Competitive bidding underlies a growing body of proposals to control costs and increase the efficiency of the Medicare program.  One of the most recent proposals for competitive bidding was released by the Bipartisan Policy Center, a distinguished group that includes two former Senate Majority Leaders (Tom Daschle and Bill Frist), a former Chair of the Senate Budget Committee (Peter Dominici), and a former Director of the Congressional Budget Office (Alice Rivlin).

The interest in competitive bidding is a good sign.  In a review of competitive bidding efforts for Medicare, we found that competitive bidding was relatively straightforward to implement for many different parts of Medicare.  We also found that all of the competitive bidding demonstrations that reached the point of bid evaluation — even those using bidding models that were watered down under provider pressure — demonstrated that they would save substantial amounts of money.

Competitive bidding thus is a proven method for bringing efficient prices to Medicare.  Unfortunately, the BPC’s proposals for competitive bidding are critically flawed.  Most important, the BPC proposal proposes a limited form of competitive bidding, restricted to Medicare Advantage plans only — the traditional Medicare fee-for-service (FFS) plan is not one of the bidders.  For reasons detailed below, that is a serious flaw.  There are other important flaws as well in the BPC proposal.

We propose a bidding arrangement for all Medicare plans, MA plans and the traditional FFS Plan.  In this post, we explain why.  We first review the BPC proposals and then describe the problems that would result from the particular form of competitive bidding BPC has proposed, and why a more comprehensive bidding arrangement would be a far more important reform for Medicare.

The BPC Proposal

The BPC proposes (see especially pp. 42ff and Appendix) to establish a standard minimum benefit for Medicare Advantage (MA) plans that includes all services covered by traditional Medicare (the statutory “entitlement benefit”), plus two enhancements: a cost-sharing limit to protect against catastrophic expenses; and slightly lower cost sharing, at least in transition. (See Note 1.)  It then proposes to pay plans for this benefit using a competitive bidding system that applies only to MA plans, not to the traditional FFS plan.  MA plans would be required to submit two bids: one bid under the current, non-competitive system for the entitlement benefit, and a second bid for the enhanced benefit package under the competitive system.  The new competitive price would take effect only in regions where there are at least two plans, and where that price is lower than the current payment rate, ensuring that the government would realize savings from this reform.

Plans would be paid a benchmark of either the enrollment-weighted average of all plan bids or, if MA enrollment exceeds 40 percent in a particular region, the 35th percentile of bids.  If a particular plan’s bid is below the benchmark, the enrollee would receive a dollar-for-dollar rebate equal to the difference; alternatively, if the plan’s bid is above the benchmark, the enrollee would pay the difference.  In other words, low bids would be translated into easily compared rebates, rather than complex benefit improvements.  For an increased premium, MA plans could continue to offer options with additional benefits, such as dental and vision, and/or reduced cost sharing.

The essential feature of the BPC’s competitive bidding proposal is that only MA plans participate in the bidding process.  FFS Medicare remains untouched by the bidding.

Summary Of Our Critique

MA-only bidding systems like the one proposed in the BPC report are seriously flawed in several ways.  Specifically, under the BPC proposals:

  1. A cartel of plans or one dominant plan can control whether the MA-only bidding system goes into effect or not.  Large parts of the country would avoid competitive bidding.
  2. Even if the BPC bidding arrangement does go into effect, it uses an overcomplicated and flawed bidding design — plans are unlikely to bid their true (average) cost, thus substantially dissipating the benefits of competitive bidding.
  3. Even if plans do bid their true cost, BPC recommends that the government pay plans a weighted average of their bids, thus ensuring that the new bidding system won’t save much money.
  4. Even if the bidding system did save some money, the BPC proposal introduces a cap on the growth in the government contribution, which eventually will turn the whole system into a voucher that is set without regard to the cost of the statutory entitlement benefit — thus reneging on the most fundamental promise of the Medicare program to provide at least some options for beneficiaries to purchase the entitlement benefit from a qualified plan at zero added cost over the Part B premium.
  5. And for all these problems, the BPC proposal ensures extreme political reactions from the MA plans which feel, with some justification, they are competing with themselves on an unlevel playing field that leaves FFS Medicare untouched and relatively more generous.

Fortunately, there is a better alternative: a comprehensive competitive bidding system that includes both MA plans and FFS Medicare would be far easier to explain and manage, would save far more money, and would reduce some of the unnecessary opposition that will surely greet the peculiar form of competitive bidding BPC has recommended.  We address each of these arguments in sequence, below.

1.  A cartel of plans or one dominant plan can control whether the MA-only bidding system goes into effect.

The BPC proposal would go into effect only where MA plans’ bids for an enhanced benefit package are less than the current benchmark payment rates.  While the purpose of this provision — to ensure that the proposal saves money — is commendable, it will leave large parts of the country under the old, flawed payment system.  This could happen in two ways.

First, a dominant MA plan or a group of plans acting in concert could submit a high bid or bids.  The dominant plan’s bid pushes the weighted average bid above the current benchmark, and this “circuit breaker” prevents the new payment system from taking effect under BPC rules.  The dominant plan has an incentive to submit a high bid because its payment under the new system would be less than its old payment.  And without FFS as one of the bidding plans, the dominant plan is less constrained in determining how high to bid.  We think it’s a bad idea to let plans determine under which system they will operate.

Second, the new payment system will fail to take effect in many areas even if plans submit “honest” bids that exceed the benchmark.  It is difficult to predict how often this would happen.  Data from MedPAC (see Table 12-3) indicate that MA plans’ bids on average were 14 percent less than the county benchmarks in 2012.  However, county benchmarks are transitioning to a new system that will substantially reduce the benchmarks when provisions of the Affordable Care Act (ACA) are fully implemented in 2017.  Additionally, while under the APC the benchmarks are being maintained at artificially high levels by quality bonus add-ons that ranged from 3 to 10 percent in 2012, the bonuses will be phased down in 2014, which will further reduce the benchmarks.  With these reductions, MA plan bids will exceed benchmarks in more areas of the county and thus will exempt more areas from competitive bidding.

We simulated the incidence of the BPC proposal under the assumptions that: (1) MA plans will submit the same bids as they do now, (2) the quality bonuses are entirely eliminated, and (3) the new ACA benchmarks are fully implemented.  We project that 29.5 million Medicare beneficiaries live in areas where the BPC proposal would be implemented, but 15.3 million — one-third of the entire Medicare population — do not. (See Note 2.)   A national reform proposal that exempts one-third of the Medicare population seems strikingly incomplete to us.

2.  Even if MA plans fail to prevent the BPC bidding arrangement from taking effect, plans are unlikely to bid their true (average) cost.

A key shortcoming of the BPC proposal is that the BPC bidding model exempts the FFS Medicare plan from the competitive bidding process.  As long as the FFS plan is protected from price competition with MA plans, the government is likely to pay the cost of supplementary benefits in the least efficient MA plan, even if the government contribution to premiums is based on the lowest bid.

Under “normal” circumstances, competitive bidding can discover the cost of the most efficient health plan in each local market.  In the commercial insurance sector, for example, the efficiency of competitive pricing is maximized when employers set their contribution to premiums no higher than the lowest bid.  In that system, plans have an incentive to maximize the out-of-pocket premium difference between themselves and other plans.  Unfortunately, that is not how BPC’s MA-only bidding system in Medicare would work.

In an MA-only bidding system, the FFS plan is protected from price competition with MA plans, in that the beneficiary’s out-of-pocket premium for entitlement benefits provided by the FFS plan cannot exceed the Part B premium — no matter whether qualified MA plans would provide the same benefit at a lower price.  This constraint causes significant problems with the bidding system and drastically reduces the likelihood that the government will experience significant savings following implementation.

In the MA-only bidding arrangement BPC proposes, the beneficiary’s cost of the FFS plan is limited to the Part B premium, even if MA plans underbid the FFS plan.  But MA plans competing with each other must consider the effect of their bids not only on the relative out-of-pocket premiums for other MA plans (their competition with each other), but also for the FFS plan (their competition with FFS).  An efficient MA plan that had the opportunity to bid low, thus distancing itself from a rival MA plan, might choose not to do so because with a higher bid it can improve its position relative to the typically large-enrollment FFS plan.

For example, suppose the entitlement benefits in the FFS plan cost $4,000 a year. The following figures provide an illustration of an MA-only bidding system where the government’s contribution to MA plan premiums is — for graphical simplicity — based on the lowest MA plan bid for the entitlement benefits. (See Note 3.)  The MA plans in this illustration are not equally efficient: the cost of the entitlement benefits is higher in one MA plan than in another.

We also assume that most beneficiaries want to purchase additional supplementary benefits — a reasonable assumption given that 90 percent of FFS beneficiaries purchase supplementary benefits in the form of Medigap coverage (see Note 4) — and we assume supplementary benefits cost both FFS Medicare (i.e., let us assume, Medigap) (see Note 5) and the MA plans $600 to produce.

In Figure 1, both MA plans bid “honestly,” submitting bids for the entitlement benefits equal to their true average costs.  The FFS costs and MA bids are as follows:

  • FFS cost is $4,000 for the entitlement ($4,000), plus a charge to beneficiaries of $600 for the supplements.  Total:  $4,600.
  • The inefficient plan bids its cost for the entitlement – $3,400 – plus a charge to beneficiaries of $600 for the supplements.  Total:  $4,000.
  • The efficient plan bids its cost for the entitlement – $3,000 – and charges $600 for the supplements.  Total:  $3,600.

The bar chart shows that by bidding honestly, the efficient MA plan maintains out-of-pocket premium equality for entitlement plus supplementary benefits between itself and FFS Medicare – since the government contribution covers the cost of entitlement benefits in each case – and a $400 premium differential between itself and the less efficient MA plan.  Beneficiaries have to pay $600 for the supplemental benefits, no matter which plan they choose.  This is the key:  a different bidding strategy allows the efficient plan to reduce its price for supplements without reducing its total revenue.

Figure 1: MA-only Competitive Bidding System:
The Efficient MA Plan Bids Its TRUE Average Cost for Entitlement Benefits


This superior strategy for the efficient MA plan is shown in Figure 2.  In Figure 2, the more efficient MA plan increases its bid for the entitlement benefits up to the level of the less efficient MA plan’s cost for entitlement benefit; and it reduces its own premium for supplementary benefits by an equivalent amount.  Now the FFS costs/MA bids are as follows:

  • FFS cost is unchanged ($4,000) as is the charge to beneficiaries for the supplements ($600).  Total:  $4,600.
  • The inefficient plan’s bid is also unchanged:  it bids its true costs for the entitlement – $3,400 – and charges the same $600 for the supplements.  Total:  $4,000.
  • The critical difference is for the efficient plan.  It doesn’t bid its true cost.  It bids the cost of the inefficient plan – $3,400 – and charges $200 for the supplements, rather than their actual cost of $600.  Total:  $3,600, the same as in Figure 1.

By adopting the bidding strategy in Figure 2, the efficient MA plan has not given up any advantage relative to the less efficient MA plan.  The total out-of-pocket premium differential for entitlement plus supplementary benefits between the two MA plans remains $400.

However, the efficient MA plan has created a new, $400 premium differential between itself and FFS, through the pricing of supplements.  Figure 2 obviously represents a dominant bidding strategy for the efficient MA plan.

But this comes at a direct cost to the Medicare program.  An MA-only bidding system is likely to save the government $600 per MA enrollee if MA plans behave as shown in Figure 2 – not the $1,000 per MA enrollee one would expect given the true cost differential between FFS Medicare and the most efficient MA plan.  This is a troubling result.  What is true of two MA plans is true of more than two MA plans.  As long as some MA plans are more efficient than others and the FFS plan is protected from price competition with MA plans, the government can expect to pay a portion of the cost of supplementary benefits, even if the government contribution to MA premiums is based on the lowest MA bid for the entitlement benefit package.  If supplementary benefits are expensive enough and most beneficiaries purchase them, the entire potential savings from competitive bidding could be eliminated.

Figure 2

One data point provides reason for optimism regarding an MA-only competitive bidding system.  In the late 1990s, the Centers for Medicare and Medicaid Services (CMS) attempted to demonstrate MA-only bidding in four U.S. cities.  Congress mandated those demonstrations, but the demonstrations later were blocked by federal court actions and members of Congress from both parties in whose districts the demonstrations were scheduled to take place.  Nonetheless, there were some interesting results.  The MA bids for the entitlement benefit package in Denver were 25 to 38 percent below what the government would have paid the plans under standard administrative arrangements at the time. (See Note 6.)

While these results are impressive, there is reason to believe they do not represent the full potential savings from a comprehensive competitive bidding system that included both MA plans and FFS Medicare.  We expect the bids would be even lower if plans were bidding against FFS and FFS was not protected from the competition.  Meanwhile, even in an MA-only bidding system, these bids do not necessarily represent how plans might bid over time, after learning better how to avoid competing against each other through repeated bidding cycles.

3.  Even if MA plans were to bid their true cost, setting the government’s premium contribution equal to the weighted average of their bids will ensure that the new bidding system won’t save much money.

We already have noted the flaw of using weighted-average bids to determine when the BPC system takes effect.  Weighted-average bids have another flaw that we discuss next: they won’t save much money.

The BPC proposal is not new.  During his campaign and transition, President Obama proposed an MA-only bidding system with the government contribution based on enrollment-weighted average bids.  Using data from 2005, we estimated that his proposal would have saved almost $6 billion, or less than 2 percent of Medicare spending in that year.  While this estimate obviously would need to be updated, it provides an “apples to apples” comparison with a true competitive bidding system that includes FFS Medicare and uses the lowest bid from any plan to set the government premium contribution.  A true competitive bidding system would have saved $27.1 billion, or approximately 8 percent of Medicare costs in 2005.

Consequently, the “MA-only, weighted-average bid” approach proposed by BPC would realize less than one-quarter of the potential savings from a more complete competitive bidding system.

4.  Even if the bidding system did save a little money, the BPC’s cap on the growth of the government contribution eventually will turn the whole system into a voucher set without regard to the cost of the entitlement benefit package.

The BPC proposal provides that, no earlier than 2020, a fallback spending cap should be imposed that would restrain annual per-beneficiary spending growth to a target of GDP per-capita growth + 0.5 percentage points (over a five-year moving average), applied separately to fee-for-service, Medicare Networks, and Medicare Advantage.  Having limited the effectiveness and reach of its competitive bidding proposal, BPC seems to deal with the likely fiscal consequences of limited savings by imposing a cap on spending.

As we have detailed at length in an earlier book and paper, spending caps are a last resort, as they shift the risk for annual cost increases to beneficiaries.  In the BPC proposal they do so without imposing competitive pressures that would reduce the imperative of the cap in the first place, by generating far more in savings as we discuss above.  In so doing, the BPC retreats from the most basic promise of the Medicare program, to provide at least some options to every beneficiary to purchase the entitlement benefit from a qualified plan at zero added cost over the Part B premium.

5. By leaving fee-for-service Medicare untouched, the BPC proposal ensures vigorous political opposition from MA plans.

MA-only bidding systems evoke intense opposition from health plans.  One reason is simple self-interest — competitive bidding threatens to bring payments in line with plans’ costs, thus ending what many have termed “overpayments” to the MA plans.

But more is at stake in this opposition than self-interest.  First, the FFS plan is “overpaid” in relation to MA plans in a large part of the country – MA plans are cheaper than FFS in almost 40 percent of all counties, in which approximately half of all beneficiaries live.  Leaving FFS out of the bidding is rightly seen as a subsidy to the FFS plan.  As Robert Berenson has noted of earlier efforts to demonstrate an MA-only bidding system:

In the context of [1990s competitive pricing demonstrations, health plans]…argued, with justification, that the competitive pricing design was tilted in favor of traditional (FFS) Medicare because plans bidding only against each other would face pricing discipline from market competition, whereas the traditional program would effectively be given a pass. Under that bidding approach, only beneficiaries selecting the traditional Medicare program would continue to be highly subsidized for their selection.

Secondly, health plans and others worry that the federal government is incapable of administering a level playing field and instead will favor its own FFS plan against MA plans locked in a competitive bidding system.  The failure of BPC to include FFS Medicare almost certainly will be viewed as confirmation of this concern.

We understand that a comprehensive competitive bidding system that includes both FFS Medicare and MA plans would represent a reduction in government-financed benefits for beneficiaries in areas where FFS is more expensive than MA plans.  (As noted earlier, roughly half of all beneficiaries live in counties where FFS would be priced above the benchmark, and thus require beneficiaries to pay at least a modest additional premium.)  However, those free supplementary benefits are not, and never have been, part of Medicare’s entitlement benefit package.  They are doled out to a subset of beneficiaries as a “reward” for living in areas where FFS Medicare is especially inefficient — in some cases, astonishingly inefficient relative to MA plans.

Our consistent position has been that FFS Medicare has an important role to play in the Medicare program.  FFS Medicare guarantees universal access to health insurance for its beneficiaries (although not universal access to health care services).  In areas where providers have amassed significant market pricing power, FFS Medicare’s administrative prices may help offset some of the deleterious effects on consumers.  Nor do we minimize the political difficulties of any approach to address Medicare’s financing problems.  Our proposal is to phase in comprehensive competitive bidding for all beneficiaries over a five- to ten-year period, using appropriate transition rules to accommodate reasonable beneficiary expectations.  But it is far better, in our view, to end up with an efficient bidding system for all plans than to institute a flawed bidding system that will almost certainly require the imposition of an arbitrary spending cap.  Any proposal that contains an arbitrary cap on Medicare spending will create a severe political backlash.

One might argue that an MA-only competitive bidding system is a “foot in the door” for full competitive bidding, but our experience suggests that in the Medicare program, a foot in the door can become a concrete boot — difficult to change and impervious to improvement, as further changes threaten a new round of politically active losers.


For all the reasons outlined above, we contend that the BPC approach to competitive bidding is a bad idea.  A comprehensive competitive bidding system that includes all health plans would:

  • be easier to manage — CMS already is doing most of what it would require;
  • be easier to explain to the public and to beneficiaries than the BPC plan;
  • save far more money, possibly avoiding the need for any cap; and
  • reduce some of the unnecessary opposition that the peculiar form of the BPC recommendations will elicit.

The proposal we recommend obviously faces political opposition of its own, but it is a more productive political challenge, as it points to a more efficient system for all beneficiaries, wherever they live, that saves far more money than the alternatives and at the same time preserves the most fundamental Medicare promise to provide the entitlement benefit for the Part B premium.  Why is the BPC proposing a complex system with flawed bidding rules and intricate embellishments when there is a clean, straightforward system that is far more effective and easier to understand?


Note 1. Benefit enhancements would “ensure that enrollees benefit from some of the savings derived from competitive bidding, while minimizing disruption, as many Medicare Advantage Plans currently have far more generous benefits than traditional Medicare.  Over time, the actuarial value of the standardized package would phase down until it is equivalent to traditional Medicare.”

Note 2.  See Roger Feldman, Robert Coulam, and Bryan Dowd “Competitive Bidding Can Help Solve Medicare’s Fiscal Crisis,” Health Policy Outlook, No. 2, American Enterprise Institute for Public Policy Research, February, 2012, for details of the simulation model.  Due to lack of data on weighted-average bids, we used the median MA bid instead.  Our estimate of the exempted population is probably too low because the quality bonuses will not be eliminated and MA plans will bid on an enhanced benefit package that costs more than the current Medicare benefit.

Note 3.  Other plausible bidding rules would not change the analysis in the text.

Note 4.  BPC proposes to trim permissible Medigap benefits to lower costs for Medicare beneficiaries and encourage more appropriate utilization of care.  Specifically, BPC proposes that, beginning in 2016, all supplemental coverage from Medigap plans and employer-provided plans (including Tricare-for-Life and the Federal Employees Health Benefits Program) should: (1) include a deductible of at least $250; (2) include a maximum on the beneficiary’s out-of-pocket cost no lower than $2,500; and (3) cover no more than half of beneficiary copayments and coinsurance.  Such changes would reduce the dollar amount of supplementary benefits in the example above.  But plans’ incentives not to bid their true cost would remain, albeit smaller in size.

Note 5.  If FFS were somehow permitted to act like a health plan, these supplements could literally be offered by FFS.

Note 6.  At the time of the Denver demonstration, health plans were paid by Medicare at a so-called average per capita cost (AAPCC) rate.  Under the AAPCC, payments were set at 95 percent of the cost of a standardized enrollee in Medicare FFS in the county where the beneficiary lived, with adjustments for a few enrollee characteristics (e.g., age and sex).  The imperfections of the system were obvious, with large overpayments in some areas (allowing plans to offer drug benefits and other substantial enhancements at no added cost) and underpayments in other areas (requiring added premiums to cover little more than the entitlement benefit).  After the Denver demonstration was stopped temporarily by the courts and then more permanently by Congress, Congress dealt with the issue of plan payments by cutting payments across-the-board in the Balanced Budget Act of 1997, so that very low and very high payments under historical methods were compressed toward the national average.  This was yet another cycle in paying private Medicare plans too generously and then, under the BBA, more stringently, but in both cases the rates were derived from FFS Medicare costs, not plans’ true costs to provide the service.