Distributional Effects of Applying Social Security Taxes to Employer-Sponsored Health Insurance Premiums
Policy Brief No. 2016-01 (released August 2016)
This policy brief analyzes how applying the Social Security tax to employer-sponsored health insurance premiums could affect Social Security beneficiaries. Specifically, the brief examines an option presented by the Social Security Advisory Board in which both employee and employer premiums would count as wages for Social Security tax calculations, and later for benefit calculations. Using the Modeling Income in the Near Term model, the results show that for most Social Security beneficiaries aged 60 or older from 2017 to 2080, benefits would gradually increase and the poverty rate would decrease faster than the rate under current law. Counting employer-sponsored health insurance premiums as wages for Social Security purposes would increase Social Security taxes for most individuals and those taxes would increase more than Social Security benefits for individuals at all earning levels.
Kathleen Romig is a senior policy analyst with the Center on Budget and Policy Priorities. Dave Shoffner is a social science research analyst with the Office of Retirement Policy, Office of Retirement and Disability Policy, Social Security Administration. Kevin Whitman is the Idaho state coordinator for the National Assessment of Educational Progress. Questions about the analysis should be directed to Dave Shoffner at (202) 358-6210.
Acknowledgments: The authors thank Joni Lavery, Joyce Manchester, Karen Smith, Gayle Reznik, Marina Kutyavina, Mark Sarney, and Natalie Lu for their helpful comments and suggestions.
The findings and conclusions presented in this brief are those of the authors and do not necessarily represent the views of the Center on Budget and Policy Priorities, the Social Security Administration, or the National Assessment of Educational Progress.
|ACA||Affordable Care Act|
|AWI||average wage index|
|MINT||Modeling Income in the Near Term|
|SSA||Social Security Administration|
Employer-sponsored health insurance premiums are exempt from Social Security payroll taxes, hereafter called “Social Security taxes.” 1 In contrast, health insurance policies purchased outside the workplace—including those purchased through health care exchanges—are subject to Social Security taxes. Some policymakers have proposed applying the Social Security tax to all health insurance premiums, including employer-sponsored health insurance premiums.2 Social Security's Trustees project a long-term funding shortfall (Board of Trustees 2016).3 Including all premiums as income subject to Social Security taxes would close over a third of the shortfall.4
This policy brief analyzes how applying the Social Security tax to employer-sponsored health insurance premiums could affect Social Security beneficiaries. Specifically, we examine an option presented by the Social Security Advisory Board in which, starting in 2017,5 both employee and employer premiums would count as wages for Social Security tax calculations, and later for benefit calculations (Social Security Advisory Board 2010).
Our analysis focuses on how this option would affect Social Security taxes, Social Security benefits, and poverty rates over time. We break down our results by earnings level so that readers can assess the progressivity of the option. This brief focuses exclusively on Social Security. We do not examine the effects of applying the Medicare Hospital Insurance (HI) tax or the federal income tax to employer-sponsored health insurance premiums, as some policymakers have proposed.
Using Social Security Administration's (SSA's) Modeling Income in the Near Term, version 7 (MINT7) model, 6 our results show Social Security beneficiaries aged 60 or older from 2017 to 2080. We compare benefits under the policy option with the benefits scheduled to be paid under current law. Our analysis assumes that employers will lower wages by the amount of additional Social Security tax payable on employer-sponsored health insurance premiums, so that their total compensation costs remain the same.7 Otherwise, our analysis does not simulate behavior changes in response to taxing employer-sponsored health insurance premiums, though it is possible that employers could respond by changing their health insurance coverage or costs, or that employees could change their health insurance selections.
- Counting employer-sponsored health insurance premiums as wages for Social Security purposes would increase Social Security taxes for most individuals. The tax increases would generally be proportionately smallest for high earners, and largest for low and middle earners.8
- Social Security benefits would gradually increase for most new beneficiaries. Benefit increases would generally be larger for low and middle earners than for high earners, and larger for later cohorts than for earlier cohorts.
- On the balance, beneficiaries' lifetime Social Security taxes would increase more than their lifetime Social Security benefits at all earning levels.
- The poverty rate among Social Security beneficiaries aged 60 or older would decrease faster than the poverty rate under current law.
Policy Option Background
Workers must pay Social Security taxes on their covered earnings. Other forms of compensation, including employer-sponsored health insurance, are exempt from Social Security taxes. In contrast, health insurance policies purchased outside of the workplace—including those purchased through health care exchanges—are generally paid for with income that has been subject to Social Security taxes.9
Employer-sponsored health care covers over half of the non-elderly U.S. population, about 147 million people in total (Kaiser Family Foundation 2015). Excluding employer-sponsored health insurance premiums from Social Security taxes cost about $100 billion in payroll tax expenditures in 2015, and will cost about $1.25 trillion over the 10-year period—2016 through 2025.10 In addition, employer-sponsored health insurance premiums have become an increasingly larger share of total compensation. Burtless and Milusheva (2013) found that employers' contributions for employees' health insurance nearly doubled in 30 years, from 3.7 percent of total compensation in 1980 to more than 7.0 percent in 2010.
The primary argument for exempting employer-sponsored health insurance is that it encourages employers to offer this benefit. Employer-paid premiums are similar in value to wages, but allow employers to avoid the additional Social Security tax they would incur from paying higher wages instead of paying premiums. However, the exemption substantially reduces taxes paid and a large proportion of the reductions in taxes go toward higher-income individuals.11
Our projections use the MINT7 model (Social Security Administration, n.d. a), which includes self-reported health insurance coverage from the Survey of Income and Program Participation. Using 2012 data from the Kaiser Foundation, MINT7 also projects at each job change whether the employer offers health insurance as well as premium costs for both family and single plans. MINT7 assumes that:
- Families select the health insurance plan(s) that cover all members of the household with the lowest total out-of-pocket premium cost.
- Children are covered until age 26 (Patient Protection and Affordable Care Act (ACA) rule).
- In the short term (up until 2022), premiums grow faster than wages. In the long term (2022 and later), premiums grow at the same rate as wages.
We also made a number of assumptions specific to this policy brief, including that the taxation of all employer-sponsored health insurance premiums begins in 2017. For Social Security tax and benefit purposes, we only counted employer-sponsored health insurance premiums as part of covered earnings if more than half of a person's earnings were covered by Social Security that year. We assumed that employers would pass on their share of the increased Social Security tax by reducing employee wages, thus, keeping employers' total compensation costs constant. However, to prevent workers with relatively low earnings and high insurance premiums from having their wages reduced too steeply, we capped the amount that wages could be reduced in a given year at 10 percent.12
Average Wage Index Assumptions
SSA uses the average wage index (AWI) to calculate annual increases in initial Social Security benefits, the taxable maximum, the retirement earnings test thresholds, and many other parts of Social Security. 13
In this brief, we assume no changes to the AWI. If Congress passes legislation applying the Social Security tax to employer-sponsored health insurance premiums, it would also have to decide how the change affects AWI calculation. If Congress increased the AWI to reflect all of the newly covered health insurance premiums, benefits would increase immediately for all new beneficiaries, whether or not they had contributed more to Social Security. If Congress did not increase the AWI, the proportion of all covered wages that would be taxable would likely fall, as covered wages (which would include employer-sponsored health insurance premiums) would grow faster than maximum taxable wages (which would not, because they are indexed to the AWI). Congress could also change the AWI more subtly, as it did when it subjected contributions to defined contribution pensions to the Social Security tax.14
Effect of ACA on Projections
MINT7 assumes the ACA would not have a significant effect on employer-sponsored health insurance coverage and premiums. While most people newly covered under ACA would receive Medicaid or buy insurance from health insurance exchanges—which would not be affected by the option—others will shift away from employer-sponsored coverage as a result of the ACA. The Congressional Budget Office and Joint Committee on Taxation assume that each year from 2019 to 2022, 3 million to 5 million fewer people will have employer-sponsored health insurance coverage, compared with current law (Congressional Budget Office 2012).15 As far as premiums, the outlook is uncertain. The ACA will impose an excise tax on high-cost employer-sponsored health insurance plans starting in 2018. Some analysts expect this tax to cause some employers to shift compensation from health insurance premiums to wages (Social Security Advisory Board 2010). If this occurs, the vast majority of those wages would be subject to the Social Security tax.
Lifetime Social Security Taxes Would Increase for Most Individuals
Including all employer-sponsored health insurance premiums in taxable wages would increase Social Security taxes for most workers. The increases would be proportionately smallest for workers with the highest earnings because employer-sponsored health insurance premiums make up a smaller share of total compensation for high earners (Burtless and Milusheva 2013). In addition, workers do not pay Social Security tax on wages above the taxable maximum wage ($118,500 in 2016)—including any premiums that would count as wages under this option. Thus, workers who already earn above the annual taxable maximum would be unaffected by the change and workers whose employer-sponsored health insurance premiums push their earnings over the taxable maximum would be shielded from some of the additional tax. Typically, about 6 percent of workers earn more than the taxable maximum wage each year (Whitman and Shoffner 2011). Workers whose jobs are not covered by Social Security (such as some state and local government employees) do not pay Social Security taxes and would not be affected by this option.
To show this option's effect on Social Security taxes, we calculated the equivalent Social Security tax increase, which is the difference between the present value of lifetime Social Security taxes paid under the option and current law, divided by the present value of lifetime taxable covered earnings under current law. This value represents the Social Security tax increase under the policy option. For example, a 1.0 percentage point increase is the equivalent of raising the combined employer and employee Social Security tax rate from the current 12.4 percent to 13.4 percent over the course of a worker's career.16
As workers pay Social Security taxes on their employer-sponsored health insurance premiums for more years, we project that their lifetime Social Security taxes will increase. Chart 1 shows the median change in equivalent Social Security taxes under this option, which is assumed to begin in 2017. In 2030, we project no change in the equivalent tax rate at the median; in 2050, 0.4 percentage points; in 2070, 0.7 percentage points.
Overall, we project that the median equivalent Social Security tax increases will be relatively lower for high earners than for low and middle earners, as shown in Chart 2.17 Earners in the second lowest quintile would receive the highest equivalent Social Security tax increase because they are both likely to receive employer-sponsored health insurance and to pay a relatively larger share of their wages for employer-sponsored health insurance premiums.18 Workers in the lowest quintile are less likely to receive employer-sponsored health insurance, and workers in higher quintiles are likely to pay a smaller share of their wages toward premiums.
Eventually, 75 Percent of Social Security Beneficiaries Would Receive Higher Benefits
The option's effect on benefits would phase in gradually, as workers spend more years contributing to Social Security under the new rules. Chart 3 shows the proportion of beneficiaries with higher benefits under this option, compared with current law.19 The proportion grows over the period shown.
Four types of beneficiaries would not be affected by the option, specifically those who in 2017 and later: do not have earnings, do not have pre-tax employer-sponsored health insurance, earn above the Social Security taxable maximum, or have earnings not covered by Social Security.
As the number of beneficiaries affected by the option increases, so too would benefits increase. Chart 4 shows the median percent change in benefits under this option.
Benefit increases vary by earnings level and would increase over time with larger benefits for later cohorts. Chart 5 shows the median percent change in benefits under the option in three snapshot years broken down by shared lifetime earnings quintiles.
Low and middle earners would receive higher Social Security benefits under the policy option for several reasons. While health insurance premiums vary across the earnings distribution, they do not vary as much as earnings, thus, employer-sponsored health insurance premiums make up a larger share of compensation for low earners than high earners (Burtless and Milusheva 2013). Thus, the relative increase in covered, taxable earnings under the option is greater for low and middle earners, who would pay relatively higher taxes as a result. Further, each additional dollar in covered earnings is replaced more generously by Social Security for lower earners than for high earners, and high earners are more likely to earn more than the annual taxable maximum.20
Beneficiaries' Lifetime Social Security Taxes Would Increase More than Lifetime Social Security Benefits
On balance, the option would increase lifetime Social Security taxes more than lifetime Social Security benefits, at all earning levels. The shared benefit/tax ratio compares projections of the present value of lifetime benefits to lifetime taxes, with benefits and taxes split evenly for married couples during years in which they are married. We project that the ratios for beneficiaries aged 60 or older under the option will be equal to or lower than the ratios under current law for all quintiles. The lower ratios show that the increase in lifetime benefits (the numerator) is smaller than the increase in lifetime Social Security taxes (the denominator). Chart 6 illustrates this by showing the median percent change in shared benefit/tax ratios under this option. The ratios shrink more for lower earnings quintiles.
Poverty Among Elderly Beneficiaries Would Decrease Faster than Under Scheduled Benefits
This option would lower the poverty rate faster among Social Security beneficiaries aged 60 or older compared with scheduled benefits (Chart 7). As the number of people projected to receive higher benefits under the option rises, the number of beneficiaries in poverty falls. By 2070, we project that the option would reduce the poverty rate for elderly beneficiaries by about 15 percent, from 2.1 percent to 1.8 percent. It is important to note that the poverty rate under current law and the option are both projected to decline over time.21 This is not because of any change in Social Security policy, but because the poverty threshold grows with inflation while we assume that household income will grow at the generally higher rate of wage growth.22
As part of comprehensive Social Security reform, some policymakers have proposed applying the Social Security tax to all health insurance premiums, including employer-sponsored health insurance premiums. Taxing employer-sponsored health insurance premiums would treat those with workplace coverage the same as those without. An increasing number of Americans are purchasing insurance through health care exchanges, and paying their premiums with income that has been taxed for Social Security. Imposing the same tax on those who receive health insurance through their employers would broaden Social Security's tax base and close over a third of the projected long-term financing gap.
We found that this option would increase Social Security taxes for most beneficiaries, and that these increases would be proportionally largest for low and middle earners and smallest for high earners. At the same time, it would increase Social Security benefits for most new beneficiaries, particularly for low and middle earners in later cohorts. Over a lifetime, beneficiaries at all income levels would pay more in additional taxes than they would receive in additional benefits.
It is important to understand the limitations of our projections. Predicting health insurance coverage and costs is uncertain, and the ACA has added uncertainty. We did not assume employers or employees would change their behavior in response to this option, though some certainly would. In addition, our study focuses exclusively on Social Security taxes, though some policymakers have also proposed applying federal income and Medicare HI taxes to employer-sponsored health insurance premiums.
1 “Social Security taxes” only includes the Social Security payroll tax that workers pay on their wages, and does not include the taxation of Social Security benefits or any other types of taxation.
2 For examples, see: Restoring America's Future: Reviving the Economy, Cutting Spending and Debt, and Creating a Simple, Pro-Growth Tax System by Washington, DC: Bipartisan Policy Center, http://bipartisanpolicy.org/library/restoring-americas-future/; and “Roadmap for America's Future Act of 2010” H.R. 4529, 111th Cong. (2010), https://www.congress.gov/bill/111th-congress/house-bill/4529. (Representative Paul Ryan [WI] introduced the bill, but it was not enacted.) These plans would subject employer-sponsored health insurance premiums to Social Security and Medicare payroll taxes as well as federal income taxes.
3 Social Security's Trustees project a shortfall of 2.66 percent of taxable payroll over the next 75 years (Board of Trustees 2016).
4 Based on the 2009 Trustees Report, SSA's Office of the Chief Actuary projected that the original version of this option would reduce the 75-year deficit by 57 percent. The actuaries noted that the option's solvency impact would depend, in part, on growth in health care premiums. A more recent estimate of a similar provision proposed by the Bipartisan Policy Center finds that the 75-year deficit would be reduced by 37 percent. https://www.socialsecurity.gov/oact/solvency/provisions/charts/chart_run278.html.
5 For this policy brief, we updated the option to start in 2017.
6 The MINT7 data used for this brief are based on the 2012 Trustees Report (Board of Trustees 2012). For more information on MINT7, see Social Security Administration (n.d. a and n.d. b).
7 Labor economists generally assume that employees bear most or all of the costs of fringe benefits in the long run (Gruber 2000; Jensen and Morrisey 2001).
8 In this brief, “earnings” refers to wages only and does not include other types of income.
9 In the 2015 enrollment period, 11.7 million Americans enrolled in health insurance plans through the Health Insurance Marketplaces established by ACA (Department of Health and Human Services 2015).
10 The Office of Management and Budget estimates that excluding employer-sponsored health insurance premiums from both Social Security and HI taxes cost $127.5 billion in 2015, and will cost $1.6 trillion over the 10-year period—2016 through 2025 (Office of Management and Budget, n.d., Supplementary Table 14-1). The Social Security Trustees estimate that the HI taxable payroll is about 25 percent larger than the Social Security taxable payroll (Board of Trustees 2015, 200). The rates are different for each tax: 12.4 percent for Social Security and 2.9 percent for HI, with an additional 0.9 percent for the highest earners. After factoring in the rate and base for each payroll tax, we estimate that Social Security taxes account for about three-quarters of the total payroll tax expenditure.
11 For more information, see Gruber, Jonathan. 2010. “The Tax Exclusion for Employer-Sponsored Health Insurance.” NBER Working Paper No. 15766. Cambridge, MA: National Bureau of Economic Research. http://www.nber.org/papers/w15766.pdf.
12 About 5 percent to 6 percent of earners with employer-sponsored health insurance would have the reduction to their wages capped at 10 percent in any given year, starting in 2017.
13 For more information on AWI, see https://www.socialsecurity.gov/oact/cola/AWI.html.
14 For more detailed information, see Clingman, Michael D. and Jeffrey L. Kunkel. 1992. “Average Wages for 1985-90 for Indexing Under the Social Security Act,” Actuarial Note No. 133. Washington, DC: Social Security Administration, Office of the Chief Actuary. https://www.socialsecurity.gov/oact/NOTES/note133.html.
15 Subsequent updates in 2013–2015 have gone up, then back down, but remain similar to the 2012 estimates. http://www.cbo.gov/sites/default/files/cbofiles/attachments/03-15-ACA_and_Insurance_2.pdf.
16 As a comparison, the 2012 Trustees Report, which corresponds to the MINT estimates included in this brief, estimates that keeping the program solvent over the 75-year projection period would require an immediate across-the-board tax increase of 2.61 percentage points (Board of Trustees 2012). Please note that the 2016 Trustees Report estimates that a tax increase of 2.66 percentage points would be required (Board of Trustees 2016). The Congressional Budget Office projects a larger long-term imbalance (Congressional Budget Office 2015).
17 Includes all earners, including those who earn above the taxable maximum and thus, would not pay higher taxes if employer-sponsored health insurance premiums were taxed. Earners are grouped by shared lifetime earnings quintile, which represents the sum of wage-indexed shared lifetime earnings. In years that an earner is married, the couple's earnings are combined and half of the earnings are attributed to each spouse.
18 While workers in the second lowest quintile would have the highest median proportional increase in taxes paid, they would not have the highest median dollar increase in taxes paid since other quintiles pay higher lifetime taxes under current law. For example, in 2070, those in the second highest quintile would have the biggest median increase in the present value of lifetime taxes paid, about $110,000, compared to a median increase of about $80,000 for workers in the second lowest quintile.
19 We characterized a beneficiary as having higher benefits if his or her benefits under the option are projected to be at least 1 percent higher than current law benefits. We project that a small number of beneficiaries would have lower benefits under the option, including some who would lose their entire benefit. In each year, about 2 percent to 3 percent of beneficiaries would have lower benefits (not shown in Chart 3), and about 0.1 percent to 0.3 percent of beneficiaries would lose their entire benefit. Beneficiaries would receive lower benefits because they are subject to the retirement earnings test (RET). The RET temporarily withholds benefits from individuals who claim benefits before their full retirement ages and continue to work, if they earn more than the RET thresholds. Benefits reduced by the RET are paid to beneficiaries after retirement. By redefining covered earnings to include health insurance premiums, this option would make more beneficiaries subject to the RET and increase RET withholdings for some beneficiaries who are already affected. Those with temporarily reduced benefits would be disproportionately in the higher income quintiles.
20 For more information on the primary insurance amount, see https://www.socialsecurity.gov/oact/COLA/piaformula.html.
21 For more information, see https://www.socialsecurity.gov/retirementpolicy/program/poverty-decline.html.
22 Though average wages have grown faster than inflation, wages have not grown equally across the earnings spectrum. For the past several decades, wage growth has been stagnant at the low end of the earnings spectrum, where people are more likely to become poor. In about the last five years, however, the change in wage inequality has leveled out somewhat. Thus, it is hard to project whether future wages will become more unequal, maintain a similar level of inequality to current wages, or become less unequal. Increased inequality has important implications for Social Security—and potential changes to it. See, for example, Favreault, Melissa R. 2009. “Rising Tides and Retirement: The Aggregate and Distributional Effects of Differential Wage Growth on Social Security.” CRR Working Paper No. 2009-7. Chestnut Hill, MA: Center for Retirement Research at Boston College. http://crr.bc.edu/wp-content/uploads/2009/02/wp_2009-7-508.pdf. Our MINT projections assume a level of future wage inequality that is similar to the current level.
Board of Trustees, Federal Old-Age and Survivors Insurance and Federal Disability Insurance Funds. 2012. The 2012 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds. Washington, DC: Government Printing Office. https://www.socialsecurity.gov/oact/TR/2012/tr2012.pdf.
———. 2015. The 2015 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds. Washington, DC: Government Printing Office. https://www.socialsecurity.gov/oact/TR/2015/tr2015.pdf.
———. 2016. The 2016 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds. Washington, DC: Government Printing Office. https://www.socialsecurity.gov/oact/TR/2016/tr2016.pdf.
Burtless, Gary, and Sveta Milusheva. 2013. “Effects of Employer-Sponsored Health Insurance Costs on Social Security Taxable Wages.” Social Security Bulletin 73(1): 83–108. https://www.socialsecurity.gov/policy/docs/ssb/v73n1/v73n1p83.html.
Congressional Budget Office. 2012. CBO and JCT's Estimates of the Effects of the Affordable Care Act on the Number of People Obtaining Employment-Based Health Insurance. Washington, DC: CBO. https://www.cbo.gov/sites/default/files/112th-congress-2011-2012/reports/03-15-ACA_and_Insurance_2.pdf.
———. 2015. CBO's 2015 Long-Term Projections for Social Security: Additional Information. Washington, DC: CBO. https://www.cbo.gov/sites/default/files/114th-congress-2015-2016/reports/51047-SSUpdate.pdf.
Department of Health and Human Services. 2015. “Nationwide nearly 11.7 million consumers are enrolled in 2015 Health Insurance Marketplace coverage.” Press release. http://www.hhs.gov/about/news/2015/03/10/nationwide-nearly-11-point-7-million-consumers-are-enrolled-in-2015-health-insurance-marketplace-coverage.html.
Gruber, Jonathan. 2000. “Health Insurance and the Labor Market.” In Handbook of Health Economics, edited by Anthony Culyer and Joseph Newhouse, 645–706. The Netherlands: Elsevier. http://www.sciencedirect.com/science/article/pii/S1574006400801717.
Jensen, Gail A., and Morrisey, Michael A. 2001. “Endogenous Fringe Benefits, Compensating Wage Differentials, and Older Workers.” International Journal of Health Care Finance and Economics. 1(3): 203–226. http://link.springer.com/article/10.1023/A:1013711501104.
Kaiser Family Foundation. 2015. “2015 Employer Health Benefits Survey.” http://kff.org/report-section/ehbs-2015-summary-of-findings/.
Office of Management and Budget. n.d. Analytical Perspectives, Analytical Perspectives, Budget of the United States Government, Fiscal Year 2017. Washington, DC: Government Printing Office. https://www.whitehouse.gov/omb/budget/Analytical_Perspectives.
Social Security Administration. n.d. a. Research, Statistics & Policy Analysis, “MINT Overview.” https://www.socialsecurity.gov/policy/about/mint.html
———. n.d. b. Office of Retirement Policy, “Projection Methodology.” https://www.socialsecurity.gov/retirementpolicy/projection-methodology.html.
Social Security Advisory Board. 2010. Social Security: Why Action Should Be Taken Soon. Washington, DC: SSAB. http://www.ssab.gov/Portals/0/Documents/Sooner_Later_2010.pdf.
Whitman, Kevin, and Dave Shoffner. 2011. “The Evolution of Social Security's Taxable Maximum.” Policy Brief No. 2011-02. https://www.socialsecurity.gov/policy/docs/policybriefs/pb2011-02.html.