Download PDF A major reason Medicare faces a severe fiscal crisis is because it pays too much for basic benefits. But chronic overpayment can be cured by harnessing market forces in the form of competitive bidding. It uses bids from private Medicare Advantage (MA) plans and the traditional fee-for-service (FFS) Medicare program to set the payment rate for all Medicare health plans. Our research shows that competitive bidding—a key feature of the Wyden-Ryan plan—could save Medicare $339 billion over ten years while maintaining basic benefits and without raising taxes. Crucially, the elderly would not be exposed to the risk of higher health care costs, as in approaches that would set fixed voucher payments toward the purchase of medical insurance.
Key points in this Outlook:
- Medicare’s fiscal crisis is at bottom an overpayment and overtreatment crisis that is the natural outcome of the program’s incentives.
- Competitive bidding is a reform proposal that uses market pressures to lower prices and raise value.
- A key feature of the bipartisan Wyden-Ryan reform, competitive bidding could save $339 billion over ten years while keeping premiums for most seniors stable.
At present, Medicare is dominated by the traditional fee-for-service (FFS) plan, under which patients can see any doctor of their choice for any covered service, with only minor limits. Medicare pays a fee for each service. Not surprisingly, doctors and other health providers have an incentive to provide more services. Medicare also offers private Medicare Advantage (MA) plans that are paid a fixed amount each month to care for enrolled seniors. MA plans have an incentive to work with doctors and hospitals to reduce the use of ineffective and costly services. In some parts of the country, these plans are cheaper than FFS Medicare and the quality of care is as good. In other parts of the country, FFS is cheaper.
For decades, Medicare has based its contributions to private health plans’ premiums on the cost of FFS Medicare in the same market area. Medicare tells private plans what they will be paid, regardless of the plans’ true costs of caring for Medicare beneficiaries—something the plans know and Medicare does not. Competitive pricing reverses this misdirected flow of information.
Here is how competitive pricing would work: Both private plans and FFS Medicare would submit bids in each market area for a standard set of covered benefits (the entitlement). The FFS bid is straightforward for the government to calculate—simply the prior year’s cost in each county for a standardized FFS beneficiary, trended forward. The government then sets its payment equal to the lowest or second-lowest bid in each area, adjusted for the health risk of each plan’s enrollees. The penalty for bidding high is predictable: beneficiaries who enroll in a plan with a high bid have to pay out-of-pocket premiums, providing an incentive for each plan to offer its best price. The reward for low bidders is the prospect of increased enrollment. For competitive pricing to work, it must encompass all Medicare plans, including FFS.
Under competitive bidding, some beneficiaries will pay higher monthly premiums if their current health plan bids high, but they can avoid that higher payment by changing plans. They pay an additional amount only if they choose to remain where they are or enroll in another high-bidding plan.
Some health plans will have substantially different monthly payments from Medicare, and that will require them to change how they conduct their business. Changes of this sort are an appropriate concern in any implementation of competitive bidding. However, the most affected plans can be identified in advance, and many ways exist to create a smooth transition to full competitive pricing. Furthermore, all seniors would have access to least one health plan offering the standard set of benefits at no more than the Part B premium that seniors currently pay.
"The additional budget savings from competitive bidding could amount to $339 billion through 2020."Versions of this proposal have been offered by Sen. Ron Wyden (D-OR) and Rep. Paul Ryan (R-WI), as well as an independent commission headed by former Sen. Pete Domenici (R-NM) and former director of the Congressional Budget Office Alice Rivlin. It is clear that this is a timely topic and may be the central architecture for structural reform of Medicare.
It is important for policymakers to have accurate estimates of the savings from competitive bidding, as well as the potential changes to beneficiaries and plans that they may have to mitigate to make those savings politically palatable.
Two years ago, we estimated that a fully implemented competitive bidding system would have saved approximately 8 percent of Medicare costs in 2005. We subsequently extrapolated this estimate to a savings of $550 billion over ten years. However, two significant changes have occurred in the Medicare program that alter the savings through competitive bidding. First, MA enrollment has grown from about 12 percent of all Medicare beneficiaries in 2005 to 25 percent in September 2011. Second, Medicare changed its method of paying MA plans from one based on (but always higher than) FFS costs to a new system that will pay more than FFS in low-cost areas and less than FFS in high-cost areas.
In this Outlook, we update our estimates of the savings from competitive bidding to account for the recent changes in Medicare. Using data from 2009, we estimate that fully implemented competitive bidding would save 9.5 percent of Medicare spending, more than our previous estimate. However, the Patient Protection and Affordable Care Act of 2010 (PPACA) is estimated to save 4.2 percent if implemented as the law requires, so competitive bidding would save 5.6 percent more than the PPACA legislation. The additional budget savings from competitive bidding could amount to $339 billion through 2020.
We also estimate the size and geographic distribution of the changes in payments that beneficiaries would face from competitive bidding if they do not choose to change plans. The overwhelming majority—approximately two-thirds—of beneficiaries face a relatively small change: 43 percent face no change, and 22 percent face a change of less than $40 per month. But some beneficiaries would face changes in payments as large as $352 per month. Measures targeted at beneficiaries who face more substantial changes may be necessary to keep short-term changes in payments within reasonable levels, though it is worth emphasizing again that beneficiaries are not locked into the higher amount; they can avoid changes in payments by switching plans. This message is reinforced by the geographic distribution of the changes, which are relatively concentrated, reinforcing the political need to introduce the changes gradually through a transition period.
Data and Methods
For our calculations, we used public data from the Centers for Medicare and Medicaid Services (CMS) on enrollment by county in the traditional FFS program and MA plans in 2009. These sources also tell us how much FFS spent per enrollee and the county-level benchmarks for MA plans. The appendix describes these files and how we merged them in more detail.
We started by estimating Medicare spending under the PPACA, which will use a new formula to determine the county-level benchmarks for MA plans: the benchmark in counties in the highest quartile of FFS costs for the previous year will be 95 percent of FFS costs; this increases to 100 percent of FFS costs in counties in the second highest quartile, 107.5 percent in counties in the third highest quartile, and 115 percent in counties in the lowest quartile. The benchmark cannot be more than the county would have been paid under the old payment system. The new benchmarks will be phased in over four years if the difference between the current and new benchmarks is between $30 and $50 per month and over six years if the difference is more than $50 per month. We set the benchmarks at their fully implemented values.
MA plans bidding lower than the benchmarks can offer additional benefits not included in the entitlement, or they can give premium rebates. Because MA plans can offer rebates, we assume that any extra benefits the MA plan includes in its bid are worth to beneficiaries at least what they cost the MA plan to provide and thus that beneficiaries would be willing to pay for those additional benefits out of their own pockets. MA plans bidding above the benchmarks must charge out-of-pocket premiums.
The PPACA made several additional changes to MA plan payments with effects we did not attempt to estimate:
- Medical loss ratio: The PPACA implements a medical loss ratio requirement for MA plans of at least 85 percent, beginning in 2014.
- Tax on low bids: The tax on bids less than the benchmark will increase from 25 percent to as much as 50 percent, depending on plan quality. We assume that all plans will earn high enough quality ratings to qualify for the lowest tax rate.
- Bonus payments: Under the PPACA, MA plans that achieve certain levels of quality will earn bonus payments that will add to PPACA costs.
A key data element in the estimation is MA plans’ bids. Plans’ bids are not available to the public—a flaw in the current payment system—so we used a summary of the bids in 2009 from a presentation by the Medicare Payment Advisory Commission. This document presented the twenty-fifth, fiftieth, and seventy-fifth percentiles of MA plans’ bids by intervals of FFS spending. For example, in areas where FFS spending is between $625 and $675 per beneficiary per month, the median MA plan bid is 1.1 times (110 percent) FFS spending. We assumed that this ratio applies to all counties in that spending range, and then we looked up all counties in the merged CMS data with FFS spending between $625 and $675 and attached a median MA plan bid of 1.1 times FFS spending to those counties.
Next, we calculated the savings from competitive bidding and compared them with the PPACA baseline. Under competitive bidding, payments to all Medicare health plans will be determined by plans’ bids. We assumed that MA plans will submit bids equal to those they submitted in 2009. In other words, competitive bidding will not result in changes in plans’ operations that increase efficiency and lead to lower bids. FFS will also submit a bid—the county-level cost in the traditional FFS program. Our favored payment system is based on the lowest bid in each county, but because the lowest MA bids were not available, we used the twenty-fifth percentile of the bids instead. FFS will be the winning bidder if the FFS bid is lower than the twenty-fifth percentile of the MA bids.
"Under competitive bidding, beneficiaries who are willing to change plans will face no increase in monthly premiums."Because we used the twenty-fifth percentile of MA bids or FFS costs to set the payment rate, the savings estimates in this report are less than those of the lowest-bid system we prefer. However, they may be similar to the Domenici-Rivlin Protect Medicare Act, which proposed that the government payment to all Medicare plans be determined by the second-lowest bid. We also note that FFS Medicare will be the lowest bidder in many counties, which would render moot the lack of data on the lowest MA plan bid in those counties.
In areas where MA bids are higher than FFS spending, MA enrollees will have to pay the difference in premiums out of their own pockets if they remain in their MA plans. Our estimates of the financial impact of competitive bidding on MA enrollees thus include not only the out-of-pocket premium the enrollees would have to pay if their MA plan costs more than the government payment, but also the additional amount they would have to pay out-of-pocket to maintain their current benefits beyond the entitlement level of coverage.
Savings from Competitive Bidding
Medicare spending in 2009, before the PPACA, was $387 billion per year. The PPACA will reduce that to $371 billion, saving 4.2 percent. Competitive bidding would reduce Medicare spending to $350 billion per year, saving 9.5 percent compared with the baseline and 5.6 percent compared with the PPACA. Another way of saying this is that the PPACA made a “down payment” on Medicare savings by reducing the MA benchmarks. Competitive bidding would continue this process by using plans’ bids to set the payment rate for all Medicare health plans.
If trends in Medicare spending continue as projected by the 2011 Trustees Report, the ten-year savings from competitive bidding would be $339 billion compared with the PPACA. We caution that all long-term projections are based on numerous assumptions; nevertheless, the cumulative savings from competitive bidding would be substantial.
Next, we examined the distribution of the savings compared with the PPACA baseline by intervals from zero (no savings) to more than $125 per beneficiary per month (PBPM). The results are shown in figure 1. Twelve percent of Medicare beneficiaries live in counties where the savings are zero; in other words, no seniors in those counties will have to pay more for their current health plans. Next, 57 percent—the largest share—live in counties with savings of $1–40 PBPM. This is followed by smaller proportions of beneficiaries living in counties with larger savings. Only 2 percent of beneficiaries live in counties with 14 percent of the total savings.
Changes in Beneficiary Payments
Under competitive bidding, beneficiaries who are willing to change plans will face no increase in monthly premiums. The overwhelming majority of beneficiaries who do not change plans will face a modest change in payments ($0–40 per month) compared with the PPACA baseline, but some beneficiaries will face changes in payments of as much as $352 per month. The distribution of these changes is more geographically extensive for MA enrollees than for FFS enrollees.
Figure 2 below shows the size distribution of changes in payments for FFS and MA enrollees. The diameters of the bubbles are proportional to the number of enrollees at each level of additional payment.
Forty-three percent of all enrollees (39 percent in FFS, 4 percent in MA plans) face no change at all, and 22 percent face higher payments of $1–40 per month. Thus, 65 percent of all enrollees face payment changes of $40 per month or less. Eighteen percent face additional payments in the $40–70 range, and 17 percent face additional payments of more than $70 per month: 16 percent in the $70–125 range, and the remaining 1 percent ranging unevenly up to $352.
It is easiest to present the information on geographic changes in payments in three parts: FFS, MA, and enrollment-weighted average. These parts correspond to figures 3–5, which compare geographic changes in payments from competitive bidding with the PPACA baseline. Note that in this part of the analysis, our unit of analysis is counties, not enrollment; when we say the change in payments is $X, that is the change for a specific county.
Figure 3 shows the geographic distribution of changes in payments for current FFS enrollees. These changes indicate that some FFS enrollees will have to pay higher monthly premiums to stay in FFS—changes they could avoid by switching to a low-cost MA plan.
The most important point of the map is that few counties, relatively speaking, present a problem: the lightest color (<$40 change in payment) covers most of the map. But the many darker areas identify greater changes. Five counties have more than $300 per month in additional charges: four of these are in Texas, with very small (<1,000) enrollment. However, Dade County (Miami, Florida), with more than 175,000 FFS enrollees, has $352 per month in additional payments. Other selected counties with more than $100 per month in additional payments are Kings (Brooklyn, New York, $102), Harris (Houston, Texas, $101), and Cook (Chicago, Illinois, $101).
Figure 4 shows the geographic distribution of changes in payments for MA enrollees. These changes indicate that some MA plan enrollees will pay more for basic benefits unless they switch to FFS; others will have to pay more for extra benefits that were free.
The story here is different than for FFS: MA plan enrollees will pay some additional amount in more than 80 percent of counties. The changes are relatively large (indicated by dark colors on the map). Compared to the FFS map, many new areas appear among the counties with the largest change in MA payments, including bands of counties in northern and mountain states.
Dade County has the ignominious distinction of having the largest change in MA payment ($258), as well as the largest FFS change. The lowest bid in Dade County would be well below current MA payments and also well below the average FFS costs on which PPACA payments are based.
Because FFS enrollment is so much greater than MA enrollment (about 76 percent for FFS versus 24 percent for MA), the enrollment-weighted map in figure 5 looks more like the FFS map, with more light or very light than dark areas. If the weighted average change in payment (across MA and FFS) prompts political action in Congress, this map would predict opposition from California, Texas, Louisiana, and the mountain states, a few pockets along the East Coast (Miami, of course, but also Maryland, Massachusetts, and New York), and a few pockets among the northern and northwestern states.
Competitive bidding would save a substantial amount of money, helping solve Medicare’s fiscal crisis. We estimate the savings at 5.6 percent of Medicare costs, compared with the fully implemented PPACA. This estimate is likely to be conservative because it is based on the second-lowest bid, not the lowest bid that we prefer. In addition, we assume that competitive bidding will not create incentives for health plans to be more efficient; factoring in the increased efficiency likely to result makes our estimates even more conservative.
About one-third of beneficiaries would face an increase in monthly payments of $40 per month or more for their current plans. However, they could avoid this increase by switching to more competitively priced plans. Abrupt changes in payments for those who do not switch can be moderated by introducing competitive prices using blends of old and new methods, or by a limit on the per-year amount of change. However, buffering strategies reduce the savings in the short run. Efforts to prevent price increases for beneficiaries in the short run conflict with the need for savings in the Medicare program.
Additional buffering strategies could be used to assist low-income beneficiaries. Both the Ryan-Wyden and Domenici-Rivlin plans have proposed this type of assistance. This was especially important in versions of these proposals that do not ensure beneficiaries will always have at least one option available at the Part B premium. Where the bidding model ensures such an option in all areas, targeted assistance is not necessary to fix this weakness, though there may be other reasons to assist low-income beneficiaries (for example, to help with affordability of the Part B premium and copayments). In these situations, some balance will need to be struck.
We used four CMS data files:
- FFS Data 2009—Aged;
- FFS Data 2009—Disabled;
- MA 2009 Ratebook; and
- MA State/County Penetration—July 2009.
We merged these data files by county, after excluding US territories and possessions. In merging the files, we lost twelve counties or areas in Alaska and one in Hawaii because of no match. Otherwise, we merged Medicare data for all US counties.
FFS enrollment is based on Part A enrollment of aged and disabled beneficiaries. FFS spending includes indirect medical education, disproportionate share payments, and graduate medical education payments. Medicare reimbursements for hospice, end-stage renal disease, and cost contracts are excluded. Medicare Part D (the outpatient prescription drug benefit) is excluded, as this part of Medicare already is offered under a form of competitive bidding.
Per-county FFS spending was based on the risk-adjusted cost of beneficiary of average risk, also known as a “1.0-risk” beneficiary. The reason for using a 1.0-risk beneficiary is that plans should not be penalized for submitting high bids if they enroll high-risk beneficiaries or rewarded for low bids if they enroll low-risk beneficiaries.
We assumed constant 2009 FFS and MA enrollment by county and no MA plan entry or exit. Shifts in enrollment are not relevant for estimating the savings from competitive bidding because the low bidder sets the payment rate for all Medicare beneficiaries in a county. Some high-cost MA plans might exit the market under competitive bidding. Enrollees in those plans would have to switch to another MA plan, if available, or to FFS Medicare.
Roger Feldman (email@example.com) is the Blue Cross Professor of Health Insurance at the University of Minnesota, Robert Coulam (firstname.lastname@example.org) is the director of the Center for Health Policy Research at the Simmons College School of Management, and Bryan Dowd (email@example.com) is a professor of health policy and management at the University of Minnesota. The authors wish to thank Peter Graven for his assistance with data analysis and Joe Antos for helpful comments.
1. Robert H. Miller and Harold S. Luft, “HMO Plan Performance Update: An Analysis of the Literature, 1997–2001,” Health Affairs 21, no. 4 (July/August 2002): 63–86.
2. US House of Representatives Committee on the Budget, “Wyden and Ryan Advance Bipartisan Plan to Strengthen Medicare and Expand Health Care Choices for All,” news release, December 15, 2011, www.budget.house.gov/bipartisanhealthoptions (accessed February 13, 2012).
3. Bipartisan Policy Center, Domenici-Rivlin Protect Medicare Act, December 16, 2011, www.bipartisanpolicy.org/sites/default/files/Domenici-Rivlin%20Protect%20Medicare%20Act.pdf (accessed February 9, 2012).
4. Robert Coulam, Roger Feldman, and Bryan Dowd, Bring Market Prices to Medicare: Essential Reform at a Time of Fiscal Crisis (Washington, DC: AEI Press, 2009).
5. Coulam, Feldman, and Dowd, Bring Market Prices to Medicare; Marsha Gold et al., “Medicare Advantage 2012 Spotlight: Plan Availability and Premiums,” Kaiser Family Foundation Data Spotlight (November 2011), www.kff.org/medicare/upload/8258.pdf (accessed February 9, 2012).
6. Patient Protection and Affordable Care Act, Public Law 111-148, 124 Stat. 119, 111th Cong. (March 23, 2010), amended by the Health Care and Education Reconciliation Act of 2010, Public Law 111-152, 111th Cong. (March 30, 2010).
7. The PPACA savings of 4.2 percent and the 5.6 percent additional savings from bidding are multiplied to get the total savings of 9.5 percent from bidding.
8. Ideally, we would like to estimate the savings from competitive bidding using actual PPACA spending as the baseline; however, data on MA plans’ bids are available only for 2009, before the PPACA was passed. This means we had to start by estimating the PPACA baseline.
9. Scott Harrison et al., “MIPPA MA Payment Report” March 12, 2009, www.medpac.gov/transcripts/MA%20pay%20rpt%203%2009%20final.pdf (accessed February 9, 2012).
10. Total benefit payments from the hospital insurance and supplementary medical insurance trust funds in 2009 were $239.3 billion and $202.6 billion, respectively, for a total of $442 billion, according to tables III.B4 and III.C8 in The 2011 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Supplementary Medical Insurance Trust Funds, May 13, 2011, www.cms.gov/ReportsTrustFunds/downloads/tr2011.pdf (accessed February 13, 2012). Our estimate of baseline Medicare spending is lower because we excluded certain expenses noted in our appendix.
11. Ibid. This projection is based on the trustees’ intermediate estimates of Medicare spending from 2011 through 2020.