September 13, 2004
ASSESSING THE HSA COALITION’S COVERAGE ESTIMATES FOR THE
ADMINISTRATION’S PROPOSED HSA TAX DEDUCTION
by Edwin Park and Robert Greenstein
PDF of analysis
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.
May 10, 2004, the Center on Budget and Policy Priorities issued an analysis of a proposal in the Administration’s 2005 budget to provide a tax deduction for the premium costs of high-deductible health insurance purchased in the individual health insurance market in conjunction with HealthSavings Accounts (HSAs). The analysis presented the findings of M.I.T. economist Jonathan Gruber, who is widely regarded as one of the nation’s leading health economists. Gruber’s analysis found that because the deduction would lead some employers to drop coverage, it would likely cause the ranks of the uninsured to increase. Gruber estimated that the increase in the ranks of the uninsured would be nearly 350,000.
Two weeks later the HSA Coalition, which promotes and lobbies for HSAs and includes firms and trade associations that profit from HSAs and conservative health policy groups, among others, issued a rebuttal challenging Professor Gruber’s analysis. The HSA Coalition’s paper contended that rather than increasing the number of uninsured, the proposed tax deduction would reduce the ranks of the uninsured by between 1.2 million and 4.3 million people.
These differences are not simply a result of two alternative but reasonable ways of analyzing this matter. As is the case with work conducted by Professor Gruber, who enjoys a strong reputation in the economics profession in general and in health economics in particular, the Gruber analysis represents careful, rigorous work. The same cannot be said of the HSA Coalition paper. As explained below, the HSA Coalition analysis is severely flawed. Because of fundamental errors of analysis and misuse of data, its conclusions must be regarded as invalid.
The Gruber Analysis
The Gruber analysis, originally conducted for the Kaiser Family Foundation, examines two effects of the proposed tax deduction. One of these effects would reduce the number of uninsured, while the other would increase the number of uninsured. Under the proposal, a tax deduction would be provided for the premium costs of high-deductible health insurance that an individual purchases in the individual health insurance market in conjunction with a
HealthSavings Account. The deduction would not apply to the employee share of the premium costs of health insurance obtained through an employer.
As a result of the tax deduction that would be provided, the proposal would reduce the effective cost of high-deductible insurance purchased in the individual market. It would thereby make such insurance somewhat more affordable, which should increase the number of people purchasing it and reduce the number of uninsured.
The deduction, however, also would lessen the incentive for employers to provide health insurance coverage. Employers deciding whether to offer coverage and to incur the costs of doing so would know that if they did not provide coverage, their workers could receive two tax benefits if they purchased coverage in the individual market — a deduction for the premium costs of high-deductible insurance as well as the HSA tax breaks already in law, which can reduce the cost of deductibles, co-payments and other out-of-pocket costs. In combination, these tax breaks would be quite substantial for people in higher tax brackets. As a result of this change in incentives, fewer employers would be expected to offer coverage. Since some people who would lose access to employer-based coverage would not secure coverage in the individual insurance market, this effect of the proposed deduction would increase the number of people who are uninsured.
Gruber examined both of these effects of the proposed deduction, using a rigorous health insurance microsimulation model. This is the type of model that analysts at the Congressional Budget Office and the Treasury Department also use.
Gruber found that nearly 1.1 million people who otherwise would be uninsured would secure insurance as a result of the deduction’s effect in reducing the effective cost of high-deductible coverage. This number is relatively low largely because a substantial majority of the uninsured does not earn enough to owe income tax or are in the 10 percent or 15 percent tax brackets. For these people, the proposed deduction either would not reduce the cost of health insurance or would reduce the cost by only 10 percent or 15 percent.
Gruber found that the effects of the deduction in inducing employers to drop coverage would be limited as well. Fewer than one percent of those insured through an employer would become uninsured. But because so many people have employer-based coverage, even a loss of coverage for one percent of them produces a sizable number. Gruber’s analysis found that 1.4 million people would become uninsured as a result of actions taken by employers in response to the deduction.
With nearly 1.1 million uninsured people gaining coverage and 1.4 million previously insured people losing coverage, the net effect of the deduction would be an increase in the number of uninsured of nearly 350,000.
Evaluating the HSA Coalition’s Claims
The HSA Coalition disputed Gruber’s analysis and produced it own estimate — that the proposed tax deduction would reduce the number of uninsured by between 1.2 million and 4.3 million. The HSA Coalition’s work on this matter, however, is not credible.
What Proportion of Those Who Use the Deduction Would Previously Have Been Insured?
A significant number of people already purchase high-deductible insurance in the individual health insurance market. Analysts agree that this number will rise in coming years as use of the generous HSA tax breaks enacted as part of last year’s Medicare drug legislation becomes more widespread. People who will purchase high-deductible insurance regardless of whether the proposed deduction is enacted will, of course, make use of the deduction if it becomes law. But such people would not otherwise be uninsured.
Gruber’s analysis finds that these individuals make up the great bulk of those who would use the new deduction. (As noted above, for most people who otherwise would be uninsured because they cannot afford coverage, the proposed deduction would have only modest effects in making insurance affordable, since it would reduce the premium costs of insurance for them by no more than 15 percent.) Gruber estimates that nearly eight million people would use the proposed tax deduction but that only 13.4 percent of them — just under 1.1 million people — would be people who otherwise would be uninsured.
The HSA Coalition disputes these figures. It says preliminary data indicate that 32.8 percent to 42 percent of HSA purchasers in the individual market are previously uninsured. It then applies these percentages to the projected number of people whom Gruber estimates would use the tax deduction (7.98 million) and contends that 2.6 million to 3.4 million individuals who otherwise would be uninsured would gain coverage. (The low-end of the Coalition’s estimate that the deduction would reduce the ranks of the uninsured by 1.2 million to 4.3 million is obtained by applying Gruber’s estimate that 1.4 million currently uninsured people would lose coverage due to actions by employers and subtracting this amount from the 2.6 million figure just cited.)
What the HSA Coalition has done here, however, is not valid. The proportion of uninsured individuals who purchase high-deductible insurance and HSAs today — in the absence of the proposed tax deduction — is not relevant to determining the effects of the deduction. The relevant question is how many uninsured people in addition to those who already purchase individual-market high-deductible policies in conjunction with HSAs would be induced to purchase coverage by the new deduction. By applying the 32.8 percent and 42 percent figures to the number of people assumed to take the proposed deduction — and asserting that the results represent the number of people who would gain coverage as a result of the deduction — the HSA Coalition has committed a fatal error that undermines its analysis.
Under his model, Gruber finds that the increase in the number of otherwise uninsured individuals who would purchase high-deductible policies as a result of the proposed deduction is modest. The bulk of those who would use the deduction — 86.6 percent — would be people who would purchase coverage anyway. The HSA Coalition provides no evidence — and no credible analysis — to use as a basis for estimating the increase in coverage that would result from the deduction. The Coalition also failed to use a microsimulation model to estimate the coverage increase; as noted above, such a model is the standard way that such estimates are made.
The Coalition’s report is marred both here and elsewhere in the paper by a conflation of the potential effect of the HSA tax breaks already in law with the effect of the proposed deduction. The Gruber analysis (and the Center on Budget and Policy Priorities paper that presented it) examine the effects of the proposed deduction, not the effects of the HSA provisions already in law. (Note: claims that 32.8 percent to 42 percent of those who will purchase HSAs under current law would otherwise be uninsured, which some HSA proponents have made elsewhere, also are invalid; those claims, too, rest on misuse of data. That matter is beyond the scope of this paper; for a discussion of it, see
Edwin Parkand Robert Greenstein, “Initial Data on Individual Market Employment Fail to Dispel Concerns About Health Savings Accounts,” Center on Budget and Policy Priorities, September 13, 2004.)
As a result of this critical mistake, the estimates in the HSA Coalition report lack foundation and must be disregarded. Although this is the report’s principal flaw, the report suffers from other weaknesses as well.
Estimates of the Number of People Who Would Make Use of the Deduction
The HSA Coalition report contends that Gruber underestimated overall participation in the HSA tax deduction. Gruber estimates that in the first year the tax deduction was fully implemented, it would cost $1.4 billion. This is less than the estimated cost of nearly $1.8 billion that the Administration and the Joint Committee on Taxation assume for tax year 2006, when they expect the proposed deduction to be fully implemented. Gruber believes this difference is likely the result of the fact that under this model, a somewhat lower total number of participants are projected to take up the proposed deduction than the Administration and Joint Committee on Taxation apparently assume. The HSA Coalition argues that Gruber’s estimate that 7.98 million tax filers will take the deduction should be increased by 28.6 percent to approximate the Administration’s and the Joint Committee on Taxation’s higher spending estimates.
The HSA Coalition analysis proceeds to increase Gruber’s estimate of 7.98 million participants by 28.6 percent to arrive at an estimate that 10.3 million people would use the deduction. The Coalition then applies the mistaken 32.8 percent and 42 percent factors discussed above to produce an estimate that 3.4 million to 4.3 million people who otherwise would be uninsured would gain coverage as a result of the deduction. In this part of its computations, the Coalition retains Gruber’s finding that there would be a 1.4 million reduction in coverage due to actions by employers, producing an estimate of the net coverage gain of between 2.0 million and 2.9 million.
In addition to mistakenly applying the 32.8 percent and 42 percent factors, this computation has another weakness. If the Coalition wished to assume higher take-up of the proposed tax deduction on the basis of the Administration and Joint Committee on Taxation cost estimates, the Coalition should also have adjusted upward Gruber’s estimate that 1.4 million workers who currently have employer-based coverage would become uninsured. Greater use of the tax deduction to purchase coverage in the individual market likely implies greater sensitivity in employer behavior in response to the tax deduction; the higher estimate suggests that more employers would drop coverage (or reduce their premium contributions) than Gruber assumed, thereby shifting more workers to the individual health insurance market. That, in turn, would mean that more currently covered workers would end up without coverage than Gruber estimated, offsetting at least some of the gains in coverage that otherwise would result from the increase in the total number of people using the deduction.
Would Some People Lose Coverage As a Result of Employer Actions?
The HSA Coalition report then goes further, disputing Gruber’s estimate that 1.4 million currently insured workers would become uninsured as a result of employers dropping health insurance coverage or reducing premium contributions. The HSA Coalition report challenges Gruber’s finding that employers would drop coverage for an estimated 2.1 million workers, with 1.2 million becoming uninsured. (Gruber also estimated that 190,000 workers would become uninsured as a result of employer actions to reduce premium contributions, in response to the new deduction, bringing to 1.4 million the total number of employees who would lose coverage.)
The Coalition report argues that employers already have the ability to drop coverage for their workers and are doing so, and that more employers will offer coverage through HSAs because of the lower premium costs of high-deductible health insurance, as compared to the costs of the comprehensive policies that employers typically offer. The Coalition produces a further set of computations under which the deduction is assumed to have zero effect on employer-based coverage and all 1.4 million workers whom Gruber estimates would lose coverage are added back. This further increases the Coalition’s estimate of the net coverage gain under the proposed deduction — to between 3.4 million and 4.3 million. This 4.3 million figure is the figure used as the high end of the Coalition’s range of the number of people who would gain coverage because of the deduction.
In this part of its report, the HSA Coalition rejects the concept that employer behavior would be sensitive to the availability of the HSA tax deduction in the individual market. This assumption by the Coalition is difficult to justify; it conflicts with basic principles of health economics. As Professor Gruber explained earlier this year:
“If you think about part of why employers offer health insurance, it’s because their employees, while taxed on their wages, are not taxed on their health insurance expenditures. As a result, this [is] what we call the employer exclusion to health insurance [and] is what makes it attractive for employers to offer health insurance. By allowing employees to deduct non-group insurance from their taxes as well, while providing some equity in the treatment of group and non-group insurance, what this does [the proposed deduction] is it reduces the incentive for employers to offer. Many employers who [do not] really want to be in this business in the first place are going to say, look, if there’s no longer a tax advantage to me offering health insurance, I’m not going to offer it anymore, I’ll let my employees who want to go take this tax deduction [and buy a] HSA.”
Some other employers may continue to offer coverage but decide to reduce the size of their premium contributions because the value of the employer tax exclusion has been reduced due to the new deduction.
Gruber’s assumption about employer response to the proposed tax deduction is consistent with the prevailing view of health economists. Other health economists also expect employers to respond in such a manner to the availability of deductions and/or tax credits for the purchase of health insurance in the individual market. For example, the Administration’s own Council of Economic Advisors has stated that tax incentives for the purchase of health insurance in the individual market could have an adverse effect on employer-based coverage. There may be disagreement as to the extent to which employers will drop coverage or reduce employee premium contributions (the effects of such actions could be larger or smaller than Gruber has estimated), but assuming there would be no change whatsoever in employer behavior is not reasonable.
It also should be noted that Gruber assumes the vast majority of 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 175 million individuals currently obtaining coverage through employers would be affected. Considering the size of the employer-based health insurance system, Gruber’s estimates are conservative.
Confusion About Whether the Deduction Would
Apply to Employer-Based Coverage
The HSA Coalition report appears to imply mistakenly that under the proposed deduction, employees could take a tax deduction for their share of the premium costs of employer-based high-deductible health insurance. In arguing that the proposed deduction is unlikely to cause employers to scale back premium contributions for their workers, the HSA Coalition report states that “the President’s HSA tax deduction is designed to make a popularly sought benefit less expensive and more flexible for both employers and employees.” The report implies that the deduction would reduce the cost of employer-based coverage.
This is not correct. Individuals with employer-based insurance would be prohibited from taking the deduction. The deduction could be used only in connection with the purchase of health insurance in the individual market. As it does elsewhere its report, the HSA Coalition appears to conflate here the effects of the proposed deduction with the effects of the HSA tax breaks already in law.
The HSA Coalition report also argues that more employers will offer health insurance to their workers through HSAs generally because of the lower premium costs of high-deductible insurance and that this factor should have been taken into account in Gruber’s estimates. This criticism is off the mark. Gruber’s analysis does not address whether the existence of the HSA tax breaks already in law will result in employers covering more of the uninsured; his analysis examines the effects of the proposed tax deduction. That deduction could be used only for the purchase of high-deductible health insurance in the individual market. The proposed deduction would not apply to employer-based coverage and would not lower the premium costs of such coverage. (It also should be noted that numerous health policy experts have concluded that the widespread availability of HSAs under the HSA tax breaks already in law is likely to weaken, rather than strengthen, employer-based coverage by leading to “adverse selection.”)
Other Issues with the HSA Coalition Report
The HSA Coalition report raises two additional issues that merit discussion.
- The report claims that the Gruber analysis on the effects of the proposed deduction fails to take into account the availability of the Administration’s proposal for a refundable tax credit for the purchase of health insurance in the individual market. The value of a tax deduction rises with one’s tax bracket. As a result, low- and moderate-income individuals who make up a large proportion of the uninsured would derive little or no tax benefit from the proposed deduction. (Workers who do not earn enough to owe income tax would receive no benefit from the deduction, while the deduction would reduce the cost of health insurance policies for the bulk of middle-class taxpayers in the 10 percent or 15 percent tax brackets by only 10 percent or 15 percent, which is likely to be too little in most cases to make health insurance affordable.) As a result, the deduction would induce only a small number of uninsured individuals to purchase health insurance in the individual market. The HSA Coalition report acknowledges that low- and moderate-income uninsured workers may derive little benefit from the deduction but argues that these workers could turn to the President’s proposed refundable tax credit for the purchase of health insurance in the individual market and that the effects of the tax credit should have been taken into account in the Gruber estimates presented in the Center on Budget and Policy Priorities paper.
The tax credit proposal is outside the scope of the Gruber analysis presented in the Center paper. The paper examined the effect of the proposed deduction, not of a broader set of proposals. This is a reasonable approach to take; when the current HSA tax breaks were enacted as part of the Medicare prescription drug bill last fall, the Administration’s proposed tax credit was not included with them even though both the HSA tax break and the tax credit were part of the President’s budget last year. Moreover, while the President’s budget this year gives lip service to the proposed tax credit, the budget includes no money for the primary component (the refundable component) of the credit and indicates that the Administration favors enacting the credit only if Congress can come up with tens of billions of dollars in offsetting savings to cover the credit’s cost. No similar stipulation is placed on the proposal in the Administration’s budget for the HSA tax deduction. Enactment of the proposed deduction clearly is not dependent upon or tied to enactment of the tax credit.
In a paper for the Kaiser Family Foundation, Gruber did estimate the combined coverage effects of the tax credit and the HSA deduction. He found that together, the proposals would produce a net coverage gain of about 1.3 million people, but that the tax credit by itself would produce a net coverage gain of 1.8 million. That Gruber analysis, as well, thus showed that the proposed HSA tax deduction would itself have a negative effect on coverage.
- The HSA Coalition report argues that in conducting his joint analysis of the deduction and the tax credit for the Kaiser Family Foundation, Gruber failed to assume that moderate-income individuals could take both the tax deduction and a tax credit in the same year, making insurance more affordable. The HSA Coalition report questions the Kaiser joint estimate on the grounds that it fails to take into account the “synergy” whereby individuals could take both the deduction and tax credit in the same year. Had Gruber taken that into account, the HSA Coalition report argues, the overall coverage effects would be higher.
This line of argument is flawed. It is unlikely that a moderate-income family would be able to claim both the proposed refundable tax credit and the proposed premium deduction in the same tax year. The
tax code generally does not permit “double-dipping;” that is, a taxpayer cannot claim both a credit and a deduction for the same expense. While the Administration does not specifically address in its budget the issue of how the refundable tax credit and the proposed deduction for the costs of high-deductible insurance in conjunction with a HSA would interact, the Administration makes clear that people using the tax credit would not be allowed to deduct contributions to a HSA. Similarly, legislation introduced in the House of Representatives that is based on the Administration’s proposed HSA deduction proposal prohibits individuals taking the deduction from claiming the premium costs of insurance as a deductible expense under either the existing tax break for the self-employed or the tax break for people with large health expenses. Despite HSA Coalition claims to the contrary, it is unlikely that under the Administration’s proposals, a family would be able to claim both the tax credit and the deduction in the same year. U.S.
The HSA Coalition report attempts to challenge M.I.T. economist Jonathan Gruber’s estimates of the coverage impact of the Administration’s HSA tax deduction proposal. As discussed above, the techniques that the Coalition uses to critique Professor Gruber’s estimates are severely flawed and rest on assumptions and computations that are not valid. Gruber’s estimates, in contrast to those the Coalition has produced, are based on rigorous economic analysis and the same type of microsimulation modeling that government analysts at the Treasury Department and the Congressional Budget Office use.
The Center on Budget and Policy Priorities is grateful to the Nathan Cummings Foundation for its support of this report.
Edwin Parkand Robert Greenstein, “Proposal for New HSA Tax Deduction Found Likely to Increase the Ranks of the Uninsured,” Center on Budget and Policy Priorities, revised April 10, 2004.
 Park and Greenstein, op cit.
 Richard Nadler and Dan Perrin, “The Center on Budget and Policy Priorities’ Study on HSA Premium Tax Deduction Misses the Point,” The HSA Coalition,
May 25, 2004. According to their website, the HSA Coalition is a coalition of organizations that support Health Savings Accounts.
 In an analysis issued in 1998, the General Accounting Office (now known as the Government Accountability 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. Similarly, in an analysis issued in 2000, Professor Gruber determined that of the uninsured who had positive tax liabilities, 90 percent were in the 15 percent tax bracket. Jonathan Gruber, “Tax Subsidies for Health Insurance: Evaluating the Costs and Benefits,” National Bureau of Economic Research, February 2000. The 10 percent tax bracket, which was carved out of the 15 percent tax bracket by the 2001 tax legislation, did not yet exist at the time of either study.
 For a separate analysis assessing this data, see
Edwin Parkand Robert Greenstein, “Initial Data on Individual Market Enrollment Do Not Dispel Concerns about HealthSavings Accounts,” Center on Budget and Policy Priorities, September 13, 2004.
 Transcript of Center on Budget and Policy Priorities May 25, 2004 media briefing call. A transcript is available from CBPP upon request.
 See, for example, a discussion of the health economics literature in Leonard E. Burman, Cori E. Uccello, Laura L. Wheaton and Deborah Kobes, “Tax Incentives for Health Insurance,” Urban-Brookings Tax Policy Center, May 2003.
 Council of Economic Advisors, “Health Insurance Tax Credits,”
February 13, 2002.
 U.S. Department of Treasury, p.25.
 See Emmett B. Keeler, et. al., “Can Medical Savings Accounts for the Nonelderly Reduce
HealthCare Costs?” Journal of the American Medical Association, June 5, 1996, p. 1666-71; Len M. Nichols, et. al., “Tax-Preferred Medical Savings Accounts and Catastrophic HealthInsurance Plans: A Numerical Analysis of Winners and Losers,” The Urban Institute, April 1996; of Actuaries, “Medical Savings Accounts: Cost Implications and Design Issues,” May 1995; Daniel Zabinski et. al., “Medical Savings Accounts: Microsimulation Results from a Model with Adverse Selection,” Journal of Health Economics, April 1999, p.195-218; and Gail Shearer, “The Health Care Divide: Unfair Financial Burdens,” Consumers Union, American Academy August 10, 2000(relying on Lewin Group estimates).
 For an analysis of the Administration’s tax credit proposal, which includes coverage estimates of the tax credit conducted by Professor Gruber, see
Edwin Park, “Administration’s Proposed Tax Credit for the Purchase of Health Insurance Could Weaken Employer-Based Health Insurance,” Center on Budget and Policy Priorities, revised April 6, 2004.
 Kaiser Family Foundation, “Coverage and Cost Impacts of the President’s Health Insurance Tax Credit and Tax Deduction Proposals,” March 2004.
 U.S. Department of Treasury, “General Explanations of the Administration’s Fiscal Year 2005 Revenue Proposals,”
February 2, 2004, p.22.
 H.R. 3901 sponsored by Rep. Crane (R-IL). Under current law, self-employed individuals can deduct 100 percent of the premium costs of their health insurance and individuals can deduct premium costs (and other out-of-pocket medical expenses) in excess of 7.5 percent of their Adjusted Gross Income (AGI).