fb-pixelCan we afford Medicare for All? - The Boston Globe Skip to main content
Opinion | Adam Gaffney, David Himmelstein, and Steffie Woolhandler

Can we afford Medicare for All?

Illustration by Globe Staff; Adobe

Would Medicare for All break the bank? Proponents contend that it could cover the uninsured — and upgrade coverage for just about everybody else — without swelling our nation’s already hefty health care bill. But critics retort that the costs of such an expansion would be exorbitant, even prohibitive. Too often, however, this debate plays out in the absence of good evidence. But our new research, published Tuesday in the Annals of Internal Medicine, may change that.

Universal coverage might increase costs for one big reason: When the uninsured get coverage (or the underinsured get better coverage), they’re likely to use more health care. For example, a study begun in the 1970s randomized thousands of people into different health care plans, and found that those with better coverage used more health care. This sort of evidence drives estimates that Medicare for All, by expanding coverage, could cause US health care spending — already at $3.5 trillion a year — to further swell.


And yet, nations with universal coverage spend much less than the United States on health care. What explains this paradox? To find out, we analyzed how the two biggest coverage expansions in US history — the implementation of Medicare in 1966 and the Affordable Care Act in 2014 — affected hospital utilization. In both of these expansions, about 10 percent of Americans gained coverage — a similar percentage as would gain coverage under Medicare Care for All. But contrary to expectations, neither Medicare nor the ACA caused an overall increase in hospital use. Before Medicare’s startup, the hospitalization rate was 12.8 per hundred Americans. Afterward, it was 12.7. Similarly, before the ACA, the hospitalization rate was 9.4 per capita. After, it was 9.0.

As expected, however, the newly-covered did get more care. With Medicare, hospitalizations increased among the elderly (about half of whom were uninsured before Medicare), particularly among low-income seniors. Yet those increases were counterbalanced by small — nearly imperceptible — reductions in hospitalizations among younger and wealthier people. Effects after the ACA were more subtle, but we again saw evidence that it shifted admissions from wealthier Americans to low- and middle-income ones, and from the healthy to the sick. But, again overall hospital use didn’t rise.


The most likely explanation for these shifts in utilization is “Roemer’s Law,” named for Milton Roemer, a pioneering health policy expert who studied the relationship between the number of hospital beds in a community and the use of those beds. As the number of beds rises, he asserted, so does the number of nights that patients spend in them. Or as he put it back in 1959, “a hospital bed built is a filled bed.” Roemer’s Law has been borne out by many studies over the years, and it implies that if the number of beds in a community is finite, increases in insurance coverage usually won’t translate into more bed use.

As physicians, we have seen Roemer’s Law in action. While some hospitalizations are due to emergencies, many are elective. And when hospitals have many empty beds, the number of elective procedures, and hospitalizations, rises. Elective procedures are not necessarily a good thing, and well-insured patients sometimes receive unnecessary, even harmful care. For instance, while implanting defibrillators or stents into diseased hearts can be lifesaving, these potentially dangerous devices are sometimes overused, inserted into many patients who won’t benefit from them. The same is true of many other tests and surgical procedures, leading the Institute of Medicine to conclude that nearly one-third of medical spending is wasted.

It’s not surprising then, that when expanded coverage caused small reductions in hospitalizations for wealthier, healthier people, no one seemed to complain, or for that matter, even notice. Indeed, by producing offsets in unnecessary care, universal coverage might actually benefit those of us lucky enough to already have good insurance.

Overall, our finding that coverage expansions do not cause a surge in hospital use is reassuring. It implies that universal coverage via Medicare for All could enhance care for those who need it, without breaking the bank. Conversely, it suggests that shrinking coverage — for instance, by repealing or overturning the ACA — would save little in health care costs, even as it deprived millions of the care they need.


Past projections of the likely costs of universal coverage for the United States assumed that expanded coverage would increase hospital use and raise costs. But that assumption is wrong. The history of Medicare and the ACA, together with the experiences of nations with national health insurance around the globe, shows that covering everyone is not just the right thing to do — it’s also something we can afford to do.

Dr. Adam Gaffney is an instructor in medicine at Harvard Medical School and a physician at the Cambridge Health Alliance. Dr. Steffie Woolhandler is a professor of public health at the University of New York at Hunter College and lecturer at Harvard Medical School. Dr. David Himmelstein is professor of public health at the University of New York at Hunter College and a lecturer at Harvard Medical School. They are also members of Physicians for a National Health Program, which advocates for Medicare for All. The organization was not involved in the study.