Adjusted Community Rate Reforms to Promote HMO Participation in Medicare+Choice

Abstract

The authors review the financial regulations imposed on health maintenance organizations (HMOs) that participate in the Medicare+Choice program and identify elements of the regulations that may discourage HMO participation in the program. Modifications of the regulations are proposed that could encourage the participation of HMOs without affording them excessive profit. The modifications include smoothing and bounding profit estimates and authorizing and encouraging expanded use of benefit stabilization funds.

Introduction

Until recently, Medicare beneficiaries had been enrolling in risk-contract HMOs at record pace. Enrollment more than doubled—from 2.3 million to 5.2 million—between December of 1994 and December of 1997 (U.S. General Accounting Office, 1998). During 1997, average monthly enrollment of Medicare beneficiaries in risk HMOs increased by approximately 93,000. By September of 1998, however, this monthly enrollment growth had declined to just over 55,000. It declined further to approximately 38,000 in December (Health Care Financing Administration, 1999). By the beginning of January 1999, a total of 99 risk HMOs had discontinued operations in 400 counties in 31 States. Approximately 407,000 HMO enrollees (roughly 1 in 14) had to seek alternative health care providers, and 60,000 of these were forced to obtain health care services from fee-for-service (FFS) providers. Overall, 800,000 Medicare beneficiaries lost all access to risk HMOs in their counties (U.S. General Accounting Office, 1999).

Although HMO enrollment growth increased in the first half of 1999, it is not yet clear if this increased growth will be sustained. The 1999 Health Market Survey (Medicine and Health, 1999) found 70 of the 301 Medicare risk HMOs are no longer accepting new enrollees, and some major plans have already announced further withdrawals for 2000 (Falk, 1999).

These statistics are disturbing, particularly if continued expansion of the managed care sector is to help limit the growth of Medicare expenditures. (Medicare expenditures are projected to grow by 80 percent over the next 10 years, from $231 billion in 1998 to $416 billion in 2007 [Smith et al., 1998].) Therefore, it is important to identify the causes of the observed slowdown in enrollment growth in the Medicare+Choice program (the alternative to Medicare FFS) and the observed termination of Medicare operations by risk HMOs.

The observed terminations and reduced enrollment growth are likely caused by many factors. Some obvious factors include the changes introduced in the Balanced Budget Act (BBA) of 1997 (Public Law 105-33). The BBA placed new limits on increases in the capitated payments that Medicare makes to Medicare+Choice managed care organizations (MCOs) on behalf of Medicare enrollees. Historically, these payments have generally increased by more than 8 percent annually. But, largely as a result of the BBA reforms, 1999 payments were only 2 percent above their 1998 levels in approximately 85 percent of counties served by MCOs (U.S. General Accounting Office, 1999). The BBA also imposed new auditing and accounting requirements on MCOs (one-third of the HMO financial reports filed in conjunction with the Medicare+Choice program must be audited annually) and imposed user fees on MCOs that amounted to $95 million in both 1998 and 1999. (These fees help to offset the costs of informing Medicare beneficiaries about the Medicare+Choice plans available to them [Medicare Payment Advisory Commission, 1999].) By reducing the revenues and increasing the costs associated with participation in the Medicare+Choice program, the BBA has reduced the financial attraction of the program. This reduced attraction may be a primary determinant of the observed withdrawal from the Medicare+Choice program.

There are other factors that may have contributed to the observed withdrawal. For instance, MCOs were required to submit their proposed 1999 premiums and benefit packages before HCFA published its final regulations implementing Medicare+Choice for 1999. When these MCOs were denied permission to raise premiums or reduce benefits after seeing the final regulations, some withdrew from some of the areas they had been serving (Medicare Payment Advisory Commission, 1999). Significant increases in the prices of medical supplies and services, more intense industry competition, less favorable enrollee selection, and growing public sentiment against the rationing of medical services may also have diminished the profitability of serving Medicare beneficiaries (Medicare Payment Advisory Commission, 1999).

Some MCOs withdrew from Medicare+Choice because they had difficulty attracting either the minimum number of enrollees required for profitable operation or the physicians required to operate a successful network (U.S. General Accounting Office, 1999; National Health Policy Forum, 1999). In some cases, an MCO withdrawal resulted from a merger between MCOs (National Health Policy Forum, 1999). Some MCOs may have decided not to be serving Medicare beneficiaries at the start of the Medicare+Choice program in order to avoid the 5-year lockout penalty for withdrawal from the program once it began, as specified in the BBA (U.S. General Accounting Office, 1999). Particularly in counties where capitated payments from Medicare are relatively low, some MCOs may have terminated operations in response to the new requirement in the BBA that MCOs deliver uniform benefits at a uniform price to all Medicare enrollees in their service areas. The Medicare Payment Advisory Commission (MedPAC) (1999) recommended that MCO flexibility to offer different benefit packages in different counties be restored in order to limit withdrawal from counties with low capitated payments from Medicare. HCFA provided some flexibility in 1999 as a transitional measure.

As MedPAC (1999) has indicated, “It is too soon to tell” exactly what has caused the recent withdrawal of MCOs from Medicare+Choice. Additional study of the impact of all of the aforementioned factors is warranted. This research, though, focuses on another possible cause of the MCO withdrawal that has been afforded relatively little attention to date: the methodology employed by Medicare to regulate the earnings of MCOs. This methodology is called the adjusted community rate (ACR) mechanism.

We focus on two features of the ACR mechanism that may create undesired and unnecessary incentives for MCOs to limit or terminate their participation in the Medicare+Choice program. The first feature is an asymmetric treatment of profits and losses. Although the ACR mechanism does not permit what is considered to be more than a reasonable level of profit, it does allow financial losses. Consequently normal variation in annual revenues and costs can impose financial hardship on MCOs. The second feature of the ACR mechanism is its reliance on profit from commercial operations as the basis for calculating a reasonable level of profit from Medicare operations. Such reliance can render Medicare operations unduly unattractive when commercial markets are characterized by excess capacity ongoing merger and acquisition activity and unusually meager rates of return.

Before proceeding, we note that most of our proposed modifications of the ACR mechanism would only promote MCO participation in Medicare+Choice if the ACR mechanism effectively constrained the profit that MCOs earn in their Medicare operations. Competition, not the ACR mechanism, is the key constraint on MCO profit in a significant number of regions (e.g., Chicago, New York, and Philadelphia), where MCOs choose to deliver substantially more benefits to Medicare enrollees than the ACR mechanism requires them to deliver (Physician Payment Review Commission, 1997). Furthermore, some experts believe that even in the many regions where competition among MCOs is limited, the ACR mechanism may not constrain profit effectively because it provides MCOs with considerable leeway to understate their net returns from Medicare operations (in ways that we identify later) (Logistics Management Institute, 1997). Such understatement is a relevant concern. However, the increased auditing standards imposed by the BBA and HCFA's requirement that MCO executives formally attest to the accuracy of their ACR proposals seem likely to render the ACR mechanism a real constraint on the profit that many MCOs earn from Medicare operations in the future, even if it is not a binding constraint today (Medicare Payment Advisory Commission, 1999).

Description of the ACR Mechanism

The ACR mechanism is intended to ensure that MCOs deliver to Medicare enrollees services that are commensurate with the payments that Medicare makes to the MCOs. These payments are not determined as part of the ACR process. Historically, these payments have been set equal to 95 percent of the estimated cost of providing the basic benefits that Medicare provides for beneficiaries who are treated by FFS providers (although one estimate suggests that the BBA payment reforms may lead to capitated payments that average 89 percent of FFS costs by 2003 [Rodgers, 1998]). Payments vary with a beneficiary's age, sex, institutional and Medicaid status, and county of residence. In 2000, payments will also begin to vary to a limited extent with a patient's health status.

To ensure service levels that are commensurate with payments from Medicare, the ACR mechanism estimates an MCO's costs of delivering basic Medicare services. If these costs equal or exceed the stipulated payment from Medicare, the MCO is not obligated to provide anything more than the basic services that are provided to Medicare beneficiaries who receive care in the FFS sector. If payments exceed the estimated costs of delivering these basic Medicare services, however, the MCO must eliminate the excess of payments over estimated costs by providing additional benefits to Medicare enrollees. The value of these additional benefits (which often include free prescription drugs and/or eye care) averaged $87.62 per enrollee per month in 1997 (Office of Inspector General, 1998). An MCO must deliver at least 85 percent of the excess of payments over estimated costs to Medicare enrollees each year in the form of additional benefits or reduced cost-sharing. As much as 15 percent of the excess can be placed in a benefit-stabilization fund. This fund acts like a savings account in that the MCO can withdraw monies from the fund in future years when payments from Medicare fall short of the MCO's estimated costs of serving Medicare enrollees. An MCO may be permitted to contribute more than 15 percent of the excess to a benefit-stabilization fund if it “can demonstrate … that the value of the additional benefits it provides to its Medicare enrollees … fluctuates substantially in excess of 15 percent from one contract period to another” (United States Code of Federal Regulation, 1998). Historically, though, MCOs have seldom made use of the benefit-stabilization fund, choosing instead to deliver any excess of payment over estimated cost to beneficiaries in the form of increased benefits or reduced cost-sharing (Logistics Management Institute, 1997; Medicare Payment Advisory Commission, 1998).

The calculation of an MCO's estimated costs of serving Medicare enrollees is a critical component of the ACR process. For an MCO that serves both commercial and Medicare enrollees, its estimated Medicare costs are based on its corresponding commercial costs, after adjusting for the fact that Medicare enrollees are typically more costly to serve than are non-Medicare enrollees, whom we refer to, for simplicity, as commercial enrollees. The methodology employed to estimate Medicare costs (as of January 1, 2000) involves the following six-step procedure.

First, an MCO's costs of serving Medicare and commercial enrollees in the most recent year (for example, year t) are divided into two components: (1) the direct costs of providing medical care; and (2) administration costs. Direct costs include the costs of inpatient hospital services, physician services, skilled nursing services, and outpatient laboratory services. Administration costs include such overhead costs as occupancy costs, general administrative, auditing, and support costs, reinsurance costs, and sales and marketing expenses. We denote an MCO's direct costs (per enrollee) of serving Medicare and commercial enrollees, respectively, by ${\mathit{\text{D}}}_t^M$ and ${\mathit{\text{D}}}_t^C$ in year t. We denote the corresponding administration costs (per enrollee) of serving Medicare and commercial enrollees, respectively, by ${\mathit{\text{A}}}_t^M$ and ${\mathit{\text{A}}}_t^C$.

Second, the profit that the MCO earned in its Medicare and commercial operations in the most recent year (t) is recorded. Profit, otherwise known as additional revenue, is the difference between revenue and cost (per enrollee). We denote by ${\Pi }_t^M$ and ${\Pi }_t^C$ the profit that the MCO earns in year t from serving Medicare and commercial enrollees, respectively. We also denote by ${\mathit{\text{R}}}_t^M$ and ${\mathit{\text{R}}}_t^C$ the average compensation (per enrollee) that the MCO receives in year t for serving Medicare and commercial enrollees, respectively. Because profit is the difference between revenue and cost, we have:

$${\mathrm{\Pi }}_t^M={\mathit{\text{R}}}_t^M-{\mathit{\text{D}}}_t^M-{\mathit{\text{A}}}_t^M,\text{and}$$ (1)

$${\mathrm{\Pi }}_t^C={\mathit{\text{R}}}_t^C-{\mathit{\text{D}}}_t^C-{\mathit{\text{A}}}_t^C.$$ (2)

Third, these cost and profit statistics are employed to calculate relative cost factors (r), one for each of direct cost, administration cost, and profit. Formally:

$${\mathit{\text{r}}}_t^D=\frac{{\mathit{\text{D}}}_t^M}{{\mathit{\text{D}}}_t^C};{\mathit{\text{r}}}_t^A=\frac{{\mathit{\text{A}}}_t^M}{{\mathit{\text{A}}}_t^C},\text{and}{r}_t^{\mathrm{\Pi }}=\frac{{\mathrm{\Pi }}_t^M}{{\mathrm{\Pi }}_t^C}.$$ (3)

These relative cost factors are designed to account for systematic differences between the costs of serving Medicare and commercial enrollees.

Fourth, the MCO's revenues, direct costs, and administration costs from commercial operations are estimated for the coming year (year t + 1). These estimates take into account planned changes in premiums and anticipated changes in operating costs, including those due to inflation. We denote these estimates by ${\stackrel{\sim }{R}}_{t+1}^C$, ${\stackrel{\sim }{D}}_{t+1}^C$, and ${\stackrel{\sim }{A}}_{t+1}^C$, respectively. These statistics enable an estimate of profit (per enrollee) from commercial operations:

$${\stackrel{\sim }{\mathrm{\Pi }}}_{t+1}^C={\stackrel{\sim }{R}}_{t+1}^C-{\stackrel{\sim }{D}}_{t+1}^C-{\stackrel{\sim }{A}}_{t+1}^C.$$ (4)

Fifth, the ACR for the coming year t+1 (denoted ACR_t+1) is calculated. This statistic is the sum of projected direct costs, administration costs, and profit in the commercial sector, scaled by the relative cost factors. Formally:

$${\text{ACR}}_{t+\text{1}}={\mathit{\text{r}}}_t^D{\stackrel{\sim }{D}}_{t+1}^C+{\mathit{\text{r}}}_t^A{\stackrel{\sim }{A}}_{t+1}^C+r_t^{\mathrm{\Pi }}{\stackrel{\sim }{\mathrm{\Pi }}}_{t+1}^C$$ (5)

Sixth, the ACR is adjusted to account for expected variation, which is the change in projected cost due to factors that are not captured in the relative cost ratios. Such factors include changes in the level or mix of mandated Medicare benefits. Correcting the ACR statistic for expected variation provides what is called the adjusted ACR. The adjusted ACR is essentially an estimate of the MCO's cost (per enrollee) of serving Medicare beneficiaries in the coming year.

An MCO's adjusted ACR is compared with the average payment it will receive from Medicare to care for Medicare enrollees in the coming year. If the average payment exceeds the adjusted ACR, the MCO must increase benefits to Medicare enrollees so as to eliminate the excess. (Recall that the MCO can place as much as 15 percent of the excess in a benefit-stabilization account, but MCOs have generally chosen not to make use of the benefit-stabilization fund.)

Before proceeding to discuss potential drawbacks to the ACR mechanism that would be present even if it provided a flawless estimate of the net financial return from serving Medicare enrollees, we note that it does not, for at least four reasons. First, the distinction between direct costs and administration costs is not always clear. Second, the proper allocation of administration costs between Medicare and commercial operations can be controversial. Third, MCOs typically provide different sets of services to Medicare and commercial patients, so a comparison of the corresponding cost estimates can be problematic. Fourth, forecasts of future commercial revenues and costs do not always match actual revenues and costs.

Potential Drawbacks of the ACR Mechanism

Equation (5) reflects a key element of the ACR mechanism that warrants emphasis. The mechanism bases its estimate of reasonable costs of serving Medicare enrollees on the corresponding costs of serving commercial enrollees (after correcting for systematic differences in the costs of serving the two populations). In doing so, the ACR mechanism attempts to replicate for Medicare operations the market forces that discipline commercial operations. The basic idea is the following. Suppose competitive forces limit profit to reasonable levels and compel MCOs to minimize operating costs in commercial markets. Then, by equating estimated cost and profit in the Medicare sector with (suitably adjusted) commercial cost and profit, Medicare can be confident that it is not compensating MCOs for excessive cost or profit in their Medicare operations.

Despite the potential merits of this approach, it is not without its drawbacks. Although the approach can import the discipline of competitive commercial markets, it can also import the vagaries of these markets. There are two such vagaries for which the ACR mechanism may not account adequately: intertemporal profit variation and market disequilibrium.

Asymmetric Treatment of Profits and Losses

Firms that operate in competitive markets earn only normal profit—the minimum profit required to induce them to employ their assets in the competitive market. If a firm is earning more than normal profit in a competitive industry, other firms will expand their operations in the industry in order to garner a share of this profit. The lower prices and/or higher quality that competitors offer to attract customers eventually serve to eliminate all but normal profit in the industry. Although they earn normal profit on average, firms in competitive markets may earn more or less than this amount in any given year, but the positive and negative variations balance out over time. The ACR mechanism does not permit a corresponding balance of natural variations in profit.

In years when costs exceed revenues from Medicare operations, MCOs suffer financial losses. In years when revenues exceed costs, MCOs must deliver extra benefits to Medicare enrollees to eliminate the excess. Consequently, this asymmetry in the ACR mechanism translates natural intertemporal variation in revenue and cost into net losses for MCOs, which limits the financial attraction of serving Medicare beneficiaries. (Conceivably, an MCO could reduce extra enrollee benefits in years when Medicare costs exceed Medicare revenues. MCOs are often reluctant to do so unilaterally, however, for fear that enrollees will seek alternative suppliers if cherished benefits to which enrollees have become accustomed are eliminated.)

In theory, the benefit-stabilization fund might be employed to mitigate this undesirable feature of the ACR mechanism. If an MCO could contribute the excess of Medicare revenues over costs to a fund in years when the excess is positive and withdraw monies from the fund to offset any excess of costs over revenues in subsequent years, a symmetric treatment of profits and losses could be fashioned. But the benefit-stabilization fund does not serve this role under the ACR mechanism for two reasons. First, monies deposited in the fund cannot be “withdrawn … to refinance losses suffered during … [a] previous contract period” (United States Code of Federal Regulation, 1998). Thus, the benefit-stabilization fund can be employed to help offset increasing costs of current benefits and/or the costs of new benefits, but it cannot be used to offset historic financial losses. Second, contributions to a stabilization fund are normally capped at 15 percent of the excess of Medicare revenues over projected costs. Therefore, even if monies in the fund could be employed to offset financial losses, the extent of the offset would be limited, and so the negative impact of natural intertemporal profit variation on an MCO's financial return from serving Medicare enrollees would persist.

Current regulations also render use of the benefit stabilization fund relatively unattractive. Monies deposited in these funds earn no interest (United States Code of Federal Regulation, 1998), thus, the real value of these deposits is eroded by inflation. An MCO may also be reluctant to contribute to its benefit-stabilization fund unilaterally. If an MCO contributes to its stabilization fund while competitors offer additional benefits to Medicare enrollees, the MCO may suffer disenrollment, as former enrollees switch to MCOs that offer more pronounced immediate benefits.

In summary, the ACR mechanism can cause MCOs to suffer financially in their Medicare operations as profits in the commercial sector vary naturally around their equilibrium (long-term) level. We now explain why the ACR mechanism may impose even more severe financial hardship on MCOs if commercial profits remain below equilibrium levels for an extended period of time.

Inadequate Correction for Commercial Market Disequilibrium

As previously noted, the ACR mechanism employs MCO performance in commercial markets as an indicator of reasonable performance in the Medicare sector. Although this approach can be appropriate when commercial markets are in competitive equilibrium, it can introduce undesired distortions and limit participation in Medicare markets when commercial markets are not in equilibrium. To see why, consider the common setting where many firms enter a new market, hoping to earn substantial profit. If the number of entrants exceeds the number of firms that can operate profitably in the long run, a consolidation and/or exit of firms will follow. Prior to and during this period of consolidation and exit, intense competition among the excessive number of competitors will often force the firms to suffer financial losses.

Now consider the effect of the ACR mechanism when commercial health care markets are characterized by considerable exit and negative profit. As equation (5) reveals, the negative profit forecast in commercial operations $({\stackrel{\sim }{\mathrm{\Pi }}}_{t+1}^C)$ becomes the standard for reasonable profit in Medicare operations. Consequently, MCOs may find the allowed returns from Medicare operations to be too meager to justify serving Medicare beneficiaries, even though they could serve these beneficiaries at substantially lower cost than can the FFS sector. In this manner, the ACR mechanism can limit MCO Medicare operations unduly at times of consolidation and exit in commercial health care markets.

It is important to note in this regard that HMOs have suffered financial losses in their commercial operations in recent years. Following aggregate profit of $1.8 billion in 1995, industry profit declined to $0.7 billion in 1996. In 1997, the industry suffered aggregate financial losses of nearly $0.8 billion. Fifty-seven percent of HMOs reported negative profit in 1997 (Managed Care Outlook, 1998). To the extent that the ACR mechanism employs these losses as the benchmark for reasonable profit in the Medicare sector and to the extent that these losses reflect more than temporary losses due to standard insurance underwriting profit cycles (Ginsberg and Gabel, 1998), the mechanism may offer unduly meager rewards for serving Medicare beneficiaries.

Possible Modifications of the ACR Mechanism

Limiting Asymmetric Treatment of Profits and Losses

Consider first how the ACR mechanism might be modified to mitigate the asymmetric treatment of unavoidable intertemporal variation in commercial profit. Six possible modifications are suggested.

First, the ACR mechanism might employ a weighted average of past and projected profit instead of only the latter. In particular, the projected commercial profit term $({\stackrel{\sim }{\mathrm{\Pi }}}_{t+1}^C)$ in equation (5) might be replaced with a weighted average of projected profit and realized commercial profit in recent years. Doing so would tend to reduce transient variation in the statistic employed to measure reasonable Medicare profit and thereby avoid an unintended mandated increase in benefits to Medicare enrollees. The ideal weights to place on past and projected profit would likely vary with perceived ability to forecast significant, predictable change in the health care industry.

A second way to reduce such transient variation in allowed Medicare profit is to employ industry profit as well as the profit of the individual MCO when calculating an MCO's ACR. If commercial health care markets are truly competitive, then industry profit will provide a good proxy for a reasonable level of profit for an individual competitor. Because average industry profit smooths out some of the variation in the profit of individual industry members, a weighted average of industry and own profit in commercial markets can provide a more stable measure of reasonable Medicare profit for an MCO than will own profit alone.

Determining the best measure of industry profit will require careful thought. Because many suppliers of health care services are not-for-profit organizations, they may not produce accurate profit statistics. Further, even for-profit health care organizations do not all employ the same accounting conventions, and some health care suppliers may enjoy particularly large profits or suffer unusually meager profits due to circumstances that are irrelevant for MCOs participating in the Medicare+Choice program. Because of these concerns and others, it may be advisable to define industry profit fairly narrowly at first (e.g., base it only on the earnings of for-profit MCOs in the Medicare+Choice program) and place a relatively small weight on industry profit until more experience and comfort is gained with the statistic.

Third, an upper bound on acceptable deviations between Medicare revenues and costs might be established. As long as its Medicare revenues did not exceed its estimated Medicare costs by more than this upper bound, an MCO would not be required to deliver additional benefits to Medicare enrollees. By setting this upper bound at a moderate but strictly positive level, the MCO would be permitted to retain modest transient profit, just as it is typically compelled to absorb modest, transient losses.

An appropriate upper bound might be one that is designed to ensure the MCO somewhat more than (e.g., 110 percent of) a fair rate of return on its investments, which is the return that is just sufficient to attract additional capital to the organization and that is commensurate with the organization's operating risk (Phillips, 1993). A fair return is designed to ensure only normal profit for an efficient organization. Although the determination of a fair rate of return is not an exact science, it has been practiced for many decades in a variety of regulated utility industries. Observed returns in the commercial health care industry would likely help to inform an estimate of a fair return on MCO Medicare operations, but they would not serve as the sole proxy for this fair return, as they do under the ACR mechanism.

Fourth, expanded use of the benefit-stabilization fund might be authorized. If MCOs were permitted to: (1) employ contributions to the fund to offset financial losses; and (2) contribute to the fund the entire excess of Medicare revenues over costs instead of only 15 percent of the excess, then natural variation in revenues and costs would not necessarily translate into lower financial returns for MCOs from their Medicare operations.

Fifth, monies deposited in benefit-stabilization funds might be afforded reasonable rates of return (e.g., at least the prevailing rate on U.S. Treasury bonds). Reasonable returns would encourage expanded use of the funds and thereby avoid unintended ratcheting up of the benefits delivered to Medicare beneficiaries.

Sixth, contributions to benefit-stabilization funds might be mandated when Medicare revenues exceed estimated Medicare costs by a moderate amount (e.g., by 10 or 20 percent of costs). When all MCOs are required to contribute to stabilization funds in this manner, the incentive that an individual MCO might otherwise have to unilaterally expand benefits to Medicare enrollees in order to increase its market share at the expense of competitors is reduced. Consequently, mandated use of benefit-stabilization funds might help to avoid any excessive competition among MCOs that may exist in Medicare markets.

Correcting for Commercial Market Disequilibrium

Now consider how the ACR mechanism might be modified to account for settings where the commercial health care industry is undergoing consolidation and exit. Recall that in such settings, realized earnings are often below competitive levels and so may not constitute the best estimate of reasonable profit from serving Medicare enrollees. Consequently, it may be better to employ an alternative estimate of reasonable profit in the ACR calculation. This alternative estimate might be set by specifying a lower bound on reasonable Medicare profit (e.g., 80 percent of the established fair rate of return). If projected profit in the commercial sector ( ${\stackrel{\sim }{\Pi }}_{t+1}^C$ in equation (5)) falls below the specified lower bound, the lower bound would replace projected commercial profit in the ACR calculation. Use of such a lower bound would limit the tendency for commercial market disequilibrium to reduce allowed returns in the Medicare market unduly. Consequently, it could encourage expanded participation in Medicare markets.

Conceivably, an upper bound on reasonable profit (e.g., one corresponding to 120 percent of the estimated fair rate of return) might also be imposed. Such a bound could serve two useful purposes. First, it could guard against allowing excessive profit in Medicare markets in settings where corresponding commercial markets are not competitive. Second, it could limit the incentives an MCO might otherwise have to drive competitors from the market and then enjoy excessive monopoly profit.

Conclusions

We have shown how the ACR mechanism is designed to harness for Medicare markets the discipline of competitive commercial health care markets. We have also pointed out why, in doing so, the ACR mechanism may import and accentuate undesired vagaries of competitive markets and thereby discourage MCOs from participating in the Medicare+Choice program. We have suggested how the ACR mechanism might be modified to mitigate undesired and unintended consequences. Suggested modifications include smoothing profit estimates, bounding these estimates, using industry data to calculate reasonable profit levels, authorizing expanded use of benefit stabilization funds, enhancing the appeal of these funds, and perhaps mandating their use under some circumstances.

These modifications of the ACR mechanism are designed to address the immediate concern of reduced participation of health care suppliers in the Medicare+Choice program. (Of course, before taking actions to limit the withdrawal of MCOs from the Medicare+Choice program, it is important to determine whether such withdrawal is undesirable. If, for example, MCOs terminate operations in regions where they are not the least-cost suppliers of health care services [but where operations were profitable because payment rates were particularly high], then such termination may be desirable). More generally, additional modifications of the ACR mechanism might be contemplated as part of a systematic evaluation of the financial regulation of MCOs. For example, uniform guidelines for projecting revenues and costs might be imposed to ensure consistency across MCOs. Uniform cost allocation procedures might also be mandated. (Refer to Glass and Sappington [1999] for an analysis of the types of cost-shifting that the ACR mechanism may invite.) In addition, realized revenues and costs might be compared with projected revenues and costs. Some reconciliation might be required in cases where the divergence between predicted and realized statistics is pronounced. Furthermore, competitive bidding for the right to serve Medicare beneficiaries might be considered as an alternative to attempts to match Medicare payments and costs. A careful assessment of these and other possible modifications of the financial regulations that are imposed on MCOs awaits further study.