Matt Weidinger’s December 2025 commentary argued persuasively that complexity has become an impediment to rational reform of the nation’s safety net. By approving scores of new programs across several decades, Congress and the states have assembled such an unwieldy and uncoordinated anti-poverty portfolio that policymakers have difficulty fully grasping what it contains much less what might happen if they attempted to reform even one or two components of it.

This general observation also applies in a specific way to Medicaid, which is far and away the most expensive entitlement provided to low-income households and contains within its remit (at least as interpreted by most states) such a diversity of missions, roles, and optional services that it is difficult to approach reforming it as if it had a single purpose (Thomas Savidge published a paper in October 2025 broadly making this same observation).

Medicaid’s original design planted the seeds for its unrelenting growth. At its heart, Medicaid is a financing program, with the federal government committed to paying for a portion of all qualified state-administered spending. The matching formula, called the Federal Medical Assistance Percentage, or FMAP, is calculated based on a comparison of average per capita income levels in the states relative to the national average. Lower-income states get more federal support per dollar spent compared to higher income states. In fiscal year 2026, Mississippi is eligible for an FMAP of 76.9 percent for its standard Medicaid costs while multiple higher-income states, including California, Colorado, Massachusetts, New Jersey, New York, and Wyoming, get the minimum allowable rate of 50.0 percent (without this floor, several states would have FMAPs below 50 percent).

Although high-income states get lower FMAPs, they still have the largest programs because they have the tax bases required to support less restrictive eligibility and benefit rules. In addition, nearly all states use provider taxes and general revenue transfers to artificially inflate their FMAPs and shift more costs onto federal taxpayers.

Current Medicaid law also adjusts the standard FMAP for certain qualified expenditures and populations, including by paying for 90 percent of the costs for persons made eligible by the Affordable Care Act’s expansion of the program. Administrative expenses are generally split 50-50.

The matching system embedded in Medicaid has been a powerful incentive for states to search for as many activities as possible that can be interpreted as falling within the program’s boundaries. Under Medicaid, states’ pay at most only half of the costs when they identify spending that will quality for FMAP support while the costs of many non-Medicaid initiatives must be covered fully by the states’ taxpayers.

Persons eligible for certain Social Security Act welfare programs were made “categorically eligible” for Medicaid at enactment. This group includes persons receiving support under the terms of the now-repealed Aid to Families with Dependent Children (AFDC) program along with Supplement Security Income (SSI) recipients. After enactment, Congress required states to provide Medicaid coverage to numerous additional populations, including low-income pregnant women, low-income children, and low-income Medicare beneficiaries.

Further, the law specified from the beginning that persons deemed “medically needy” were to be eligible too. This pathway meant that elderly persons who were receiving Social Security but otherwise had limited means could often qualify for coverage of their nursing home expenses, which was listed among the mandatory services states would be required to provide to qualified beneficiaries.

With both Congress and the states mostly aligned in favoring ever-higher enrollment in the program, the Medicaid-covered population steadily increased over the past half century as shown in Figure 1. The ACA expansion, which as approved in 2010 and was not supported by both parties, dramatically increased enrollment by offering to states a generous FMAP (the originally plan was for mandatory adoption by the states, but the Supreme Court intervened). Enrollment reached a peak of nearly 30 percent of the entire U.S. population in 2023 with pandemic-era rules temporarily suspending eligibility checks, but has since fallen back to levels more consistent with the pre-2020 trend.

Figure 1. Medicaid Enrollment as a Percentage of Total U.S. Population

Sources: MACPAC and Census Bureau

Combined federal and state spending on the program has risen in tandem with enrollment, as shown in Figure 2. In the 1970s, total Medicaid spending was less than 1.0 percent of annual GDP. Today, it is approaching 3.5 percent of GDP, and could go much higher in future years with the aging of the population.

Figure 2. Total (Federal and State) Medicaid Spending (% of GDP)

Sources: MACPAC and CBO

As Medicaid spending has grown, it has become a sprawling enterprise with layers and layers of complexity. For instance:

• Through spending on Disproportionate Share Hospital (DSH) and favorable treatment of federally-qualified health centers (FQHCs), Medicaid is seen as providing a financial backstop to a vast network of service providers. Any effort to reduce the financial burden associated with these providers is described as threatening access to care for millions of Medicaid beneficiaries.
  
  
• Medicaid is the nation’s largest payer of nursing home claims, with more than 60 percent of residents qualifying for program support of their costs.
  
  
• Medicaid provides coverage to 13.6 million people who are also eligible for Medicare. Providing effective coordination of the two sources of insurance for these beneficiaries has proven to be an elusive objective, with scores of efforts attempted going back three decades. The results have not been stellar, with only about 10 percent of the “dually eligible” in arrangements that actively coordinate their Medicaid and Medicare benefits.
  
  
• Medicaid spent $196 billion for 6 million beneficiaries who qualify for home and community-based services (HCBS) focused on non-medical support. These are persons with various disabilities or the frail elderly who need assistance with activities of daily living (ADLs) on an on-going basis.
  
  
• Nearly every states has a federally-approved section 1115 waiver which allows them to administer various aspects of their Medicaid programs outside of the standard terms. For instance, states can use waivers to expand eligibility beyond what is allowable without a waiver, or to pay for services not normally covered by the program.
  
  
• Medicaid is a major purchaser of non-emergency medical transportation (NEMT) services, with an expected cost of nearly $16 billion in 2028. This spending pays for transportation to medical appointments for qualified beneficiaries.
  
  
• Medicaid pays for the premiums and cost-sharing of millions of low-income Medicare beneficiaries irrespective of the need for long-term services and supports.
  
  
• In some states, Medicaid is also now paying for housing support and food assistance.

When considering future changes to Medicaid, federal policymaker might consider designing proposed changes based on the program’s multiple constituencies. There are five main groups:

• Health Insurance Enrollees. For non-elderly and non-disabled adults and their children, Medicaid is their health insurance program in much the same way that millions of workers enroll in coverage offered by their employers. Medicaid provides financial security in case of major expenses and access to the types of routine and primary care services healthy populations require. Reforms to Medicaid for these beneficiaries should focus on better coordination of this coverage with employer-sponsored insurance (ESI) and the ACA markets to allow for less disruptive transitions when individuals get better paying jobs. The enrollment systems for each of these platforms should be coordinated and easier to navigate for the applicants.
  
  
• Non-Elderly Disabled Living in Communities. Medicaid is a major source of support for the population with significant long-term disabilities. It pays for their medical care but also for other support that allows many of them to live in their communities rather than in institutions.

• Nursing Home Residents. As noted, Medicaid is the primary source of insurance coverage for 60 percent of the disabled and elderly requiring 24-hour institutional care. Reform of this aspect of the program requires prospective changes that might take a decade or more to deliver savings. For instance, states could tighten enforcement of draw down rules aimed at preventing non-poor individuals from qualifying for Medicaid through the dissipation of assets.
  
  
• The Frail Elderly Living in Their Communities. There are many frail elderly persons who qualify for both Medicare and Medicaid but are not yet residing in nursing homes. The programs still do not provide good coordination of care for this population to prevent conditions from worsening to the point that a nursing home becomes the only answer. Reform should focus on forcing coordination of coverage for this population so that there is less unnecessary use of expensive institutionalized care.
  
  
• Medicare Savings Program Beneficiaries. Although many of the dually eligible are frail and need social support to remain in their homes and communities, some are eligible only for Medicaid coverage of some or all of their Medicare out-of-pocket costs. Qualified Medicare Beneficiaries (QMBs) are eligible for Medicaid coverage of their premiums and cost-sharing, while other groupings get coverage for their Medicare part B premiums. Reforms to this aspect of Medicaid should focus on what might be done to help the beneficiaries take more control of the financial support provided on their behalf.

Before the changes approved by Congress in July 2025, the Congressional Budget Office (CBO) estimated federal Medicaid spending over the next decade at $8.2 trillion. The new law is expected to reduce that total by $911 billion, or about 11 percent. It is unlikely that these changes represent the final word on the future of the program, if for no other reason than some of the saving might never be realized.

In future efforts, it will be important for federal policymakers to approach Medicaid with its diverse constituencies in mind. As a starting point, it may be useful to think about the program’s role as a source of health insurance for the non-disabled, non-elderly population as running on an entirely different track from its support for the disabled and frail elderly.