Update PolicyEngine US to 1.676.1

[policyengine]

Summary

Update PolicyEngine US from 1.653.3 to 1.676.1.

What changed (1.653.3 -> 1.676.1)

Added

• Added the Maine millionaire income tax surcharge, affordability payments, and 2026 Property Tax Fairness Credit cap increase.
• Pennsylvania Unemployment Compensation.
• Added Michigan State Supplementary Payment (SSP) program.
• Implement Louisiana Optional State Supplement (OSS).
• Connecticut, Nebraska, and North Dakota stillbirth tax provisions effective 2022.
• Added Idaho conformity to the federal enhanced senior deduction through Idaho's qualified business income and federal Schedule 1-A deduction path starting in 2025.
• Add Maine LD 229 income tax bracket expansion reform (3 to 5 brackets).
• Added Basic Health Program (ACA §1331) coverage modeling. Creates a shared gov/hhs/basic_health_program coverage path covering MN MinnesotaCare, NY Essential Plan, OR OHP Bridge (2025+), and DC Healthy DC Plan (2026+). BHP-eligible people are excluded from ACA Premium Tax Credit eligibility, so per-state PTC projections now zero out for BHP enrollees. Eligibility enforces the statutory 133–200% FPL band (250% FPL in NY from 2025). Moves the state adult/parent Medicaid income ceilings back to Medicaid-only levels in BHP states so people above the Medicaid line fall through to BHP rather than silently losing coverage. Refs #8112.
• • Added 2026 ACA lowest-cost bronze plan (lcbp) rating-area premiums and bronze companion variables.
• Add shared CPS tax-unit construction helper rules.
• North Carolina Child and Dependent Care Credit contrib reform (refundable, 30% match of federal credit, 2026+).
• Add Florida Optional State Supplementation (OSS).
• Add Washington millionaires tax (9.9% on income above $1M) and WFTC age expansion (18+) per ESSB 6346. Removes contributed reform since bill is now law.
• Maryland military retirement income subtraction (Md. Code Tax-General § 10-207(q)) from 2021 onward, including the Keep Our Heroes Home Act increase to $12,500 / $20,000 effective 2023.
• Add California AB 2591 contributed reform: standard deduction at the federal poverty line.
• Added Kansas State Supplemental Payment Program (SSPP) for SSI recipients in Medicaid-approved institutions.
• Add DC Optional State Supplemental Payment (OSSP) program.
• Added medicaid_premium federal aggregator variable plus Indiana HIP POWER Account contribution (in_hip_power_account_contribution), and wired medicaid_premium into spm_unit_medical_out_of_pocket_expenses. Collection is currently paused in Indiana; the encoded schedule is reform-ready.
• Added marketplace_net_premium variable — annual Marketplace plan premium net of applied PTC.
• Added per_capita_chip_gross and chip_gross variables that reconstruct gross CHIP service value by adding CMS-21 state cost-sharing offsets to the net-of-premium MACPAC spending.
• Added household_health_costs aggregate; chip_premium now reduces household_net_income when health benefits are counted as income.
• Added chip_premium variable with CHIP premium and enrollment-fee implementations for 17 states (AL, CT, DE, FL, GA, IA, ID, IL, IN, KS, LA, MA, MI, MO, NY, TX, WI).
• Model CHIP enrollment separately from eligibility via chip_enrolled, takes_up_chip_if_eligible, chip_take_up_seed, and gov.hhs.chip.takeup_rate (default 0.93), matching the Medicaid takeup pattern. chip, chip_federal_share, chip_federal_cost, and chip_state_cost now gate on enrollment rather than raw eligibility.
• Federal/state cost attribution for Medicaid (FMAP per 42 USC 1396d) and CHIP (enhanced FMAP per 42 USC 1397ee), plus cross-program federal_benefit_cost and state_benefit_cost aggregates.

Changed

• SPM unit medical out-of-pocket expenses now uses rules-based income_adjusted_part_b_premium (base + IRMAA) in place of the imputed medicare_part_b_premiums, so reforms to the Medicare Part B base premium or IRMAA thresholds propagate through SPM resources.
• Added spm_unit_medical_out_of_pocket_expenses that combines person-level imputed MOOP with rules-based tax-unit-level premiums (currently chip_premium). spm_unit_spm_expenses now reads the wrapper instead of listing MOOP and each premium separately.
• Include chip_premium in spm_unit_spm_expenses so CHIP premium reforms propagate to SPM resources.
• Bump policyengine-core floor to 3.24.4 (closes #8066).

Fixed

• Reconciled FLSA overtime regular-rate calculations with annual employment income.
• Fix state benchmark tests affected by modeled Medicare enrollment inputs.
• Split medical expenses into program-specific variables and remove the generic medical out-of-pocket expense aggregate.
• Fix Marketplace selected-plan and net premiums for unsubsidized households.
• Exclude children with employer-sponsored insurance from CHIP eligibility.
• Fix farm operations income, business self-employment tax, and labor supply response inputs.
• Backfilled Medicaid and CHIP income limit parameters to 2018 using official CMS sources, including thresholds stored with the 5 percentage-point MAGI disregard.
• Removed the one-time Connecticut child tax rebate from the ongoing state Child Tax Credits catalog, eliminating a phantom $250 per child value in taxsim state CTC output for 2023-2026.
• Cap the Oklahoma earned income tax credit at remaining tax liability in years where it is non-refundable (2016-2021).
• Update SSI state supplement program registry to mark CT, DC, ID, IN, NM, SC, and TX as complete.
• Fix test_batched.py silently dropping --batches N and --exclude for folders without subdirectories, which caused duplicated and under-parallelized CI test runs.
• Fixed MT state_income_tax returning $0 for 2024+ by replacing min_(indiv, joint) with where(filing_separately, indiv, joint) in mt_income_tax_before_refundable_credits_unit.
• Fixed OH senior citizen credit to check both head and spouse age per ORC 5747.055.
• Split heavy CI test jobs into parallel matrix runners with 60-min timeouts to avoid memory exhaustion on 16 GB GitHub Actions runners.
• Rewrite federal_eitc_without_age_minimum to compose the federal EITC without mutating the shared tax-benefit system, eliminating an order-dependent test pollution bug in Maryland EITC calculations.
• Reverted the 2029 sunset from senior_deduction/amount.yaml and removed its corresponding test — the sunset is handled at the deduction aggregator level, and zeroing the amount parameter broke the CRFB senior_deduction_extension tests.
• Fixed the federal senior deduction phase-out to apply per-senior rather than to the combined total, matching IRS Schedule 1-A Part V.
• Split the Full Suite YAML batch grouping in tests/test_batched.py so each heavy folder runs in its own isolated subprocess. Previous grouping (3 folders per contrib batch, paired gov/ folders in the baseline-other job) pushed peak memory to ~8-9 GB per subprocess, which OOMed the 16 GB ubuntu-latest runner once the policyengine-core 3.24+ overhead landed and surfaced as The runner has received a shutdown signal mid-batch. Each batch now targets ≤5 GB peak, eliminating the intermittent Full Suite failures without changing test coverage.
• Added historical effective-date brackets for MO and NY CHIP premium schedules so chip_premium returns accurate 2020-2026 values for the 2024 data-imputation baseline anchor.
• Bump policyengine-core minimum to >=3.25.0 to pick up the set_input preservation fix from Preserve set_input values across apply_reform policyengine-core#475. Addresses the state_fips / tax_unit_itemizes / household-dataset regression (Add regression test for parameter breakdown mismatch (closes #8055) policyengine-us#8058) where apply_reform wiped user-provided dataset inputs during structural reforms.
• Excluded earnings of CT Care 4 Kids family members under 18 who are not parents of the applicant child from countable income, per RCSA 17b-749-05(b)(2)(E).
• Excluded minor children's earnings from NJ CCAP countable income, per the CCAP application form (CC-1) which collects earnings only from applicant/co-applicant.
• Excluded earnings of VA CCSP household members under 18 from countable income, per 8VAC20-790-40(H).
• Documented that VA CCSP family day home rates vary by region only, not by age group, with new regression tests.

---

Generated automatically by GitHub Actions

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.66%. Comparing base (1690e20) to head (4558326).
⚠️ Report is 2 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #3532   +/-   ##
=======================================
  Coverage   76.66%   76.66%           
=======================================
  Files          63       63           
  Lines        3446     3446           
  Branches      621      621           
=======================================
  Hits         2642     2642           
  Misses        629      629           
  Partials      175      175           

☔ View full report in Codecov by Sentry.
📢 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.