跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
开发与测试

policyengine-healthcare

US healthcare domain knowledge (Medicaid, ACA premium tax credit, CHIP, Medicare) for analysts using the policyengine package. Load for: why ACA PTC and Medicaid changes show $0 in household_net_income (the include_health_benefits_in_net_income gotcha) and how to measure health impact instead; the aca_ptc / medicaid / medicaid_category / slcsp / is_aca_ptc_eligible / per_capita_chip variables; Medicaid category precedence (42 CFR 435.119); the _fc/_nfc split; why a flat reform dict fails on required_contribution_percentage; slcsp returning $0 for the ineligible; and what "IRA reform" means. Triggers: Medicaid, ACA, premium tax credit, PTC, aca_ptc, SLCSP, CHIP, Medicare, MAGI, coverage gap, Medicaid expansion, IRA subsidy extension, required contribution percentage, age curve, family tier, medicaid_category. NOT for: non-health US programs (use policyengine-us); new health variables (use policyengine-model-development); core calculate_household / Simulation mechanics (use the policyeng

PolicyEngine · 已有静态检查记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add PolicyEngine/policyengine-claude --skill policyengine-healthcare

技能说明


name: policyengine-healthcare description: | US healthcare domain knowledge (Medicaid, ACA premium tax credit, CHIP, Medicare) for analysts using the policyengine package. Load for: why ACA PTC and Medicaid changes show $0 in household_net_income (the include_health_benefits_in_net_income gotcha) and how to measure health impact instead; the aca_ptc / medicaid / medicaid_category / slcsp / is_aca_ptc_eligible / per_capita_chip variables; Medicaid category precedence (42 CFR 435.119); the _fc/_nfc split; why a flat reform dict fails on required_contribution_percentage; slcsp returning $0 for the ineligible; and what "IRA reform" means. Triggers: Medicaid, ACA, premium tax credit, PTC, aca_ptc, SLCSP, CHIP, Medicare, MAGI, coverage gap, Medicaid expansion, IRA subsidy extension, required contribution percentage, age curve, family tier, medicaid_category. NOT for: non-health US programs (use policyengine-us); new health variables (use policyengine-model-development); core calculate_household / Simulation mechanics (use the policyengine skill). metadata: category: domain

PolicyEngine US healthcare domain knowledge

Medicaid, the ACA marketplace (premium tax credit), CHIP, and Medicare are the hardest US programs to analyze correctly, because they are modeled as in-kind health benefits rather than cash. This skill is the domain layer for analysts using the policyengine package; for non-health US programs use policyengine-us, for the shared calculation/reform mechanics use the policyengine skill, and for building new health variables use policyengine-model-development.

Verified against policyengine 4.21.0 / policyengine-us 1.764.6 (2026-07). Re-verify variable names and parameter values before reporting.

The one gotcha that dominates everything: health benefits are excluded from net income

gov.simulation.include_health_benefits_in_net_income defaults to False. ACA PTC, Medicaid, CHIP, and Medicare cost flow through household_health_benefits, which the default household_net_income does not include. So a reform that changes aca_ptc or medicaid produces a $0 change in household_net_income — the single biggest time-sink in health analysis. If you score a PTC or Medicaid reform by net-income change and get zero, this is why.

Measure health impact one of these ways instead:

  • household_net_income_including_health_benefits — the same net income with health benefits folded in. Its difference from household_net_income is exactly the household's health-benefit value.
  • The program variable directly — aca_ptc (tax_unit), medicaid (person), summed or differenced across baseline and reform.
  • Population cost: pe.us.calculate_budgetary_impact. Its total already adds the shared-funding health-program cost (Medicaid / CHIP / Medicare Savings Programs) on top of Δtax − Δbenefits, so it captures the health split automatically — do not hand-roll it from household_net_income. See the policyengine skill for the population flow.

Verified end to end (single adult, $35k, Texas — ACA-eligible, not Medicaid-eligible):

import policyengine as pe

r = pe.us.calculate_household(
    people=[{"age": 45, "employment_income": 35_000}],
    tax_unit={"filing_status": "SINGLE"},
    household={"state_code": "TX"},
    year=2026,
    extra_variables=["aca_ptc", "household_net_income_including_health_benefits"],
)
assert round(r.tax_unit.aca_ptc, 2) == 6_250.38          # a real PTC...
# ...but it is NOT in default net income; the gap is exactly the PTC:
gap = (r.household.household_net_income_including_health_benefits
       - r.household.household_net_income)
assert round(gap, 2) == round(r.tax_unit.aca_ptc, 2)

And the toggle itself:

from policyengine_us import CountryTaxBenefitSystem
p = CountryTaxBenefitSystem().parameters
assert p.gov.simulation.include_health_benefits_in_net_income("2026-01-01") == False

Healthcare variables

Entity and period are load-bearing (verified). Note aca_ptc and slcsp are tax-unit; Medicaid/CHIP eligibility is person; slcsp is monthly.

VariableEntityPeriodMeaning
aca_ptctax_unityearACA premium tax credit (annual)
is_aca_ptc_eligiblepersonyearPTC eligibility (income band, no disqualifying coverage)
aca_magitax_unityearMAGI for ACA — adds medicaid_magi
aca_required_contribution_percentagetax_unityearrequired premium contribution rate
slcsptax_unitmonthsecond-lowest silver plan (benchmark) premium
medicaidpersonyearmodeled Medicaid benefit value (per-capita cost when enrolled)
is_medicaid_eligiblepersonyearoverall Medicaid eligibility
medicaid_enrolledpersonyeareligibility × take-up
medicaid_categorypersonyearenum (see precedence below)
medicaid_magitax_unityearMAGI for Medicaid
medicaid_income_levelpersonyearMAGI as a fraction of FPL
per_capita_chippersonyearCHIP benefit per capita
is_chip_eligiblepersonyearCHIP eligibility

Medicaid and ACA income tests use medicaid_magi, not raw AGI (aca_magi delegates to it). The programs are mutually exclusive by construction — is_aca_ptc_eligible excludes people who are Medicaid- or CHIP-eligible — so a Medicaid eligibility change cascades into ACA eligibility and can open a coverage gap (below 100% FPL, no Medicaid and no PTC in non-expansion states).

Medicaid category precedence (42 CFR 435.119)

medicaid_category assigns each person the highest-precedence category they qualify for. Federal law requires mandatory groups before optional ones, and the ACA "not otherwise eligible" rule (42 CFR 435.119) makes adult expansion the residual among mandatory groups. The model's current evaluation or

阅读完整来源 ↗