November 13, 2013

Mandatory SpendingOption 1

Function 550 - Health

Impose Caps on Federal Spending for Medicaid

(Billions of dollars) 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2014-2018 2014-2023
  Caps on Overall Spending
Change in Mandatory Outlays                        
  Base growth of caps on the CPI-U 0 0 -10 -25 -36 -48 -61 -74 -90 -106 -71 -450
  Base growth of caps on per capita NHE 0 0 * -9 -13 -14 -15 -16 -18 -20 -22 -105
 
  Caps on Spending per Enrollee
Change in Mandatory Outlays                        
  Base growth of caps on the CPI-U 0 0 -28 -45 -57 -69 -81 -94 -108 -124 -130 -606
  Base growth of caps on per capita NHE 0 0 -16 -28 -34 -37 -38 -40 -43 -46 -78 -282

Notes: This option would take effect in October 2015.

* = between -$500 million and zero; CPI-U = consumer price index for all urban consumers; NHE = national health expenditures.

Overview of the Issue

The Medicaid program covers acute and long-term care for low-income families with dependent children, elderly people, people with disabilities, and, at states’ option starting in January 2014, all nonelderly adults with family income up to 138 percent of the federal poverty guidelines. Under current law, the federal and state governments share in the administration of Medicaid. The federal government is responsible for establishing broad statutory, regulatory, and administrative parameters for state Medicaid programs to operate within, including determining which individuals and medical services must be covered and which may be covered at a state’s discretion. The federal government also monitors states’ compliance with the parameters it sets. For their part, states decide which of the eligibility and service options to adopt and are responsible for administering the daily operations of the program. Because of the discretion that states have, their Medicaid programs vary widely in terms of the optional eligibility groups and services covered, the rates used for paying health care providers, and other program elements.

Medicaid is also financed jointly by the federal and state governments; in 2012, states received $251 billion from the federal government for Medicaid and also spent $181 billion of their own funds on the program. Under current law, almost all of the federal funding is provided on an open-ended basis, meaning that increases in the number of enrollees or in costs per enrollee automatically generate more federal payments to states. For people now enrolled in Medicaid, the federal government pays about 57 percent of program costs, on average (that share varies by state from 50 percent to a current high of 73 percent). For the optional Medicaid expansion beginning in 2014, the federal share of costs will start at 100 percent in all states and phase down to 90 percent by 2020.

Spending on the Medicaid program has grown rapidly over time, consuming an increasing share of the federal budget and representing a growing percentage of gross domestic product (GDP)—trends that the Congressional Budget Office projects will continue in the future. Over the past 20 years, federal Medicaid spending has risen at an average rate of a little over 6 percent a year, because of general growth in health care costs, mandatory and optional expansions of program eligibility and covered services, and states’ efforts to increase federal payments for Medicaid. CBO expects federal Medicaid spending to grow at a higher rate over the next decade, an average of 8 percent a year, largely because of the optional coverage expansion authorized by the Affordable Care Act (in which many, though not all, states are expected to participate). By comparison, GDP is projected to increase by about 5 percent a year over the next decade, and general inflation is expected to average about 2 percent a year. Under current law, CBO projects, Medicaid will go from accounting for 8 percent of the federal government’s noninterest spending in 2013 to accounting for 11 percent in 2023.

Lawmakers could make various structural changes to Medicaid to decrease federal spending for the program. Those changes include reducing the scope of covered services, eliminating eligibility categories, repealing the Medicaid expansion due to start in 2014, lowering the federal government’s share of total Medicaid spending, or capping the amount that each state receives from the federal government to operate the program. This option focuses on that last approach, although the other approaches’ effects on federal and state spending or on enrollees could be similar to the effects of caps on federal Medicaid payments, depending on how states were allowed to, and decided to, respond to such a policy change.

Capping federal payments for Medicaid could have several advantages relative to current law. It could generate savings for the federal budget if the caps were set below current projections of federal Medicaid spending. (Caps that were significantly lower than current projections could produce large savings.) Setting an upper limit on spending would also make federal costs for Medicaid more predictable. In addition, federal spending caps would reduce states’ current ability to increase federal Medicaid funds—an ability created by the open-ended nature of federal financing for the program and by the relatively high share of costs paid by the federal government. Because the relative benefit of state spending on an open-ended program such as Medicaid is higher than the relative benefit of state spending on other programs that do not receive federal funds, states have considerable incentive to devote more of their budgets to Medicaid than they would otherwise and to shift activities that had been funded entirely by the states themselves to Medicaid. Finally, if spending limits were accompanied by significant new flexibility for states—as many proposals for Medicaid caps envision—such flexibility might give states the opportunity to develop their own strategies for reducing program costs.

Caps on federal Medicaid spending could also have several disadvantages relative to current law. If the limits on federal payments were set low enough, they would shift additional costs—perhaps substantial costs—to states and cause state Medicaid budgets to become less predictable. In response, states would have to commit more of their own revenues to Medicaid or reduce services, restrict eligibility or enrollment, cut payment rates for health care providers, or (to the extent feasible) develop ways to deliver services more efficiently, each of which would raise various concerns. Moreover, depending on the structure of the caps, Medicaid might no longer serve as a countercyclical source of federal funds for states during economic downturns (meaning that a state might not automatically receive more federal funding if a downturn caused more state residents to enroll in Medicaid). In addition, because states differ significantly in the size of their Medicaid programs—and because spending varies widely (and grows at varying rates) for different types of enrollees within a state—policymakers could find it difficult to set caps at levels that accurately reflect states’ costs. Finally, it might be difficult to set caps that balanced the competing goals of creating incentives for program efficiency and generating federal budgetary savings, on the one hand, and providing enough funding that states could generally maintain the size of their current Medicaid programs, on the other hand.

Key Design Choices That Would Affect Savings

A wide variety of design specifications could significantly affect the amount of savings that caps on federal Medicaid spending would produce. The key specifications include the following: whether the caps would be set on an overall or a per-enrollee basis; what portions of Medicaid spending and what eligibility categories would be included in the spending limits; what year’s spending the initial caps would be based on and what percentage rate (or growth factor) would be used to increase the caps over time; how much new flexibility states would be given to make changes to Medicaid; and whether the optional coverage expansion authorized by the Affordable Care Act (ACA) would be subject to the caps (which would create some special complexities because that expansion has not yet been implemented). Those various design choices could interact in complicated ways.

Overall Cap or per-Enrollee Cap. Two principal ways to limit federal Medicaid spending through caps would be to cap overall federal spending for the program or to cap spending per enrollee. In general, overall spending caps would consist of a maximum amount of funding that the federal government would give each state to operate Medicaid. Once established, those caps would generally not change in response to changes in enrollment or (depending on how the caps were set to increase over time) in response to changes in the cost of providing medical services.

Per-enrollee spending caps—sometimes referred to as per capita caps—would consist of an upper limit on the amount that states could spend per Medicaid enrollee, on average. Under that type of cap, the federal government would provide funds for each person enrolled in the program but only up to a specified amount per enrollee. As aresult, total federal funding for each state would be limited to the number of enrollees multiplied by the per-enrollee spending limit. (Individual enrollees who incurred high costs could still generate additional federal payments, as long as the total average cost per enrollee was less than the per capita cap.) Unlike overall spending caps, this approach would provide additional funding to states if Medicaid enrollment rose (as it does when states choose to expand eligibility or during an economic downturn) and would provide less funding to states if Medicaid enrollment fell (as it does when states restrict enrollment or when the economy is strong).

Overall caps on federal Medicaid spending could be structured in two main ways. The federal government could provide states with fixed block grants that, in general, would not increase if states’ costs rose or decrease if states’ costs fell. Alternatively, the federal government could maintain the current financing structure of Medicaid, in which it pays for a specific share of total spending, but it could set a limit on the amount of federal funding that could be sent to the states. In that case, states would bear all of the additional costs for any spending that exceeded the federal caps, but both the states and the federal government would share the savings if spending was less than the caps. However, if caps were lower than current projections of federal Medicaid spending, such savings would be unlikely, in CBO’s view. Given states’ incentives to maximize federal funding, CBO expects that states would generally structure their Medicaid programs so as to qualify for all of the available federal funds up to the amount of the caps.

Per-enrollee spending caps could also be structured in different ways. One method would be to establish fixed federal payments per enrollee per month, similar to the capitation payments that managed care companies receive from public or private payers for each enrollee. Another method would be to base caps on average federal spending per enrollee for each of the four principal categories of people eligible for Medicaid: the elderly; the blind or disabled; children; and nonelderly, nondisabled adults. To determine the spending limit for each eligibility category, the federal government would count the number of enrollees in a category and multiply it by the specified per-enrollee spending amount for that category. In effect, the overall limit on Medicaid spending for each state would be the sum of the four limits for the four groups. A similar but more flexible approach would be to set one total limit based on the sum of the limits for the four groups as above, but allow states to cross-subsidize groups (spend more than the cap for some eligibility groups and less than the cap for others) as long as a state’s overall cap was maintained.

Spending Categories Included Under the Caps. Policy options to cap federal Medicaid spending could target all of that spending or spending for specific types of services. In Medicaid, most federal spending covers acute care ($152 billion in 2012) and long-term care ($71 billion in 2012), both of which could be broken into various subcategories. Other types of federal Medicaid spending include payments to hospitals that serve a disproportionate share of Medicaid enrollees and uninsured patients (known as DSH payments); spending under the Vaccines for Children (VFC) program; and administrative costs. (Together, those three categories totaled $28 billion in 2012.) In general, the more spending categories included under the caps, the greater the potential for savings to the federal government.

Eligibility Categories Included Under the Caps. Besides determining what types of Medicaid spending to cap, policymakers would face choices about which groups of enrollees to include. In general, the more eligibility categories covered by spending limits, the greater the potential for savings to the federal government. For example, caps could limit federal Medicaid spending on children and certain adults (either on an overall or on a per-enrollee basis) but could leave spending on the elderly and the disabled uncapped. However, because the elderly and the disabled currently account for about 65 percent of Medicaid spending—and are projected to account for about 50 percent in 2023, after the ACA’s expansion of coverage for nonelderly, nondisabled adults—caps that did not apply to those two groups would save far less than caps that covered all eligibility groups (assuming that the other characteristics of the two sets of caps were the same).

Base-Year Spending. Establishing caps on federal spending for Medicaid would generally begin with selecting a recent year of Medicaid outlays—the base year—and calculating that year’s total spending for the service categories and eligibility groups to be included in the caps. Those spending totals would then be inflated (as described in the next section) to calculate the spending limits in future years. Thus, for both overall and per-enrollee spending caps, the selection of the base year is important because the level of spending in that year would help determine future spending caps: A higher base-year amount would lead to higher caps (and lower federal savings) than a lower base-year amount would.

Another important choice in selecting a base year is whether to use a past or future year. Most cap proposals that include base years use a past year for which Medicaid expenditures are known. The main reason for using a past year is that states cannot raise payment rates for providers, make additional one-time supplemental payments, or move payments for claims from different periods into the base year to maximize Medicaid spending and thereby boost their future spending limits. However, policymakers might want to choose a future base year in situations in which a past year would not adequately reflect an upcoming program change, such as the implementation of the optional coverage expansion starting in 2014.

Another consideration is that using a prior base year would essentially lock in states’ past choices about their Medicaid programs and perpetuate those choices. (As an example of the differences among state Medicaid programs, in 2010, federal spending per disabled enrollee ranged from a low of about $5,000 in Alabama to a high of about $17,600 in the District of Columbia.) Once caps were set on the basis of states’ prior choices, it would be increasingly difficult over time for states to significantly raise their payment rates or voluntarily add covered services because, unlike under current law, such changes would not lead to higher federal payments. (One way to address that issue would be to add supplemental amounts to base-year spending levels for states defined as “low spending,” which would give them more room to expand their programs over time. That approach would reduce the savings from the caps, however.)

Growth Factor. The growth factor is the annual rate of growth that would be applied to base-year spending to determine the caps on (and rate of increase for) federal Medicaid spending in future years. The growth factor could be set to achieve different purposes and different levels of savings. For example, a growth factor that was roughly equal to the growth rate that CBO projects for Medicaid under current law would result in little or no budgetary savings relative to CBO’s spending projections, but it could achieve other policy aims. Alternatively, a growth factor could be set to make the increase in federal Medicaid spending—overall or per enrollee—consistent with the general rate of inflation (as measured by the consumer price index for all urban consumers, or CPI-U, for example), consistent with the growth rate of health care costs per person (as measured by the increase in national health expenditures, or NHE, per person, for example), or consistent with the rate of economic growth per person (as measured by the increase in per capita GDP). However, growth factors tied to price indexes or overall economic growth would not generally account for increases in the average quantity or intensity of medical services of the sort that have occurred in the past.

For overall spending caps, which would not provide additional funds automatically if Medicaid enrollment rose, the growth factor could include a measure of population growth (such as the Census Bureau’s state population estimates) to account for increases in enrollment. The growth factor could also be any legislatively specified rate designed to produce a desired amount of savings.

In general, the lower the growth factor relative to CBO’s projected growth rate for federal Medicaid spending under current law, the greater the federal budgetary savings. But the lower the growth factor, the greater the possibility that it would not keep pace with increases in costs per Medicaid enrollee and (in the case of overall caps) with increases in Medicaid enrollment, thus raising the likelihood that states would not be able to maintain their current levels of services or coverage.

Using a growth factor that incorporated the annual change in the CPI-U or in per capita NHE would mean that changes in federal Medicaid funding for states could vary considerably from year to year—although such funding could still vary less than it does under current law. As inflation, overall economic growth, or the growth of health care costs changed over time, growth factors based on those measures would cause federal Medicaid payments to rise and fall in tandem with those changes. Policymakers could address that potential volatility by using a three-year or five-year average of the growth factor in question, or they could limit the amount of annual fluctuation by allowing the growth factor to change by no more than a certain percentage.

Efforts to reduce the degree of variability in the growth factor, however, would diminish the factor’s responsiveness to changes in economic conditions. For example, if a period of low inflation, which caused only modest increases in a growth factor based on the CPI-U, gave way to a period of higher inflation, using a multiyear average for the growth factor or limiting annual changes in that factor would delay the full increase in federal Medicaid payments to states that would otherwise occur when inflation picked up. That delay would leave states with higher costs but not commensurately higher federal payments. Conversely, during a period when inflation declined—as it did in the most recent recession—mechanisms to dampen the volatility of the growth factor would slow the decrease in federal payments that would otherwise occur with per-enrollee caps. Overall, a range of adjustments are possible to mitigate those effects, but none would completely counter the effect of increased volatility without some loss of responsiveness to current economic conditions.

New Flexibility for States. Another important consideration in capping federal funding for Medicaid is how much new flexibility states would be granted. States have considerable flexibility under current Medicaid law to choose among optional services and eligibility groups; set payment rates for providers; and establish methods for delivering care, such as managed care and home- and community-based long-term care. However, states’ flexibility under current law is limited in significant ways, and obtaining waivers from certain program rules can be cumbersome and time-consuming even if the waivers are ultimately granted. In principle, the structure of Medicaid’s financing and the degree of state flexibility are separate issues: With a federal spending cap, the flexibility available under current law could remain the same or be altered to give states more or fewer options, and states’ flexibility could be increased or decreased under the current financing structure. Nonetheless, some proponents of caps consider additional state flexibility an essential feature of proposals to limit Medicaid spending.

If spending caps were coupled with new state flexibility, states could be given more discretion over a number of program features, such as administrative requirements, ways to deliver health care, cost-sharing levels, and covered eligibility categories and medical services. New flexibility would make it easier for states to adjust their Medicaid spending in response to a limit on federal funds. The degree of new flexibility that states received would be particularly important if the federal spending caps were significantly lower than CBO’s projection of Medicaid spending under current law.

Alternatively, federal spending caps could include a “maintenance of effort” requirement that would prevent states from changing the eligibility categories and medical benefits they covered before the caps took effect. That approach would ensure that key characteristics of the program in the base year—such as eligibility criteria, covered services, and the amount, duration, and scope of those services—would continue, preventing states from significantly curtailing their Medicaid programs after caps had been set.

Although the degree of new state flexibility included with caps could have a significant impact on states’ ability to adjust their programs in response to the caps, it would affect federal savings on Medicaid only if three things happened: states had enough flexibility to scale back their programs to the point where federal spending was less than the caps; federal funding remained linked to the level of state funding, as under current law; and some states chose to do such scaling back. If, instead, all states drew federal payments up to the amount of the caps—as CBO expects would generally happen—the degree of state flexibility would not affect the federal savings from the caps (although it might alter the scale and effectiveness of the Medicaid program, as discussed below).

The Optional Medicaid Expansion. Beginning in 2014, states have the option to expand eligibility for Medicaid to most individuals with income below 138 percent of the federal poverty guidelines. The federal government will cover a much higher share of the cost for those people than for other types of Medicaid enrollees: 100 percent initially, phasing down to 90 percent by 2020. That optional expansion creates added complexities for federal spending caps. Data from a past base year would reflect spending only for current eligibility groups, which, when increased using the growth factor, would fail to account for future spending for the expansion group (in states that adopt the optional expansion). Average per capita amounts also could differ for new eligibility groups.

In designing Medicaid caps, lawmakers could address those issues in several ways:

  • Select a base year far enough in the future to allow time for states to adopt the expansion (if they choose to) and for enrollment to reach a fairly stable level. Using a future base year, however, would give states the opportunity to inflate spending in that year, thus increasing their federal spending limits and reducing federal savings.
  • Leave spending attributable to the optional expansion group uncapped and limit spending only for nonexpansion enrollees. That approach would remove most of the complications created by the optional coverage group; however, it would not account for future spending for people already eligible for Medicaid who are not enrolled now but who are expected to enroll starting in 2014 (either because of the ACA’s mandate to obtain health insurance coverage or because of publicity about the Medicaid expansion). One way to account for the enrollment of that group would be to add an amount to the growth rate in the early years of the expansion. Another way would be to adjust the cap levels after several years of experience to account for the additional enrollees who were previously eligible but not enrolled, although knowing how much spending was attributable to that group would be difficult.
  • Cap spending for all enrollees but add a large enough amount to the growth factor to account for the enrollment of both newly eligible people and those who were previously eligible but not enrolled. Determining the size of those add-on factors would be challenging, however, and would be unlikely to provide the precise amounts of additional cap room needed to match those enrollees’ costs (the caps could end up being too low or too high).

Another issue related to the optional expansion is that capping federal Medicaid spending might cause some states that would otherwise expand coverage to reject the option instead. Limits on federal Medicaid payments represent a potential shifting of costs to states, which would affect their budget processes and decisions. One of the ways in which states could lower their Medicaid costs and reduce their financial risks would be to drop the optional expansion or fail to adopt it in the future (if not already implemented). CBO anticipates that the more that caps reduced federal funding below the level projected under current law, the greater the likelihood that states would turn down the optional expansion.

To the extent that states responded to caps by declining the optional expansion, some people would lose access to Medicaid coverage, although some of them would gain access to the health insurance exchanges as a result. Specifically, people with income between 100 percent and 138 percent of the federal poverty guidelines who lost their Medicaid eligibility would qualify for premium assistance tax credits to buy coverage through the exchanges. Of the people with income below the federal poverty guidelines who no longer had access to Medicaid, most would become uninsured, and the rest would enroll in other types of coverage, principally employment-based insurance. The net budgetary effect would be to increase the federal savings from the cap policy, CBO estimates, because the savings from the reduction in Medicaid coverage would be larger than the increase in spending for exchange subsidies for the share of people who would qualify for those subsidies.

Specific Alternatives and Estimates

CBO analyzed two types of limits on federal Medicaid spending: overall spending caps and per-enrollee caps. For both types, CBO assumed that the caps would take effect in October 2015 and would be based on spending in 2013 (excluding Medicaid’s DSH and VFC spending because the former is already capped and the latter provides vaccines for some children who may not be enrolled in Medicaid). In addition, for both types of caps, CBO excluded projected spending for the optional Medicaid expansion beginning in 2014 to avoid the complications discussed above. To illustrate a range of possible savings, CBO used two alternative growth factors for each type of cap: the annual change in the CPI-U or in per capita NHE. Other than the caps on spending, financing for the program would remain the same as under current law, with the federal government basing its share of total Medicaid spending on states’ expenditures (up to the caps). Under all of the alternatives, states would not receive any new programmatic flexibility but would retain the flexibility they have now to make decisions about optional benefits, optional enrollees, and payment rates for providers.

For the overall spending caps, CBO added 1 to 3 percentage points per year to the growth factors in 2014 through 2016 to account for previously eligible people who were not enrolled but would be induced to enroll by the changes introduced by the ACA. (CBO anticipates that most such effects would be fully in place by 2016.) Those add-on factors represent the percentage of Medicaid program growth under CBO’s baseline attributable to enrollment by that group. Those overall caps would save the federal government $450 billion between 2016 and 2023 using the CPI-U growth factor or $105 billion using the per capita NHE growth factor, CBO estimates. Those amounts represent savings of about 12 percent and 3 percent, respectively, of CBO’s projection of total federal Medicaid spending in that period under current law. By 2023, annual savings from the two varieties of overall caps would represent about 19 percent and 4 percent, respectively, of projected federal Medicaid spending in 2023 under current law.

For the per-enrollee spending caps, CBO assumed that separate spending limits would be set for each state for each of the four main Medicaid eligibility groups: the elderly; the blind or disabled; children; and nonelderly, nondisabled adults. States would not be permitted to cross-subsidize groups. CBO used the same growth factors as for the overall caps but did not include add-on factors for the previously eligible but not enrolled because per-enrollee caps would allow for additional payments on behalf of those enrollees. With those design parameters, the per-enrollee caps would save the federal government $610 billion through 2023 using the CPI-U growth factor or $280 billion using the per capita NHE growth factor, CBO estimates. Those amounts represent savings of about 17 percent and 8 percent, respectively, of total projected federal spending for Medicaid between 2016 and 2023 under current law. By 2023, the savings would represent about 23 percent and 8 percent, respectively, of projected federal Medicaid spending in that year under current law.

CBO’s estimate that per-enrollee caps would save more than overall caps on Medicaid spending (holding other factors equal) reflects some unusual economic circumstances. Under more typical economic conditions, overall caps would save more than per-enrollee caps because, with overall caps, Medicaid spending would increase only by the specified growth factor, whereas with per-enrollee caps, spending would rise by both the growth factor and increases in Medicaid enrollment. In its baseline forecast for the 2014–2023 period, however, CBO projects that Medicaid enrollment by nonexpansion adults and children will decline in some years because of the relatively rapid economic growth that is expected to occur as the U.S. economy recovers from its recent weakness. Those projected declines in enrollment lead to less Medicaid spending under per-enrollee caps but do not alter CBO’s estimate of federal payments under overall caps, thus increasing the relative savings from per-enrollee caps.

Other Considerations

Limits on federal Medicaid spending would affect not only the federal budget but also the operations of the Centers for Medicare & Medicaid Services (CMS), states' role in the Medicaid program, and, potentially, enrollees' Medicaid eligibility and the extent of covered services.

Implementation Issues. For both the overall and per-enrollee spending caps, CMS would have to establish new enforcement mechanisms to ensure compliance with the spending limits. The nature of those enforcement mechanisms would depend on the way in which authorizing legislation directed CMS to establish the caps.

If the caps were based on the actual values of the CPI-U or per capita NHE, CMS would not know the final spending limits until after the end of the fiscal year, when the growth rates for those measures were finalized. In addition, for per-enrollee caps, CMS would need to wait until final Medicaid enrollment for the year was known to determine the spending limits for Medicaid's four main eligibility groups. Because it currently takes up to two years to finalize states’ reports of enrollment, CMS would need to establish more timely reporting of enrollment to avoid large adjustments well after the close of the year. Regardless of how long it took to determine the final spending limits, CMS would need to adopt a reconciliation process to enforce compliance with the caps, either disallowing expenditures over the caps or lowering the following year's caps by the same amount.

As an alternative to waiting to finalize a given year's caps until after the end of the year, the caps could be based on projections of the CPI-U or per capita NHE. That way, states would know their cap amounts well before the end of the fiscal year and could plan accordingly, although then the caps would not account for changes to those measures that might occur later in the year.

Effects on States. Capping federal Medicaid spending would fundamentally change the federal-state financial relationship in the program. A capped federal commitment would mean that the responsibility for any growth in the program's costs that exceeded the growth factor (in this case, the increase in the CPI-U or per capita NHE) would be shifted to the states. CBO expects Medicaid costs to grow faster than the CPI-U or per capita NHE between 2015 and 2023, so the federal payments to states under this option would be lower than the payments projected under current law. Those savings to the federal government would represent lost revenues to states, and the losses would increase over time as the gap between federal payments under a capped program and under the current program grew larger.

Besides shifting some of the federal government's existing financial responsibility to the states, caps on federal payments would leave states at greater risk than they are now for changes in the health care marketplace and in the broader economy—elements over which they have limited control. In the case of overall spending caps, if the economy went into a recession, the growth of federal Medicaid payments would fail to keep pace with the rising need for services. (Between 2007 and 2010, for example, Medicaid enrollment increased by a total of about 14 percent.) With per-enrollee caps whose growth was based on the CPI-U, federal payments would rise in response to increases in enrollment, but payments would not respond when the growth of health care costs exceeded the growth of the CPI-U. With per-enrollee caps whose growth was based on per capita NHE, payments would adjust to average changes in the nationwide health care system but not to idiosyncratic changes in states' health care systems—and the federal savings from that alternative would be much smaller than from the approach examined here that would use the CPI-U.

With less federal funding and more budgetary uncertainty, states would have a stronger incentive than under current law to lower the cost of their Medicaid programs. To help states reduce costs, some proponents of Medicaid caps consider new programmatic flexibility for states to be an essential feature of such a policy. That flexibility could take several forms. States could be permitted to run their programs without having to meet some or all of CMS's current administrative requirements, they could be given more autonomy to experiment with new ways to deliver health care to Medicaid enrollees, and they could be granted discretion to reduce coverage of mandatory services and eligibility groups.

Proponents of caps point to several ways in which additional administrative flexibility could enable states to operate their Medicaid programs more efficiently. Depending on the nature of the flexibility provided, states might be able to implement administrative procedures that would require fewer employees or reduce the number of reports submitted to CMS for oversight purposes. However, administrative costs accounted for only about 5 percent of states’ total Medicaid spending in 2012, which suggests that even significant administrative efficiencies would save only modest amounts relative to total state spending on Medicaid. Proponents of caps also argue that giving states more flexibility could help them create incentives for Medicaid enrollees to use fewer services, such as through the use of increased cost sharing or of higher deductibles coupled with health savings accounts. In addition, some states might use extra flexibility to adjust the level of benefits provided to some enrollees so that, instead of receiving comprehensive benefits, as required under current law, those enrollees would receive a smaller set of targeted services to meet critical needs.

Under alternatives that would lead to significant reductions in federal funding, many states would find it difficult to offset the losses solely through the potential efficiencies described above. Such states would have three potential approaches open to them: raise additional revenues, cut other state programs to devote a greater share of their resources to Medicaid, or produce additional savings by lowering payment rates to providers, reducing covered services, or decreasing the number of enrollees. States already have some ability to adjust those elements of their Medicaid programs, but more flexibility would give them the opportunity to offset the larger losses of federal funding estimated under this option without having to raise additional revenues or cut other state programs. CBO expects that states would adopt a mix of those various approaches. Whether states would have enough flexibility to prevent declines in the number of people served by Medicaid or in the services that people received would depend largely on the size of the spending cuts that states would have to make to stay below the caps.

Effects on Enrollees. The ways in which Medicaid spending caps would affect individual enrollees would depend greatly on how an enrollee’s state responded to the caps. In states that chose to leave their Medicaid programs unchanged by finding other ways to offset the loss of federal funds, enrollees would experience little or no noticeable change in their Medicaid coverage. By contrast, in states that opted to reduce payment rates for providers, covered services, or Medicaid eligibility within the parameters of current law—or to a greater extent, if given the flexibility—enrollees would probably face several consequences. If states reduced payment rates, enrollees might find fewer providers willing to accept Medicaid patients, especially given that Medicaid already pays significantly lower rates than Medicare or private insurance in many cases. If states reduced the optional benefits they covered, some enrollees might pay out of pocket for those services or might forgo them entirely. And if states reduced the optional eligibility categories they covered (including the optional expansion slated to begin in 2014), those optional enrollees would lose access to Medicaid coverage.