In March of 2010 President Obama signed into law the Patient Protection and Affordability Care Act (ACA). The legislation requires each state to create, by 2014, an exchange for the buying and selling of health insurance policies, including Medicaid. Through these exchanges, citizens and eligible non-citizens with income below 138% of the Federal Poverty Line (FPL) will be eligible to receive Medicaid coverage. This 138% threshold is considerably higher than the eligibility threshold prior to ACA. Individuals with income between 138% and 400% of the FPL will be eligible for a refundable tax credit for the purchase of private health insurance through the exchange. ACA specifies a uniform income definition for both Medicaid eligibility and health insurance subsidies. This income concept relies on tax code definitions for the determination of eligibility. ACA income, more commonly called Modified Adjusted Gross Income (MAGI), is adjusted gross income (AGI) from IRS Form 1040 plus: (i) tax-exempt interest; (ii) excludable foreign earned income; and (iii) non-taxable Social Security income. In this paper we estimate the distribution of ACA-income using both the well known Current Population Survey (CPS) and administrative tax return data.  Specifically, we compare the number and fraction of people below 138% of FPL and between 138-400% of FPL from the two data sets. We then explore how the ACA legislation might affect the number of people who are Medicaid eligible and health insurance subsidy eligible by current health insurance status.

Estimates from the tax data show that in an expansion year like 2007, 28.0% or 73.3 million individuals under the age of 65 have family ACA-income below 138% of FPL and 39.7% or 103.9 million individuals have family ACA-income between 138 and 400% of FPL. Our CPS-based estimates, after adjusting the data for tax unit structure and ACA income definitions, shows 25.2% or 66.2 million individuals under the age of 65 with family income below 138% of FPL and 38.7% or 101.4 million individuals with family income between 138 and 400% of FPL. This result suggests that using survey response data yields a higher ACA-income distribution than the administrative tax data, implying that estimates using survey data alone will undercount Medicaid eligibility. Estimates using 2009 data show that during the Great Recession the number of individuals with family ACA-income below 138% of FPL increases by about 5 percentage points or 13 million individuals in the tax data.

Although the CPS undercounts the number of people with family ACA-income below 138% of FPL relative to tax data, it contains detailed health insurance status information.  Using the CPS, we find that in 2007 the uninsured comprise about 32% (21.1 million) of the individuals with family ACA-income below 138% of FPL and about 18.1% (18.3 million) of the individuals with family ACA-income between 138% and 400% of FPL.  The uninsured seem to be a sizeable portion of both the Medicaid eligible group and the potential exchange subsidy pool under the ACA legislation.