Calibrate ACA take-up contract

[daphnehanse11]

Summary

• derive takes_up_aca_if_eligible from reported Marketplace coverage or a data-supplied/simulated take-up switch
• add simulated_aca_take_up_if_eligible so Populace and other datasets can provide calibrated non-reported Marketplace take-up without writing the final formula-owned take-up variable
• document selected_marketplace_plan_benchmark_ratio as an explicit user/data input, defaulting to the benchmark-plan assumption when absent
• add tests for observed Marketplace take-up and an unsubsidized reported Marketplace buyer using a selected-plan ratio below the benchmark

Closes #8168.

Data framework note

Under the Populace-style contract, has_marketplace_health_coverage_at_interview remains the observed coverage input, simulated_aca_take_up_if_eligible is the data-supplied/calibrated take-up input for tax units without reported Marketplace coverage, and takes_up_aca_if_eligible is computed by PolicyEngine-US. Populace should therefore write simulated_aca_take_up_if_eligible, not takes_up_aca_if_eligible, once this lands.

Tests

• uv run python -m policyengine_core.scripts.policyengine_command test policyengine_us/tests/policy/baseline/gov/aca/takes_up_aca_if_eligible.yaml policyengine_us/tests/policy/baseline/gov/aca/person_receives_aca.yaml policyengine_us/tests/policy/baseline/gov/aca/ptc/marketplace_net_premium.yaml policyengine_us/tests/policy/baseline/gov/aca/ptc/selected_marketplace_plan_premium_proxy.yaml policyengine_us/tests/policy/baseline/gov/aca/ptc/used_aca_ptc.yaml policyengine_us/tests/policy/baseline/gov/aca/ptc/assigned_aca_ptc.yaml -c policyengine_us
• uv run python -m policyengine_core.scripts.policyengine_command test policyengine_us/tests/policy/baseline/gov/aca -c policyengine_us
• uv run pytest policyengine_us/tests/test_parameter_files.py -q
• uv run --extra dev ruff check policyengine_us/variables/gov/aca policyengine_us/tests/policy/baseline/gov/aca
• uv run --extra dev ruff format --check policyengine_us/variables/gov/aca/takes_up_aca_if_eligible.py policyengine_us/variables/gov/aca/simulated_aca_take_up_if_eligible.py policyengine_us/variables/gov/aca/ptc/selected_marketplace_plan_benchmark_ratio.py
• git diff --check

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (206b9e7) to head (09b160d).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##              main     #8352   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            8         3    -5     
  Lines          127        30   -97     
=========================================
- Hits           127        30   -97     

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.

[MaxGhenis]

Reviewed in depth (agent-assisted). Technically clean (green CI on the current head, clean merge against today's main, 22 ACA tests pass post-merge, single-household default stays deterministically true) — but holding merge for a contract decision:

1. The seam contradicts the newer direction. The Feb 2026 deterministic-country-package change made take-up variables formula-less with data-package randomness, and Retire legacy Marketplace coverage input #8693's partner guidance says simulated ACA take-up should write takes_up_aca_if_eligible directly — the opposite of this PR's simulated_aca_take_up_if_eligible pass-through, which exists nowhere else on main or in the newer health PRs. Whichever contract survives should be decided once, and this PR reconciled with Retire legacy Marketplace coverage input #8693 rather than landing a competing seam.

2. No calibration despite the title — no probability parameter, target, or citation; the diff delivers the explicit-input path only. Fine on its own, but Calibrate Marketplace take-up and selected-plan premium ratios #8168's calibrated-take-up half remains open either way.

Suggest folding the observed-coverage-informs-take-up behavior into the #8693 line of work.

[MaxGhenis]

Closing as superseded by the author's own newer contract direction: the Feb 2026 deterministic-country-package architecture made take-up variables formula-less with data-package randomness, and #8693's partner guidance states simulated ACA take-up should write takes_up_aca_if_eligible directly — the opposite of this PR's simulated_aca_take_up_if_eligible pass-through seam, which nothing else on main or in the newer health PRs references. The observed-coverage-informs-take-up behavior is worth folding into the #8693 line of work (detailed findings in the review comment above). Happy to reopen if the contract decision goes the other way.