Reduce Tax Subsidies for Employment-Based Health Insurance

Overview of the Issue

The federal tax system provides preferential treatment for health insurance that people buy through an employer. That treatment applies to payments and contributions made both by employers and by employees. Unlike cash compensation, employers' payments for their employees' health insurance premiums are excluded from income and payroll taxes. In most cases, the amount that workers pay for their own share of health insurance premiums is also excluded from income and payroll taxes. Contributions made to certain accounts by employers to pay for employees' health care costs are excluded from income and payroll taxes as well. In all, that favorable tax treatment cost the federal government about $300 billion in forgone revenues in 2018, and that cost will probably rise over time as the price of health care increases. Although a new excise tax will go into effect in 2022, somewhat reducing the tax exclusions' consequences, those exclusions will continue to have significant implications for the federal budget.

Further reducing the tax exclusion for employment-based health insurance—as outlined in this option—would raise federal revenues. However, it also would reduce the number of people with employment-based coverage, boost enrollment in the health insurance marketplaces established under the Affordable Care Act, and increase the number of people without insurance. Total spending on health care would be lower than it would have been otherwise because fewer people would have insurance.

Current Law. The federal tax system subsidizes employment-based health insurance both by excluding employers' premium payments from income and payroll taxes and by allowing employees at firms that offer "cafeteria plans"—which allow workers to choose between a taxable benefit, such as cash wages, and nontaxable fringe benefits—to pay their share of premiums without that share's being subject to income or payroll taxes. The tax system also subsidizes health care costs not covered by insurance by excluding from income and payroll taxes the contributions made to various accounts that employees can use to cover those costs. Examples include employees' contributions to flexible spending arrangements (FSAs), employers' contributions to health reimbursement arrangements (HRAs), and employers' and employees' contributions to health savings accounts (HSAs). On average, people with higher income (and thus higher tax rates) or more expensive health insurance plans receive larger subsidies.

The favorable tax treatment of employment-based health benefits is the federal government's largest single tax expenditure. (Tax expenditures are exclusions, deductions, preferential rates, deferrals, and credits in the tax system that resemble federal spending in that they provide financial assistance for specific activities, entities, or groups of people.) Including effects on both income taxes and payroll taxes, those exclusions are projected to equal 1.5 percent of gross domestic product in 2018.

The excise tax that is scheduled to start in 2022 will effectively reduce the tax subsidy for employment-based health insurance. It will be levied on employment-based health benefits—consisting of employers' and employees' currently taxable and tax-excluded contributions for health insurance premiums and contributions to FSAs, HRAs, or HSAs—whose value exceeds certain thresholds. The excise tax will thus curtail the current open-ended tax exclusions. Even when the excise tax is in effect, however, employment-based health insurance will still receive a significant tax subsidy, and that subsidy will still be larger for people with higher income.

The excise tax will equal 40 percent of the difference between the total value of tax-excluded contributions and the applicable threshold. The Congressional Budget Office and the staff of the Joint Committee on Taxation (JCT) project that some employers will change their offers of coverage to reduce their exposure to the tax. In the agencies' projections, which do not account for such changes, roughly 15 percent of people enrolled in an employment-based plan in 2022 have some taxable and tax-excluded contributions in excess of the thresholds. That share is projected to increase to roughly 25 percent by 2028. (However, CBO and JCT expect people's responses to the tax to reduce that share, as discussed below.)

In 2022, the thresholds are projected to be $11,200 for individual coverage and $30,100 for family coverage. (Those thresholds will be slightly higher for retirees who are 55 to 64 years old and for workers in certain high-risk professions. Further adjustments will be made for age, sex, and other characteristics of an employer's workforce.) After 2022, the thresholds will be indexed to the growth of the chained consumer price index for all urban consumers (chained CPI-U), which measures inflation in a way that accounts for individuals' changing consumption as prices increase. Because health insurance premiums will probably continue to rise faster than inflation, the excise tax will most likely affect a growing number of people over time. As a result, revenues stemming from the tax are projected to rise from $8 billion in 2022 to $39 billion in 2028.

Effects of Current Law. The tax exclusions currently in effect encourage the use of employment-based insurance, making it likelier that healthy people will buy health insurance (which lowers the average cost of insurance and helps to limit a phenomenon known as adverse selection). At the same time, the subsidy for health insurance provided by the exclusion is likely to increase total spending on health care. Another effect is that higher-income workers receive larger subsidies than lower-income workers do.

Encouraging the Use of Employment-Based Insurance. By subsidizing employment-based health insurance, the tax exclusions encourage firms to offer it and workers to enroll in it. Such insurance would be attractive to employers and employees in any case because it pools risks within groups of workers and their families and reduces the administrative costs of marketing insurance policies and collecting premiums. But the tax exclusions give employment-based insurance additional appeal. In 2017, according to the Medical Expenditure Panel Survey, 85 percent of private-sector employees worked for an employer that offered health insurance coverage; 77 percent of those employees were eligible for that coverage (the rest were ineligible for various reasons, such as working only part time); and 73 percent of the eligible workers chose to enroll in the plan offered by their employer.

Reducing Adverse Selection. A major problem that can occur in insurance markets is adverse selection, in which less healthy people are likelier to buy health insurance (or to buy certain types of plans) than healthier people are. Adverse selection occurs because insurance provides more benefit to enrollees with above-average costs—and is therefore more attractive to them—and less benefit to people with below-average costs. As premiums increase to cover the less healthy enrollees, the healthier ones may stop buying insurance, which results in another price increase—a spiral that may continue until the market is very small or nonexistent. Adverse selection also can reduce markets' efficiency by making it harder for insurers to predict costs for a group of potential enrollees.

Employment-based health insurance and the tax subsidies that encourage its use limit adverse selection in several ways. Employers generally select a workforce on the basis of criteria other than health care costs, so most workforces consist of a mix of healthier and less healthy people. Pooling risks across such a workforce reduces the variability of average health care spending for the group. Also, once employers offer health insurance, they tend to pay a large share of premiums in order to encourage employees to enroll—making the employees' share small in relation to their expected health care costs, encouraging them to buy insurance, and reducing adverse selection. The tax exclusions also limit adverse selection by reducing the after-subsidy price of insurance, encouraging even the healthy to enroll.

Increasing Total Health Care Spending. For workers and their families who enroll in such coverage, the tax subsidies for employment-based health insurance encourage more spending on health care than would be the case without subsidies. That occurs because the subsidies encourage workers to favor health care over other goods and services that they could purchase and also because the tax exclusions encourage employers to compensate their workers with a combination of health insurance coverage and cash wages rather than entirely with cash wages (which the employees would be unlikely to spend on health care to the same extent). Furthermore, the tax exclusions are currently open-ended (and will be until the excise tax takes effect in 2022). That is, their value increases with an insurance plan's premium, encouraging people to enroll in plans that cover a greater number of services, cover more expensive services, or require enrollees to pay a smaller share of costs. As a result, people use more health care—and health care spending is higher—than would otherwise be the case.

That effect may have been lessened somewhat in recent years because employment-based insurance plans that require workers to pay a higher share of health costs have become more common. For example, 29 percent of people with employment-based coverage were enrolled in a high-deductible health plan in 2017, up from 8 percent in 2008.

Subsidizing Workers With Different Income Differently. Another concern about the tax exclusions is that they subsidize workers with different amounts of income differently. The value of the exclusions is generally larger for workers with higher income, partly because those workers face higher income tax rates (although they may face lower rates of payroll taxation) and partly because they are more likely to work for an employer that offers coverage. Because larger subsidies go to higher-income workers, who are more likely to buy insurance even without the tax exclusions, and smaller subsidies go to lower-income workers, who are less likely to buy coverage, the exclusions are an inefficient means of increasing the number of people who have health insurance, and they are regressive in the sense that they provide larger benefits to people with higher income.

The forthcoming excise tax will somewhat reduce the regressive nature of the tax exclusions. The excise tax will be levied on insurers and on employers who offer their own insurance plans, but economic theory and empirical evidence suggest that the cost will be passed on to workers. CBO and JCT expect that, in many cases, the tax burden will shift when employers and workers decide to avoid paying the tax by switching to health plans with premiums below the thresholds. In those cases, the money that would otherwise have been used to pay for the more expensive premiums would generally increase either workers' wages or employers' profits, both of which are taxable. Because workers with higher income will pay higher marginal tax rates on those increased wages, the regressive nature of the tax exclusions will be reduced. When employers and workers do not shift to lower-cost health plans to avoid the excise tax, the costs of that tax will be spread equally among affected workers, JCT and CBO expect. However, workers with higher income are more likely to be enrolled in high-cost plans and thus more likely to have their subsidies reduced by the excise tax. Nevertheless, most workers will have health benefits whose value is below the thresholds and therefore will be largely unaffected by the excise tax. Consequently, the existing tax subsidies and the new excise tax will continue to subsidize employment-based health insurance and to provide larger subsidies to higher-income people.

Key Design Choices That Would Affect Savings

Lawmakers who wanted to design laws to reduce the tax subsidies for employment-based health insurance could take various approaches. In general, reducing the tax subsidies for employment-based health insurance would tend to lower the number of people with such insurance. It also would increase out-of-pocket payments by people enrolled in employment-based insurance, which would decrease spending on health care and increase the financial burden on people with substantial health problems. The precise effect, however, would depend on the specific features of any policy change.

Lawmakers could cancel the excise tax that is scheduled to take effect under current law and instead subject contributions for health insurance premiums, along with contributions to various health-related accounts, to income or payroll taxation. If lawmakers did that, they would have to decide whether to tax all of the contributions or only some of them. For example, the exclusions could be retained, but with an upper limit that applied to all taxpayers, or the exclusions could be phased down for higher-income people. Such limits also could be allowed to vary according to other characteristics of employees—such as age, sex, or occupation—that are associated with average health care costs. (The forthcoming excise tax includes several adjustments of that sort. For instance, the threshold above which health care costs are taxed is higher for some groups of people whose average costs are high because they work in dangerous occupations.)

Lawmakers also would need to decide whether to subject the contributions to income taxation, payroll taxation, or both. On average, enrollees in employment-based plans face slightly higher federal income tax rates than payroll tax rates. Specifically, CBO and JCT estimate that those workers' average marginal income tax rate—that is, the rate that applies to the last dollar of their earnings—would be about 18 percent in 2022, whereas their average marginal payroll tax rate (including both the employer's and the employee's shares of payroll taxes) would be about 14 percent. Therefore, subjecting contributions to income taxation would raise slightly more revenues than subjecting them to payroll taxation, all else being equal, and doing both would raise the most revenues.

Even if the average income tax rate and the average payroll tax rate for enrollees in employment-based plans were the same, subjecting contributions to income taxation and payroll taxation would have very different effects on the tax liability of people in different income groups. Higher-income people are likely to have higher marginal income tax rates but lower marginal payroll tax rates than lower-income people. Among people with employment-based insurance, therefore, subjecting contributions to income taxation would raise the tax liability of higher-income people more than that of lower-income people. The opposite would be true if contributions were subject to payroll taxation.

Subjecting contributions to taxation would increase the after-tax price of people's employment-based health insurance and therefore reduce insurance coverage. However, CBO and JCT estimate that subjecting contributions to income taxation would yield smaller reductions in the number of people with health insurance than subjecting contributions to payroll taxation would (provided that the same upper limit applied in each case). As discussed above, for lower-income people, the average marginal tax rate is smaller for income taxes than for payroll taxes. Therefore, lower-income people would face smaller increases in the after-tax price of their employment-based health insurance if the contributions were subject to income taxation than if contributions were subject to payroll taxation. Consequently, low-income people would be less likely to forgo insurance if their contributions were subject to income taxation rather than payroll taxation. At the same time, because higher-income people, on average, face a higher marginal income tax rate than marginal payroll tax rate, more higher-income people would forgo insurance if their contributions were subject to income taxation than if they were subject to payroll taxation. However, that reduction in insurance coverage for higher-income people would be smaller than the reduction for lower-income people. (Higher-income people are less responsive to price changes in health insurance because they tend to have more assets to protect and higher demand for health care services.)

Specific Alternatives and Estimates

CBO and JCT analyzed three alternatives for reducing the tax subsidies for employment-based health insurance. Each alternative would take effect in 2022 and would replace the excise tax on high-cost plans with a limit on the tax exclusions. The first and second alternatives would limit the exclusions from income and payroll taxation; the third would limit the exclusion from income taxation but continue the unlimited exclusion from payroll taxation. Those policy changes would increase the tax liability and affect the behavior of people with high premiums for employment-based health plans, but the specific increases in taxes and changes in behavior would be different under each approach.

Each alternative was estimated using CBO and JCT's microsimulation models. Those models use a combination of detailed survey and administrative data to construct a nationally representative sample of employers and individuals in order to estimate the national distribution of health insurance coverage and taxes under current law and different policy scenarios. The advantage of using those models is that they simulate how employers and individuals make decisions about what type of health insurance coverage to offer and purchase on the basis of their income, firm or family size, expected health care spending, and the relative price and generosity of each health insurance option available to them. The models are also interactive in that they allow the choices firms and individuals make to affect the choices of other firms and individuals within the model. That allows the models to estimate the initial effects of a policy change and the subsequent behavioral responses to those changes. For example, if the costs of employment-based health insurance increased because of a change in the tax treatment of employers' premium contributions, microsimulation models are particularly useful for capturing the alternative insurance options and subsidies available to workers and for estimating the changes in coverage that would result from that type of price increase. As a result of those features, microsimulation models can better approximate the full range of behavioral responses that different types of employers and individuals would make in response to policy changes, as compared with other types of static economic models.

Replace the Excise Tax With a Limit on the Income and Payroll Tax Exclusions Set at the 50th Percentile of Premiums. The first alternative would eliminate the excise tax and instead impose a limit on the extent to which employers' and employees' contributions for health insurance premiums—and to FSAs, HRAs, and HSAs—could be excluded from income and payroll taxation. Specifically, starting in 2022, contributions that exceeded $7,800 a year for individual coverage and $18,500 for family coverage would be included in employees' taxable income—that is, they would be subject to both income and payroll taxes. Those limits, which are equal to the estimated 50th percentile of health insurance premiums paid by or through employers in 2020, would be indexed for inflation by means of the chained CPI-U, a measure of inflation that attempts to account for the effects of substitution on changes in the cost of living. The same limits would apply to the deduction for health insurance available to self-employed people. Because the limits would be lower than the thresholds scheduled to take effect for the excise tax—for example, $11,200 for individual coverage in 2022—federal tax subsidies would be lower as well.

This alternative would decrease cumulative federal deficits by $638 billion by 2028, CBO and JCT project. On a net basis, $51 billion in additional revenues would be collected in 2022, and that amount would grow to $132 billion by 2028. The increasing amount of revenues that would be collected under this alternative would be the result of indexing the exclusion thresholds to the chained CPI-U, which would increase the threshold amounts at a lower rate than the projected growth of health insurance premiums. Over time, that would increase the share of insurance contributions subject to taxation. Those revenues would be slightly offset by $32 billion of additional outlays—the majority of which would be attributable to more people enrolling in subsidized nongroup insurance. By reducing the appeal of employment-based health insurance, it also would cause about 3 million fewer people to have such coverage in 2028 than would have it under current law. Of those people, about 2 million would buy coverage directly through the nongroup market (that is, either in the health insurance marketplaces or from insurers outside of the marketplaces); fewer than 500,000 would enroll in Medicaid; and about 1 million would be uninsured. (Those numbers do not add up to the total because of rounding.)

The reduction in the deficit would stem from several changes in revenues and outlays that partially offset one another. Income and payroll tax revenues would rise by $841 billion through 2028 because the number of people with employment-based coverage would decline and because many of those who retained such coverage would receive a smaller benefit from the tax exclusion. (For example, in 2028, the capped tax exclusions would reduce the combined federal income and payroll tax liability of policyholders with employment-based coverage by an average of $4,450; that reduction would be $6,242 under current law.) Because large employers are required by law to provide health insurance to their employees or pay certain penalties, additional penalty payments by large employers that no longer offered health insurance coverage to their employees also would increase revenues, although only by a small amount. However, additional tax credits for coverage purchased through the marketplaces would reduce revenues, as would the repeal of the excise tax. In all, revenues through 2028 would be $670 billion higher than under current law. The alternative also would boost federal outlays by $32 billion through 2028, primarily because of increased subsidies for insurance purchased through the marketplaces, increased spending on Medicaid, and Medicare "disproportionate share hospital" payments to inpatient facilities that serve a higher percentage of low-income patients.

Replace the Excise Tax With a Limit on the Income and Payroll Tax Exclusions Set at the 75th Percentile of Premiums. Like the first alternative, the second alternative would eliminate the excise tax and impose limits on the extent to which contributions could be excluded from income and payroll taxation. Under this alternative, however, the limits would be higher: $9,900 a year for individual coverage and $25,000 for family coverage. Those limits are equal to the estimated 75th percentile of health insurance premiums paid by or through employers in 2020 and inflated by the chained CPI-U.

The second alternative would decrease cumulative federal deficits by $256 billion by 2028, CBO and JCT estimate. Specifically, it would increase revenues by $270 billion and outlays by $15 billion. Under this alternative, the government would collect, on a net basis, $19 billion in additional revenues in 2022 and an additional $57 billion in 2028 compared with current law. Fewer revenues would be collected under this alternative than under the first alternative because the tax exclusion threshold would be higher. The amount of additional annual revenues collected would grow substantially by 2028 because the thresholds would grow at a lower rate than the projected growth of health insurance premiums, and those revenues would be offset by $15 billion in additional outlays. Also, like the first alternative, this one would reduce the appeal of employment-based health insurance, causing slightly more than 1 million fewer people to have such insurance in 2028 than would have it under current law. In that year, of those people affected by this alternative, slightly less than 1 million more people would buy coverage through the nongroup market, fewer than 500,000 people would enroll in Medicaid or the Children's Health Insurance Program, and fewer than 500,000 would be uninsured.

Replace the Excise Tax With a Limit on Only the Income Tax Exclusion Set at the 50th Percentile of Premiums. The third alternative would eliminate the excise tax and impose a limit on the extent to which contributions could be excluded from income taxation; exclusions for payroll taxation would remain unlimited. Specifically, starting in 2022, contributions that employers or workers made for health insurance—and for medical expenses through FSAs, HRAs, and HSAs—that exceeded $7,800 a year for individual coverage and $18,500 for family coverage would be included in employees' taxable income and subject to income taxes. Those are the same limits as the ones described in the first alternative, and once again, they would be indexed for inflation by means of the chained CPI-U. As noted above, limiting the tax exclusion for income taxes only would raise more revenues, and reduce insurance coverage less, than would limiting the exclusion for payroll taxes only (as long as the same limit applied in each case).

The third alternative would decrease cumulative federal deficits by $438 billion by 2028, CBO and JCT estimate: Revenues would be $452 billion higher, and outlays would be $14 billion higher. The amount of revenues collected would be lower than under the first alternative because health insurance contributions would still be exempt from payroll taxation. Outlays would offset revenues to a lesser degree than under the first and second alternatives because fewer people who gave up employment-based insurance would enroll in subsidized health insurance. This alternative would cause about 1.5 million fewer people to have employment-based insurance in 2028 than would have it under current law. Of those people, about 1 million would buy coverage through the nongroup market, fewer than 500,000 would enroll in Medicaid, and about 500,000 more would be uninsured.

Sources of Uncertainty

These estimates rely on the complex interaction of many variables and are therefore inherently uncertain. The stability of nongroup insurance markets under current law is one source of uncertainty. If the nongroup market was less stable than projected in CBO's baseline, nongroup coverage would be more expensive and less attractive as an alternative to employment-based coverage. Consequently, CBO and JCT would expect more employers to offer such coverage (and more employees to take up such offers). This would increase, all else being equal, the amount of revenues collected under each alternative because more people would be enrolled in employment-based insurance. If nongroup insurance markets remained stable, or if they became more competitive over time, fewer people would enroll in employment-based coverage, which would reduce the amount of revenues that would be collected under these alternatives.

These estimates are also sensitive to changes in the price of employment-based health insurance. For example, if premiums for such coverage grew faster than in CBO and JCT's baseline projections, all else being equal, fewer people would enroll in such coverage, but an increased number of plans and employers would have premiums above the excise tax's threshold under current law. Under the alternatives discussed here, higher growth in premiums would increase the revenues collected by the federal government because a higher share of workers would have premiums that exceeded the alternatives' thresholds for tax exclusions. However, because a smaller number of workers would have employment-based coverage both under current law and under the option if premiums for employment-based coverage grew at faster rates than in CBO and JCT's baseline projections, the net effect of the option on federal revenues could be higher or lower than the estimates presented here.

Another source of uncertainty is employers' willingness to continue offering health insurance coverage. In recent years, offers of employment-based health insurance have generally remained stable, on average. Employers' decisions to offer coverage are affected by a variety of factors, including the availability of alternative sources of coverage, changes in the after-tax price of such coverage, and competition in the labor market. Firms may become more willing to drop offers of employment-based coverage as each of those factors changes over time. If that were the case, the change in coverage resulting from this option and the associated reduction in the deficit would both be larger.

Changes in the larger economy, such as a recession that resulted in increased unemployment, could also affect these estimates. In such a scenario, fewer people would be enrolled in employment-based health coverage, which would reduce the amount of revenues that would be collected under each alternative. By contrast, faster than expected economic growth could increase the number of people with employment-based coverage, thereby increasing the amount of taxes that would be collected under these alternatives.

Other Considerations

Reducing tax subsidies for employment-based health insurance would affect many aspects of health care in the United States, including the growth of health care costs, the health of the population, the decisions that employers and workers make about health insurance coverage, and the number of people without health insurance.

Effects on Health Care Costs. Replacing the excise tax with a limit on the tax exclusions that is lower than the excise tax thresholds would make health care spending lower than it would be under current law. The current tax subsidies for employment-based insurance give health insurance plans an incentive to cover more services, to cover more expensive services, and to require enrollees to pay a smaller share of the costs than would be the case otherwise. The excise tax will effectively scale back those tax subsidies. The alternatives examined here would increase taxes for a larger share of employment-based plans than the excise tax would—giving employers and their workers less incentive to buy expensive health insurance, reducing upward pressure on the price and use of health care, and encouraging more cost-effective use of care.

Effects on People's Health. By reducing the incentive to buy expensive coverage and increasing the incentive to buy insurance plans that require people to pay more out of pocket, all three of the alternatives analyzed here would reduce the amount of care received and worsen some people's health. That conclusion is supported by an experiment conducted by the RAND Corporation from 1974 to 1982 in which nonelderly participants were randomly assigned to health insurance plans (Newhouse and the Insurance Experiment Group 1993). The experiment showed that plans requiring more out-of-pocket payments reduced the use of both effective and less-effective care, as defined by a team of physicians. Differences in out-of-pocket requirements had no effect on most participants' health, but among the poorest and sickest participants, those who faced no requirements of that kind were healthier by some measures than those who did.

Effects on Employers' and Workers' Decisions About Health Care Coverage. By increasing the tax liability of people enrolled in high-cost employment-based plans more than the excise tax would, the alternatives considered here would probably increase the financial burden on some people with substantial health problems. In particular, some employers and workers would avoid the increased tax liability by shifting to plans with lower premiums and more limited benefits, which would increase costs the most for people who used the most services.

In general, workers with higher income face higher income tax rates and are more likely to enroll in plans with high premiums. Therefore, limiting the exclusion from income taxation, as the third alternative would do, would reduce that benefit more for people with higher income. The first and second alternatives, which would limit the exclusion not only for income taxation but also for payroll taxation, would still increase tax liabilities more for higher-income people, on average, because they tend to enroll in plans with higher premiums.

Under all three alternatives, employees of firms that had a less healthy workforce or that operated in an area with above-average health care costs would be more likely to see their tax liability increase. In higher-cost areas, those increases in people's tax liability might exert pressure on health care providers and insurers to reduce prices or decrease unnecessary care.

Although these alternatives would reduce total spending on health care, they would increase after-tax premiums for some people enrolled in employment-based insurance, particularly those whose premiums were above the limits imposed by each alternative and who therefore would be paying taxes on that portion of their premiums for the first time. In addition, because all three alternatives would impose a limit on the exclusion that was lower than the excise tax thresholds that are scheduled to go into effect under current law, employers would have a heightened incentive to keep premiums low, which could cause them to refrain from hiring older workers (who tend to spend more on health care and to raise average premiums) or to reduce the compensation of older workers. That effect would be particularly likely among employers with fewer employees over whom to spread risks.

Effects on the Number of Uninsured People. The tax increases in these alternatives would lead fewer employers to offer health insurance. Although most people whose employers stopped offering health insurance would instead buy coverage in the nongroup market, either in the health insurance marketplaces or elsewhere, CBO and JCT anticipate that some workers would forgo coverage.