In the Balanced Budget Act of 1997, Congress inserted a provision that froze the number of Medicare-funded graduate medical education residency positions at each hospital at its 1996 level. The provision was intended as a cost-control measure — a way to prevent hospitals from expanding residency programs and billing Medicare for the training costs. Nearly three decades later, that single legislative decision remains the most consequential structural cause of the physician shortage in the United States, including the 8,000-physician gap in emergency medicine documented by ACEP in 2026.
The United States trains its physicians through a system of graduate medical education — residency and fellowship programs in which medical school graduates spend three to seven years in supervised clinical training before they are eligible for independent practice. The federal government, through Medicare, pays the majority of the cost of this training. In 1997, Congress decided that it was paying for too many residents and capped the number it would fund. The U.S. population has grown by 70 million people since then. The number of Medicare-funded residency slots has barely moved.
Graduate medical education in the United States is funded through a complex system of federal, state, and private payments. The largest single source of GME funding is Medicare, which pays teaching hospitals two types of payments: direct graduate medical education (DGME) payments, which cover the direct costs of resident training (salaries, benefits, supervision), and indirect medical education (IME) payments, which compensate teaching hospitals for the higher costs associated with treating complex patients in an academic environment.
In fiscal year 2024, Medicare spent approximately $16.7 billion on GME — roughly $150,000 per resident per year in DGME payments alone, with IME payments adding substantially more. This funding is critical to the financial viability of residency programs: without it, most hospitals could not afford to train residents, and the physician pipeline would collapse.
The number of Medicare-funded residency slots at each hospital is determined by its "cap" — a number established by the Balanced Budget Act of 1997 and based on the hospital's 1996 resident count. Hospitals can train more residents than their cap, but Medicare will not pay for the additional slots. In practice, most hospitals train at or near their cap, because training unfunded residents is expensive and produces no Medicare reimbursement.
The Balanced Budget Act of 1997 was one of the most significant pieces of fiscal legislation in a generation. Passed by a Republican Congress and signed by President Clinton, it was designed to balance the federal budget by 2002 — a goal it largely achieved. It contained hundreds of provisions affecting Medicare, Medicaid, and other federal programs. The GME cap was one of the smaller provisions, tucked into a bill that ran to hundreds of pages.
The rationale for the GME cap was straightforward: in the mid-1990s, many health policy analysts believed that the United States was training too many physicians. The managed care revolution was expected to reduce the demand for physician services by emphasizing primary care and preventive medicine over specialist care. The Graduate Medical Education National Advisory Committee had warned in 1980 that the country faced a physician surplus. By 1997, that warning seemed prescient: HMO enrollment was surging, and some analysts projected that the country would have 165,000 more physicians than it needed by 2000.
The GME cap was intended to prevent hospitals from expanding residency programs in anticipation of continued growth in Medicare payments. It was not intended to permanently freeze physician training at 1996 levels — the expectation was that Congress would revisit the cap as circumstances changed. Congress did not revisit the cap in any meaningful way for more than two decades.
The Managed Care Forecast That Never Materialized
The physician surplus projections of the 1990s were based on the assumption that managed care would fundamentally restructure how Americans received healthcare — reducing specialist utilization, emphasizing prevention, and dramatically cutting the demand for physician services. The managed care revolution stalled. HMO enrollment peaked in 1999 and declined sharply after a patient backlash. The physician surplus never materialized. The GME cap remained. The country that had frozen its physician training pipeline in anticipation of a surplus it never had was now facing a shortage it had not prepared for.
The consequences of the GME cap were not immediately apparent. In the late 1990s and early 2000s, the physician workforce was adequate to meet demand — the baby boom generation had not yet reached the age of peak healthcare utilization, and the physicians trained before the cap was imposed were still in active practice. The shortage built slowly, masked by the large cohort of physicians trained in the 1970s and 1980s who were still working.
By the mid-2010s, the trajectory was unmistakable. The Association of American Medical Colleges began publishing annual workforce projections showing a growing gap between physician supply and demand. The 2016 AAMC report projected a shortage of between 61,700 and 94,700 physicians by 2025. The 2024 report revised that projection upward: a shortage of between 86,000 and 124,000 physicians by 2036, driven by the retirement of the baby boom physician cohort and the aging of the patient population.
Emergency medicine has been particularly hard hit. The specialty grew rapidly in the 1990s and 2000s, driven by increasing emergency department utilization and the recognition of emergency medicine as a distinct specialty requiring specialized training. But the GME cap limited the growth of emergency medicine residency programs, and the specialty has never had enough trained physicians to staff the nation's emergency departments. The 2026 ACEP Workforce Report documented a current shortage of approximately 8,000 emergency physicians, with the gap projected to widen to 12,000 by 2030 absent significant intervention.
Congress has made several attempts to address the GME cap, with limited success. The most significant was the Resident Physician Shortage Reduction Act, which has been introduced in multiple sessions of Congress since 2019. The bill would add 14,000 new Medicare-funded residency slots over seven years — a meaningful increase, but one that critics note would still leave the country far short of the physician supply needed to meet projected demand.
The bill has passed the House of Representatives twice but has never received a floor vote in the Senate. Opposition has come from fiscal conservatives concerned about the cost (approximately $10 billion over ten years), from some managed care advocates who argue that more physicians will drive up healthcare spending, and from hospital systems that benefit from the current scarcity of physicians and the leverage it gives them in employment negotiations.
The Consolidated Appropriations Act of 2021 included a provision adding 1,000 new Medicare-funded residency slots — the first significant increase in the cap since 1997. The slots were distributed to hospitals in rural areas and those serving high proportions of low-income patients. The increase was welcomed by physician workforce advocates but described as "a drop in the bucket" by AAMC, which estimated that the country needed 10 to 15 times as many new slots to meaningfully address the shortage.
The GME cap's impact on emergency medicine is compounded by several factors specific to the specialty. Emergency medicine residencies are three to four years long — shorter than many other specialties — which means that the pipeline can respond more quickly to increased slot availability than, say, neurosurgery or cardiothoracic surgery. But the cap has prevented emergency medicine programs from expanding to meet demand, and the shortage of emergency physicians has been filled in part by physicians trained in other specialties who practice emergency medicine without formal EM training.
The ACEP 2026 Workforce Report estimates that approximately 30% of physicians currently working in emergency departments are not emergency medicine residency-trained. This is not necessarily a quality problem — many of these physicians are highly competent — but it reflects the degree to which the EM workforce has had to improvise in the face of a training pipeline that cannot meet demand.
The shortage also has geographic dimensions. Rural and underserved communities, which already struggle to attract physicians, are disproportionately affected by the EM shortage. The closure of rural emergency departments — more than 180 since 2010 — is driven in part by the inability to recruit and retain emergency physicians in communities that cannot offer the compensation packages available in urban and suburban markets.
The consensus among physician workforce researchers is that lifting the GME cap is necessary but not sufficient to address the physician shortage. Even if Congress were to add 14,000 new residency slots tomorrow, it would take seven to ten years for those slots to produce practicing physicians — and the shortage is acute now. Short-term solutions include expanding the scope of practice for nurse practitioners and physician assistants, increasing the use of telemedicine to extend physician reach, and reforming immigration policies to make it easier for internationally trained physicians to practice in the United States.
Longer-term solutions require sustained investment in the physician training pipeline: more residency slots, more medical school capacity, and better distribution of physicians to underserved communities through loan repayment programs and service obligations. The cost of these investments is real — but so is the cost of the shortage. Emergency departments that cannot staff their shifts, rural communities that lose their only hospital, and patients who wait hours for care that should take minutes: these are the costs of the 1997 decision, compounding with interest for nearly three decades.
1997
Year Congress froze Medicare-funded GME slots at 1996 levels
8,000
Current emergency physician shortage (ACEP 2026)
12,000
Projected EM physician shortage by 2030
1,000
New GME slots added by Congress in 2021 — first increase since 1997
14,000
New slots proposed by Resident Physician Shortage Reduction Act
$16.7B
Annual Medicare GME spending (FY 2024)