This Article Abstract Submit a response to this article Services E-mail this article to a friend Alert me to new issues of the journal

S T A T E  I S S U E S : 
F E D E R A L I S M
W E B  E X C L U S I V E 

16 July 2003

John Holahan, Alan Weil, and Joshua M. Wiener

ABSTRACT:

The current balance of responsibility between states and the federal government for low-income people’s health coverage has achieved a great deal. It covers many of the neediest people, supports the safety net, responds to emerging needs, and supports some experimentation. However, it leaves more than forty million people uninsured, allows excessive variation across states, places unsustainable pressure on state budgets, creates tension between the two levels of government, and yields too few benefits from experimentation. This mixed record argues for a significant simplification of and increase in eligibility for public programs, with the federal government either providing extra funds to states to meet these needs or assuming full responsibility for insuring the poor.

Financing for health and long-term care for low-income Americans is a joint federal and state responsibility. The states assume major financial and administrative responsibilities while the federal government provides substantial funding and oversight. How to balance these responsibilities has been debated for decades.

Controversy over state and national roles in health policy mirrors broader debates over federalism that trace their roots to the founding of the United States.¹ For philosophical reasons, some view states as the appropriate locus of authority to reflect local values and priorities. Others prefer a strong national role to achieve national objectives and, as the U.S. Constitution says, “to promote the general welfare.” Theories of public choice and public finance also shape views of federalism. To some, a strong state role encourages competition to develop the best and most efficient policies. Others argue for a stronger federal role because income redistribution is best carried out at higher levels of government and because interstate competition can lead to a race to the bottom as states cut programs for the poor so they can lower taxes to attract businesses and high-income taxpayers. Despite these philosophical positions, most Americans are pragmatists. When they see a problem, they will turn to whichever level of government they believe will do the best job of solving it.

This paper examines federalism in health policy from a pragmatist’s perspective. Drawing upon seven years of quantitative and qualitative research conducted as part of the Assessing the New Federalism project at the Urban Institute, it assesses the current balance of responsibility between states and the federal government. How have states responded to the combination of financial support and programmatic flexibility they now have? How innovative have states been, and how effectively have good ideas been shared? Might other models of federalism yield better health policy? The findings presented here are discussed in greater depth in Federalism and Health Policy, a new book edited by the authors of this paper and published by the Urban Institute Press.

The nation is engaged in a debate over how best to structure federalism in health policy. State revenues are declining, and the Congressional Budget Office (CBO) projects that Medicaid costs (the second-largest item in most state budgets) will increase on average 8.5 percent each year over the coming decade.² States are calling for fiscal relief, while the Bush administration has proposed offering states voluntary block grants that link increased flexibility with predetermined levels of federal funding. Meanwhile, the number of uninsured Americans rose in 2001 after two years of decline; when data are available for 2002, they will likely show another substantial rise.

What Have We Gained From The Current Federalism Balance?

The current division of responsibility for health care has given the system three major strengths: access to health care for many of the neediest Americans, the ability to evolve to meet emerging needs, and the capacity to experiment.

Health care for the neediest Americans. Medicaid is the backbone of the current system of care for low-income Americans. The program is an open-ended individual entitlement that covers about thirty-eight million children and their parents, pregnant women, frail elderly people, and people with disabilities. Covered benefits include hospital care; physician, laboratory, and radiological services; prescription drugs; and long-term care. States administer Medicaid under federal supervision. The federal government pays 50–77 percent of the program’s costs, depending upon the state. Combined federal and state Medicaid expenditures in fiscal year (FY) 2002 were over $257 billion, about equal to those of Medicare. Medicaid expenditures are expected to exceed Medicare’s in FY 2003.³

Medicaid, the State Children’s Health Insurance Program (SCHIP), and smaller state-funded programs together cover about 35 percent of Americans with incomes below the federal poverty level ($18,400 for a family of four in 2003) and 25 percent of those with incomes of less than twice the poverty level. These programs cover about half of children and one-quarter of adults with incomes below the poverty level.⁴ Medicaid pays for more than one-third of all births in the United States and provides prenatal care for low-income pregnant women and preventive services for children.

Medicaid has improved access to health care services for millions of low-income Americans. People on Medicaid have better access to care and use more services than people without insurance coverage, and they fare as well on these measures as people with private insurance coverage do, after differences in population characteristics are controlled for.⁵ Medicaid has also reduced out-of-pocket spending on health care, relative to a comparable uninsured population.⁶

Medicaid covers people with a wide range of disabilities, including mental illness, developmental and physical disabilities, and HIV/AIDS. It supports institutional care for low-income older people and people with disabilities, paying for about half of all nursing home costs and two-thirds of all nursing home residents.⁷ Increasingly, Medicaid is paying for long-term care provided in people’s homes and communities. It also pays Medicare premiums, deductibles, and coinsurance for 5.2 million elderly or disabled persons, as well as for some acute care services not covered by Medicare. In particular, Medicaid finances the purchase of prescription drugs for many of the nation’s sickest people, spending about $16 billion in 2001 on prescription drugs for older people and people with disabilities.⁸

Medicaid provides coverage to people who would be hard-pressed to obtain or afford private coverage that meets their medical needs. Medicaid enrollees are considerably less healthy than average. Beneficiaries are far more likely than either the privately insured or the uninsured are to report being in only fair or poor health and to have chronic conditions and activity limitations. This is the case not only for people enrolled because they have a disability, but also for the Medicaid population as a whole.⁹

In 1996 Congress established SCHIP to extend coverage to near-poor children ineligible for Medicaid.¹⁰ Like Medicaid, SCHIP is administered by the states within federal guidelines, and funding is shared between the two levels of government. However, federal funding for SCHIP is capped, there is no federal entitlement to benefits for individuals, and states have more flexibility and control over the program while they pay a smaller share of the costs relative to Medicaid.

States responded enthusiastically to SCHIP, and most implemented the new program rapidly. Enrollment reached 3.6 million by June 2002. States’ aggressive SCHIP outreach also brought in many children eligible for Medicaid and is no doubt responsible for some of the increase in children’s enrollment in Medicaid between 1998 and 2002.

Medicaid and other programs also support safety-net providers—hospitals, clinics, and individual practitioners—delivering care to the uninsured. While uninsured Americans have less access to care than those with coverage have, and their health suffers as a result, some care is available when needed.¹¹ For example, 77 percent of parents of uninsured children with family incomes below 200 percent of poverty report that their children have a regular source of care other than a hospital emergency room, and an even larger share (89 percent) feel confident that they can get their children care if they need it.¹² Adults fare worse on these counts, but these data demonstrate that a functioning safety net meets many of the uninsured’s needs.

Responding to emerging needs. The current system addresses new issues as they arise. It has covered new populations as gaps in coverage have been identified, and it has adapted to and sometimes initiated shifts in health care delivery.

Medicaid eligibility has greatly expanded since the program began. In response to evidence showing the importance of prenatal care, the program expanded mandatory coverage of pregnant women. Children’s coverage has also expanded dramatically, consistent with the view that preventive and therapeutic services for children promote healthy development. Medicaid has become the largest payer of services for people with AIDS. Medicaid programs are gradually extending eligibility to disabled people who have some earnings; thus, Medicaid can be a source of comprehensive benefits that are rarely available through employers.

Medicaid has evolved with changes in the health care system. The largest shift came with the widespread adoption of managed care in the 1990s. Medicaid has also initiated changes in health care delivery, especially in long-term care, where it is the dominant purchaser of services. Medicaid has gradually increased spending on home and community-based services, shedding its original bias toward institutional care. In response to demands by people with disabilities for more control over their care, a handful of states are now experimenting with consumer-directed approaches, which give people with disabilities and their families far more control over hiring, training, scheduling, and supervising caregivers.¹³

Supporting experimentation. The current system gives states room to experiment, allowing them to function as the “laboratories of democracy.” States have revamped data systems and reimbursement methodologies, developed preferred drug programs, and extended drug coverage to older people.

States have several mechanisms at their disposal to extend coverage to low-income populations. These include Section 1115 research and demonstration waivers, SCHIP, family coverage provisions in the 1996 welfare reform legislation, and more recent Health Insurance Flexibility and Accountability (HIFA) waivers. Many states, albeit a minority, have used these mechanisms to greatly expand coverage, demonstrating considerable creativity and political will.¹⁴

States are experimenting with managed care models that reflect local circumstances and states’ political constraints.¹⁵ Examples of innovation include California’s two-plan model, which was designed as a way to contract with mainstream plans while protecting safety-net providers. Other states have found ways to maintain access to a wide range of providers by establishing a system of competing health maintenance organizations (HMOs) and primary care case managers. Florida’s experience with health plan marketing abuses led states to use enrollment brokers to help beneficiaries choose plans, one of Medicaid managed care’s most important innovations. Colorado, Washington, and Maryland have pioneered payment methodologies that protect plans from the financial risk of enrolling a large number of sick and disabled beneficiaries and thereby encourage more appropriate care for these vulnerable populations. Massachusetts has added to its primary care case management program many of the desirable and effective features of more comprehensive delivery systems.

States such as Oregon, Washington, and Wisconsin have used Medicaid waivers to develop community-based systems of long-term care for adults with disabilities.¹⁶ Minnesota, Massachusetts, Texas, and Wisconsin have attempted to integrate health and long-term care services for beneficiaries who are dually eligible for Medicare and Medicaid. These programs improve coordination of health and long-term care services and thus could reduce the use of inappropriate health care. States also have been leaders in involving consumers in managing Medicaid and state-funded services. As noted above, ever more states, including Washington, Wisconsin, Michigan, and California, allow disabled beneficiaries or their designees to choose and direct independent providers.

Weaknesses Of The Current System

Although the current system has accomplished much, several problems in the system make it hard to build upon and perhaps even to sustain. Serious weaknesses in its allocation of federal-state responsibilities limit the nation’s ability to meet the needs of low-income families. Millions of Americans remain uninsured, and variations in coverage across states are dramatic and difficult to justify. The current allocation of fiscal responsibility has created friction between states and the federal government, slowing progress toward meeting the needs of the uninsured. In addition, despite the range of policies states have adopted, the nation learns little from actual experimentation.

The unsolved problem of the uninsured. The current federal system’s most serious weakness is leaving more than forty million people without health insurance. To be fair, Medicaid and SCHIP were never designed to provide universal health insurance coverage. Yet the test of the federal system is not just whether its programs have met their goals but also whether shared responsibility engenders solutions to underlying problems. In this respect, the record is mixed.

Enrollment in Medicaid has increased from four million in 1966 to almost forty million in 2002.¹⁷ In the absence of Medicaid, the number of Americans without health insurance would surely be larger still. Yet the current system, patched together over many decades, ultimately leaves many people without the coverage they need. One-quarter of poor children remain uninsured—21.3 percent of those with incomes below twice the poverty level.¹⁸ The figures are far worse for adults, whose eligibility for public coverage is limited. Nearly half of all poor adults and 38 percent of adults with incomes at 100–200 percent of poverty are uninsured.

Unfortunately, the programs designed to give low-income people access to health care are administered or implemented in ways that fail to reach many of them.¹⁹ Most uninsured children—and all poor children, except for many immigrants—are eligible for public coverage. In addition, while the Medicaid benefit package is comprehensive, coverage does not always translate into access to high-quality services. Reports of barriers to access among Medicaid enrollees are common. Concern about the quality of care in nursing homes, often paid for by Medicaid, is substantial.²⁰ Thus, while the federal government provides a floor of eligibility and benefits, some Medicaid enrollees still go without needed services.

Excessive variation. Today’s system of federalism in health care leaves large variations in coverage across states. The nation’s major health programs are structured to encourage cross-state equity, with higher matching rates in Medicaid and SCHIP for poorer states and larger SCHIP budget allocations to states with more poor children. Despite these mechanisms, levels of coverage vary dramatically by state. Rates of uninsurance for children vary by a ratio of more than 3 to 1, with just over 8 percent uninsured in Minnesota and more than 25 percent uninsured in Texas.²¹ For people ages 18–65, uninsurance rates range from 11 percent in Minnesota to 27 percent in Texas. Uninsurance rates tend to mirror variations in employer-sponsored insurance: That is, states with high rates of employer coverage have low uninsurance rates, and vice versa. Accordingly, employer-sponsored insurance covers 75 percent of people under age sixty-five in Minnesota and 59 percent in Texas. Public coverage affects uninsurance rates as well, particularly for low-income people, and varies greatly among states. Eleven states have designed creative programs that extend coverage in significant ways to parents and adults without children, and another ten have expanded eligibility to parents to a substantial degree. Still, more than half have done little beyond meeting federal minimum standards.²² Between 1997 and 1999 Massachusetts covered 35 percent of its population with income below 200 percent of poverty, while Colorado covered only 15 percent.

States’ willingness to fund health care programs also varies dramatically. Medicaid spending on acute care services per low-income person under age sixty-five varies by roughly a factor of three among states—even more if only the states’ shares of spending are considered.²³ States with higher per capita income have a stronger base of employer coverage and spend more of their own revenue on Medicaid and SCHIP. States that spend less on these programs also spend less on other health programs—that is, they do not increase other health spending (for example, on public hospitals and clinics) to compensate.

As with many other issues, the question of how much variation across states and populations is appropriate is a matter of social values. If health insurance is considered what economists call a “normal” good, one would expect and accept that those with more resources spend more on it. Arguably, it is inefficient and unfair to force those with fewer resources to spend more on health insurance than they desire, leaving them with less to spend on other basic needs.

However, when polled, Americans consistently and overwhelmingly reject the notion that health care should be treated as any other economic good.²⁴ While there is disagreement about whether or not health care should be treated as a “right,” there is consensus that it should not be doled out solely based upon ability to pay. From a practical standpoint, providing health care to people without the means to pay helps them be productive members of society, reduces the incidence of public health crises that arise from untreated conditions, and reduces the inefficiencies of “cost shifting” in the health care system. From a moral perspective, providing health care to the needy reflects a fundamental respect for the worth of each human being and for individual autonomy. The degree of variation across states (and the even greater variation across local areas and among different subpopulations) is too large, since it means that very poor people in some states are far more likely to go without needed health care than elsewhere.

Spending growth stresses state budgets. The current allocation of responsibility between states and the federal government places a burden on states that is proving hard to sustain. Medicaid has grown steadily as a share of state spending and is now one of the largest components of state budgets.²⁵ Between 1990 and 2000 Medicaid spending per capita increased by 88 percent, adjusted for inflation, while total state spending net of inflation increased by 32 percent. By 2002 Medicaid spending accounted for 12 percent of state and local general fund expenditures.

Medicaid spending is projected to grow by about 8.5 percent per year over the next decade as a result of a combination of factors. Between rising health care costs and slower economic growth, employer coverage is likely to decline, pushing Medicaid and SCHIP enrollment up even without changes in eligibility standards. Hospital costs and prescription drug spending are likely to continue to increase fairly rapidly. Medicaid managed care is no longer providing the savings it did in the 1990s and will not provide states with the tools they need to constrain spending on acute care. Long-term care costs are also likely to rise due to the aging of the population, labor-force shortages, and efforts to improve nursing home quality. The U.S. Supreme Court’s Olmstead decision, in some circumstances, requires states to make available community-based services for people with disabilities, which will have serious fiscal implications for some states.

At the same time, states are under pressure to increase other expenditures—most notably education—by raising education standards, reducing class sizes, and raising teachers’ salaries.²⁶ Although growth in K–12 enrollment has slowed, the number of students in higher education is increasing. In several states, revenues are earmarked for education or state law mandates increases in education spending. As a result, Medicaid must compete for a piece of the pie that has already had some big bites taken out of it.

States are least able to afford Medicaid costs precisely when demands on the program are greatest. The burden Medicaid imposes on state budgets looms particularly large in FY 2004, when total state budget deficits are estimated at $60– $85 billion, or 13–18 percent of expenditures. The problems of Medicaid cost growth and cyclicality are most apparent now, when the economy is not performing well, but the fundamental cost problems are lasting. State fiscal resources are unlikely to grow quickly when the economy recovers, and Medicaid cost growth is projected at rates that exceed even healthy budget growth.

Fiscal friction. The current shared fiscal relationship has also resulted in states’ perpetually trying to shift a larger portion of their costs onto the federal government through various Medicaid accounting schemes, such as disproportionate-share hospital (DSH) and upper payment limit (UPL) programs.²⁷ Under these arrangements, Medicaid agencies have obtained money from providers through donations or taxes or from state and local government agencies through intergovernmental transfers. States have used these funds to make DSH or UPL payments to providers under Medicaid, thereby obtaining federal matching funds. State or locally generated funds have then been returned to the state or locality, along with some or all of the federal payments. Providers benefit when they retain some of the federal funds. Between 1990 and 1992 DSH spending increased from $1.4 billion to $17.5 billion, although it subsequently fell to $14.4 billion in 2000. UPL programs, which began in the mid-1990s, increased from $313 million in 1995 to $1.4 billion in 1998, then to $10 billion in 2000.

These programs pose several problems. First, while some payments are used as intended to meet the needs of providers serving substantial numbers of Medicaid and uninsured patients, some are primarily strategies for obtaining more federal funds for state government without chipping in more state funds. This tactic causes Medicaid expenditures to be overstated and effectively redefines the matching rate set in statute. Second, when providers retain federal dollars, state and local subsidies often are reduced or erode over time. Third, DSH and UPL payments are not distributed equitably. The amount of money coming into states depends on states’ creativity and willingness to exploit these mechanisms. For example, DSH spending in 1998 varied from 23 percent of Medicaid spending in Louisiana, 20 percent in Missouri, and 19 percent in South Carolina to less than 1 percent in Arkansas, Nebraska, and Wisconsin.

These financing mechanisms have eroded trust between the federal government and the states. The federal government has used laws and regulations to curtail DSH and UPL spending, but unraveling existing games is difficult because states come to rely upon these funds and are able to convince their congressional delegations to grandfather historical practices even as new rules are established for states going forward. Federal fear of state financial manipulation is a barrier to using Medicaid to expand health care for the uninsured.

States respond to charges that they are gaming the system with two arguments. First, they say that everything they have done is legal and part of a state plan approved by the federal government. Second, they argue, these initiatives were launched largely in response to new federal mandates to cover additional populations, federal court rulings requiring higher payments to hospitals and nursing homes, and other federal practices that limit states’ ability to implement program efficiencies, such as managed care and prescription drug formularies. These arguments might be true, but they in no way answer concerns that the Medicaid financing structure has lost its integrity.

One response to this distrust has been the approach taken in SCHIP—capping the federal appropriation so states have less incentive to play fiscal games and not creating an individual entitlement to benefits so costs are easier to control. Fiscal distrust is also one reason that some parties, including the Bush administration, endorse proposals to convert Medicaid into a block grant. Yet budget caps and block grants are draconian prescriptions for a fiscal management problem.

Limited benefits from experimentation. One potential benefit of variation in state policies and practices is that states can operate as laboratories of democracy. The idea is that states choose varied approaches and evaluate those that do and do not work, and then other states or the federal government makes better decisions based on lessons learned. Given the vast variation across states in administrative mechanisms, reimbursement methods, outreach and enrollment systems, organization of delivery systems, and other factors, the list of successful innovations that have been replicated by other states is disappointingly short. Most innovations are not evaluated, and other states do not learn from the experiments elsewhere or cannot implement innovations even when they are proved effective.

Take the example of long-term care. Few other states have duplicated the well-documented successes of such states as Oregon, Washington, and Wisconsin in changing the balance between institutional and home and community-based services. The primary barriers to following the leaders seem to be state concerns over costs, limited political appetite for change, lobbying power by nursing home providers, and bureaucratic inertia. These barriers suggest that while experimentation may yield valuable information, the current model of federalism does not easily translate results of experiments into practice. The federal government is attempting to speed dissemination through Systems Change Grants, but this provides money only for administration, not program costs.

Another example of experimentation is Medicaid managed care.²⁸ For political and cost containment reasons, states adopted this idea rapidly—perhaps too rapidly. They had little time to learn from each other, so opportunities for benefiting from experimentation were limited. Yet more learning has occurred over time and the story of Medicaid managed care, with its good and bad points, may represent the best case for state experimentation.

Innovations designed to expand health insurance coverage generally require spending money and redistributing resources—tasks that states find difficult. While a few leadership states have developed creative approaches to covering their populations, the failure of more states to follow suit most likely reflects the fact that reducing the number of people without health insurance is not as high on the agenda for them. Most states hold other goals as higher priorities: keeping taxes low, limiting the government’s involvement in the health care market, or funding other programs.

Which Way For Federalism In Health Care For The Low-Income Population?

Our review of the evidence suggests that the record of the current balance between state and federal responsibility is mixed. Some states are providing extensive health and long-term care coverage to the low-income population in innovative ways, but many others are just meeting basic requirements. States face varying burdens in financing coverage depending upon the income of their citizens and their base of employer-sponsored insurance. The current system is straining under a burden that will not disappear when the economy recovers. Policy debates focus on how to minimize the harm of program cuts, not how to expand coverage. Like so much of health policy, money or the lack of it is driving decisions.

Given this hard reality, it is time to rethink the allocation of responsibility between states and the federal government in coverage policy. We are guided by three objectives. First, we seek to fill in the glaring gaps in coverage for the poorest Americans. Second, we seek to reduce interstate variations in coverage, and to the extent that variations remain we consider them more defensible at moderate income levels than for people living in or near poverty. Finally, we seek a system that is stable and sustainable in its financing.

A fundamental conclusion from the evidence is that increasing the federal role in financing coverage is necessary if these three objectives are to be met. Providing health care to low-income people is expensive, and the federal government is better positioned than the states are for a task such as this, which requires redistributing income. The federal government is also better positioned to provide countercyclical financing. Thus, it is also apparent that converting the Medicaid program to a block grant, which shifts a larger share of the cost and risk to the states, runs directly counter to the objectives stated above.

Below we present two options that illustrate better allocations of responsibility between the states and the federal government. These options attempt to realign existing public responsibility, not fundamentally alter the role of government in health care. Either would improve the current system, but neither is designed to yield universal coverage. The options are presented in enough detail to illustrate distinct approaches, but many other features would need to be defined to make them complete proposals (Exhibit 1).

Option 1: improve Medicaid and SCHIP. Recognizing the strengths and weaknesses of Medicaid and SCHIP, one option is for the federal government to redefine the base of Medicaid and SCHIP coverage by simplifying and raising eligibility criteria, while giving states stronger incentives to go further.²⁹

Income eligibility could be set at 200 percent of poverty for children and pregnant women and 100 percent of poverty for all other adults. These standards would replace today’s multiple eligibility categories. The federal entitlement to benefits would remain, as would the federal definition of comprehensive benefits for eligible people. States would be subject to federal standards regarding payment to providers. Minimum payment levels would be established to ensure that enrollees have access to services, while maximum payment levels would protect the program’s fiscal integrity.

The current melange of Medicaid optional eligibility groups and optional services and SCHIP would be replaced by a simplified set of options that states could adopt to extend coverage beyond the federally defined core. States would have substantial leeway in structuring the options—they could impose copayments, premiums, limits on benefits, and limits on the number of enrollees, all within federal guidelines designed to protect low-income beneficiaries. States also could extend coverage as far up the income scale as they wish, and they could add services. They could expand eligibility to targeted groups, permitting people with high medical costs or in need of long-term care to receive or buy into Medicaid coverage. Federal matching funds would be available to cover the costs of these options.

States would receive the same federal matching rate for the core program and for all optional populations and services. This rate would be about 15 percent higher than the current rate under Medicaid, but below the SCHIP program’s 30 percent enhancement—appropriate since the enhancement applies to the entire program, not just to the additional enrollees, as in SCHIP. Setting the same matching rate for all populations and services improves fiscal integrity by eliminating incentives for states to enroll people in one program rather than another.

The federal government would take over all financial responsibility for limitations in Medicare’s acute care benefit package. This would include prescription drug costs and the full cost of premiums and cost sharing for Medicaid enrollees who are also enrolled in Medicare. Medicare also would assume the costs states now bear for the Medicare savings programs (Qualified Medicare Beneficiaries, Specified Low-Income Medicare Beneficiaries, and Qualified Individuals).

A new waiver process would allow states to experiment with programs that go beyond the federally defined options. Experiments would be allowed in delivery systems, financing systems, and benefit design. The experiments would not be expected to be cost-neutral—in fact, federal funds would be available on a competitive basis to test new approaches. Waivers would be for a limited time and would be permitted only for proposals that include rigorous evaluation of results.

The Medicaid home and community-based services waiver system would be replaced by a flexible program modeled on SCHIP, with the same 15 percent higher enhanced matching rate, substantial state flexibility in service design, and new federal funds. Such a program reflects the need for new investment to encourage the move away from institutionalization.

The DSH program would be eliminated. While the program has been an important source of funds for safety-net providers, it has come at the cost of program integrity. Fiscal integrity in a matching-grant program can be ensured only if the program has a defined population, set of benefits, and payment structure. DSH has provided an opportunity for fiscal games because it operates outside that structure. The federal government should reassess all remaining uncompensated care and consider developing a new grant program to meet those needs.

Option 2: a new federal program. A second option draws upon the lessons of Medicare.³⁰ Medicare is financed and administered entirely by the federal government. Eligibility and benefits are uniform around the country. Medicare financing is stable because the federal government has the advantage of being able to incur budget deficits during economic downturns. Yet Medicare can be rigid and is less able than Medicaid is to keep up with changes in the health care system.³¹ For example, Medicare still lacks a prescription drug benefit and protection against catastrophic costs, despite years of debate.

Recognizing these strengths and weaknesses, a second option is to shift responsibility for a large portion of the health insurance safety net for low-income people to the federal government while allowing states to go further. The federal government would assume full responsibility for covering the acute care costs of the same group described above: children and pregnant women with incomes up to twice the poverty level, and all other adults with incomes up to the poverty level. The federal government also would finance and run this program in a nationally uniform manner, as it does Medicare. The new program would operate as an entitlement to individuals, with a federally defined, comprehensive package of benefits and a national system for paying providers.

States could still cover populations beyond the federally defined minimum and could use the fiscal relief they would obtain from the new federal program to do so. The same would be true if states wanted to provide their low-income populations with benefits that are not covered by the new federal program. Presumably, many states would accept the base of federally provided acute care coverage for their low-income populations as sufficient and go no further. However, some states—particularly those with current programs that extend to people with higher incomes—would continue to operate programs on top of the federal base.

The existing Medicaid program with its matching structure would remain in place for long-term care.³² As in the first option, the federal government would assume full financial responsibility for acute care services in the Medicare benefit package for people enrolled in both Medicare and Medicaid. It would also assume all state costs associated with the Medicaid savings programs. The new program for home and community-based services described above, with an enhanced match, would also be included in this option. The DSH program would be eliminated, and possibly reconstituted, as in the first plan.

Comparing the approaches. Both of the approaches described above create a sturdy federal floor of coverage for low-income populations with the possibility of expansion. Both would provide new coverage to many people who are now uninsured. And both expand the federal role in financing, relative to the states.

The first proposal provides generous matching funds for expansions that states can design with the sort of flexibility they now have in SCHIP. The second proposal leaves expansions entirely in the hands of the states with no federal rules (and no federal matching funds). Under either proposal, the federal government could take additional steps. It could layer tax credits on top of this base of coverage, could provide additional funding to states in the form of block or matching grants, or could raise the eligibility levels for the basic program.

The approaches differ in the amount of state variation in how care is delivered to low-income populations. The first option essentially retains the state-administered system, although to improve access and quality, states would be subject to federal oversight regarding provider payment rates. On such matters as patient education, provider relations, use of managed care, and access to the enrollment system, states would continue to vary. The second approach puts the delivery system under federal control, bringing about greater uniformity. Along with federal control and responsibility would likely come greater political pressure on the federal government to provide enough funds to support adequate provider payment rates. A federal commitment to administration would allow for more rapid dissemination of innovations, such as risk-adjusted payments.

The fiscal implications of these options are complex, and we have not fully simulated them. Both proposals impose substantial new costs on the federal government, although savings to states offset much of this. Option 2 has a larger federal cost because assuming the full cost of the core population is more expensive than increasing the matching rate even if several states expand coverage. Altering the specifics of the options would change the federal cost, the savings for states, and the allocation of those savings across states.

Under either option, states that have taken more steps to provide coverage (high-coverage states) would receive more fiscal relief than states that have taken fewer steps (low-coverage states). In option 1, high-coverage states receive an enhanced match on a larger base of spending than low-coverage states receive; low-coverage states must pay a portion of the cost of covering the new populations made eligible in this option. In option 2, the federal government assumes the full cost of more people in high-coverage states than in low-coverage states.

This logic applies to all aspects of coverage. A state that now extends coverage further up the income scale, covers more of the Medicare population, or pays its providers at a higher rate will receive more fiscal relief from either option than a state that is less generous would receive. The one exception is the DSH program—its elimination will impose the greatest burden on states that have taken maximum advantage of the program and retained the funds for their own use.³³

By contrast, low-coverage states would receive the largest benefits associated with expanding health insurance coverage to their low-income populations. These benefits include the reduced burden of untreated illness and disease, the economic benefits of new federal funds supporting coverage for people who are now uninsured, and the easing of financial pressure on safety-net providers.

The first approach retains the cyclical effects of Medicaid on state budgets, but at a lower and presumably more manageable base. The second smoothes out cycles in states’ budgets by eliminating state responsibility for the eligibility group that grows during economic downturns when state budgets are under pressure.

Conclusion

The current balance of federal and state responsibilities for health insurance coverage has achieved a great deal. But it has failed to insure forty million Americans, it expands coverage in small steps only, and under it no state, much less the nation as a whole, has developed and implemented a comprehensive approach to covering the uninsured. The states are now largely playing defense against an eroding employer base of coverage and a fiscal future in which Medicaid expenditures are likely to grow faster than revenues. There is no reason to believe that the current federal structure will ever yield universal coverage or even come close. Indeed, the late 1990s may turn out to have been the high-water mark for health insurance coverage within the parameters of the current system.

Decades of experience show that major progress in covering the uninsured will require a substantial new investment by the federal government. Heavy reliance upon state financing, on top of large differences in employer coverage, is the primary reason for dramatic interstate variations in coverage and, ultimately, for the large gaps in coverage that remain. While states have substantial financial capacity, that capacity is more limited than that of the federal government; it falls with economic downturns, at precisely the same time that health care needs increase; and its funding sources are less progressive than the federal government’s are, making it harder to redistribute funds to services for low-income families.

Yet the state role in financing serves important purposes. It is impossible to completely separate financing and administration, so the benefits of state administration can only be gained if there is some state financing. State funding, even when highly leveraged as in SCHIP, gives states ownership that encourages good performance. Finally, state funding reduces the federal government’s fiscal burden, making expansions more likely.

The challenge for federalism is to devise a financing role for states that continues to prompt administrative innovation while minimizing the inequities that arise when they must bear an overwhelming fiscal burden. Our two approaches take different paths toward striking this balance. The first retains a large role for states; the second diminishes that role. Both would maximize the base of coverage and provide opportunities for moving beyond it while encouraging innovation and experimentation. Neither approach eliminates interstate variation or the inequities it implies. However, since both start from a higher coverage base, the inequities are more defensible than are those in the current system.

The nation should take advantage of its diversity by encouraging true experimentation and learning across states. This, the greatest potential benefit of interstate variation, can be realized only if state policies are documented, examined, and evaluated and the results disseminated. Too many opportunities for learning are squandered in the current environment, where state flexibility is valued for its own sake rather than for its contribution to better public policy.

Proposals that require substantial new federal funding could be unrealistic in the current fiscal climate. However, without additional federal funds, the states are unlikely to sustain their current coverage levels, let alone increase them. Changing the balance of federalism involves risk. However, there is no other way to return stability and sustainability to our system so that we can build upon it to greatly reduce the number of uninsured Americans.

This paper was prepared with the financial support of the Robert Wood Johnson Foundation and the other funders of the Assessing the New Federalism project: the Annie E. Casey, W.K. Kellogg, John D. and Catherine T. MacArthur, and Ford Foundations. The views expressed are those of the authors and do not necessarily reflect those of the Urban Institute, its board, or its sponsors. The authors are grateful for the valuable comments provided by two anonymous referees on an earlier draft.

NOTES

1. R. Bovbjerg, J. Wiener, and M. Houseman, “State and Federal Roles in Health Care: Rationales for Allocating Responsibilities,” in Federalism and Health Policy, ed. J. Holahan, A. Weil, and J. Wiener (Washington: Urban Institute Press, 2003), 25–58.
2. Authors’ calculations from Congressional Budget Office 2003 baseline (available from the authors by e-mail to jholahan{at}ui.urban.org).
3. Ibid.
4. Authors’ calculations from the 2002 Current Population Survey.
5. L. Dubay and G. Kenney, “Health Care Access and Use among Low-Income Children: Who Fares Best?” Health Affairs (Jan/Feb 2001): 112–121; P.W. Newacheck et al., “Health Insurance and Access to Primary Care for Children,” New England Journal of Medicine (19 February 1998): 513–519; and M.L. Rosenbach, “The Impact of Medicaid on Physician Use by Low-Income Children,” American Journal of Public Health (September 1989): 1220–1226.
6. A. Davidoff et al., “Children Eligible for Medicaid but Not Enrolled: Health Status, Access to Care, and Implications for Medicaid Enrollment,” Inquiry (Summer 2000): 203–218.
7. C.M. Cowles, Nursing Home Statistical Yearbook 2001 (Montgomery Village, Md.: Cowles Research Group, 2002); and Centers for Medicare and Medicaid Services, “Table 3: National Health Expenditures by Sources of Funds and Types of Expenditures, Select Calendar Years 1996–2001,”
www.cms.hhs.gov/statistics/nhe/historical/t3.asp (20 June 2003).
8. Authors’ calculations from HCFA 2082 and HCFA 64 data.
9. J. Holahan, “Health Status and the Cost of Expanding Insurance Coverage,” Health Affairs (Nov/Dec 2001): 279–286.
10. A. Weil and I. Hill, “The State Children’s Health Insurance Program: A New Approach to Federalism,” in Federalism and Health Policy, ed. Holahan et al., 293–324.
11. Institute of Medicine, Coverage Matters: Insurance and Health Care (Washington: National Academies Press, 2001); and IOM, Care without Coverage: Too Little, Too Late (Washington: National Academies Press, 2002).
12. Authors’ calculations from the 1999 National Survey of American Families” (Washington: Urban Institute, 2001); and G. Kenney, L. Dubay, and J. Haley, Health Insurance, Access, and Health Status of Children, Snapshots of America’s Families II: A View of the Nation and Thirteen States from the National Survey of America’s Families (Washington: Urban Institute Press, April 2001).
13. J. Wiener and J. Tilley, “Long-Term Care: Can the States Be the Engine of Reform?” in Federalism and Health Policy, ed. Holahan et al., 249–292.
14. J. Holahan and M. Pohl, “Leaders and Laggards in State Coverage Expansions,” in Federalism and Health Policy, ed. Holahan et al., 179–214.
15. R. Hurley and S. Zuckerman, “Medicaid Managed Care: State Flexibility in Action,” in Federalism and Health Policy, ed. Holahan et al., 215–248.
16. Wiener and Tilley, “Long-Term Care.”
17. Authors’ calculations from CBO baseline, 2003, adjusted by the ratio of full-year enrollees to average monthly enrollees from the HCFA 2082 data.
18. Authors’ calculations from the 2002 CPS.
19. L. Dubay, G. Kenney, and J. Haley, “Children’s Participation in Medicaid and SCHIP: Early in the SCHIP Era,” Assessing the New Federalism Policy Brief B-40 (Washington: Urban Institute, 2002).
20. IOM, Coverage Matters; IOM, Care without Coverage; and G.S. Wunderlich and P.O. Kohler, eds., Improving the Quality of Long-Term Care (Washington: National Academies Press, 2001).
21. All insurance data in this paragraph are from J. Holahan, “Variation in Health Insurance Coverage and Medical Expenditures: How Much Is Too Much?” in Federalism and Health Policy, ed. Holahan et al., 111–144.
22. Holahan and Pohl, “Leaders and Laggards.”
23. Holahan, “Variation in Health Insurance Coverage.”
24. Henry J. Kaiser Family Foundation and NewsHour with Jim Lehrer, “Health Care Should Be Provided Equally to Everyone,” Kaiser Health Poll Report, January/February 2003, www.kff.org/healthpollreport/templates/detail.php?page=6&feature=feature3 (20 June 2003.)
25. D.J. Boyd, “Health Care within the Larger State Budget,” in Federalism and Health Policy, ed. Holahan et al., 59–110.
26. J. Holahan et al., “The State Fiscal Crisis and Medicaid: Will Health Programs Be Major Budget Targets?” (Washington: Kaiser Commission on Medicaid and the Uninsured, January 2003).
27. T.A. Coughlin and S. Zuckerman, “States’ Strategies for Tapping Federal Revenues: Implications and Consequences of Medicaid Maximization,” in Federalism and Health Policy, ed. Holahan et al., 145–178.
28. Hurley and Zuckerman, “Medicaid Managed Care.”
29. An approach that requires states to meet higher standards as a mechanism for achieving national objectives is consistent with recent trends in insurance regulation. R. Bovbjerg, “A Different Approach to Federalism: Growing National Authority in Health Insurance Regulation,” in Federalism and Health Policy, ed. Holahan et al., 361–398.
30. M. Moon, “Making Medicaid a National Program: Medicare as a Model,” in Federalism and Health Policy, ed. Holahan et al., 325–360.
31. L. Brown and M. Sparer, “Poor Program’s Progress: The Unanticipated Politics of Medicaid Policy,” Health Affairs (Jan/Feb 2003): 31–44.
32. We propose to retain the current Medicaid matching design for long-term care even though the financing system for long-term care in the United States is in need of reform. Given this paper’s focus on inequities across states in coverage for the low-income population, we have chosen not to attempt to redesign the long-term care system as well.
33. The existing variable match rate structure also makes option 2 a better deal for high-income states than for low-income states. When the federal government assumes the full cost of a population, the high-income state with a 50 percent match rate saves 50 percent of the cost. A low-income state, with a 75 percent match rate, saves only 25 percent of the cost, because the other 75 percent was already borne by the federal government. Since high coverage correlates with high income, option 2 provides an added bonus to high-coverage states.

John Holahan (jholahan{at}ui.urban.org) directs the Health Policy Center at the Urban Institute in Washington, D.C.; Alan Weil directs the Assessing the New Federalism project there. Joshua Wiener was a principal research associate in the Health Policy Center and now is director of Aging, Disability, and Long-term Care at RTI International.

©2003 Project HOPE–The People-to-People Health Foundation, Inc.