Mandatory Spending

Multiple Budget Functions

Consolidate and Reduce Federal Payments for Graduate Medical Education at Teaching Hospitals

CBO periodically issues a compendium of policy options (called Options for Reducing the Deficit) covering a broad range of issues, as well as separate reports that include options for changing federal tax and spending policies in particular areas. This option appears in one of those publications. The options are derived from many sources and reflect a range of possibilities. For each option, CBO presents an estimate of its effects on the budget but makes no recommendations. Inclusion or exclusion of any particular option does not imply an endorsement or rejection by CBO.

Billions of Dollars 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2019-
Change in Outlays  
  Establish a grant program, with growth of grant based on the CPI-U 0 -1.4 -1.9 -2.5 -3.1 -3.7 -4.3 -4.9 -5.6 -6.6 -8.9 -34.0
  Establish a grant program, with growth of grant based on the CPI-U minus 1 percentage point 0 -1.4 -2.0 -2.8 -3.5 -4.3 -5.1 -5.9 -6.7 -7.9 -9.7 -39.5

This option would take effect in October 2019.
CPI-U = consumer price index for all urban consumers.


Under certain circumstances, hospitals with teaching programs can receive funds from Medicare and Medicaid for costs related to graduate medical education (GME). Medicare's payments cover two types of costs: those for direct graduate medical education (DGME) and those for indirect medical education (IME). DGME costs are for the compensation of medical residents and institutional overhead. IME costs are other teaching-related costs—for instance, those associated with the added demands placed on staff as a result of teaching activities and the greater number of tests and procedures ordered by residents as part of the educational process. As for funding provided by Medicaid, the federal government matches a portion of what state Medicaid programs pay for GME. The Congressional Budget Office estimates that total mandatory federal spending for hospital-based GME in 2018 was more than $15 billion, of which roughly 80 percent was financed by Medicare and the remainder by Medicaid. That spending is projected to grow at an average annual rate of 5.5 percent from 2020 through 2028 (about 3 percentage points faster than the average annual growth rate of the consumer price index for all urban consumers, or CPI-U). Teaching hospitals also receive funding from other federal agencies—which is discretionary rather than mandatory spending—as well as funding from private sources.

Medicare's payments for DGME are based on three factors: a hospital's costs per resident in a base year, indexed for subsequent inflation; the hospital's number of residents, which is subject to a cap that was first enacted in the Balanced Budget Act of 1997; and the share of total inpatient days at the hospital that is accounted for by Medicare beneficiaries. Payments for IME are made under Medicare's hospital inpatient prospective payment system as a percentage add-on to the base payment and reflect a hospital's teaching intensity (such as its ratio of full-time equivalent residents to the number of beds). In the Medicaid program, GME payments are considered to be a part of supplemental payments and states are allowed, but not required, to make Medicaid payments for GME. Each state determines its own level of Medicaid payments for GME and how those payments will be made. For example, some states base their GME payments on Medicare's methodology or on a modified form of that methodology, whereas other states provide lump-sum payments for GME. Those payments are subject to the same federal matching rates as other Medicaid spending and are subject to upper payment limits for Medicaid spending.


Beginning in October 2019, this option would consolidate all mandatory federal spending for GME into a grant program for teaching hospitals. Payments would be apportioned among hospitals according to the number of residents at a hospital (up to its existing cap) and the share of the hospital's inpatient days accounted for by Medicare and Medicaid patients. Total funds available for distribution in 2020 would be fixed at an amount equaling the sum of Medicare's 2018 payments for DGME and IME and the federal share of Medicaid's 2018 payments for GME. Total funding for the grant program would then grow at the rate of inflation. CBO examined two alternative measures of inflation. Under the first alternative, funding for the grant program would grow with the CPI-U; and under the second alternative, funding for the grant program would grow with the CPI-U minus 1 percentage point per year.

Effects on the Budget

In CBO's estimation, the first alternative would reduce mandatory spending by $34 billion between 2020 and 2028. Using the amount of federal funding for GME in 2018 to establish the total funding available in 2020 would cause a downward shift in the funding stream—relative to CBO's projection of federal spending on GME under current law—that would reduce federal spending by $17.5 billion between 2020 and 2028. Increasing GME funding at the rate of the CPI-U, rather than at the rate of growth CBO projects under current law, would yield an additional $21.4 billion reduction in federal spending over that period. However, CBO expects that those savings would be partially offset by a $4.8 billion increase in federal Medicaid spending. Many states make supplemental payments to hospitals that serve as safety-net hospitals (medical facilities that provide care regardless of a person's ability to pay) and to those that provide charity care or other types of community benefits. Those supplemental payments are eligible for the same federal matching payments as other types of Medicaid-covered services. CBO anticipates that some states would make separate supplemental payments to replace a portion of lost hospital revenue for some or all of their teaching hospitals, which would partially offset the reduction in federal spending for Medicaid.

CBO estimates that the second alternative would reduce spending by $40 billion between 2020 and 2028. Under that alternative, the reduction in spending associated with the downward shift in the funding stream would be the same as under the first alternative, $17.5 billion. Increasing federal GME funding at the rate of the CPI-U minus 1 percentage point per year would yield a greater reduction in spending than would the first alternative, or $27.6 billion between 2020 and 2028. The offsetting increase in federal Medicaid spending over that period would also be larger than under the first alternative and is estimated to be $5.4 billion.

By 2028, the savings associated with the first alternative would represent about 16 percent of projected federal spending for GME under current law, whereas savings associated with the second alternative would represent about 19 percent. By consolidating federal funding for medical education, this option could reduce the federal government's costs of administering the program. Any such administrative efficiencies would accrue to discretionary spending and therefore are not included in the estimate of changes to mandatory spending described above.

The option would not change the existing caps on the number of subsidized slots for residents. Altering those caps would not change the budgetary effects because total federal payments for GME under this option would not depend on the number of residents. Removing those caps might allow the existing slots to be allocated more efficiently among hospitals, but it also would create an incentive for hospitals to expand their residency programs in an attempt to receive a larger share of the total. The net effects on hospitals' residency programs would be difficult to predict.

Two sources of uncertainty in the estimates relate to the projected payment amounts for GME and the projected growth in the CPI-U from 2019 through 2028. In the event that the actual growth rates for either DGME or IME were higher or lower than the projected rates, the estimated savings would be greater or lesser than those using CBO's current baseline projections. Also, to the extent that the difference between actual growth in the CPI-U and the growth in projected payments for GME occurring under current law turned out to be greater than CBO has estimated, the savings under the option would be larger, and vice versa. A third source of uncertainty is anticipating and projecting the extent to which states would offset the reductions to GME payments, for example, by making separate supplemental payments to teaching hospitals that experience reductions in GME funding.

Other Effects

An argument for reducing the overall subsidy for GME is that federal payments under current law exceed hospitals' actual teaching costs. The Medicare Payment Advisory Commission (MedPAC) has consistently found that the IME adjustment is greater than hospitals' estimated indirect costs of providing medical education. In a 2016 analysis, MedPAC estimated that an IME adjustment about one-third the size of the current one would reflect the indirect costs that teaching hospitals actually incur (MedPAC 2016). That analysis suggested that a smaller subsidy would not unduly affect hospitals' teaching activities. A smaller subsidy also would reduce the incentive for hospitals to hire a greater number of residents than necessary. Another argument in favor of consolidating GME funding to hospitals is that unifying the funding for GME could allow for a broader policy discussion about the ways in which medical education is funded.

An argument against the option is that reducing the federal subsidy for GME could lead teaching hospitals to shift the composition of their residency programs toward specialists and away from primary care residents. Hospitals made such a shift after the caps on Medicare-funded residency slots were enacted because employing specialists tends to be more profitable. If hospitals responded to further reductions in federal GME subsidies in the same way, they could exacerbate concerns about a shortage of primary care physicians in the future. Alternatively, hospitals might respond to the reduced subsidy by lowering residents' compensation and making them responsible for more of the cost of their medical training.

Another argument against the option is that some teaching hospitals might use part of their GME payments to fund care for uninsured people. The option could therefore disproportionately affect teaching hospitals that treat a larger number of uninsured patients. Furthermore, states could lose some discretion to direct Medicaid GME payments to hospitals because the federal government would be administering the grant program. Under those circumstances, states would no longer receive federal matching for those funds and might choose to reduce their GME payments to hospitals. However, that reduction would be mitigated if states instead shifted their GME payments to other types of supplemental payments (which are subject to federal matching). Finally, if hospitals' costs grew faster than GME payments, hospitals and residents might bear an increasing share of the costs of operating a residency program over time.