Context
While reviewing health model cleanup candidates, a few objects appear to be unconsumed by current PolicyEngine-US formulas or data generation, but some are still exposed through metadata/API seed paths. We should audit them before deleting or renaming anything.
Candidates
Old IRS PTC parameter path
gov.irs.credits.premium_tax_credit.eligibility
gov.irs.credits.premium_tax_credit.phase_out.starting_rate
gov.irs.credits.premium_tax_credit.phase_out.ending_rate
Current ACA formulas appear to use gov.aca.required_contribution_percentage.* and gov.aca.ptc_income_eligibility instead. Local searches did not find formula/data consumers for the old IRS path. However, policyengine-api-v2-alpha testing seeds still whitelist gov.irs.credits.premium_tax_credit, so this may be part of an API metadata contract.
ACA enrollment/spending state parameters
gov.aca.enrollment.state
gov.aca.spending.state
These are exposed in generated app metadata, but local searches did not find current formula/data consumers. We should decide whether these are intended calibration/reference targets, public API metadata, or stale parameters.
Medicaid geography table / hidden input follow-up
policyengine_us/parameters/gov/hhs/medicaid/geography/second_lowest_silver_plan_cost.csv
medicaid_rating_area
second_lowest_silver_plan_cost is imported but did not appear to be used by current formulas. medicaid_rating_area is exposed/seeded externally but appears to have no current model consumers. These may be old ACA/Medicaid geography artifacts and should be audited before cleanup.
Suggested resolution
- Confirm whether any API partners depend on these names through metadata, reforms, or explicit inputs.
- If no consumers exist, remove or deprecate them with a changelog entry.
- If they are API contracts, mark them clearly as deprecated/compatibility-only and migrate consumers to the active ACA parameter path.
- Consider adding a lightweight stale-parameter check that distinguishes internal formula usage from public API exposure.
Context
While reviewing health model cleanup candidates, a few objects appear to be unconsumed by current PolicyEngine-US formulas or data generation, but some are still exposed through metadata/API seed paths. We should audit them before deleting or renaming anything.
Candidates
Old IRS PTC parameter path
gov.irs.credits.premium_tax_credit.eligibilitygov.irs.credits.premium_tax_credit.phase_out.starting_rategov.irs.credits.premium_tax_credit.phase_out.ending_rateCurrent ACA formulas appear to use
gov.aca.required_contribution_percentage.*andgov.aca.ptc_income_eligibilityinstead. Local searches did not find formula/data consumers for the old IRS path. However,policyengine-api-v2-alphatesting seeds still whitelistgov.irs.credits.premium_tax_credit, so this may be part of an API metadata contract.ACA enrollment/spending state parameters
gov.aca.enrollment.stategov.aca.spending.stateThese are exposed in generated app metadata, but local searches did not find current formula/data consumers. We should decide whether these are intended calibration/reference targets, public API metadata, or stale parameters.
Medicaid geography table / hidden input follow-up
policyengine_us/parameters/gov/hhs/medicaid/geography/second_lowest_silver_plan_cost.csvmedicaid_rating_areasecond_lowest_silver_plan_costis imported but did not appear to be used by current formulas.medicaid_rating_areais exposed/seeded externally but appears to have no current model consumers. These may be old ACA/Medicaid geography artifacts and should be audited before cleanup.Suggested resolution