 and from 2026, the PR assigns the allowed SALT deduction to real estate tax first and counts the remainder as income tax. The 2025 instructions and the 2026 draft order the taxes the same way.
    • Before 2018, the PR follows the 2017 line 14 worksheet and leaves the protected deductions out of the proration. Main has since added gambling losses to the two federal lists the formula sums (Count gambling winnings in federal gross income and fix the wagering losses deduction #10070 and Correct historical federal taxes and add annual 2017–2026 tests #10854), so the merged formula protects them too, as line 2 of the worksheet does.
    • The model caps SALT at AGI less exemptions, so a low-income itemizer's federal total can fall below the income tax paid, and the 2018 through 2025 base can turn negative. For example, the first review's 2024 single filer with $5,000 of wages, $10,000 of real estate tax and $8,000 of income tax gets a base of −$3,000. The final credit floors at zero.
    • A federal total of zero triggers the existing reconstruction from the other deductions plus allowed SALT, so the model can't tell a true zero from a missing total.
  • Gates checked live at merge:
  • Impact: we estimate that this branch raises Utah income tax by $899,308 in 2025 on the default dataset (5 tax units, 4,411 weighted) and lowers household net income by $899,306, and that it changes neither in 2026. The federal deductions that Utah counts toward the taxpayer credit, before the credit's income phase-out, fall by $202.5m in 2025 (59 tax units, 64,278 weighted) and by $275.5m in 2026 (18 tax units, 23,623 weighted). Main aead96b524 against c209c50ae0.
  • Next:
  • Axiom: Repair Utah Code 59-10-1018: date the taxpayer credit's state or local income tax subtraction by tax year (before 2018, 2018 through 2025, from 2026) (PE-US #10843) TheAxiomFoundation/rulespec-us#2287 queued, a new pe-parity issue to date Utah Code 59-10-1018(1)(e) and (1)(f) by tax year in us-ut/statutes/59-10-1018.yaml, with companion rows from this PR's YAML. That module applies the 2026 text to every year and floors the difference at zero. The pinned corpus release holds the current section alone, and the corpus lacks the edition superseded in May 2018 and both bills, so the signed encoder can't date the earlier text yet.

🤖 Generated with Claude Code

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant