Available in Adobe Acrobat PDF format

The dispute continues in Washington regarding the costs of the recently enacted Medicare reform bill. The Bush administration now estimates that the program will cost more than a half trillion dollars over the next decade, and long-term costs will place an even heavier burden on future generations. Fundamental reform is required that improves health care delivery and provides better incentives to patients, providers, and health plans. The new law takes some initial steps to improve Medicare, but the issue needs to be reopened--perhaps as early as next year.

The ink was barely dry on the new Medicare prescription-drug law before the latest round of controversy inflamed passions on both sides of the aisle in Congress. In a remarkable act of candor, the White House announced in early February that the Medicare bill that was supposed to cost no more than $400 billion would actually cost $534 billion between 2004 and 2013. That triggered a storm of protest from conservative Republicans, who felt that there had been a commitment to keep a lid on additional Medicare spending. Even liberal Democrats, who would have been happy to spend twice as much as Republicans, were unhappy, suspecting that the new higher cost estimated for Medicare could be just an excuse to cut domestic spending.

Speculations are swirling about what key members of the Bush administration knew and when they knew it. Passing the Medicare legislation was clearly the top priority last year for Republican leaders at both ends of Pennsylvania Avenue. A high price tag would have killed the bill in the Senate, which could not spend more than $400 billion unless Republicans could muster more than sixty votes--a dubious prospect at best. Surely the administration's actuaries would have known months ago that the drug benefit was going to bust the budget. Is this a case of don't ask, don't tell? Or is there less here than meets the eye?

Dueling Estimates

It is no surprise that the Congressional Budget Office (CBO) and the actuaries at the Centers for Medicare and Medicaid Services (CMS) disagree about the cost of the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 (also known as the "Medicare Modernization Act," or MMA). The potential for widely varying cost estimates was apparent to some observers early on, and important points of disagreement became public last summer.

The political sensitivity of the cost estimates was made clear in a widely reported incident in late June. The administration denied a request from Rep. Pete Stark (D-Calif.) for a preliminary cost estimate of the Senate version of the Medicare bill, which reportedly showed that the Senate bill would greatly exceed the $400 billion that had been allocated for the legislation. Stark claimed that the CMS chief actuary had been threatened with dismissal if he released the estimate. Tom Scully, then administrator of CMS, claimed that Congress had no right to demand the report.

Disagreements between CBO and CMS about technical estimating assumptions were aired in late July, when CBO issued its first formal cost estimate of the House and Senate versions of the bill. CBO assumed that the new private health plan options that were being considered would not prove popular with seniors. Even under the more ambitious incentives of the House bill, only about 11 percent of Medicare beneficiaries would enroll in private plans, according to CBO.

That contrasted sharply with the administration's view. CMS claimed enrollment might be as high as 43 percent of Medicare beneficiaries, according to reports citing an unpublished June 26 memo from the chief actuary. Tom Scully was widely quoted as saying he had never seen "such a big difference in judgment."

Although CMS did not release a formal report, that difference in judgment was clearly a warning that the cost of the Medicare bill could be substantially larger than the $405 billion estimated by CBO for the House version. If CBO was correct, the private health plan provisions would be ineffective but the bill could slide in under the $400 billion limit. If CMS was correct, the private health plan provisions would work but the cost might balloon.

Concern about the cost of the new drug benefit remained high throughout the fall as congressional negotiators from the House and the Senate worked to iron out differences over provisions of the bill. The final version of the bill was scored by CBO as costing $395 billion--within the limit set by the Senate. As noted above, we now know that Bush administration officials expect costs to substantially exceed $500 billion, something they surely knew in November when Congress was taking the final votes.

Who Is Right?

Much of the press coverage of this controversy leads the reader to think that the administration's $534 billion is the real cost of the Medicare drug benefit. In fact, that number is no more reliable than CBO's lower estimate, although the administration's estimate points in the right direction. The actual cost of MMA will be much higher than either estimate reveals because the drug benefit represents a permanent commitment of resources to seniors rather than a benefit that will expire in a decade. It is likely that even the ten-year cost will be much higher than today's estimates indicate. Those facts were widely known from the beginning of the debate but little acknowledged by the administration or Congress.

Cost estimates are likely to be unreliable precisely when it matters most: when the legislation is novel and far-reaching, making major changes in a big program like Medicare. The cost estimate for MMA has many moving parts and is driven by multiple layers of assumptions about how seniors, health care providers, pharmaceutical companies, employers, and others will respond to new policies that depart significantly from past experience. Under those circumstances, good analysts can make reasonable assumptions and arrive at wildly different results.

This seems to be the case with the Medicare drug benefit. CBO explained the key differences between their cost estimate and CMS's in a February 2 letter to Rep. Jim Nussle (R-Iowa), chairman of the House Budget Committee. For such a controversial issue, the explanation is delightfully bland, showing that relatively small differences in numerous assumptions can snowball into a big dollar difference in estimates.

The biggest disagreement between CBO and CMS is the popularity among seniors of private Medicare Advantage plans. CBO assumes that only 9 percent of seniors will choose a private plan instead of traditional Medicare, while CMS assumes that 32 percent will enroll in such plans. That difference in assumptions accounts for only about a quarter of the $139 billion difference in the estimated cost, however. The rest of the cost difference is attributed to more modest differences in assumptions: CBO assumes slightly lower enrollment in the Part D drug benefit and slightly lower drug costs per enrollee. CBO assumes that seniors will be attracted to the low-income subsidies offered to them, but less quickly than does the administration. CBO also uses slightly higher projections of spending in Medicaid. These kinds of small differences add up to more than a $100 billion discrepancy between CBO and CMS cost estimates.

How many actuaries can dance on the head of a pin? There are vast uncertainties in estimating the cost of MMA, and trying to specify how one estimate differs from another is interesting but ultimately not useful. Even if one or the other estimate was correct, they assume no future legislative changes in Medicare. A more reasonable assumption, but one impossible to quantify, is that Congress will make the drug benefit more generous and make a host of other changes over the next few years that on balance are likely to raise federal outlays significantly. If that happens, the actual cost of the drug benefit could greatly exceed the $534 billion estimate from the administration.

The congressional budget process makes any argument about whose estimate is better moot. Long-standing budget rules put CBO in the driver's seat, making their estimates the basis for congressional action. That imposes some consistency on the analysis of different proposals and largely eliminates the risk that Congress will shop around for estimates that favor one policy position over another. (On rare occasions, Congress has directed CBO to score a proposal using assumptions from the president's Office of Management and Budget.)

In addition, the budget process demands a single cost estimate presented with a ludicrous degree of apparent precision. The uncertainty in projecting budget deficits that CBO emphasizes in its annual reports extends to cost estimates of individual proposals. But congressional budgeters need one number, since that can determine how much they have left to spend elsewhere. One cannot plausibly argue that the Medicare drug benefit will increase federal outlays by $44.6 billion in 2008 instead of, say, $43.8 billion, but that is the minimum level of accuracy required. If CBO offered a range of cost estimates for a specific proposal, Congress would demand that they settle on one number.

Finally, and most critically, the budget process is focused on the short term and does not account for financial commitments made outside the ten-year window. Arguments about even fairly large differences in ten-year cost estimates overlook the fact that such differences pale in comparison with the cost beyond the budget window. Douglas Holtz-Eakin, director of CBO, recently pointed out that the $400 billion Medicare drug benefit likely becomes a $2 trillion obligation in the decade after 2013. If we were to fund the new benefit in perpetuity, an immediate investment of some $13 trillion would be necessary.

The Medicare drug benefit has added significantly to the financial pressures facing the country in the decades ahead as the baby boomers age into retirement. Medicare, Medicaid, and Social Security will drive future spending as rapidly rising health and pension obligations to seniors outstrip growth in the economy. CBO, CMS, the General Accounting Office, and many others have repeatedly pointed out the risks of allowing those long-term commitments to continue to grow unimpeded.

CBO's latest report on the long-term outlook for the federal budget makes the point unambiguously. Unless Congress enacts major policy changes, CBO projects that Medicare, Medicaid, and Social Security will account for 17.4 percent of GDP by 2030, or the equivalent of $2 trillion today. By 2050, spending for those programs is projected to reach 27.6 percent of GDP, or more than $3 trillion in today's terms--tripling the cost of the three major entitlement programs. According to the report, current spending policies will be financially unsustainable over the next fifty years unless taxes are raised to unprecedented levels.

Long-term projections are fraught with even more uncertainty than ten-year budget estimates, but they indicate the risk of following the current fiscal path. The new drug benefit is the largest entitlement since the creation of Medicare in 1965. There are, however, other provisions in MMA that could begin to slow the growth of Medicare spending. Actions to promote consumer choice and competition among health plans, enhance efforts to improve the quality and appropriateness of care, and encourage the use of information technology are promising. But there is little doubt that program spending will continue to grow rapidly in the near term, creating pressure to revisit the contentious debate over reform and the future of Medicare.

Political Fallout

The furor over competing cost estimates has put new life into demands from both sides of the political spectrum for policies intended to constrain Medicare spending for prescription drugs. Such arguments are not likely to reopen the Medicare law, but political contentiousness could detract from actions needed to implement the complex new drug benefit.

Conservatives want a cap on spending to assure that federal outlays for prescription drugs would not exceed $400 billion. Liberals want price controls, calling for repeal of the provision prohibiting the federal government from negotiating prices directly with pharmaceutical companies. Those negotiations would be one-sided, however. The federal government determines who has access to the Medicare market and can force agreement on any price it chooses. Failing that, some liberals and conservatives continue to call for allowing drug importation from Canada, an indirect way of establishing price controls.

Neither caps nor price controls make sense. A cap on drug spending might rein in that part of Medicare spending, but ignores increases in spending for other health services. Thus, a cap could distort treatment decisions, favoring non-pharmaceutical approaches that could be less effective and result in higher overall program cost. Spending caps have not been successful when tried in the past since Congress often makes exceptions in the face of pressure for fiscal relief from constituents.

Price controls on pharmaceuticals might reduce federal outlays in the short run, but they would surely also have unintended consequences. If prices were set too low, patients might lose access to their drugs, or other aspects of their treatment might change for the worse. Over the long term, price controls would be no more effective in constraining federal outlays than the system of competing private drug plans that was legislated. However, such price setting would distort the financial incentives facing pharmaceutical companies. Price controls create additional uncertainty about the potential returns to innovation, discouraging research and development that could lead to new drugs for cancer, dementia, or other diseases of the elderly.

The case of cancer drugs provided on an outpatient basis (and covered under Medicare Part B) illustrates some of the difficulties with federal price setting. There is widespread agreement that Medicare overpaid for Part B drugs, although oncologists argued that those overpayments helped compensate for the extra costs of administering the drugs and caring for patients that were not reflected in fees paid by Medicare for office visits. MMA reduced payments for those drugs, leading almost immediately to cutbacks in services by oncologists. Recent stories in the press document a shift of patients out of the doctor's office and back to inpatient hospital care, which reduces patient satisfaction and could increase federal outlays. The extent of that disruption in the way health services are delivered may lessen over time as providers adjust to the new pricing system. Once the dust settles, however, we may find that this attempt to lower a few inflated drug prices has resulted in higher program spending and less effective treatment for some Medicare patients.

Political arguments will continue, but there is little chance that Congress or the administration will impose either spending caps or price controls in any guise on the Medicare drug benefit this year. The focus will shift to the difficult job of implementing the thousands of provisions contained in MMA. An important step in that implementation is to educate beneficiaries about the new benefits and health plan options, but even that effort has generated controversy.

The announcement in early February of a television advertising campaign to build public support and awareness of the new Medicare law was met with storms of protest from Democrats. The administration claims that this is fundamentally an educational effort; Democrats claim that the campaign is politically motivated. Both statements are undoubtedly correct. The television advertisement is on hold until the General Accounting Office decides whether the government can legally spend the money.

While politicians rail against each other, many seniors remain confused. A recent Philadelphia Inquirer story found at least one elderly couple who think they know enough about the Medicare discount card program to reject it. The only problem is that none of their options have been announced yet, and new industry alliances are forming every day that could improve the value of the program to people just like the couple in Pennsylvania. Eli Lilly, for example, announced that it would offer all of its products to low-income beneficiaries for $12 a month. Merck indicated that it would provide medications at no cost to such beneficiaries once they had used their $600 annual subsidy. Here is clear evidence that an effective information campaign is needed now, but that might be impossible in this highly political year.

By any estimate, Congress has made a very substantial and permanent commitment of new resources to Medicare. The effects of that commitment are being felt today at least in part through a tight domestic budget proposed by the president for 2005, but the immediate consequences are minor compared to the fiscal burden that has been incurred for future generations. Unfortunately, no quick policy fixes can responsibly and effectively control Medicare spending. The key to cost savings is fundamental reform that changes the incentives facing patients, providers, and health plans, and that improves the efficiency of health care delivery. MMA takes some initial steps toward that reform. Congress should rise above partisanship to support implementation of the new law and prepare for the next round of reform efforts that may come as soon as next year.

Joseph Antos is the Wilson H. Taylor Scholar in Health Care and Retirement Policy at AEI.