Apply the IRC § 21(e)(2)/(4) filing-status rules to Wisconsin, New York, and Hawaii CDCC provisions

[hua7450]

Summary

Fixes the three state CDCC bugs confirmed by the statute verification in #8826: Wisconsin, New York, and Hawaii pay their child and dependent care provisions to married-filing-separately filers whose statutes deny them, and Hawaii additionally over-applies its deemed earned income floor. All three are pre-existing gaps — these code paths bypass the federal variables that #8704 gated — surfaced by the post-#8704 state-coupling audit.

Addresses the WI, NY, and HI items in #8826.

Wisconsin — 2024+ credit and subtraction miss the § 21(e)(2)/(4) gate

Wis. Stat. § 71.07(9g)(c)4. and § 71.05(6)(b)43.e. each provide that a claimant "is subject to the special rules in 26 USC 21 (e)(2) and (4)". 2023 Act 101 replaced only the § 21(c) expense cap with Wisconsin's own $10,000 limit — the rest of § 21, including the joint-return requirement, is imported, and the 2024 Schedule WI-2441 instructions limit eligibility to single/HoH/MFJ filers.

The model's pre-2024 path multiplies the federal cdcc_potential and inherited the gate from #8704, but the 2024+ standalone branch of wi_childcare_expense_credit_potential and the wi_childcare_expense_subtraction compute from WI's own caps with the ungated cdcc_rate. Both now multiply by cdcc_filing_status_eligible. Also repairs the subtraction's reference metadata (two URLs were concatenated into one string) and adds the statute link.

New York — replica credit misses the filing-status gate

ny_cdcc rebuilds the federal credit from components (cdcc_relevant_expenses × ny_cdcc_rate × cdcc_rate) without the gate, so an MFS filer kept a NY credit that § 21(e)(2) denies. The IT-216 instructions require qualifying for the federal credit and state the married-filing-separately rule with the considered-unmarried exception (living apart the last six months, qualifying person in the home more than half the year, over half the cost of keeping up the home) — § 21(e)(2)/(4) exactly. ny_cdcc now multiplies by cdcc_filing_status_eligible; nyc_cdcc reads ny_cdcc and inherits the fix.

Hawaii — missing joint-return gate; deeming over-applied; wrong disability standard

HRS § 235-55.6:

• (e)(2): "If the taxpayer is married at the close of the taxable year, the credit shall be allowed … only if the taxpayer and the taxpayer's spouse file a joint return", with the (e)(4) living-apart exception. hi_cdcc_eligible only checked count_cdcc_eligible > 0; it now also requires cdcc_filing_status_eligible (HI's qualifying-individual definition tracks federal § 21(b)(1)/(e)(5), so the federal gate is exact).
• (d)(2): deems earned income only for "a spouse who is a student or a qualified individual described in subsection (b)(1)(C)" — spouse-only, on a married return. hi_cdcc_income_floor_eligible was person-level (is_disabled | is_full_time_student) with no marriage requirement, so a single disabled filer's $0 earnings were floored to $2,400, manufacturing a credit HRS gives as $0. It now mirrors the federal cdcc_income_floor_eligible: head/spouse, married, student or incapable of self-care.
• Disability standard: HRS says "incapable of caring for oneself" — the variable now uses is_incapable_of_self_care (which HI's qualifying-individual count already used via the federal chain) instead of the broader is_disabled.

Confirmed correct and unchanged: HI's $2,400/$4,800 floor amounts ((d)(2)(A)/(B) $200/$400 monthly) and the two-incapacitated-spouses-count-as-one inheritance.

Reproductions (before → after)

Scenario	Before	After
WI 2024, MFS filer, one child, $2,000 expenses	$400	$0
WI 2021 subtraction, MFS filer	$2,000	$0
NY 2022, MFS filer, three children, $10,000 expenses	$3,850	$0
HI 2022, MFS filer with qualifying individuals	positive	$0
HI single filer incapable of self-care, $200 earnings	floored to $2,400	$200
Each state, separated filer maintaining a home for a qualifying child (§ 21(e)(4))	—	credit preserved

Tests

• WI credit (2024 branch), WI subtraction, NY, and HI each gain an MFS → $0 case and a separated-filer § 21(e)(4) carve-out case.
• PA gains the end-to-end MFS regression test deferred from Improve federal CDCC conformance with IRC § 21 (deemed income, qualifying-individual count, filing status) and fix CO/MA/CA coupling #8704 (PA statute verification confirmed Act 53 couples to § 21, including the joint-return rule and the (e)(4)-mirror exception).
• HI floor tests rewritten with explicit head/spouse roles and is_incapable_of_self_care; new regressions for the single-filer and disabled-but-capable-spouse cases; the HI integration suite swaps to the correct disability variable with expectations unchanged and adds a SEPARATE case.
• Ran locally: HI CDCC tree, WI credit/subtraction/credit-order, PA, and the full federal gov/irs/credits/cdcc tree — 132 passed. NY suites left to CI (locally slow per NY EITC/CTC formulas clone the full tax-benefit system on every call — CI memory blowup #8114; the NY change is a single multiplicative gate).

Out of scope (tracked in #8826)

The SC § 21(e)(4) decision item, the VA/ID/OR/MN/WI expense-base conformity gap, and the AR/VT/MD tangential flags.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (f9e58e7) to head (e891e79).
⚠️ Report is 46 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##              main     #8855   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            3         5    +2     
  Lines           55        82   +27     
  Branches         0         1    +1     
=========================================
+ Hits            55        82   +27     

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.