Health Affairs, 25, no. 6 (2006): 1568-1579 doi: 10.1377/hlthaff.25.6.1568 © 2006 by Project HOPE	New Online   House Health Reform Bill   Paying for Reform   Vetting AHIP's Report   HIV/AIDS Costs   Brief: Insurance Reform   HA Blog Top 10

This Article Abstract Reprint (PDF) Technical Appendix Submit a response to this article Alert me when this article is cited Alert me when Comments are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to Citation Manager Reprints & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Selden, T. M. Articles by Gray, B. M. Search for Related Content PubMed PubMed Citation Articles by Selden, T. M. Articles by Gray, B. M. Related Collections Health Reform Insurance Coverage Insurance - Employer-Based System Business Of Health Consumer Issues Insurance Market

Policy

Tax Subsidies For Employment-Related Health Insurance: Estimates For 2006

Abstract

 
Employment-related health insurance is subsidized through exemptions from federal and state income taxes, as well as from taxes for Social Security and Medicare. Proposals to modify this subsidy are a perennial subject of policy debate. We present tax-subsidy projections from a new data resource constructed using a statistical linkage between the establishment and household components of the Medical Expenditure Panel Survey (MEPS). We project that the total federal and state tax subsidy in 2006 for employment-related coverage of active workers will exceed $200 billion. We present per worker tax-subsidy estimates and an analysis of insurance incidence by establishment characteristics.

Tax subsidies have helped strengthen incentives for employment-based risk pooling during a period of rapidly rising costs.³ Nevertheless, economists have long criticized the design of the tax subsidy for encouraging excessive levels of coverage and excessive medical care consumption.⁴ Moreover, numerous researchers have criticized the subsidy for primarily benefiting higher-income families, raising concerns about both equity and the efficiency of providing the largest subsidies to families that are the least likely to be uninsured or on public coverage.⁵

Developing sound public policy regarding the tax treatment of employment-related coverage requires accurate estimates of the subsidies’ cost and incidence. This paper presents estimates from a new model that synthetically links the establishment and household components of the Medical Expenditure Panel Survey (MEPS) and projects those data forward to 2006.

Our aggregate estimate of the federal and state tax subsidy for employment-related health insurance premiums in 2006 is $208.6 billion. This estimate is approximately in line with other federal government estimates for 2006, as well as estimates for 2004 (appropriately inflated) by John Sheils and Randall Haught.⁶ In addition to providing updated estimates, the paper’s main contribution is that the MEPS data support detailed estimates of premiums, enrollment, and tax subsidies for subgroups of establishments categorized by firm size, full-time versus part-time employment, wage level, and most states. The results thereby provide information that is critical to gauging the differential impact of proposals that would alter the tax treatment of employment-related health insurance.

Study Data And Methods

Top Study Data And Methods Study Findings Discussion NOTES

 
MEPS database. Estimating the tax subsidy for employment-related health insurance requires data on employer and employee premium contributions, as well as data to support marginal tax rate simulations, including each worker’s income by type from all sources (including spouses), family composition, and itemizable deductions. Unfortunately, no single nationally representative data source contains all of these data elements, so researchers must combine information from multiple sources. In contrast to previous models, which typically rely on several sometimes outdated data sets, our access to confidential MEPS Insurance Component (MEPS-IC) data offered a relatively simple approach. In essence, we used data on workers from the MEPS Household Component (MEPS HC) to form synthetic workforces for each establishment in MEPS-IC. Although other approaches yield reliable national estimates, our model preserves to the greatest possible extent the correlations between establishments’ and workers’ characteristics, enabling us to provide estimates by detailed establishment characteristics.

MEPS-IC is an annual survey containing data on more than 32,000 private and public establishments in 2002. It is conducted by the U.S. Bureau of the Census under the sponsorship of the Agency for Healthcare Research and Quality (AHRQ).⁷ Because of the large sample size and high response rate, MEPS-IC is the leading source of data on employment-related insurance, including eligibility, enrollment by plan type, employer and employee premiums, and whether employee contributions are tax-preferred under Section 125 of the Internal Revenue Code.⁸ MEPS HC is an annual, nationally representative sample of the U.S. civilian, noninstitutionalized population sponsored by AHRQ and the National Center for Health Statistics (NCHS).⁹ It provides detailed information on insurance coverage, medical care spending, and a wide range of other health-related and socioeconomic characteristics. To obtain a sufficient sample of workers, we pooled MEPS HC data from 2000–2002.

Data linkage. For each MEPS-IC establishment, we selected MEPS HC workers who matched as nearly as possible the establishment’s region, industry, and size and whether the firm had multiple establishments. We then fine-tuned each establishment’s synthetic workforce by adjusting workers’ sampling weights to align as closely as possible with the establishment’s reported percentage of employees who were eligible for coverage, enrolled in each plan type, in each of three wage groups, full time, age fifty and older, male, and unionized.¹⁰ We projected this enhanced data set to 2006 from 2002, the MEPS-IC sample year, adjusting MEPS-IC sampling weights to account for employment growth and changes in share of covered workers.¹¹ We also adjusted employer and employee premiums for single and nonsingle coverage to reflect premium growth over this period.¹²

Tax rate simulation. We simulated marginal tax rates using MEPS HC data on each worker’s family and the National Bureau of Economic Research (NBER) TAXSIM model.¹³ Our focus was the tax subsidy for employer coverage, and we did not estimate tax subsidies on contributions to health savings accounts (HSAs), medical expenses paid for through flexible spending accounts, or deductible medical expenses in excess of 7.5 percent of adjusted gross income.¹⁴ Any change in the tax treatment of employment-based health insurance would likely affect these other subsidies. In particular, we did not adjust our estimates for the likely increase in medical expense deductions that would occur if the tax subsidy were removed entirely while leaving the existing itemization rules intact. We believe that our approach is appropriate when studying the magnitude and distribution of the tax subsidy—in contrast to analyzing the budgetary impact of a specific legislative proposal.

All results were weighted to be nationally representative of active employees in U.S. establishments. Retiree and self-employment coverage were excluded. All standard errors and statistical tests were adjusted to account for the complex design of MEPS-IC. Standard errors do not, however, reflect the variation introduced by statistical matching with MEPS HC, errors in simulating marginal tax rates, or adjustments made in projecting the data to 2006. Thus, reported standard-error estimates are likely to overstate the precision of our results.

Study Findings

Top Study Data And Methods Study Findings Discussion NOTES

 
Aggregate tax subsidy. The total tax subsidy or "tax expenditure" for the employer coverage of active, non-self-employed workers for 2006 is $208.6 billion (in 2006 dollars) (Exhibit 1). The federal income tax and FICA components are $111.9 billion and $73.3 billion, respectively. In contrast, exemption from state income taxation represents only $23.4 billion, or 11.2 percent of the total subsidy.

Average tax subsidy per worker. The average subsidy for all workers (regardless of eligibility or coverage) is $1,585 (Exhibit 2). Among eligible workers—the group most likely to bear the incidence of employer contributions—the average subsidy rises to $2,279.¹⁵ Among enrolled workers, the average tax subsidy is $2,778, or 35.4 percent of premiums. Also, as shown in Exhibit 2, the average tax subsidy for single coverage, $1,573, is less than half the $3,825 subsidy for nonsingle coverage (both family and employee-plus-one). This is not surprising, given that nonsingle coverage covers at least one person in addition to the worker.

Firm size Of particular interest are the private-sector results by establishment characteristics. Clearly, the subsidy disproportionately flows to workers in large firms. The average subsidy per worker in firms with 1,000 or more employees is $1,886, which is roughly two and a half times the $770 subsidy per worker in firms with fewer than ten employees. This is primarily because small firms are less likely than large firms to offer coverage. More than three-fourths of workers in the largest firms were eligible for coverage, versus only 38.4 percent in the smallest firms. Workers in the largest firms also have higher take-up rates conditional on being eligible, higher average subsidy rates per premium dollar, and higher average subsidies per covered employee. These differences, while statistically significant, are small relative to the difference in the share who are eligible.

Full-time versus part-time Differences in the subsidy per worker are even more pronounced when we compare establishments by the percentage of workers who work full time versus part time. In establishments where less than 25 percent of all workers work full time, the average subsidy per worker is only $244. This is primarily attributable to the small share of eligible workers in these establishments. In addition, the average subsidy among covered workers in these establishments is $2,212—21 percent less than the average subsidy to covered workers in establishments with the highest percentage of full-time employees.

Distribution by income level As noted in the literature, the incidence of the tax subsidy is regressive, because of the fact that higher-income workers have higher marginal tax rates, higher coverage rates, and higher premiums. Our results for workers in establishments grouped by predominant wage level are consistent with this finding.¹⁷ In establishments where more than half of workers are high-wage (earning more than $23.07 per hour in 2006 dollars), the per worker subsidy is $2,525, and the subsidy per covered worker is $3,283. In contrast, in establishments where workers are predominantly low-wage (earning less than $10.43 per hour), the per worker subsidy is $637, and the subsidy per covered worker is $2,268.

Distribution across industries The tax subsidy is also unevenly distributed across industries. The industries with the lowest per worker tax subsidies (ranging from $781 to $1,189) are agriculture, fisheries, and forestry; other services; retail trade; and construction (Exhibit 3). In contrast, average subsidies in the other industries listed in Exhibit 3 are much higher, ranging from $1,751 to $2,289. Once again, the primary driver of these differences is the share of workers who are eligible for coverage, although differences in the take-up and tax-subsidy rates also play a role.

Multivariate results. To provide additional insights into these private-sector differences, Exhibit 4 presents multivariate results from worker-weighted linear regressions of per worker subsidies on the firm characteristics in Exhibit 3. In the first column, the dependent variable is the average tax subsidy per worker (regardless of coverage). All of the differences we highlighted in discussing Exhibit 3 continue to play important roles in determining the per worker incidence of the tax subsidy. Thus, firm size appears to matter even when we control for differences in wages, the percentage of workers who work full time, and industry. Similarly, we continue to observe large differences by the percentage full-time, wage rate, and industry, holding other factors constant. In the second column in Exhibit 4, the dependent variable is the tax subsidy per covered employee. The coefficient estimates from the covered worker regression generally have the same sign as those in the first column. As in Exhibit 3, the magnitudes of the differences among covered workers are much smaller than those among all workers. Nevertheless, we continue to observe important and statistically significant differences by firm size, the percentage of workers who work full time, wage rate, and industry (although many of the industry coefficients have large relative standard errors).

Top Study Data And Methods Study Findings Discussion NOTES

Our aggregate estimates are approximately in line with published U.S. government projections. Our estimate of federal income tax expenditure is $111.9 billion—$20.6 billion less than the most recent Treasury estimate of $132.7 billion for 2006.¹⁸ That estimate, however, includes tax expenditures on retiree coverage, which Sheils and Haught estimated were 6.9 percent of total federal income tax expenditure in 2004.¹⁹ Adjusting our estimate for retiree coverage explains nearly half of the difference between our estimate and Treasury’s, the remainder being largely due to Treasury’s inclusion of tax expenditures for flexible spending accounts, TRICARE, workers’ compensation, and in-plant employer health care. In contrast, our federal income tax expenditure estimate is $20.3 billion larger than the $90.6 billion estimate by the Joint Committee on Taxation (JCT).²⁰ In contrast to our approach and Treasury’s, the JCT assumes that absent the existing tax exclusion, workers could treat employer and employee premium contributions as out-of-pocket medical expenses in itemizing medical expenses exceeding 7.5 percent of adjusted gross income. Thomas Selden and John Moeller estimated that adjusting for increased itemization reduces federal tax expenditure estimates for 1996 by 27.9 percent.²¹ Adjusting our 2006 estimate using the increased itemization and retiree coverage percentages yields a federal income tax subsidy estimate of $94.0 billion—much closer to the JCT estimate. We conclude that although basic differences complicate direct comparisons, our results are approximately in line with published government estimates on federal income tax expenditures.

Long-term increase. Our estimates highlight the substantial long-term increase in tax expenditures for health insurance. Looking back nearly two decades, the total tax subsidy in 1987 was only $81 billion (inflated to 2006 dollars).²² The corresponding estimate for 1996 is $105 billion (inflated to 2006 dollars).²³ Part of the 157.6 percent real increase between 1987 and 2006 can be attributed to the 33.4 percent increase in aggregate U.S. employment.²⁴ Most of the increase over time, however, is due to rising premiums. Indeed, the average tax subsidy rate per dollar of premiums held virtually constant over this period. The average subsidy rose a statistically insignificant amount, from 35.3 percent in 1987 to 35.4 percent in 2006, as declining federal marginal income tax rates were offset by rising marginal FICA tax rates, as well as increased tax exclusion of employee contributions.

Employment-related risk pooling. Employment-related risk pooling entails sizable and growing transfers of income from workers with low expected costs to higher-risk workers. Using 1987 data from the National Medical Expenditure Survey (NMES), Alan Monheit and colleagues showed that the tax subsidy at that time offset a large share of the gap low-risk workers faced between premiums and expected claims, thereby providing an inducement for them to participate in employer coverage.²⁵ Given the rapid growth in health costs and premiums, a similar analysis in 2006 would likely find a much larger gap between the premiums and claims of low-risk employees, even after accounting for the tax subsidy. Net benefits from employer coverage can be especially small for low-risk workers with low wages, given that these workers have lower average tax subsidy rates combined with greater access to free or low-cost public and "safety-net" coverage.

THIS PAPER ADDS TO THE LITERATURE finding that the tax subsidy is poorly targeted if the goal is to help stem the tide of rising uninsurance and public coverage in this country. Whether at least part of the public policy prescription for the private insurance market is to retarget tax incentives toward workers with the highest rates of uninsurance and public coverage is a public policy question that is likely to become more pressing over time.

Editor's Notes

 
Thomas Selden (tselden{at}ahrq.gov) is an economist at the Center for Financing, Access, and Cost Trends, Division of Modeling and Simulation, Agency for Healthcare Research and Quality (AHRQ), in Rockville, Maryland. Bradley Gray is an economist in the Health Care Operations and Policy Research Center of the CNA Corporation in Alexandria, Virginia.

This research was largely completed while Bradley Gray was employed at the U.S. Census Bureau. The authors appreciate the helpful comments and assistance of Jessica Banthin, Joel Cohen, Steven Cohen, Philip Cooper, Tim Dowd, Daniel Feenberg, Gillian Hunter, Alexandra Minicozzi, and Jessica Vistnes. All views expressed in this paper are those of the authors, and no official endorsement by the U.S. Department of Health and Human Services, the Agency for Healthcare Research and Quality, the U.S. Bureau of the Census, or the CNA Corporation is intended or should be inferred.

NOTES

Top Study Data And Methods Study Findings Discussion NOTES

 

1. Authors’ calculations for coverage at any point during the calendar year using the 2003 Medical Expenditure Panel Survey Household Component (MEPS HC) Public Use File HC-079, http://www.meps.ahrq.gov/Puf/PUFLookup.asp (accessed 12 April 2006). Employment-related coverage for these estimates includes TRICARE and private coverage held by people outside the household.
2. Office of Management and Budget, Analytical Perspectives, Budget of the United States Government, Fiscal Year 2007 (Washington: U.S. Government Printing Office, 2006), Tables 19-2 and 19-3.
3. J. Gruber, "The Impact of the Tax System on Health Insurance Decisions," International Journal of Health Care Finance and Economics 1, no. 3–4 (2002): 293–304; A.C. Monheit, L.M Nichols, and T.M. Selden, "How Are Net Health Insurance Benefits Distributed in the Employment-related Insurance Market?" Inquiry 32, no. 4 (1995/1996): 379–391; and M. Chernew, D.M. Cutler, and P.S. Keenan, "Increasing Health Insurance Costs and the Decline in Insurance Coverage," Health Services Research 40, no. 4 (2005): 1021–1039.[CrossRef][Web of Science][Medline]
4. M. Feldstein, "The Welfare Loss of Excess Health Insurance," Journal of Political Economy 81, no. 2 (1973): 1–22; M. Pauly, "Taxation, Health Insurance, and Market Failure in the Medical Economy," Journal of Economic Literature 24, no. 2 (1986): 629–675[Web of Science][Medline]; and J.P. Newhouse, "Medical Care Costs: How Much Welfare Loss?" Journal of Economic Perspectives 6, no. 3 (1992): 2–33.
5. L. Burman and R. Williams, "Tax Caps on Employment-based Health Insurance," National Tax Journal 47, no. 3 (1994): 529–545; J.F. Moeller, "Gainers and Losers under a Tax-based Health Care Reform Plan," Inquiry 32, no. 3 (1995): 285–299[Web of Science][Medline]; L. Burman and J. Gruber, "Tax Credits for Health Insurance," Discussion Paper no. 19 (Washington: Urban Institute, 2005); T. Selden and J. Moeller, "Estimates of the Tax Subsidy for Employment-related Insurance," National Tax Journal 53, no. 4, Part 1 (2000): 877–888; and T.M. Selden and D.M. Bernard, "Tax Incidence and Net Benefits in the Market for Employment-related Health Insurance: Sensitivity of Estimates to the Incidence of Employer Costs," International Journal of Health Care Finance and Economics 4, no. 2 (2004): 167–192.[CrossRef][Medline]
6. J. Sheils and R. Haught, "The Cost of Tax-Exempt Health Benefits in 2004," Health Affairs 23 (2004): w106–w112 (published online 25 February 2004; 10.1377/hlthaff.w4.106)[Abstract/Free Full Text]; Joint Committee on Taxation, Estimate of Federal Tax Expenditures for Fiscal Years 2004–2008, Pub. no. JCS-8-03 (Washington: JCT, 22 December 2003); JCT, Estimate of Federal Tax Expenditures for Fiscal Years 2005–2009, Pub. no. JCS-1-05 (Washington: JCT, 12 January 2005); and Selden and Moeller, "Estimates of the Tax Subsidy."
7. J.P. Sommers, List Sample Design of the 1996 Medical Expenditure Panel Survey Insurance Component, Methodology Report no. 6, July 1999, http://www.meps.ahrq.gov/papers/mr6_99-0037/mr6.pdf (accessed 16 August 2006); J.P. Sommers, Construction of Weights for the 1996 Medical Expenditure Panel Survey Insurance Component List Sample, Methodology Report no. 8, November 1999, http://www.meps.ahrq.gov/papers/mr8_00-0005/mr8.pdf (accessed 16 August 2006); J.P. Sommers, Imputation of Employer Information for the 1996 Medical Expenditure Panel Survey Insurance Component, Methodology Report no. 10, June 2000, http://www.meps.ahrq.gov/PAPERS/mr10_00-0039/mr10.pdf (accessed 16 August 2006); and MEPS-IC Technical Appendix, http://www.meps.ahrq.gov/MEPSDATA/ic/2002/techappendix.htm (accessed 24 April 2006).
8. Federal premium data are from U.S. Office of Personnel Management, The Fact Book, http://www.opm.gov/feddata/factbook/html/fb-p74.htm (accessed 8 December 2005); and OPM, "OPM Announces Smallest Average FEHB Premium Increase in Nine Years," Press Release, 15 September 2005.
9. J.W. Cohen et al., "The Medical Expenditure Panel Survey: A National Health Information Resource," Inquiry 33, no. 4 (1996/1997): 373–389[CrossRef]; and S. Cohen, Sample Design of the 1996 Medical Expenditure Panel Survey Household Component, Methodology Report no. 2, July 1997, http://www.meps.ahrq.gov/papers/mr2_97-0027/mr2.pdf (accessed 16 August 2006).
10. A similar method is used in J. Sheils and R. Haught, "Cost and Coverage Analysis of Ten Proposals to Expand Health Insurance Coverage," Report to the Robert Wood Johnson Foundation (Falls Church, Va.: Lewin Group, October 2003), Appendix A. For further analytic details, see our online technical appendix at http://content.healthaffairs.org/cgi/content/full/25/6/1568/DC1.
11. From 2003 MEPS-IC and Bureau of Labor Statistics, Current Employment Statistics, http://www.bls.gov/ces/home.htm#overview (accessed 12 April 2006); and G. Claxton et al., Employer Health Benefits: 2005 Annual Survey, September 2005, http://www.kff.org/insurance/7315/index.cfm (accessed 16 August 2006).
12. From 2003 MEPS-IC data, Claxton et al., Employer Health Benefits; J. Gabel et al., "Health Benefits in 2005: Premium Increases Slow Down, Coverage Continues to Erode," Health Affairs 24, no. 5 (2005): 1273–1280[Abstract/Free Full Text]; and Mercer Human Resource Consulting, "Health Benefit Cost Growth Will Slow in 2006 as Employers Continue Trimming," Press Release, 13 September 2005, http://www.mercerhr.com/summary.jhtml?idContent=1193570 (accessed 12 April 2006).
13. D. Feenberg and E. Coutts, "An Introduction to the TAXSIM Model," Journal of Policy Analysis and Management 12, no. 1 (1993): 189–194. TAXSIM version 6.0 was used. See National Bureau of Economic Research, "TAXSIM Related Files at the NBER," http://www.nber.org/~taxsim (accessed 15 August 2006).[CrossRef]
14. For estimates of these other tax expenditures as of 2004, see Sheils and Haught, "The Cost of Tax-Exempt Health Benefits."
15. See discussion in Selden and Bernard, "Tax Incidence and Net Benefits."
16. Authors’ calculations using 2003 MEPS HC data show that 59.9 percent of insured federal workers have family coverage. In contrast, 53.7 percent of insured workers in private establishments had nonsingle coverage in 2001, of which 32.2 percent was employee-plus-one coverage with much lower premiums. AHRQ, Employer-sponsored Single, Employee-Plus-One, and Family Health Insurance Coverage: Selection and Cost, 2001, Statistical Brief no. 23, August 2003, http://www.meps.ahrq.gov/papers/st23/stat23.htm (accessed 25 April 2006).
17. Hourly wage categories inflated to February 2006 dollars using the all-goods Consumer Price Index (CPI) are as follows: low wage, under $10.43; medium wage, $10.43–$23.07; and high wage, over $23.07.
18. OMB, Analytical Perspectives.
19. Sheils and Haught, "The Cost of Tax-Exempt Health Benefits in 2004."
20. JCT, Estimates of Federal Tax Expenditures for Fiscal Years 2006–2010, Pub. no. JCS-2-06 (Washington: U.S. GPO, 2006).
21. Unpublished estimates developed during preparation of Selden and Moeller, "Estimates."
22. Ibid.
23. Middle-case estimates from ibid.
24. Authors’ calculations using BLS estimates of seasonally adjusted March total nonfarm employment estimates, available at http://data.bls.gov/PDQ/outside.jsp?survey=ce (accessed 19 April 2006).
25. Monheit et al., "How Are Net Health Insurance Benefits Distributed?"

                      What's this?

This article has been cited by other articles:

				M. R. Mulkey and M. D. Smith The Long And Winding Road: Reflections On California's 'Year Of Health Reform' Health Aff., May 1, 2009; 28(3): w446 - w456. [Abstract] [Full Text] [PDF]

				T. M. Selden The Impact of Increased Tax Subsidies on the Insurance Coverage of Self-Employed Families: Evidence from the 1996-2004 Medical Expenditure Panel Survey J. Human Resources, January 1, 2009; 44(1): 115 - 139. [Abstract] [PDF]

				J. Oberlander The Politics Of Paying For Health Reform: Zombies, Payroll Taxes, And The Holy Grail Health Aff., November 1, 2008; 27(6): w544 - w555. [Abstract] [Full Text] [PDF]

				T. M. Selden and M. Sing The Distribution Of Public Spending For Health Care In The United States, 2002 Health Aff., September 1, 2008; 27(5): w349 - w359. [Abstract] [Full Text] [PDF]

				J. Hadley, J. Holahan, T. Coughlin, and D. Miller Covering The Uninsured In 2008: Current Costs, Sources Of Payment, And Incremental Costs Health Aff., September 1, 2008; 27(5): w399 - w415. [Abstract] [Full Text] [PDF]

				S. Y. Sessions and P. R. Lee A Road Map for Universal Coverage: Finding a Pass through the Financial Mountains Journal of Health Politics Policy and Law, January 1, 2008; 33(2): 155 - 197. [Abstract] [PDF]

				J. L. Hudson and T. M. Selden Children's Eligibility And Coverage: Recent Trends And A Look Ahead Health Aff., September 1, 2007; 26(5): w618 - w629. [Abstract] [Full Text] [PDF]

				H. S. Luft Universal Health Care Coverage: A Potential Hybrid Solution JAMA, March 14, 2007; 297(10): 1115 - 1118. [Full Text] [PDF]