Using the Federal Employees' Model: Nine Tests for RationalMedicare Reform

Outdated Political Centralization

Today, the Medicare program is overwhelmingly statist. Medicare uses political fiat, price controls, and centralized bureaucratic processes to try to regulate an infinitely complicated health care program. Many Medicare rules--such as record keeping, ownership rules, and quality standards--directly regulate almost all health care providers.

Most regulatory decisions made by Medicare are compromises that are wrong--often deeply wrong--for many enrollees and providers. Medicare is like a government-designed automobile: designed by committee, changed too late, with final details set by legislative or bureaucratic fiat, and based on the principle that "one size fits all" and the corollary ethical proposition that everyone should get an identical benefit because anything else is "unfair." Like the government-designed automobile, Medicare fits very few as well as the plan they would choose for themselves if given a choice.

Medicare's failure to provide coverage abroad is one example. This issue, while less important than prescription drug coverage and many others, is important to those who wish to retire abroad without giving up their health insurance or travel abroad without buying an exorbitantly expensive Medigap supplement. But a program run on the bureaucratic regulatory model necessarily fails to deal optimally with many problems like this, both large and small.

Indeed, a major political impediment to a Medicare drug benefit has been the program's use of draconian price controls, the level of which has not been seen in most of the American economy since World War II. Price controls are anathema not only to the pharmaceutical industry, but also to all Americans who expect the cures for Alzheimer's disease--and many other diseases--that are likely to come only from a research industry able to invest in creating life-saving medicines.

A Modern Progressive System

In contrast, the Federal Employees Health Benefits Program (FEHBP) uses the mildest forms of government direction and oversight to allow the forces of choice and competition to determine health plan costs, benefits, provider choice, administrative convenience, and a host of details. In the Medicare Prescription Drug and Modernization Act of 2003 (H.R. 1), recently enacted by the House of Representatives, Subtitle C of Title II provides for a transition to an FEHBP-style competitive system, beginning in 2010. Thus, authors of the House bill, in contrast to the Senate sponsors, make the accomplishment of an FEHBP-style reform an explicit objective of the Medicare legislation.¹

Not every detail of the FEHBP should be incorporated into a reformed Medicare program. However, Congress can test the effectiveness of its reform by ensuring that the plan:

• Ensures that government functions as a good business partner with health plans;
• Establishes reasonable and predictable financing;
• Allows health plans flexibility in benefit design;
• Allows flexibility in service areas;
• Exempts health plans from state mandates;
• Establishes a sensible budget strategy;
• Authorizes the program in brief and simple statutory language;
• Encourages broad participation by employer plans; and
• Allows FEHBP plans to participate.

Performance: Medicare Versus FEHBP

Careful analysis shows that the FEHBP clearly outperforms Medicare in three vital areas: (1) providing access to physicians, health plans, and rural health coverage; (2) providing innovative benefits and satisfying consumers; and (3) controlling costs. These points have been discussed at length in a companion paper.² However, for Medicare reform, it is particularly important to understand that private plans can and do outperform Medicare in controlling costs despite paying providers more than the rock-bottom Medicare rates.

Cost Control
The FEHBP and Medicare programs have virtually identical records from 1975 to 2003, with Medicare showing a cumulative advantage of only 1 percent over 28 years.³ However, this analysis does not account for the substantial benefit improvements in the FEHBP. After accounting for benefit improvements, the FEHBP clearly outperforms Medicare in cost control.

It should not be surprising that the records are broadly similar since both programs operate in the context of the American health care system, with the same underlying structure of hospitals, doctors, costs, technological changes, and a myriad of other commonalities.

However, viewed another way, it is a surprise. The Medicare Administrator operates a system of price controls. As Congress has so amply demonstrated in its recent flip-flop attempts to set physician, hospital, and Medicare+Choice (M+C) reimbursements at the "right" levels (determined in large part by the decibel level of the political outcry), price controls can be set arbitrarily within a fairly broad range. Thus, Medicare could outperform the FEHBP in reducing premium costs through cutbacks in provider prices and income, benefit reductions, and other government-mandated reductions. (Health care resources, both human and bricks and mortar, are not perfectly mobile in the short run.) Ultimately, the Medicare budget is set by what the political system tolerates, not by the market or any objective method.

A recent analysis published by The Heritage Foundation uses the National Health Account data together with data from the National Medical Care Expenditure Survey and other sources. It demonstrates that when cost increases are adjusted for benefit improvements, the private sector has outperformed Medicare over the past 30 years.⁴

In other words, whether looking at private spending in general or the FEHBP in particular, benefit-adjusted private-sector costs have increased less than Medicare costs over most or all of the life of the Medicare program. In the case of the FEHBP, its cost growth is so superior that it ties or slightly outperforms Medicare even without adjusting for benefit improvement over time.

Promoting Innovation
This cost-control performance has come despite (or because of) higher administrative costs for the FEHBP, paying physicians and other providers more than Medicare,⁵ and the near absence of direct managerial controls. One reason, of course, is that Medicare lurches from one crisis to another as both consumers and providers find ways to game the system. In the FEHBP, plans are always looking for ways to control unnecessary spending, relying on a wide range of techniques. The Office of Personnel Management (OPM) can urge plans to adopt useful innovations by simple requests, unencumbered by the Federal Register process used by the Centers for Medicare and Medicaid Services (CMS), which on average requires years from inception to final publication of binding rules.

For example, it took years of regulatory indecision, and ultimately an act of Congress, to stop Medicare from paying for unnecessarily expensive seat-lift chairs, once routinely prescribed by doctors for patients who saw beautiful and expensive lounge chairs advertised on television as covered by Medicare. In the FEHBP, the OPM was not involved, and plans simply agreed to pay for only the most austere models of seat lifts, relying on "reasonableness" clauses in their policies.

Care Management
In addition, FEHBP plans use care management, benefit design, and other tools that enable them to keep constant pressure on costs. For example, the FEHBP plans increasingly seek to assure that enrollees take the proper medications and thereby obviate the need for costly hospitalizations. These kinds of techniques can save billions of dollars. Medicare has virtually no existing tools that use disease management techniques to reduce costs, but it does have a tradition of "penny wise and pound foolish" restrictions on administrative costs that virtually preclude use of such tools on any scale larger than small demonstrations.⁶

The FEHBP Record Compared to Medicare+Choice

Some Members of Congress claim that consumer choice has been tried but has failed because Medicare+Choice had a rocky start and failed to reduce overall Medicare costs. However, M+C never had a chance to perform properly because it was tied to the yo-yo of annual changes in Medicare spending levels and used the Medicare reimbursement formula, which relied on the fundamentally flawed Adjusted Average Per Capita Costs estimates of geographic variability in health costs.⁷

An important study published in Health Affairs demonstrated that geographic variations in managed health care costs are minimal, rarely exceeding a 10 percent variation above or below the national average anywhere in the country.⁸8 A well-designed defined contribution program, using rolling averages or all-plan averages and minimal geographic adjustments (if any), would have functioned far better.

In addition, a set of draconian and unreasonable mandates made participation expensive and burdensome for any fee-for-service (FFS) or preferred provider organization (PPO) plan, and for most health maintenance organizations (HMOs). One regulatory mandate for language interpreter services paid by each plan is arguably illegal in at least three different ways.⁹ Incredibly, despite these problems, M+C still manages to attract about 150 plans and almost 5 million enrollees, about one in eight Medicare clients.

Regulatory mandates, unreliable funding levels, constant change, unrealistic government expectations, and other rocky issues have led to what may well be M+C's most fundamental problem. Health plans do not regard Medicare (both CMS and Congress) as a good business partner. Even the promise of substantial additional business has proven a weak incentive in the face of the underlying distrust--distrust based on the bureaucracy's well-known track record.

Nine Tests for Rational Medicare Reform

On each of the major dimensions of performance against which fundamentally different approaches to health insurance programs can be compared, the FEHBP is arguably at least equal to--and usually superior to--Medicare.

However, this does not lead to any simple conclusion on the best way to reform Medicare. The issues are many and complicated, and the FEHBP certainly has many important problems. Several of these are embodied in the Medicare statute. A senseless and costly restriction in Medicare law prohibits FEHBP plans (but no other employer plans) from paying Part B premiums. This is costly to both Medicare and the FEHBP because it forces plans to offer unusually high benefit wraparounds rather than lower premiums, creating incentives for unconstrained use of health care benefits.

A second problem is the needless penalty of 10 percent per year imposed on late enrollment in Medicare Part B. This penalty is imposed even if the enrollee is covered by comprehensive insurance and the possibility of adverse selection is remote. Lifting this restriction for those covered by comprehensive plans would induce more elderly to remain in employer-sponsored retirement plans, thereby directly reducing Medicare costs.¹⁰

Obviously, the FEHBP model cannot and should not be adopted in every detail, or even in every major feature, in designing Medicare reform. Designing a viable Medicare program modeled on the FEHBP will require many carefully analyzed decisions. However, certain pitfalls and solutions are obvious. The following nine tests provide a scorecard against which to judge legislation on Medicare reform.

Test 1: Ensure that the government is a good business partner
The rules of the game should be few, robust, and rarely changed. In sharp contrast, the next wrenching Medicare reversal is rarely farther away than the next session of Congress. To succeed, Medicare reform must provide for reasonable assurance of stable and growing payment rates; stable plan participation (no risk of being evicted from the program if a plan's premiums go up just slightly "too much" next year); and freedom from costly mandates, nuisance regulations, and other significant burdens of current Medicare practice.

Test 2: Establish a reasonable and predictable level of financing
Medicare+Choice has been seriously hampered by its reliance on the annual level of Medicare spending. Reliance on competitive bidding may introduce equally arbitrary results, with plan participation varying unpredictably depending on the annual bidding decisions of other plans.

But a better approach can be devised, such as basing the government payment level on an enrollment-weighted rolling average of costs in either a new Medicare Advantage program or traditional Medicare. Ideally, such a payment would be level, or nearly so, throughout the country.

Stability would be enhanced by using a rolling average (past years adjusted upward to account for recent cost increases) to give plans more certainty as to future year premium contributions. Most important, stability would be greatly enhanced if plans could participate regardless of how they set their bid offer. Thus, a particularly bad year might cost the plan some enrollment but would not force it out of the program.

This kind of stability is important not only to plans, which could otherwise be reluctant to invest in a program from which they might be thrown out, but also to enrollees, who would benefit because they could count on continued participation by a plan they like and would not be forced to change physicians, which might happen if their plans were changed annually through a bidding process. This stability also encourages plans to emphasize cost-reducing care management, since they will reap the savings of preventing future hospitalizations.

Test 3: Allow health plans to decide benefit and coverage details
Many otherwise astute students of reform have suggested that competing plans should have identical benefits, specified in detail by the government.¹¹ While this may be an attractive idea in terms of simplifying decisions for enrollees, it would be a fatal mistake. It would transform otherwise private decisions on benefit details into government decisions on the uniform benefit structure, just as under traditional Medicare. Because those government decisions would be made through bureaucratic processes and often on political grounds, rather than through evolving consumer choices and plan responses, the essential mechanisms of timely benefit innovations and cost control would be destroyed.

Requiring all plans to adhere to government-specified benefit details is comparable to requiring all automobile manufacturers to follow uniform, one-size-fits-all government specifications regarding size, seats, horsepower, cup holders, paint colors, and all the other features that distinguish one car model from another. There are obvious alternatives to detailed benefit specification, such as providing benefits that, in total, meet a simple and straightforward actuarial test of overall dollar equivalence.¹² This test should be applied to core benefits, not just extra benefits. No plan should have to meet the precise--and often unnecessarily costly, limiting, and arbitrary--parameters of Medicare Parts A and B for hospital and outpatient services.

As a simple example, almost all national plans and many HMOs in the FEHBP use a dual benefit structure for paying doctors and hospitals. Many private plans offer the same arrangement. Under these structures, benefits are significantly better when enrollees use preferred providers, but enrollees may nonetheless use out-of-network providers. In most cases, enrollees pay 25 percent of reasonable costs for these out-of-network providers and a low copayment for using preferred providers. This two-tiered structure provides a fail-safe for persons who need or want to use a particular physician who does not participate in their plan. It is particularly valuable in rural areas in which providers are scarce or unwilling to become preferred providers and accept fee restrictions. By definition, a legal requirement that plans offer the Medicare benefit structure Most important, no plan should have to meet the precise parameters of a "Part D" drug benefit added to traditional Medicare. This would freeze early design decisions into arbitrary patterns not used by any private plan and prevent innovations to control costs, improve benefits, and attract potential customers. Moreover, the proposed Part D benefit, as specified in both House and Senate-passed bills, is designed primarily to meet budget targets rather than rational health plan design.

Test 4: Allow flexible service areas and preferred provider depth
Some analysts have suggested requiring that plans competing in a reformed Medicare program have identical service areas specified by the government.¹³ The rationale is that this requirement will prevent plans from cherry-picking the healthiest, serve rural areas better, and simplify employee choice.

However, these are purely hypothetical solutions to nonexistent problems, and the employee choice argument fails even the laugh test. Modern Internet technology allows every single enrollee (or family and friend advisers) to receive or create plan comparisons by ZIP code without regard to what other areas the plans cover. The CHECKBOOK and OPM Web sites for federal employee plan choices organize and present plan comparisons by geographic area.

Furthermore, in the real world, plans serve and enrollees live in reasonably well-defined areas. Anyone can understand that Plan A covers all of New Jersey, Plan B all of New York and New Jersey, and Plan C the metropolitan New York area in those states and in Connecticut. The cherry-picking argument deals with a nonexistent problem that has never emerged in the history of the FEHBP.

Uniform boundaries could create an administrative disaster, and rigidly enforced boundaries would preclude participation by small plans specializing in particular areas. In effect, the government would be telling Kaiser and every other HMO and PPO that it must cover a named multi-state area even if Kaiser cannot and will not build a network to cover those precise areas. These problems might be less if uniform areas were applied only to PPOs rather than HMOs. However, even then, problems could abound if plans were prevented from providing service outside these areas or forced to expand networks in unnatural ways.

Under such a system, the government presumably would require West Coast plans to cover Alaska and Hawaii, Hawaii plans to cover the West Coast, and Puerto Rico plans to cover the mid-Atlantic region. A Pittsburgh plan in the mid-Atlantic region might be forbidden to cover Ohio residents just down the Ohio River because they would be located in the Midwestern region. Even the Blue Cross system, with its ever-evolving boundaries, might have to restructure its service areas throughout the nation to meet the Medicare boundaries. The only thing that is certain about government-prescribed geographic boundaries is that they would be wrong for every single private plan existing today.

These are not hypothetical issues. The Organ Procurement and Transplantation Network, the government-sponsored system for organ allocation, is plagued with problems created by its system of geographic regions. For example, patients on waiting lists in the Omaha metropolitan area who live on the Iowa side of the Missouri River must travel to distant Iowa cities to obtain organs simply because Nebraska and Iowa fall in different regions. The Medicare Prospective Payment System for hospitals has extensive problems in determining boundaries among reimbursement areas.

No system of service or payment with strict geographic boundaries can avoid anomalies like these. Moving the boundary from one place to another simply moves the locus of error and controversy. Allowing for exceptions (e.g., Puerto Rico plans can appeal to Medicare not to have to cover the mainland) simply creates another burdensome bureaucratic process and ultimately leads to a tangled mess. Furthermore, almost any geographic restrictions make it far more difficult than it otherwise would be for plans to build and maintain networks.

Traditional Medicare will not cover health care abroad, except in Canada and Mexico, and most of the government-mandated Medigap options preclude coverage abroad. However, every single FEHBP plan covers health care anywhere in the world, and HMOs offer emergency care anywhere outside the plan area. This is because few consumers would voluntarily enroll in a plan that did not offer this feature even if they had no immediate travel plans.

If this feature were expensive, some plans would decline to offer it and seek to attract the "stay at home" group. The fact that hundreds of health plans do not act this way demonstrates that the extra costs of this feature are small or may even save money since many foreign countries provide Americans with free or reduced-price health care. Geographic restrictions could, as under the existing Medicare program, defy the ability of plans to serve these enrollee needs.

Nor are geographic restrictions needed to promote rural access. Nothing in either logic or FEHBP experience suggests that every single plan need provide the same depth of provider networks in every geographic subunit. The robust FEHBP rural performance shows that there is no compelling reason why every plan in any area has to offer equally broad provider networks to assure good rural access. In most remote areas, several plans will offer good provider panels even if all do not. At the very least, some plans would provide for FFS benefits along with preferred provider benefits, as in the FEHBP.

Most important, requiring every plan to offer equal access or meet identical boundaries will restrict the number of plans willing to offer services in a given area and reduce the ability of plans to manage their networks efficiently. In other words, requirements for one-size-fits-all geographic coverage and minimum access standards for every county would deprive, not foster, enrollee choice of plans and providers while driving enrollee costs higher than necessary.

Moreover, restricting the number of plans allowed to compete in any area to three, as both the House and Senate bills would do, is the single provision most destructive of rural access. In the FEHBP, it takes perhaps a dozen plans available in each county to assure rich access to essentially all federal employees and retirees in America.

These arguments suggest not that geography plays no role in reform, but that any provisions need to be crafted very carefully to assure that boundaries do not create more problems than they solve. In summary, any sensible Medicare reform would have the following features:

• PPO plans should be encouraged to participate by permitting preferred provider networks that vary in depth from county to county, allow for access by reasonable travel, and are not uniformly strong in every single location.
• Enough plans should be allowed to participate in a given area so that almost all potential enrollees (except persons in truly remote areas) will always have one, and often several, plan choices with reasonably strong networks in or near their areas of residence.
• Enrollees should be allowed to use preferred providers and out-of-network providers across any government boundary lines, whether in adjacent metropolitan areas or at centers of excellence some distance away.
• Enrollees who live or travel abroad or winter in the South should have at least fee-for-service access, if not full preferred provider networks.

Test 5: Exempt competing plans from state mandates
The FEHBP limits state regulation of HMO benefits to those of the plan's home state. Thus, the Kaiser plan for the mid-Atlantic enrolls federal members from six jurisdictions, but must meet only Maryland--not Delaware, District of Columbia, Virginia, Pennsylvania, and West Virginia--mandates. National FEHBP plans are completely exempt from state regulation.

Private health insurance in the United States is bizarrely regulated. Each of 50 states can and does impose its own standards on every plan offered in the state. In effect, private-sector national plans must comply with 50 sets of benefit mandates.

No other national service or business is subject to such a draconian set of competing regulations. For example, corporations generally must meet only the standards of the state in which they are incorporated, not the standards of every state in which they do business. Likewise, in areas ranging from auto safety to food safety to copyrights and trademarks, federal laws preempt potentially disparate state standards to allow national marketing and sale of uniform products.

Whether exemption from state mandates is complete or partial, as in the FEHBP, it will be essential to "Medicare Advantage" to prevent state mandates from foreclosing flexibility in plan and enrollee decisions on benefit and coverage details and provider participation.

Test 6: Establish a sensible budgetary strategy
Medicare reform must meet both short-term and long-term budgetary objectives. Painful compromises on the generosity of benefits are necessary. However, if cost constraints force an unduly parsimonious reform package, along with "hole in the doughnut" prescription drug benefits, the entire purpose of reform may be jeopardized. Highly constrained and geographically bounded competitive bidding systems may have the unintended result of zero cost simply because no sensible health plan will want to participate.

In this regard, large employers may reap windfall reductions in post-retirement health insurance costs with the introduction of prescription drug coverage into traditional Medicare. There are, however, sensible ways to make that windfall smaller. For example, the tax deductibility of health insurance contributions to these firms could be conditioned on at least partial maintenance of effort for retirees, with the firms essentially being obliged to bear part of the cost of premium supplements for both old and new Medicare plans.

There are also less draconian ways to entice participation, such as making the program design attractively generous (even at higher premium levels for enrollees) and thereby pressuring employers to pay all or most of the employee share. One recent study concluded that the combination of what the government now spends on Parts A and B, plus the amounts typically spent on Medigap plans (whether by employers or retirees), would finance a generous defined contribution program.¹⁴

Further, there are the enrollees themselves. In a Medicare program in which long-term insolvency looms ever closer, increasing the proportion of costs borne by the elderly from its current small fraction seems desirable. Moreover, the higher the nominal premium borne by the elderly, the higher the level of subsidy that large employers will find themselves encouraged to bear in subsidizing that premium. Low-income elderly can and should be protected through premium subsidies, either by improving current arrangements under Medicaid or through direct discounts based on prior year tax returns.

In addition, Congress faces choices as to whether the government contribution should be designed to be higher, lower, or approximately the same as its cost under traditional Medicare. One option might be to set it slightly higher in early years to stimulate plan participation but have it gradually decrease over time to encourage long-term savings. The precise decision made does not matter (within reasonable bounds) so much as having a coherent strategy that makes sense and is understood by both plans and enrollees.

Test 7: The authorizing statute should be brief and simple, and regulations should be almost nonexistent
The length of the statute is a rough indicator of the program's complexity and the micromanagement imposed by the government. The FEHBP authorizing statute is only a handful of pages long, exclusive of eligibility standards and requirements. A Medicare reform statute establishing a new private-sector plan alternative could be written in several dozen pages. The statute will be hundreds of pages long only if it includes details that should be left to consumer and plan decisions rather than to the government.

The reform statute must not replicate the incredible morass of regulatory requirements that have been imposed on Medicare+Choice plans and on Medicare itself. Instead, private health plans should be allowed to operate as they do now without being forced to become mere Medicare contractors like the carriers and intermediaries that operate Medicare today. This means, for example, that Medicare coverage decisions, payment rules, benefit designs, provider rules, claims processing, and reporting requirements would not be imposed on these plans. Instead, they should meet overall tests of solvency, actuarial fairness, and information provision without being bound in any detail by current Medicare practice.

Test 8: Encourage employer plans to participate fully
One of the best ways to reform Medicare is to encourage retirees to remain in the same well-established employer plans that they used before age 65. Why should any American be forced to give up his or her health insurance plan, with its settled expectations and known providers, simply because he or she has turned 65?

This goal can be achieved through continued employer sponsorship of participation in the very same group plan(s) as in the FEHBP. This would be a radical change for many employers, whose current practice is to sponsor different plans post-retirement. Alternatively, existing plans could simply be allowed to participate in the new program and enroll individuals through Medicare Advantage. Of course, neither option would work if plans were forced to change benefits, operate by region, or change preferred provider panels. Limitations on the number of competitors in an area would preclude almost all of these plans from participating.

To make this reform most effective, the Part B penalty for late enrollment should be repealed for persons enrolled in a comprehensive health plan with a benefit package that is as actuarially valuable as Medicare's. The new Part D drug benefit should be handled in the same way. In addition, the specific prohibition against FEHBP health plans paying the Part B premium should be repealed.

Test 9: Encourage and allow FEHBP plans to participate
As a final criterion, FEHBP plans should be allowed to compete for Medicare business under a reformed system. The plans would continue their same networks, benefits, and management practices. These plans could readily segregate finances and enrollment information for the two enrollee groups. Of course, they would not compete if forced to comply with cumbersome rules that would affect their benefits and coverages, provider networks, administrative costs, ability to participate over an extended period of time without eviction from the program, or autonomy under the FEHBP. If the system that Congress ultimately chooses does not accommodate participation by these plans, it will likely fail.

Conclusion

A reform modeled around these nine tests would be likely to entice plan participation, contain costs in both the short and long run, and meet participant needs. Of particular importance in the debate over reform, rural participants would be well-served. In the FEHBP, rural access is enhanced by a combination of features including multiple plans (never less than 12 in recent years) in every county in America; dual benefit structures that reduce dependence on preferred providers; and flexible boundaries that fit medical delivery patterns. Through the participation of many plans, and dual benefit structures, all enrollees can find at least some reasonable alternative in their communities even if every plan does not have the provider panel they would prefer.

As a measure of this success in the FEHBP, a recent study found that in 87 percent of America's rural counties, federal employees and retirees sign up for six or more plans.¹⁵ Three or more plans operated successfully in 98 percent of rural counties. No competing model comes even close to this kind of performance, but limitations on plan participation would likely deny residents in many rural areas any choice of plans.

These nine tests summarize the features that should be included in any significant reform of the Medicare program through creation of a "Medicare Advantage" program to enlist robust and successful private-plan alternatives to traditional Medicare. Failure to include these features will simply perpetuate the statist Medicare program with its inadequate coverage, inferior benefits, and draconian cost controls. American retirees and taxpayers deserve better. The President and Congress should reject any Medicare reform bill that fails these tests.

Walton Francis is a self-employed economist and policy analyst and has authored the annual CHECKBOOK's Guide to Health Plans for Federal Employees for the past two decades. This paper is based largely on the author's testimony before the Senate Special Committee on Aging on May 6, 2003, and the Senate Finance Committee on June 6, 2003.