Block grants and per capita caps have been proposed as mechanisms for controlling Medicaid expenditures. Block grants would allocate money to states based on current overall spending levels in each state, and per capita caps would allocate funds based on current spending per enrollee. In this brief we show that federal spending (adjusted for the size of each state’s low income population) varies across states by more than 5 to 1 and spending per enrollee varies on the order of 2 to 1. In general, high income states will get larger block grants and higher spending per enrollee caps because they spend more today. These disparities would be locked in under these kinds of proposals.
Over the past 25 years, various Congressional leaders have called for Medicaid financing to be reformed with either block grants or per capita caps. Block grants would give states a fixed allotment based on aggregate, historical spending levels multiplied by a predetermined growth rate. Per capita caps are similar: they would set the allotments for specified enrollment groups based on historical spending per enrollee multiplied by a predetermined growth rate. Supporters of either approach would set growth rates lower than would be expected under current law in order to reduce federal spending. These proposals have been seen as problematic because of their potential to limit enrollment, reduce benefits, and shift spending to states, providers, and individuals. But the issue we address in this brief is that setting state allocations based on their historical spending levels would lock in the current huge variation in federal dollars sent to each state. We estimate that spending per low-income person varies by a factor of about 5 to 1 across the states, and spending per enrollee varies by a factor of at least 2 to 1. Thus, some states would get far higher block grant allotments relative to the size of their high-need populations or much higher spending per enrollee than other states. Despite federal matching grants that vary inversely with state per capita income, higher-income states spend more and would in general receive the larger allotments or caps. These differences reflect current spending variations, which are the result of state policy decisions. But block grants and per capita caps would lock them in place in perpetuity as federal policy decisions. In particular, block grants would largely prevent states from expanding coverage and benefits in the future. Any federal attempt to reduce current differences in Medicaid funding to states would require substantial disruptions to health systems in some states or additional federal spending.
Block grants for Medicaid have been proposed this year by several Republican presidential candidates. They have also been proposed off and on over the past 25 years by Republican legislators (Holahan and Liska 1995; Holahan and Weil 2003). A block grant proposal was passed by a Republican Congress in 1995 but vetoed by President Bill Clinton.1 A new variant of block grants, known as per capita caps, has been proposed by House Republicans and the American Enterprise Institute (Antos, Capretta, and Wilensky 2015; Rudowitz, Garfield, and Young 2016). The recent proposal announced by House Republicans, which would repeal the Affordable Care Act (ACA) and institute an array of policy changes, would give states a choice between block grants and per capita caps.2
In this brief we show how much variation there would be across states in federal dollars available under block grant and per capita approaches. Our key point is that establishing block grants or per capita caps, whatever their merits, would be difficult because of the large disparities across states that have built up over time in the federal share of Medicaid expenditures. These disparities reflect not only state policies toward coverage, benefits, and payment rates, but also federal matching rates. The problem is that either a block grant or a per capita cap would be based on current federal expenditure levels for each state and then multiplied by a uniform growth rate. This practice would produce huge differences in the distribution of federal dollars among states and would generally favor high-income states because they have historically spent more on Medicaid.
These differences have developed because of state decisions on Medicaid eligibility, benefits, and efforts to encourage high levels of participation (Kenney et al. 2016). With a block grant or per capita caps, these differences would be locked in through federal law. Federal expenditures would reflect federal decisions, not state decisions, on the allocation of tax dollars. In fact, states would lose the flexibility they now have to expand coverage, improve benefits, or increase participation. These policies can create other problems that have been well documented, such as forcing states to make undesirable cuts in spending (reducing benefits or enrollment) and placing greater financing burdens on states, households, and providers (Holahan et al. 2011; Holahan et al. 2012).
What Are Block Grants and Per Capita Caps?
Block grants would end Medicaid’s open-ended matching structure whereby states receive federal matching payments based on their expenditures. Federal matching rates vary inversely with state per capita income, thus favoring lower-income states. Currently, federal matching rates range from 75 percent in Mississippi to 50 percent in 12 states, including California and New York. But the amount of federal dollars, as well as total spending, depends on the amount states choose to spend on their programs. Although there are minimum eligibility and benefit standards, states are given a considerable amount of discretion over which optional populations to cover and what additional benefits to offer. Some states cover more of their low-income populations than others, have broader benefit packages, and pay providers more; some have more extensive long-term care payments. Block grants would allocate federal dollars for each state based on some measure of current expenditures, and the size of the grant would be allowed to grow over time by a predetermined growth rate, such as the consumer price index. In all public proposals in which a growth rate has been specified, the block grant approach has been designed to reduce federal expenditures below that projected by the Congressional Budget Office or the Centers for Medicare & Medicaid Services, or CMS (Holahan and Weil 2003). In this way, this type of policy would provide savings to the federal government. Such policies could also provide some fiscal relief to states, depending on whether states are required to maintain a certain level of effort. As noted, however, these proposed policies could also increase costs to states, depending on their decisions on coverage and benefits. States could be given one overall block grant or separate block grants for specific groups of eligibles or for categories of services provided (e.g., aged and disabled and adults and children or acute care and long-term care).
The intent is to provide states with a fixed budget, forcing them to become more efficient. At the same time, states would be given flexibility that they either currently lack or must work hard to obtain through the Section 1115 waiver process. Some states appear to desire flexibility in using premiums, imposing cost sharing, introducing health savings account–type arrangements, and imposing work requirements (Wishner et al. 2015). Some states also object to the requirement that benefits offered to one group of eligibles must be offered to all, and many would prefer more benefits flexibility. Although some states argue for more flexibility, a great deal of money is spent on Medicaid that is currently optional for states, including optional eligibility groups and optional benefits. For example, prescription drugs and adult dental care are optional benefits, and many long-term care services are optional. Because states already constrain spending relatively aggressively, staying beneath this lower rate of growth in federal dollars would often be difficult (Clemans-Cope, Holahan, and Garfield 2016). Indeed, Medicaid spending per enrollee and rates of growth over time have been lower than what has been experienced in private health insurance.
The fear is that with block grants in place, states would have to reduce enrollment or increase state expenditures to maintain their current programs (Holahan et al. 2011; Holahan et al. 2012). Such a financial squeeze would be exacerbated in economic downturns. When incomes fall, more people become eligible for Medicaid, but federal payments would remain fixed. Medicaid spending can also increase unexpectedly with the need to address public health emergencies and related problems (e.g., Hurricane Katrina, 9/11, and HIV).
Per Capita Caps
Per capita caps would constrain Medicaid spending on a per enrollee basis instead of in aggregate, as would be the case with block grants. The starting point for the caps would begin with each state’s current federal expenditures per enrollee, either overall or, more likely, separately for particular groups of eligibles (e.g., the aged, disabled, adults, and children). Like block grants, the per capita caps would grow by a predetermined growth rate, one presumably lower than projected by the Congressional Budget Office or CMS. Unlike block grants, however, per capita caps are intended to protect against unexpected enrollment increases, such as in a period of recession when people lose jobs and Medicaid eligibility and enrollment increase. Although per capita allotments would stay fixed in bad economic times, the total federal allotment to a state would increase with greater enrollment and thus states would not be penalized for such downturns. However, if the predetermined cap is insufficient to provide services at current levels, states would have to choose among increasing state revenue, limiting enrollment, reducing covered benefits, and lowering provider payments (or some combination of these actions).
Data and Methods
Because current spending is the starting point for establishing either block grants or per capita caps, our analysis provides estimates of variation in current federal Medicaid spending across states. To assess potential ramifications of a block grant, we provide estimates of variations in aggregate spending, spending per low-income resident (to control for differences in the size of state populations) overall, and spending for subgroups. For our analysis of per capita caps, we provide estimates of overall spending per enrollee as well as for four subgroups. Our estimates are projected for 2017 and assume that Medicaid is structured as it was without the Medicaid expansion. We draw on data from multiple sources to estimate Medicaid enrollment and federal spending levels.
To match Medicaid administrative data as closely as possible, we reweighted the basic data underlying the Health Insurance Policy Simulation Model to replicate the 2013 enrollment estimates prepared by the Kaiser Family Foundation and Health Management Associates (Snyder et al. 2014). We then used the model’s growth factors to project Medicaid enrollment in 2014 and 2017. These projections were consistent with the increases in Medicaid enrollment from 2013 to 2016 that were reported by CMS for each state. Enrollment for 2017 was created from enrollment data from the beginning of 2016 aged by one year.
We began with the latest available data for Medicaid spending by eligibility type (aged, disabled, nondisabled adults, and nondisabled children) from the 2012 Medicaid Statistical Information System. For a few states, 2011 was the latest year for which data were available. We then incorporated this spending data by state and eligibility type into the Health Insurance Policy Simulation Model by aging the costs to 2014 by using the projected per capita increases from the National Health Expenditure Accounts. We also adjusted costs to account for differences in the health risk of the simulated Medicaid population under the ACA versus Medicaid enrollees in 2012, as well as differences in the federal share of the total costs of enrollees under the ACA in 2014. We then adjusted our estimates to be compatible with the 2014 CMS-64 at the state level. This data set is the definitive measure of total federal Medicaid spending in each state, but it has no information about enrollees. We divided the CMS-64 totals by our simulated enrollment to compute the federal spending per enrollee in 2014. We were not able to compute per capita spending for each eligibility type in each state from the CMS-64 data, so we adjusted the Medicaid Statistical Information System–based per capita spending for each eligibility type so that the overall per capita spending matched the CMS-64–based estimate. In most states, this adjustment was very small. National per capita spending estimates for 2014 based on the Medicaid Statistical Information System and CMS-64 differed only by about 2 percent. Finally, expenditures were aged from 2014 to 2017 based on National Health Expenditure Accounts projections and differences in the federal share of the total costs of enrollees in 2017.
Table 1 provides data for analyzing the implications of state variations under a block grant proposal, and table 2 provides data for analyzing the effect of the variations under a per capita cap approach.
Table 1 shows differences in federal Medicaid spending by state, both in aggregate and per low-income state resident. Income below 150 percent of the federal poverty level is used to control for differences in the size of state populations. Estimates are provided for all enrollees combined and, in case distinct block grants for different eligibility groups are contemplated, separately for the aged plus disabled and for nonelderly adults plus children. We provide results assuming that the ACA Medicaid expansion would be repealed and that spending on the expansion populations would not be included in a block grant or a per capita cap. However, the amount of spending variation across states that we show would be unchanged, if not exacerbated, if recent additional federal spending was included. We assume block grants would be based on historical spending because that is how they have usually been designed (Rudowitz, Garfield, and Young 2016).
Federal Medicaid spending per low-income state resident varies by a factor of 11.3 to 1, or by 5.4 to 1 not counting the District of Columbia, resulting from a host of state Medicaid spending and other policy decisions, health care costs, program participation rates, and other factors. This variation means that high-spending states, such as the District of Columbia ($11,917 per low-income resident), Vermont ($5,438), New York ($5,646), and Connecticut ($4,432), would get a rich allotment of federal dollars per low-income person and could sustain most of the comprehensive coverage and benefits they provide, at least in the near term, until the predetermined slower federal spending growth began to bite. (The District of Columbia’s spending is so high because it offers a comprehensive program but also benefits from a 70 percent federal matching rate.3) California would receive about half as much per low-income resident ($2,979) as New York. States such as Virginia ($1,778 per low-income resident), Utah ($1,696), New Hampshire ($1,599), and Nevada ($1,051) would be allotted substantially less per low-income state resident and would not have the option to expand coverage or benefits without financing the expansion themselves. Eight states are used to illustrate this wide variation alongside the national average in figure 1. In general, higher-income states would receive larger block grants (per low-income person) than lower-income states.
If separate block grants were used instead of a single aggregate one, the same result would occur. Spending on the aged and disabled per low-income person varies by a factor of almost 10 to 1, or 5 to 1 not counting the District of Columbia, from $40,757 in the District of Columbia and $23,100 in New York to $4,696 in New Hampshire and $4,520 in Nevada. Spending per low-income person on adults and children also varies by a factor of more than 10 to 1 (4 to 1, not including the District of Columbia), from $4,894 in the District of Columbia and $2,902 in Vermont to $703 in New Hampshire and $502 in Nevada. California ($1,033) would receive about half as much as New York ($2,251) per low-income person.
Freezing these aggregate payments in place and having them all grow by the same rate would therefore lock in place current differences in federal spending per low-income individual. Although these differences exist today, states have the flexibility to modify their policy decisions (e.g., they can choose to expand coverage and benefits or increase provider payment rates). In the future this flexibility would no longer exist.
Designing a block grant to reduce differences in current federal spending levels by state is difficult. Some analysts argue that setting the aggregate amount of each state’s block grant to a single national amount per low-income state resident times the number of such residents would be fairer. The Healthcare Accessibility, Empowerment, and Liberty Act of 2016 bill, for example, incorporates the principle of moving to a more equitable distribution of federal funds.4 But such changes, although arguably more equitable, would be hugely disruptive. Although a simple benchmark average amount would eliminate differences, roughly half the states would receive less in federal dollars than they do today under such an alternative. The approach would create huge problems for high-spending states, requiring substantial increases in state taxes to sustain their current programs or creating enormous disruptions for individuals’ coverage and for health systems. Pegging aggregate block grants in each state to national average spending per low-income resident instead of state-specific spending would mean, for example, that New York would receive $18.8 billion instead of $38.0 billion in 2017; Massachusetts would receive $5.2 billion rather than $8.2 billion that year. Moreover, states that have spent less historically would get far more than they are currently spending or may want to spend. Under the same approach, Texas, for example, would receive $28.8 billion rather than the $18.7 billion it would receive if payments were based on state-specific spending; Florida would receive $21.7 billion rather than $15.9 billion. These latter states would be better positioned to absorb the slower increases in federal spending in subsequent years, but they could instead use these additional federal dollars for other nonhealth state financing priorities or tax relief.
A less radical approach to dealing with differences in funding levels across states is to apply slower growth rates to allotments for high-spending states and faster growth rates to allotments for low-spending states. This approach would also create the same number of state winners and losers, but it would take a significant period of time for the full impact to be felt.5
Per capita caps face similar problems. Table 2 shows estimated 2017 federal Medicaid spending per enrollee by state for all eligibility groups combined and by eligibility category; these data are the basis for computing state-specific per capita caps. As table 2 shows, spending per enrollee overall and for each eligibility group varies considerably. The differences in allotments based on this varied spending would therefore be substantial. The base for a single aggregate per capita cap would be as high as $10,459 per enrollee in the District of Columbia, $6,540 in New York, and $6,300 in Connecticut and as low as $3,312 per enrollee in Nevada, $3,262 in Illinois, and $3,084 in New Hampshire (figure 2). Per capita caps specifically for aged Medicaid enrollees would vary from $12,789 in the District of Columbia and $12,267 in Connecticut to $3,772 in Utah and $2,608 in Illinois. Per capita caps specific to persons with disabilities would range from $32,871 in the District of Columbia and $20,374 in New York to $8,597 in Alabama and $8,312 in New Hampshire. Per capita caps for adults would vary from $5,947 in Kentucky and $5,908 in the District of Columbia to $1,725 in Arkansas and $1,999 in New Hampshire. Per capita caps for children would range from $3,510 in Vermont and $3,428 in the District of Columbia to $1,320 in Wisconsin and $1,359 in New Hampshire. Per capita caps would freeze these state-level differences in place, eliminating states’ current flexibility to make changes in their spending per enrollee by either modifying benefits or changing provider payment rates. The advantage of the per capita caps relative to block grants, however, is that the approach does not penalize states financially for enrollment growth.
There are simply no politically easy ways to implement block grant or per capita proposals in Medicaid given the substantial variation in federal spending across states. The case for adopting them is weak on other grounds as well. For all the rhetoric, Medicaid is not a high-spending program. As we have shown recently in an extensive literature review, Medicaid spending is well below private spending levels when health status is controlled for (Clemans-Cope, Holahan, and Garfield 2016). Hadley and Holahan (2003) found that it would cost about 18 percent more to enroll Medicaid-covered individuals in private health insurance plans. The Congressional Budget Office (2014) came to similar conclusions when assessing Medicaid expansion provisions of the ACA. In addition, Medicaid spending growth has been lower than that of other payers and low relative to gross domestic product or inflation (Clemans-Cope, Holahan, and Garfield 2016). The primary problem that block grants seek to remedy is, at best, not a serious one. Many of the same people who argue that Medicaid is wasteful and a huge fiscal drain also highlight the program’s weaknesses by citing low rates of provider participation in response to inadequate payment rates. States already constrain Medicaid spending, generally preferring to spend limited state resources on K–12 and higher education, corrections, and infrastructure. States have successfully controlled costs by limiting provider payment rates, providing alternatives to nursing home care, and controlling prescription drug care and utilization. Neither block grant nor per capita cap proposals are likely to improve program efficiency. Block grants and per capita caps would successfully address problems such as the use of provider taxes to draw in new federal dollars through higher provider reimbursement rates. But these problems can be addressed in other ways without the same potential adverse consequences (Iritani 2014).
If anything, it could be argued that Medicaid should receive additional resources to improve provider payment rates, a change that would be made substantially more difficult for states to make under either a block grant or a per capita cap approach (Zuckerman, Skopec, and McCormack 2014).
Block grants and per capita caps attempt to reduce the level and the rate of increase of federal outlays on Medicaid. The policies would also reduce states’ authority to make policy decisions over their own programs. These mechanisms would threaten current coverage levels and benefits that low-income people often need yet cannot afford. In addition to slowing federal spending growth in Medicaid to reduce spending in real terms over time, the approaches would cement current levels of spending by state when those levels vary tremendously across the country. There is a much higher level of federal spending, both overall (adjusted for the size of each state’s low-income population) and on a per enrollee basis in some states than in others. Variations in spending per low-income resident can be more than 5 to 1, and spending per enrollee varies on the order of 2 to 1. In general, higher-income states would get larger block grants and higher spending per enrollee caps because they spend more today and thus receive larger allocations from the federal government despite having lower federal matching rates. These proposals would lock in these spending differences. Forcing equity in the current distribution of federal dollars without increasing spending in aggregate would require deep cuts to the health systems in currently higher-spending states in order to provide increases in funds to other states that are well below the average today. As in the past, it is likely to prove very difficult to enact mechanisms like block grants and per capita caps into law because of the widespread differences in the current distribution of federal dollars that would become locked into the system. These variations now reflect state decisionmaking; with either block grants or per capita caps, the differences in allocation of federal dollars would reflect federal decisionmaking.
- Alison Mitchell, “With First Veto, Clinton Rejects Budget-Cut Bill,” New York Times, June 8, 1995, http://www.nytimes.com/1995/06/08/us/with-first-veto-clinton-rejects-budget-cut-bill.html.↩
- “Health Care,” A Better Way, accessed September 6, 2016, http://abetterway.speaker.gov/_assets/pdf/ABetterWay-HealthCare-PolicyPaper.pdf.↩
- The 70 percent federal matching rate was agreed to in the Balanced Budget Act of 1997. Section 4725(b) of the Balanced Budget Act of 1997 amended section 1905(b) such that the federal medical assistance percentages for the District of Columbia will be 70 percent.↩
- 114th Congress, Healthcare Accessibility, Empowerment, and Liberty Act of 2016 (HAELA), May 2016, http://www.goodmaninstitute.org/wp-content/uploads/2016/05/SESSIO_007_xml.pdf.↩
Antos, Joseph, James C. Capretta, and Gail Wilensky. 2015. “Replacing the Affordable Care Act and Other Suggested Reforms.” Washington, DC: American Enterprise Institute.
Clemans-Cope, Lisa, John Holahan, and Rachel Garfield. 2016. “Medicaid Spending Growth Compared to Other Payers: A Look at the Evidence.” Washington, DC: Kaiser Commission on Medicaid and the Uninsured.
Congressional Budget Office. 2014. “The 2014 Long-Term Budget Outlook.” Washington, DC: Congressional Budget Office.
Hadley, Jack, and John Holahan. 2003. “Is Health Care Spending Higher under Medicaid or Private Insurance?” Inquiry 40: 323–42.
Holahan, John, Matthew Buettgens, Caitlin Carroll, and Vicki Chen. 2012. “National and State-By-State Impact of the 2012 House Republican Budget Plan for Medicaid.” Washington, DC: Kaiser Family Foundation.
Holahan, J., M. Buettgens, V. Chen, C. Carroll, and E. Lawton. 2011. “House Republican Budget Plan: State-by-State Impact of Changes in Medicaid Financing.” Washington, DC: Kaiser Family Foundation.
Holahan, John, and David Liska. 1995. “The Impact of the ‘Medigrant’ Plan on the Federal Payment to States.” Washington, DC: Kaiser Family Foundation.
Holahan, John, and Alan Weil. 2003. “Block Grants Are the Wrong Prescription for Medicaid.” Washington, DC: Urban Institute.
Iritani, Katherine M. 2014. “Completed and Preliminary Work Indicate That Transparency around State Financing Methods and Payments to Providers Is Still Needed for Oversight.” Testimony before the Subcommittee on Energy, Policy, Health Care and Entitlements, Committee on Oversight and Government Reform, House of Representatives. Washington, DC: United States Government Accountability Office.
Kenney, Genevieve M., Jennifer M. Haley, Clare Wang Pan, Victoria Lynch, and Matthew Buettgens. 2016. “Children's Coverage Climb Continues: Uninsurance and Medicaid/CHIP Eligibility and Participation under the ACA.” Washington, DC: Urban Institute.
Rudowitz, Robin, Rachel Garfield, and Katherine Young. 2016. “Overview of Medicaid per Capita Cap Proposals.” Washington, DC: Kaiser Family Foundation.
Snyder, Laura, Robin Rudowitz, Eileen Ellis, and Dennis Roberts. 2014. “Medicaid Enrollment: June 2013 Data Snapshot.” Washington, DC: Kaiser Family Foundation.
Wishner, Jane B., John Holahan, Divvy Upadhyay, and Megan McGrath. 2015. “Medicaid Expansion: The Private Option and Personal Responsibility Requirements.” Washington, DC: Urban Institute.
Zuckerman, Stephen, Laura Skopec, and Kristen McCormack. 2014. “Reversing the Medicaid Fee Bump: How Much Could Medicaid Physician Fees for Primary Care Fall in 2015?” Washington, DC: Urban Institute.
About the Authors
John Holahan is an Institute fellow in the Health Policy Center at Urban, where he previously served as center director for over 30 years. His recent work focuses on health reform, the uninsured, and health expenditure growth, developing proposals for health system reform most recently in Massachusetts. He examines the coverage, costs, and economic impact of the Affordable Care Act (ACA), including the costs of Medicaid expansion as well as the macroeconomic effects of the law. He has also analyzed the health status of Medicaid and exchange enrollees, and the implications for costs and exchange premiums. Holahan has written on competition in insurer and provider markets and implications for premiums and government subsidy costs as well as on the cost-containment provisions of the ACA.
Holahan has conducted significant work on Medicaid and Medicare reform, including analyses on the recent growth in Medicaid expenditures, implications of block grants and swap proposals on states and the federal government, and the effect of state decisions to expand Medicaid in the ACA on federal and state spending. Recent work on Medicare includes a paper on reforms that could both reduce budgetary impacts and improve the structure of the program. His work on the uninsured explores reasons for the growth in the uninsured over time and the effects of proposals to expand health insurance coverage on the number of uninsured and the cost to federal and state governments.
Matthew Buettgens is a senior research analyst in the Health Policy Center at the Urban Institute, where he is the mathematician leading the development of Urban’s Health Insurance Policy Simulation Model (HIPSM). The model is currently being used to provide technical assistance for health reform implementation in Massachusetts, Missouri, New York, Virginia, and Washington as well as to the federal government. His recent work includes a number of research papers analyzing various aspects of national health insurance reform, both nationally and state-by-state. Research topics have included the costs and coverage implications of Medicaid expansion for both federal and state governments; small firm self-insurance under the Affordable Care Act and its effect on the fully insured market; state-by-state analysis of changes in health insurance coverage and the remaining uninsured; the effect of reform on employers; the affordability of coverage under health insurance exchanges; and the implications of age rating for the affordability of coverage.
Buettgens was previously a major developer of the Health Insurance Reform Simulation Model—the predecessor to HIPSM—used in the design of the 2006 Roadmap to Universal Health Insurance Coverage in Massachusetts.
This brief was funded by the Urban Institute. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders. Funders do not determine research findings or the insights and recommendations of Urban experts. Further information on the Urban Institute’s funding principles is available at www.urban.org/support.
We are grateful for the comments from Genevieve Kenney and Linda Blumberg and the research assistant support from Clare Pan and Patricia Solleveld.