Under current law, the government estimates that the U.S. will spend about one-sixth of G.D.P. on health care this year, with those costs divided between the federal government, individuals, employers and state governments.
This estimate, from an economist close to the Bernie Sanders 2016 campaign, estimates the largest savings from converting to Medicare for all.
This estimate assumes that Medicare for all will pay all medical providers the same amounts Medicare pays now. That decision means it would lower total health spending, but its author thinks the real system would have to pay higher prices.
This estimate assumes that Medicare for all would need to pay all medical providers higher rates than Medicare pays them now.
The Urban Institute estimate includes a limit on how many more doctors’ visits people will be able to make. Even so, it projects a substantial increase in spending under Medicare for all.
Even without including all the costs for long-term care, which some Medicare for all proposals include, this estimate still finds that Medicare for all would cost substantially more than the current system.
How much would a “Medicare for all” plan, like the kind endorsed by the Democratic presidential candidates Bernie Sanders and Elizabeth Warren, change health spending in the United States?
Some advocates have said costs would actually be lower because of gains in efficiency and scale, while critics have predicted huge increases.
We asked a handful of economists and think tanks with a range of perspectives to estimate total American health care expenditures in 2019 under such a plan. The chart at the top of this page shows the estimates, both in composition and in total cost.
In all of these estimates, patients and private insurers would spend far less, and the federal government would pay far more. But the overall changes are also important, and they’re larger than they may look. Even the difference between the most expensive estimate and the second-most expensive estimate was larger than the budget of most federal agencies.
The big differences in the estimates of experts reflect the challenge of forecasting a change of this magnitude; it would be the largest domestic policy change in a generation.
The proposals themselves are vague on crucial points. More broadly, any Medicare for all system would be influenced by the decisions and actions of parties concerned — patients, health care providers and political actors — in complex, hard-to-predict ways. But seeing the range of responses, and the things that all the experts agree on, can give us some ideas about what Medicare for all could mean for the country’s budget and economy.
These estimates come from:
Gerald Friedman, a professor of economics at the University of Massachusetts, Amherst, whose estimates were frequently cited by the Bernie Sanders presidential campaign in 2016.
Charles Blahous, a senior research strategist at the Mercatus Center at George Mason University, and a former trustee of Medicare and Social Security.
Analysts at the RAND Corporation, a global policy research group that has estimated the effects of several single-payer health care proposals.
Kenneth E. Thorpe, the chairman of the health policy department at Emory University, who helped Vermont estimate the costs of a single-payer proposal there in 2006.
Analysts at the Urban Institute, a Washington policy research group that frequently estimates the effects of health policy changes.
Right now, individuals and employers pay insurance premiums; people pay cash co-payments for drugs; and state governments pay a share of Medicaid costs. In a system like one introduced as a bill by Mr. Sanders or another from Representative Pramila Jayapal and the Congressional Progressive Caucus, nearly all of that would be replaced by federal spending. That’s why some experts describe such a system as single-payer. (Other Democrats who are supporting coverage expansion through Medicare have offered more modest proposals that would preserve some out-of-pocket spending and a role for private insurance.)
The economists made their calculations using different assumptions and methods, and you can read more about those methods at the bottom of this article.
These two estimates, for example, from the Mercatus Center and the Urban Institute, differ by about $730 billion per year, roughly 3 percent of G.D.P. The two groups don’t often agree on public policy — Mercatus tends to be more right-leaning and Urban more left-leaning.
The biggest difference between the Mercatus estimate and the Urban one is related to how much the new system would pay doctors, hospitals and other medical providers for health services. Mr. Friedman’s estimate, the least expensive of the group, assumed that the government could achieve the largest cost savings on both prescription drugs and administrative spending.
Pay too little, and you risk hospital closings and unhappy health care providers. Pay too much, and the system will become far more expensive. Small differences add up.
| friedman | blahous | thorpe | urban | rand |
|---|---|---|---|---|
| 6% | 0% | 5% | 7% | 9% |
In our current system, doctors, hospitals and other health care providers are paid by a number of insurers, and those insurers all pay them slightly different prices. In general, private insurance pays medical providers more than Medicare does. Under a Medicare for all system, Medicare would pick up all the bills. Paying the same prices that Medicare pays now would mean an effective pay cut for medical providers who currently see a lot of patients with private insurance.
For a Medicare for all system to save money, it needs to reduce the health care industry’s income somewhat. But if rates are too low, hospitals already facing financial difficulties could be put out of business.
Neither Mr. Sanders’s legislation nor the Jayapal House bill specify what the Medicare for all system would pay, but they say that Medicare would establish budgets and payment rates. So our estimators offered their best guess of what they thought such a plan might do.
Mr. Thorpe said he picked a number higher than current Medicare prices for hospitals, because he thought anything lower would be unsustainable. Mr. Blahous said he constructed his starting estimate at precisely Medicare rates, though he thought the real number would most likely be higher. He also reran his calculations with a more generous assumption: At 111 percent of Medicare, around the average amount all health insurers pay medical providers now, the total shot up by hundreds of billions of dollars, about an additional 1.5 percent of G.D.P.
By negotiating directly on behalf of all Americans, instead of having individual insurance companies and plans bargain separately, the government should be able to pay lower drug prices.
| friedman | blahous | thorpe | urban | rand |
|---|---|---|---|---|
| 31% | 12% | 4% | 20% | 11% |
Patients in the United States pay the highest prices in the world for prescription drugs. That’s partly a result of a fractured system in which different payers negotiate separately for drug benefits. But it also reflects national preferences: An effective negotiator needs to be able to say no, and American patients tend to want access to the widest array of cutting-edge drugs, even if it means paying more.
A Medicare for all system would have more leverage with the drug industry because it could bargain for the whole country’s drug supply at once. But politics would still be a constraint. A system willing to pay for fewer drugs could probably get bigger discounts than one that wanted to preserve the current set of choices. That would mean, though, that some patients would be denied the medications they want.
All of our economists thought a Medicare for all system could negotiate lower prices than the current ones. But they differed in their assessments of how cutthroat a negotiator Medicare would be. Mr. Friedman thought Medicare for all could reduce drug spending by nearly a third. The Urban team said the savings would be at least 20 percent. The other researchers imagined more modest reductions.
By expanding coverage to the uninsured, adding new benefits and wiping out cost sharing, Medicare for all would encourage more Americans to seek health care services.
| friedman | blahous | thorpe | urban | rand |
|---|---|---|---|---|
| 7% | 11% | 15% | — | 8% |
Medicare for all would give insurance to around 28 million Americans who don’t have it now. And evidence shows that people use more health services when they’re insured. That change alone would increase the bill for the program.
Other changes to Medicare for all would also tend to increase health care spending. Some proposals would eliminate nearly all co-payments and deductibles. Evidence shows that people tend to go to the doctor more when there’s no such cost sharing. The proposed plans would also add medical benefits not typically covered by health insurance, such as dental care, hearing aids and optometry services, which would increase their use.
The economists differ somewhat in how much they think people would increase their use of medical services. (Because of the way the Urban Institute team’s estimate was calculated, it couldn’t easily provide a number for this question.)
Right now, the health care system is complicated, with lots of different payers and ways to negotiate prices and bill for services. A single payment system could save some money by simplifying all that.
| friedman | blahous | thorpe | urban | rand |
|---|---|---|---|---|
| 2% | — | 6% | 6% | 5% |
The complexity of the American system means that administrative costs can often be high. Insurance companies spend on negotiations, claims review, marketing and sometimes shareholder returns. One key possible advantage of a Medicare for all system would be to strip away some of those overhead costs.
But estimating possible savings in management and administration is not easy. Medicare currently has a much lower administrative cost share than other forms of insurance, but it also covers sicker people, distorting such comparisons. Certain administrative functions, like fraud detection, can have a substantial return on investment.
The economists all said administrative costs would be lower under Medicare for all, but they differed on how much. Those differences amount to percentage points on top of the differing estimates of medical spending. On this question, there was rough agreement among our estimators that administrative costs would be no higher than 6 percent of medical costs, a number similar to the administrative costs that large employers spend on their health plans. Mr. Blahous said a 6 percent estimate would probably apply to populations currently covered under private insurance, but did not calculate an overall rate.
All of these estimates looked at the potential health care bill under a Sanders-style Medicare for all plan. In some estimates, the country would not pay more for health care, but there would still be a drastic shift in who is doing the paying. Individuals and their employers now pay nearly half of the total cost of medical care, but that percentage would fall close to zero, and the percentage paid by the federal government would rise to compensate. Even under Mr. Blahous’s lower estimate, which assumes a reduction in overall health care spending, federal spending on health care would still increase by 10 percent of G.D.P., or more than triple what the government spends on the military.
How that transfer takes place is one of the least well explained parts of the reform proposals. Taxation is the most obvious way to collect that extra revenue, but so far none of the current Medicare for all proposals have included a detailed tax plan. Even if total medical spending stayed flat over all, some taxpayers could come out ahead and pay less; others could find themselves paying more.
Raising revenue would require broad tax increases that are likely to be partly borne by the middle class, potentially impeding passage. Advocates, including Mr. Sanders, tend to favor funding the program with payroll taxes.
For some people, any increase in federal taxes might be more than offset by reductions in their spending on premiums, co-payments, deductibles and state taxes. There is evidence to suggest that premium savings by employers would also be returned to workers in the form of higher salaries. But, depending on the details, other groups could end up paying more in tax increases than they save in those reductions.
After Mr. Sanders’s presidential campaign released a tax proposal in 2016, the Urban Institute tried to calculate the effects on different groups. But it found that the proposed taxes would pay for only about half of the increased federal bill. That means that a real financing proposal would probably need to raise a lot more in taxes. How those are spread across the population would change who would be better or worse off under Medicare for all.
Our economists differed somewhat in their estimation methods. They also examined a couple of different Medicare for all proposals, though all the plans had the same major features.
Gerald Friedman calculated the cost of Medicare for all by making adjustments to current health care spending using assumptions he derived from the research literature. His measurements didn’t capture the behavior of individual Americans, but estimated broader changes as groups of people gained access to different insurance, and as medical providers earned a different mix of payments. A 2018 paper with his analysis of several different variations on Medicare for all is available here.
Kenneth E. Thorpe calculated the cost of Medicare for all by making adjustments to current health care spending using assumptions he derived from the research literature. His measurements didn’t capture the behavior of individual Americans, but estimated broader changes as groups of people gained access to different insurance, and as medical providers earned a different mix of payments. A 2016 paper with more of his findings on Mr. Sanders’s presidential campaign proposal is available here.
The Urban Institute built its estimates using a microsimulation model, which estimates how individuals with different incomes and health care needs would respond to changes in health insurance. The model does not consider the effects of policy changes on military and veterans’ health care or the Indian Health Service, so its totals assumed those programs would not change. It also measures limits on the availability of doctors and hospitals using evidence from the Medicaid program. The team at Urban that prepared the calculations includes John Holahan, Lisa Clemans-Cope, Matthew Buettgens, Melissa Favreault, Linda J. Blumberg and Siyabonga Ndwandwe. Its detailed report on Mr. Sanders’s presidential campaign proposal from 2016 is available here.
Charles Blahous calculated the cost of Medicare for all by making adjustments to current health care spending using assumptions he derived from the research literature. His measurements didn’t capture the behavior of individual Americans, but estimated broader changes as groups of people gained access to different insurance, and as medical providers earned a different mix of payments. His calculations were made based on Mr. Sanders’s 2017 Medicare for All Act, which indicated that states would continue to pay a share of long-term care costs. A 2018 paper with more of his findings is available here, and includes both sets of estimates for Medicare provider payments.
The RAND Corporation built its estimates by making adjustments to previous single-payer analyses. The original estimates used a microsimulation model, which estimates how individuals with different incomes and health care needs would respond to changes in health insurance. The RAND model, which it uses to estimate the effects of various health policy changes, is called RAND COMPARE. Calculations were made assuming a Medicare for all plan that offers coverage with no cost sharing and long-term care benefits. The RAND team that prepared the estimate includes Christine Eibner and Jodi Liu. A copy of the report is available here; Ms. Liu’s 2016 study of how different approaches to single-payer might affect its costs is here.