The Cadillac Tax: Targeting "Luxury" in a High-Cost World

Starting in 2018, employer-sponsored health insurance benefits with annual premiums exceeding $10,200 for single coverage and $27,500 for family coverage will be subject to an excise tax of forty percent on amounts above these thresholds.  The tax is commonly known as the “Cadillac Tax,” because it is intended to tax overly generous plans that may create incentives for enrollees to over consume medical care.  The imposition of the Cadillac Tax is one mechanism by which the Affordable Care Act attempts to “bend the medical care cost curve” and contribute to the financing of subsidies for Marketplace coverage.  A widely cited analysis by Herring and Lentz (2011/12), however, raises important issues related to the targeting of the tax and the potential for increasing incidence of the tax over time.   

In this paper we use a larger, more recent data set - the 2012 Medical Expenditure Panel Survey Insurance Component (MEPS-IC) - to re-examine these issues and to extend the analysis along several dimensions.  We begin by examining the incidence of the Cadillac Tax for both private and public employers by identifying plans that exceed the single and family thresholds, including employee-plus-one plans which face the family threshold and are, therefore, less likely to be taxed.  To provide information on potential changes in the incidence of the tax over time, we project our data forward to the years 2018-2029, and use a range of assumptions on premium growth rates.  Since the Cadillac Tax targets plans with high premiums, rather than high actuarial values, it may have the unintended effect of imposing penalties on enrollees with high expected medical costs (e.g., older enrollees and those living in high cost areas) or who face high administrative costs.  The large sample size and detailed plan-level data available in the MEPS-IC allow us to examine the targeting of the Cadillac Tax by plan generosity, by workforce and employer characteristics, and by regional and state-level market conditions. 

In the second part of our analysis, we use the EMPSIM model developed in Miller and Selden (2012) to examine the per-worker tax and the total revenue generated by the Cadillac Tax. The EMPSIM model uses data on workers from the MEPS-Household Component to construct synthetic workforces for each establishment in the MEPS-IC.  The resulting database contains information on the full distribution of workers’ Federal, state and FICA marginal tax rates and the premiums for the health insurance policies they hold.  We use this information to estimate the baseline tax subsidy to employer-sponsored health insurance, the revenue generated by the Cadillac Tax and the net tax subsidy.  In addition to national estimates, we examine the incidence of subsidies and taxes by employer, workforce and geographic characteristics.  

Herring and Lentz (2011/12) and other commentators have raised important issues about the Cadillac Tax both in terms of appropriate targeting at a point in time and the increasing proportion of plans and enrollees subject to the tax over time. Our analysis seeks to provide additional information to policymakers as they address these concerns.