Overcoming Obstacles Facing the Uninsured: How the Use of Medical Savings Accounts, Flexible Spending Accounts, and Tax Credits Can Help

May 8, 2003 • Testimony

Good afternoon, Mr. Chairman and Members of the Subcommittee. My name is Tom Miller. I am director of health policy studies at the Cato Institute. It is a pleasure to appear before you today to examine how potential changes to tax‐​advantaged mechanisms for financing health care spending can address problems faced by the uninsured.

Medical savings accounts (MSAs), flexible spending accounts (FSAs), and tax credits represent three types of tax‐​advantaged vehicles that could play a larger role in improving access to affordable health insurance and health care for millions of uninsured Americans. Removing or redesigning restrictions on the availability of each one of them would help reduce the number of uninsured individuals and families. Even more progress could be achieved by providing broader parity in the tax treatment of health insurance financing for all purchasers; reducing artificial tax and regulatory barriers to market‐​based, value‐​maximizing choices; and empowering all health care consumers to match their own needs and preferences to a wider variety of affordable options.

Before focusing more specifically on the three tax‐​advantaged vehicle options noted above, I first would like to frame the overall issue of health insurance availability and affordability within the context of the small‐​employer market and then the health care system as a whole.

During the recent period of annual double‐​digit percentage increases in the health insurance premiums, the cost issue has driven everything else within our health system. Health insurance premium expenses have been particularly difficult to afford for the many smaller employers that teeter on the edge of decisions whether to drop current coverage, reduce its scope and scale, or offer some sort of coverage for the very first time.

Small employers — those with fewer than 50 workers — employ almost one‐​third of the private sector workforce. The smaller the employer, the less likely it will be to sponsor health insurance coverage — particularly at the threshold levels of below 50 workers, and below 10 workers.

The relatively higher cost of small group insurance coverage for comparable benefits is one of the primary reasons why many small employers don’t offer it and/​or some of their employees don’t demand it. It’s generally more expensive to sell and administer insurance for small employers. About 20 to 25 percent of small employers’ group insurance premiums go toward expenses other than health benefits — compared to about 10 percent for large employers. However, the administrative cost differential for small employers has been reduced over the last decade or two — it used to be closer to 40 percent of group insurance premiums, but growth in managed care and improvements in information technology narrowed some of that gap.1 Even though basic demographic and self‐​reported health characteristics of individuals insured through the average small employer and the average large employer do not vary significantly, it remains harder for an insurer to predict the health status and the accompanying health costs of any particular small group. Hence, various degrees of medical underwriting are used for small groups because the law of large numbers does not kick in for actuarial predictions of future costs until a group approaches about 500 covered lives — although limited versions of experience rating begin to replace medical underwriting well below that level.2

Although some states do not allow private insurers to adjust premiums for the health characteristics of employee enrollees, GAO found that small employers in those states had average premiums about 6 percent higher, compared with other states, when adjusted for geographic differences in the cost of physician services. Yet those states prohibiting insurers from setting premiums based on health status did not gain anything in terms of covering more high‐​risk employees. Those states did not have a higher proportion of high‐​risk individuals insured through small employers than states with more flexible restrictions.3

Lower‐ income workers in small firms bear the brunt of this and other forms of excessive state health insurance regulation. Their employers generally are unable to selfinsure and thereby gain ERISA protection from costly state benefit mandates, restrictions on rating and underwriting, and other regulatory burdens.

Market‐​based public policy cannot, and should not, overcome all of the natural cost differentials between large and small employers that may be due to the stability, size, and composition of a given employer’s pool of insurable workers, as well as any relative economies of scale in benefits administration. But tax and regulatory policy should, at a minimum, avoid making those cost differentials greater than they need to be.

We should also consider whether improving health outcomes and health status for lower‐ income individuals, rather than just reducing the total number of uninsured Americans, should be the primary goal of health care system reform. Increasing health insurance coverage levels per se remains at best an imprecise tool of limited effectiveness in achieving that objective. Hence, I have suggested elsewhere that it may well be more efficient, on balance, to improve the overall affordability of insurance coverage by encouraging it to be less comprehensive, while selectively targeting expansion of safety net care, than to subsidize expansion of conventional health insurance coverage.4

Making the market‐​based cost of care more transparent to all parties involved in health spending decisions would encourage its more efficient consumption and delivery. Reducing the long‐​term rate of growth in the cost of health care remains more important than (and, beyond a certain point, operates at cross‐​purposes to) expanding the scope and scale of subsidized health insurance coverage. Past levels of health insurance subsidies have increased medical costs and the demand for health insurance, creating net welfare losses estimated at 20 percent to 30 percent of total insurance spending.5 The consequences of failing to reduce, let alone further increasing, the effects of comprehensive third‐​party insurance in raising costs and limiting access to health care would do more overall harm than any modest disincentives to obtain insurance protection that may be caused by direct provision of more charity care. When rising health care expenditures outpace wage increases, their strongest effect is to reduce health insurance coverage for low‐​income workers. Hence, at the margin, increasing incentives to purchase less‐​comprehensive health insurance and filling in urgent gaps in direct delivery of health care through safety net mechanisms may produce more affordable and accessible health care.

If the rate of cost growth dips lower, other complaints about access, quality, and equity become much more manageable.


One of the primary factors driving health care costs higher in the past has been the increased share of medical bills paid by third‐​party payers such as private health insurers, employers, and government health program administrators. Whereas out‐​of‐​pocket payments by individual consumers accounted for about 50 percent of total health care spending in 1960, the share of third‐​party payments (by private health insurers, employers, and government agencies) for health care has grown to over 80 percent. Third‐​party payment of health bills insulates individual consumers from the real cost of their health care decisions and treatment. Consumers have less reason to avoid unnecessary care, question costs, or shop around for the best treatment available at a reasonable price, but they have every incentive to demand more services. Excessive third‐​party coverage with low deductibles also increases administrative costs, because every small bill must be submitted for review and checked for accuracy.

The centerpiece of market‐​oriented health care that can reverse this trend remains MSAs. They combine two elements‐​a savings account controlled by the insured individual to be used to pay for routine health care expenses and a high‐​deductible (catastrophic) insurance policy to cover more substantial health care needs. The savings account is controlled by the insured person and used to pay routine health care expenses. The accompanying catastrophic insurance policy covers more substantial health care costs. Because the cost of such a policy is usually significantly less than the cost of a low‐​deductible policy, the money saved may be used to increase contributions by an individual (or his employer) to an MSA administered by a designated trustee or custodian. Unspent MSA funds, including any interest or investment earnings, accumulate from year to year, providing money to cover possible medical expenses in the future. This feature also expands the time horizons of MSA consumers to look beyond a single year at a time of insurance coverage in order to maximize the value of their health care dollars.

The evidence is overwhelming that increased cost sharing reduces health insurance premiums substantially. For example, Jason Lee and Laura Tollen noted in a June 2002 article in Health Affairs that increasing cost sharing from a plan with $15 copays and no deductible to one with 20 percent coinsurance and a $250 deductible reduces premiums by about 22 percent; and a combination of 30 percent coinsurance and a $1000 deductible would reduce premiums by 44 percent.6

Because MSA plans are linked to high‐​deductible insurance that covers health claims that are more catastrophic in nature, they make the cost of insurance coverage more affordable for most Americans. Less‐​comprehensive coverage will mean lower premiums for a larger fraction of people with low incomes. The majority of standardized insurance policies currently available are generous and expensive — making them unaffordable to low‐ income people. On the other hand, catastrophic insurance for very large, less‐​predictable health care expenses forces consumers to bear the full marginal costs of health care up to the point where their use of health care exceeds the deductible.

Because MSAs funnel a much smaller share of health care spending through third‐​party insurance, they provide workers strong market incentives to control the costs of their health care. Account holders are effectively spending their own money for routine health items. That, in turn, stimulates real cost competition among and price disclosure by doctors and hospitals.

Offering less comprehensive insurance plans with greater enrollee financial responsibility is designed to encourage enrollees to be smarter consumers of health care services, limit demand for less beneficial “discretionary” care, seek out higher‐​value options, and save money for more critical medical needs in the future. Under many thirdparty health benefit arrangements, consumers have little incentive or ability to become more knowledgeable about health care. MSAs stimulate consumer demand for information about the quality and price of health care.

A number of studies have illustrated that MSAs help control costs, improve access to health care, expand consumers’ choice in and control of health care, and increase savings. Moreover, MSAs improve health plan options not just for affluent and health individuals, but for all Americans.

In April 2000, RAND Corporation researchers examined the effects of making MSA options available to small businesses. RAND rejected the assumption that MSAs appeal most to the wealthiest and healthiest workers. It found that HMOs would remain more attractive to higher‐​income workers, primarily for tax reasons, and exceptionally good health risks would be more likely to decline any insurance at all than to select the MSA option.7

A 1996 study by National Bur eau of Economic Research analysts (including current FDA commissioner Mark McClellan) concluded that most workers would end up retaining a substantial portion of the contributions they made to MSAs by the time they retired. Approximately 80 percent of employees would have retained over 50 percent of their MSA contributions by the time of retirement, and only 5 percent of workers would have saved less than 20 percent of their contributions. Although workers with high health care expenses in one year tend to have lower but still higher‐​than‐​average expenses in the next few years, the concentration of annual expenditures declines continuously as more and more years of expenditures are cumulated. High expenditure levels typically do not last for many years.8

Another 1996 study of Ohio‐​based firms that offered MSAs that did not qualify for tax advantages under the Health Insurance Portability and Accountability Act (HIPAA) determined that the employer’s total cost for family coverage under those MSA plans averaged 23 percent less than traditional family coverage, yet the average employee with family coverage also would be $ 1355 better off under the worst‐​case (maximum out‐​of‐​pocket liability) scenario.9

So, if MSAs are so great, why don’t we see more of them in the marketplace?

In 1996, HIPAA authorized up to 750,000 “tax‐​qualified” MSAs over a fouryear period (later extended to December 31, 2003). Unlike previous MSAs, those socalled Archer MSAs featured tax‐​deductible treatment of MSA deposits and tax‐​exempt treatment of investment earnings accumulated with the MSAs. However, the potential of Archer MSAs has been hampered by eligibility limits and other design flaws mandated by HIPAA.

Congress still needs to permanently authorize federally qualified MSAs; lift the enrollment cap and allow an unlimited number of people to have MSAs; expand MSA eligibility to include employees in businesses of all sizes, as well as employees without employer‐​sponsored insurance (ESI); allow MSA plans to offer a much wider range of deductibles; allow MSA holders to fund fully their MSAs each year, up to 100 percent of the insurance policy deductible; allow employers and employees to combine their contributions to MSAs at any time within a given year; and either preempt first‐​dollar state‐ mandated benefits or provide the flexibility for MSA plans to adjust to comply with those conflicting mandates.

Even under the current handicaps imposed by HIPAA rules, MSAs have become particularly attractive to the uninsured. Last October, the Internal Revenue Service reported that 73 percent of taxpayers filing MSA tax returns for the taxable year 2001 previously lacked any health plan coverage at some time within the six‐​month period before their high‐​deductible/​MSA coverage began. Archer MSA eligibility is not only prohibited for individuals who work for companies with more than 50 employees; it also does not extend to workers in companies, of any size, that do not offer MSAs as a benefits option or even in companies that do not offer any kind of employer‐​sponsored health insurance at all!

Expanding MSA eligibility to all health insurance purchasers and permanently authorizing tax‐​advantaged treatment for Archer MSA coverage would deepen the potential market for MSA plan vendors and encourage larger insurers to invest the necessary resources to develop and market such products. It also would increase the level of consumer knowledge of the uninsured, or “cash market,” prices for more discretionary health care services and products. Finally, it would help facilitate the potential convergence of MSA options with other tax‐​advantaged vehicles for health care spending, such as health reimbursement accounts (HRAs) and FSAs.


FSAs have a longer history than MSAs. These tax‐​advantaged health spending vehicles were first approved by Congress in 1978. FSAs allow an employee to authorize a pre‐​tax salary reduction amount, generally in the first month of a given calendar year, that will be transferred into an account to be used to pay directly for health care expenses, as defined quite broadly by section 213(d) of the Internal Revenue Code. FSA contributions are exempt from both income and payroll taxes.

Access to FSAs and greater use of them are limited by several rules. Eligible employees must work for a company that decides to offer an FSA benefit option; they cannot set up a personal FSA on their own. They also must lock in their annual contribution amount early in the year and cannot adjust it later. Any unspent balances in a worker’s FSA at the end of the calendar year are subject to “use it or lose it” penalties. Forfeited funds revert back to their employer.

According to James Cardon and Michael Showalter of Brigham Young University, major U.S. firms are the most likely employers to offer FSAs.10 Actual participation rates by employees remain relatively low. Decisions regarding whether to participate, as well as how much to contribute, tend to be quite conservative, and they are primarily based on employees’ foreknowledge of their likely health expenditures in a given year. Participants tend to spend their FSA amounts relatively early in the year (Cardon and Showalter note that January 1 appears to be busiest day of the year for FSA charges and that nearly 11 percent of FSA expenditures for the year are incurred in the first week of January). They then run a negative account balance for most of the year until their periodic contributions to the FSA catch up.11 This incentive to use FSAs as interest‐ free loans from bene fits administrators also places a sponsoring employer at risk for the portion of any upfront FSA expense reimbursements that would not be fully repaid if an employee left the firm before the end of the year.

Although the median exha ustion date for the FSAs in the Cardon/​Showalter sample was early October, some portion of all FSA holders are likely to miscalculate their ability to hit their health spending targets. As the end of year nears for them, the use it or lose it penalties provide strong incentives to spend remaining funds on less valuable health services (triggering the so‐​called “designer sunglasses in December” syndrome).

Changing FSA rules to allow unspent balances to carry forward into the next year without being forfeited or subject to taxes would remove the most flawed incentives under current rules. Eligible workers would become more likely to participate in FSA options and to dedicate more funds to them.

Increasing the size of average FSA contributions also would help absorb more of their relatively high administrative costs for claim filing. Ideally, expanded use of smart card technologies by FSA administrators and improved disclosure of comparative price information in the market for out‐​of‐​pocket health spending would provide other compensating efficiencies and increased point‐​of‐​sale convenience.

Although FSA critics may presume that wealthier workers take greater advantage of current FSA options than less affluent ones, Cardon and Showalter find that is not true. They also conclude that an increase in a worker’s level of risk aversion increases the likely amount of their FSA contribution. In that case, FSA benefits operate in a manner similar to stop‐​loss provisions in standard health insurance policies.12

In any case, the full potential of FSAs should be unleashed. Allowing FSA holders to accumulate tax‐​advantaged funds on a multi‐ year basis, enabling workers to invest fund balances, and providing full portability as employees change jobs would encourage a further restructuring of private health insurance markets. FSAs could become the primary vehicle for financing medical procedures and health services that involve relatively predictable expenses and are easily foreseeable in the near future.13 Value‐​conscious employers and employees also could insist that insurers “spin off” (not insure) those items for which the typical treatment cost is relatively low compared with the paperwork required to process the claim. Self‐ financing those items through expanded FSAs would be more efficient than covering them under comprehensive health insurance plans.

Multi‐ year, portable FSAs also could provide a backdoor route to greater firstparty control of health spending decisions for workers who continue to lack access to full‐​strength MSA products. Even the currently authorized version of FSAs could be cleverly combined with employer‐​financed HRAs to allow joint employee/​employer funding of much larger tax‐​advantaged health savings accounts to handle higher levels of out‐​of‐​pocket spending and encourage greater use of high‐​deductible insurance coverage.

Tax Credits

The last few years of political debate over the relative merits of refundable, fixeddollar tax credits versus expanded public program coverage for assisting the uninsured have generated a superficial consensus to “do something” with little agreement on which path to take. Rather than simply repeat the steps of others down that well‐ worn policy path, I would prefer to offer a few cautionary no tes regarding what has been overlooked in that discussion and then suggest an alternative approach that both limits and expands the conventional tax credit approach.

First, we should remember that the primary goal of tax policy in this area should not be to steer health care consumers toward financial arrangements that make most health care spending choices appear to cost significantly less than their market value. Extending the longstanding distortions of the tax exclusion for employer‐​sponsored group insurance to other forms of health spending may become necessary in a world of thirdbest and fourth‐​best policy alternatives, but we should try harder to aim at a version of tax parity for health insurance purchasers that holds the resulting collateral damage to a minimum.

More than fifty years of health policy history should remind us of the costs, as well as the benefits, of the special tax treatment of ESI. The tax exclusion forced many working Americans to accept the only health plan offered by their employer, or pay higher taxes. Job‐​based tax benefits for health care spending put employers, instead of employees, in charge of selecting health care benefits. They also raised the comparative after‐​tax price of other non‐​employer‐​based insurance alternatives.

The tax exclusion continues to distort health care purchasing choices today by favoring the financing of medical services through insurance and providing the greatest tax benefits for the most costly versions of employer‐​sponsored coverage. It encourages workers to think that someone else (their employer) pays for their health care, and it reduces their sensitivity to the cost of health insurance choices. It disconnects the consumption decisions of insured workers and their families from the payment decisions of employers and their insurers. Tax subsidies for health insurance overstimulate the demand for health care and, perversely, increase its total cost. Because the current tax subsidy for health insurance is inefficient and unfair, it should be reformed to place individuals, not employers or government, in charge of choosing something as personal as health care.

The best way to remove tax policy distortions in the health insurance market would be to eliminate tax subsidies for employment‐​based health insurance altogether.

Employers that continued to offer ESI should be required to report the value of the employer‐ financed share of that coverage to individual employees on their regular periodic pay statements and annual W-2 forms. The default setting for such disclosure would assume that workers in employer‐​group plans are community rated within the firm and the employer contributions for coverage are identical for each worker (such as the periodic equivalent of the firm’s per‐​employee COBRA premium). In the event that employers were allowed to adjust health plan contributions to reflect factors specific to individual workers, they could report those different amounts instead. The new tax credits should be assignable to insurers and advanceable, but not refundable. The maximum tax credit available to any eligible individual would be no greater than that individual’s total federal income tax and FICA payroll tax liability (including both the employee and employer shares) for the previous calendar year. Only taxpayers would receive tax credit “relief” for health insurance costs.

The net effect of the above tax reform would be to encourage workers and their families either to move from ESI coverage to individually purchased insurance or to ensure that the ESI plan they select represents the best competitive value they can find.

Even if an employer were no longer “paying” directly for an employee’s insurance coverage under an employer‐​sponsored group plan, the employer still could facilitate delivery of health tax credit assistance through several mechanisms. An employer could choose to include in an employee’s gross wages an amount equivalent to what the employee otherwise would have received as the non‐​taxable employer’s contribution to the employee’s applicable share of any group coverage offered under the employer’s health benefits plan. An employer could “list bill” and allow employees using the tax credit option to pay their individual insurance premiums through payroll deduction (that is, an insurer bills the employer for a list of designated employees who have opted for nonemployer‐​sponsored coverage). In the latter case, the workers would bear the entire premium, but apply their tax credit to reduce the net payment due.

Congress should consider using the new tax credit option to leverage other market‐​opening reforms. In that case, consumers wishing to use the tax credit would have to purchase an insurance package that covered a minimum set of health services and included a minimum, but significant, front‐​end deductible (along with maximum out‐​ofpocket “stop‐ loss” limits). Qualified insurance policies might provide separately priced guaranteed renewal options in return for exemption from HIPAA’s guaranteed renewal requirements. Those policies also should be exempt from individual state benefit mandates.

New voluntary purchasing pools also could be authorized to accept tax credit funds to pay for such qualified insurance in return for federal preemption of state benefit mandates, fictitious group laws, or rating laws that would otherwise interfere with their operations. Those pools would need to meet certain minimum criteria that include capital and solvency requirements. They would have to provide annual open seasons and be open to all willing purchasers who use the health tax credit option.

Such purchasing pools would have the potential to provide an efficient mechanism for workers to gain a wider choice of health plans than many employers (particularly smaller ones) can offer on their own. Indeed, they might provide effective alternatives to poorly performing employer‐​selected health plans. For that reason, pool participation should not be limited just to business firm buyers making collective decisions for all their employees. Membership in “voluntary” purchasing pools should reflect the preferences of individual workers and other health care consumers, not just the interests and convenience of employers.

The role of purchasing pools would be to provide a single, stable source of ongoing coverage. They would ease the burden of choosing and buying coverage, particularly for people seeking insurance without the assistance of an employer. Pool administrators would help design benefits packages offered to individual pool participants. They would negotiate contracts and premiums with the health plans choosing to sell to pool members. In short, pool administrators would and should be effective purchasers and advocates on their behalf.

Providing a new tax credit option could jump‐​start the evolution toward an employee benefits environment in which workers more directly control their healthcare benefits and insurance choices. It would ensure sufficient consumer demand for individually selected insurance arrangements and provide a competitive alternative to ESI coverage.

Any new tax credits for health care should not try to finance comprehensive insurance for all uninsured, low‐​income Americans. When most refundable tax credit proposals award tax “cuts” to individuals who pay little, or no, federal taxes, they blur necessary policy distinctions between how to set the appropriate level of income‐​based welfare assistance and how to neutralize the many distortions caused by our complex tax system. The politics of refundable tax credit proposals also have unfortunately steered recent health care debates away from broad, individual‐​empowerment tax reforms and toward a narrow, cramped version of targeted handouts to smaller slices of the lowerincome, uninsured population.

Refundable tax credits combine bad tax policy, bad welfare policy, and bad health policy. They reinforce the mistaken stance of those who argue that cuts in marginal tax rates are somehow ”unfair” when they provide most of their benefits to those who pay the largest share of federal income taxes. Refundable credits also are prone to carrying the lumpy baggage of complex income‐​based, phase‐​out levels; tight restrictions on the contents of eligible health benefits packages; and narrow rules for eligible insurers.

Making health tax credits refundable would endorse expansion of current taxpayer‐ financed ”entitlements” to health insurance coverage. It would adopt the view that health insurance is a ”merit good” for everyone and that necessary access to health care cannot be adequately financed without even greater subsidies from taxpayers for insurance coverage. Many lawmakers who salute the remarkable benefits gained from limiting the magnitude and duration of cash assistance to low‐​income beneficiaries on the welfare rolls nevertheless appear poised to dole out a new round of permanent “welfare” checks to the working poor, hidden beneath a refundable health tax‐​credit label.

For low‐ income individuals lacking access to health insurance, the better policy solutions include safety net reforms that strengthen state high‐​risk pools and encourage charitable contributions to provide health services through nonprofit intermediaries. Dollar for dollar, investing in safety net assistance that directly delivers care to the uninsured is more effective and productive than trying to coax them to purchase health insurance with modest tax subsidies.

The current market for charity care operates quite rationally in mimicking the effects of private catastrophic insurance policies. The proportion of health care paid outof‐ pocket by the uninsured decreases considerably as utilization and total “spending” increases. Proportionately more charity care is available for uninsured individuals who incur larger medical expenses. And the low‐​income uninsured with high medical bills (above $10,000) pay about half as much out‐​of‐​pocket for their care as do high‐ income uninsured individuals with similarly sized bills. (The low‐​income uninsured with such high bills receive 90 percent of their care for free.)14

In the long run, improving the quality of education that lower‐ income individuals receive, expanding their personal control of health care decisions, and reversing regulatory policies that increase the cost of their health care will yield even greater returns in improved health outcomes.

If Congress nevertheless feels compelled to provide refundable tax credits to lower‐ income Americans for reasons that blur universal coverage goals with income redistribution objectives, it should first answer the question: Why not simply provide equivalent income support through more targeted delivery of fungible cash vouchers and then concentrate on implementing broader health policy reforms that improve the availability in private markets of more affordable health insurance purchasing options? Sorting out the parameters of income support policies from the opportunities for broader health care market reform would target better the appropriate policy instruments for each of these distinct objectives.

Refundable tax credit subsidies directed primarily at uninsured lower‐​income workers obviously would provide some increase in overall health insurance coverage. But they still will need to be financed at a higher enough rate — perhaps 50 percent or more of premiums for relatively limited benefits — to have any substantial effect on levels of uninsurance in the target population. Fixed‐​dollar credits might be easier to score for budget purposes and to provide higher coverage take‐ up rates, but proportional tax credits would be more effective in targeting uninsured individuals at the highest risk and in the greatest need.

In any case, new tax subsidies for health insurance purchasing should pass through directly to individual consumers, rather than be used as initial sweeteners to induce employers to maintain or establish job‐​based coverage. The current tax code already allows employers sufficient opportunity to deduct the full cost of providing group insurance coverage to their employees, if they chose to do so, as an ordinary business expense.

Other Policy Tools to Help the Uninsured

The inevitable tendency of conventional politics is to latch on to facile fixes for chronic and complex problems by channeling seemingly “free” sources of budget dollars into narrow program silos. But attempts to provide sufficiently generous tax credits for the uninsured, with an end goal of near‐​universal health insurance coverage, will quickly bump up against other competing claims on public resources. Expanding the availability of MSAs and acknowledging the tradeoffs between catastrophic insurance protection and subsidies for early‐​dollar health spending would allow us to accomplish more by aiming to do less. Other promising policy tools beyond the scope of today’s hearing might include providing “competitive federalism” incentives to deregulate state‐​based insurance regulation, facilitating the convergence of various tax‐​advantaged savings vehicles for future health care needs, and clearing away remaining policy barriers to consumer‐​driven health care options that can travel through all terrains.15

It would also help to keep the chronic problem of the “uninsured” in perspective. Last September’s annual Census Bureau estimate of the number of uninsured people in the U.S. found that Americans lacking health coverage during 2001 increased by 1.4 million, to 41.2 million, or 14.6 percent of the population. That number by itself was disappointing and needs to be improved. However, it occurred during a year in which an economic recession began, it represented a smaller share of the population being uninsured than was the case from 1992 through 1998, and more than half (approximately 58 percent) of the year‐​to‐​year increase in the uninsured came from members of households with annual incomes of $75,000 or more.

Our society may well decide to spend ever greater shares of our nation’s resources on health care in future years — for the uninsured, perhaps, but certainly for the worried well of the middle class — as long as someone can be found somewhere to foot the bills. But American consumers will receive more value for each dollar they spend only if the distorting effects of government’s multiple roles as a regulator, purchaser, and subsidizer of health care are reduced. Our objective should be neither to artificially keep spending levels higher, nor to keep them lower, than their market‐​determined costs. Instead, we should allow individual consumers to seek the best value that balances their spending preferences and priorities with the resources that they can command.

  1. U.S. General Accounting Office, “Private Health Insurance: Small Employers Continue to Face Challenges in Providing Coverage,” GAO-02–08, October 2001, pp. 13–15.
  2. Ibid., pp. 10–16.
  3. Ibid., pp. 16, 17.
  4. Tom Miller, “Improving Access to Health Care without Comprehensive Health Insurance Coverage: Incentives, Competition, Choice, and Priorities,” in Covering America: Real Remedies for the Uninsured, ed. E.K. Wicks and J.A. Meyer. Washington: Economic and Social Research Institute, 2002.
  5. See, e.g., Martin Feldstein. Hospital Costs and Health Insurance. Cambridge, MA, and London: Harvard University Press, 1981, pp. 88, 201–03, 239–44; Martin Feldstein. “The Welfare Loss of Excess Health Insurance.” Journal of Political Economy 81 (2) (1973): 251–80. See also Edgar A. Peden and Mark S. Freeland. “A Historical Analysis of Medical Spending Growth, 1960–1993,” Health Affairs 14 (2) (1995): 236–47 (finding that about half the growth in real per capita medical spending from 1960 to 1993, and twothirds of its growth from 1983 to 1993, resulted from either the level or growth of insurance coverage).
  6. Jason S. Lee and Laura Tollen, “How Low Can You Go? The Impact of Reduced Benefits and Increased Cost Sharing,” Health Affairs web exclusive, June 19, 2002, www​.healthaf​fairs​.org/​W​e​b​E​x​c​l​u​s​i​v​e​s​/​L​e​e​_​W​e​b​_​E​x​c​l​_​0​6​1​9​0​2.htm.
  7. Dana P. Goldman, Joan L. Buchanan, and Emmett B. Keeler, “Simulating the Impact of Medical Savings Accounts on Small Business,” Health Services Research 35 (1) (2000): 53–73.
  8. Matthew J. Eichner, Mark B. McClellan, and David A. Wise, “Insurance or Self‐​Insurance? Variation, Persistence, and Individual Health Accounts,” National Bureau of Economic Research Working Paper no. 5640, June 1996.
  9. Michael T. Bond, Mary W. Hrivnak, and Brian P. Heshizer, “Reducing Employee Health Expenses with Medical Savings Accounts,” Compensation and Benefits Review 28, no. 5 (September‐​October 1996): 51- 56.
  10. James H. Cardon and Mark H. Showalter. “An Examination of Flexible Spending Accounts,” Journal of Health Economics 20 (6) (2001): 935–54.
  11. Ibid., p. 944.
  12. James H. Cardon and Mark H. Showalter, “Flexible Spending Accounts as Insurance,” Journal of Risk and Insurance 70 (1) (2002): 49–50.
  13. Cardon and Showalter suggest that such items might include orthodontic care, prescription drugs for chronic conditions, dental care, optical care, and other elective health services. “An Examination of Flexible Spending Accounts,” pp. 936, 953.
  14. Bradley J. Herring. “Does Access to Charity Care for the Uninsured Crowd Out Private Health Insurance Coverage?” Yale University Institution for Social and Policy Studies working paper. September 7, 2001, p. 10.
  15. See Miller, 2002.
About the Author