Income Responses to the Affordable Care Act: Evidence from the Premium Tax Credit Cliff

This paper examines the extent to which taxpayers responded to marginal income incentives that are implicit in the subsidy schedule of the Affordable Care Act (ACA) for taxpayers who purchase health insurance through an ACA marketplace.  Under the ACA, a subsidy is available for the purchase of insurance through a Federal or state marketplace in the form of a tax credit for families with income up to 400% of the Federal Poverty Line (FPL).   This subsidy is calculated by comparing the price of the second-lowest cost silver plan in the marketplace to the maximum a family is expected to pay for health insurance.  If the former is greater than the latter, taxpayers with Modified Adjusted Gross Income (MAGI) below 400% of the Federal Poverty Line (FPL) receive a subsidy for the difference.  Because the subsidy drops discontinuously to zero at 400% FPL, however, this subsidy schedule creates a cliff in the budget constraint for taxpayers who would be eligible for a positive subsidy just below 400% FPL.  Taxpayers, in turn, have an incentive to keep their MAGI below 400% FPL to remain eligible for the positive subsidy amount.

To examine the extent to which taxpayers’ incomes bunched below 400% FPL, we analyze a sample of taxpayers from 2013-2014 drawn from data maintained by the Internal Revenue Service’s (IRS) population of U.S. individual income tax returns. We find a significant amount of bunching below 400% FPL in 2014 that is only apparent among those who purchased insurance from a marketplace, and so would be potentially eligible for a subsidy.  Further, we do not find such bunching among this same group of taxpayers in 2013, when the subsidies were not available.  We see evidence of bunching at the cliff both among the self-employed and among wage and salary workers.  We also calculate potential subsidies at 399% of FPL, and find that those whose income fell right below 400% FPL tended to be eligible for larger subsidies than those whose income fell above the cliff. 

Finally, we examine the manner in which taxpayers lowered their income to keep it below the cliff, and examine whether such a response is consistent with a change in real economic activity, tax avoidance, or evasion. We see some weak evidence of reduction in wages on the left side of the cliff which is suggestive of labor response.  However, the bunching appears to primarily have been driven by increases in contributions to IRA accounts.  Finally, missing tax credit reconciliation forms for taxpayers who received an advance premium tax credit and whose incomes fell above 400% FPL seem to be mostly due to misunderstanding, and not evidence of evasion.