How to Save and Fix Medicare

Yevgeniy Feyman

Spring 2017

It is often said that the American health-care system is an example of the free market failing. One can quibble about what failure means in this context. But no amount of Marx and Engels should be able to convince anyone that the American health-care system is a "free market." In particular, Medicare — our socialized health-insurance scheme for the elderly and disabled — covers 55 million people. That's 17% of the American population, or roughly the population of England. The program accounts for 15% of the federal budget and 3% of our economy.

While popular, this system is beginning to show its age and poses an ever-growing burden to the country's finances. Changes to the program in the coming years, whether through legislation or through pilot programs created by executive action, are unavoidable. What is less inevitable is the direction that these changes will take. Some wish to see a "Medicare-for-all" system, which would implement a universal socialized health-insurance scheme. In the 2016 presidential election, this was a key pillar of Senator Bernie Sanders's health-care platform. Others would sooner see a full privatization of Medicare. A better way might straddle the middle. Whatever the economic or social benefits of either approach, political reality dictates that a more incremental, targeted strategy has a much better chance of succeeding.

For those who see value in a larger private-sector role in the Medicare program, this incremental approach may very well be the ideal. Indeed, it would exploit innovations from the private sector to protect a program that is — and will remain — a vital part of the American safety net. The key will be shaping these changes in a way that will maintain the core goal of the program — protecting seniors from catastrophic medical costs — while enabling private-sector innovations to make the program more efficient.

Successfully navigating these politically challenging waters would pay dividends for taxpayers, Medicare beneficiaries, and America's health-care system as a whole. 


Medicare is one of the American political system's confused, rambling answers to the call for socialized health insurance akin to that of other countries. But that doesn't mean that Medicare's design problems today are the result of accident or carelessness. Rather, they are the result of messy attempts to balance various concerns of interest groups, political factions, and American culture.

Attempts to create a socialized health-insurance scheme in the United States date back to the Progressive Era of Teddy Roosevelt. The Progressive Party platform in 1912 endorsed socialized health insurance, and the Bull Moose himself lobbied for sickness benefits as a state program. More formalized efforts resumed under Franklin Roosevelt, first as part of the broader Social Security legislation and later as the Wagner National Health Bill (both of which were opposed, of course, by the American Medical Association). These efforts, and later ones by Harry Truman, also failed to garner the needed support from Congress or from doctors.

The Great Society era, however, allowed for a unique compromise that helps explain what Medicare looks like today. The moral case for covering vulnerable populations, along with the growth in hospital admissions, the use of new and expensive medical technologies and techniques after World War II, and a growing economy (of which employer-sponsored coverage was a strong pillar), combined to spur the demand for health-insurance reform. The program that ultimately emerged from Congress in 1965 was the result of a series of compromises. These included efforts to protect the Social Security trust fund from being drained by Medicare (which resulted in a separate dedicated payroll tax) and an alternative called Eldercare — a state-administered voluntary program that included outpatient visits — proposed by the AMA and congressional Republicans. These two ideas were joined with a series of further compromises that nationalized the program's administration, offered concessions to physicians (reimbursement based on usual and customary fees) and hospitals (cost plus reimbursement), and created a separate program for coverage of low-income populations with matching funds for states (Medicaid).

Initial changes to the program were quick to be implemented as it became clear that its design was — to put it mildly — flawed and that costs would run far higher than projected. The program's retrospective reimbursement system (which essentially amounted to asking hospitals after the fact what their costs were) was a particular problem since it allowed hospitals to raise costs at the taxpayer's expense without much pushback. After a series of efforts to combat this, one of the biggest and longest-lasting changes to Medicare was implemented. Rather than paying hospitals based on what they claimed were their costs, Congress mandated the creation of an inpatient prospective payment system to be effective starting in October 1983.

This system grouped various procedures into diagnosis-related groups (DRGs) and paid a fixed rate for each DRG, regardless of the length of stay and regardless of what was actually provided during the hospital stay. This had the intended effect of slowing cost growth (in the 10 years after implementation, per-beneficiary costs grew 7% annually while in the 10 years prior they grew nearly 15%) but created new incentives for "upcoding" patients into higher-value DRGs.

This was far from the only change that Medicare has undergone in its 50 years. Individuals under age 65 with long-term disabilities were brought into the program under President Nixon. Hospice coverage came in the '80s, and a catastrophic benefit was created in 1988 (though it was repealed a year later).

The 1990s formalized the inclusion of private plans as an option in Medicare (then called Medicare+Choice) — which now stand to serve as the primary vehicle for further modernizing reforms. In 2003, a major overhaul of the program once again took place: Prescription-drug coverage was added through private insurers in the Part D program, and Medicare+Choice was substantially transformed and renamed Medicare Advantage (MA). Finally, in 2010, Obamacare made further changes to reimbursements in the program and reformed how MA plans are paid.

After all of these changes, today's Medicare program looks radically different than it did at its inception. Sixteen percent of the Medicare population is covered due to disabilities rather than age (up from 7% in 1973); over 20% are dually enrolled in Medicare and Medicaid; and roughly one-third of enrollees receive coverage through the MA program. Most beneficiaries, however, still face a benefit design based on mid-20th-century health insurance. 


None of the changes the program has gone through over the half-century since its creation has come easily. Because Medicare covers one vocal constituency (seniors) and funds several others (pharmaceutical companies, hospitals, and physicians), proposed changes often meet with stiff resistance.

The infamous "doc fix" offers a classic case. Prior to the 1990s, physician payments in Medicare were (as hospital payments once were) based on prevailing charges in the market. This had the same result as it did with hospital payments — everyone raised their prices. In 1989, legislators enacted a so-called "volume performance standard" (VPS), which modified payment growth rates based on whether service volume grew faster or slower than a target rate. Even this didn't put enough of a brake on cost growth to satisfy lawmakers' desires, however. From 1990 to '97 (the VPS's seven years of operation), per-beneficiary cost growth in Medicare exceeded real GDP by over four percentage points. The VPS was soon replaced with the "sustainable growth rate" (SGR) mechanism. The SGR took cost-growth calculation a step further, tying growth in physician payments to costs, the number of Medicare fee-for-service beneficiaries, changes in benefits, and the 10-year average growth rate of real GDP per capita.

Initially, the SGR functioned as designed, with per-beneficiary costs growing relatively slowly. By 2002, however, cost growth had picked up, and the Department of Health and Human Services was required to cut physician payments by 4.8%. They did just that, for the first and last time. In every subsequent year, physicians demanded relief from such cuts and threatened to stop accepting Medicare patients if relief was not forthcoming. Congress responded with the doc fix — "temporarily" overriding each year's required payment-rate reduction.

It took until 2015 to put an end to this pantomime, and that measure was as complex and cynical as the doc fix itself. The Medicare Access and CHIP Reauthorization Act of 2015 (MACRA) replaced the SGR and ended the doc-fix era by tying physician-payment growth to participation in value-based payment models created by the Centers for Medicare and Medicaid Services, and allowing payment reductions for physicians who don't meet certain quality goals. While paying physicians for value rather than volume is likely to be a significant improvement, it remains to be seen whether actual reductions in payments (which are not set to begin until 2019) will be any more politically realistic than the SGR's cuts. It is within the realm of possibility that MACRA will become yet another political football for lawmakers to toss around while deciding how best to placate interest groups and constituencies. 

Though the doc fix and physician payments are far from the only political footballs in Medicare, the SGR experience underscores the challenges of fixing problems in the program. It was clear for many years that the SGR hadn't been a legislative success. But the fact that many powerful interest groups relied on the predictable changes (or lack thereof) in payments under the law, and the ease with which Congress could override scheduled reductions, made reform difficult.

The doc-fix era and Medicare's broader history offer important lessons for future reforms that set lofty goals, which might fall prey to Congress's legislative whims. Lasting reform would need to address the program's flaws (and protect its successes) while taking seriously the simple reality that multiple, diverse constituencies are invested in Medicare and will rise up in force against any perceived harm.


There are two broad reasons for reforming Medicare. The first is to reduce costs in the program. This saves money for taxpayers and extends the program's solvency. Typically, this points to changes in benefit structures and payment schedules or to increases in revenue. The second reason for reform is to deliver better value to beneficiaries. Doing so might involve some benefit changes, but it also can include the various experiments being conducted to incentivize higher-value care.

Reforms that save money are incontrovertibly necessary in the near term. According to the 2016 annual report of the Medicare trustees, Medicare's Hospital Insurance (HI) trust fund, used to pay for inpatient expenditures, will exhaust its funds by 2028. Despite the much-heralded slowdown in Medicare spending per person, growth in Medicare's income (for the HI trust fund, this mostly means revenues raised through payroll taxes) is still expected to be slower than growth in total expenditures. From 2020 to 2025, for instance, the trustees expect expenditures to grow about 10 percentage points faster than income.

After the trust fund's exhaustion, Medicare would only be able to pay for 87% of required benefits. Medicare's actuaries note that, as of the issuance of their report, closing the program's 75-year actuarial deficit would require an immediate 25% increase in Medicare's payroll-tax rate (from 2.9% to 3.63%) or an immediate reduction of expenditures by 16%. Given that painful policy changes of this sort are usually implemented on some delay, these numbers would likely be larger in magnitude in a more realistic scenario.

Medicare's trust fund for Supplementary Medical Insurance (SMI), which is used to pay for care in a physician's office and for retail prescription drugs, doesn't face the same problem. Statutorily, the SMI trust fund is required to balance each year, with a portion coming from premiums and the rest from general revenue. Future growth in spending, however, will require increases in general revenue devoted to the SMI trust fund, as well as increases in beneficiary premiums. In 2015, for instance, revenue devoted to the SMI trust fund accounted for about 13.5% of personal and corporate federal income taxes. By 2030, it is expected to hit over 21%. This is driven mainly by the trustees' assumption that, again, expenditures will grow faster than the revenue base.

The point is simple: Without changes that either reduce cost growth or increase revenues, under current law, one major part of Medicare will be unable to pay out the benefits that it is expected to owe in the near future, and the other major part of Medicare will eat up an ever-larger share of general federal revenues. This means that, without far more deficit spending, a growing share of general tax revenue will be unavailable for other federal government priorities.

Reform that delivers better value to beneficiaries isn't necessarily urgent. Certainly there aren't any particular deadlines for such changes. However, Medicare's role as one of the largest health-care purchasers in the country means that the way the program functions has a disproportionately large effect on health-care markets more broadly. For instance, many insurers base reimbursements to physicians and hospitals on Medicare's payment schedule, often paying some multiple of Medicare payments. Medicare's prohibition on refusing coverage for therapies on the basis of cost effectiveness similarly bleeds into the rest of the health-insurance system, encouraging coverage of an unnecessarily broad range of drugs and procedures in the private insurance market.

Politically speaking, reforms that both reduce costs and improve value for beneficiaries are also those most likely to draw support from across the political spectrum. This means that comprehensive reform of the program will have to carefully address both challenges.


There is actually no shortage of reforms that are both politically feasible and could hit the dual targets of improved value and lower costs. These include everything from changing Medicare's benefit design to more accurately model modern-day private insurance, to extending mandatory rebates for certain drugs to Medicare patients. One idea in particular — commonly known as premium support or competitive bidding — has the potential to more radically change the structure of the Medicare program to improve its effectiveness and reduce its costs while still providing the benefit it now does to seniors. And this would not be a new idea in Medicare — it is key to how Medicare Advantage works today.

As noted earlier, individuals who become eligible for Medicare can opt to receive their benefits through a private insurance plan instead of the government-run program. That roughly one-third of beneficiaries receive coverage through a private plan is itself a testament to the popularity of these plans. An enrollee who chooses private coverage may pay an additional premium on top of the premium for traditional Medicare, or may have to pay no additional premium, depending on the plan he chooses. Recent research suggests that nearly 50% of MA plans offer "zero-premium" coverage. And MA plans typically also offer retail prescription-drug coverage for an additional premium as well.

Contrary to traditional Medicare, MA plans are permitted to use a variety of utilization-management strategies. For instance, most plans have provider networks that restrict which hospitals and physicians enrollees can use. Plans also typically have different benefit structures that look more like those in the private insurance market, with co-pays and co-insurance that can vary for different services. Out-of-pocket costs for beneficiaries are also required to be capped for each MA plan. Lastly, MA plans can also vary contract designs and payment schedules; recent research suggests that MA plans pay about 6% less than traditional Medicare for hospital services.

The methods used to calculate plan payments have changed significantly over time. Indeed, actual bidding, at the county level, was only introduced in 2006. Prior to that, plan payments were typically either paid at the rates of traditional Medicare (or slightly below), given "floor payments" that would guarantee at least a certain level of payment, or in some years simply received a fixed increase in payments. While a crude risk-adjustment mechanism was used initially, this accounted only for enrollee demographics. Since 2004, MA payments have become increasingly adjusted for enrollee health status to discourage plans from selecting only healthy enrollees.

Today, MA plans submit bids based on the expected cost of covering an average Medicare beneficiary. This bid is then adjusted for enrollee demographics and risk score (a measure used by the Centers for Medicare and Medicaid Services to quantify how sick an individual is) and is compared to a statutorily determined benchmark.

Prior to the Affordable Care Act, these benchmarks typically grew every year at a fixed rate, which often resulted in benchmarks coming in significantly higher than traditional Medicare costs. Under the ACA, however, county-level benchmarks are calculated based on where the county ranks in terms of its traditional Medicare costs. Counties then receive a benchmark equal to some percentage of traditional Medicare costs — ranging from 95% to 115%, with the highest-cost counties receiving the biggest reduction and lowest-cost counties receiving an increase.

Plans that bid above the benchmark are required to charge beneficiaries a premium equal to the difference between the benchmark and the bid. Plans bidding below this benchmark cost nothing extra for beneficiaries and receive a rebate based on the difference between the benchmark and their bid. This rebate must be used to offer supplemental benefits, to reduce the beneficiary's Part B or Part D premiums, or to reduce premiums for supplemental benefits.

Understanding the deficiency of this system is critical to understanding how a relatively straightforward reform could result in a substantial improvement.

A major flaw one might spot immediately is that the bidding system does not fully hold high-cost plans accountable for higher-than-average bids. Indeed, the fact that benchmarks are established based on administrative calculations discourages true competition among plans. Because lower- or higher-than-average bids do not affect benchmarks, plans face less incentive to bid competitively. This is likely one reason that, prior to Obamacare, MA benchmarks were routinely greater than traditional Medicare costs. Similarly, payments to MA plans were also consistently higher than bids (114% versus 102% on average in 2009, for instance).

Another flaw, which may be less apparent, involves the dynamics of decisions faced by Medicare beneficiaries. With an administrative benchmark, the beneficiary is not held accountable for making inefficient, costly decisions. Consider what happens when an MA plan bids below the cost of providing traditional Medicare coverage and is under or at the benchmark. Enrolling in that MA plan would mean paying the same premium that would be required for traditional Medicare, possibly with some supplemental benefits. At the margin the enrollee wouldn't face a penalty for choosing to receive coverage through traditional Medicare even though taxpayers would save money if that beneficiary chose the MA plan.

This has continued under the ACA, as well. Under the new rules for paying plans, for instance, a quarter of all counties will receive a benchmark set at 115% of traditional Medicare costs. This helps explain why, while MA plans are bidding 10% below traditional Medicare costs, they will be paid 100% of traditional Medicare costs in 2017. In these counties, beneficiaries who choose MA plans bidding at the benchmark will pay no higher premiums than those who choose traditional Medicare, despite the higher cost to taxpayers.

There are several reasons why this set-up makes little sense. For one, it is unlike any other element of our health-insurance system. If you work for an employer that offers a choice of insurance plans, you can bet that the higher-cost plans will demand a higher premium contribution from enrollees. Federal employees choose among several different plans in a marketplace and pay at least 25% of the cost of the plan that they select — again, requiring a larger contribution for more expensive plans. On the ACA's exchanges, too, competition between plans has determined how much an enrollee pays. While the law caps the cost of premiums relative to enrollees' incomes, enrollees choosing a more expensive plan are always required to pay more. And the Medicare Part D program, which provides retail prescription-drug coverage to Medicare beneficiaries, pays plans based on the difference between their estimated costs of providing coverage and enrollee premiums. Enrollees, however, pay a premium that factors in the difference between the plan's bid and the nationwide average bid. While enrollees in MA plans pay a higher premium for plans that bid above the benchmark, that is the only instance where enrollees are held accountable for selecting higher-cost plans.

Secondly, because benchmarks are set administratively, competition between plans has markedly less room to reduce costs. Of course, plans are disincentivized from bidding above the benchmark because enrollees would pay a higher premium (except in areas where plans may want to limit enrollment, such as those with unusually sick patients). But aside from that, plans don't have much leeway to vary premiums in order to attract enrollment. So while the MA program has created a façade of competition, it is very limited in practice.


Implementing a premium-support system in Medicare would be challenging in practice, since it would require some major design and funding decisions that would affect costs to taxpayers and beneficiaries. But the overall approach is theoretically simple.

In each market area, Medicare beneficiaries would face the choice of enrolling in Medicare Advantage or traditional Medicare coverage, as they do today. The cost of this coverage would vary, however. Benchmarks wouldn't be explicitly set in statute, and would instead vary with the overall costs in that market. For instance, one approach would set benchmarks equal to the second-cheapest plan in the market area. This means that beneficiaries would pay some standard premium, and would generally face no increase in costs as long as they chose either the second-cheapest plan or the lowest-cost plan.

Under this approach, both traditional Medicare and MA would be treated equally, in that traditional Medicare would remain a competitor with MA. If traditional Medicare turns out to be the benchmark plan, then enrolling in a more expensive MA plan would cost beneficiaries more money. Plans would be paid based on their bids, which would reflect the cost of providing coverage to an enrollee of average health status (these payments would then be adjusted for the actual health status of the plan's enrollees). But taxpayers' share of this payment would be limited to the benchmark value — anything above that would be reflected in higher enrollee premiums.

This is a very simplified version of what is in reality a highly complicated policy idea. A few key decisions would have a major impact on how such a program would work. For starters, there isn't one single "correct" approach to calculating the benchmark. Setting the benchmark at the second-cheapest plan would follow a logic similar to the ACA's benchmark structure, which ties benchmarks to the value of the second-cheapest silver plan (a plan that covers, on average, 70% of an individual's health-care expenses). If policymakers were concerned about offering more protections to beneficiaries, they might set the benchmark at the median or average plan's value. This would offer more plans to beneficiaries for the benchmark premium, and would be more likely to include traditional Medicare as an equal-cost option. One important tradeoff, however, is that the higher the benchmark is set, the smaller the cost savings to the program.

Another important consideration is how the "standard premium" is calculated. In a 2016 report, the Medicare Payment Advisory Commission offered three potential options for doing this. In one, the beneficiary pays a premium tied to the national cost of providing traditional Medicare coverage (this is how premiums for beneficiaries are currently set). In the second option, the standard premium is set at either some share of the national cost of traditional Medicare coverage or the benchmark MA bid in the market area, whichever is lower. In the final option, the standard premium is determined by either local traditional Medicare costs or the benchmark MA bid, depending on which is lower. In the first option, beneficiaries are guaranteed traditional Medicare coverage, and pay more only if they enroll in a more expensive MA plan. The second option remains agnostic about what choices would require the beneficiary to spend more money — it would depend on whether the traditional program or MA was less expensive. In this case, the standard premium is still consistent across the country. The last approach would vary the base premium, making it lower in markets where either traditional Medicare or MA is less expensive than national traditional Medicare costs, but making it higher in markets where both are more expensive than national traditional Medicare spending. Again, these design decisions trade off beneficiary protections and costs, as well as savings to taxpayers. There is no free lunch.

There are numerous other factors for policymakers to consider. Some of these include whether the system should cover dual-eligible beneficiaries (those with Medicaid and Medicare coverage) or leave them in traditional Medicare, what (if any) cap should be imposed on total growth of benchmarks, and how (if at all) MA benefits should be standardized relative to those in traditional Medicare.

Ultimately, decisions on these fronts would require bipartisan negotiations and numerous analyses to get them right. The difficulty in making these decisions is best highlighted in the different premium-support proposals that have been offered up in recent years. But the fact that premium-support proposals have come from a variety of sources (the idea first appeared in a 1995 Health Affairs article by left-of-center health economists Henry Aaron and Robert Reischauer), and even made their way into President Obama's FY 2017 budget proposal, suggests that bipartisan agreement may be possible.

One incontrovertible truth is that premium support would have significant effects on Medicare's finances. In a 2013 estimate by the nonpartisan Congressional Budget Office, analysts projected that an option with a second-cheapest benchmark would save $275 billion over a six-year period. While the analysts also projected an increase in beneficiary premium payments, depending on the structure of the program, beneficiaries may be able to avoid some of these increases by selecting a benchmark plan.


President Trump has not been enthusiastic about Medicare reform and has suggested he does not want to touch the program. But reform is possible without reducing the value of the benefit seniors now get. And the fiscal condition of the Medicare program necessitates action in the near future. The longer reform is delayed, the more drastic the changes that will be needed to stabilize the program.

A slew of ideas has emerged over the years to help address these challenges. Among those, a model based on premium support offers the potential to save more than $200 billion dollars over a 10-year period, while encouraging more efficiency both in the government-operated traditional Medicare program and among private plans that offer coverage through Medicare Advantage.

Taking advantage of competition among insurers to rein in the cost of Medicare while protecting core benefits for future beneficiaries should be part of every policymaker's reform agenda. Of course, like all changes to Medicare, such a reform is unlikely to come easily. But with appropriate safeguards, this approach would be a win-win for both taxpayers and beneficiaries.

Yevgeniy Feyman is a senior research assistant at the Harvard T.H. Chan School of Public Health, and an adjunct fellow in health policy at the Manhattan Institute for Policy Research.


from the


A weekly newsletter with free essays from past issues of National Affairs and The Public Interest that shed light on the week's pressing issues.


to your National Affairs subscriber account.

Already a subscriber? Activate your account.


Unlimited access to intelligent essays on the nation’s affairs.

Subscribe to National Affairs.