Recent changes in the U.S. political environment have once again stirred up discussions of major reforms to the healthcare market. While a main topic in news discussions has been proposed reforms to health insurance exchanges created by the Patient Protection and Affordable Care Act (ACA), Medicaid reform has the potential to affect more people than any other source of coverage.
Republican Medicaid reform proposals have thus far focused on converting federal funding from the current approach of proportional federal and state financing to either block grants or per capita caps. While these funding approaches may sound relatively straightforward, understanding the implications of such changes requires consideration of several factors.
In this paper, we have broken down the detailed considerations into two primary categories: initial benchmark development and annual growth rates. Defining the assumptions and methodologies used to establish benchmarks and growth rates is key to aligning service cost with funding under alternative federal financing for Medicaid. Without consideration of these concepts, the actual cost of Medicaid relative to the federal budget for Medicaid will begin to diverge, and the gap may become wider over time. As this theoretical funding gap emerges, states will be at increased risk for funding additional program cost.
Figure 1 identifies detailed assumptions to consider for each key category. Additional details for each are included in the last section of this paper. Figure 2 illustrates state and federal expenditure growth risks and considerations for current funding, block grants, and per capita caps.
Figure 1: Considerations to alternative funding
|Key categories of consideration|
|Initial benchmark development||Annual growth rates|
|Category of aid||Medical cost & utilization trends|
|Age, gender, & care settings||Emerging medical treatment cost|
|Geographic cost variance||Historical or prospective trends|
|Base data period & source||Aging demographics|
|Benefit design||Population reliance on Medicaid|
|Federal medical assistance percentage||Economic growth rates/indices|
Figure 2: Potential risk by funding source
|Funding attribute||Current||Block grant||Per capita cap|
|Funding limit||None, as long as regulatory requirements are met.||Established in advance, unchanged with population growth or environmental factors.||Established in advance, varies based on population size, but unchanged for environmental factors.|
|State vs. federal medical growth rate||Consistent growth rates.||Federal growth defined in advance. State growth leveraged based on overall growth.||Federal growth is mitigated. State growth may be leveraged if cost per enrollee is more than projected.|
|Enrollment mix change risk||Federal risk varies by FMAP: if populations with higher federal match increase at a faster rate than the overall population, state share of bill is lower. For states with low/no expansion enrollment, match is relatively steady.||Federal government transfers risk to states.||Depends on structure. If cap is per capita on an aid category basis, then risk is similar to current. If not based on aid category, mix of members by aid category could negatively impact states as population groups age and LTSS become more prevalent.|
|Enrollment growth mix||Consistent risk state versus federal.||Federal government transfers risk to states.||Consistent risk state versus federal, as long as new members don't have higher-than-average cost.|
Medicaid was originally established as an assistance program for medical coverage of low-income children and disabled citizens under Title XIX of the Social Security Act (the Act) in 1965. It offers comprehensive healthcare coverage for a range of federally mandated and state-optional services. Each state administers its own program and has some autonomy over eligibility criteria and benefit packages. The program is regulated federally by the Centers for Medicare and Medicaid Services (CMS). Medicaid coverage has been revised over time, with the two most notable expansions being Title XXI of the Act, creating the State Children’s Health Insurance Program (CHIP)—covering children of families with higher income levels—and the optional extension of coverage under the Patient Protection and Affordable Care Act (ACA), effectively covering adults up to 138% of the federal poverty level (FPL).1 Medicaid and CHIP covered an average of 74.6 million people in federal fiscal year (FFY) 2015, as the largest single source of healthcare coverage in the country. Figure 3 illustrates a breakdown of enrollment and expenditures on the financial outlook for Medicaid, published by CMS and based on the two most recently available actuarial reports.2,3 It should be noted that the managed care expenditure value includes both acute and long-term services and supports (LTSS). LTSS expenditures appear to decrease in FFY 2015, however this is related to a shift from FFS to managed care delivery of these services. Values also include nonclaim costs such as Medicare premiums/cost sharing and Part D clawback; however, we have excluded disproportionate share hospital (DSH) payments as well as adjustments and administration cost.
Figure 3: Medicaid enrollment and expenditures
|Average monthly enrollment (# millions)|
|Population group||FFY 2014||FFY 2015|
|Title XXI CHIP4||5.9||6.1|
|Annual expenditures ($ billions)|
|Expense category||FFY 2014||FFY 2015|
|Title XXI CHIP4||$13.2||$14.6|
Medicaid is jointly funded by state and federal governments. The federal medical assistance percentage (FMAP) varies by state and is updated each year based primarily on state per capita income relative to the national average. FMAP rates range between 50% and 75% of traditional Medicaid service cost (as of federal fiscal year 2017), and states must comply with federally mandated eligibility and covered service requirements to receive federal funding.5 Federal participation also varies for different cohorts of the population, providing enhanced FMAPs for CHIP-eligible members under the CHIP Reauthorization Act of 2009 (CHIPRA) and for newly eligible adults under ACA expansion.6 Under the current financing system, states pay all medical cost incurred by Medicaid enrollees and submit quarterly expenses on a cash basis to CMS to draw down federal funds at the established FMAP rate.7 Figure 4 illustrates historical annual federal and state/local Medicaid expenditures, federal Medicaid funding as a percentage of total Medicaid expenditures, and the federal and state/local Medicaid expenditure growth rates from calendar year 2010 to 2015.8 It should be noted that the American Recovery and Reinvestment Act of 2009 (ARRA) provided for enhanced federal funding from October 2008 through June 2011.9 The increase in federal funding for 2014 and 2015 is primarily linked to expansion of eligibility for low-income adults under the ACA.
Figure 4: Medicaid spending by funding source
|CY||Federal||State/local||% Federal||Federal growth rate||State growth rate|
There is no fixed limit to Medicaid spending as long as states meet regulatory requirements for approved populations and services, so federal and state spending will increase proportionally when enrollment grows or medical costs trend upward. This open-ended financing system is difficult to forecast, and is a key reason that alternative funding proposals have been introduced from time to time. With the current transition to Republican control of the White House and Congress, Medicaid reform has again become a key topic of discussion.
Two alternative federal funding methods have been proposed by current Republican leadership: block grants and per capita caps. This paper discusses these methods at a high level, offering important considerations in setting up alternate funding.
Block grants are a funding mechanism that has been proposed at various times for Medicaid, and it serves as the current funding methodology for some nonmedical assistance social programs, e.g., Temporary Assistance for Needy Families (TANF).10 Under this proposal, each state would receive a predetermined amount of funds each year to provide Medicaid coverage. Unlike the current funding system, states would be responsible for funding all costs in excess of the federally established block grant budget amount rather than receive a proportional federal match for all cost. From a federal perspective, this makes budget planning more predictable, as the amount of funding provided to the states is formulaic and known in advance each year.
To establish block grant funding, historical medical cost would be the most likely place to start in establishing a baseline for first-year funding. Updates would be made annually for subsequent years based on formulaic trend factors intended to account for growth in both enrollment and cost of care as well as potential adjustments related to FMAP changes. In an effort to constrain federal spending on the Medicaid program, annual trend rates may be set lower than historical Medicaid trends.
Although a trend methodology has not been defined at this point, it is likely that the funding growth would not tie directly to the many complex factors that drive the growth of Medicaid expenditures. The gross domestic product (GDP) has been discussed as a potential growth rate, but may not reflect trends in aggregate future medical costs. For example, in times of recession, Medicaid enrollment often increases as unemployment increases and more people meet the income-based eligibility criteria. Additionally, the growth of block grant funding may not reflect ever-changing factors that drive per enrollee costs of healthcare, such as the emergence of new, expensive (but innovative) therapies and the aging demographics of the U.S. population.
It is a common expectation that if federal funding changes to block grants, states are likely to be given more flexibility to design more cost-effective programs, such as establishing state-determined eligibility requirement minimums and covered services.11 Each state is currently responsible for the administration of its Medicaid program. States have some latitude in designing their programs. However, in order to receive federal funding they must comply with mandated eligibility and benefit coverage requirements. If a block grant methodology is employed, based on previously proposed models and without modifying current Medicaid State Plan benefits, federal costs will increase at a defined rate, while state cost increases may be leveraged disproportionately to subsidize remaining cost as total program cost increases. To the extent that program cost requires additional state funding, the removal of certain CMS requirements could mitigate budget concerns. Some examples of added flexibility include:
- Establish wait lists instead of immediately enrolling qualified individuals.
- Eliminate retroactive coverage for periods prior to enrollment.
- Eliminate coverage entirely for certain populations.
- Benefit reductions:
- Reduce benefits below current federally-mandated levels.
- Allow alternative benefit plans with limited services for certain cohorts.
- Member engagement:
- Introduce health savings accounts, marginal premiums, or cost sharing for certain services.
Block grant funding may serve as an upper limit to federal funding, working in a manner consistent with current reporting and reimbursement. For example, total expenditures would be reported quarterly, and states would draw down funds up to the maximum allowable amount, based on FMAP rates. This structure would eliminate the incentive for states to make drastic cuts and use this federal funding for other purposes.
History of proposals
Figure 5 illustrates a history of proposals for funding Medicaid using a block grant or per capita cap funding mechanism.
Figure 5: Timeline of block grant proposals
Example of block grant funding for medical services
A prominent example of using block grants to finance a public healthcare system is the U.K.'s National Health Service (NHS), which currently provides comprehensive healthcare coverage for more than 64 million people across the U.K.16 Each year, Parliament decides on the amount of money that will be allocated to fund the program, and most of this funding is ultimately passed on to locally focused Clinical Commissioning Groups (CCGs), which purchase care from providers participating in the system.
However, over the last several years, a lack of funding to appropriately compensate providers has become an increasingly exacerbated issue. Overall in fiscal year 2016, NHS providers recorded a deficit of approximately GBP 2.45 billion, as costs for providers outpaced the total financing allocated from Parliament through the NHS. Furthermore, many individual CCGs and their corresponding local providers realized deficits that were even larger proportionally, as the formulas used to allocate funding to each CCG did not necessarily match the needs of the providers in the CCG. These formulas utilize information such as age/gender, poverty levels, and population size in order to decide how much healthcare funding each CCG should need to pay providers. However, to the extent that actual healthcare costs differ from the costs predicted by these formulas, there will be a disconnect between funding for providers and their actual costs.17
The funding issues surrounding the NHS have been well publicized and are a major focus of the current political discussion in the U.K. The experiences of the program highlight the importance of assumptions in determining the overall growth of block grant funding, as well as how that funding is allocated to localized purchasers of healthcare, where states, managed care entities, and medical providers will all be at risk for funding deficits.
Per capita caps
Another proposed methodology for determining federal Medicaid financing involves appropriating funds based on per capita caps. Under this proposal, a maximum baseline amount of funding is established per Medicaid enrollee, and this per enrollee cost cap would grow based on formulaic cost of care trend factors consistent with block grant funding. Also consistent with block grant funding, per capita cap funding would require states to cover all spending in excess of the cap. Unlike block grant funding, however, per capita caps allow for enrollment growth without penalizing state budgets. While the per capita cap mitigates state risk of higher-than-expected enrollment growth, it also means that federal funding amounts are not as predictable as they are under a block grant system.
Like block grant funding, a baseline per capita amount would be established for each state in the first year, and then the per capita amount would be calculated using a predetermined formulaic growth methodology. The applied growth factors would be designed to reflect an estimated increase in cost per enrollee. If the growth rates were to be set lower than historical Medicaid cost trends, it may reduce federal spending over time.
Although the per capita cap system is designed to allow for adjustments in funding as the number of people enrolled in Medicaid changes, it is not yet known whether the growth methodology would account for changes in factors such as the mix of members enrolled in Medicaid. Healthcare utilization and the average cost of services incurred by members vary by the demographics of the member, such as age, gender, or institutional care needs. For example, members requiring LTSS will be much more expensive than an average healthy child.
Figure 6 summarizes proposals that have been introduced for per capita cap funding.
Figure 6: Per capita cap proposals
Example of per capita cap funding
Section 1115 demonstration waivers are a long-standing example of how per capita funding could operate within Medicaid. Currently, Section 1115 of the Act allows the Secretary of the U.S. Department of Health and Human Services (HHS) to approve experimental programs that provide services or eligibility for populations not traditionally covered by Medicaid.20 In order to attain approval, states must show budget neutrality to the federal government, meaning that required federal funding must be no more than the estimated federal cost without the program.
Typically, budget neutrality is demonstrated by establishing a benchmark per capita cost based on historical experience, which is trended forward using a calculated growth rate assumption. The actual per capita cost under the waiver program is reported on a regular basis and must prove lower than the trended benchmark cost to satisfy neutrality requirements. If actual per capita spending exceeds the trended benchmark amount, states must either cover the excess cost or submit a formal request to modify the benchmarks, based on extenuating circumstances.
In applying for Section 1115 waiver approval, states establish benchmark per capita cost and growth rates using historical experience by Medicaid population. This mitigates the risk of varying growth rates in populations that have significantly different per capita costs. This process is analogous to how a per capita funding mechanism could work, although it is not clear whether states would be responsible for providing the initial assumptions or if the federal government would determine these assumptions.
We have outlined several technical and general considerations for stakeholders involved in converting the federal funding to an alternative proposal. If overlooked, these factors could cause inequities among states or a divergence in medical expenditure and funding growth rates over time.
Initial benchmark rates
Initial benchmarks must be set under either a block grant or per capita cap federal funding scheme. In developing benchmarks, there are many assumptions that must be addressed.
- Category of aid: Medicaid enrollees qualify for coverage based on age, income, and disability requirements, and each category has a different utilization and cost profile (e.g., low-income adult, aged, disabled, child). There are currently federal minimum requirements for mandatory coverage, and many states also extend coverage to optional groups. As a result of state demographics and varying eligibility standards, each state has a different mix of participants by category of aid. Average costs across category differ because of differences in health status, dual status (both Medicare and Medicaid coverage), disability status, or covered services. For example, the average cost of a low-income adult was approximately $340 per member per month (PMPM) whereas the average cost of a disabled adult was approximately $1,540 PMPM, based on national FFY 2011 data.21 Current FMAPs also vary by category of aid, which creates additional differences in funding by state.
- Age/gender: The demographic makeup of individual state populations varies, causing differences in each state’s Medicaid enrollment demographics. Even within a particular category of aid, costs can differ substantially by age and/or gender. For example, the average cost for children under the age of 2 can be four times as much as for children between 2 and 18.22
- Geographic differences in cost: The average cost of Medicaid services tends to be higher in urban areas relative to rural areas. Additionally, there are definite regional differences in healthcare markets across states because of provider or service availability, provider practice patterns, local healthcare purchasing nuances, and differences in covered populations or benefits. Medicaid reimbursement levels vary significantly by state as well, ranging from 38% of Medicare to 141% of Medicare rates for physician services.23 For example, the average annual Medicaid cost for a child in FFY 2011 was greater than $3,950 in five states, three of which are in the Northeast (one is Alaska), while the average annual cost for a child was less than $2,000 in six states, all but one of which are in the Midwest or Mountain West regions.24
- Base data period: In developing benchmark rates, the time period of historical base data will be critical. There are regular disruptions in state Medicaid programs, such as economic recessions, eligibility changes, benefit coverage changes, delivery system changes, and reimbursement manual changes that may happen at any time. It is difficult to establish a clean historical data period, and adjustments for disruption will vary by state and time frame.
- Benefit design: Each state currently defines the covered benefits for each aid category, subject to federal minimums. States may offer optional services to enrolled members, and this coverage may vary from year to year (e.g., adult dental or vision services). It is unclear whether historical data will be adjusted to establish a benefit minimum across all states to establish coverage consistency for developing benchmark rates, or if all states will be considered at their currently defined benefit levels.
- State or national data: A central ideological consideration to the development benchmarks is whether national or state-specific historical experience will be used. We have outlined demographic and economic reasons for variance in current Medicaid spending by state. However, even after adjusting for these known differences, spending by state may still differ significantly because of current program administration and local healthcare market considerations. In a recent letter to MACPAC commissioners, Republican leaders have requested that MACPAC “immediately initiate work to report on optional eligibility groups covered and optional benefits in each state Medicaid program…” focusing on federal and state expenditures for each.25
- Long-term care settings: One key difference in program administration across states is the management of enrollees receiving LTSS. Some states provide comprehensive home and community-based services (HCBS) in an effort to reduce long-term institutional costs. The mix of institutionalized versus community-based care settings varies by state, which lends to the variable average LTSS cost by state. If initial benchmarks are set based on current spending, then states with reduced LTSS spending will receive less funding, which is due to efforts they have already undertaken, and they will have lower opportunity for additional savings.
- FMAP: FMAP rates currently differ by category of aid (discussed above) and by state, based on each state’s per capita income. It is not clear how federal funding will be allocated among states under a block grant or per capita cap arrangement. Will state-specific funding remain consistent with historical federal expenditures for each state, or will changes in state per capita income influence the funding formula?
Annual growth rate selection
Once benchmark rates have been established, they will need to be trended forward to the funding period and updated annually thereafter.
- Cost and utilization trends: Cost (inflation) and utilization trends tend to vary by service category, and various cohorts of the population have different service mix needs. It may not be appropriate to choose a single trend that applies equally to all populations and services, because it may create winners and losers across states. Trends that reflect a state-specific mix of services and population demographics could facilitate funding that tracks more consistently with expenditures.
- Emerging treatments: New prescription drug treatments have been a major component of recent healthcare trends. Two recent examples are Harvoni® and Orkambi® , which are newly developed and expensive treatments for hepatitis C and cystic fibrosis, respectively. In 2015, Medicaid spent more on Harvoni® than any other pharmaceutical product.26 High-cost treatments such as these can have a significant impact on Medicaid spending. The impact will vary by state, depending upon the prevalence of the treated condition within the Medicaid population.
Historical vs. prospective trends: Historical trends are not always appropriate indicators of changes in future healthcare costs. Medical trends can change significantly over time because of emerging treatments, patent expirations for brand-name drugs, changes in medical practice patterns, changes in patient preference, or regulatory changes. However, historical trends are objective and can be simple to calculate.
Prospective trends applied to block grant or per capita cap funding may better capture future changes in Medicaid cost. However, prospective trends must be estimated in advance, are imprecise, and require judgment. They are subject to variability because of random fluctuation and unforeseen events.
- Aging demographics: Based on an analysis of recent American Community Survey (ACS) data, the size of the 65-to-74 age group increased 23% nationally between 2011 and 2014.27 Over the next decade, that population cohort will age into the 75+ age group and increase the demand for LTSS, an expensive component of current Medicaid spending. The increased cost of providing LTSS for this population could vary dramatically by state. Based on the ACS data, the five states with the largest growth in the 65-to-74 population experienced a 30% increase in that age group, while the bottom five states experienced a 17% increase. Changes in post-retirement geographic migration patterns and proximity of family members could cause additional variation.
- Reliance on Medicaid: Many external factors could increase the reliance on Medicaid. For example, an economic downturn could increase the unemployment rate and reliance on Medicaid for healthcare coverage. Another example is the recent rate increases in the private long-term care (LTC) insurance market, which may make private coverage less prevalent.28,29 Reduced rates of private coverage would put even more burden on Medicaid programs to fund LTSS expenditures.
- Growth rate comparison: Recent proposals for Medicaid funding changes have identified non-healthcare inflationary trends to be used as potential growth rates applied to block grant or per capita cap benchmarks. In "A Better Way," Speaker Ryan has proposed linking Medicaid funding growth to GDP growth. Other proposals have suggested linking growth to consumer price index (CPI) growth rates. Figure 7 illustrates national healthcare expenditure (NHE) growth, Medicaid expenditure growth, GDP growth rate, annual CPI for All Urban Consumers (CPI-U) change, annual CPI for medical costs (CPI-M) change from 2010 through 2015. Note that Medicaid trends vary by state from year to year, and the trends in the first three columns below include population growth, with Medicaid enrollment growth exceeding the overall population growth underlying NHE and GDP growth rates.
Figure 7: Medical and nonmedical annual change
Impact on state programs
Medicaid spending accounts for approximately 20% of individual state budgets costs, second in size only to education spending.35 Because of how much state spending is tied to Medicaid, there tend to be significant pressures on state lawmakers to reduce Medicaid spending when budgets are tight. Moving to a fixed federal funding formula rather than the current proportional federal funding could increase state responsibility and introduce additional variability to state funding requirements. As pressures to reduce state spending continue, significant policy decisions will need to be made to reduce budgetary requirements. Some examples of budgetary actions include:
- Managed care implementation or expansion
- Encouraging provider engagement through accountable care arrangements or delivery system reform incentive payment (DSRIP) programs
- Reducing provider reimbursement rates
- Eliminating optionally covered populations or benefits
- Implementing service limits
- Introducing member wait lists for coverage
In addition to the potential policy changes already noted, states may begin turning to alternative benefit designs for administering Medicaid. Some states have applied for Section 1115 demonstration waivers to provide such coverage to the Medicaid expansion population and other nondisabled adults. Two examples of these demonstrations include the Healthy Indiana Plan (HIP) and Healthy Ohio.
- HIP 2.0: Indiana expanded its Medicaid program to cover low-income adults in 2014. Rather than establishing coverage under the Medicaid State Plan, the state introduced a demonstration waiver with a benefit design that includes patient financial responsibility unlike standard Medicaid benefits. In February 2015, the demonstration was expanded to cover all nondisabled adults. HIP 2.0 introduced member contributions to a Personal Wellness and Responsibility (POWER) account, requiring members to make monthly contributions or face a coverage lockout (six months for members who have income above the federal poverty level) or a reduced benefit package (for members who have income below the federal poverty level).36
- Healthy Ohio: Ohio expanded its Medicaid program to cover low-income adults under the Medicaid State Plan. However, in 2016 it applied to establish a demonstration waiver that would have modified the benefit plan for Ohio’s Medicaid expansion population to add health savings accounts (HSAs), annual deductibles, copayments, monthly premiums, a healthy behavior program, a workforce requirement, and disenrollment from coverage for noncompliance. The demonstration application was denied by CMS with the reported concerns that it undermined affordability, leading to a reduction in access to coverage by the Medicaid expansion population.37
Some states have also begun to explore delivery system reforms that focus on incentivizing providers to promote the health of the population while finding efficiencies in medical care. Two examples of such reforms include Oregon’s Coordinated Care Organizations (CCOs) and New York’s Medicaid Reform Transformation (MRT) Waiver.
- Oregon CCOs: Oregon used an 1115 waiver to implement system reform by creating provider-owned entities that are responsible for physical and behavioral health needs of Medicaid patients. Under Oregon’s waiver, CCOs are assigned a global budget to cover medical service cost based on the enrolled population. A percentage of the budget is withheld until the end of a measurement period, after which a CCO can earn back withheld funds by meeting certain quality indicators.38
- New York MRT: New York implemented a DSRIP program in 2015 under the authority of an 1115 waiver. Funding is used to incentivize provider groups to develop coordinated care networks to achieve improvements in patient outcomes and overall population health. Provider groups choose clinical outcomes which are measured over time, and DSRIP funding pools are shared among providers based on improvement in quality based on chosen measures.39
As states seek out ways to continue offering Medicaid coverage under more limited federal funding, reformed coverage terms such as those introduced in the Indiana and Ohio demonstration waiver applications may become more widespread, increasing the financial participation and engagement of Medicaid enrollees in their healthcare purchasing. H.R. 277, a new ACA repeal bill, has been released by the Republican Study Committee and would permit states to offer HSA-like accounts for Medicaid enrollees.
The materials in this document represent the opinion of the authors and are not representative of the views of Milliman, Inc. Milliman does not certify the information, nor does it guarantee the accuracy and completeness of such information. Use of such information is voluntary and should not be relied upon unless an independent review of its accuracy and completeness has been performed. Materials may not be reproduced without the express consent of Milliman. All information is current as of the time of publication. This report should not be interpreted as an endorsement of any particular legislation by Milliman or the authors. As legislation develops and regulations change, answers may emerge that prompt new questions or considerations. We ask that this report be distributed only in its entirety because extracts of this report taken in isolation may be misleading.