Premium Subsidies, the Mandate, and Medicaid: Coverage Expansion and Labor Market Impacts Under the Affordable Care Act

Using a combination of subsidized premiums for Marketplace coverage, an individual mandate, and expanded Medicaid eligibility, the Affordable Care Act (ACA) has significantly increased insurance coverage rates, with a drop in the non-elderly uninsured rate from 17.0% in 2013 to 13.8% in 2014 according to the Census data.  The relative contributions of different policy levers to these coverage gains is unknown, and there has also been significant speculation about the ACA’s labor market effects.  We assessed the impacts on insurance coverage and labor market outcomes of these different ACA provisions in the law’s first full year.

Our analysis combined rating-area level premium data for all 50 states plus Washington D.C. with details on the law’s premium subsidy scheme, mandate penalties, and Medicaid eligibility. The primary dataset was the publicly-available microdata from the 2012-2014 American Community Survey.  We mapped the ACA’s insurance rating areas onto the Census survey’s public-use microdata areas (PUMA), population centers of roughly 100,000 individuals each.  We then applied state-specific age-adjusted premium curves for each family unit and the ACA’s premium rules to construct two distinct measures of premium subsidies, both pegged to the second-lowest cost silver plan in each rating area: net premiums in dollars (the per-person cost of coverage to a family net of subsidies, including Medicaid eligibility); and the percentage subsidy (1 minus the ratio of net premium to the full unsubsidized cost of coverage).  We derived the family-level mandate penalty based on family income and the ACA’s various mandate exemptions.  Finally, we modeled both pre-ACA Medicaid eligibility for adults and children, as well as the expansion eligibility in participating states for 2014. 

We examined the following dependent variables: any insurance, Medicaid, private insurance, labor force participation, and hours worked per week (the latter two only for those ages 19-64).  The identification strategy used within-rating area variation in subsidies, mandate penalty, and eligibility, and then differenced out the baseline relationship between those policy variables and the dependent variable using interaction terms between each policy and the year 2014.  Insurance analyses controlled for family income as a percentage of federal poverty (with polynomial terms up to the 4^th power and dummy indicators for ACA-relevant income bands), household structure, demographics, year, and rating area/PUMA.  Labor market analyses controlled for the same set of covariates but used predicted income, rather than observed income. 

Our preliminary results show that overall coverage changes in 2014 were more responsive to the percentage subsidy than net premiums, with each 20% increase in subsidy reducing the uninsured rate by 0.35 percentage points (or 1.0 million people).  The mandate penalty had modest and inconsistent effects on coverage, primarily encouraging insurance take-up among those whose incomes were too high to receive subsidies.  The Medicaid expansion increased Medicaid coverage among newly-eligible individuals by 4.5-5.0 percentage points in 2014, with moderate private insurance crowd-out.  Finally, we found no significant labor market changes associated with the ACA’s policies on either the extensive or intensive margins.