The Tax Exclusion for Employer-Provided Health Insurance: Policy Issues Regarding the Repeal Debate

The Tax Exclusion for Employer-Provided
Health Insurance: Policy Issues
Regarding the Repeal Debate
November 21, 2008
Bob Lyke
Specialist in Social Legislation
Domestic Social Policy Division

The Tax Exclusion for Employer-Provided Health
Insurance: Policy Issues Regarding the Repeal Debate
Employer-provided health insurance is excluded from the determination of
employees’ federal income taxes, resulting in significant tax savings for many
workers. Comparable exclusions apply to federal employment taxes and to state
income and employment taxes. Since employment-based health insurance covers
three-fifths of the population under age 65, the exclusions also result in considerable
revenue loss to the government. Ending them could raise several hundred billion
dollars a year, depending on exactly what is repealed and how workers and employers
adjust. Some see this revenue as a source for financing health care reform without
explicitly raising taxes.
The federal income tax exclusion — the focus of this report — is criticized for
several reasons. Since it reduces the after-tax cost of insurance in ways that are not
transparent, it likely results in people with insurance obtaining more coverage than
they otherwise would. Not being explicitly capped or limited, it does little to restrict
the generosity of the insurance or annual premium increases. These attributes
contribute to what some economists argue is a welfare (or efficiency) loss from
excess health insurance for those with coverage and also contribute to rising health
care costs and spending. In addition, the income tax exclusion often is criticized
since it gives greater tax savings to higher income individuals and families, an
outcome that strikes many observers as wasteful and inequitable.
These arguments about the exclusion merit careful consideration as Congress
is starting to debate broad health care reform for the first time in 15 years. However,
the arguments involve complex issues, and other points and perspectives might be
taken into account. The welfare loss may be difficult to gauge considering how
consumers react to higher cost-sharing. Determining alternative tax benefits to
replace the exclusion that would not adversely affect people with high costs could be
challenging. The larger tax savings to higher income people might not be an
inequitable subsidy but only a consequence of the proper treatment of losses under
a progressive income tax.
The income tax exclusion has been in the tax code over 50 years, and its repeal
could have unintended consequences. For example, unless exceptions were made,
repeal would also terminate the exclusion for employer-paid disability insurance,
health care flexible spending accounts, and other benefits some consider useful.
The exclusion and regulatory decisions in the 1940s sometimes are said to be
the reason why employer-paid coverage is the predominant form of private health
insurance in the United States. There is something to this argument, but there are
other reasons why employment-based insurance arose and why it remains attractive.
These reasons make it difficult to predict the effect of ending the exclusion on the
future of employment-based insurance, a major policy issue.
This report will not be updated.

Scope of the Exclusion..............................................3
Coverage ....................................................3
Benefits .....................................................4
Some Policy Issues.............................................5
Origins ..........................................................7
Debate Over Consequences......................................9
Excess Insurance.................................................12
Some Policy Issues............................................14
Tax Equity......................................................17
Conclusion ......................................................21
Appendix .......................................................23
Congressional clients with questions regarding this report may contact the listed
author, Bob Lyke, or Janemarie Mulvey, Specialist on Aging Policy.

The Tax Exclusion for Employer-Provided
Health Insurance: Policy Issues
Regarding the Repeal Debate
Under current law, employer-provided health insurance coverage is excluded1
from employees’ income for determining their federal income taxes. Exclusions
also apply to federal employment taxes (Social Security, Medicare, and
unemployment taxes) and to state income and payroll taxes as well. Considering the
average cost of employment-based insurance, now around $4,750 a year for single
coverage and $12,700 for family coverage, these exclusions result in significant tax
savings for many workers.2
Since employment-based health insurance covers more than three-fifths of the
population under age 65, the exclusions also result in considerable revenue loss to
the government. Ending the exclusions could raise several hundred billion dollars
a year, depending on exactly what would be repealed and how workers and
employers adjust.3 Revenue effects of this magnitude make the exclusions a

1 An exclusion allows an item of income to be left out of tax calculations. Terms such as
exclusion, deduction, and credits are shown in the federal income tax formula that appears
in the Appendix.
2 Employer Health Benefits: 2008 Annual Survey. The Kaiser Family Foundation and
Health Research and Educational Trust, p. 14. Employers typically pay on average 84% of
the cost of single coverage and 73% of the cost of family coverage (p. 72); it is these shares
to which the tax exclusion applies, assuming employees pay their share with after-tax
3 In March 2007, the Joint Committee on Taxation estimated that the Administration’s
proposal to terminate the federal income and Social Security and Medicare exclusion for
employer coverage, the tax deduction for self-employed taxpayers, and the itemized
deduction for individuals not enrolled in Medicare would increase revenues by $1,228.3
billion over the FY2009-FY2012 period, an average of more than $300 billion a year. Tax
expenditure estimates (which unlike revenue estimates do not include the effects of
behavioral responses) of the exclusions alone typically are smaller. On July 30, 2008, the
Joint Committee on Taxation estimated calendar year 2007 tax expenditures for the
employer coverage exclusion to have been $143.3 billion for the federal income tax and
$100.7 billion for FICA (Social Security and Medicare) taxes. Tax Expenditures for Health
Care, JCX-66A-08. Also see Thomas M. Selden and Bradley M. Gray, “Tax Subsidies for
Employer-Related Health Insurance: Estimates for 2006,” Health Affairs, v. 25 no. 6
(November/December 2006), pp. 1568-1579.

tempting target for policy makers and other advocates who seek to finance health care
reform without explicitly raising tax rates.4
The federal income tax exclusion — the focus of this report — has been targeted
for other reasons as well. Since it reduces the after-tax cost of insurance to the worker
in ways that are not transparent, it likely results in people with insurance obtaining
more coverage than they otherwise would. Not being explicitly capped or limited in
some other manner, it does little to restrict the generosity of the insurance or annual
premium increases. These attributes contribute to what many economists argue is a
welfare (or efficiency) loss from excess health insurance for those with coverage.
These attributes also contribute to rising health care costs and spending throughout
the country.
In addition, the federal income tax exclusion often is criticized as unfair since
the workers’ tax savings depend on their marginal tax rate. High income workers
generally have greater savings than middle income workers, and the latter usually
have more than low income workers.5 When these tax savings are viewed simply as
an economic subsidy, this pattern strikes many people as wasteful and inequitable.
These three assertions about the income tax exclusion — that the revenue loss
could instead finance health care reform, that it contributes to inefficiency and rising
health care costs, and that higher income taxpayers unfairly get greater tax savings
— are the principal arguments put forth for repeal. Each is discussed in this report.
Before turning to them, however, the report discusses the scope and origins of the
exclusion, both of which are more complex than is generally recognized. The income
tax exclusion has been in the tax code for over 50 years, and its repeal could have
unintended consequences unless policy makers understand what transactions it covers
and what role it has had in the development of employment-based insurance. A
discussion of the scope and origins will also reveal some of the uncertainties repeal
would raise for both tax and health care policy.
Limited to these topics, the report does not address all issues that might be
raised about the exclusion. A comprehensive analysis would be difficult because of
limited knowledge about who actually pays for employer-provided health insurance
and how workers value insurance at different ages and income levels. While there is
general understanding about these matters — it is reasonable to assume that much of
the employer contribution is actually borne by workers through reduced wages —

4 A number of 110th Congress bills would eliminate the exclusion for employer-provided
health insurance, including S. 334 (Wyden), S. 1019 (Coburn), S. 1783 (Enzi), S. 1875
(DeMint) and H.R. 3163 (Baird). S. 397 (Martinez) and H.R. 914 (Ryan) would limit the
exclusion to specified amounts.
5 The marginal tax rate is the rate that applies to the last dollar of income received by a
taxpayer. Often it is the same as the statutory tax rate that applies to the highest band of
taxable income for the taxpayer. For married couples filing joint returns for 2008, the
statutory tax rate for taxable incomes not exceeding $16,050 is 10%, whereas the rate for
taxable incomes over $357,700 is 35%. An exclusion reduces the income that is taxed at
the highest rate; thus the tax savings on an exclusion of $1,000 generally would be $100
(i.e., $1,000 x .10) for someone in the lowest tax bracket and $350 (i.e., $1,000 x .35) for
someone in the highest bracket.

that is unlikely to be sufficient, and could be misleading, for drawing conclusions
applicable to the diversity of employment arrangements throughout the country. For
legislation, differences among types of employers (by size and industry, for example)
and types of workers (by gender, family status, and income) often are important.
Scope of the Exclusion
The statutory provision allowing an income tax exclusion for employer-provided
health insurance is short but complex. It covers more than health insurance (and
health plans that often are described as insurance) and applies to most but not all
workers, including members of the workers’ families. How other forms of coverage
would be affected by repeal of the exclusion for health insurance might be
considered. In order to understand the tax issue the exclusion resolves, it should be
considered along with another exclusion applying to benefits.
When employees receive something of value from their employer, the general
presumption under an income tax is that it will be taxable. Whether taxes will
actually be paid depends on the deductions and credits a taxpayer might claim, but
at least the starting position is that whatever is received should be taken into account
for purpose of determining an employee’s tax liability.
This rule is reflected in Section 61(a) of the Internal Revenue Code, which
provides that all forms of income from whatever source are taxable unless there is an
express exception. With respect to employer-provided health insurance, there are
two exceptions that are relevant, an exclusion for coverage and an exclusion for
benefits received. These exceptions reflect two different tax questions that arise with
this insurance, whether the provision of coverage by itself should be taxable
(regardless of whether the taxpayer actually uses insurance benefits) and whether the
insurance benefits used should be taxable (regardless of whether the taxpayer paid
for the coverage). Proposals to end the exclusion for employer-provided health
insurance involve the former question and not the latter, but there are interactions that
might be noted. In addition, because the exclusion for coverage applies to more than
health insurance as it is commonly understood, this section of the report also
discusses whether it might be appropriate to retain the exclusion for some purposes.
The exclusion for coverage can be found in Section 106(a) of the Code:
General rule — Except as otherwise provided in this section, gross income of an
employee does not include employer-provided coverage under an accident or
health plan.
The provision allows employees to exclude (that is, leave out of their income
tax calculations) employer-paid premiums or contributions to a trust or other fund for
their accident or health plan. There is an equivalent exclusion for employment taxes

elsewhere in the Code that pertains to both employees and the employer.6 The term
“accident or health plan” includes not only health insurance but also accidental death
and dismemberment insurance, short-term and long-term disability coverage, and
coverage through reimbursement arrangements such as health care flexible spending
accounts (FSAs) and health reimbursement accounts (HRAs).
Although usually described as an employee exclusion, the provision applies as
well to coverage of a spouse and dependents of the employee, whether prior to or
after the latter’s death. Coverage of former employees is also included.
The exclusion applies to both insured and self-insured plans. An insured plan
typically involves the purchase of coverage from a commercial carrier, while under
a self-insured plan the employer retains the financial risk.7 In addition, the exclusion
applies to coverage employees obtain separately in the individual insurance market,
provided employers pay the premiums directly (list-billing) or reimburse employees
under a secure arrangement that does not allow reimbursements to be diverted to
another purpose. Individual market insurance is not common with this exclusion.
Finally, the exclusion applies to employee-paid premiums under premium
conversion plans. Under these arrangements, which must be set up by employers,
employees reduce their taxable wages in exchange for their employers using the
money to pay health insurance. From an accounting perspective, the premiums are
no longer considered to be paid by the employees.8
Section 106(a) does not apply to self-employed individuals since they are not
considered employees under the Code.9 However, under current law they can obtain
roughly the same income tax savings from the above-the-line deduction authorized
under Section 162(l).
Exclusions for benefits received can be found in Sections 105(b) and 104(a)(3).
Section 105(b) allows taxpayers to exclude benefits they receive from employer-
financed accident or health plans. It applies to benefits for their spouse or dependents
as well. The exclusion is limited to expenses for medical care as defined in Section

213(d), except to the extent they were claimed as an itemized deduction in a prior

6 Section 3121(a)(2) for Social Security and Medicare taxes and Section 3306(b)(2) for
federal unemployment taxes.
7 In many self-insured health plans, the employer purchases stop-loss insurance to insure
against unexpected large losses.
8 Premium conversion plans are allowed under cafeteria plan provisions in Section 125.
9 Self-employed individuals include sole proprietors, general partners in a partnership,
limited partners who receive guaranteed payments, and individuals who receive wages from
S-corporations in which they are more than 2% shareholders.

year.10 Thus, the exclusion does not apply to disability benefits attributable to
employer payments since these benefits need not be spent on medical care.
In contrast, Section 104(a)(3) allows taxpayers to exclude amounts received
through accident or health plans except for amounts financed by an employer. The
taxpayers themselves might pay for this insurance, but that is not necessary; the only
requirement is that the employer not pay. The exclusion here is not limited to
expenses for medical care as is the exclusion in Section 105(b). Thus, it might
include disability benefits.
Section 104(a)(3) applies to individual and group policies, including
employment-based plans. If both employers and employees pay for the coverage, the
benefits might have to be apportioned to determine their tax treatment. This is not
necessary for medical expenses from health insurance, which would be excluded
under either Section 105(b) or Section 104(a)(3). However, disability benefits
attributable to employer payments would be taxable since they are not covered by
Section 105(b). Those attributable to employee payments would be exempt under
Section 104(a)(3).
Self-insured plans covering highly compensated employees must comply with
nondiscrimination rules. If they do not, the Section 105(b) exclusion may be limited
for these employees. The nondiscrimination rules do not apply to insured plans.
Self-employed people are covered by Section 104(a)(3), not Section 105(b).
Some Policy Issues
If Section 106(a) were repealed, employer-paid coverage under accident and
health plans would become taxable to employees for the federal income tax. States
would likely adopt the same treatment, as they do for most income tax provisions.
Repealing 106(a) would not by itself end the exclusions for employment taxes, but
most proposals would abolish those as well, if only to obtain additional revenue for
health care reform.
However, repealing Section 106(a) would also affect disability insurance and
health care reimbursement arrangements, among other employer benefit plans. These
possible changes have not received much attention, and they might not be what is
intended. Consideration might be given to retaining the exclusion for these benefits,
though, as will be seen, doing so may raise other complications.
Employer-paid disability insurance may be needed to prevent the adverse
selection that occurs when employees purchase coverage themselves. When left to
individual choice, people who think they might become disabled are more likely to
purchase insurance than those who do not; this drives up premiums as insurers adjust

10 Section 213(d) defines medical care broadly, including amounts paid “for the diagnosis,
cure, mitigation, treatment, or prevention of disease, or for the purpose of affecting any
structure or function of the body.” Internal Revenue Service publication 502, Medical and
Dental Expenses, provides further guidance.

for their anticipated cost. Employment-paid disability insurance may also be needed
to provide coverage to workers who will need but would fail to purchase it. Many
people underestimate the likelihood of becoming disabled during their working years
or do not think about the issue at all.
Of course, employers could purchase disability coverage for all employees even
if the exclusion were repealed; while the coverage would be taxable, the benefits
would then be exempt under Section 104(a)(3). However, many employees would
likely object to being taxed for coverage they would not choose in exchange for
receiving tax-free benefits most will not need.
For disability insurance, unlike health insurance, there are few proposals for
alternative tax subsidies to encourage people to purchase coverage. An expanded tax
deduction or a tax credit might provide some of the incentive as the exclusion for
employer-paid coverage, though neither would likely overcome the adverse selection
Health care reimbursement arrangements allow employees to pay out-of-pocket
medical expenses (deductibles, copayments, and things not covered by insurance) on
a pre-tax basis. The most common form is a health care flexible spending account
(FSA), though some employers offer health reimbursement accounts (HRAs).11 Most
health care FSAs are funded through salary reduction agreements under which
employees agree to receive lower take-home pay in exchange for benefits, though
others are funded at least in part by employers. Either way, Section 106(a) applies.12
Health care FSAs are popular with employees who have families or otherwise
anticipate having lots of medical and dental bills. They can help families manage
their expenses and cash flow. Employers find them useful for helping employees
deal with rising health care costs and for getting them to accept insurance with higher
deductibles and copayments as well as restrictions on certain services and products.
While the Section 106(a) exclusion raises many of the equity and efficiency concerns
for FSAs as it does for health insurance, retaining it for FSAs might be justified as
a way to hold down rising premiums, arguably the greater problem.13
At the same time, retaining the exclusion for these arrangements could raise
technical and enforcement issues. If FSAs and HRAs retained tax preferences but
insurance did not, over time it is likely the former would expand and the latter would
contract, at least in terms of covered benefits if not cost. In effect, FSAs and HRAs
would assume some of the functions now borne by insurance. This would counteract

11 For information on health care FSAs, HRAs, as well as health savings accounts (HSAs)
and medical savings accounts (MSAs), see CRS Report RS21573, Tax-Advantaged Accounts
for Health Care Expenses: Side-by-Side Comparison, by Bob Lyke and Chris L. Peterson.
12 Although salary reduction agreements are allowed under the cafeteria plan provisions of
Section 125, health care benefits under an FSA are exempt under Section 106(a) because
they are technically considered to be provided by the employer.
13 In the FY2008 and FY2009 budgets, President Bush proposed using Health Savings
Accounts (HSAs) to serve some of these functions. However, there are a number of
unresolved policy issues regarding HSAs, and they remain controversial.

some of the anticipated effects of ending the exclusion. It would also result in
disparate treatment for small employers, who often cannot establish FSAs due to
statutory nondiscrimination requirements.
Similar issues might arise with respect to the health care employers provide with
on-site doctors and nurses and through wellness programs. If Section 106(a) were
repealed, coverage for some of these services might become taxable, discouraging
their use.14 This would not further what some observers think is a useful way to
improve workers’ health. If instead they were not taxable, however, over time they
would likely expand, replacing some of the services now covered by insurance.
Attempting to delineate the line between what should be taxable and what should be
exempt could be challenging.
The origins of the tax exclusion for employer-provided health insurance have
become part of the debate over whether the exclusion should be repealed. To some,
initial determinations about a narrow tax issue, coupled with contemporaneous
regulatory decisions, account for the predominance of employer-provided coverage
in the United States. However, the history of the exclusion is not unambiguous in
this respect, and it reveals concerns about tax policy questions that remain today.
Section 106 was enacted in 1954 as part of a comprehensive revision of the
Internal Revenue Code. At the time, it consisted only of what later became Section

106(a), though the wording has changed several times. Current law subsections (b),

(c), and (d), which are not relevant to the present discussion, were added relatively15
Prior to 1954, there was no statutory provision that explicitly allowed an
exclusion for coverage under employer-provided accident and health insurance.
Regulatory rulings shortly after the modern income tax began had conflicting
outcomes, first providing that premiums on life, accident and health insurance were
considered income to employees but then saying they were not, at least in the case
of group life coverage.16 A 1943 ruling held that an employer contribution for group

14 Some might initially remain exempt as a working condition or de minimus fringe benefit
under Section 132.
15 Subsections (b) and (d) of Section 106 allow for an exclusion for employer contributions
to Medical Savings Accounts and Health Savings Accounts; they were added in 1996 and
2003, respectively. Subsection (c) disallows an exclusion for employer-provided coverage
for long-term care provided through flexible spending account or similar arrangement; it was
added in 1996 when long-term care premiums and other expenses were explicitly allowed
as part of the itemized deduction for medical expenses. For a time, the nondiscrimination
provisions for self-insured plans now found in Section 105(h) were included under Section


16 The initial ruling was in Reg. 45, Art. 33 as revised on April 17, 1919; the later ruling was
issued on January 28, 1921. For information on these and other rulings prior to 1954, see

medical and hospitalization insurance issued by a commercial insurer was exempt
income to workers.17 Since group coverage of this sort had started spreading in the
1930s, the need for resolution had become important. Nonetheless, apparently the
distinction between which insurance arrangements qualified for the exclusion and
which did not remained unclear.
Employer payments for individual coverage were always taxable prior to 1954.
Various rationales were advanced for the different treatment of group plans (or at
least certain group plans) and individual insurance. While the economic value of
employer payments for individual insurance could be determined from the premiums
individuals were charged, the economic value of group coverage would vary among
individuals, perhaps substantially. It was not at all clear — as it still is not today —
how to value coverage for someone who would not be insurable in the individual
market. In addition, there was parallel treatment at the time for life insurance:
employer coverage for individual life insurance (or life insurance with cash value)
was taxable, while group coverage for term insurance was not, at least to a point.
Finally, health insurance at the time often included wage continuation payments for
periods of illness; since rights for this were forfeited when employment was
terminated, it was not clear whether coverage by itself (in contrast to actual receipt
of payments) constituted income.
The enactment of Section 106 provided a basis for excluding employer
payments for individual insurance and certain other coverage such as union plans.18
It also clarified a number of other points, though other ambiguities remained.19
Sections 104(a)(3) and 105(b) were also enacted as part of the 1954 Code and
have remained largely unchanged since. The predecessor to Section 104(a)(3) can
be traced back to the Revenue Act of 1918, when it was added to clarify that amounts
received through accident or health insurance or from workmen’s compensation acts
would be exempt. Under that provision as well as the successor Section 22(b)(5) of
the 1939 Code, the exclusion was limited to amounts received through accident and
health “insurance,” which included commercially insured arrangements but not most
private employer or employee association plans. Benefits from private employer
plans were sometimes exempt when they were established pursuant to state law.
Further uncertainty about the definition of “insurance” arose from the 1952 Seventh
Circuit decision in Epmeier v. U.S., which applied the term to a self-insured plan

16 (...continued)
“Employer Health or Accident Plans: Taxfree Protection and Proceeds,” The University of
Chicago Law Review, vol. 21 (1953-1954), p 277-286.
17 Ruling letter August 26, 1943.
18 The exclusion for individual coverage is based on language in the House and Senate
reports on the legislation. Individual coverage still had to be provided through plans.
19 For example, the scope of the term “plan” remained unclear. “Taxation of Employee
Accident and Health Plans Before and Under the 1954 Code,” Yale Law Journal, vol. 64
(1954-1955) p. 222-247.

paying benefits based upon salary and length of employment, that is, a form of wage
The 1954 Code clarified these issues by carrying over the language in Section

22(b)(5) of the 1939 Code to present law Section 104(a)(3) and adding a new section,

105(b), that provided an exemption for benefits from self-insured plans. The
legislation allowed a limited exemption for wage replacement benefits in Section

105(d), but this was repealed by the Tax Reform Act of 1976.

Debate Over Consequences
By removing much of the uncertainty about these tax questions, the 1954 Code
furthered the growth of group insurance. Not only did the statutory exclusion reduce
the effective price of insurance, but it made coverage easier to obtain as employers,
having clearer guidance, became more willing to establish plans. One study shows
how the change both increased the amount of coverage obtained and engendered a
shift from individual to group policies.21
Even before 1954, however, group insurance coverage had been expanding.
One reason for this is that the Stabilization Act of 1942, which otherwise imposed
strict limits on wage increases, expressly exempted insurance and pension benefits.22
Employers eager to attract workers in a tight labor market expanded their benefit
plans instead. In addition, the National Labor Relations Board in 1948 held that
fringe benefits could be covered under collective bargaining agreements; this
furthered the growth of union plans.23
To some observers, these tax changes and regulatory decisions largely explain
the predominance of employment-based insurance in the United States. Coming at
a time of dramatic growth in the number of people with insurance, the policies were
both a response to changes in how insurance was provided and the cause of their

20 One reason for the complexity and uncertainty prior to 1954 was that it was not unusual
for health insurance at that time to have wage continuation benefits. Although arguably
these could be exempt if the taxpayer had purchased the insurance with after-tax earnings,
an exemption did not appear justified when the employer paid for the coverage, particularly
if that were excluded from taxation as well. Since wages from working were taxable, the
argument goes, why should continuation wages from not working be exempt?
21 Melissa A. Thomasson, “The Importance of Group Coverage: How Tax Policy Shaped
U.S. Health Insurance,” The American Economic Review, vol. 93, no. 4 (September 2003)
pp. 1373-1382.
22 P.L. 729 of the 77th Congress, 56 Stat. 765, Section 10.
23 The NLRB rulings were the subject of two federal court cases, Inland Steel Company v.
NLRB 170 F. 2d 247 and W.W. Cross and Company v. NLRB 174 F. 2d 875. The NLRB
order in the former case dealt with retirement and pension plans though it made passing
reference to insurance. The order in the latter case dealt with group accident and health

perpetuation.24 In retrospect, it might seem that a handful of legal developments,
each of which was aimed at resolving a narrow issue, was responsible for both the
way in which a majority of the population now finances its health care and the
problems associated with it.
There is something to be said for this argument. At the very least, it shows that
tax policies currently at issue were largely shaped during a particular period in
history, responding to what were perceived to be the needs at that time. An
implication might be that current policies should be reviewed in light of today’s
needs, and that further changes might now be appropriate.
At the same time, the historical argument about the importance of tax and
regulatory policies may be overstated. Enrollment in group plans expanded steadily
from 1939 onwards, starting prior to the tax ruling and regulatory changes mentioned
above and increasing roughly in tandem with increases in enrollments in individual
policies.25 If employment-based insurance supplanted individual market coverage,
it was not evident at the time.
Moreover, other forces possibly contributed to the rise of employment-based
insurance. According to one analysis, a range of influential groups — doctors,
hospitals, insurers, and employers — saw it as a way of achieving their own goals
while simultaneously reducing pressure for a government solution in the form of
compulsory social insurance. The latter had been advocated by some during the years
before World War Two.26
More important, the historical argument does not explain why employment-
based insurance has persisted as the predominant form of coverage for people under
age 65. While the proportion of this age group covered under employment plans has
declined somewhat in recent years, it is still about the same as it was 20 years ago.
Much of the recent decline has occurred in smaller businesses, where the principal
problem has less to do with financing, let alone the tax benefits, than with lack of
access to stable insurance pools in which risk and administrative costs are spread
over large numbers of participants.
Employers are not legally required to provide health insurance to their workers,
let alone pay for it. There is little doubt they do so partly because of tax advantages
from the Section 106(a) exclusion. It is uncertain how much employers gain from
the exclusion except from reductions in employment taxes, but even if they gained
nothing directly they would likely provide coverage in order to give their workers tax

24 The proportion of the population with hospitalization insurance increased from 10% in
1940 to 71% in 1957. By 1954, around half the population had group coverage. Herman
M. Somers and Anne R. Somers, “Private Health Insurance,” California Law Review, vol.

46 (1958), pp. 376 and 378.

25 The President’s Commission on the Health Needs of the Nation, Building America’s
Health, Financing a Health Program for America, vol. 4. Washington, 1953. Table 11.6.
26 Jacob S. Hacker, The Divided Welfare State: The Battle over Public and Private Social
Benefits in the United States, Cambridge (2002), p. 219.

savings. In a competitive labor market, workers’ tax savings on one form of
compensation might allow employers to reduce other forms.
However, it is likely that employers provide health insurance for other reasons
as well. One is that insurance is an attractive benefit to most workers, both for the
coverage it provides and for the time it saves them in shopping for policies on their
own. Given these preferences, when other employers competing for the same
workers offer health insurance, it is difficult for one employer not to do so. Second,
employers have an economic interest in healthy workers and, to some extent,
workers’ healthy families. According to some studies, healthy workers are more
productive. These advantages are more likely to be found in large firms employing
higher-skilled workers; they diminish as firm size shrinks and skill levels decline.27
But even if employers could readily dismiss ill employees and replace them with
others, there would be turnover costs.
For these reasons, whether employer-provided coverage would erode if the
exclusion were repealed cannot be forecast with any certainty, particularly given the
diversity of employment arrangements in the country. Much would depend on the
availability of alternative tax benefits that might replace the exclusion as well as
whether these would be linked to attempts to create stable insurance pools.
Regardless of how the exclusion came to be part of the tax code, these present factors
are likely to be more significant.
The future of employment-based insurance is one of the most important issues
in health care reform. Considering its dominant role in providing coverage for
people under age 65, changes to the insurance, whether intentional or not, ought not
be taken lightly. Most employment-based insurance is thought to have larger and
more stable risk pools than individual market insurance, and barring improvements
in the latter some would oppose policies that might threaten the former.28 Some
would argue that one of the first decisions to be made about reform is whether
employment-based insurance should be strengthened, weakened, or left alone, and
that decisions about the tax exclusion should be based upon that choice.

27 Ellen O’Brien, Employers’ “Benefits from Workers’ Health Insurance.” The Milbank
Quarterly, vol. 81 no. 1 (2003), pp. 5-40. The article puts forth a business case for
employer-provided health insurance and summarizes a number of relevant studies. Also see
Paul Fronstin and Ray Werntz, “The ‘Business Case’ for Investing in Employee Health: A
Review of the Literature and Employer Self-Assessment.” Issue Brief no. 267, Employee
Benefit Research Institute (EBRI), March 2004.
28 However, some analyses argue that there is considerable risk pooling in the individual
market, even where it is largely unregulated. Mark V. Pauly and Bradley Herring, “Risk
Pooling and Regulation: Policy and Reality in Today’s Individual Health Insurance Market,”
Health Affairs, vol. 26, no. 3 (2007), pp. 770-779.

Excess Insurance
One criticism of the exclusion for employer-provided health insurance is that
it reduces the after-tax cost of insurance to workers in ways that are not transparent,
likely resulting in their obtaining more coverage than they otherwise would. Not
being explicitly capped or limited in some other manner, it does little to restrict the
generosity of the insurance or annual premium increases. The exclusion thus
contributes to what some economists consider an excess of insurance coverage and
a significant welfare (or efficiency) loss for insured individuals and society as a29
whole. How repealing the exclusion would affect this welfare loss is a complicated
question, however, depending on how consumers react to higher cost-sharing. It
could be challenging to determine alternative tax benefits to replace the exclusion
without adversely affecting people with high costs.
The welfare loss from excess insurance, particularly insurance with low
deductibles and copayments, occurs because people pay more for health care services
than they would if everyone assumed more of the cost themselves.30 This outcome
is caused partly by the increased demand attributable to insurance (people generally
use more services when they have coverage because their effective price at the time
of service drops) and partly by the increase in market prices for services due to higher
aggregate demand. Increased market prices in turn encourage people to purchase
more insurance in order to avoid or minimize the additional financial risk from
higher prices.
Low deductibles and copayments reduce the financial risk insured people face,
and this welfare gain must be offset against the losses described above. Even taking
this into account, estimates of the net welfare loss from health insurance indicated
that it could be a large sum. In an article on this issue in 1973, Martin Feldstein
estimated that raising the coinsurance rate from one-third to two-thirds of private
hospital expenditures could have saved $4 billion a year out of a base of $12.6 billion
in 1969.31 A later revision of Feldstein’s work in a 1991 study estimated that the net
welfare loss from excess health insurance could have been $109.3 billion in 1984
dollars, assuming that insurance induces a large increase in gross prices and
consumers place large marginal value on medical spending. Assuming constant
prices and high marginal valuation, the estimate would still have been $33.4 billion.32
These figures obviously would be much higher in today’s dollars.

29 The term excess health insurance may sound inappropriate when there is widespread
concern about the growing number of people who are uninsured. However, the argument
pertains to people with coverage, not without it. One reason that some people do not obtain
insurance is that excess coverage increases the cost for everyone.
30 Ibid. To some observers, the country has serious problems of both excess insurance and
growing numbers of uninsured.
31 Martin S. Feldstein, The Welfare Loss of Excess Health Insurance, Journal of Political
Economy, vol. 81 (March/April 1973), pp. 251-280.
32 Roger Feldman and Bryan Dowd, “A New Estimate of the Welfare Loss of Excess Health
Insurance,” The American Economic Review, vol. 81, no. 1 (March 1991), pp. 297-301. For

Health insurance and health care have changed since the 1980s, and some of the
excesses at that time might have been attributable to unrestricted fee-for-service plans
that no longer are common. In 1988, 73% of workers with employment-based
coverage were enrolled in conventional insurance plans while in 2007 only 3% were.
In the latter year, 57% were enrolled in preferred-provider organization (PPO) plans
and 21% in health maintenance organizations (HMOs); in 1988, only 27% were
enrolled in either.33 Today it is not unusual for health plans both to control utilization
directly (requiring approval by gatekeepers for access to specialists, for example) and
to limit providers’ charges as a condition of their participating in a network. In these
respects, employment-based health insurance is subject to restrictions even though
the tax exclusion itself is not.
Nonetheless, for the most part employment-based insurance does not have high
deductibles. In 2007, over 80% of workers in HMOs did not have any general annual
deductible, which was also the case for just under 30% of workers in PPO plans and
over 50% of workers in point-of-service (POS) plans. For those with a general
annual deductible, the average amounts for single coverage were only $401 in
HMOs, $461 for PPO plans, and $621 for POS plans. While the average deductible
in high deductible plans with savings options was $1,729, only 5% of covered
workers were in these plans.34 (Enrollees in all of these plans would likely have had
copayments for most services.)
Whether moving to higher insurance cost sharing would reduce health care
spending is not at issue; notwithstanding measurement difficulties, economic theory,
actuarial experience, and empirical studies all indicate that it does. Probably the
most frequently cited research demonstrating this point is the RAND Health
Insurance Experiment (HIE), a carefully designed study of nearly 6,000 people
between 1974 and 1982. Among other things, the study showed that per capita
expenses for patients with a 95% coinsurance requirement for outpatient services
were 31% lower than those for patients without cost-sharing. Reductions were also
present but somewhat smaller for patients with lower coinsurance requirements, as
they were for those with deductible policies. Reductions occurred for a broad range
of conditions, especially for ambulatory care but also for hospitalizations. 35

32 (...continued)
other work on the concept of excess health insurance see Mark V. Pauly, Taxation, Health
Insurance, and the Market Failure in the medical Economy, Journal of Economic Literature,
vol. 24 (June 1986), pp. 629-675 and Thomas Rice, The Economics of Health Reconsidered,
Chicago, Health Administration Press, 1998, pp. 82-91.
33 Employer Health Benefits: 2007 Annual Survey. The Kaiser Family Foundation and
Health Research and Educational Trust. p. 63.
34 Employer Health Benefits: 2007 Annual Survey, pp. 63 and 87. Savings options
associated with the high deductible plans were either health savings accounts or health
reimbursement accounts. Coinsurance requirements of the different plans are difficult to
summarize here; for information, see pp. 87-122.
35 Willard G. Manning, et al. “Health Insurance and the Demand for Medical Care:
Evidence from a Randomized Experiment,” The American Economic Review, vol. 77, no.

3, June 1987. p. 258.

More debatable is what effect reductions in spending have on individuals’
health, which could affect measures of the welfare loss. A common reading of the
RAND HIE is that the health outcomes of those with high cost sharing were not
different from those having conventional coverage, with several exceptions.36 (The
exceptions included high blood pressure and vision imperfections in adults and
anemia in children.37) Although more health problems might have arisen for the high
cost sharing group had the experiment continued longer, there is no way to prove or
disprove this now.
However, the similarity in outcomes for participants in the HIE did not occur
because individuals with high cost sharing chose to forgo only services of limited
clinical value. Instead, it reflected a beneficial reduction in harmful medical services
that offset a detrimental reduction in useful services. Bad care and good care were
both forgone, resulting in outcomes similar to those having conventional coverage.
Today there may be more effective treatments for chronic diseases that higher cost
sharing would place at risk, and formerly untreatable or acute illnesses may now be
considered chronic and subject to treatment.38
A critical question here is whether greater cost sharing can be combined with
other reforms so that individuals reduce their spending but maintain or even increase
effective care. Possible changes include reducing or eliminating cost sharing for
services likely to impart high value, providing better information and helping people
use it, and making prices for services and products more available and transparent.
Some steps have already been taken in these directions (in consumer-driven health
care plans, for example), though it is too early to tell how far they will spread and
whether they will make a significant difference.
Some Policy Issues
Ending the tax exclusion could be helpful in reducing health care spending.
Since the exclusion reduces the net price of insurance, it likely encourages workers
to obtain more coverage than otherwise; since it is uncapped, tax savings and the
incentive to purchase more coverage grow unchecked as the cost of insurance rises.
In contrast, while an alternative tax benefit such as an expanded tax deduction or a
refundable tax credit might have no effect on the former problem (since it too would
reduce the net price of coverage), it could provide a brake on the latter, depending on
how it is designed.

36 See for example the statement in John F. Cogan et al., Healthy, Wealthy, and Wise: Five
Steps to a Better Health Care System, Washington, American Enterprise Institute, 2005, p.


37 Kathleen N. Lohr, et al., “Use of Medical Care in the RAND Health Insurance
Experiment,” Medical Care, vol. 24, no. 9, September 1986. pp. S72-S87.
38 Michael E. Chernew and Joseph P. Newhouse, “What Does the RAND Health Insurance
Experiment Tell Us About the Impact of Patient Cost Sharing on Health Outcomes?”
American Journal of Managed Care, vol. 14, no. 7 (July 2008) p. 412. The authors state
that “the RAND findings should not be used to justify higher cost sharing across the board.”

For example, President George W. Bush proposed terminating the exclusion and
replacing it with a standard deduction for health insurance (SDHI) of $7,500 for self-
only coverage or $15,000 for family coverage. The Joint Committee on Taxation has
estimated that adoption of the SDHI coupled with other parts of the proposal would
result in a net revenue increase of over $440 billion from FY2009 through FY2018.
The increase would occur largely because the SDHI would be indexed to changes in
the Consumer Price Index while the exclusion it replaces would have grown at the
higher rate for health insurance costs. The reduced tax savings from the change
presumably would affect the amount of coverage purchased.39
However, it may be somewhat speculative to assume that a reduction in tax
savings from indexing alternative benefits to the CPI would materialize to the extent
envisioned. If large savings were to occur in the out-years, the reduction in spending
they encourage might lead to calls for Congress to adjust the limits.40 It may also be
speculative whether ending the exclusion would also encourage more effective care.
One possibility is that a deduction or credit would be allowed only if the insurance had
incentives to this effect. Another is that if the cost of insurance exceeds what could
be claimed as a deduction or taken into account as a credit people would become
more cost-conscious and concerned about effectiveness. Whether these things would
occur is unknown, however.
There are other issues with potentially ending the exclusion that merit attention.
For one thing, in replacing the exclusion with a capped deduction or credit (for
example, the $7,500 and $15,000 figures in the President’s proposal), it is not
immediately clear where to set the cap. Setting it too low might erode needed
insurance benefits for some people; if many were affected, the level might not be
sustainable. Setting it too high may lock in current welfare losses, delaying significant
savings for years to come. (The same dilemma applies to proposals to cap the
exclusion rather than replace it with a deduction or credit.)
The issue is complicated by the wide range of health care costs throughout the
country. As analysis by the Congressional Budget Office shows health care spending
per capita in 2004 varied from approximately $4,000 in Utah to $6,700 in
Massachusetts. While differences in spending reflect more than price differentials
(they are also influenced by differences in health status, preferences, and other
matters), the magnitude of the differences suggests price disparities might not be
inconsequential.41 Even if they were, spending may be the better indicator for setting

39 Description of Revenue Provisions Contained in the President’s Fiscal Year 2009 Budget
Proposal. Joint Committee on Taxation (March 2008), JCS-1-08, p. 317. The $440 billion
estimate also reflects other parts of the proposal: repeal the health insurance deduction for
self-employed taxpayers and the itemized deduction for people not enrolled in Medicare.
A footnote states that the “estimate is very preliminary and subject to change upon
clarification of the proposal. Many of the details of the President’s health proposals are
40 Of the $440 billion estimate cited above, nearly 70% occurs in the last three fiscal years,

2016 through 2018, and 28% occurs in the last year alone.

41 Congressional Budget Office, Geographic Variation in Health Care Spending, February,

a cap unless it is plausible to assume that usage can be reduced in high cost areas. A
study by Milliman of 14 major metropolitan areas shows similar spending
differences. 42
In addition, health care costs vary among people with employment-based
insurance regardless of where they live. In 2005, median health care spending was
$2,525 for people ages 55 through 64 and $623 for those ages 25 through 34. For
women ages 25 through 64 with this insurance, median spending was $1,086, while
for men it was $530. People ages 25 through 64 with this insurance who reported
being in fair health had median expenditures of $3,777, while those who reported
being in very good health had median expenditures of $893.43
To the extent that health insurance reflects these differences, its cost will vary.
The largest differentials might occur among small employer plans, particularly in
states where there are few restrictions on premium variations in insured plans. (At the
same time, small employers might drop coverage that becomes too expensive, which
large employers would be unlikely to do.) Workers who incur very high costs might
have illnesses or injuries that require them to leave their jobs, reducing the
differentials, though sometimes the high cost person is a family member.
Conceivably, tax limits on insurance costs might be adjusted to compensate for
some of these differences. For example, the ceiling on a deduction or credit (or the
cap on the exclusion) could be set higher for people ages 55 through 64.
Alternatively, high cost people might be allowed an additional deduction based upon
some portion of their income (similar to the current itemized deduction, though here
not limited to itemizers). However, these variations would add complexity, and they
might perpetuate differentials that ideally should be reduced.
Ending the exclusion would also raise the question of what amount should be
included in employees’ income for tax purposes. An average premium (for example,
the COBRA premium) has been suggested.44 However, the economic value of group
health insurance varies widely among workers, as was recognized in debates over the
exclusion prior to its codification in 1954 (see the discussion above in the “Origins”
section). The value of insurance to a young, healthy worker is likely far less than the
value to an older worker in poor health, or who perhaps has a family member in poor
health. It would seem unfair to require both workers to recognize the same additional

41 (...continued)

2008, pp. 1 and 10.

42 2008 Milliman Medical Index, p. 4.
43 The estimates were done by CRS using the Medical Expenditure Panel Survey. Reflecting
the high cost of people with serious health care needs, the mean expenditure for people
reporting being in fair health was $11,294, whereas for people reporting very good health
it was $2,342.
44 Under Title X of the Consolidated Omnibus Budget Reconciliation Act of 1985 (COBRA,
P.L. 99-272), employers with 20 or more employees must offer separated employees the
option of continuing the employer’s health plan coverage for a temporary period.
Employers may be charged up to 102% of the total premium for this coverage.

amount in the income on which they must pay taxes. Age-based variations might
seem more equitable, but the opposite might be true if the younger worker had poor
health and the older worker were in good health. Moreover, if the total amount of the
employer contribution had to be recognized, small increases for young workers would
mean very large increases for older workers, likely to provoke their opposition to the
policy change.
Determining a fair allocation of the employer contribution that workers should
recognize could become complex. For example, some might think that income should
be taken into consideration. Income often increases with age, so older workers
arguably could afford the larger tax obligation. However, average incomes increase
only until the late middle ages; then they start to decline just when health care
expenses rise more sharply.45 In addition, some research suggests that some older
workers already pay for their higher health care costs through lower wages.46 To the
extent this is true, it might not seem appropriate to increase the amount of the
employer contribution older workers must recognize.
The current exclusion avoids all these technical issues; everything is just left out
of the tax calculations. To some observers, this is a virtue. At the same time, while
the simplicity of omission has obvious attractions, it does not show variations in the
allocation of tax subsidies that some would consider important. Differences in the
amount of subsidy by age, geography, and health condition might seem inappropriate
to some. Ending the exclusion would improve transparency and allow these
difference to be assessed and debated. At the same time, it might create new
controversies that divert attention from other health care issues.
Tax Equity
Two equity issues arise with respect to the tax exclusion of employer-provided
coverage. One is how people with employer-provided coverage are treated for tax
purposes compared to otherwise similar people who have coverage purchased in the
individual market. The other is how people with employer-provided coverage are
treated at different income levels. The former issue, which is addressed first, is easier
to analyze, and there is an obvious solution to what some people consider a problem.
The latter issue is more complicated than is usually recognized and may defy simple
The tax exclusion sometimes is criticized for providing tax savings when
employers pay for the insurance, while coverage purchased in the individual market
generally has no tax savings. To some, this creates an unwarranted distortion in the

45 Income, Poverty, and Health Insurance Coverage in the United States: 2006, U.S. Census
Bureau, Current Population Reports P60-233 (August 2007), Table 1, p. 5. Median income
increased from $30,937 for householders ages 15 through 24 to $64,874 for householders
ages 45 to 54; it then declined for those ages 55 to 64, in part because they were no longer
in the labor force.
46 Louise Sheiner. Health Care Costs, Wages, and Aging. Finance and Economics
Discussion Series. Federal Reserve Board, 1999.

insurance markets. In their view, it is not obvious why health insurance, which is
essentially a personal matter, should be tied to one’s job. It is argued that the tie limits
workers’ health plan choices and penalizes workers who might be able to work more
productively elsewhere.
Under current law, there are two provisions that provide tax savings to people
who purchase insurance in the individual market. One is the 100% deduction allowed
self-employed taxpayers who buy policies for themselves and their family members;
this applies only to a small number of people, less than 3% of all who file returns.47
The other is the itemized deduction for unreimbursed medical expenses, which is
available only to taxpayers who itemize their deductions and only to the extent the
expenses exceed 7.5% of their adjusted gross income. For most taxpayers, the
standard deduction is larger than the sum of their potential itemized deductions, and
of those who itemize, most do not have extensive unreimbursed medical expenses.
In 2005, about 35% of all returns had itemized deductions, and of these, less than 21%
(about 7% of all returns) claimed the medical expense deduction. Most people who
purchase insurance in the individual market cannot claim either of these deductions.
This imbalance in individuals’ tax savings could largely be remedied by allowing
an above-the-line deduction for premiums that is available to everyone without
employer-provided coverage. This type of deduction, which could be claimed whether
or not one itemized, might be a standard deduction like President George W. Bush
proposed or limited to premiums actually paid; in the latter case, the income tax
savings to individuals would be approximately the same as those from the exclusion
for equivalently-priced employer-provided insurance.48 The income tax deduction just
described would not provide savings from the exclusion for employment taxes, but an
additional deduction might be allowed people with employment income for that
Some might oppose this solution since employer-provided insurance generally
has broader and more stable risk pooling; in their view, this should be protected and
rewarded despite the apparent inequity in tax savings.
The second tax equity issue is how people with employer-provided coverage are
treated at different income levels. Income tax savings from the exclusion for
employer provided health insurance depend on taxpayers’ marginal tax rates. For low
income taxpayers, the savings on federal income taxes might be 10% or as little as

47 IRS data for tax year 2005. The health insurance deduction for self-employed taxpayers
is claimed on about 3% of all returns. In addition to being self-employed, the taxpayer must
meet a number of additional conditions. The deduction cannot exceed the net profit and any
other earned income from the business under which the plan is established, less deductions
taken for certain retirement plans and for one-half the self-employment tax. It is not
available for any month in which the taxpayer or the taxpayer’s spouse is eligible to
participate in a subsidized employment-based health plan (i.e., one in which the employer
pays part of the cost). These restrictions prevent taxpayers with little net income from their
business (which is not uncommon for a new business) from deducting much if any of their
insurance payments.
48 The savings would be the same if the employer paid for all the coverage or the workers
could pay their share with pre-tax dollars under a premium conversion arrangement.

none, depending on their income. In 2008, for example, single tax filers generally will
not have a regular tax liability until their income exceeds $8,950, the sum of their
standard deduction and personal exemption. Until their income exceeds $16,975,
their taxable income (the difference between $16,975 and $8,950) would be taxed at

10%. Thus, if single taxpayers with wage income of $13,000 were given employer-

paid insurance worth $3,000, their savings from the income tax exclusion would be
$300 (i.e., $3,000 x 10%).49
For higher income taxpayers, the income tax savings would be greater. For
single filers in the top tax bracket of 35%, the tax savings from the exclusion of
$3,000 of coverage would be $1,050, appreciably more than the savings for low
income workers. For middle income taxpayers in the 15% or 25% brackets, the
income tax savings would be $450 or $750, respectively. (In 2008, the 15% bracket
for single filers applies to taxable incomes of $8,026 to $32,550; the 25% bracket to
taxable incomes of $32,551 to $78,850; and the 35% bracket to taxable incomes over
$357,700). Different figures apply to taxpayers filing as married or head of
households, but the pattern just described would be the same.
Taxpayers often get additional savings from a parallel exclusion on their state
income taxes. Currently, 43 states and the District of Columbia impose taxes on
individual incomes. Rates vary from state to state, and while some of the tax savings
would be offset by a reduction in the itemized deduction for state income taxes,
typically higher income taxpayers would get greater savings.
This picture is complicated by tax savings from the exclusion for employment
taxes. Medicare taxes of 1.85% of wages (considering just the employee’s share)
apply to workers regardless of income; thus all would be affected equally by an
exclusion for insurance worth $3,000. For Social Security taxes, however, the
exclusion benefits low and middle income workers but not those with wages above
$102,000, the wage base ceiling in 2008. The exclusion would save the former 6.2%
(considering just the employee’s share), or $186 for $3,000 worth of insurance.
Employment taxes aside, the figures above show that middle income taxpayers
generally benefit more from the exclusion than low income taxpayers, and that high
income taxpayers generally benefit more than those with middle incomes. When these
tax savings are viewed purely as an economic subsidy, this pattern seems unfair if not
wasteful. It is unlikely that an insurance subsidy program using appropriated funds
would be designed in this manner.
The equity problems are likely to be exacerbated in several respects. For one
thing, higher income workers are more likely to be eligible for employer-provided
health insurance. One study showed that 89.6% of workers with family incomes
above 400% of the federal poverty level were eligible for employer coverage in
contrast to 39.8% of workers with incomes below the poverty level in 2005.50 The

49 A full-time employee working 2,000 hours a year at the new national minimum wage of
$6.55 an hour would earn $13,100 for the year.
50 Lisa Clemans-Cope et al., Changes in Employees’ Health Insurance Coverage 2001-

same study also showed that 83% of those eligible in the higher income group actually
accepted coverage, compared to 63.5% of those eligible in the lower income group.
Where these patterns occur, the disparities in tax savings based on the $3,000 figure
in the discussion above would understate differences among income groups.
Moreover, higher income taxpayers may benefit more from the exclusion if their
employment-based insurance on average provides greater economic benefits than
identical coverage provides lower income taxpayers. As described in the previous
section, income on average increases with age, at least until the late middle ages, and
age usually is associated with higher health care expenditures.
The preceding analysis is not the only way of looking at the tax equity issue. To
some extent, the tax savings shown above might not be an inequitable subsidy but
only a consequence of the proper treatment of losses under a progressive income tax.
In a progressive tax system, when gains are taxed at higher and higher rates depending
on income, then it is conceptually appropriate to deduct losses at those same rates. To
the extent that health insurance is viewed as a method of smoothing losses over time
and individuals, then the greater tax savings for higher income people from the
exclusion would not be inappropriate or, for that matter, unfair.
The argument about progressivity would have greater force if health insurance
covered only catastrophic expenditures that everyone would clearly see as losses (for
example, hospitalization for accidents or heart attacks). However, since a portion of
its cost is for anticipated expenditures for routine care, not all of the exclusion can be
associated with a loss of this nature.51
The root issue is whether health care is like other forms of personal consumption.
On reflection, many people might conclude that it is not. Health care takes time,
provides no personal pleasure, and, most important, usually is designed to return
patients to a condition prior to an accident or illness that no one would choose to have
had. Absent the latter, it does not result in gratification or enhancement of one’s well
being. 52
On the other hand, health insurance, which is the subject of the exclusion,
provides additional value because people recognize they cannot predict their health
with certainty and worry that they might not be able to afford care when they need it.
It buys peace of mind as well as services. In this respect, insurance may be like other

50 (...continued)

2005. Issue Paper, Kaiser Commission on Medicaid and the Uninsured. 2006.

51 The distinction between catastrophic and routine expenses is not always an indication of
what might be a deductible loss. Some high expenditures are voluntary (for example,
childbirth of a planned baby), whereas some routine expenditures are incurred only to
forestall losses (for example, statin drugs).
52 This was the argument put forth in a once well-known article by William D. Andrews,
“Personal Deductions in an Ideal Income Tax,” Harvard Law Review, vol. 86 (1972), p. 309.
For a more recent discussion, see Jay A. Soled, “Taxation of Employer-Provided Coverage:
Inclusion, Timing, and Policy Issues,” Virginia Tax Review, vol. 15 (1995-1996), p. 447.

forms of personal consumption in that it is not directly associated with a loss that
should be recognized under an income tax.
Since it is difficult to determine what portion of an insurance premium merits
treatment as a tax loss, it is challenging to think of what might be an appropriate
remedy. The issue is compounded by the diversity of insurance arrangements and
uncertainty about the extent to which particular workers ultimately bear the cost of
employer-provided coverage.
There is much to be said for moving towards a health care financing system that
is more transparent, equitable, and efficient. Given the extraordinary complexity of
American health care and increasing concerns about rising costs and the growing
numbers of uninsured, new policy approaches are widely being given careful
Whether ending the exclusion should be part of these approaches depends on
what the replacement policies are. Problems attributable to the exclusion might also
occur with new measures, or other problems could arise. It is difficult to evaluate one
step without knowing the other. Given the origin of the exclusion, it would not be
inappropriate to consider revisions in light of today’s problems. However, one might
move carefully before dismantling a policy that has been in place for over 50 years.
Some policy-makers appear to be interested in the exclusion primarily to raise
money for health care reform. In order to make a substantial reduction in the number
of uninsured, for example, there would likely have to be significant subsidies, either
through the tax system with refundable tax credits, through the insurance system with
premium subsidies and rating limitations, or through some other means. Revenue
increases from ending the exclusion would be one way to pay for these initiatives.
However, if the sole objective were to raise money, it might be simpler just to
increase taxes generally. For many people who currently have employment-based
insurance — over three-fifths of the population under age 65 — the mathematical
effect might be roughly the same, at least for income taxes. For people under age 65
paying for other private insurance, an offsetting deduction could provide tax equity.
People under age 65 with other public insurance might not be affected, considering
that their income often is too low to be taxed.
Nonetheless, some argue that the exclusion should be ended not to raise money
but to improve efficiency and limit costs and spending. In the absence of alternative
proposals to achieve these goals, termination might be appropriate.
A principal policy decision appears to be whether to maintain and possibly
strengthen the employment-based system of health care. If that is the goal, then
maintaining the exclusion might be appropriate since it is unclear what the effects of
termination would be over time. If instead the goal were to move towards individual

market insurance or an expansion of public coverage, then ending the exclusion
should be given greater consideration.

The general formula for calculating federal income taxes appears below. The list
omits some steps, such as prepayments (from withholding and estimated payments)
and the alternative minimum tax.

1.Gross income (everything counted for tax purposes)

2.Minus deductions (or adjustments) for determining adjusted gross income
(AGI) — “above the line deductions”
3.Equals AGI
4.Minus greater of standard or itemized deductions
5.Minus personal and dependency exemptions
6.Equals taxable income
7.Times tax rate

8.Equals tax on taxable income (i.e., “regular tax liability”)

9.Minus credits

10.Equals final tax liability