Association for Public Policy Analysis & Management

Starting in 2020, employer-sponsored health insurance (ESI) plans with total health benefits (annual premiums plus contributions to tax-preferred accounts) that exceed about $10,750 for self-only coverage and about $29,000 for other-than-self-only coverage will be subject to an excise tax of forty percent on amounts above these thresholds.  A number of studies, including Herring and Lentz (2011/12) and Miller and Vistnes (2015), have raised important issues related to how well the tax targets plans with overly generous benefits and the potential for increasing incidence of the tax over time.  In this paper, we use data from the 2012 Medical Expenditure Panel Survey Insurance Component (MEPS-IC) to simulate the incidence of the excise tax on ESI enrollees in the private sector.  The richness of the MEPS-IC data allows us to examine single, family and employee-plus-one plans, to include contributions to tax-preferred accounts (HSAs, HRAs and FSAs) in our measure of total health benefits, and to adjust tax thresholds for high-risk professions and the age/sex composition of each employer’s workforce.  We use these data to extend the previous research along two major dimensions.

First, we examine the implications of an alternative approach to determining the cost of other-than-self-only plans.  Previous research has examined the incidence of the excise tax on non-self coverage by focusing solely on family plans (e.g., Herring and Lentz) or by separately analyzing employee-plus-one and family plans (Miller and Vistnes).  However, Treasury and the IRS have outlined an approach that would allow employers to determine the applicable cost of of a health plan by aggregating across all types of other-than-self-only coverage.  To simulate the effect of this proposed approach, we calculate an enrollee-weighted average premium for employee-plus-one and family coverage for each health plan and project the incidence of the tax for this ‘blended’ premium for each year from 2020 to 2029.  In addition, we examine the reduction (increase) in incidence for family (employee-plus-one) enrollees resulting from the aggregation of premiums for the two types of coverage.   

Next, we extend previous analyses of the relationship between plan generosity and the likelihood of exceeding the thresholds for the excise tax.  We begin by examining the distribution of deductibles in plans that exceed the threshold to assess the mix of low and high-deductible plans from 2020 to 2029.  Then, we examine the extent to which plans that are projected to exceed the tax thresholds could avoid the tax if they changed their benefit structure to be in line with that of the 'average' plan. To do this, we use a regression-based approach to predict premiums for each plan with cost-sharing parameters (deductibles, coinsurance rates, out-of-pocket limits, etc.) set to average values.  We use this method to predict premiums for both single and ‘blended’ non-single health plans and project the change in the incidence of the tax should employers alter benefits in this way.  Finally, we examine the distribution of premiums that are above the threshold to provide information on the potential burden of the excise tax.