Premium Rebates and the Quiet Consensus on Market Reform for Medicare

Abstract

Premium rebates allow beneficiaries who choose more efficient Medicare options to receive cash rebates, rather than extra benefits. That simple idea has been controversial. Without fanfare, however, premium rebates have become a key area of agreement in the debate on Medicare reform. Moreover, in legislation in late 2000, it became official policy: Medicare+Choice (M+C) plans will be allowed to offer rebates beginning in 2003. This article explores the economic rationale for premium rebates, provides a historical perspective on the rebate debate, discusses some of the implementation issues that need to be addressed before 2003, and reviews the implications of premium rebates for current legislative proposals for Medicare reform.

Introduction

M+C plans, almost entirely health maintenance organizations (HMOs), are paid a fixed amount per member, per month, by Medicare. Medicare regulations require efficient plans—plans that can produce the Medicare entitlement benefits for less than the amount Medicare pays—either to make up the difference in added benefits or to return the excess payment to the government (an unlikely alternative). M+C organizations have not been permitted a third alternative: to give premium rebates to beneficiaries, i.e., to give any or all of the excess back to beneficiaries, in the form of cash payments. The underlying premise, largely implicit, was that beneficiaries needed to be protected from themselves—that beneficiaries would make bad health care decisions when faced with the tantalizing prospect of increasing their incomes by receiving a rebate check. Efficient plans thus were allowed to signal their efficiencies by offering more generous benefits, but that was considered signal enough. Economists (Dowd, Feldman, and Christianson, 1996) criticized this limitation, arguing that cash rebates would be more efficient than extra benefits and would permit a more direct form of price competition between M+C plans and fee-for-service (FFS) Medicare.

Public policy is about to change in the direction advocated by the critics. In the process, a consensus on Medicare reform has quietly emerged, with premium rebates as its common link. The reform proposal of the previous administration, the revised Breaux-Frist bill in the Senate, and the Medicare Competitive Pricing Demonstration all proposed to allow beneficiaries to receive rebates. By late 2000, premium rebates received formal legislative approval in the Benefits Improvement and Protection Act (BIPA) of 2000. Starting in 2003, plans will have the option to elect payment reductions of up to 125 percent of the Part B premium, with plan enrollees receiving 80 percent of any reductions (i.e., a maximum of 100 percent of the Part B premium), with the government slated to receive the remainder.

The current consensus shows premium rebates to be the key area of agreement on introducing market-oriented pricing to the Medicare program. Premium rebates improve the structure of the M+C program, but they will require some care in implementation and will not have the same effects everywhere. This article explores the economic rationale for allowing Medicare health plans the flexibility to offer rebates, provides a historical perspective on the rebate debate, and discusses some of the implementation issues that will need to be addressed for Medicare to permit M+C plans to offer rebates. In the final section, we review the implications of this discussion for current legislative proposals for Medicare reform.

Background

Medicare, the national health insurance program for the elderly and disabled, allows beneficiaries to receive coverage through M+C plans. About one-seventh of the 40 million Medicare beneficiaries have chosen this system.¹ Initially, payments to M+C plans were based on Medicare FFS costs in the same area. Year-to-year increase factors are now based on methods set forth in the Balanced Budget Act (BBA) of 1997 (Public Law 105-33) and later refinements, rather than FFS costs per se. With the exception of minimum payment or “floor” counties, however, the pattern of M+C payment levels across the country still reflects historic FFS costs.

M+C plans must cover all services to which Medicare beneficiaries are legally entitled. M+C plans also are permitted to offer enhanced benefits not included in the entitlement, and to charge an out-of-pocket premium for whatever benefits they offer (whether or not enhanced beyond the basic entitlement). These premiums and benefits appear to vary across the country in certain characteristic ways:

• In 2000, out-of-pocket premiums varied from zero in some counties to over $100 per month in others. This variation appears inversely related to the government's monthly payment rate—not surprisingly, if the government pays more, the beneficiary usually pays less (McBride, 1998; Pizer, Frakt, and Doksum, 2001).

• As with out-of-pocket premiums, benefit enhancements² offered by M+C plans also vary by the payment level and other factors. The Physician Payment Review Commission (1997) found that risk plans in Miami, one of the highest paid areas in the county, offered additional benefits worth $125 per month in 1995 for no premium; in contrast, plans in low-paid non-metropolitan areas of Florida offered benefits worth much less. More recently, McBride (1998) and Pizer et al. (2001) confirmed that access to additional benefits is often better in urban counties than in rural counties. The last study also found that the intensity of competition (measured by the number of plans in the county) was positively related to the generosity of additional benefits offered by M+C plans.

By 2000, the benefits plans offered and the premiums plans charged were reflecting the strains of lower payment increases.³ But it still remained true that M+C plans in high-payment areas were competing by offering benefit enhancements, at little or no cost to enrollees. And it also remained true that the premiums and benefit enhancements varied with factors other than payment, notably, urbanicity and the intensity of competition in a county.

Benefit variation in a national entitlement program like Medicare raises an important equity issue. Medicare was founded on the principles that the program's tax rates and benefits would be uniform nationally. The principle of benefit equity has been undermined substantially already, and the variations in benefits could become even greater if M+C plans are allowed to add a potent additional enhancement—extra cash—to the benefits already available in high-payment areas.

Before reviewing the politics and history of rebates, we provide an economic analysis of the current competition over benefits rather than over price. We show that this type of competition is inefficient, meaning that enrollees could be made better off without any additional burden to taxpayers by allowing plans to offer cash rebates in lieu of some or all of the benefit enhancements they currently offer.

M+C Plan Competition Model

Our model of M+C plan competition is taken from research by Feldman and colleagues (1993), Dowd and Feldman (1996), and Feldman and Dowd (1998). The model focuses on the incentives of M+C plans in a payment system where the government determines the M+C capitation rate based on administrative calculations of FFS costs. M+C plans can offer benefits (e.g., coverage of outpatient prescription drugs) that enhance the required Medicare benefits and can charge an out-of-pocket premium for them.

Suppose that a profit-seeking M+C plan faces a downward-sloping demand curve for enrollment. In other words, as the plan cuts its out-of-pocket premium, more beneficiaries will switch to the plan from FFS Medicare and from other M+C plans. The M+C plan also factors the government premium contribution into its profit calculation. If the government contribution rises, for example, the plan calculates that it will be profitable to enroll more beneficiaries. Therefore, we predict that the plan will respond to an increase in the government contribution by cutting its premium, as found by Feldman and colleagues (1993).

It is less clear what happens to the level of benefit enhancements offered by premium-charging M+C plans as the government contribution increases. Even if all current and potential enrollees prefer more generous over less generous benefits, the plan may not offer a richer benefit package as the government contribution increases. The plan's decision depends on whether new enrollees who are attracted by lower premiums place a higher or lower value on benefits than the plan's current enrollees. Three cases are possible:

• The current enrollees may demand more additional benefits than the new enrollees who join as the premium falls. The M+C plan would respond to decreasing demand for benefits by reducing the benefit enhancements as its enrollment increases. Although this outcome (rising enrollment and falling benefit enhancements) seems counter-intuitive, it is not implausible: the plan's current enrollees may well demand more benefit enhancements than the new enrollees.

• All enrollees may have equally strong demand for benefits. A plan catering to these preferences would hold the level of benefit enhancements constant as its enrollment grows.

• New enrollees may demand more benefits than current enrollees. The M+C plan would offer more generous benefits in this last case if the government contribution increased.

Despite their differences, all three cases share a common assumption that premium-charging M+C health plans use both premiums and benefits to compete for new enrollees. Common sense suggests that a plan will be indifferent to cutting its premium by $1 or increasing the cost of the benefit package by $1 if these strategies have the same effect on the plan's profit. It is equally sensible that the plan would not offer $1 of additional benefits if it were more profitable to reduce the premium by $1.

In selecting which benefits to offer, the plan also will consider the relationship between risk selection and the nature of additional benefits. A more generous prescription drug benefit may attract an unhealthy group while an exercise program is more appealing to a healthier crowd. In a sense, the balance is between how much to lower premiums and how much and what type of benefits to add. As the government raises the payment rate, regardless of which benefits the premium-charging plan uses to compete for enrollees, eventually the plan will cut its premium to zero. If the government contribution continues to increase, the plan will want to grow, but the only way to grow when the premium is zero is to offer more benefits. Because the benefits that most consumers are willing to pay for out-of-pocket already are being offered (as evidenced by the out-of-pocket premiums being charged prior to the increase in the government contribution), the new benefits offered at zero premium are not likely to be valued at their cost by most consumers. Feldman and Dowd (1998) refer to these latter benefits as inefficient benefits. Beneficiaries would not purchase inefficient benefits with their own money because they do not value the benefits as much as they cost to produce. In fact, they would not purchase inefficient benefits even if they were spending taxpayers' money, if they had the option of taking the cash instead. In contrast, the benefits offered by M+C plans that charge out-of-pocket premiums are efficient—if they were not, the plans would (or, at least, should) drop them and reduce premiums.

Premium Rebates

Premium rebates probably will not matter in areas that currently are offering an efficient level of benefits—broadly, the counties in which all M+C plans currently charge a premium for their basic benefit package. In 2001, 61 percent of all beneficiaries lived in counties with no zero-premium plans,⁴ but for the remaining 39 percent of beneficiaries, premium rebates may be a potent catalyst for increasing enrollment in M+C plans, compared with providing inefficient benefits. Rebates would probably attract beneficiaries from FFS Medicare as well. Most beneficiaries supplement FFS Medicare with additional medigap insurance (Rice and Bernstein, 1999). Beneficiaries are sensitive to the cost of medigap premiums (McLaughlin, Chernew, and Taylor, 2001). A premium rebate would increase the cost of FFS-plus-medigap in relation to the cost of M+C plans. This increase should attract at least some beneficiaries from FFS to the M+C plans that offered rebates.

The option of offering cash rather than benefits would present health plans with an interesting choice. There may be significant variation in the demand for rebates by health status (healthy beneficiaries may prefer rebates), so plans might try to use rebates to select favorable risks. But healthy beneficiaries also are attracted by inefficient benefits such as health club memberships, so it is not clear how rebates would affect risk selection, in comparison to at least some enhanced benefits a plan might offer. In addition, rebates may attract low-income beneficiaries, who are often less healthy than higher-income beneficiaries. Plans and beneficiaries need to determine the appropriate balance of benefits and rebates in a market environment (and CMS should monitor changes in benefits and the apparent impact on selection, when rebates are offered). If some benefits currently offered are inefficient, beneficiaries will prefer rebates. On the other hand, some benefits may be sufficiently valuable that beneficiaries will prefer those benefits instead of rebates. One indication of the appropriate balance between benefits and rebates is the current experience of plans in low-payment areas that offer additional benefits and charge a premium. The preventive care benefits offered by such plans might be the best prediction of what all M+C plans would offer if the Medicare program allowed premium rebates, as the benefit-premium mix is most likely to represent efficient combinations in these areas.

Meanwhile, it is worth noting the results, previously mentioned, describing the relationship of payment levels, premiums, and benefits. One of the most important findings of the research described is that premium levels and benefit generosity have the expected relationship to payment levels, but that other variables—in particular, the intensity of competition and whether the county is urban or rural—also have an important influence on premiums and benefits. Those other variables are also likely to influence the presence and size of premium rebates. For example, plans would likely feel pressure to increase the size of rebates in more competitive market areas, and in urban areas, just as they appear to do with benefits generally.

In truth, it remains to be seen how plans and beneficiaries will respond to the new option. Rebates could have special significance because they are a cash incentive in a program that has not used this incentive to date. On the other hand, rebates may be a minor option. M+C enrollees may prefer their current benefit enhancements, especially if the government taxes rebates at high rates. Even if rebates are offered, few beneficiaries may switch from FFS Medicare to M+C plans that offer rebates. For these reasons, it may be prudent to demonstrate the idea, before implementing it systemwide.

History

If premium rebates are so desirable, it is worth asking why they have not been introduced before (Dowd, Coulam, and Feldman, 2000). To date, the Medicare managed care program has prohibited premium rebates. When asked about this prohibition by one of the authors, a CMS manager replied, “We don't give away toasters.” This response reflected a paternalistic uneasiness with the choices that beneficiaries might make if health plans offered cash, as well as benefits, to potential enrollees. Offering premium rebates means creating a new benefit, but it also requires a change in thinking—toward accepting a more central role for market incentives in the Medicare program—that CMS was not initially prepared to take.

While CMS did not wish to implement premium rebates in the Medicare program, the agency, in fact, worked very hard in the 1990s to introduce competitive pricing in Medicare, in the form of demonstrations of competitive pricing for HMOs, durable medical equipment (DME), and clinical laboratory services.⁵ Our principal concern here is with the efforts to introduce competitive pricing for Medicare HMOs, and the relation of those efforts to premium rebates.

Starting in the mid-1990s, CMS developed a bidding model for HMOs, using technical consultants and expert panels. The model required all HMOs to bid on a standardized benefit package (plans could offer optional supplements as they chose). The government would base its premium contribution on the distribution of bids received from the plans. Plans that bid high—over the cutoff price set by the government—would have to charge a premium to enrollees, equal to the difference between the cutoff price and the plan's bid.

Two major features of CMS's bidding model concerned us: the exclusion of FFS Medicare and the exclusion of premium rebates from the demonstration. With respect to FFS, CMS decided at the outset that the agency did not have the legal authority to include FFS Medicare in the demonstration. One reason for CMS's concern was political. Including the FFS sector would have meant higher premiums for FFS beneficiaries in demonstration areas. The reason: FFS was expected to be more expensive than M+C plans in the high-payment areas where the demonstration was likely to be sited (i.e., any FFS bid was expected to be higher than most M+C plan bids). Charging a new premium to FFS beneficiaries in demonstration areas would have increased enormously the political opposition to the demonstration.

With respect to premium rebates, CMS decided not to allow plans to give cash rebates directly to enrollees, even a rebate limited to the amount of the Part B premium, as an incentive for choosing particular plans. This decision appeared to be based on the agency's longstanding position against using financial incentives to influence beneficiary choice in the Medicare program. As the demonstration design evolved, agreement was reached to allow low-bidding health plans to enrich their benefit packages. This decision was consistent with past agency policy against cash rebates, but it reinforced the prevailing method of restricting the reward to additional benefits, rather than cash.

With the exclusion of FFS bids and rebates from the demonstration, the possibility of direct price competition between FFS and M+C was substantially reduced. The FFS price could not be raised by bidding, and the M+C price could not be lowered by rebates. The only direct price competition would be among M+C plans, with penalties for bids above the cutoff payment level.

CMS attempted to implement this design (without FFS or premium rebates) in Baltimore in 1996 and Denver in 1997. In each case, opposition from local health plans and State congressional delegations (plus, in Denver, a lawsuit) stopped the effort.

Although Congress blocked the Denver demonstration, there still was considerable support for the idea of competitive pricing among some influential members. Senator John Breaux and colleagues strongly committed to the idea successfully pushed for a competitive pricing amendment to what became BBA 1997. The BBA required the Department of Health and Human Services to design and implement a series of demonstrations for payment of HMOs (which BBA now grouped with other M+C plans). CMS was required to work under the guidance of a national expert committee (CPAC) and local expert committees (Area Advisory Committees) that included members from all the important stakeholder groups.

CPAC began its work in May 1998. CPAC designed a demonstration similar to the model previously proposed for Baltimore and Denver. As in those sites, the CPAC design excluded FFS Medicare, given the lack of legal authority to include it. CPAC raised the issue of premium rebates early in its discussions, as rebates promised to address some of the political problems that had thwarted the demonstration to that point. One of the main reasons health plans gave for their opposition to the demonstration was that FFS had been excluded. The American Association of Health Plans, in particular, claimed that a demonstration excluding FFS would be guilty of “tilting competition unfairly against private plans” (Weinstein, 1999) because FFS would not be subject to the same competitive pressures that CMS proposed to put on health plans. Thus, exclusion of FFS from the demonstration increased industry and congressional opposition to the project. Premium rebates might serve to moderate that opposition, as rebates would create an important link between FFS Medicare and the competitive prices determined in the demonstration. If a low-bidding plan offered a premium rebate, it would reduce the price of that M+C plan below the Part B premium that FFS enrollees still would pay for their coverage, thereby creating direct price competition between M+C plans and FFS. To be sure, this price competition would come in the form of opportunities for low-bidding M+C plans to offer reduced premiums, rather than increased premiums for FFS, as would follow from complete inclusion of FFS at most sites. But given the absence of congressional authorization, complete inclusion of FFS in the demonstration was out of the question.

Although premium rebates presented an opportunity to change the politics of competitive pricing, it was not an opportunity that initially could be embraced. After making early inquiries, CPAC was briefed by CMS on the substantial administrative difficulties of including premium rebates and by the Department of Health and Human Service's Office of General Counsel (OGC) on possible legal difficulties as well (e.g., conflicts with the anti-kickback statutes). Accordingly, CPAC set aside the rebate option and created an initial design for the demonstration that excluded premium rebates.

A vector of considerations combined to put rebates back on the table. First, on June 29, 1999, the previous administration proposed a significant reform of the Medicare program that included premium rebates. That proposal, dubbed the Competitive Defined Benefit, would pay health plans for covering Medicare's defined benefits, including a new subsidized drug benefit. The Competitive Defined Benefit would provide beneficiaries with a rebate of $.75 for every dollar of savings that resulted from choosing lower-cost plans (the government would get the other $.25). Beneficiaries opting to stay in the traditional FFS program would be able to do so without an increase in premiums. This proposal legitimized CMS in taking another look at premium rebates.

Second, Congress also became interested in Medicare premium rebates. By late 1999, the demonstration had run into substantial local opposition, and the CPAC and the Area Advisory Committees at the initial demonstration sites (Kansas City and Phoenix) were exploring ways to make the demonstration design more attractive to beneficiaries. In private conversations, key congressional staff stated to CPAC members that premium rebates were consistent with CMS's demonstration authority. This position was being written into law in Section 533 of the BBA Refinement Act of 1999, which expressly authorized premium rebates recommended by CPAC. CPAC decided to revisit the rebate issue.

At this point, OGC presented the opinion that providing rebates up to the amount of the Part B premium would be legal under the demonstration. This decision removed any remaining opposition to rebates from CMS except with respect to low-income beneficiaries whose Part B premium is paid by Medicaid. (For this latter group, any rebates would have to be returned to the State Medicaid program rather than directly to the beneficiary.)

A subcommittee of CPAC considering the issue of Part B premium rebates in the demonstration concluded that this option would give efficient health plans another tool to make their product more attractive vis-a-vis their competitors and FFS Medicare. The subcommittee's recommendation was discussed and approved unanimously at the CPAC meeting in fall 1999, just before Congress blocked appropriations to implement the demonstration in Kansas City and Phoenix.

The Competitive Pricing Demonstration itself may never take place, due to political and other factors. Nevertheless, by late 1999, premium rebates had become a key issue for future discussions of Medicare reform, as a result of converging factors: the previous administration's reform proposal based in part on rebates, the congressional authorization for CPAC to include premium rebates in a demonstration, and CPAC's acceptance of the rebate option. We maintain that the later inclusion of rebates in the revised Breaux-Frist proposal has solidified the consensus in favor of Medicare premium rebates.

Rebates Become Official: Bipa 2000

Premium rebates received a definitive boost in late 2000, when Section 606 of BIPA authorized a premium rebate option. Starting in 2003, M+C plans will be allowed to elect a reduction in their basic payment, up to 125 percent of the Part B premium. Eighty percent of the payment reduction that a plan elects will be applied to reduce each enrollee's Part B premium. The government will share in the payment savings, by retaining 20 percent of the payment reduction. For a plan that elects a full 125 percent reduction in payment, enrollees will pay no Part B premium. Once it was decided that the savings should be shared in an 80/20 ratio between beneficiaries and taxpayers, the 125 percent cap became a mechanical formula to ensure that the maximum rebate would be equal to the Part B premium.

With this provision, premium rebates became not merely a consensus, but an authoritative plan for the future—the closest thing to M+C payment reform. But if the authority and schedule are clear, the details of implementation are not.

Implementation

Implementation issues for offering rebates include the administrative challenge of processing rebates, tax implications of receiving a rebate, and the question of who receives the rebate for low-income beneficiaries eligible for Medicaid coverage or assistance under the qualified Medicare beneficiary, specified low-income Medicare beneficiary, or qualified individuals programs. In addition to these technical issues of implementing premium rebates, there is the larger issue of likely political problems surrounding rebates.

Administrative Challenges

Currently, Part B premiums are deducted from beneficiaries' Social Security checks by the Social Security Administration (SSA). Some suggest that it would be more straightforward to pay premium rebates through SSA. Indeed, Section 606 of BIPA seems to envision exactly that kind of arrangement:

“In the case of an individual enrolled in a Medicare+Choice plan, the Secretary shall provide for necessary adjustments of the monthly beneficiary premium to reflect 80 percent of any reduction [in payment elected by a plan]…. To the extent to which the Secretary determines that such an adjustment is appropriate, with the concurrence of any agency responsible for the administration of such benefits, such premium adjustment may be provided directly, as an adjustment to any social security, railroad retirement, or civil service retirement benefits….”

Given the magnitude and complexity of SSA's administrative systems, incorporating a BIPA-directed mechanism to adjust Part B premiums for qualified beneficiaries might take years to implement. Meanwhile, the procedures would not be simple. A rebate process utilizing adjustments to Social Security payments would require efficient information exchanges between SSA and M+C plans. One complication in these exchanges is that beneficiaries may change plans in the middle of the year, requiring a change in their rebate adjustments and making the accounting a continuous process rather than an annual event. At the same time, enrollees' status under forms of public assistance varies over time, and with that variation would come changes in how SSA would need to adjust the premium rebate amount (for example, shifts in and out of Medicaid would require changes in the payee from the State to the beneficiary). While these obstacles may be surmountable in the long run, two simpler implementation options that may be consistent with section 606 are worth considering: CMS or the M+C plans themselves could process the rebates.

A scenario in which CMS sends the checks could be adapted from the demonstration of medical savings accounts. With relatively minor adjustments, CMS could modify this system to send checks directly to beneficiaries' bank accounts on a monthly, quarterly, or annual basis, with or without a lag in payments. Quarterly or annual rebates would be more cost effective and a lag in payments would allow for any reconciliation that might be needed if a beneficiary changes plans mid-year. This approach does require a complex administrative process, however. Beneficiaries would need to notify CMS of their designated bank accounts, and the agency would need to link the rebate amounts for each beneficiary to each change in enrollment.

A variation on this approach that CMS has explored is to adapt the system used by them to bill beneficiaries who are not eligible for Social Security benefits for their Part B premium. Because this billing system focuses on Part B enrollment only and receives all of its information from SSA, however, it could not be easily adapted to track whether a beneficiary has enrolled in a health plan that provides a rebate. Significant investment and time would be needed to make these substantial changes.

As a second alternative, health plans that elect a payment reduction could give beneficiaries the check or cash themselves, on a monthly or quarterly basis. This option has the advantage of linking the beneficiary's choice of plan directly with receipt of a rebate. It also has the advantage of relieving CMS or SSA from virtually all of the administrative burden and system changes associated with the first option. SSA would have no new role in the rebate system, and CMS would need only to process the reductions in payment rates that the plans elect.

If plans issue the checks themselves, the current regulatory structure appears to be adequate to address any concerns about ensuring that beneficiaries receive the correct rebate amount. Health plans already have the authority, and often have the systems in place, to bill enrollees for out-of-pocket premiums (for basic or supplementary benefit packages) or to provide refunds for any reversed appeals. CMS monitors the plans' marketing materials and conducts biennial compliance reviews that should detect any misrepresentation or failure to comply with the terms of the rebate. In any event, the terms of the rebate are virtually self-enforcing, once they are publicized—CMS is almost certain to learn if plans fail to pay, through ordinary publicity, if not through more formal grievance and appeals processes.

Plans might find this structure advantageous if it is attractive to beneficiaries and does not encourage adverse selection. The cost of developing or expanding these disbursement and accounting mechanisms would need to be offset by the success of this new marketing tool in order for rebates to appeal to health plans.

Tax Implications of Rebates

To our knowledge, the Internal Revenue Service (IRS) has not ruled officially on whether premium rebates constitute income to the recipient. The only IRS consultations of which we are aware were informal conversations between CMS, the IRS, and the Department of the Treasury in 1999, in conjunction with the deliberations of CPAC. These obviously are not binding on the agency, with respect to the BIPA rebate option.

However, all parties to the premium rebate discussions at each stage—from the CPAC discussions to the BIPA legislation—have made the assumption that a premium rebate limited to the value of the Part B premium was not income to the recipient, but rather a change in the net price of Medicare Part B. Rebates exceeding the Part B premium would be considered income and therefore have to be reported. Assuming the IRS does not take issue with these assumptions, rebates limited to the Part B premium would not affect tax liabilities for most beneficiaries. However, even on that simplifying assumption (which remains to be confirmed as premium rebates are implemented), there is a problem for those beneficiaries who have sufficient medical expenses to take an itemized deduction for them. These beneficiaries could deduct their full Part B premium payments one year, but receive a premium rebate in the next year. In this case, beneficiaries would have to consider the rebate as income in the next year.

There are basically two ways this potential problem could be resolved. First, the problem could be avoided if premium rebates were made in the same calendar year as the Part B premium they offset, because then the original deduction could reflect the Part B premium net of the rebate. Second, if the problem is not avoided, it could be treated in the same way itemized deductions are treated more generally. Taxpayers who itemize deductions have to report any changes in those deductions in a subsequent period due to rebates, refunds, and so on (e.g., when they claim State income taxes as a deduction on their Federal return for a given year, but receive a State tax refund for the same year). Since taxpayers are already accustomed to such transactions, it should not be a problem for them to handle a similar transaction with respect to premium rebates.

If these methods would take care of the problems of beneficiary computation of tax liabilities, there remains a question of whether premium rebates would create new reporting requirements for CMS, health plans, the IRS, or others. The best advice during the 1999 deliberations of CPAC was that no new reporting requirements would be created for a demonstration rebate that could not exceed the value of the Part B premium. But it is not clear that that would also be true for a premium rebate option available across the entire M+C program. This is another tax-related issue that remains to be resolved, as implementation of the premium rebate option approaches.

In view of these preliminary opinions, the tax implications of premium rebates appear to be straightforward. But that conclusion assumes that the IRS follows the assumption that rebate advocates have been making—that rebates limited to the Part B premium are not income—and it assumes that the IRS's answers to some remaining questions (e.g., concerning reporting requirements) do not create new complexities.

Impact on Low-Income Beneficiaries

Beneficiaries below 175 percent of the Federal poverty level are eligible for complete or partial assistance with the Part B premium. CMS's OGC determined in the course of CPAC deliberations that beneficiaries who receive Medicaid coverage would not be eligible to receive the rebate directly because Medicaid pays their Part B premium. Section 606 of BIPA reaches a similar conclusion:

“…such premium adjustment may be provided directly, as an adjustment to any social security, railroad retirement, or civil service retirement benefits, or, in the case of an individual who receives [Medicaid assistance] … for medicare costs … as an adjustment to the amount otherwise owed by the State for such medical assistance.”

For beneficiaries eligible for partial assistance with the Part B premium, considered qualified individuals, some partial rebate to the beneficiary could be considered. Sending the rebate check to the State Medicaid program would present additional administrative complexities, but these should not be insurmountable. CMS knows which Medicare beneficiaries are receiving Medicaid assistance and reports this information to health plans monthly, although retroactive adjustments to eligibility are quite common. Whatever the arrangement selected for crediting beneficiaries (i.e., SSA adjustment of Part B premiums, CMS issuance of checks, or health plan issuance of checks), a mechanism will have to be created to transmit rebate checks to the States. While sending the check to the States rather than the enrollee is likely to dilute the intended effect of the rebate on the dually eligible population, Section 606 of BIPA leaves no alternative, were one desirable.

Many beneficiaries who might be eligible for assistance with the Part B premium have not enrolled in the programs that offer assistance. These beneficiaries would be eligible for the full rebate. However, they (and others who are spending down to become eligible for public assistance) may be disadvantaged by receiving the rebate. For the purposes of determining eligibility for Medicaid, supplemental security assistance, food stamps, and other public aid programs, the rebate would be considered income. This could make it more difficult for beneficiaries to qualify for these programs, depending on how close they were to the eligibility thresholds. It is worth keeping in mind, however, that the rebate is true cash to the beneficiary and that its effect on eligibility for such programs is no different than other, ordinary cash income. The disincentive effects of any tax that these other programs place on the premium rebate is due to the design of these other programs. There is no way to design a premium rebate to avoid them.

Political Issues

The political issues of implementing rebates may turn out to be more important than the technical issues. While this discussion is necessarily speculative, history suggests that CMS would not be given the necessary resources or time to implement rebates effectively. Beneficiary education, outreach, etc., could be formidable problems. These problems might be addressed to some extent by a demonstration of rebates, as outlined in the following section.

Even if these problems can be solved, there is likely to be political opposition to the availability of unfair cash payments in high-payment areas. This is the M+C equity problem in a new guise, in this case because plans in high-payment areas can provide cash rebates, not just enhanced or free benefits, as before. Congress may be uncomfortable seeing beneficiaries in high-payment areas receiving checks. At the same time, however, other effects of BIPA—notably, its substantial increase in the minimum payment levels—might reduce these concerns. But it is at least worth asking whether the unique attraction of cash as an incentive to choose efficient plans might not be mirrored by the uniquely provocative spectacle of plans writing checks to some beneficiaries while others get nothing to offset their Part B premium each month.

Rebates and Medicare Reform

The major Medicare reform proposal before the 107th Congress has been the Medicare Prescription Drug and Modernization Act of 2001 (Senate 358), a revised version of the Breaux-Frist bill, first introduced in the previous Congress. This bill represents a modification of the unapproved draft proposal of the National Bipartisan Commission on the Future of Medicare. The latter included certain politically difficult provisions, e.g., folding FFS Medicare into the bidding process with health plans, in such a way that beneficiaries could end up paying an additional out-of-pocket premium to stay in FFS. The previous administration's appointees to the Bipartisan Commission were unable to support a recommendation that included FFS Medicare in the competitive fray. Considerable opposition to higher FFS premiums remains in Congress and among Medicare advocacy groups.

Senate 358 is a “more incremental alternative,” developed in an “effort to pass Medicare reform legislation that includes a prescription drug benefit,” (Breaux, 2001). It reflects a series of compromise moves in the direction of the administration's previous Medicare reform proposal, specifically:

• It excludes FFS Medicare from the bidding process.

• M+C plans are paid according to a benchmark amount based on the average price of traditional FFS Medicare in each area.

• Plans that bid below the benchmark amount can offer premium rebates, with 75 percent of the amount below the benchmark price going to beneficiaries and the balance going to the government.

The convergence of reform proposals may represent an historic opportunity for Medicare reform. Conflicts unresolved at the time of the Bipartisan Commission now appear to have been compromised to an important extent. Premium rebates are the device that permitted some level of agreement between these two points of view. This agreement was definitively expressed with the passage of BIPA, which put the Medicare program on a schedule to introduce the premium rebate option in 2003. While the Bush Administration has not, to our knowledge, made any specific commitments to premium rebates, it has apparently expressed some agreement with the Bipartisan Commission's approach to Medicare reform, with S. 358 as a starting point (Frist, 2001).

Despite plan withdrawals from the M+C program from 1998-2001 that were in part prompted by payment reductions in the BBA, M+C payments still vary widely among markets and allow M+C plans in many areas to provide generous benefit enhancements at no out-of-pocket cost to enrollees. Rebates may provide powerful incentives for beneficiaries to choose low-cost plans, and for plans to make their benefits more efficient. Or rebates may be a minor option, with little change anywhere in the system.

Given this uncertainty, it arguably would be prudent to demonstrate the use of premium rebates before instituting the option programwide. Such a demonstration would have at least two purposes (1) it would provide a pilot test of key features of the impending systemwide introduction of a premium rebate option, and (2) it would provide information on how plans and beneficiaries treat benefit enhancements versus premium rebates, in order to determine whether premium rebates are a minor option with little effect or a major change in incentives and behavior.

A third objective for the demonstration would be to test the effects of different tax rates on beneficiary behavior. A recent simulation (Thorpe and Atherly, 2001) suggests that the 25 percent tax imposed on rebates in Senate 358 is too large in relation to their supposed efficiency advantage over enhanced benefits, so that most beneficiaries would keep their current benefits rather than receive a rebate. Smaller tax rates would make rebates more attractive to beneficiaries but would reduce the government savings from a rebate program. Since BIPA and Breaux-Frist already envision different tax rates (20 versus 25 percent), and the Competitive Pricing Demonstration did not tax the rebates at all, it would be important to test this design feature of a rebate proposal.

A premium rebate demonstration could be fashioned relatively quickly, since the current M+C program would have to change in only one respect—to allow M+C plans the option to take a reduction in payments in order to give rebates to enrollees. The demonstration would have to address the implementation complexities previously noted but the basic steps are not difficult (1) choose (or solicit) sites; (2) announce that, after a given date, M+C plans in the selected sites could take payment reductions up to 125 percent of the Part B premium, with 80 percent of any amount chosen being rebated to enrollees; and (3) reach provisional understandings with the IRS and others, as needed, for the duration of the demonstration. There would be a few administrative details, but only a few. For a demonstration, at least, it would be essential to simplify the issuance of rebate checks (there would be little time, for example, to alter SSA or CMS systems), so the checks would have to be issued by the M+C plans themselves. CMS would need to institute special outreach and education activities, to inform beneficiaries about the demonstration. They would also need to work with the plans and CPAC to determine the rules for rebates (e.g., minimum length of enrollment, timing of rebates, etc.).

Whether or not a demonstration is used to reduce some of the uncertainties of the method, it now appears possible that the unheralded consensus that emerged in 2000 soon will create an interesting, and possibly very important, market reform of Medicare. Indications thus far are that premium rebates have passed key political tests that blocked other efforts at reform to date. Premium rebates leave current policy untouched, except to offer these incentives as price signals to beneficiaries. What remains to be seen is how M+C plans, beneficiaries, and other parts of the Medicare program react to a premium rebate option. Will premium rebates provide a powerful incentive, or a minor option? Will premium rebates bring about major changes in behavior? With or without a demonstration, that test will soon begin.

Technical Note

• Starting January 1, 1998, BBA set Medicare's monthly M+C capitation rates at the largest of a blend of area-specific and national rates, a minimum payment amount ($367 in 1998), and a minimum increase of 2 percent from the previous year's rate. To avoid tightening payments at a time of plan withdrawals and reactions against the BBA limits, the 1999 Balanced Budget Refinement Act provided for special payment increases for plans entering underserved, largely rural areas. In BIPA 2000, Congress again eased the BBA payment limits in three ways: (1) creating a minimum payment of $525 for urban areas and increasing the payment floor for other areas to $475; (2) raising the minimum increase to 3 percent in 2001; and (3) authorizing premium rebates beginning in 2003.

• BBA's lock-in provisions might simplify the enrollment process somewhat. The act provides that, beginning in 2003 (just as the premium rebate option takes effect), changes in enrollment will be allowed once during the first 3 months of each year, and during an annual coordinated election period. If this provision goes into effect, the volume of enrollment switches will decline substantially and to that extent will reduce the administrative problem of tracking enrollment changes for rebate adjustments.