Follow-up from PolicyEngine/policyengine-us#8099 and PolicyEngine/policyengine-us#8523.
Problem
PolicyEngine-US currently has state CHIP premium formulas, including tax-unit/family mechanics such as household premiums and family caps. But estimating aggregate CHIP premium collections for #8099 requires a calibrated CHIP-enrolled population, not just modeled eligibility or child counts.
The key work is in policyengine-us-data:
- CPS already exposes current CHIP coverage through
NOW_PCHIP.
policyengine-us-data maps that to reported_has_chip_health_coverage_at_interview.
- CHIP take-up is not currently wired like Medicaid take-up.
- CMS Medicaid/CHIP enrollment raw inputs already include
Total CHIP Enrollment by state.
Without CHIP take-up calibration, summing chip_premium by state can overstate collections because takes_up_chip_if_eligible defaults to true in PolicyEngine-US.
Proposed data workflow
- Use CPS
reported_has_chip_health_coverage_at_interview / NOW_PCHIP as a reported-first anchor for CHIP take-up among eligible people.
- Add CHIP to the take-up generation workflow, analogous to Medicaid where reported coverage anchors are prioritized.
- Add CMS
Total CHIP Enrollment by state as a state-level calibration or diagnostic target.
- Generate/validate
takes_up_chip_if_eligible in enhanced CPS outputs.
- Aggregate PolicyEngine-US
chip_premium by state after CHIP take-up calibration.
- Produce a comparison table for CT, DE, MA, MI, and NY:
- CMS Line 49 proxy values from #8099,
- direct CMS M-CHIP cost-sharing offset rows,
- calibrated modeled
chip_premium aggregate,
- and #8099 expected ranges.
Cross-repo dependency
PolicyEngine-US may need a small prerequisite input variable for reported CHIP coverage at interview, analogous to existing health coverage inputs such as has_medicaid_health_coverage_at_interview. That companion issue is PolicyEngine/policyengine-us#8523.
The main calibration and estimation work belongs here because it depends on CPS NOW_PCHIP, persisted take-up inputs, CMS enrollment targets, and the enhanced CPS build/calibration pipeline.
Acceptance criteria
- CHIP take-up generation in US-data prioritizes CPS-reported CHIP coverage where available.
- CHIP enrollment is calibrated or at least reported against CMS
Total CHIP Enrollment by state.
- Enhanced CPS outputs include usable CHIP take-up inputs for PolicyEngine-US simulations.
- A validation/reporting artifact quantifies calibrated modeled CHIP premium collections by state, especially CT/DE/MA/MI/NY.
- The final premium collection estimate uses tax-unit
chip_premium, not premium schedules multiplied by child enrollment counts, so household premiums and family caps are preserved.
Follow-up from PolicyEngine/policyengine-us#8099 and PolicyEngine/policyengine-us#8523.
Problem
PolicyEngine-US currently has state CHIP premium formulas, including tax-unit/family mechanics such as household premiums and family caps. But estimating aggregate CHIP premium collections for #8099 requires a calibrated CHIP-enrolled population, not just modeled eligibility or child counts.
The key work is in
policyengine-us-data:NOW_PCHIP.policyengine-us-datamaps that toreported_has_chip_health_coverage_at_interview.Total CHIP Enrollmentby state.Without CHIP take-up calibration, summing
chip_premiumby state can overstate collections becausetakes_up_chip_if_eligibledefaults to true in PolicyEngine-US.Proposed data workflow
reported_has_chip_health_coverage_at_interview/NOW_PCHIPas a reported-first anchor for CHIP take-up among eligible people.Total CHIP Enrollmentby state as a state-level calibration or diagnostic target.takes_up_chip_if_eligiblein enhanced CPS outputs.chip_premiumby state after CHIP take-up calibration.chip_premiumaggregate,Cross-repo dependency
PolicyEngine-US may need a small prerequisite input variable for reported CHIP coverage at interview, analogous to existing health coverage inputs such as
has_medicaid_health_coverage_at_interview. That companion issue is PolicyEngine/policyengine-us#8523.The main calibration and estimation work belongs here because it depends on CPS
NOW_PCHIP, persisted take-up inputs, CMS enrollment targets, and the enhanced CPS build/calibration pipeline.Acceptance criteria
Total CHIP Enrollmentby state.chip_premium, not premium schedules multiplied by child enrollment counts, so household premiums and family caps are preserved.