Reconstruct gross CHIP per-capita value for premium netting

[MaxGhenis]

Prerequisite for cliff-watch CHIP premium handling and for #8089's MOOP architecture.

Problem

per_capita_chip is sourced from MACPAC Exhibit 33 "CHIP Spending by State". MACPAC aggregates net CHIP expenditures — specifically, federal + state shares equal gross expenditures minus the "Premiums: Cost Sharing Offset" lines on CMS-21 (the form states file quarterly). So per_capita_chip already has household premium revenue netted out.

Consumers that track chip_premium separately as a household-paid cost (e.g. the in-flight cliff calculator, SPM-style MOOP accounting) would double-count if they subtracted chip_premium on top of a chip value that already netted it.

Fix

Add CMS-21 cost-sharing offset figures as calibration parameters and expose a gross counterpart:

• parameters/calibration/gov/hhs/cms/chip/cost_sharing_offsets/separate_chip.yaml — FY2024 family-paid collections by state, separate-CHIP sheet
• parameters/calibration/gov/hhs/cms/chip/cost_sharing_offsets/medicaid_expansion_chip.yaml — zeros placeholder for M-CHIP states (CT, DE, MA, MI, NY) pending a second-pass parse of the M-CHIP sheets
• parameters/calibration/gov/hhs/cms/chip/cost_sharing_offsets/total.yaml — sum (currently equals separate_chip until M-CHIP lands)
• per_capita_chip_gross — per-enrollee gross = net spending + offset, divided by enrollment
• chip_gross — person-level sum, counterpart to existing chip

Data notes

• Top charging states (FY2024 separate-CHIP): FL $19.9M, MO $12.7M, WA $9.6M, AL $5.5M, KS $4.7M, TX $2.8M, IA $2.7M.
• National total ~$63M/yr from separate CHIP (FY2024).
• M-CHIP placeholder: 5 states (CT, DE, MA, MI, NY) run CHIP as M-CHIP and file their premium collections on the M-CHIP CMS-21 sheets, not the separate-CHIP sheets. Zeros here now, to be filled by a follow-up PR that reads from FY 2024 FMR CHIP NET EXPENDITURES.xlsx M-CHIP tabs. NY Child Health Plus especially should have substantial revenue.
• States with premium schedules but zero FY2024 collections (AZ, IL, CA M-CHIP, GA through Sept 2024, UT after July 2024): pandemic suspensions / program transitions / recent eliminations. Noted in description.

Source

CMS Medicaid Financial Management Report — FY 2024 CHIP Net Expenditures Excel workbook: https://www.medicaid.gov/medicaid/financial-management/state-expenditure-reporting-for-medicaid-chip/expenditure-reports-mbescbes

CMS-21 form layout (cost-sharing offset lines): https://www.medicaid.gov/medicaid-chip-program-information/by-topics/data-and-systems/downloads/collection-systems/chip-cms21-expenditure-forms.pdf

Testing

3/3 new tests pass. 57/57 CHIP tests pass. make format / ruff check clean.

Test plan

• Tests pass locally
• Format / lint clean
• CI passes

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (05d5d97) to head (745af58).
⚠️ Report is 32 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##              main     #8094    +/-   ##
==========================================
  Coverage   100.00%   100.00%            
==========================================
  Files           18         2    -16     
  Lines          353        32   -321     
  Branches         4         0     -4     
==========================================
- Hits           353        32   -321     

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 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.