AMERICAN ENTERPRISE INSTITUTE
From the book "A Safety Net That Works," edited by Robert Doar
In 1965, the authors of Medicaid thought they were creating a program that would provide federal structure, uniformity, and some funding for the many state programs that were already providing relatively inexpensive "indigent care" services to low-income households. They did not foresee the transformation they were setting in motion. Medicaid has grown into the largest health care program in the country by enrollment, with 66 million participants and with annual federal and state costs of more than $550 billion. It is by far our nation's largest program serving low-income households.1
Medicaid's purpose is to provide access to medically necessary health care to persons who, because of limited resources and lack of health insurance coverage, have less capacity to secure care for themselves. The definition of the services paid for by Medicaid has expanded over the years through legislative changes, administrative decisions, and court cases. Today, the program pays for hospital and physician care, prescription drugs, screening and preventive services, and long-term care services and supports for elderly and disabled persons.
Medicaid spending has increased rapidly nearly every year since the program was enacted, creating significant pressure in federal and state budgets. Medicaid, along with other major entitlement programs, is a primary reason for today's large federal budget deficit and for the massive deficits projected for the coming decades. Its growth has forced federal and state policymakers to limit spending on other priorities, including many programs that also serve low-income households.
There are several reasons for Medicaid's growth, including eligibility liberalizations enacted by Congress and cost trends that have rapidly increased health spending for all insurance plans. But Medicaid's original federal-state design is also an important factor, especially the "matching" system through which the federal government pays for more than half of state Medicaid spending.
Medicaid is a shared federal-state responsibility. The program was created in federal law, but it is largely administered by the states. States pay for the medical and social services that enrollees use and then receive funds from the federal government to partially pay for the program's costs.
Most federal funding provided to the states under Medicaid is determined by a standard state-specific formula called the federal medical assistance percentage (FMAP). The FMAP is based on the ratio of per capita income in the states relative to the national average. States with lower per capita incomes get higher FMAPs. For instance, in 2016, the FMAP for Arkansas is 70 percent, meaning that for every $1.00 the state spends on standard Medicaid benefits, the federal government pays for $0.70 of the bill. Medicaid law puts a floor on the standard FMAP at 50 percent, so states with higher per capita incomes, such as California, Massachusetts, and New York, which might otherwise have FMAPs below 50 percent, get half their Medicaid bill paid for by the federal government.2
Special matching rates apply to different subparts of Medicaid. For instance, the federal government generally applies a uniform FMAP of 50 percent for all state administrative expenses. In addition, the Affordable Care Act (ACA) allows states to expand Medicaid eligibility to all households with incomes below 138 percent of the federal poverty line. For three years (2014 through 2016), the federal government is paying 100 percent of the costs for newly eligible Medicaid enrollees. After 2016, the FMAP for this "expansion population" will be 90 percent.3 On average, the federal government pays for between 62 and 64 percent of total Medicaid spending, depending on the year.4
The Medicaid FMAP is the fundamental flaw in the program's current design and the main reason it is so costly. States can initiate new spending in Medicaid—spending that often will boost economic activity in the state—and federal taxpayers pay for at least half the cost. At the same time, savings from state-initiated Medicaid-spending cuts are also shared with federal taxpayers.
For instance, in a state where the FMAP is 60 percent, the governor and state legislators face the unattractive prospect of keeping only $1.00 of every $2.50 in Medicaid savings they can identify and implement. The other $1.50 goes to the federal treasury. Put another way, governors and state legislators are reluctant to impose $2.50 in budgetary pain for a $1.00 gain to their bottom line.
Powerful health-sector interests, including hospital systems, nursing-home operators, physician groups, and insurers, also make it difficult for states to restrict total funding on the program, although payment rates for individual services can be quite low.
Medicaid's current federal-state design also undermines political accountability. Neither the federal government nor the states are fully in charge. As a result, each side has tended to blame the other for the program's shortcomings, and neither believes it has sufficient power to unilaterally impose effective reforms.
Federal Medicaid spending has no upper limit. The federal government continues to make matching payments to the states as long as the spending is within the boundaries of allowable Medicaid expenditures.
The only check on runaway expenditures is the state contribution to the program. The FMAP establishes both the federal government's matching payments for Medicaid costs and the state share (1 – FMAP). The requirement of a state contribution for every extra dollar of spending should serve as a disincentive for wasteful and excessive expenditures.
Unfortunately, the discipline that comes from requiring a state contribution to program costs has been partially undermined by schemes that have effectively lowered states' shares of Medicaid expenditures. These schemes generally involve grossly inflating state payment rates to certain narrowly defined sets of providers of services. These high payments draw additional federal matching funds. The state then imposes a special, narrowly drawn tax on those same providers of services, which has the effect of reimbursing the state for its presumed share of the higher Medicaid payment rates. The net effect is more federal spending on Medicaid and no additional burden on taxpayers in the state.5
States can also reduce their budgetary costs if they are able to move programs traditionally financed with state-only funds under the Medicaid programmatic umbrella, thus drawing partial federal support. Not surprisingly, this has been a common practice in states for many years as well.6
Congress enacted a series of program liberalizations in the 1980s and 1990s that fueled program enrollment growth. These changes extended public insurance coverage to tens of millions of people nationwide, mainly women and children.
But it was not just federal action that expanded the program. The pressures that pushed Congress to broaden Medicaid's reach also pushed states to adopt some of the optional coverage expansions allowed in federal law. The FMAP paved the way for these state decisions by lowering the cost barrier to higher program enrollment.
The combined effect of these factors has been a remarkable and uninterrupted increase in the percentage of the US population participating in Medicaid. As shown in Figure 1, in 1980, national Medicaid enrollment was 8.7 percent of the US population. By 2013, it had risen to 18.3 percent.
Medicaid-spending growth over the past 50 years is an important reason the federal government is under fiscal stress today. Moreover, continued growth in program spending is expected to contribute substantially to large deficits and growing federal debt over the coming decades.
As shown in Figure 2, federal spending on Medicaid in 1972 was only 0.4 percent of gross domestic product (GDP). Today, it is 2 percent of GDP.
The Congressional Budget Office expects spending on the program to continue to rise rapidly. Current long-term projections show combined spending for Medicaid, the Children's Health Insurance Program, and the subsidies for health insurance provided under the ACA increasing from about 2.2 percent of GDP today to 3.2 percent in 2050.7
There are different program beneficiaries within Medicaid, with different service needs. Specifically, Medicaid provides insurance to lower-income households needing access to traditional medical services, and it provides assistance to the severely disabled and the frail elderly in need of nursing-home care or other support services to help them with their activities of daily living.
As shown in Figure 3, Medicaid eligibility is dominated by children and nondisabled adults, including pregnant women. In 2015, these beneficiaries made up 79 percent of all persons enrolled in the program. The elderly and the disabled comprised just 21 percent of enrollment.
The distribution of Medicaid spending is a different matter. In 2015, more than half of all federal Medicaid spending went toward services used by the elderly or the disabled, and less than half went to services used by children and nondisabled adults. The higher concentration of spending on the elderly and disabled is due to the much more expensive and intensive services needed for these populations. In 2012, the federal and state governments spent, on average, $17,848 for every disabled Medicaid enrollee, but only $2,679 for every child enrolled in the program.8
Medicaid's cumbersome rules and federal mandates have led many states to seek more flexibility in running the program through program waivers. These waivers, authorized by Section 1115 of the Social Security Act, allow the Department of Health and Human Services (HHS) to set aside certain Medicaid requirements as part of "demonstrations," or tests, of new approaches to providing health benefits to lower-income households. States also seek waivers to help them manage their programs outside of the normal and lengthy federal constraints that otherwise apply to the program.
The federal government has approved many state waiver requests, but not all that have been submitted. Some have been approved only after a lengthy and contentious negotiation between federal and state officials over the content of the waiver program. According to the HHS online database, 407 current waiver programs of all types have been approved by HHS and are in operation, of which 40 are existing Section 1115 waivers.9 Twenty-six waiver requests are now pending at HHS.
The most important consideration in any significant waiver request is federal funding. More specifically, the waiver requests from states are assessed by federal officials to determine whether they are budget neutral, meaning whether the federal government would pay more to the state under the waiver than it would without it. Not surprisingly, this is the source of frequent disagreements between states and the federal government.
The concept of budget neutrality in Medicaid waiver assessments dates back to the early 1980s. Before that time, HHS could approve state Medicaid requests without regard to the waivers' impact on federal spending. The statute never mentions budget neutrality as a requirement for federal approval of the demonstration programs. As HHS approved more and more requests by the states to waive certain Medicaid statutory provisions, the White House Office of Management and Budget (OMB) became concerned that the state programs being approved by HHS were actually costing the federal government substantially more than the regular Medicaid program.
In 1983, the OMB and HHS came to an agreement that all future Section 1115 waivers must be budget neutral to the federal government over the life of the demonstration.10 This agreement also gave the OMB the authority to reject demonstration requests that did not meet the test of neutrality. Since then, the OMB has played a central role in virtually all federal-state negotiations over significant waiver requests.
The Medicaid program struggles to provide its enrollees with sufficient access to care. The easiest way for states to slow cost growth has been to limit payment-rate increases for hospitals and physicians, so much so that payments are now well below what private insurers and the Medicare program pay for the same services. As shown in Figure 4, in 2013, Medicaid's payment rates were only 61 percent of what private insurers paid for the same inpatient hospital services. Similarly, Medicaid's payment rates for physicians were only 58 percent of what private insurers paid on behalf of their enrollees.11
Physicians and other service providers respond to these low payment rates by explicitly limiting the number of Medicaid patients they will see or by employing other business practices, such as the location of their offices and facilities, to cater to patients with higher-paying commercial insurance. Medicaid's inadequate network of willing providers makes it difficult for some patients to access care when they need it, or from the most qualified practitioners.
Several academic studies have documented the problems that occur when access to care is inadequate. A 2010 study published in the Annals of Surgery found that, after controlling for important demographic and risk factors, Medicaid patients fared much worse than their private-insurance counterparts in terms of outcomes from major surgical interventions. The study examined nearly 900,000 cases from a large patient-care database compiled from hospitals nationwide. The authors found that patients with Medicaid coverage were much more likely to die from surgical interventions than the privately insured and that Medicaid patients even had higher mortality rates than those who were uninsured.12
Medicaid's low payment rates for services would seem to contradict the program's high and rapidly rising costs. But Medicaid's costs have been driven mainly by large increases in program enrollment and a growth in the volume and intensity of the services provided to each patient. States have responded by limiting what they pay for each service, but that has only served to erode the quality of care provided to each enrollee without keeping spending growth to acceptable levels.
Medicaid reform needs to start with a restructured relationship between the federal government and the states and a reformed approach to financing the program that does not rely on open-ended federal matching funds.
Medicaid funding should be divided into two funding streams, one for the disabled and elderly and the other for everyone else. The services provided to these two populations vary greatly, as do the state approaches to administering Medicaid for them. Breaking Medicaid funding up into two streams would begin the process of implementing different kinds of reforms for the different populations of Medicaid participants.
The other important structural change would be the switch to some form of fixed federal funding to states. The federal government would continue to heavily support the Medicaid program, but the commitment would have a limit, which would give states a strong incentive to manage the program for efficiency and cost control.
One approach would be a block grant. Under a block grant, the federal government would make fixed, aggregate payments to the states based on historical spending patterns. Cost overruns at the state level would require the state to find additional resources within the state budget. Conversely, states that were able to control costs would enjoy the full benefits of their efforts. The block grant would not be reduced when states found ways to root out waste and inefficiency from the program.
The key issue in converting to a block grant is establishing the basis by which the federal government will make payments to states. One option would be to examine the federal government's historical Medicaid-spending levels in various states over a particular number of preceding years. The block grant's first year could then be calculated as the average of federal Medicaid spending in the state per year during that time, inflated to the year in question by the national Medicaid-spending growth rate.
Once the first year is settled, the question becomes how to inflate the federal Medicaid block-grant amounts in future years. The indexing options include using the consumer price index, which historically is well below medical inflation; the growth rate of the national economy as measured by GDP; or perhaps a measure of national or regional health spending growth. The decision on indexing is highly consequential because alternative approaches can result in large differences in federal spending over time. If the block grant is pursued in part to help ease the nation's severe, long-term budgetary challenges, then indexing the block-grant amounts to something below the historical rate of growth for Medicaid can produce significant savings estimates, especially over the long term.
Opponents of the block-grant concept argue that it will necessarily reduce services for vulnerable populations. But that is far from certain. The current program, with open-ended federal matching payments, provides strong incentives to states to move as much spending as possible under the Medicaid umbrella. With a block grant, states would have strong incentives to eliminate waste without undermining coverage for those who truly need it.
In 1996, similar arguments were made about the block granting of welfare funding, with predictions that it would lead to significant hardship for the program's enrollees. Instead, states reviewed who was on the cash assistance program and quickly found that many of them were capable of entering the workforce and improving their household incomes from wages instead of government assistance. By 2000, the cash welfare rolls had fallen by about half, even as the population in the bottom fifth of the income distribution experienced substantial gains in their real incomes.13
Health coverage is more complicated than cash welfare, but there is every reason to expect that substantial inefficiency exists in Medicaid and that a block grant would provide the incentive to find and eliminate it. Among other things, states would seek to remove from the program persons who are erroneously enrolled today. The Centers for Medicare and Medicaid Services reports that the program has an improper payment rate of 9.4 percent and that much of the wasteful spending is associated with persons who are ineligible for Medicaid enrollment.14 Further, states would seek to design their programs so that households that could enroll in more privately financed coverage, including employer plans, do so. Taking steps to minimize erroneous or unnecessary Medicaid spending would allow states to concentrate on ensuring high-quality care for those who remain on the program.
Still, concerns about the effect that a block grant might have on health services for the vulnerable has led to proposals that mitigate against some of the financial risks a block grant would entail. The most prominent example of such a proposal is per capita caps.
Under per capita caps, the federal government would establish for each state a per-person payment for each of the main eligibility categories in the Medicaid program: the elderly, the blind and disabled, nondisabled adults, and children. The federal government would then make payments to the states based on the number of Medicaid enrollees in each of these categories. The per capita payment would be based on historical spending rates for the various categories of beneficiaries in each state and, again, would be indexed to a predetermined growth rate.15
Per capita caps in Medicaid would have the same advantages as a block grant in that the states would have strong incentives to use federal funding wisely. The amount of the federal payment per person would be the same regardless of how much the state spends on each enrollee. The only difference with the block grant is that the states would not be at risk for increased Medicaid enrollment because the federal government's payments to the states would be made on a per-enrollee basis, including for enrollees who might not have been expected to sign up and would have been excluded from the block-grant formula. This could be important in times of slow economic growth or during a recession, when Medicaid enrollment typically surges.
Perhaps most important, per capita caps have enjoyed bipartisan support in the past. In 1995 and 1996, the Clinton administration proposed Medicaid per capita caps as part of a larger balanced-budget plan. That proposal was explicitly endorsed by 46 Senate Democrats in a letter to the president in December 1995.16
The per capita allotment approach to federal funding allows for more enrollment flexibility and perhaps more bipartisan support. For these reasons, policymakers should make implementation of per capita federal payments to states the centerpiece of a Medicaid reform plan.
The Medicaid program is an important component of the nation's safety net. The program serves tens of millions of people who would otherwise struggle to secure access to needed medical care and long-term care services. But it is no longer possible to ignore the immense budgetary pressures the program has created for the federal government and the states. Indeed, state governors and legislators, on a bipartisan basis, have been calling for fundamental Medicaid reform for many years because they see the program crowding out their ability to address other important state priorities.
What is needed is a reworking of the federal-state relationship concerning the program's financing and governance. The federal government should step back from micromanaging every aspect of program administration and instead provide the overall framework and a predictable funding stream to the states. States should pay for a portion of the program's cost. They should also be given the freedom and responsibility to make the major decisions over program design and be held accountable for the results.
To read the complete collection of essays, please click here.