Health Affairs

This Article Abstract Submit a response to this article Services E-mail this article to a friend Alert me to new issues of the journal

S P E N D I N G  &  P L A N  C H O I C E:
M E D I C A R E  R E F O R M
W E B  E X C L U S I V E  

10 October 2001

Reforming Medicare: Impacts On Federal Spending
And Choice Of Health Plans

Reductions in Medicare spending are likely to occur
only if beneficiaries' premiums go up.

by Kenneth E. Thorpe and Adam Atherly


ABSTRACT:

The rising cost of Medicare and well-documented problems plaguing Medicare+Choice (M+C) have increased interest in "reforming" the program. To improve efficiency, most reform proposals would rely on competitive bidding to establish payments to M+C plans. At the same time, beneficiaries would be given financial incentives to select low-cost M+C plans. A major unknown is the extent to which Medicare reforms would generate federal budgetary savings. To examine this issue, we develop three illustrative Medicare reform options that differ greatly in how Medicare would establish its payments to plans. Our results highlight the fact that Medicare should expect modest savings from reforming the program. However, other goals of reform, such as establishing more efficient payments to plans, would be achieved.

Growing bipartisan interest in the need to "reform" Medicare has been driven by three problems: the rising share of the federal budget and gross domestic product (GDP) consumed by Medicare; an outdated benefit package; and distortions and inefficiencies in the payment methodology for Medicare+Choice (M+C) plans. It is hoped that solutions to the third problem can help to provide solutions to the first two. Several recent proposals for reforming payments to M+C plans have suggested abandoning the current administrative pricing system in favor of some form of competitive bidding among M+C plans. It has been suggested that competition among plans, combined with price incentives for beneficiaries to select lower-price plans, may serve to slow the overall growth in health care costs and Medicare payments.

Under the current system, payments for M+C plans are unrelated to plans' underlying costs. Instead, payments are derived from costs in the fee-for-service (FFS) sector. Prior to the 1997 Balanced Budget Act (BBA), managed care payments were 95 percent of average costs in the FFS sector (the adjusted average per capita cost, or AAPCC). Since the BBA, M+C payments have been established by a complicated formula that is the greater of a minimum payment (floor), a minimum update from the prior year's payment, or a blend of local and national rates, all of which are related to some degree to the 1997 AAPCC. The blend is subject to a budget-neutrality constraint, and no plan received the blended payment in 2001.

This system has a number of problems. First, for plans above the floor (65.6 percent of projected M+C enrollment in 2002 is in counties where the payment rate is above the floor), payments are still linked to the 1997 AAPCC. Second, from the plan perspective, year-to-year adjustments in payments are volatile; as a result, some plans exit the Medicare market and others are deterred from entering new markets. Finally, payments vary by county. Commercial plans and those in the Federal Employees Health Benefits Program (FEHBP) base payment on larger geographic units, such as metropolitan statistical areas (MSAs) or plan service areas. Plans operating in large MSAs, with a single provider network, face multiple reimbursement rates. For example, in 2002 the reimbursement rate in Bronx County, New York, will be $812 per beneficiary; however, in nearby Queens plans will receive only $735.

Competitive bidding could alleviate these problems by assuring that plan payments reflect underlying costs and that, over time, payment increases reflect the costs of efficient health plans. Properly structured, the use of competition in Medicare should equate Medicare payments in the M+C sector with the (efficient) cost of delivering services by M+C plans.

Despite these advantages, the extent to which M+C payment reform will generate federal budget savings is unknown. In concept, competition could reduce growth in Medicare spending in two ways. First, the competitive bidding process could result in additional efficiencies and lower payments to health plans than those established through regulation. Savings then could be realized if Medicare beneficiaries enroll in the lower-price health plans. Second, Medicare could collect additional premiums from beneficiaries choosing to remain in more expensive options. Most reform proposals provide financial incentives for beneficiaries to choose lower-price plans, although many will choose to remain in traditional FFS Medicare. Whether a competitive bidding process would reduce overall Medicare spending depends on the resolution of key design issues and how beneficiaries respond to them. This paper examines these issues using three illustrative Medicare reform options. Relying on estimates of how Medicare beneficiaries respond to financial incentives, we examine the likely effect of these changes on both Medicare spending and beneficiaries' premium payments.

Beneficiary premium. Under current law, Medicare beneficiaries pay a uniform Part B premium ($50 per month in 2001). Enrollees in an M+C plan also may pay a supplemental premium charged by the plan. Under most competitive bidding proposals, Medicare would provide a contribution toward the price of the M+C premium up to a predefined maximum. We refer to the premium associated with this maximum contribution as the reference premium (RP). Health plans will receive their premium bid regardless of where the RP is set, but the premium paid by beneficiaries joining either M+C plans or FFS Medicare will be determined by the RP. Beneficiaries joining plans with bids equal to the RP would simply pay the current-law Part B premium. If an M+C plan bid a premium above the RP, the beneficiary joining that plan would pay the Part B premium plus the difference between the plan premium and the RP. If an M+C plan bid below the RP, some portion of the difference between the RP and the bid would be returned to the enrollee in the form of a reduced Part B premium.¹ Thus, Medicare savings are the sum of changes in beneficiaries' premium contributions and changes in payments to health plans for Medicare-covered services.

Two broad reform approaches. There are two broad approaches for structuring Medicare reform. The first would link Medicare's contribution to FFS Medicare and return a portion of these savings through lower Part B premiums (or supplemental benefits) and reserve a portion of the savings for the Medicare program. If the "savings" (that is, the difference between the RP and the bid submitted by the M+C plan) generated through competitive bidding are larger than under current law, then both beneficiaries and the Medicare program could be better off.

The second would link Medicare's contribution to a low-price M+C plan, providing financial incentives for beneficiaries enrolled in FFS Medicare to switch to a lower-price plan. Savings to the Medicare program would occur if beneficiaries do not switch and therefore pay higher monthly premiums to remain in FFS Medicare or if they do switch, reducing Medicare payments to M+C plans. The impact on beneficiaries and the program in both cases depends critically on the savings or efficiencies (bids submitted) generated through the reforms and the willingness (price-sensitivity) of Medicare beneficiaries to enroll in lower-price plans.

Few studies have examined beneficiaries' price-sensitivity. Those that have suggest that Medicare beneficiaries are relatively insensitive to price. If this is correct, Medicare beneficiaries not currently enrolled in an M+C plan would require substantial financial incentives to switch to such a plan, and those currently enrolled in an M+C plan would require incentives worth at least the value of the additional benefits they receive today to remain enrolled under reform. However, providing beneficiaries with large financial incentives runs the risk of exhausting the efficiency gains produced by M+C plans, leaving little (if any) savings for the government. Thus, savings to the Medicare program will depend on the rules establishing the government's contribution, beneficiary payments, and the price-sensitivity of beneficiaries. To illustrate the importance of these design issues, we compare three different approaches for setting Medicare's and beneficiaries' contributions.

Establishing What Medicare And Beneficiaries Pay: The Reference Premium

The method used to establish the RP is among the most important design issues in Medicare reform. Most, although not all, reform proposals would solicit bids from M+C plans based on their cost of providing the core Medicare benefit package. This is the approach we explore here. Under such reforms, Medicare could set its contribution equal to the lowest (plan) bid in a market, the median (either nationally or in a market), or even traditional Medicare.

Under current law, there is wide geographic disparity in M+C plan payments. County-level variation in plan payments is intended to reflect differences in underlying costs but actually overstates these differences, leaving high-cost counties overpaid and low-cost counties underpaid. This raises equity issues, since M+C plans in high-cost/high-payment counties tend to offer generous supplemental benefits and not charge a premium. In contrast, M+C plans in low-cost/low-payment counties tend to offer few, if any, additional benefits and charge a monthly premium. Changes to the system therefore may seek to reduce geographic disparity. But geographically smoothing M+C payments without increasing overall spending would lead to payment reductions in the high-cost/high-payment counties-the same counties where the bulk of M+C enrollment now resides.

The illustrative plans. We present three illustrative plans to highlight the trade-offs involved in the establishment of the RP (Exhibit 1). Each proposal will have a different impact on Medicare program savings and what beneficiaries pay for M+C plans and traditional Medicare. The first two illustrative plans are similar to proposals developed by the National Bipartisan Commission on the Future of Medicare and, more recently, the Medicare proposals (S. 357 and S. 358) advanced by Senators John Breaux (D-LA) and Bill Frist (R-TN) in the 107th Congress. The third, which is similar to an approach outlined by Roger Feldman and Brian Dowd, would limit Medicare's contribution to the average bid from a qualified plan in a market.²

Under Option 1 the RP is county-specific and explicitly linked to the FFS sector (similar to S. 358). This approach is similar to the current Medicare program. Beneficiaries enrolling in lower-cost plans would pay lower premiums. In Option 2 the RP is based on a national enrollment-weighted average premium (similar to the FEHBP design and S. 357). Following the logic in S. 357, beneficiaries would pay 80 percent of the difference between the cost of the plan they enroll in and 85 percent of the national enrollment-weighted average premium.

Option 2 differs from Option 1 in two important respects. First, it breaks the explicit link between FFS costs and M+C plan payments. The RP is weighted for enrollment and is therefore a weighted average of costs in the FFS and managed care sectors. To the extent that managed care plans generate savings, this will lead to a lower overall RP. Second, the link between county costs and the RP is severed. Under both current law and Option 1, M+C plans in high-cost counties receive much higher plan payments. Under Option 2, some of the geographic variation in the RP is removed.³

Option 3 would set the RP equal to the average M+C bid in the county. For rural counties, where no M+C plans operate, this would be the FFS plan. Like Option 2, Option 3 breaks the link between FFS costs and M+C plan payments. However, as in Option 1, payments under Option 3 would be determined county by county. Geographic variation in managed care costs would lead to variation in plan bids and hence plan payments. The government's contribution under Option 3 would be lower than under Option 1 because it is linked to the average M+C bid (for Medicare-covered services) and not the higher costs associated with traditional FFS Medicare.

Potential Impact Of The Proposals On Medicare And Beneficiary Spending

Data and methods. We calculated reference premiums and the cost of providing Medicare-covered services for both traditional Medicare and M+C plans for fiscal year 2002. From this information, we estimated annual premiums that beneficiaries would pay under each of the three options. Our approach assumes that bids submitted by M+C plans will approximate the plans' cost of providing Medicare-covered services in 2002. A key unknown factor is how quickly M+C plan bids will reflect underlying costs. Our analysis assumes an immediate competitive response; thus, estimated M+C costs for providing core benefits and bids are the same.

To estimate county-specific FFS costs, we began with the 1997 AAPCC. In 1997 the Centers for Medicare and Medicaid Services (CMS, formerly HCFA) estimated that the AAPCC was equal to 98.1 percent of FFS costs.⁴ We then inflated our estimate of 1997 FFS costs by the FFS growth rate reported by the CMS to produce a county-specific estimate of FFS costs in 2002. We also removed 100 percent of all graduate medical education expenses (direct and indirect) from FFS cost calculations to produce a county-specific estimate of the cost of providing the basic benefit package in the FFS sector.

We based our estimate of this cost on an unpublished study from the Congressional Budget Office (CBO) and the CMS, which reported that in 1997 managed care plans' costs to provide core Medicare benefits were approximately 80 percent of the AAPCC. A confirmation of this figure was recently reported by the CMS, which revealed that the bids submitted by managed care plans in Denver for the equivalent of the FFS benefit package ranged from 25 percent to 38 percent below the BBA payment rates.⁵ We trended these costs to the year 2002 using data from major M+C plans (PacificCare and Aetna) on growth in medical costs (not plan payments) for their Medicare products.⁶ On average, we assumed that managed care costs for Medicare-covered services increased 5 percent annually during 1997-2002.

We first examine markets that currently have M+C plans, using data from the CMS.⁷ In these markets we estimate that it costs traditional Medicare, on average, approximately $7,096 per year to provide the basic benefit package (weighted by Medicare enrollees). In those same counties it would cost the average M+C plan $5,947 to provide the same set of core Medicare benefits for the same population, although we estimated that it cost M+C plans $6,456 to provide services for their enrollees because M+C enrollment is concentrated in high-cost/high-payment counties. M+C plans are paid an average of $7,272 per year for their enrollees (again reflecting the concentration of enrollment in high-cost counties). Thus, today M+C plans are providing, on average, approximately $816 more in additional benefits than they are charging in additional premiums.⁸

Option 1. Under Option 1 the RP is established at the higher of the county FFS costs or the M+C payment rate. Under this option, which mimics current law, much of the current distribution of Medicare payments to M+C plans would be retained, and the RP would vary across counties. The average RP would be equal to the average current M+C reimbursement ($7,272). Beneficiaries choosing to remain in traditional Medicare under this option would pay a premium ($688 per year in 2002)-just as under current law (Exhibit 2). Beneficiaries choosing a lower-cost plan, such as a typical M+C plan, would pay a lower premium-$76 in the average county (the traditional Medicare premium less 75 percent of the difference between $7,272 and $6,456) per year in our example. Beneficiaries could use their savings (at least $612 per year relative to current law) to purchase supplemental benefits such as prescription drugs, vision care, and other services not covered by Medicare.⁹

Option 2. The RP for Option 2 is the average of FFS Medicare and each M+C plan, weighted by enrollment. In our illustration this weighted (using 2001 enrollment) average is $6,599. The RP for Option 2 is far lower than that for Option 1, for two reasons. First, it is partially based on the lower-cost M+C sector. Second, it is based on the national average in the FFS sector, which includes many low-cost counties where no M+C plans now operate. The premium for beneficiaries joining plans with a bid equal to the RP is higher than the current-law Part B premium-$792 per year (80 percent of the difference between the RP and 85 percent of the RP). Under Option 2 beneficiaries choosing to remain in traditional Medicare would pay $792 per year plus the full difference between the RP and the national average cost of Medicare FFS, which amounts to $817 annually-nearly a 19 percent increase. Following the discussion from the Bipartisan Commission on the Future of Medicare, we limit the variation in plan payments to 70 percent of today's variation in county-level FFS payments.

By establishing a single RP, Option 2 would raise plan payments in low-cost/low-payment areas and lower them in high-cost/high-payment areas. In areas that have M+C plans, the premium would be equal to $677 (80 percent of the difference between average costs in the M+C sector and 85 percent of the RP). In areas without M+C plans, the average cost in the FFS sector ($5,676) is only slightly above 85 percent of the RP ($5,608). If M+C plans bid FFS costs in the county, the beneficiary premium would be close to zero. The FFS premium in these non-M+C counties would be identical to that in M+C counties: $817. Because the RP in Option 2 is based on a national average, beneficiaries living in markets with an M+C plan today would pay more for FFS Medicare as well as for a managed care plan, compared with Option 1.

Option 3. The RP for Option 3 is equal to the average M+C plan bid in a county. For beneficiaries enrolling in the average M+C plan, the premium would be the same as the current-law Part B premium: $688 annually. Beneficiaries choosing to remain in the FFS sector would pay $688 plus the difference between the cost of traditional Medicare and the average M+C plan in the county. The out-of-pocket cost for FFS Medicare would vary depending on whether an M+C plan operated in the county. If there were no M+C plan, then the FFS plan would be the low-cost plan, and the premium would equal $688. However, in the typical market in which M+C plans now operate, the premium would be equal to the Part B premium ($688) plus the marginal cost of the FFS sector, for a total of $1,216 per year-a 77 percent increase. So, while the price of the basic benefit package would be uniform nationally, the monthly premiums associated with FFS Medicare would vary across counties.

Results: Medicare+Choice Enrollment And Medicare Spending

The key to our analysis is to compare changes in the dollar value of additional benefits provided by M+C plans and changes in what beneficiaries pay for the plans they choose. We calculate these dollar changes and multiply them by a measure of price-responsiveness (elasticity) similar to those used by the CMS and the CBO. If the change in dollar benefits offered Medicare beneficiaries under reform is small (and beneficiaries are price-insensitive), few additional beneficiaries will enroll in M+C plans. Under these three proposals, the additional benefits offered by plans today would be eliminated, and M+C plans would compete with the FFS sector for beneficiaries by offering a lower out-of-pocket premium. This amounts to trading dollars for benefits. Enrollment in M+C plans will depend in large part on which is greater.

Enrollment. To estimate the enrollment in M+C plans, we used methods similar to those used by the CBO and the CMS to estimate the impact of President Bill Clinton's Medicare reform proposal in the 106th Congress.¹⁰ This approach compares the change in the dollar value of the additional benefits (above the basic Medicare package) offered by M+C plans (this year versus last year) with the change in the annual premiums when purchasing an M+C plan or FFS Medicare. Using the standard economic assumption that consumers value cash more than in-kind benefits, we discounted the dollar value of benefits gained or lost in our estimates. The switching elasticities are similar in magnitude to unpublished work we have completed, as well as to recently published estimates that focused on the Medicare population (see Note 10). These switching elasticities indicate that compared with the population under age sixty-five, the Medicare population is not very price-sensitive.

Option 1. Under Option 1, if health plans passed their savings along to beneficiaries in the form of lower premiums, M+C enrollment would only rise from 5.60 million to 5.63 million (Exhibit 3). This is because, from the beneficiary's perspective, the value of the extra benefits lost almost exactly offsets the value of the lower premium. Medicare would save through lower payments to health plans- -approximately $4 billion annually when fully phased in. However, virtually all of these savings ($3.4 billion) would be returned to beneficiaries in the form of lower monthly premiums. On balance, we estimate that this approach would reduce Medicare expenditures by a very modest amount-approximately $1 billion when fully implemented.

Although Option 1 would generate slight savings, it appears that the Medicare reform actually proposed in S. 358 would increase Medicare spending (see Note 10). This is because that proposal sets the RP for plans at 100 percent of FFS plus an adjustment for costs incurred by Medicare beneficiaries at Department of Veterans Affairs and Department of Defense facilities. These adjustments would increase the RP to approximately 103 percent of FFS costs rather than 100 percent under Option 1.

Option 2. Option 2 would produce slightly more, although still modest, savings for the Medicare program. A slightly higher number of beneficiaries would enroll in M+C plans-6.2 million compared with 5.6 million today. Most of the additional M+C enrollment is expected to come from markets without an M+C plan available now. These are generally markets with a commercial health maintenance organization (HMO) presence but no M+C plans. We expect new enrollment in these areas because the national premium will greatly increase M+C plan payments in low-cost areas. Medicare would collect higher premiums from beneficiaries choosing to remain in traditional Medicare, more than offsetting the lower premiums from beneficiaries choosing lower-price plans. In this option, plan payments would be unchanged from current law despite the increased enrollment (due to lower plan payment rates), FFS expenditures would fall by nearly $4 billion, and beneficiary premiums would rise by $1 billion-generating about $5 billion in savings.

Option 3. Option 3 would produce much greater savings. The number of beneficiaries enrolling in M+C plans would be similar to today's enrollment-about 5.6 million. This is because the value of the free benefits in M+C plans "taken" from beneficiaries is similar to the increase in the traditional FFS premium. The savings are much larger-more than $16 billion annually. However, almost all of these savings are derived from higher premium payments-some 77 percent above current-law premiums-by beneficiaries in the FFS sector.

Each option generates approximately the same relatively small dollar reduction in Medicare payments to health plans-approximately $1-$5 billion when fully implemented. The options differ dramatically, however, in the amount of premium contributions collected from beneficiaries. Thus, the major difference in Medicare savings across the options is higher premiums rather than lower plan payments.

The substantial increase in the cost of remaining in traditional Medicare in Options 2 and 3 highlights the importance of risk adjustment. As the costs of FFS Medicare rises, lower-cost beneficiaries will face increased incentives to migrate to M+C plans. Absent an adequate risk adjuster, this would further increase the cost of FFS Medicare.

Summary And Discussion

Our results highlight the difficulty of greatly increasing the number of Medicare beneficiaries enrolling in M+C plans. Given beneficiaries' relative price-insensitivity, enticing them to enroll in M+C plans will require a substantial financial incentive. We now provide this incentive by passing on M+C savings in the form of additional benefits. However, our analysis reveals the difficulty of simultaneously eliminating these additional benefits, inducing additional plan switching, and generating savings to the Medicare program. This task is even more daunting since most reform proposals would include a prescription drug benefit with a uniform subsidy available to beneficiaries in either FFS Medicare or an M+C plan.

Under current law, Medicare beneficiaries receive approximately $1,000, on average, in additional benefits when enrolling in an M+C plan. For these additional benefits, beneficiaries pay an average of $276 per year in supplemental premiums.¹¹ Most Medicare reform proposals would convert these additional benefits to cash, in the form of lower premiums. However, to entice additional Medicare beneficiaries to enroll in an M+C plan will require an even larger financial reward than under current law. These financial incentives may be structured as a financial "carrot" or a financial "stick." Option 1 provides financial incentives in the form of lower premiums. Beneficiaries now receive all of the savings generated by M+C plans in the form of lower benefits. Under Option 1 they would receive 75 percent of the savings in the form of lower monthly premiums, approximately the same "value" provided by M+C plans today. Thus, this approach is unlikely to result in additional M+C enrollment and produces very limited savings to the Medicare program.

Options 2 and 3 are more substantial reforms, as they break the link between M+C payments and the FFS market and provide a financial "stick" to move beneficiaries into M+C plans. Beneficiaries choosing to remain in traditional Medicare would face a substantial hike in monthly premiums-ranging from 19 percent to 77 percent-in our illustrative options. However, even under these approaches relatively few additional Medicare beneficiaries would be likely to enroll in M+C plans. This is in part traced to the small relative difference in premiums to purchase Medicare-covered services in the form of an M+C-delivered benefit or through traditional Medicare (see Exhibit 3). These dollar differences are less than the value of additional benefits provided by M+C plans today. On the other hand, these approaches would generate program savings, as they increase premiums paid by beneficiaries when enrolling in Medicare Part B.

None of our results mean that Medicare reform and competitive bidding are bad ideas. These reforms would provide solutions to several important problems facing Medicare. Our results, however, highlight the fact that reforming the M+C payment system is unlikely to generate major federal budget savings absent substantial changes in the program. Each of the options examined produces virtually the same small savings in Medicare spending. The key difference across the proposals is the share of Medicare spending financed by Medicare beneficiaries. As our final option highlights, reductions in net Medicare expenditures under reform are likely to occur only through substantial increases in premiums paid by beneficiaries.

The authors acknowledge the comments from three anonymous referees. In addition, their analysis benefited from conversations with staff from Congress and the Congressional Budget Office.

NOTES

1.Some Medicare reform proposals would allow M+C plans to pass savings along in the form of lower premiums, additional benefits, or both. Under the provisions outlined in H.R. 5661, the Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act of 2000 (106th Congress), M+C plans may spend their savings in the form of either lower Part B premiums, additional benefits, or both starting in 2003. Our analysis assumes that plans would pass all savings along to consumers in the form of lower premiums. As described in Note 10, this assumption produces higher estimates of plan switching compared with assumptions that plans pass these savings to consumers in the form of additional benefits.
2.R. Feldman and B. Dowd, "Structuring Choice under Medicare," in Medicare: Preparing for the Challenges of the Twenty-first Century, ed. R. Reischauer, S. Butler, and J. Lave (Washington: National Academy for Social Insurance, 1998), 75-123.
3.This option would have plans bid based on Medicare-covered services. For actual plan payments, however, Medicare would publish geographic adjusters (that may only reflect wage differences) and risk adjusters. Plans would use these data when preparing their bids. If wage differences are reflected in the geographic adjusters, there would be less variation in the geographic adjusters than in actual premiums or M+C payments under current law. This would result in lower payments than today in high-cost areas and higher payments in low-cost areas. In our illustration we passed through 70 percent of the underlying variation in current-law payments in the geographic adjustment.
4.Centers for Medicare and Medicaid Services, M+C (AAPCC) Payment Rates Information, www.hcfa.gov/stats/hmorates/aapccpg.htm (13 July 2001).
5.B. Dowd, "More on Medicare Competitive Pricing" (letter), Health Affairs (Jan/Feb 2001): 306-307. Data indicating that the cost of providing the core Medicare benefit package is approximately 80 percent of M+C payments were reported by Nancy-Ann Min DeParle, as part of the session on Government Payors and the Future of Medicare and Medicaid HMOs at the Council on the Economic Impact of Health System Change conference on the Future of Managed Care, Princeton, New Jersey, 17-19 May 2001. We assume that plans (in 1997) with more generous benefits have lower costs, and those with less generous benefits have higher costs, relative to the 1997 AAPCC. We normalize the average ratio of costs to the AAPCC at 80 percent. Thus, the underlying distribution of the adjusted community rate (ACR) to the AAPCC for 1997 is apportioned across plans according to the generosity of supplemental benefits provided in 1997. These underlying costs are then increased by 5 percent per year from 1997 to 2002 (see Note 10).
6.Our estimates were based on M+C cost increases reported by PacifiCare and Aetna, two major health plans: www.pacificare.com/corporate/news/pressrel and www.aetna.com/presscenter.index.htm (13 July 2001). These reports include both Medicare-covered services and additional benefits (prescription drugs). To develop estimates for Medicare-covered services, we excluded the estimated growth in prescription drugs from these two published reports using data from the CMS, www.hcfa.gov/stats/NHE-Proj.
7.CMS, M+C (AAPCC) Payment Rates Information.
8.This is an average figure. M+C plans provide more than this figure in additional benefits since they, on average, charge a monthly premium averaging of $23 per month; see www.medicare.gov/mphcompare/home.asp under the Medicare Compare data set. Our calculations indicate that by 2002 the cost of M+C plans to provide the core Medicare services will be approximately 88 percent of M+C payments, a higher figure than the 80 percent estimated in 1997. This difference stems from the fact that the underlying costs of providing core benefits have increased faster than M+C payments since 1997. Reductions in the value of additional benefits provided by M+C plans over time reflect these trends.
9.Most Medicare reform proposals also include a voluntary Medicare drug benefit. Depending on the proposal, some Medicare beneficiaries (such as those under 175 percent of the federal poverty level) would receive this benefit at no or a greatly reduced cost, while higher-income seniors would pay 50-75 percent of the drug premium. Thus, some of these savings could be used to finance the cost of purchasing a drug benefit.
10.These analyses compare changes in M+C benefits in the reform year relative to the prior year, as well as changes in M+C premium and FFS premiums. We quantify M+C benefits as the difference between the M+C payment rate and our estimate of their costs. As noted in the text, these costs are based on 1997 ACR reports, trended forward over time using published data on Medicare cost trends. This approach is similar to that used in the CBO analysis. We use a similar formulation, but we link the changes in market share to published elasticity measures (similar to the published work noted below) to develop our estimates. Our internal results are very similar (although the switching elasticities are higher) to recent findings from T. Buchmueller, "The Health Plan Choices of Retirees under Managed Competition," Health Services Research (December 2000, Part 1): 949-976. Both our approach and the CBO method generated similar findings. The price elasticities from this literature are approximately -0.2. We use a range of utility factors, varying from 0.5 to 0.8, in our analysis. Thus, our analysis tabulates the change in dollar benefits (both premiums and additional benefits) in an M+C plan, plus the change in M+C premiums and the change in FFS premium as a percentage of the value of the core Medicare benefit package. We multiply this figure by an elasticity to calculate the change in market share for each M+C plan.

For Option 2, we assumed that in markets with a commercial HMO, without M+C plans, and receiving a payment increase above FFS levels, M+C plans would have an incentive to enter. Entry into these markets would occur over time, so actual savings in this option would occur over time as well.

The exhibits show savings associated with full implementation. Following CBO assumptions, we assumed that plans would phase in their bids (relative to underlying costs) in the initial years of the program. In addition, following the work of W.P. Welch, "The Elasticity of Demand for Health Maintenance Organizations," Journal of Human Resources (Spring 1986): 252-266, we assume that the ultimate enrollment shifts would occur over a ten-year period, with a third of the ultimate effects occurring in the first year. Thus, over a ten-year budget period the change in Medicare spending would actually increase by $12 billion with Option 1 to $55 billion in Medicare savings with Option 3. This assumes the language concerning the reference premium in S. 358. The RP includes the costs of Veterans Affairs and military health plan payments for Medicare beneficiaries. This increases the RP from 100 percent of FFS in a county to 103 percent of FFS costs on average. Thus, S. 358 would result in higher Medicare expenditures.
11.The average difference in the M+C payment rate and costs is about $816. On average, M+C plans also charge a $276 annual supplemental benefit. Thus, the approximate dollar value of all benefits provided by a "typical" plan is approximately $1,000 per year.