Skip to content
Publicly Available Published by De Gruyter November 20, 2015

Could Improving Choice and Competition in Medicare Advantage be the Future of Medicare?

Alice M. Rivlin and Willem Daniel


About 30 percent of Medicare beneficiaries enroll in private Medicare Advantage (MA) plans but do so at a relatively high-cost. This paper explores the advantages and challenges of introducing competitive bidding among MA plans (Plan One) or among MA plans and Fee-for-Service (Plan Two or Premium Support). We conclude that competitive bidding could reduce the cost of Medicare, especially in densely populated urban areas. However, there would be serious challenges in rural areas and risk adjustment methodology would have to be substantially improved. In Plan Two, sicker beneficiaries might move to Fee-for-Service and beneficiaries might have to pay more to stay with a preferred provider or broader network. If these problems are addressed, we believe that premium support can be a meaningful improvement to the MA program.

1 Introduction

A curious political anomaly has emerged in the last several years of ferment over health care reform. Democrats enacted the Affordable Care Act (ACA), which features consumer choice among competing private health plans combined with government subsidies to make the insurance affordable. Republicans vociferously attack the ACA and vote repeatedly to repeal it. At the same time, they proposed converting Medicare to a premium support model, which would combine consumer choice among competing private health plans with government subsidies to defray the cost of Medicare coverage. Republicans, while rejecting the ACA, believe that giving Medicare beneficiaries a subsidized choice among competing health plans could result in higher quality care, more efficient delivery, and lower costs. Democrats, while supporting the ACA, believe just as strongly that converting Medicare to a premium support model, which they call a voucher program, would lead to deteriorating care, hardship for low-income beneficiaries, and the eventual destruction of “Medicare as we know it.”

In fact, however, changes to “Medicare as we know it” are necessary to ensure that this popular, successful program is able to deliver higher quality care at sustainable cost to the much larger population of older beneficiaries who will be eligible by 2030. Moreover, the outcome will not likely involve a clear choice between a competition model and improving traditional Medicare, but will combine elements of both. Many changes in Medicare are already underway and other papers in this volume explore possible next steps.

This paper will attempt to cut through the partisan ideological rhetoric surrounding premium support proposals and explore the advantages and disadvantages of strengthening competition in Medicare Advantage (MA). MA already offers beneficiaries a choice among private plans and currently enrolls about 30 percent of the Medicare population, so it is a natural starting point for introducing more competition into Medicare.

First, we examine the feasibility and consequences of basing the government payment to MA plans on the results of competitive bidding among qualified plans in local areas. We conclude that such a system could reduce the costs of MA plans, especially in high cost urban areas. At the same time it could reduce the number of plans and enrollment in low-cost areas, especially in sparsely populated rural areas where there are few providers.

Second, we explain the pros and cons of having traditional Fee-For-Service Medicare compete with MA plans. Such competition would put all Medicare beneficiaries in a position to choose a plan (including a Fee-For-Service plan) on an exchange. The government contribution (which would be risk-adjusted to account for differences in beneficiaries’ health) would be determined by a weighted average of the bids in the area, or alternatively by the second lowest bid, as in the ACA. The government contribution would be determined by the cost of delivering that package in the area. This proposal would preserve the Medicare entitlement to a defined package of benefits, but beneficiaries desiring Fee-For-Service might have to pay extra.

2 Brief history of MA

Capitated plans participated in Medicare from the beginning, although it was not easy to fit them into the evolving Medicare payment system, which was primarily geared to Fee-For-Service payments. Starting in 1982, capitated plans were paid 95 percent of risk adjusted Fee-For-Service cost of the Medicare population in the area. Then as now, private plans could compete well in higher cost, primarily urban areas, but not in lower cost, rural areas. Continuing tension between efforts to subsidize payments so that private plans could compete more effectively and desire to limit costs led to repeated changes in payment rules. Enrollment in private plans has varied in response to subsidies that resulted from changes in payment rules.

2.1 Enrollment in Private Plans

The Medicare Modernization Act of 2003 created MA, which based the payment rules for private Medicare plans based on a bidding process and allowed plans to add supplemental benefits. The revamped program featured complex payment formulas designed to encourage private health plans to compete in more markets and resulted in very substantial subsidies for these plans, especially in low cost areas. In part because of the subsidies, average payments per Medicare beneficiary were higher in MA than traditional Medicare. Also MA plans were able to attract healthier, lower cost patients. Attempts to improve risk adjustment and block overt efforts to attract healthy patients were only partially successful.

The ACA, which passed in 2010, reduced the subsidies to MA plans and used the savings to help fund the ACA. These payment rules are still being phased in. Enrollment rates in MA were expected to fall as a result of the lower subsidies, but have held up surprisingly well. Currently about 30 percent of Medicare beneficiaries are enrolled in MA (see Chart 1), although MA enrollment varies enormously by state. At the extremes, 51 percent of beneficiaries are enrolled in MA in Minnesota in 2014, compared with only three percent in Wyoming.

Chart 1: MA Enrollment.Source: Kaiser Family Foundation, MA Spotlight.Note: Data include MA Cost Plans, Special Needs Plans, and other MA Plans.

Chart 1:

MA Enrollment.

Source: Kaiser Family Foundation, MA Spotlight.

Note: Data include MA Cost Plans, Special Needs Plans, and other MA Plans.

As shown in Chart 1, the enrollment trend in MA resembles a rollercoaster. In 1999 in the aftermath of the Balanced Budget Act, about 18 percent of the Medicare population was enrolled in MA. However, enrollment immediately declined. Many private plans withdrew from the MA market, particularly in rural areas. In 1998 there were 346 plans operating in MA; by 2003 there were only 146 plans. But once the Medicare Modernization Act began subsidizing private plans to enter the market both the number of plans and the enrollment in MA rebounded.

3 How MA Works Now

Currently, MA plans are paid a fixed premium in order to provide the Medicare Part A and Part B (and sometimes D)[1] benefit packages, which gives them an incentive to control costs. If MA plans are successful at lowering costs, they are allowed to return a portion of the cost-savings to the beneficiaries in the form of supplemental benefits or reduced cost-sharing.

The Centers for Medicare and Medicaid Services (CMS) determines the geographic reference area for MA plans. The MA plans then submit bids to CMS. The bids submitted are supposed to represent the cost of covering the average beneficiary within the geographic reference area. The bids represent the per-person, per-month revenue needed to provide Medicare Parts A and B benefits, possibly Medicare Part D benefits, and supplemental benefits or cost sharing reductions.

MA plans can offer extra benefits in order to attract enrollees. Since the majority of private plans are HMOs with their own networks and ability to manage care, MA plans design their supplemental benefits to offset the disutility of networks and care management. The supplemental benefits can take the form of reduced cost-sharing for beneficiaries or benefits such as eyeglasses that are not part the Medicare package. Famously, one MA plan offered gym membership in an attempt to attract fitter beneficiaries.

CMS compares each bid submitted by the MA plans to a benchmark in order to calculate the monthly payments made to the plan. The ACA simplified the method used to calculate the MA benchmarks, and reduced, but did not eliminate the previous subsidy to MA plans. All counties are divided into quartiles according to the cost of Fee-For-Service Medicare in that area. The MA benchmark is equal to a fixed percent of the Fee-For-Service in that county. The lowest cost Fee-For-Service counties are benchmarked at 115 percent of the Fee-For-Service costs; the other quartiles are benchmarked at 107.5 percent, 100 percent, or 95 percent of the Fee-For-Service costs, respectively.

Each plan’s bid is compared to the benchmark. If a MA plan’s bid is lower than the county benchmark, the government and the private plans share the savings that result from the difference in payment rates. The MA plan gets to keep 75 percent of the difference between its bid and the benchmark; the Medicare program retains 25 percent of the difference. The plans are required to spend their share of the cost-savings on supplemental benefits or reducing cost-sharing. On the other hand, if the plan’s bid exceeds the country-level benchmark, the beneficiaries in that plan are required to pay a premium equal to the difference.

Medicare’s monthly payments to MA plans are primarily based on the bids submitted by the plans. However, several additional adjustments are made to the bids in order to calculate a final payment rate. Individuals in MA are assigned risk-scores in order to account for differences in expected health expenditures. CMS uses an econometric model to calculate a beneficiary’s expected costs based on age, demographics, and health conditions.[2] The capitation rate for MA beneficiaries is adjusted based on the predicted expenditures.

Over time, research has shown that MA plans “up-code” their beneficiaries.[3] Beneficiaries in MA tend to have higher risk-scores than equivalent individuals in Fee-For-Service Medicare because MA firms have a financial incentive to diagnose as many conditions as possible in their beneficiaries. Additionally, as beneficiaries in MA grow older, their risk-scores increase faster than equivalent beneficiaries in Fee-For-Service Medicare. The risk-scores for MA beneficiaries are deflated by a coding intensity adjustment in order to account for the up-coding problem. The ACA increases the coding intensity adjustment.

The ACA also requires CMS to adjust the payments made to the MA plans according to the plans’ quality. Each MA plan is assigned a quality rating on a scale of one to five stars. The star rating affects both the rebate rate and the plans’ benchmark. A MA plan that received four stars would be given 65 percent of the difference between its bid and the county benchmark; additionally, the benchmark itself would be increased by 5 percent for the purposes of calculating the benchmark for that plan. On the other hand, a plan that received only three stars would receive 50 percent of the difference between its bid and the benchmark; and the benchmark itself would not be adjusted. Since the quality ratings can produce bonuses, but no penalties, they tend to raise the MA payments relative to Fee-For-Service.

4 Which Costs More – MA or Fee-For-Service?

MA was supposed to reduce the cost of delivering Medicare benefits and improve the patient experience by creating competition between plans. However, opponents of MA are quick to point out that average per beneficiary payment to MA plans has been higher than the average per beneficiary Fee-For-Service costs. This is true, but supporters of MA counter with the equally valid fact that the average MA plan bid was below Fee-For-Service costs. How can both be true?

Much of the discussion about MA has been confused by the differences between the payment to MA plans and the bids submitted by MA plans. Bids are supposed to represent the cost of providing the required Medicare benefit package, whereas payments to MA plans add the cost of providing supplemental benefits and risk-adjustments for differences in health. Some analysts of MA treat bids and costs as equivalent – as we do in this paper–while others assume that bids do not perfectly represent a plan’s actual costs. However, whichever measure is used, average MA costs or bids are lower than average Fee-For-Service costs. Thus the arguments by the proponents and opponents are MA are both true depending on whether the cost of supplemental benefits are included.

For example, one recent study of MA and Fee-For-Service Medicare costs by county illustrates these points.[4] The authors calculated the cost of delivering the Medicare benefit package alone (without any supplementary services) and compared the costs of MA plans and traditional Medicare nationally. Nationally, the costs were the same. MA costs by county were 99 percent of traditional Medicare. For rural counties the MA costs were 15 percent higher than Fee-For-Service costs. But MA plans were consistently cheaper than Fee-For-Service in the higher cost urban areas.

There is significant variation in relative costs of MA and Fee-For-Service Medicare, which suggests some scope for reducing MA payments. The costs for MA plans – particularly the HMO plans – do not vary nearly as much by area as Fee-For-Service costs do.

MedPac and Mathematica[5] separately studied the geographic variation in the relationship between MA costs and Fee-For-Service costs. In the lowest Fee-For-Service cost quartiles (by county) the average payment to MA plans is 118.2 percent of the average Fee-For-Service costs. But in the highest Fee-For-Service cost quartiles, the average MA payments are only 95 percent of the Fee-For-Service costs.

5 Competitive Bidding in MA

Competitive bidding in MA has been suggested many times as a solution to the high cost of Medicare.[6] In 1996, a demonstration project for competitive bidding in MA was proposed for Baltimore, but political pressure prevented its implementation. In 1997, another demonstration project was attempted in Denver but legal objections stopped it. Further attempts to launch a demonstration project took place in 1999 and 2010.

Despite these setbacks, health policy scholars across the political spectrum have been attracted to the potential benefits of competitive bidding in Medicare. The American Enterprise Institute has published several policy briefs advocating competitive bidding. The Obama Administration advocated competitive bidding in the 2010 budget proposal. The Center for American Progress issued a somewhat similar proposal. In what follows, we first explore the possible advantages and uncertainties associated with plans for limiting the bidding to MA plans, leaving traditional Medicare as it is. Then we consider including traditional Medicare in the bidding process – a form of premium support.

5.1 Plan One: Bidding within MA

Instead of competing on the richness of their benefits, CMS could require plans to name the price at which they would agree to supply the Medicare package of benefits in an area. The concept would be the same as health plans offering to supply, say, the ACA silver plan in the area at a particular price. Plans would submit sealed bids. Bidders would have to meet quality standards and show that they had sufficient capacity to meet demand. These bids – the prices at which plans agreed to deliver the Medicare benefit package – would determine the benchmark, which would be the federal payment for MA in the area.

The benchmark payment could be calculated in different ways. For example, the benchmark could be set at enrollment-weighted average of all bids; or the benchmark could be set at the second-lowest bid. The benchmark on the ACA exchanges is the second lowest silver plan. Since the cost of the second lowest plan might bounce around from year to year, an enrollment weighted average of all the bids might yield a more stable benchmark.

Under the current bidding system the actual payments to the MA plans may be higher than the plans’ bids – and sometimes significantly higher – because the benchmarks are determined by the Fee-For-Service costs in the relevant area. Under Plan One the cost of traditional Medicare in the area would not be part of the bidding process. The benchmark would be determined by the MA plan bids alone. It would be the competitively determined price of private plans providing Medicare benefits in the area.

Medicare beneficiaries who wanted to enroll in MA would then be able to choose among plans on an electronic exchange, possibly an expansion of the ACA exchange. Plans would be required to provide accurate information on the providers in their network, including health outcomes and other quality information that would be refined as the system improved. A beneficiary who chose a plan that cost more than the benchmark – either because it was less efficient or offered supplementary benefits – would have to pay extra. A beneficiary who chooses a cheaper plan, would receive a cash rebate. The plan would receive the benchmark payment or their bid, whichever was less. The payment would be adjusted for the health risk of the beneficiary, and possibly for up-coding and quality, as at present. Traditional Medicare would continue to be an option.

Such a bidding system would likely bring down MA payments and improve quality in higher cost urban areas where a substantial fraction of the Medicare population lives. However, without subsidies many MA plans would not maintain a presence in low-cost areas, especially sparsely populated rural ones. One policy option would be simply to use competitive bidding to improve competition and lower costs in places where it is most likely to be able to do so but not to force MA into areas where it is unlikely to thrive.

A variety of studies have examined the implications of these alternative benchmarks by looking at bids submitted by the MA plans and comparing the bids to the current cost of MA. For example, Zuiri Song, David Cutler, and Michael Chernew[7] estimated the cost of MA under a competitive bidding system by looking at the bids submitted between 2006 and 2009. Their analysis indicated that a competitive bidding system could lower the MA payment rates to about 90 percent of traditional Fee-For-Service costs.

According to the historical pattern of bids and payments analyzed by Song and his coauthors if the payment rates were set equal to the second-lowest bid then the cost of MA would be about 91 percent of Fee-Service costs. Alternatively, the proposal in the President’s budget would set the MA payment rate equal to the average of all Medicare bids so the cost of the MA program would be considerably lower if the payment rates were also lowered substantially.

These calculations suggest that competitive bidding in MA could produce substantial cost-savings under either method of defining the benchmark. However, these studies assume that MA plans will submit bids according to their historical tendencies. Unfortunately, it is hard to know what the MA bids would be under a new system not anchored to the cost of Fee-For-Service Medicare. One reason for the uncertainty is that MA markets are highly concentrated and MA plans often have considerable market power. Another is that withdrawal of the subsidy for MA plans in low cost areas would likely trigger substantial withdrawal of health plans from the bidding process.

5.1.1 Concentration and Market Power

Brian Biles, Jonah Pozen, and Staurt Guterman,[8] examined the level of concentration in MA markets by calculating a Herfindahl Index (HHI) for each county, The HHI represents the extent to which MA enrollment is concentrated among a small number of firms. The majority of the MA markets – 71.5 percent of them – have high levels of concentration. These counties also contained nearly 75 percent of the total MA population. Only 2.1 percent of the MA markets have low level of concentration and only 1.6 percent of MA beneficiaries live in these counties. However, it is not clear that such a stringent definition of concentration is relevant to the question of how vigorously insurance companies will compete for market share. There may be vigorous competition in markets with two or three dominant firms, especially if profits are substantial and other firms stand ready to contest the market.

Several studies have shown that when the MA benchmarks increase the bids submitted also increase – an indication that the plans have market power. Ziuri Song, Mary Landrum, and Michael Chernow[9] show that when the payments to MA HMO increased by $1 the HMOs increased their bids by $0.49 and only $0.34 was passed along to beneficiaries in the form of rebates. The correlation between the benchmarks and other types of MA plans was smaller but still significant. Other studies have shown smaller but still significant relationships.

Taking a similar approach, Mark Duggan, Amanda Starc, and Boris Vabson[10] show that when the payment rates to MA plans increase MA plans keep almost half of the increase in payments and pass about one-third of the increase onto the beneficiaries. Numerically, this finding aligns closely with the finding by Song and his coauthors. The authors of this study also show that when the government’s payment to MA plans increase, a significant amount of that payment is devoted to advertising. Naturally therefore, the government’s payments tend to correlate with an increase in the number of beneficiaries in MA – not surprising in a concentrated market.

High concentration in MA markets makes it difficult to have confidence in estimates about the outcomes of competitive bidding. For example, it is possible plans will engage in predatory pricing in order drive other firms out of the market before increasing their bids. If this happens, invoking the anti-trust laws would be appropriate, although such actions tend to be lengthy and cumbersome.

5.1.2 Exit from Low Cost Areas

Another concern with a competitive bidding system is that MA firms are likely to exit the market as subsidies are withdrawn. Before the subsidies were increased in 2003, a large number of private HMO plans left the Medicare market because they found it difficult to operate in rural areas. Providing access to at least one MA plan in market has been a political objective for some time.

The policy of “over-paying” MA plans to enter in low-cost areas has been successful. Almost 99 percent of MA beneficiaries have a choice of two or more MA plans. Unsurprisingly, studies of MA entry and exit decisions have concluded that high payment rates heavily induce private plans to enter the market. Conversely, if the MA rates decline, plans are expected to exit.

Austin Frakt, Steven Pizer, and Roger Feldmen[11] studied twelve large insurance firms as these firms decided whether or not to enter different MA markets. They conducted a simulation study to determine how entry and exit decisions will change under the new ACA benchmarks. They estimate that under the new benchmarks, the probability that a private insurer enters the market declines by 40 percent. Another simulation by Shiko Maruyama[12] estimated the change in the number of MA HMOs rather than the individual probability that a firm enters the market. Maruyama simulated the effect of equating the MA benchmarks with the average Fee-For-Service costs; this is different from the scenario considered by Frakt and his coauthors but the simulation suggests that HMOs will exit from markets where the initial number of HMO plans is small. But in markets with more than 5 HMOs, plans actually enter the market.

We expect that the ultimate effect would be that uncompetitive MA markets become more uncompetitive and competitive markets become more competitive. If a competitive bidding system causes the payment rate for MA plans to decline significantly then there is likely to be substantial exit by MA plans. In some markets, firms may exit the market leaving no viable MA plan for beneficiaries; in other markets, firms may exit the market leaving only one or two MA firms with considerable market power. In short, it is hard to predict the outcome.

5.2 Plan Two: Both Fee-For-Service Medicare and MA Plans in Competitive Bidding

A more drastic proposal would bring traditional Fee-For-Service Medicare into the same bidding system with private MA plans and give Medicare beneficiaries the full range of choices on an exchange. In such a system, the government would submit a bid on behalf of Fee-For-Service Medicare. The bid would equal the average Fee-For-Service cost in each geographic reference area. Traditional Medicare would then function just like another MA plan in the competitive bidding mechanism. The competitively determined benchmark would define the government’s contribution.

The benchmark could be set as the second lowest bid or the enrollment-weighted average of the bids as in Plan One. In some places the Fee-For-Service plan would cost more than the benchmark and beneficiaries who chose it would have to pay a premium in order to obtain Fee-For-Service Medicare. In other areas, Fee-For-Service might be below the bench mark and those who chose it would get a rebate. If no private MA plans entered the market, Fee-For-Service would be the only option.

Such a system would preserve the entitlement of Medicare beneficiaries to the Medicare package of benefits. However, unlike Plan One, Plan Two would not guarantee that the Medicare benefits would be available in a Fee-For-Service setting without extra charges. In areas where Fee-For-Service was more costly than MA, a beneficiary might want to stay with a particular provider and that provider might not be in a plan available at or below the benchmark price. Such a beneficiary would have to pay extra for that option.

Plan Two is form of Medicare premium support – a program in which the government makes a specified contribution to Medicare and allows the beneficiary to choose among competing plans, including Fee-For-Service Medicare. In this case, the government contribution would be the local cost, determined by competitive bidding, of delivering the Medicare benefit package.

In areas where Fee-For-Service Medicare was more costly than efficient MA plans, Plan Two would clearly generate more savings than Plan One. The Congressional Budget Office (CBO) analyzed one option that allowed traditional Fee-For-Service Medicare to complete with MA plans. The analysis found that that the total cost of Medicare would be reduced by about 11 percent relative to baseline spending.

The CBO also estimated the savings from introducing competitive bidding only among MA plans (a proposal similar to Plan One). They expect that MA plans will submit bids roughly equivalent to the bids that they have submitted in the past, similar to the assumption made by Song, Cutler, and Chernew. Under this assumption, the CBO estimated that competitive bidding would reduce the cost of MA by $158 billion when the benchmark was set equal to the average off all bids submitted.[13] President Obama’s 2009 budget estimated total savings of $176 billion. OMB’s estimate assumed that MA plans would reduce their bids in the face of greater competitive pressures so the estimated savings were greater.

5.2.1 The Controversy over Premium Support

The term “premium support” goes back to a proposal by Henry Aaron and Robert Reischauer in 1995.[14] The Aaron-Reischauer plan called for Medicare to pay a defined sum for a defined benefit. Individuals would then have the option to enroll in private health insurance plans. Their plan would have equated the growth rate in the premium contribution with the growth rate of health care for the non-elderly. The basic idea was featured in the National Bipartisan Commission on the Future of Medicare (Breaux-Thomas Commission)[15] in 1999, but failed to gain the support of the Democrats on the Commission.

Premium support resurfaced in the political debate in 2008 when Congressman Paul Ryan put forward the first of a series of controversial proposals to convert Medicare to a premium support plan. Under the Ryan plan, the government would make a defined contribution (initially equal to the average cost of Medicare) and allow Medicare beneficiaries to choose among private plans. Various versions of the Ryan proposal were incorporated into House-passed Budget Resolutions after Ryan became Chairman of the House Budget Committee. Early versions of the Ryan plan would have phased out traditional Medicare by putting all new beneficiaries in the premium support plan, while existing ones could stay in traditional Medicare if they wanted to. Substantial savings were achieved by allowing the government contribution to grow only with the increase in the consumer price index (CPI). Since health care spending was then projected to rise much faster than prices, indeed considerably faster than the growth in GDP, this strict growth cap implied substantial cuts in future Medicare benefits.[16]

The Ryan plan provoked an explosion of opposition among Democrats, including an effective political ad in which a tall, lanky Ryan-like figure was shown pushing a sweet-faced old lady off a cliff in her wheel chair. Democrats characterized Republican plan as a “voucher,” an end to “Medicare as we know it,” and a medical/financial disaster for the elderly and disabled.

While this political battle raged, more moderate versions of premium support emerged. The Bipartisan Policy Center’s Debt Reduction Task Force, co-chaired by Pete Domenici and Alice Rivlin, proposed a premium support plan involving competitive bidding among private plans on a Medicare exchange, but preserved traditional Medicare as an option. The proposal was similar to Plan One except that it included a cap on the growth in of the government contribution to Medicare spending, albeit a more generous cap than Ryan’s. The cap was included because of uncertainty about the success of competitive bidding in holding down costs and the need for savings that the CBO would score. Another similar premium support proposal was a compromise between Republican Congressman Paul Ryan and Democratic Senator Ron Wyden that also capped the increase in the per capita government contribution at the growth of per capita GDP+1 percent. This compromise produced smaller savings than Congressman Ryan’s premium support proposal in his 2011 “Path to Prosperity” Budget that linked the government’s contribution to a price index (the CPI-U). Much of the public discussion of these plans focused on their spending caps rather than the details of the competitive bidding.

For a time, the Democrats’ total rejection of the original Ryan proposal (which phased out traditional Medicare and reduced later year benefits) appeared to rule out any bipartisan consideration of premium support–even a version that kept traditional Medicare as an option and preserved beneficiaries’ entitlement to the Medicare benefit package at no extra cost. The whole concept seemed politically toxic. But the success of the ACA exchanges – as well as the growth of private exchanges and the older population’s experience with choosing among private plans in Part D of Medicare–has familiarized Americans with the idea choosing among health plans and with electronic exchanges. Payment reforms in traditional Medicare may also make more people familiar with provider-led groups, such as Accountable Care Organizations (ACOs) that are evolving in the direction of capitated plans. In this atmosphere, introducing competitive bidding into MA could be seen as an effort to improve the efficiency of an existing – and well-accepted part of Medicare, rather than a scary new departure that might cut back expected benefits.

Unlike the Domenici-Rivlin proposal, our Plans One and Two do not involve a cap on growth in the government contribution. They are more in the spirit of the Aaron-Reischauer proposal, in that changes in the government contribution would reflect actual costs. While a global cap provides an additional policy level for the government to control the cost of Medicare, we have concluded that it is either unnecessary or unsustainable, depending on what happens to Medicare spending over time. If increased competition in MA combined with other reforms in traditional Medicare were successful in keeping Medicare cost growth to a moderate rate (say, no faster than the growth of GDP), the cap would be unnecessary. If on the other hand, Medicare spending began growing rapidly again, the cap would require cuts in the Medicare benefit package. The cap would quickly become politically unsustainable and legislative action would be required. Hence, adding the cap seems undesirable. Besides, CBO estimated that there would savings from competitive bidding even without caps on total spending. The Risk-Adjustment Challenge of Integrating Fee-For-Service into a Competitive Bidding System

Besides the political challenges of reforming Medicare, technical challenges need to be resolved before MA and Fee-For-Service can compete on a level playing field. Currently, MA beneficiaries are healthier than traditional Fee-For-Service beneficiaries. Sicker beneficiaries tend to enroll in Fee-For-Service, rather than MA. As a result, the Fee-For-Service and MA plans cover different populations.

An analysis by MedPac showed that the average risk score of those enrolling in MA was an average of only 84–87 percent of those individuals enrolling in Fee-For-Service Medicare, indicating that the MA population is substantially healthier than the Fee-For-Service population. Additionally, data from the Centers for Medicare and Medicaid and MedPac show that about 10 percent of the MA population switches plans each year. Each year only about two percent of the MA population switches into Medicare Fee-For-Service but these individuals tend to be disproportionally less healthy than the remaining MA population.[17]

For example, Lauren Nicholas[18] showed that individuals with serious comorbidities were likely to choose Medicare Fee-For-Service coverage. In fact, minorities, the previously disabled, and dually eligible beneficiaries are especially likely to leave MA. On the other hand, individuals with easily treatable hypertension and diabetes will remain in MA. Gerald Riley[19] estimated that beneficiaries who dis-enroll from MA cost $1201 per month compared to $798 per month for “similar” individuals who stayed in MA.

It is not terribly surprising that MA enrolls healthier beneficiaries and that the difference widens as individuals switch out of MA and into Fee-For-Service Medicare. After all, sicker beneficiaries want access to broader networks and higher cost facilities, such a teaching hospitals. In principle, risk adjusting the payments is the answer to this problem. In practice, however, current techniques for risk adjustment might not be adequate to prevent a death spiral in which Fee-For-Service attracted the sickest patients, its costs escalated out of control, and low-income beneficiaries were unable to pay for care. Recent improvement in health information technology is making more accurate risk adjustment increasingly feasible. Recent Reductions in MA Benefits

Recent information shows that MA plans have become less generous, presumably as a result of declining subsidies. MA plans[20] have slowly increased the amount of out-of-pocket costs their beneficiaries paid. A report by the Kaiser Family Foundation[21] showed that in 2011, 51 percent of health plans capped beneficiaries’ out-of-pocket spending at less than $3400. However, by 2015 only 9 percent of MA plans limited plans’ out-of-pocket spending by that much. The result has been a substantial increase in the amount of out-of-pocket costs paid by MA beneficiaries over the same time period.

Similarly, in recent years, MA plans have slowly increased premiums charged to beneficiaries. As a consequence, MA beneficiaries are losing access to plans that do not charge a premium to beneficiaries. Prior to 2011, the vast majority of beneficiaries, about 90 percent, had access to a MA plan that covered prescription drugs. However, by 2015 a smaller majority of beneficiaries, only 78 percent of plans, had access to that covered drugs without an additional premium.

MA plans have also reduced the breadth of their provider networks in order to reduce their costs. Anecdotal reports suggest that plans have significantly reduced the breadth of their networks as a cost-saving measure in response to the payment reductions included in the ACA. Unfortunately, the research on MA provider networks is sparse and there are no systematic studies on MA provider networks.[22]

The timing of these changes in the MA system coincides with the beginnings of the ACA’s effect on MA payment rates. The competitive bidding systems that we have described above are designed to reduce the costs of MA. Therefore, it is possible that a competitive bidding system would accelerate the trends towards higher out-of-pocket costs, higher premiums, and narrower provider networks.

6 Conclusions

The question posed by the title of this paper is, “Could Improving Choice and Competition in MA be the Future of Medicare?” Our conclusion is a qualified, “yes.” Choice and competition can play a major role in improving Medicare, starting with MA, but mechanisms for implementing choice and competition have to be very carefully designed to realize their promise and avoid adverse outcomes.

An attractive scenario for the future of Medicare could involve two parallel tracks that eventually converge. One track would start with MA and build on its strengths. It would use competitive bidding, improved quality measures and consumer choice on user-friendly exchanges to reduce costs and improve health outcomes in areas where MA plans can compete effectively without subsidies. The second track would use payment reforms to lower costs and improve health outcomes in traditional Medicare. It would involve giving beneficiaries incentives to enroll in ACOs that accepted risk and received capitated payments. If both tracks were successful in improving health outcomes and the efficiency of delivery, they might eventually converge. Medicare beneficiaries – perhaps consumers of health care at all ages – would make well-informed choices among integrated health care organizations. Some of these would be run by insurers and others by provider groups that accepted insurance-like risk, but all would endeavor to keep their enrollees as healthy as possible in exchange for a periodic per capita payment.

Introducing competitive bidding into MA would be a good first step toward this future. The current pricing system often results in the MA plans being paid more than it costs them to deliver the Medicare benefit package and meet quality standards. A system in which payments to MA plans are set by competitive bidding among plans would save taxpayers money (Plan One). With reliable measures of care quality, health outcomes and patient satisfaction available on a user-friendly exchange, informed MA consumers, as well as tax payers could enjoy higher value than they do under the current system, while traditional Medicare remained as an option for those who preferred it.

Even this first step, however, would require resolving difficult policy issues. First, experience has shown that MA plans compete very well in higher cost, mostly urban areas, but not in low-cost, sparsely populated areas. One option would be simply to accept this fact, expect the benefits of competition to be largely confined to higher cost, more populous areas and rely on traditional Medicare to serve the others. Alternatively, taxpayers could continue to subsidize MA plans in the low cost areas in order give those residents a wider choice of plans. We know such subsidies are effective, but are they worth paying?

Another issue would be whether the predicted benefits of competition would actually be realized, given health insurers in many markets considerable market power. This problem goes beyond MA – it raises questions about all health insurance markets, including the employer marked and the ACA Marketplaces. Not taking market power in health insurance markets seriously and failing to design a strategy to mitigate it could prove detrimental to both taxpayers and customers.

The more ambitious step of bringing traditional Medicare into structured competition with MA plans could yield additional cost savings (Plan Two). It would allow beneficiaries to choose among health plans on an exchange and could bring costs down substantially in areas where Fee-For-Service Medicare is most expensive. Two other issues, in addition to the two just mentioned, would have to be resolved before these benefits could be realized. One is the difficult problem of risk adjustment. MA plans have historically attracted healthier beneficiaries and those who switch from MA to traditional Medicare tend to be in worse health than those who stay. Risk adjustment techniques are improving and may advance faster with as more timely and accurate individual health information becomes available. Without rapid improvement in risk adjustment, however, competitive bidding that included both MA and traditional Medicare could result in traditional Medicare attracting the sickest and most expensive patients and becoming inaccessible to those with modest incomes.

The final policy issue goes to the basic question of what the Medicare entitlement actually means. Both Plans One and Two would guarantee seniors the package of Medicare benefits without additional premiums. But Plan Two would not guarantee that the Medicare benefit package would be available in a Fee-For-Service setting without paying extra. The benchmark plan in the area might well be an MA plan. Moreover, experience shows that competition to reduce cost and attract beneficiaries often results in narrower networks. To stick with a particular provider or have access to a broader network would likely cost the beneficiary more. However, if payment reforms in traditional Medicare also lead to narrower networks and capitated payments, beneficiaries may come to regard them as the new normal.

None of these issue policy issue is easily solved, but, if they can be resolved satisfactorily, well-designed competitive bidding could be part of a viable, well-funded Medicare that delivers quality care to the seniors of 2030. Achieving such a result is worth a lot of effort, since maintaining the status quo – “Medicare as we know it” – is not a viable option.

Corresponding author: Alice M. Rivlin, Brookings Institution, Economic Studies, 1775 Massachusetts Ave, Washington DC, USA, e-mail:

Published Online: 2015-11-20
Published in Print: 2015-12-1

©2015 by De Gruyter