May 10, 2004

by Edwin Park and Robert Greenstein

PDF of report
Press Release: HTM | PDF

View Related Analyses

If you cannot access the files through the links, right-click on the underlined text, click "Save Link As," download to your directory, and open the document in Adobe Acrobat Reader.


At the cost of $25 billion over ten years, the President’s budget seeks to expand enrollment in Health Savings Accounts by allowing HSA participants to claim a tax deduction for the premium costs of high-deductible health insurance policies that they purchase in the individual health insurance market.  Analysis by M.I.T. economist Jonathan Gruber, one of the nation’s leading health economists, finds that because of its adverse effects on employer-based coverage, this deduction would likely cause the ranks of the uninsured to increase by 350,000.

Health Savings Accounts were created as part of the Medicare drug legislation.  Under that law, individuals who enroll in high-deductible health insurance plans — whether through their employers or on their own — may establish tax-favored savings accounts.  These accounts provide a lucrative tax shelter for those in higher tax brackets — contributions to the accounts are tax-deductible, earnings on funds in the accounts accrue tax-free, and withdrawals from the account are not taxed if they are used to pay for out-of-pocket medical costs.[1]  The tax-free withdrawals may be used for deductibles, co-payments, and uncovered medical costs.  They may not be used to pay the premium costs of health insurance.[2]

The establishment of HSAs represents a major event not only in health policy but in tax policy as well; the tax-sheltering opportunities that HSAs provide are unprecedented.  HSAs are the only feature of the tax code that provides for both tax-deductible deposits into accounts and tax-free withdrawals from the accounts.  Even so, the Administration is proposing to make HSAs still more lucrative as a tax break, especially to those in higher tax brackets.  The Administration’s budget proposes that individuals who use HSAs be allowed to take a separate, additional deduction for the premium costs of a high-deductible health insurance policy, as long as the policy is purchased in the individual health insurance market.  (The deduction would not be available to HSA participants who obtain high-deductible coverage through their employer.)  This deduction would be available without regard to whether an individual itemizes deductions. 

The Joint Committee on Taxation estimates the cost of the proposed deduction at $25.1 billion over ten years.  The Administration’s estimate of the costs is almost identical — $24.8 billion over ten years.

The proposed deduction would be of significant benefit to high-income taxpayers; they would receive the largest tax benefit from the deduction because they are in the highest tax brackets.  The vast bulk of people in the higher tax brackets who would use the deduction, however, will have health insurance regardless of whether the deduction is established. 

As a result, the deduction poses a problem: it would not do much to help people who cannot afford insurance to secure it, but would encourage some employers to drop employer-based coverage or not to offer it in the first place.  In ongoing research conducted for the Kaiser Family Foundation, Professor Gruber analyzed the coverage effects of the proposed deduction.  He found that the number of workers who would lose coverage because of actions by employers to drop coverage (or to decrease employer contributions toward health insurance premiums) would likely exceed the number of uninsured individuals who would gain coverage as a result of the deduction.[3]

Workers who do not earn enough to owe income tax would receive no benefit from the deduction.  For moderate- and middle-income taxpayers in the 10 percent or 15 percent tax brackets, the deduction would reduce the cost of health insurance policies by only 10 percent or 15 percent, too little in most cases to make health insurance affordable.  This is significant because about three-quarters of all U.S. households — and something like 90 percent of the uninsured — are either in the 10 percent or 15 percent tax bracket or earn too little to owe income tax.[4] 

Table 1
Projected Effects of Fiscal Year 2005
Administration HSA Deduction Proposal
in Reducing the Number of Uninsured

Projected number of total participants in the tax deduction

7.98 million

Number of participants who would previously have had health insurance coverage

6.91 million (86.6%)

Number who would previously have been uninsured and would gain coverage

1.07 million (13.4%)

Number who would previously have had employer-based coverage but would become uninsured as their employers dropped coverage or reduced their premium contributions.

-1.41 million

Net effect on number of individuals with health insurance coverage


*  Communication with Professor Jonathan Gruber, March 12, 2004.  Numbers may not add due to rounding. 

The Deduction’s Budgetary Impacts

The deduction also would affect the budget.  Unless its costs were fully offset, the proposal would increase federal deficits.

Rising Cost Estimates for the HSA Tax Breaks in the Medicare Drug Legislation

The Joint Committee on Taxation (JCT), Congress’ official “scorekeeper” on tax legislation, estimated last year that the HSA provisions of the new Medicare drug law would cost $6.4 billion over ten years. The Congressional Budget Office incorporated this JCT estimate into its overall estimate of the cost of the law.

The Joint Committee on Taxation estimate was based on an assumption that HSA use would start at one million participants in 2004 and rise to three million by 2013. Most analysts now believe, however, that this estimate substantially understates likely HSA use, given the widespread attention that HSAs are receiving and the intention of various insurance and financial investment companies to offer HSAs and high-deductible policies and market them heavily.

The Administration now estimates that the HSA provisions of the new Medicare law will cost $16 billion over ten years, two-and-a-half times what Congress assumed when the law was enacted.* This difference is one part of the gap between the estimate upon which Congress relied — that the legislation would cost about $400 billion over ten years — and the Administration’s current estimate that the legislation will cost $534 billion.

The deduction that the Administration is now proposing would further increase HSA costs in two ways. First, the deduction itself carries direct costs. Second, by causing participation in HSAs to rise, the deduction would further increase the cost of the HSA tax breaks included in the Medicare drug law.

* Office of Management and Budget, Analytical Perspectives: Fiscal Year 2005, p. 292. The OMB budget documents show that the existing Health Savings Accounts will have ongoing costs of $7 billion over the next five years. Treasury Department staff have indicated to Congressional staff that the 10-year cost is approximately $16 billion over 10 years.

The remainder of this analysis examines these issues in more detail.


Value of the Deduction Would Rise with Income and Primarily Benefit High-Income Taxpayers

With any deduction, the higher an individual’s tax bracket, the greater the subsidy the deduction provides.  This is true of the Administration’s proposed HSA deduction as well.

Low-income families that do not earn enough to incur income tax liability would receive no tax benefit.  For families in the 10 percent or 15 percent tax brackets, the deduction would defray only 10 cents or 15 cents of each dollar they would have to spend to purchase a high-deductible health insurance policy in the individual market, generally too little to make insurance affordable for such families.  For individuals in the top tax brackets, however, the deduction would subsidize as much as 35 percent of the cost of high-deductible health insurance purchased in the individual market.

Leading Urban Institute Health Policy Researcher Testifies
to the Risks Posed by Health Savings Accounts and the Proposed Deduction

Linda Blumberg, a health policy expert at the Urban Institute, recently testified on the likely effects of Health Savings Accounts and the proposed deduction.* She made the following observations about HSAs.

Greatest Benefit to High-Income Individuals. “The tax subsidy [included in HSAs] is greatest for those in the highest marginal tax bracket and is of little or no value to those who do not owe income tax. Higher income individuals are also better able to cover the costs of a high deductible, should significant medical expenses be incurred.”

Adverse Selection. “By providing incentives for healthy individuals and groups to purchase HSAs with high deductible policies, insurance risk pools can be further segmented by health status. The average medical costs of those purchasing the new [high-deductible] plans will be substantially lower if the high risk population is left in more traditional comprehensive plans. The practical effect, however, is that the most vulnerable populations (the sick and the low-income) are left bearing a greater burden of their health expenses.”

Higher Administrative Costs. “Moving individuals into higher deductible policies actually increases the share of [insurance] premiums attributable to administrative costs. The administrative “load” charged by insurers is simply the total administrative costs divided by the total benefits paid.... Because many administrative costs are fixed, lowering the actuarial value of the benefits requires the insurers to increase the administrative load. Consequently a larger share of premiums paid for high deductible policies will be attributable to administrative charges than when comprehensive coverage is purchased.”
Blumberg also warned that the proposed deduction would have deleterious effects.

Further Weakening of Employer-Based Coverage. “This new proposal increases the incentive for individuals to purchase health insurance in the private non-group insurance market, as opposed to acquiring it through employers. Making the private non-group market more attractive may lead to a decline in the availability of coverage available through small firms.... The proposal would provide a non-group insurance product whose tax advantage is almost as great as that available in the group market and which is most attractive to those with high incomes and low-health care risk.... But as low cost purchasers leave the group market, the average cost of those staying in the group market will rise, making group insurance more difficult to afford for higher risk and low-income populations. In addition, since employers and key employees will be able to get tax breaks for their high-deductible health insurance even if they do not provide it to their employees, there will be even less incentive for them to take on the hassle, expense, and risk of offering insurance to their workers. The net result could be less insurance coverage among small businesses.” Blumberg also noted that “the tax subsidy [that the deduction provides] would be worth most to those who least need assistance.”

Blumberg concluded that “the federal funds necessary to fund this legislation could more effectively be redirected towards approaches designed to address the explicit problems facing small businesses or to expansion of eligibility in existing State Children’s Health Insurance Programs (S-CHIP) or Medicaid.”

* Linda J. Blumberg, Testimony before the Subcommittee on Workforce, Empowerment and Government Programs, U.S. House Committee on Small Business, March 18, 2004.

Assume, for example, that a moderate-income, healthy uninsured family of four living in the Washington D.C. metropolitan area is considering whether it can afford a high-deductible family plan available in the individual health insurance market that carries a $3,350 deductible.  Such a policy could carry an annual premium cost of approximately $4,000.[6]

If the family is in the 10-percent tax bracket, it would receive a tax benefit of $400 from the proposed deduction.  The family would have to pay $3,600 in premium costs itself, which likely would cause the high-deductible coverage to remain unaffordable.  (The premium costs are likely to be even higher for families that contain members who are older or in poorer health.  The individual market is largely unregulated, and insurers generally vary premiums based on age and health status.)

By contrast, a high-income family of four purchasing the same high-deductible policy would receive a premium subsidy of $1,400 (35 percent of the premium cost of $4,000).  This tax subsidy would be in addition to the tax benefits of nearly $1,200 that the high-income family already would be deriving from making annual tax-deductible contributions to its HSA equal to the deductible amount of $3,350.  This family’s combined HSA-related tax benefits thus would total nearly $2,600.  (The family’s overall HSA tax subsidy actually would be even greater than this, since earnings on the amounts in the HSA would accrue tax free.)

The high-income family in the top tax bracket thus would more than double its already substantial HSA-related tax benefits if the proposed deduction were enacted.  Yet such a family does not need large government subsidies to be able to afford health insurance.

High-income individuals are the people already most likely to take advantage of HSAs under current law; HSAs provide them with unique tax advantages.  Unlike other tax-advantaged savings accounts, HSAs allow both tax-deductible contributions and tax-free withdrawals, as long as the withdrawals are used to pay for out-of-pocket medical costs.  In addition, unlike with Individual Retirement Accounts, there are no income limits on who can participate in HSAs.  Providing an additional deduction to HSA participants for the premium costs of health insurance purchased in the individual market would make HSAs even more advantageous for high-income individuals.

Would Widespread Use of HSAs Reduce Overall Health Care Costs?

Proponents of HSAs argue that high-deductible policies would discourage unnecessary utilization of health care services by requiring individuals to bear a greater portion of the costs of their care. As a result, supporters argue HSAs would produce substantial reductions in overall health care spending over time.

It is unlikely, however, that HSAs would provide significant cost containment. As Henry Aaron, a Senior Fellow at the Brookings Institution and a leading expert in the areas of health care and tax policy, recently wrote in Tax Notes:

“The reason is that most medical spending occurs during high-cost episodes in which the total cost of care charged to patients greatly exceeds the limits of any plausible high-deductible plan. Stop-loss limits have become the hallmark of adequate health insurance. Their importance and attractiveness would increase under high-deductible insurance. Once patients enter the stop-loss range of their insurance, they would, by definition, be as free of financial discipline to attend to health care costs as they are under low-deductible insurance. The direct effects of high-deductible insurance on health care costs are therefore likely to be small.”*

It should also be noted that research indicates that higher cost-sharing discourages utilization of both necessary and unnecessary services among low-income individuals. If a medical condition or illness goes untreated because individuals are unable to pay for appropriate care out-of-pocket, this can eventually lead to greater use of expensive complex services like hospitalization. As a result, the high deductible insurance policies required under HSAs could result in increases in health care costs over time for such individuals.

* Henry Aaron, “HSAs — The ‘Sleeper’ in the Drug Bill,” Tax Notes, February 23, 2004.

But while the proposed deduction would be valuable for high-income individuals, it would be unlikely to have large effects in enabling uninsured people to gain coverage, since it would do little to make insurance affordable for most people who lack it.  Jonathan Gruber’s analysis indicates that only about 1.1 million of the eight million households that would use the deduction would previously have been uninsured.  Gruber also concludes that this gain of 1.1 million individuals with insurance would be more than offset by the loss of coverage for 1.4 million currently insured people as a result of actions that employers would take in response to the deduction. 

Also of note, the small number of low- and moderate-income people who might gain coverage through the deduction would generally be underinsured, due to the less comprehensive high-deductible policies they would have.  The more restrictive coverage that such policies provide can be problematic, especially for older or sicker workers.  A Commonwealth Fund study found that older individuals who enroll in the less comprehensive high-deductible health insurance plans commonly found in the individual market are twice as likely as older people with comprehensive employer-based coverage to fail to see a doctor when a medical problem develops or to skip medical tests or follow-up treatment.[7]


Deduction Could Induce Some Employers to Drop Coverage

The proposed deduction is likely to induce some employers to drop existing health insurance coverage or to decide not to offer coverage in the first place, on the grounds that their employees could use the tax benefits of the new deduction in tandem with those of HSAs to obtain coverage through a high-deductible plan in the individual market.  As noted, Gruber estimates that about 2.1 million people will lose employer-based coverage as a result of employers dropping coverage (or electing not to offer it in the first place) if the proposed deduction is enacted, with about 1.2 million of these people becoming uninsured.[8]  Gruber estimates that an additional 190,000 people would become uninsured because they would drop out of employer-based coverage after their employers scaled back employer contributions for health insurance premiums in response to the deduction.  With only about 1.1 million previously uninsured individuals gaining coverage through the deduction, according to Gruber's analysis, the net effect would be an increase in the number of uninsured individuals of about 350,000 (see Table 1).

The employer-dropping would likely occur primarily among small businesses.  Among firms with fewer than 200 workers, the costs of health insurance premiums rose by 15.5 percent between 2002 and 2003.  Due in part to premium increases and in part to financial pressures resulting from the economic slump, the percentage of firms with fewer than 200 workers that offer health coverage declined from 68 percent in 2000 to 65 percent in 2003.[9]  Among small firms that do offer coverage, nearly one-third provided less than a 50 percent subsidy for the cost of family coverage in 2003.  If the proposed deduction is created, its availability would provide an incentive for more small employers to drop or scale back coverage. 

In response to Gruber’s analysis as conducted for the Kaiser Family Foundation, some proponents of Administration health tax proposals have argued that employer dropping on this scale is unlikely to occur under the proposed HSA deduction because the availability of the deduction would not cause most employers to change their decision to offer coverage to their workers or to reduce their premium contributions.  Gruber’s analysis, however, fully takes this into account.  Gruber assumes that most employers would not drop coverage in response to the deduction.  His estimate that employers would no longer offer coverage to 2.1 million individuals indicates that only 1.2 percent of the total number of individuals currently obtaining coverage through employers, which amounts to about 175 million people, would be affected.  Considering the size of the employer-based health insurance system, Gruber’s estimates thus are conservative.

The deduction also could encourage a significant number of healthy, more affluent workers to switch from employer-based coverage to the individual market.  Healthy, higher-income individuals who work for a firm that offers comprehensive health insurance and requires employees to pay a significant share of the premium costs could decide to shift to the individual market to take advantage of the tax benefits that HSAs and the proposed deduction would provide.  Gruber estimates that the deduction would induce approximately 790,000 workers to switch voluntarily from employer-based coverage to the individual market.

Those who leave employer-based coverage for HSAs and the individual market would primarily be younger, healthier, more affluent workers.  Older and sicker workers would be unlikely to follow this path; the high-deductible policies that accompany HSAs would be risky for them.  They also would tend to face very high premium costs for insurance policies in the individual market, as a result of their age or less-healthy status.

If substantial numbers of healthy, affluent workers opt out of comprehensive employer-based coverage while less healthy workers remain in it, however, premiums for comprehensive employer-based coverage will necessarily rise, since those who remain in such coverage will constitute a pool that is sicker and more costly to insure.  The resulting increases in premiums for comprehensive insurance could, in turn, drive additional young and healthy workers to abandon employer-based coverage for the individual market, making the pool of workers left in comprehensive employer-based coverage yet more expensive to insure.  This process, known as “adverse selection,” could ultimately lead significant numbers of employers to drop coverage altogether or to drop comprehensive coverage and replace it with high-deductible insurance and HSAs.



The benefits of the proposed deduction would be concentrated among higher-income individuals who already are insured.  Moreover, Professor Gruber’s analysis finds that the deduction likely would result in a net increase in the ranks of the uninsured.

Four likely effects of the deduction stand out: 1) it would further enlarge the tax benefits that HSAs offer to high-income individuals who purchase insurance in the individual market, despite the fact that the tax benefits HSAs offer such people are extremely generous under current law and these people do not need additional government subsidies to be able to afford insurance; 2) the deduction would induce more employers to drop coverage or not to offer it in the first place; 3) as a result, it would likely increase the ranks of the uninsured; and 4) it would increase budget deficits even as it made the problem of the uninsured worse.

The Center on Budget and Policy Priorities is grateful to the
Nathan Cummings Foundation for its support of this report.

End Notes:

[1] For an analysis of the Health Savings Accounts, see Robert Greenstein and Edwin Park, "Health Savings Accounts in Final Medicare Conference Agreement Pose Threats Both to Long-Term Fiscal Policy and to the Employer-Based Health Insurance System," Center on Budget and Policy Priorities, revised December 1, 2003.

[2] There are limited exceptions to the rule that tax-free withdrawals from HSAs may not be used to pay premiums for health insurance.  Tax-free withdrawals from HSAs may be used to pay for health insurance premiums under COBRA or while an individual is unemployed, for long-term care insurance premiums, or for premiums for private supplemental coverage under Medicare.

[3] Professor Gruber assumes a somewhat lower participation rate in the proposed deduction and as a result, he estimates that the deduction would cost less than the Administration and the Joint Committee on Taxation project.  Gruber assumes that in the first year when the proposed deduction is fully implemented, it would cost $1.4 billion.  This is lower than the nearly $1.8 billion that the Administration and JCT assume for tax year 2006, when the proposed deduction appears to be fully implemented.  According to Gruber, this cost estimate difference is primarily caused by higher participation rates for the deduction assumed under the Administration and JCT estimates.   

[4] In an analysis issued in 1998, the General Accounting Office found that more than 90 percent of the uninsured had no tax liability or were in the 15 percent tax bracket.  General Accounting Office, Letter to the Honorable Daniel Patrick Moynihan, June 10, 1998.  The 10 percent tax bracket, which was carved out of the 15 percent bracket by the 2001 tax legislation, did not yet exist. 

[5] Communication with Professor Jonathan Gruber, March 12, 2004.  See also Kaiser Family Foundation, “Coverage and Cost Impacts of the President’s Health Insurance Tax Credit and Tax Deduction Proposals,” March 2004.  The analysis that Professor Gruber conducted for the Kaiser Family Foundation report determined the simultaneous coverage effects of both the deduction and the Administration’s proposal to provide a refundable tax credit for the purchase of health insurance in the individual market.  The analysis cited here and communicated to CBPP shows the effects of the deduction separately.

[6] An illustrative high deductible health insurance policy being marketed by the Golden Rule Insurance Company as a plan meeting HSA requirements was located on for both the Northern Virginia and Maryland suburbs of Washington D.C.  The policy is a high-deductible health insurance plan available in the individual market for a hypothetical non-smoking healthy family of four (two 35 year-old adults with two 8 year-old children).  This high-deductible plan includes a deductible of $3,350, zero percent cost-sharing after the deductible is met, and a maximum out-of-pocket limit of $3,350.  The policy does not include maternity coverage or well-baby care and has a lifetime limit of $3,000 in mental health benefits per family member.

This premium quote assumes the family is in excellent health.  In most states, the premium is likely to be adjusted higher, in some cases significantly higher, to reflect the health status of family members if any are in less-than-excellent health.  In some cases, families with members in poor health would not be able to purchase such a high-deductible health insurance plan at any price.

[7] Elisabeth Simantov, Cathy Schoen and Stephanie Bruegman, “Market Failure?  Individual Insurance Markets for Older Americans,” Health Affairs, July/August 2001.

[8] Communication with Gruber.

[9] Kaiser Family Foundation and Health Research and Educational Trust, “Employer Health Benefits: 2003 Annual Survey,” September 2003.