NCHC Writer
Default Image

Steven D. Findlay, M.P.H.,
Joel E. Miller, M.S.Ed.,Introduction

In 1998, 44.3 Americans had no health insurance (U.S. Census Bureau, 1999), up 2.5 million from 1996 and 10 million over the past decade. About 30 percent of the population – 81 million people – can expect (in the absence of policy changes) to experience a gap in their coverage lasting at least one month over the next three years. One of the chief culprits causing the erosion in health insurance coverage is that employment-based coverage is declining. In 1987, 69.2 percent of the non-elderly population had employment-based coverage. By 1997, that proportion had declined to 64.2 percent (EBRI, 1998).

Stated simply, health insurance has become too expensive for many businesses and unaffordable for many low-income and middle-income working Americans.

After several years of neglect, this mounting “health insurance gap” has once again emerged as a hot political issue. Several Presidential candidates and lawmakers have put forth proposals to expand health insurance coverage in recent months. And many observers believe that the year 2000 will see a full-fledged renewed debate on whether America can and should have universal health insurance coverage, and, if so, what the best ways to achieve it may be.

Several key forces are driving this renewed debate. The most important are these:

  • Despite unprecedented good economic times and several targeted attempts by the federal government and states to expand coverage, the number of people without health insurance continues to rise;
  • After several years of slowdown in health care inflation, costs are rising sharply again, including the price of health insurance; and
  • Aging baby boomers and middle income Americans are increasingly anxious about the stability of the health care system and their insurance coverage over the long term.

Almost all the proposals put forth in recent months to deal with this problem – including those of Vice President Albert Gore and former Senator Bill Bradley – contain elements that use the federal income tax system to subsidize the voluntary purchase of health insurance. It is not a new idea. Proposals dating back 15 years have included – and in many cases relied on – tax subsidies to expand health insurance coverage. These ideas were swept aside in the period 1990-1994 as the nation, led by President Clinton, debated reforms that relied on mandatory measures (either a requirement that employers provide health insurance to workers or that individuals buy insurance) to achieve universal coverage. After the failure of that effort in 1994, lawmakers and policy makers turned their attention to other, narrower approaches. Most prominent among those is the Children’s Health Insurance Program (CHIP), enacted in 1997. CHIP, run by states, has enrolled about one million children to date, according to data from the Health Care Financing Administration. In part, the limited potential impact of such programs (as currently financial and structured) has led advocates and lawmakers to set their sights higher and on other means to expand coverage.

The prevailing wisdom today is that a combination of expansions of both public health insurance (Medicaid and Medicare) and voluntary private coverage (through tax subsidies to individuals and perhaps employers) is the most politically palatable mean to increase coverage in the short run. But critics of this approach assert that voluntary measures alone are very unlikely to achieve coverage for all Americans.

Much has been written over the years of the origin of the current tax treatment of health insurance. It is beyond the scope of this paper to reprise that history in detail. The following synopsis gives the essential background:

Prior to World War II, few Americans had health insurance. During the war, employment-based coverage began to expand. This was in large part because wages were frozen during the war and employers sought ways to attract and retain workers. Offering health insurance became a popular “benefit,” widely supported by labor unions. But in addition, the federal government gave a huge boost to this new benefit. In 1943, the War Labor Board ruled that health insurance was not subject to taxation if supplied through an employer. Specifically, the cost of health insurance was exempted from taxation as income to employees. (It was already permitted as a business expense.) The IRS put this into practice immediately, and in 1954 made it official with a formal detailed ruling and guidelines. Since then, the IRS has extended the exemption and Congress has enacted laws that amend it.

Current lawToday, the basic federal tax exemptions and exclusions for health insurance and health care costs are these:

  • The portion of health insurance costs paid by the employer for all their covered workers and retirees (and dependents) is a fully deductible business expense. Thus, employers subtract all their health insurance costs from their revenue in calculating taxable revenue. Companies do this for all allowable expenses, including wages. The policy on health insurance in this regard is not unusual. And no proposal today suggests altering it. The deduction works this way: Say a company has 200 workers, a health insurance cost of $500,000 a year and $10 million in revenue. It would deduct the $500,000 from the $10 million (along with many other business expenses, of course). This means that the company does incur a cost for health insurance; it’s far from “free,” as some people believe. If the company in this example were taxed at a 15 percent rate, for instance, the tax exclusion for health benefits would save it $75,000. The company would still have paid $425,000 for health insurance for its workers. It is worth noting that businesses pass their health insurance costs along to both workers and customers in the form of lower wages to workers and higher prices for products. The cost is built-in. But some companies also choose to take lower profits in order to subsidize health insurance for workers.
  • The portion of health insurance costs paid by the employer is not counted as part of employees’ taxable income for the purposes of federal income tax, Social Security and Medicare taxes. If a worker’s annual salary is $30,000, for example, and the company pays $3,500 for their family health insurance, the employee’s tax is not computed on an income of $33,500. This is the generous and unusual tax break that most Americans are not aware they have. (And it is a central focus of the accompanying study by The Lewin Group.) People essentially get this workplace benefit tax-free whereas many other forms of compensation are taxed. (Stock options, company cars, etc.) Money put aside in pension plans (such as 401(k) plans) is tax sheltered until it is drawn on in retirement.
  • The employee share of health insurance premiums is deducted from wage and salary income – and thus not taxed – at employers who sponsor so-called “cafeteria plans.” In the example above, if the $30,000 a year worker pays $1,000 a year as their share of the health insurance premium, their taxable income would be reduced to $29,000. In a cafeteria plan, employees can choose from a range of benefits, including health insurance, and pay their share of the cost on this “pretax” basis. About half of workers in small firms and 75 percent of workers in medium and large firms get their health insurance through such cafeteria plans.
  • Employee “out-of-pocket” expenses for medical care not covered by their insurance plan may be excluded from income – up to a certain preset dollar limit each year – in employer-sponsored “flexible spending accounts” (FSAs). Again, in the example above, if the $30,000 a year employee knows he or she is going to have about $2,000 in direct medical costs in a given year (that won’t be covered by their health plan — infertility treatments, for instance), they can elect to have this deducted from their taxable income. Their taxable income thus becomes $27,000. (Companies that sponsor FSAs put the money aside and employees must submit their medical bills to claim it. Money unspent at the end of the year can not be recouped. Some experts argue this promotes unnecessary health spending.)
  • Self-employed persons may deduct a portion of the amount they pay for health insurance. In 1999 that portion is 60 percent. It is currently scheduled to rise to 100% by 2003. Both chambers of Congress this year passed legislation that would accelerate the 100% deduction in 2001, and expand it to all people who do not have access to employer-sponsored coverage. What kind of savings does this tax break generate? A self employed person with an income of $60,000 who pays $5,000 for a family health insurance policy and whose tax rate is 28% would save $840 in 1999 – 16.8% of their health insurance cost. In 2003 (or 2001 if the pending legislation passes), the tax break would be worth $1,400 if the health insurance plan still cost $5,000.
  • Direct medical expenses in excess of 7.5 percent of adjusted gross income are tax-deductible. Again, take the $30,000 a year employee. Let’s say he or she has an adjusted gross income, after normal deductions, of $25,000. He or she would be able to deduct only medical expenses that exceed $1,875 in that taxable year. Thus, if a person without an FSA underwent a procedure costing $10,000 that was not covered by health insurance, he or she would be able to deduct $8,125 of the cost. Their taxable income would then decrease to $16,875.
  • Money contributed to a “medical savings account” (MSA) is excludable from taxable income in the same way as money contributed to an individual retirement account (IRA). Congress approved such MSA accounts in 1996. There are many rules surrounding them. Generally, only people who buy health insurance with a very high up front deductible ($1,500 and up) will find it practical to set up an MSA. Between 50,000 and 100,000 Americans have set up such accounts so far.

As alluded to above, these tax exclusions and deductions have promoted the growth of private employer-sponsored health insurance in the U.S. since World War II. Specifically, the policy that permits employees to get health insurance tax-free has been a major inducement to employers to offer this benefit. Almost all (85%, an estimated $106 billion in 2000) of the tax revenue the federal government gives up as a result of the above exclusions derives from the income tax, Social Security and Medicare exclusion. But this huge benefit has several built-in inequities and has produced adverse side effects:

  • Because of their structure, the exclusions benefit upper-income Americans to a larger degree than lower-income Americans. That’s because higher income people are more likely to have employer-sponsored health insurance, richer benefits, and by definition they are in higher income brackets. A person with an income of $80,000 who has coverage for which the company pays $3,500 and who is in a 28% income bracket gets a tax break of $980. If the person is in a cafeteria plan and has $1,000 deducted for his or her share of the premium, they save another $280. Total: $1260. By comparison, a worker with a $30,000 income in a 15% tax bracket with the same insurance coverage gets a tax break of $675 ($525 plus $150).
  • People who buy coverage on their own do not get the employee tax break.
  • The tax subsidy encourages the purchase of richer benefits by both companies and employees. That in turn has led to, and sustained, higher health care spending. Some argue that it has also led to an inefficient allocation of a scarce resource, dollars to be spent on health care.
  • The tax subsidy by its very nature supports an employment based health insurance system in the United States. That system results in a significant amount of lost coverage when people lose jobs, become temporarily unemployed, shift to a job where insurance is not offered, or are self-employed but can not afford health insurance.

These problems have helped spur the move to consider changes in the tax treatment of health insurance – both to address the inequities of the current system and as a mechanism for expanding coverage to uninsured Americans. A range of proposals exist, including a host of new ones in Congress and from presidential candidates. Several of these are examined in depth by the accompanying Lewin study. What follows is a very basic explanation of the kind of changes being contemplated and some of the issues that surround them.

Expanding the income tax deduction for health insurance

Both the Senate and the House have passed measures this year that would allow the self employed and people buying health insurance on their own to deduct 100% of the cost starting in 2001. This is a straightforward measure and builds entirely on current policy. It carries no serious administrative complexities. However, because of the limitations of the actual dollar benefit, it would not induce a significant number of low-income uninsured people to buy coverage. It benefits primarily people who have coverage now. In addition, the benefits of expanding the deductibility of health insurance accrue chiefly to those who pay taxes at a high rate, and those benefits increase as income increases.

Tax credits

In contrast to a tax deduction, a tax credit is a reduction in the amount of tax owed. For example, say a person with a taxable income of $30,000 owes $4,500 in federal income tax. If they had a tax credit of $500, they would owe $4,000. Simple? Yes, for people who actually owe taxes and if the credit is the same for everyone. But not simple at all for people who do not owe taxes – which includes a sizable number of low income and uninsured Americans – or if the credit is different for different income groups. Tax credits are also simpler if administered only for people buying health insurance on their own, and far more complicated if administered through employers for people getting coverage on the job.

The problem of people not owing taxes – and thus unable to reduce the amount they owe – is addressed by what advocates of this approach call a “refundable” tax credit. Put simply, this means that a person not owing any federal income tax would be able to send in a return and get a “refund” – in truth, a subsidy – for part of the cost of their health insurance. The trick of course, is to make that process simple enough for people to actually file for the refund.

A refundable tax credit would work like this: A family with a taxable income of $25,000 might owe $2,000 in federal taxes. If they purchased health insurance for $4,000 and qualified for a credit of $2,400 their tax bill would be reduced to 0 and they would get a $400 refund.

A major question with tax credits is who would be eligible for them. Some plans target only individuals who don’t have access to employer-based coverage. Some allow people to get a credit even for coverage purchased at work. Others target people with incomes below a certain level. And still others, more sweeping, would let all Americans get a credit – replacing or limiting the current tax exclusion for employment-based coverage. Similarly, credits can be structured in various ways. Some permit eligible people to take a credit for a fixed percentage of their insurance costs – say 25%, 30%, or 50%. Others create a flat dollar credit – for example, $800 for an adult and $400 for a child, up to maximum of a certain amount per year (say $2,400 or $3,000). Still others target credits to small firms to insure low-wage workers.

Vice President Gore has proposed a tax credit to offset 25% of the costs of health insurance for individuals who do not have access to employer sponsored coverage. Senator Bradley proposes

a broad and more complex tax credit plan that would make all families with income below $49,000 eligible. For example, a family of four with an income below $33,000 could get a tax credit of $1,200 per child. The adults would get a smaller credit for their coverage. Both Gore’s and Bradley’s plans make the credit “refundable.” A plan put forth last summer by the American College of Physicians/American Society of Internal Medicine (ACP/ASIM) would provide a $2,800 credit to all people with income below the poverty line (around $12,000 for an individual and up to around $18,000 for a family of four). The value of the tax credit would get reduced (in a stepped fashion) to $2,400 for uninsured adults up at 150% of the poverty level.

The impact of these approaches varies widely, as the accompanying Lewin study makes clear. For proposals that use tax credits to entice more low-income uninsured to purchase health insurance (and don’t propose other large-scale changes in the current system), the most important factor is the size of the credit. Researchers have modeled the effect of various tax credit levels and found that any credit that pays for less than half the cost of a health insurance policy or a health plan’s premiums attracts fewer than 25% of those eligible. The reason is simple: they still have to pay the balance of the premium and that can be very expensive. For example, a credit of $2,400 for a family of four may only cover a third to 50 percent of their annual health insurance premium. Even the very generous ACP/ASIM plan mentioned above (which would pay about 85% of the cost of coverage for an individual) is estimated to entice only 75% participation of the uninsured at the poverty level and an estimated 40% at incomes around 150% of poverty level. (One third of uninsured Americans have incomes between 100 and 200 percent of the poverty level.)

Dr. Kenneth Thorpe of Emory University examined the percent of single workers at or near 150 percent of poverty that would purchase health insurance with different tax credits. He assumed a cost of $2,800 per adult. He found that about 16% of eligible people would purchase health insurance with a $200 tax credit and 22% with an $800 credit (Thorpe, 1999).

Not surprisingly, rates of participation in health-related tax subsidy programs vary considerably by income level. Eighty-five percent of families with incomes over 400 percent of poverty take advantage of the tax exclusion of employer-provided health insurance, for example. On the other hand, only about 26% of those eligible for the Health Insurance Tax Credit (HITC – a component of the Earned Income Tax Credit (EITC) that existed in 1991 and 1992, and was later repealed), chose to participate in that program (GAO, 1994). A brief experiment, the HITC covered less than 10 percent of the cost of family coverage.

The EITC, enacted in 1975 and amended several times since, allows certain low-income workers to get an advance from their employers against taxes they’ll owe that year. These advances or “refunds” can be received month by month. The EITC is further discussed below since it serves as a potential model for the administration of a refundable health insurance tax credit.

Other Tax Credit Design Issues

  • When is the tax credit received?
  • Who or what entity receives the tax credit?
  • How is the availability of the tax credit marketed?
  • Who administers the program?

The principle reason most families and individuals do not have health insurance is that they can not afford it, even when it is offered through the workplace. That is, they literally don’t have the cash to pay the premium or if they did so, they could not afford to pay other expenses, such as the rent or utility bills. Therefore, it is useless for these families to get a tax credit or an IRS refund check in May every year. They need the money up front.

One idea proposed by many advocates of tax credits is to administer the credit through employers who offer coverage. That is, the employer would “front” the credit to employees to help them pay for coverage. (Ideally, the employer would be paying for a share of the coverage, too.) The employer would then recoup the credit from the IRS in its quarterly tax payments. This concept would be workable for low-wage workers at large firms but could be an administrative burden – and perhaps nightmare – for the nation’s small employers.

Another idea would be to create a new federal loan program that would let people borrow from banks against the credit. But that could be cumbersome to administer and possibly impractical. Yet another idea is to permit recipients to receive periodic advances against their future tax credit. The EITC can already be received in this way and could provide a model. But it’s a model with serious problems. It requires complex retrospective income determination, and applying for it is cumbersome. That’s why very few recipients (1/2 of 1 percent) actually choose to receive the EITC through periodic payments. One reason they do not may be specific to the cash payments obtained through the EITC (and hence irrelevant in the health insurance situation). Low-income taxpayers appear to prefer a large lump sum payment instead of periodic small sums. They also fear that fluctuations in their income (or health insurance coverage status) will leave them owing part of the credit back to the government after the year has ended (Glied, 1997).

Administering the program through employers leaves out people who are self-employed or unemployed or retired early. For this reason, many advocates link their tax credit proposals with programs that would permit small employers and individuals to purchase health insurance though collectives. These collectives or purchasing alliances could administer the credit as well as perform other administrative functions.

The EITC experience is also relevant to the marketing of a health insurance tax credit. A small percentage of people eligible for the EITC file for it. Other subsidized health insurance programs – offered through states – are also plagued by low participation rates. For example, some four million children who are eligible for Medicaid are not enrolled. Lack of knowledge of the programs is the chief culprit. A second reason is cumbersome application procedures.

Even educating recipients of one benefit about the availability of (and rules governing) a second benefit requires careful thought and effort. One reason for the relative failure of the HITC, which targeted EITC recipients, was a low level of awareness of the program. Advocacy groups, which might have helped in marketing the plan, were reluctant to support the HITC, in part because of their interest in protecting the EITC.

Finally, who would administer a tax credit program and does it matter? Some advocates believe a separate entity should be established within the Treasury Department. But most think a credit would have to be run by the IRS and overseen by Treasury. The Department of Labor also may have to be involved if some elements of a credit were administered through employers. Who, for example, would determine non-tax related rules of participation. For consumers, whoever runs it should be transparent.

Most importantly, if a tax credit program subsidized the purchase of individual insurance – coverage obtained outside of the workforce – it adds to problems already inherent in the individual insurance marketplace. This is a market run by brokers and small insurance operators who are adept at a process called “medical underwriting.” The aim is to sell insurance to mostly healthy people and avoid those who have a greater chance of getting sick. States regulate this market now, but with inconsistent effectiveness. If a tax credit prompted hundreds of thousands of Americans to switch from employer-based coverage to individual coverage, the pressure would be great to regulate the individual market.

That is why proposals for tax credits in the past have generally included measures for insurance market reform and closer regulation. These reforms usually included some type of “community rating” whereby all individuals who wish to enroll in a health plan are charged the same premium regardless of employment, family or health status. A main goal of previous proposals was to limit health insurers’ ability to charge different premiums to groups on the basis of risk.

Final Thoughts

Using the tax system as a vehicle for expanding health insurance coverage has advantages and disadvantages. It could enfold Americans who now lack health insurance – who, after all, pay taxes – into a well-developed system that already serves millions of people who have health insurance. It also may minimize the need to expand coverage through government entitlement or “welfare” programs, such as Medicaid – which hasn’t always delivered optimal access to health care services. In short, it “mainstreams” a portion of the uninsured population into the private health insurance system. A tax credit needn’t require annual appropriations by Congress and can garner some of the political protection already accorded the existing exclusion for employer-provided health benefits.

But unless it is designed carefully, there is a high risk that adopting a system of tax credits for health insurance will benefit mostly people who already have insurance and undermine positive

aspects of the current system. Unless they are very generous – as the accompanying study shows – tax credits induce very few uninsured people to buy coverage. As envisioned by many, tax credits allow millions of Americans who don’t have access to employer based coverage to buy coverage on their own. That has a major downside. The individual market serves the healthy very nicely, but often leaves the less healthy out in the cold. In addition, such tax credits undermine the existing – albeit arbitrary – insurance “pool” created by employers. People who get insurance at mid-sized and large companies are pooled in a system that keeps rates stable over time and benefits all. People who work in and get their health insurance through small firms have less of such a pooling benefit. Stable health insurance coverage, expanded through tax credits that moved people into private individual coverage, would almost certainly have to be accompanied by increased federal and state regulation of the health insurance marketplace.

In addition, tax credits in no way address the issue of the escalating cost of medical care and health insurance, nor do they address problems in the quality of care. Tax credits could even exacerbate both problems by inducing more people to leave employment-based coverage – where the aggregate clout of employers still has potential to hold down costs over the long run and improve quality through negotiation with managed care plans. Consumers purchasing health insurance on their own certainly gain choice, and many may choose coverage that reduces their utilization of care – such as high deductible policies. But they have little say over the aggregate cost of care or the quality of that care. Individually purchased insurance in America today is a gamble.

Finally, by definition tax credits alter the landscape of federal taxation. It is unclear at this point how they might fit in with overall tax policy, especially with any future attempt to simplify the tax system. It is also unclear how they fit in with long-term federal budget priorities. These matters would have to be addressed. For example, a tax credit of a fixed amount per family is consistent with tax simplification. A complex tax credit that requires income determination or gives a different level of tax break as health expenditures rise would seem to be inconsistent with tax simplification. Likewise, if over time the tax treatment of health care were gradually shifted from today’s exclusion and deduction system to a flat credit, it could be more compatible with a flat tax than the current system. In essence, the health tax credit could be subsumed into the general exemption for families in a flat income tax. (Butler, 1999).

The uninsured today are a large, growing, and diverse population. As attractive as tax credits are as one means to expand coverage, it is highly unlikely that they will achieve universal coverage in America. They are also unlikely to contain costs, make the health system easier to use, and propel the quality improvements so desperately needed to prevent premature deaths, ease human pain and suffering, and enhance the outcomes of treatment.


  1. Butler, S. A Health Tax Credit to Assist the Uninsured. Testimony before the House Ways and Means Committee, June 16, 1999.
  2. Custer, W.S. The Tax Preference for Employment-Based Health Insurance Coverage. Paper presented at the Employee Benefit Research Institute – Education and Research Fund Policy Forum, Washington, DC, May 5, 1999.
  3. Employee Benefits Research Institute (EBRI). Sources of Health Insurance and Characteristics of the Uninsured. Washington, D.C.,December 1998.
  4. Fronstin, P. Employment-Based Health Insurance: A Look at Tax Issues and Public Opinion. Washington, D.C.: Employee Benefits Research Institute, July 1999.
  5. General Accounting Office. Tax Policy – Health Insurance Tax Credit Participation Rate was Low. GAO/GGD-94-99, 1994.
  6. General Accounting Office. “Health Insurance for Children: State and Private Programs Create New Strategies to Insure Children.” GAO/HEHS-96-35, 1996.
  7. Glied, S. An Assessment of Strategies for Expanding Health Insurance Coverage. The Henry J. Kaiser Family Foundation, July 1997.
  8. Marquis, M.S. and Long, S.H., “Worker Demand for Health Insurance in the Non-Group Market,” Journal of Health Economics, 14(1):47-63, May 1995.
  9. McLaughlin, C.G. and Zellers, W.K., “The Shortcomings of Voluntarism in the Small-Group Insurance Market,” Health Affairs 11(2):28-40 Summer 1992.
  10. Sheils, J. and Hogan, P., “Cost of Tax-Exempt Health Benefits in 1998,” Health Affairs 18(2):176-181, March/April, 1999.
  11. Sheils, J., Hogan, P., and Haught, R. Health Insurance and Taxes: The Impact of Proposed Changes in Current Federal Policy. Washington DC: National Coalition on Health Care, 1999.
  12. Thorpe, K.E., “Changing the Tax Treatment of Health Insurance: Impacts on the Insured and Uninsured.” Paper presented at the Employee Benefit Research Institute – Education and Research Fund Policy Forum, Washington, DC, May 5, 1999.
  13. U.S. Bureau of the Census. Health Insurance Coverage:1998. (October 1999). Data from the March 1999 Current Population Survey.
  14. U.S. Bureau of the Census. Dynamics of Economic Well Being: Health Insurance 1993 to 1995, Who Loses Coverage and for How Long. (August 1998). Data from the Survey of Income and Program Participation. The calculation here is made on the basis of the current U.S. population of 270 million people.

More from releases