Example notebook: federal vs. state Medicaid fiscal impact

Context

PolicyEngine/policyengine-us#8076 adds federal/state cost attribution for Medicaid (FMAP) and CHIP (eFMAP), plus cross-program aggregates federal_benefit_cost / state_benefit_cost. The variables are there; a demonstration on real reforms makes the capability discoverable.

Proposal

Add a notebook to analysis-notebooks that walks through a realistic Medicaid reform (e.g., ACA expansion rollback, or raising the expansion FMAP from 90% back to a pre-ACA split) and shows:

1. Baseline vs. reform aggregate Medicaid cost
2. Breakdown into federal vs. state cost, at the national level and by state
3. Comparison with the single-number budgetary_impact — showing how the aggregate hides the state-level fiscal shift
4. Per-state choropleth of state-borne cost change under the reform

This surfaces the new variables to analysts and establishes a pattern for future fed/state reform scoring (OBBBA SNAP state match starting FY2028, etc.).

Related

• PolicyEngine/policyengine-us#8075 (scoping)
• PolicyEngine/policyengine-us#8076 (Medicaid + CHIP attribution — blocker for this work)
• PolicyEngine/policyengine-api#3481 (API exposure — not a blocker since policyengine.py hits the model directly)

Review the proposed notebook location in analysis-notebooks and the Medicaid and CHIP attribution work in PolicyEngine/policyengine-us#8076 before starting. Use policyengine.py to build a realistic reform comparison, then verify that the notebook shows aggregate, federal, state, single-number budgetary impact, and per-state choropleth results.