Model NY IT-214 2025+ flat credit-amount tables; fix elderly-dependent tier and gate rate

[DTrim99]

Summary

Models the New York IT-214 real property tax credit's 2025+ flat credit-amount tables and fixes two related defects. Closes #9310.

⚠️ Stacked on #9301 (fix/ny-it214-eligibility). This builds on #9301's 2025 federal-AGI income measure and touches the same file, so the diff shows #9301's commit too until that merges. Merge #9301 first, then this rebases to a clean diff.

Changes

1. 2025+ flat amount tables (Part RR of Ch. 59 of the Laws of 2025)

For tax years beginning on/after 2025-01-01 the credit amount is a flat lookup by federal-AGI bracket, with "property tax > FAGI × rate" as a pure gate — replacing the pre-2025 "50% of excess, capped by the 18-row chart" formula. New parameters (dated 2025-01-01):

• amount/elderly.yaml — Table A (65+): $375 / 330 / 300 / 260 / 230 / 200 / 150 by FAGI bracket.
• amount/non_elderly.yaml — Table B (<65): $75 / 70 / 60 / 50 (wider, distinct brackets).
• uses_flat_amount_table.yaml — bool flag flipping true at 2025-01-01 (mirrors Fix NY IT-214 real property tax credit eligibility #9301's uses_household_gross_income).

The formula branches on the flag: 2025+ returns the gated flat amount; ≤2024 is unchanged. Source: 2025 IT-214 Tables A/B.

2. Elderly tier now includes a 65+ dependent

meets_age_condition dropped the ~is_dependent term. Both the 2024 and 2025 forms grant the elderly amount if "you, your spouse, or a dependent you claimed" was 65+ (Line 7). Previously a unit whose only 65+ member was a claimed dependent wrongly got the non-elderly amount. Sources: 2024 IT-214-I, 2025 IT-214.

3. Gate rate fix — excess_real_property_tax single rate (pre-existing bug, folded in)

While implementing the 2025 gate I found excess_real_property_tax was missing type: single_amount, so its .calc() summed the bracket rates (0.12 at $6k, 0.35 at $16k) instead of returning the single Table-1 rate the form specifies (0.045, 0.065). This understated the credit via an inflated income threshold — and it affected 2024 as well as 2025. Added type: single_amount so the gate uses the correct rate. The existing 2024 cases were maximum-cap-bound so their outputs are unchanged, but this corrects the underlying gate for marginal cases in both years.

FAGI	before (summed)	after (Table 1)
$6,000	0.12	0.045
$10,000	0.225	0.055
$16,000	0.35	0.065

Tests

NY RPTC suite 19/19 (.venv-py312). Coverage added: Table A/B brackets across the ladder, the $X,001 boundary transitions, a gate-fails → $0 case, the dependent-65+ case for 2025 and 2024 (locks item 2), a FAGI-over-$18k → $0 case, and a single-rate-gate regression case (FAGI $16k, property tax $2k → $150 under the fix, $0 under the old summed gate).

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (9cff969) to head (a68c725).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##              main     #9313    +/-   ##
==========================================
  Coverage   100.00%   100.00%            
==========================================
  Files            8         1     -7     
  Lines          147        37   -110     
  Branches         0         2     +2     
==========================================
- Hits           147        37   -110     

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.

[hua7450]

Reviewed against the primary sources and pushed a follow-up commit (24272a7, plus a merge of main).

Verified as correct

• Table A/B values and bands match the 2025 IT-214 exactly (page 3), the line 19 gate is the strict inequality the form specifies, and $18,000 FAGI exactly still qualifies ("more than $18,000, stop").
• Dropping ~is_dependent for all years is right: the 2025 form's line 7 says "you, your spouse (if married), or a dependent," and the 2024 form keys on "the qualifying household member 65 or older."
• All test arithmetic re-derives by hand, including the 2024 dependent case (imputed SSI 11,316 → HGI 13,316 → cap 154).

Fixed in the follow-up commit

Band-boundary off-by-one. Every IT-214 income table uses closed integer bands ($0 to 3,000, then 3,001 to 5,000 — Table 1, Tables A/B, and the pre-2025 credit-limit chart, see 2024 IT-214-I pages 2–3), but single_amount scales put a value exactly at a threshold into the next bracket. So FAGI exactly 3,000 got rate 0.040 instead of 0.035 — with $110 of property tax the gate wrongly failed and the credit was $0 instead of $375; at HGI exactly 1,000 the 2024 cap was 73 instead of 75. Fixed by looking up all five scales with calc(..., right=True) over band-top thresholds, with a -.inf first threshold so income of exactly $0 stays in the first band rather than falling below the scale (same pattern as the NYC household credit). Tables A/B are re-keyed from 3_001-style to band tops to match, and a latent 1_0000 (10,000) typo threshold in the non-elderly chart is corrected.

Whole-dollar rounding. IT-214-I page 1: "enter whole dollar amounts only… drop amounts below 50 cents and increase amounts from 50 to 99 cents to the next dollar." Income is now rounded half up (floor(income + 0.5)) before the rate/amount lookups and the $18,000 limit, so e.g. FAGI $3,000.49 correctly stays in the 0–3,000 band and $18,000.30 remains eligible.

Test corrections. Nine regression tests added (five boundary, four rounding). One pre-existing 2022 expectation encoded the boundary bug: at HGI exactly $5,000 the 2022 chart row "4,001 to 5,000" gives a non-elderly cap of 67, not 65; updated accordingly.

Changelog. Rewritten to list the changes; the clause about applying the single Table-1 rate in the gate was removed since that fix already landed on main via 4d669d1 (#9301).

All 29 tests in the credit's test file pass locally after merging main.

[hua7450]

Verified against the 2025 IT-214 form and the 2022/2024 IT-214-I instructions; Table A/B values, the line 19 gate, and the elderly-household-member tier all match the sources. Band-boundary and whole-dollar-rounding corrections pushed directly to the branch with regression tests (see review comment above). All 33 checks green after merging main.