Bump policyengine-us to >=1.711.0; pin otherprop→NIIT behavior

[PavelMakarchuk]

Summary

• Bump policyengine-us dependency from >=1.552.0 to >=1.711.0 in pyproject.toml. The 1.711.0+ line includes Unwire UT Homeowner/Renter Relief from income tax refundable credits policyengine-us#8564, which unwires the Utah Homeowner/Renter Relief credit from TC-40 refundable credits.
• Add tests/test_otherprop_niit.py pinning the otherprop → rental_income mapping and NIIT behavior introduced in Route otherprop to rental_income for correct NIIT, suppress auto-QBID #961 to TAXSIM-35 and statute.

Why the bump

PE-US 1.710.6 (our current pin) carried a stale gov.states.ut.tax.income.credits.refundable parameter that listed ut_homeowner_renter_relief. That program is administered on Form TC-90CB (Renter) / Form TC-90H (Homeowner) — separate refund applications under Utah Code §§ 59-2A-205 and 59-2A-305 — not on the TC-40 income-tax return. The Utah Tax Commission's 2025 TC-40 instructions confirm: no reference to TC-90CB/TC-90H anywhere in the income-tax flow.

This caused a $1,412 phantom refundable credit for low-income UT seniors in the eCPS comparison (surfaced while investigating PR #961's regressions). The fix landed in PE-US main on June 2, 2026 as PR #8564. The credit continues to flow correctly via the household-level state_property_tax_credits aggregate, so TAXSIM sptcr (v40) is unaffected.

Why the new test

PR #961 routed TAXSIM otherprop to PE-US rental_income so passive rents/royalties enter the NIIT base, and disabled the auto-QBID gate. The behavior is anchored to:

• IRC § 1411(c)(1)(A)(i): NII includes "gross income from interest, dividends, annuities, royalties, and rents".
• Form 8960 Line 4a: "Income From Trades/Businesses/Farming, Rental Real Estate, Royalties, Partnerships, S Corporations, and Trusts".
• IRC § 199A(d)(1): QBID requires a "qualified trade or business". Passive individual rentals typically do not meet the § 162 trade-or-business standard absent the § 1.199A-1(b)(14) safe harbor that TAXSIM input cannot signal.
• TAXSIM-35 binary smoke test (single, age 45, otherprop=$1M, 2024): fiitax=\$353,188, niit=\$30,400 (= 3.8% × ($1M − $200K)).

Three pinning tests:

1. test_otherprop_drives_niit_at_3_8_percent — asserts NIIT = $30,400 ±$10. If the YAML mapping ever drops otherprop from rental_income, this fails.
2. test_otherprop_does_not_trigger_qbid — asserts fiitax ≈ $353,188 ±$50. If the QBID gate override regresses, PE applies ~$170K of QBID and the fiitax drops well outside tolerance.
3. test_no_otherprop_no_niit_on_wages — control: wages-only filer must owe no NIIT.

Verification

• All 3 new tests pass locally
• Full pytest tests/ suite passes locally on the bumped PE-US (1.720.6 resolved from >=1.711.0)
• Direct check on the two UT records that regressed in PR Route otherprop to rental_income for correct NIIT, suppress auto-QBID #961's eCPS run: siitax returns to \$0 (matching TAXSIM)

Test plan

• Three new tests pass
• Full suite passes
• UT regression closes on the specific records that surfaced it

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
policyengine-taxsim	Ready	Preview, Comment	Jun 4, 2026 6:59pm