G. Edward Miller Thomas M. Selden Jessica P. Vistnes AHRQ Conference September 10, 2012

1. Affordability of Insurance: Application of ACA Definitions in a Linked Employee-Employer Data Set G. Edward Miller Thomas M. Selden Jessica P. Vistnes AHRQ Conference September 10, 2012

2. Affordable Care Act (ACA) • The ACA will expand access to health insurance by: • Subsidizing Exchange coverage (139-400% FPL) • Expanding Medicaid (≤138% FPL) • Research has found a relationship between employers’ insurance decisions and alternative forms of coverage for employees. • Spousal offers of ESI • Public coverage • The ACA will present employers with important new sources of insurance coverage to consider.

3. Important Information to Assess the ACA • Important categories of Modified Adjusted Gross Family Income (MAGI) in the ACA. • ≤ 138% FPL (Medicaid eligible) • 139-250% FPL (Large Exchange Subsidy) • 251-400% FPL (Small Exchange Subsidy) • > 400% FPL (No Subsidy) • Important to know: • Within-employer distribution of workers’ MAGI. • Alternative sources of coverage. • Premiums • No nationally representative data set has all required data.

4. Employer-Sim Model • Synthetic workforces: • Use household data on workers to ‘populate’ establishment-level data on employers and their health insurance plans. • MEPS Insurance Component (IC): • 2010 data on private establishments • Out-of-Pocket (OOP) premiums for all plans offered • MEPS Household Component (HC): • 2005-2007: pooled data on workers and their families • MAGI: constructed per ACA rules, CPI-adjusted to 2010 $s

5. Linking Workers to Establishments Draw a sample of 300+ workers that match each establishment on these characteristics. Include 100+ low, medium and high wage workers. Workers are sampled with replacement.

6. Raking Workers Weights 0/1 variables for MEPS HC workers Percent distributions in MEPS IC establishments. Iteratively adjust MEPS HC sample weights until worker characteristics match estab. % distributions.

7. Synthetic Workforces Synthetic workforces (in principle) can be used to examine any employee/family characteristic from the MEPS HC. Quality depends on correlation between the HC characteristic and variables used in linking and raking. We do not provide standard errors, which could be quite large given the multiple sources of data and imperfections in the synthetic match Modified adjusted gross incomes are strongly related to many of the variables available to construct synthetic workforces.

8. Goals • Estimate how many employers have a majority of workers who are income-eligible for ACA-related coverage. • Within predominantly subsidy-eligible firms, examine the degree of variation in workers’ incomes. • Examine issues related to the affordability of insurance.

9. Stratify Analysis by Firm-Size • Large firms (50+ full-time workers) • Subject to fines if at least one full-time employee gets subsidized Exchange coverage • Small firms ( < 50 full-time workers) • No penalties for employees accessing Exchange coverage • Tax credits for some firms with < 25 workers (started in 2010)

10. Distribution of establishments, by workers' family income, 2010 Source: Authors’ calculations using data from the Employer-Sim Model Estimates are weighted by MEPS-IC establishment weights. We did not estimate standard errors for point estimates from our simulation model.

11. Distribution of incomes for full-time, eligible workers in establishments that offer ESI, 2010 Source: Authors’ calculations using data from the Employer-Sim Model Estimates are weighted by MEPS-IC establishment weights. We did not estimate standard errors for point estimates from our simulation model.

12. Affordability of Insurance • ACA definition of affordability: • OOP premium for single coverage < 9.5% of MAGI • Our definition of affordable dependent coverage • Dependent coverage if offered • OOP premium for dependent coverage < 9.5% of MAGI • Use the Employer-Sim Model to examine affordability. • Re-weight the data to produce worker-level estimates • Focus on workers with MAGI between 139-400% FPL • Potentially eligible for subsidized Exchange coverage.

13. Affordability (continued) • For full-time workers with MAGI 139-400% FPL, who are offered insurance, we examine the percent of workers with: • Affordable single coverage (legal ACA definition) • ‘Affordable’ dependent coverage from own employer • Examine options for workers with “unaffordable” dependent coverage: • Spousal offer of ESI • Children eligible for public coverage with premium

14. Affordability of Dependent Coverage • Determine needed coverage for each family member: • Assume persons ≤138% FPL, children with free public coverage, and persons with Medicare or Tricare do not need ESI • Others are assumed to need ESI provided by the worker • Count the number of persons in workers’ health insurance eligibility unit (HIEU) who need ESI: • Caveat: MEPS HIEU definition does not account for new ACA rules for dependents up to age 26 • Calculate OOP premium, from own employer, to cover all who need ESI. • Determine whether OOP premium < 9.5% of MAGI.

15. Distribution of OOP Premiums: Workers with MAGI 139%-400% FPL, 2010 Source: Authors’ calculations using data from the Employer-Sim Model Estimates are weighted by MEPS-IC worker-level weights. We did not estimate standard errors for point estimates from our simulation model. **Premiums for dependent coverage are conditional on an offer of dependent coverage.

16. Affordability of Single ESI: Workers with MAGI 139%-400% FPL, 2010 Source: Authors’ calculations using data from the Employer-Sim Model Estimates are weighted by MEPS-IC worker-level weights. We did not estimate standard errors for point estimates from our simulation model.

17. Affordability of Single ESI: Workers with MAGI 139%-400% FPL, 2010 Source: Authors’ calculations using data from the Employer-Sim Model Estimates are weighted by MEPS-IC worker-level weights. We did not estimate standard errors for point estimates from our simulation model.

18. Affordability of Dependent ESI: Workers with MAGI 139%-400% FPL, 2010 Source: Authors’ calculations using data from the Employer-Sim Model Estimates are weighted by MEPS-IC worker-level weights. We did not estimate standard errors for point estimates from our simulation model.

19. Alternative sources of dependent coverage: Workers with MAGI 139%-400% FPL, 2010 Source: Authors’ calculations using data from the Employer-Sim Model Estimates are weighted by MEPS-IC worker-level weights. We did not estimate standard errors for point estimates from our simulation model. **Workers may have both a spousal offer and eligible children, so percentages do not sum to 100%.

20. Limitations • We do not calculate standard errors (SEs). • Appropriate SEs must account for multiple sources of variance: • Matching of MEPS-HC workers to MEPS-IC establishments • Using samples of MEPS-HC workers to estimate summary statistics for each establishment • Sampling variation in the MEPS-IC • Relative standard errors (RSEs) for our point estimates are likely to be large. • Additional sources of error are attributable to the HC-IC linkage process.

21. Summary • ACA-related incentives raise important considerations for private employers regarding the provision of ESI. • Many establishments have a majority of workers who are income-eligible for Medicaid or Exchange coverage. • Within-firm heterogeneity in workers’ incomes complicates employers’ decisions. • Most workers with MAGI 139-400% FPL have access to affordable single coverage • We estimate that about 750,000 workers lacked affordable single coverage in 2010. • For those with lack of access to ‘affordable’ dependent coverage • Nearly half lacked an alternative source of dependent coverage.

22. Further Work • Evaluate the affordability of dependent coverage using all available sources: • Worker’s ESI • Spouse’s ESI • Children’s eligibility for public insurance • Evaluate workers’ net benefits of Exchange coverage by comparing costs of moving to the Exchange with the costs of the most affordable alternative. • Examine the within-establishment concentration of workers who would benefit from Exchange coverage.