What Can We Expect from the "Cadillac Tax" in 2018 and Beyond?

Beginning of article

One controversial aspect of the Patient Protection and Affordable Care Act is the provision to impose a 40% excise tax on insurance benefits above a certain threshold, commonly referred to as the "Cadillac tax." We use the Employer Health Benefits Survey, sponsored by the Kaiser Family Foundation and Health Research and Educational Trust, to examine the number and characteristics of plans that likely will be affected. We estimate that about 16% of plans will incur the tax upon implementation in 2018, while about 75% of plans will incur the tax a decade later due to the indexing of the tax thresholds with the Consumer Price Index. If the Cadillac tax is ultimately implemented as written, we find that it will likely reduce private health care benefits by. 7% in 2018 and 3.1% in 2029, and will likely raise about $931 billion in revenue over the ensuing 10-year budget window from 2020 to 2029.

**********

One of the more controversial provisions within the 2010 Patient Protection and Affordable Care Act (ACA) is the excise tax on high-cost private health insurance plans, commonly referred to as the "Cadillac tax." Starting on Jan. 1, 2018, a 40% excise tax will be imposed on employment-based health insurance plans whose value exceeds $10,200 for single coverage and $27,500 for family coverage; both premiums and amounts deposited into tax-preferred accounts will count towards the value of the benefits. The tax will equal 40% of the difference between the value of the benefits and these threshold levels, though the thresholds will be increased for certain groups. While the excise tax payments are expected to come to the U.S. Treasury directly from private health insurers and self-insured employers, the private health plans are almost certain to indirectly pass along the costs of paying the excise tax to employers as relatively higher premiums.

During the health care reform debate, the Cadillac tax was heavily opposed by many progressives and by labor union members who historically had negotiated for more generous benefits in lieu of higher wages. (Many Republicans also appeared to oppose the Cadillac tax, but it is difficult to isolate their opposition to the tax from the overall bill.) The excise tax was in the initial Senate version of the legislation but was not included in the initial House of Representatives version of the legislation. The amendment to reconcile the Senate and House versions of the legislation postponed the implementation of the Senate's excise tax proposal from 2013 to 2018. The White House reportedly advocated for keeping the Cadillac tax, primarily because of its desire that the health care legislation, over the long term, be scored to slow the growth in health care spending and reduce the federal budget deficit (Montgomery and Shear 2010).

As with other excise taxes, a main objective of the Cadillac tax, besides simply raising revenue, is to change consumer behavior. Just as cigarette taxes aim to discourage people from smoking, the Cadillac tax aims to discourage people from obtaining overly generous, "luxurious," or "gold-plated" health insurance coverage. Health economists have long argued that the tax subsidy for employment-based insurance is both inefficient and inequitable (Pauly 1986). Because employment-based health insurance premiums are not subject to income and payroll taxes, workers have an incentive to shift their compensation from taxable wages to tax-exempt benefits. The more generous health benefits that result then have the potential to increase the utilization of low-value medical care by further insulating patients from the true costs of that care (although research to date has not consistently shown that higher cost-sharing reduces the use of low-value care more than high-value care, where "value" here denotes the underlying cost effectiveness of the health care service). The inequity of the current tax subsidy results from higher-income people receiving a larger subsidy towards their health insurance premiums because of their higher marginal tax rates.

Proponents of the Cadillac tax have argued that imposing this excise tax on high-cost plans would mitigate the incentive of the current tax subsidy to obtain generous plans because the excise tax's increase in the net premium would offset the decrease received through the tax subsidy. With that underlying objective, though, one might argue that a more transparent alternative would have been to simply limit the value of the tax subsidy directly (e.g., a cap on the premium allowed to receive the subsidy as initially considered in the Senate Finance Committee). (1) Politically, however, the 40% excise tax could be framed as a tax increase on private insurers rather than a tax increase on middle-class workers. Regardless of these political optics, some have raised concerns that the Cadillac tax is imperfectly targeted, since health insurance premiums vary due to a number of factors beyond a plan's actuarial-related attributes (Gabel et al. 2010). (2)

Another controversial aspect of the Cadillac tax is how the thresholds will increase over time after it is implemented in 2018. The thresholds will rise by the Consumer Price Index (CPI) plus 1% in 2019, and then will increase by the CPI alone in 2020 and beyond. Given that historical growth rates for private health insurance premiums are considerably higher than general price inflation in the economy, more and more people will be affected by the excise tax over time, perhaps leading to growing public discontent with the tax.

To explore the likely effects of the Cadillac tax surrounding these various issues, we examine recent data for employment-based health insurance benefits projected to 2018 and onward. We estimate how many plans will likely cross the excise tax threshold in 2018 and beyond, and we examine the characteristics of the types of plans that are expected to be taxed. We then explore the extent to which employers and employees will respond to the excise tax by choosing less-expensive plans with lower actuarial values. Finally, considering the Congressional Budget Office (CBO) has tentatively suggested that the Affordable Care Act could reduce federal budget deficits by roughly $1.35 trillion in the ensuing decade after the initial 10-year budget window, we seek to determine how much projected deficit reduction might be attributed to the Cadillac tax. (3)

How Many Plans Will Cross the Threshold?

We begin our analysis of the potential future effects of the Cadillac tax by examining plan-level data for private health insurance benefits to see what proportion of plans are likely to exceed the excise tax threshold. The Cadillac tax will be applied to premiums as well as to health savings accounts (HSAs), health reimbursement accounts (HRAs), and flexible spending accounts (FSAs). (4) The excise tax will be applicable to employment-based plans purchased directly from insurers, employment-based plans purchased through state-based exchanges, and plans provided by employers who self-insure. The underlying objective is to have the excise tax apply to all employment-based tax-exempt health care spending, including HSAs, HRAs, and FSAs.

We primarily use data from the Kaiser Family Foundation and Health Research and Educational Trust's (KFF-HRET) Employer Health Benefits Survey for years 2008 and 2009. The KFF-HRET survey collects data from a nationally representative random sample of about 2,000 employers per year. Included in the KFF-HRET survey are characteristics of the employer, its workforce, and each health insurance plan it offers. Premium data are collected for single coverage and family-of-four coverage for each type of plan a firm offers (e.g., the premium for the preferred provider organization [PPO] and the premium for the health maintenance organization [HMO] if a firm offers both). We show our results for single coverage and family-of-four coverage separately; there are limitations to our analysis associated with a lack of data in the KFFHRET data on so-called "employee-plus-one" plans, and the nature of these limitations are highlighted where applicable. (5) We transform the firm-level data to plan-level data to account for firms that offer multiple plans; we weight each plan by its enrollment, yielding a pooled sample across the two years of 5,369 single plans and 5,369 family-of-four plans.

The KFF-HRET survey includes data for total premiums and, starting in 2008, for employer contributions to both HSAs and HRAs. For average employee contributions to HSAs in 2008 and 2009, we incorporate estimates from the Employee Benefit Research Institute (2009). For employee contributions to FSAs, we incorporate unpublished estimates for FSA offer rates by establishment size from the 2008 Medical Expenditure Panel Survey's Insurance Component (MEPS-IC) and estimates of FSA take-up rates and average contributions from Mercer's 2008 National Survey of Employer-Sponsored Health Plans. (6)

We inflate the values for these premiums and accounts to values in 2018 and beyond by assuming a 6% annual growth rate, consistent with historical averages seen in the MEPS-IC and future projections in the National Health Expenditure Accounts data from the Office of the Actuary at the Centers for Medicare and Medicare Services (CMS). We also demonstrate the sensitivity of our results to different assumptions for projected growth rates equal to 4.5% and 7.5%.

We then compare the projected value of these benefits (i.e., premiums plus the HSA, HRA, and FSA amounts) for single and family-of-four coverage from the KFF-HRET survey to the excise tax thresholds specified by the Affordable Care Act: $10,200 for single coverage and $27,500 for family coverage in 2018. The thresholds are to be increased by $1,650 and $3,450, respectively, for plans that cover high-risk professions and in a yet-to-be-specified manner for firms that pay higher premiums due to the age and gender composition of the firm's workforce. We therefore use the KFFHRET survey's 10 industry indicators to make an upward adjustment to the threshold for plans covering high-risk professions. (7) And we use the survey's measure of the proportion of workers under age 26 to make an upward adjustment to the thresholds for plans with an older …