Association for Public Policy Analysis & Management

The ACA’s so-called “Cadillac tax” has been criticized because the flat 40-percent excise tax is not progressive, like the federal income tax schedule. One alternative to the Cadillac tax is an “exclusion cap,” under which individuals enrolled in employer-sponsored plans would be able to exclude premiums from their taxable income only up to a dollar limit (i.e., the cap), with premiums in excess of the cap subject to federal and state income taxes. The goals of the analysis are, first, to define an exclusion cap scenario in the year 2020 that would produce the same amount of federal tax revenues as the Cadillac tax and, second, to compare the impacts of the exclusion cap and the Cadillac tax on families in different income ranges. The analysis uses RAND’s COMPARE microsimulation model, and incorporates the concept of the “tax price” of health insurance described by Gruber and Lettau (2004). The analysis follows these steps: (1) simulate firm offering behavior and premiums in 2020 in a “no-tax” scenario; (2) simulate the Cadillac tax attachment point in 2020; (3) simulate the increase in federal revenues in 2020 from the Cadillac tax; (4) identify an attachment point for an exclusion cap that produces the same increase in federal revenues; and (5) compare impacts on workers’ health benefits and take-home pay under three scenarios: no tax, Cadillac tax, and exclusion cap. The analysis produces three key results. First, under both the Cadillac tax and the exclusion cap scenarios, most of the increase in federal revenues come from the wage passback, rather than revenues from payment of the Cadillac tax or taxes on excess premiums per se. As a result, the impacts of the Cadillac tax and the exclusion cap on affected families are quite similar, both for low- and high-income families. This weakens the progressivity-based argument for replacing the Cadillac tax with an exclusion cap. Second, the simulated impacts of the Cadillac tax depend critically on a relatively obscure provision, which is the prohibition on insurers deducting Cadillac tax payments in the calculation of corporate income tax liabilities. In our main results, we assumed that employers who are exposed to the Cadillac tax had the option of shifting to not-for-profit insurers in order to avoid the issue with corporate tax liabilities. Absent such a shift, the tax price of health insurance above the Cadillac tax attachment point would be prohibitively high. Third, the fact that most of the federal revenues come from wage passbacks complicates the assessment of federal revenue impacts. The Congressional Budget Office (CBO) has recently reduced its estimates of federal revenues from the Cadillac tax, citing lower-than-expected growth in recent years in employer-sponsored insurance premiums. But, that slower growth in premiums may be due in part to anticipatory responses to the Cadillac tax—if premium growth has been slowed by the impending Cadillac tax, then estimates of federal revenues should not necessarily be reduced.