Will Congress Contain Medicare's Exploding Costs?

ROBERT MOFFIT : Recently, David Walker, the Comptroller General of the United States, noted that the official debt of the United States is more than $7 trillion, which is about $24,000 for every man, woman, and child in America. Mr. Walker told the National Press Club, however, that if you count the unfunded liabilities--in other words, the promised benefits of entitlement programs, including the new $8 trillion unfunded liability on the prescription drug benefit alone--you are talking about $42 trillion, equal to about $140,000 for every man, woman, and child in America.

Medicare is the toughest problem substantively. It is also the most difficult problem politically. The question is: Can Congress contain those costs? Will Congress contain those costs? How can Congress contain these costs? We have a distinguished panel of former Congressional Budget Office (CBO) officials with us today. They are going to try and answer these tough questions.

Our first speaker, Joseph R. Antos, is the Wilson H. Taylor Scholar in Health Care and Retirement Policy at the American Enterprise Institute. Joe is also an adjunct professor with the School of Public Health at the University of North Carolina. He was previously an assistant director for the CBO. Joe was also a staff economist at the Council of Economic Advisers and a senior economist at the Office of Management and Budget (OMB). He got his bachelor's degree in mathematics at Cornell University and his Ph.D. in economics at the University of Rochester.

Our second speaker is Jeff Lemieux. Jeff is the founder and executive director of Centrists.org, a new think tank in Washington with a decidedly "moderate" ideological stance on public policy issues. He is the author of a number of pieces about Medicare reform, balanced budget issues, and entitlement spending. In the area of Medicare reform, Jeff served as a senior staffer with the National Bipartisan Commission on the Future of Medicare, headed by Senator John Breaux (D-LA) and Congressman Bill Thomas (R-CA). The Commission produced an enormous body of analytical work on the Medicare program and its future. From 1992 to 1998, Jeff also served as a principal analyst at the Congressional Budget Office, where he helped to project costs for health care reforms, including the changes in the far-reaching Balanced Budget Act of 1997. Before his service at CBO, he was with the Office of the Actuary at the Health Care Financing Administration, the agency now known as the Centers for Medicare and Medicaid Services (CMS).

Finally, we are joined by Dan Crippen, the former director of the Congressional Budget Office. Dan was a domestic policy adviser to President Ronald Reagan and he also served as the chief counsel and economic policy adviser to Senate Majority Leader Howard Baker (R-TN).

Robert E. Moffit, Ph.D., is Director of the Center for Health Policy Studies at The Heritage Foundation.

JOSEPH R. ANTOS : David Walker is a great Comptroller General, but he is a little conservative in his estimates. He recently estimated that the drug benefit would cost $8 trillion over the next several decades. That is an understatement. My former colleague at AEI, Jagdeesh Gohkale, estimates that the Medicare drug benefit will have an unfunded liability that is more like $13 trillion. That assumes that things proceed along in a normal fashion without Congress expanding the Medicare benefit.

Can Congress contain Medicare's explosive growth? There is a 39-year track record on this question and it does not look too good. One thing is clear: The answer to getting control over the excessive spending in the Medicare program is not to keep doing the same thing that we have been doing. It is to do something else.

I think that the proper focus on Medicare costs should be the whole program, not just drug costs. However, there has been quite a mania here on Capitol Hill surrounding prescription drug costs and thinking those are the problem. Certainly, adding a big, expensive drug benefit adds to the overall burden of the taxpayer, but the broader program is also on an unsustainable path. There are obviously very positive aspects to covering prescription drugs in a modern health insurance program. Unfortunately, Medicare is not a modern health care insurance program. That is a major problem in itself.

Drug Cost Estimates. Let's talk about drugs. The recent debates focused on what the Medicare Modernization Act (MMA) would cost. We think we know the numbers--$395 billion coming from CBO, which is the only relevant estimate when it comes to congressional deliberation. The OMB estimate that came out in January or February was $534 billion. Yet that is the wrong perspective. Comptroller General David Walker, the head of the General Accounting Office, was reaching for the right number. The new benefit is a permanent entitlement: It is not going to stop in 10 years. This is a permanent commitment of resources on a massive scale to pay for prescription drugs under Medicare.

One might ask how much Medicare is going to spend on drugs alone, because there are other things in that bill. If you look carefully at CBO's November 20, 2003, letter to Senator Don Nickles (R-OK), you can find the answer. What you see is that there are three drug-spending components in the bill--not counting Part B drugs, and apart from the new Medicare discount card.

What are the three components of the 10-year CBO estimate? Benefits amount to $507 billion, which is money spent primarily through the prescription drug plans or through the new Medicare Advantage plans that will offer a drug benefit. Next, there is an employer subsidy of $71 billion to support drug purchases. Finally, there is a low-income subsidy of $192 billion. Note that this is not a payment to people: It is a payment to drug plans. That money might, in fact, go to any of the sponsors of drug benefit programs. When you add up all of these numbers, you get $770 billion in total federal spending for prescriptions under the new drug benefit.

There are some offsets. Only one offset directly relates to purchasing prescription drugs, and that is the amount of premiums paid by beneficiaries enrolled in the drug benefit. Those premiums come to about $131 billion. You could think of that as beneficiaries paying back some of their drug benefits. If you look at it that way, total federal spending for Medicare drugs under the benefit may be $640 billion instead of $770 billion.

Two other big offsets literally have nothing to do with buying prescription drugs for people in the Medicare program. They include $88 billion that the states will kick back to the federal government. That is not Medicare money. That is actually Medicaid money. In addition, Congress reduced federal Medicaid spending by $155 billion. This is the result of shifting dual-eligible persons from Medicaid to Medicare for their prescription drugs.

What is the real number? It might be $400 billion; it might be $500 billion; it might be $800 billion; and it might be $13 trillion in promised, but unfunded, benefits over the long term. Whatever the total might be, it is an unlimited amount because this is an entitlement.

Cost Containment Provisions. What about "cost containment" in the new Medicare law? I put cost containment in quotes because there really is no cost containment provision in the cost containment title. There are cost containment elements in the bill, just not in the section labeled "cost containment." There is the trigger provision: Congress and the President are supposed to take action when general revenues exceed 45 percent of total Medicare outlays. Congress has decided that that is an indication that Medicare is in serious financial trouble, and the level of general revenue financing will be predicted by the Medicare trustees.

They will look out over the next seven years and if they see that general revenues will account for 45 percent (or more) of Medicare outlays, they will serve notice. If they see that problem two years in a row, they will issue a Medicare-funding warning.

What is supposed to happen as a result? The President, by law, is required to submit proposals to Congress to address the issue. However, those proposals might not reduce excess spending. The proposals could increase revenue, decrease spending, or go the other way. On balance, it is not entirely clear what the President must propose, but he has to propose something.

Although Congress wrote in some expedited procedures for consideration of these presidential proposals, it is not obligated to do anything. At least from a political standpoint, this is a good cost containment provision because there is no political harm in it--at least not immediately.

How quickly will we see a warning from the trustees? This year's Medicare trustees' report projects that 2012 will be the first year that we hit the 45 percent level of total Medicare spending from general revenues. It could happen sooner. Our first notice that this is a problem may be given in next year's trustees' report. The year after that--if they are right about this--we will have our first official funding warning and then we will have such a warning every year for the rest of our lives.

Old Medicare Strategies. Let's talk about how you put a lid on Medicare spending. Essentially there are three ways that people have thought of to do this. The first way--the traditional method for Medicare--is to put tight limits on payments for individual services. We have used this approach over the last 30 years or so, and some argue that it has been effective. However, it is not permitted for prescription drugs by the Medicare Modernization Act. Under the new Medicare law, the Secretary of Health and Human Services is not allowed to directly negotiate drug prices with drug companies. The Secretary is supposed to let private competing plans do the negotiation: He or she is not supposed to interfere. The bill did not fundamentally change anything about importing drugs from foreign countries. Therefore, at least as the bill is written, that is not an option.

I do not think that Congress will seriously consider re-importation in 2004--and probably not next year, either. However, time is young.

The second method of limiting costs is something some fiscal conservatives wanted to see in the bill. Why not just cap overall Medicare spending? A number of people have been thinking about this and we are liable to see yet another proposal come down the road--possibly from Senator Judd Gregg (R-NH)--sometime later this year.

Finally, we come to the only way that, in my opinion, really works: market-based Medicare reform. Improve the efficiency of the delivery system. That is hard work. It is not an easy political fix, but it is what we have to do.

Spending Caps. Before I get into that, however, I want to say a few negative things about spending caps. I do not believe in them. If they were effective, they would do more harm than good. Go back through the track record on these budget gimmicks--it hasn't been very good. The Gramm-Rudman-Hollings Balanced Budget and Emergency Deficit Control Act of 1985 is an example. That act was an attempt, government-wide, to limit the growth in federal spending. It basically did not work and it fell into disuse. On the other hand, one might call the sustainable growth rate formula for physician payment in Medicare a success. That has been effective, to the dismay of many members of Congress who voted for it.

There are real downsides to spending caps. The sustainable growth rate formula for physician reimbursement illustrates the problems. In 2002, CMS took a 5.4 percent across-the-board cut on physician fees. There were all sorts of complaints from physicians. In some parts of the country there were reports that Medicare beneficiaries were having trouble getting appointments with specialists or providers other than their family physicians.

However, the physician caps did not work in a more fundamental sense. Spending grew by $3 billion in physician payments that year. If you talk with the Medicare actuaries, you know that this is remarkable, and it was likely driven by a big increase in the volume of physician services in the Medicare program. You can cap certain things, but you cannot control behavior. That is one of the lessons.

Caps do not provide lasting cost containment, but they can cause enormous political problems. I think Congress is going to endure those problems for many years. The budget rules make it difficult to change.

The Balanced Budget Act of 1997 succeeded in cutting Medicare spending, mainly through cuts in provider payments. Those cuts helped to drive managed care plans out of Medicare. If payment cuts were effective, you could have this same effect again, particularly with prescription drugs and private plans. Cutting back capitation payments by 30 percent is probably not the way to encourage private entities to participate in the program. We did it in 1997, 1998, 1999, and so on, and we could do it again.

Finally, how do you weigh immediate budget savings against the future cost of forgoing some research and development--and potentially forgoing the creation of useful new drugs? There are all sorts of views about the pharmaceutical industry and research, but economists agree that if you severely cut back the profit potential in any industry, the innovators in that industry find someplace else to put their money.

What can we do about rising Medicare spending? We have to do the hard thing. We have to make structural changes to improve the system's efficiency. Fortunately, the Medicare Act does have some components of meaningful cost control. There is more competition in the Medicare program because of this bill than there has been to date. We will see how that works out because there are problems, but nonetheless, there is a chance to see effective competition under MMA.

Moving away from the adjusted average per capita cost as a system of paying health plans and setting rates on the basis of competitive bids is a giant step forward in thinking. Yet again, we have to see how that is going to work out in its implementation.

Finally, there are some small initiatives that begin to get at the heart of health care delivery, which is what we ultimately ought to be concerned about. These are disease management, quality initiatives, information technology, and value purchasing. My concern is that these have already become political buzzwords as opposed to the real thing. These ideas are not quick fixes. These are good ideas that need to be developed into effective measures. That is going to be tricky in the context of the existing Medicare program, but these are good things to do and to promote.

Real Reform. What else can we do? What else must we do? The answer: what everybody has talked about for the last four years. The Federal Employees Health Benefit Program (FEHBP) should be adopted as a model for a new Medicare program.

The payment formula in that model does something that the Medicare Modernization Act doesn't succeed in doing: It puts a limit on federal spending, but in a sensible way. That limit recognizes that health care costs do rise even in an efficient system. It does not penalize beneficiaries dollar-for-dollar just because the inherent costs of health care have gone up from one day to the next. This is a problem with some versions of defined contribution plans in health policy. We have to allow reasonable growth for reasonable expenses, but we do not want to remove the incentives for plans and providers to try to do a better job in serving Medicare patients.

The FEHBP program does that. It offers consumers realistic choices, not promises that ultimately cannot be kept. That is where we are in Medicare today. We need to solve this problem, and there is a time-tested way to proceed.

JEFF LEMIEUX: Let me tell you a personal story of how even the best-intended price controls can go awry. When I was at CBO, I was very closely involved with the development of the sustainable growth rate (SGR) formula for updating physician payment in Medicare. Through that formula, we tied updates in physician Medicare payment to a set of variables, including the growth in the economy. It was a real eye-opener. It was the staffers at the Committees on Ways and Means and Senate Finance who were the key innovators. However, I was the one who had to do the estimate for that SGR. I think I got the estimate basically right. I said, at the time, it would save some money. Yet all I was thinking about at the time was the estimate of savings, not the possibility that the SGR system could be very volatile. What I did not say was that it would cost a lot of money in a few years and then save a ton of money a few years after that. We predicted much too smooth of a curve and did not predict the political fallout that would occur when doctors saw big rate cuts. That was a lesson to me. When Congress tries to control costs in Medicare, it can, but in the long run it can also be very difficult.

The Role of Private Health Plans. I would like to say a few words about the budget situation and Medicare spending growth, and then discuss how price controls can go in unintended directions. Third, I'll say a few words about private health plans and competition in Medicare, and how policy misconceptions can have unintended consequences. Next, I will say a few words about a theory that I have been trying to develop about how private plans can form a political alternative; and examples for how Medicare can evolve and possibly save money--even if they do not necessarily save money on Medicare in a strict formulaic sense. Finally, I would like to say a few words about the discount card, which affords some hope as a model for controlling Medicare costs.

First, let's talk about the federal budget, the surplus or deficit: The latest projections are that the federal deficit is going to be about $440 billion this year. That is almost 4 percent of gross domestic product (GDP). It will probably range somewhere between 3 percent and 3.5 percent of GDP for the next several years as the economy recovers. After that, things just go downhill.

If we have an extension of the tax cuts, which I am assuming, and we do not add any new spending or any particular spending cuts, the deficit stays within a workable range--not a good range, but a workable range of 3 percent to 4 percent GDP for the next several years. However, as the Baby Boomers start to retire and really kick in the Medicare and Social Security spending, the situation becomes completely untenable.

There are four big entitlements, and Medicare is only one of them. Social Security is going to grow by about 2 percentage points of GDP over the next 25 years, Medicare by a little more than that. Medicaid has really become a long-term care program and a program for disabled people. As you can see, the long-term prospect is that Medicaid is going grow very rapidly if it stays under its current structure.

Finally, there is the entitlement that conservatives always forget about--interest. We pay interest on the national debt. That spending entitlement has shrunk the most in the 1990s as we switched from budget deficit back to surplus, as interest rates fell, and especially with the start of the recession in 2001. Now that the economy is recovering, interest rates are going to increase and we are tacking on debt of over $400 billion per year. The "interest entitlement" could conceivably become our largest entitlement after 2030. If we do not start to control our deficits, we are going to have a major entitlement problem on our hands.

Medicare spending is growing pretty fast, with double-digit rates in the early nineties. In the mid-1990s, we started passing anti-fraud measures, which had a pretty big impact--a bigger impact than a lot of people expected. There was some over-billing in some hospitals. The growth in Medicare spending started to come down.

The Balanced Budget Act of 1997 (BBA) caused the growth of Medicare spending to fall a lot in conjunction with the anti-fraud measures. Some of the BBA cuts were more than providers could swallow and we started to give back money in the late 1990s and the beginning of this decade. Medicare spending is expected to stay under 7 percent for another year or so. Then the Medicare Modernization Act really begins in 2006 and there will be a big spike in the growth of Medicare spending.

We are growing the Medicare program as an ever-larger percent of GDP. This process is accelerating with the Medicare Modernization Act of 2003.

Price Controls. Price controls will not be a very durable way to control Medicare spending. It is exactly like the physician fee debacle that we just discussed. The formula gave physicians some pretty big raises right after we implemented it and that caused physician spending to go really high. Yet we also included in the formula a hard cap on physician spending and once we hit that cap the physician fee updates were scheduled to be negative for almost a decade.

Measured as dollars per unit of service, the physician fee schedule will go from in excess of $36 to under $28 by the time that the Baby Boomers are well into retirement. Of course, what Congress is going to do is fix this each year. Yet throughout the course of this next decade, it could easily cost $120 billion to $150 billion to fix.

Congress fixed the Balanced Budget Act of 1997 for a little while and that cost us some money. However, it also saved us some money in the "out years" because Congress did not really fix it. They just kicked the problem down the road, which actually made the physician fee cuts worse a few years later.

The HMO Issue. I would like to address briefly Medicare health maintenance organizations (HMOs). We had a big boom in Medicare HMO enrollment in the mid-1990s to late 1990s. Actually, after the enactment of BBA, HMO enrollment started to fall at a pretty rapid rate.

We have often been told by health policy experts that Medicare HMOs get the healthiest seniors; therefore, they are costing the program a lot of money. By that theory, when Medicare HMO enrollment went up a lot in the late 1990s, you should have seen Medicare spending go up because you are taking healthy people out of the traditional Medicare fee-for-service program--the very people who didn't cost us anything--and putting them into an HMO in which we are paying a capitation rate that is very close to the average Medicare spending and vice-versa: Once Medicare HMOs almost started to fall after the enactment of BBA, the logical thing that you would expect to see would be overall Medicare spending starting to fall.

The Big Surprise. It did not look like it really worked out that way. When the growth of HMO enrollment jumped up into the 20 percent range, Medicare spending started to trend down. When the growth of HMO enrollment turned negative in about 2000, that was roughly when Medicare spending started to come back up.

Of course, there is more going on here as we have already heard. There are the budget cuts, the givebacks, and all sorts of things. Yet all other things being equal, we would have expected the opposite of what did, in fact, happen. It is hard for me to imagine that the Medicare budget cuts and then the Medicare " givebacks" were so powerful that they could have negated the predicted theoretical outcome that so many health policy experts expected. Therefore I concluded that perhaps the Medicare HMOs were not, in fact, enrolling such healthy people. They might, in fact, be enrolling people with normal healthiness or they may be enrolling sicker-than-average people.

I did a little bit more work trying to determine whether or not their payment rates have changed corresponding to Medicare fee-for-service, and it turns out they have. After 1997, we essentially paid private firms in Medicare about the same rates of increase as the growth in fee-for-service spending. Yet it turns out it was really skewed. In the plans where people actually live--in the large urban areas--the payment rates have grown a lot less and that is why you have seen such a decline in enrollment. HMOs in areas where there are fewer people received the so-called floor update. Those HMOs who received the floor update actually got higher updates than Medicare fee-for-service.

In the last several years, there have been more payment adjustments, especially to some of the large low-cost urban areas, such as Minneapolis and Seattle. We are paying HMOs about the same as we paid them in 1997, but there has been this dichotomy on how we pay them.

As a CBO estimator, my theory about these HMOs is that if they come back into the Medicare program as a result of the payment increases we enact to bring them back up to par--as we do in the Medicare Modernization Act--it will certainly cost us money. However, I also have a theory that having a fairly large, workable, noticeable HMO sector in Medicare causes Medicare administrators to think very differently than when they simply operate the traditional Medicare fee-for-service program. Suddenly there is an example out there of people doing things differently. If you see a big surge in enrollment into this alternative sector--as we did in the mid-1990s--I think that caused Medicare administrators to take notice and to try to find ways to ratchet down Medicare fee-for-service spending in order to make Medicare fee-for-service more efficient.

The fact that a lot of these Medicare HMOs have drug benefits really caused politicians to ask, "Why doesn't Medicare have it? Shouldn't we do something about this to modernize the program?"

I think, both in terms of keeping the benefits up to date and in terms of cost control, having this alternative sector out there is very helpful to the Medicare program. I am somewhat optimistic that the MMA has brought HMO payments back up to par--back up to where they were in 1997. I think this will cause more HMOs to come into the program. There is a provision in the bill to try to bring more preferred provider organizations (PPOs) into the program, but sometimes laws that we pass have the opposite effect of what is intended. Yet I think this one probably does have a chance to bring more of the "less-managed" HMOs into the program and I think that will be very helpful.

I don't predict that HMOs and PPOs are going to take over the program or come to anywhere near 30 percent or 40 percent of market share. It may be 20 percent or 25 percent over the next several years or the next decade. Yet it will be very helpful to have that robust alternative program out there for us to look at and to learn from as we continue to work to improve the program.

The Potential of the Drug Card Finally, my only other source of hope about Medicare spending is this drug discount card. You have probably heard from the critics that the seniors are calling up saying, "It is too confusing on the Web," and "I couldn't get through to 1-800-MEDICARE," and "There are so many choices. What am I going to do?" Those are all valid concerns. There are some parts of the new Medicare law that were drafted poorly. I don't understand why they lock seniors into one card until the next open enrollment period. It makes no sense at all.

That, of course, causes seniors to think, "This must be a very serious choice." I initially couldn't get my mother to sign up. She is age 66 and pretty Web-savvy, but she could not get through the bit where you had to type in all your drugs. She just did not want to give the government that information, I guess. Yet she finally called 1-800-MEDICARE and got a card. I said, "Which card did you get?" She said, "I got a card that I have to drive all the way down to Wal-Mart [to use], but it's a dollar cheaper per month than the CVS card that's right next to where I live." That's interesting.

All told, it took a month to get her enrolled in this thing. She is saving $45 per month at Wal-Mart when she would have been saving $44 at CVS. Yet she spent all this time worrying about which card she got. She could have picked any card and would have saved a fair bit of money. She is not one of these people who shops smart and gets on the Web and goes to Internet sites or anything like that. It is those non-smart shoppers who save the most from the discount card.

I think this model is potentially a very good model. I think that one of the problems we have (other than seniors thinking it's such a serious deal) is that seniors think it is the final drug benefit and the choice they make now is going to be with them for 30 years. Then the enrollment is confusing. You have to find the right card. You have to type in your drugs. It should be much easier. You should just be able to make a selection at any time. Yet I think one year from now, people are going to really start to like this idea.

I think that as the people in CMS and in Congress start to think of it more as a public-private partnership that Medicare is promoting, that will help a lot. The discount card provision has a chance to be popular enough that when we implement the 2006 drug benefit, ordinary people are going to say, "Now wait a minute. You guys have finally got this discount card thing up and running and we like it. It's saving us a lot of money, especially if we weren't very careful shoppers before. Now you're telling me I have to go get a new stand-alone drug benefit, and this time, it is kind of serious. Instead of being free or essentially free, now I have to pay a serious premium and there are serious implications to how this reacts with my retiree coverage and other things."

New Options for Drug Coverage? I think there will be political groundswells to keep those discount cards going. We could perhaps add to the low-income benefits and boost that amount above $600. I would like to think that we could add a catastrophic benefit for people who have devastatingly high costs. I think we are capable of doing that through the drug discount card. It might be a lot less traumatic than trying to switch to a whole new system with the 2006 benefit. Maybe you do both and give people the option. Do you want to stick with a mostly free discount card with a few extra things that we are going to put on it? Or would you like to pay a premium and go to a more elaborate system and see how it works out?

My fear--because you see how hard it is to get the discount cards going--is that the big drug benefit will be harder. My fear is that if the drug benefit's implementation goes badly, it gives this whole idea of public-private partnership (and using the government to put together better choices and efficient options for seniors) a bad name or a black eye. I think that next year we need to be very careful about writing new legislation to improve the discount cards in order to smooth over the transition to this large drug benefit--and make the discount card another option as part of the process.

DANIEL L. CRIPPEN: Let me begin by saying that we need to think about an entirely new Medicare program. I recall something often attributed to Albert Einstein. I'm not sure he ever said it--it doesn't sound quite like him--but it goes like this: "The definition of insanity is doing the same thing over and over and expecting a different result."

That is essentially what we have been doing with Medicare for many years--trying to adjust provider payments--with ultimately little to show for it. Although occasionally we have changed something for beneficiaries, it has generally only been at the margins. In neither case have these changes ended up providing much relief from the inexorable upward spending imposed on the American taxpayers, nor have they provided a comprehensive health benefit for our seniors and disabled citizens.

We had a budgetary aberration, a slight downturn in spending, in the aftermath of the Balanced Budget Act of 1997. As Jeff said, there are lots of reasons why that might have happened. It certainly has not been a successful outing when it comes to controlling costs. On average, since the inception of the Medicare program, per capita real costs have grown much faster than the economy.

As we come to the cusp of my generation's retirement, we will face a doubling of the number of retirees for the program--from roughly 40 million to 80 million. Costs are obviously going to go up a lot simply because of the raw demographics.

Yet more important to the overall outlook for this program is this underlying trend of Medicare spending that is growing much faster than the economy. As Herb Stein might have said: This cannot go on forever.

From the very beginning of the program in 1965, we predicted that hospital costs would come into line with wage growth and economic growth. Obviously, it still has not happened. We are still waiting for the lines to converge. However, I am very much afraid that in the process of waiting, the current spending trends will have a very detrimental effect on my kids and the generations to come.

A New Paradigm. Some of you are old enough to remember that there was a fellow who did a lot of work in the first Bush Administration, Jim Pinkerton. He coined, or at least used the phrase, "a new paradigm." It was a way of framing an argument for major change. Certainly in the case of Medicare, we are due for a new paradigm. I do not know if Pinkerton is available, but maybe we can bring him back to give us a hand: Just don't tell Dick Darman!

In my opinion, too many commentators, too many pundits, and even too many health care analysts are issuing pronouncements about what we can do, how we can do it, and how much government policy or spending can move things along without regard for the context of the recent past and the foreseeable future.

Disappearing Surpluses. First, there is the issue of available budgetary resources. In that context it is instructive to ask where the budget surpluses came from and where they went.

We at CBO initially thought there were going to be surpluses for a long time. Many people would like to claim that it was the dramatic congressional actions taken in the late 1990s--the changes in budget policy, including the BBA--that yielded the strong projections of budget surpluses. In fact, the change in budget outlook was not due so much to change in policy as it was to change in the economy. The economy performed better than anybody expected at that time. Also, the economy produced more revenues than anybody expected or had seen in the past. That was due to the stock market run-up, capital-gains revenues, and also stock options being cashed in, which were taxed at normal rates (and for most people at quite high rates). Therefore, the revenue increases were greater than you would have otherwise expected. It was largely the change in the economic performance and the production of revenues that caused the change in outlook--not dramatic changes in legislation or policy.

What happened to the surpluses? The tax cuts, which are often blamed for the entire change in the budget outlook, are certainly part of it--particularly in 2003 and 2004--and they are significant. In addition to the tax cuts, spending has grown rapidly, largely in response to 9/11 and the war in Iraq. Yet there has been a myriad of other spending increases as well, including Medicare givebacks, tuition reimbursement, and other domestic discretionary spending. In fact, spending is larger than the revenue changes in terms of this altered budget outlook. Again, the economy is a major piece of the budget story. What the economy gives, the economy can take away.

In the context of the economy, we can push our analysis back in time, say, for 50 years or so, and examine our country's propensity to "tax and spend." We have averaged about 18 percent of GDP as a federal tax-take during this entire period. Obviously it has been much higher in some periods--often the periods immediately before a tax cut--and lower in others, but it is has oscillated around 18 percent for much of this period.

Looking to the end of this decade, there is a chance that the economy, and therefore revenues, will grow enough that we will get back to something close to balance. I am not predicting that we are going to end up at balance, but we could be within shouting distance of balance by the end of the decade if the economy continues to perform, spending is held in check, and no further tax cuts are enacted.

In my analysis, I assumed per capita Medicare growth of 3 percent after inflation and I incorporated a drug benefit that amounts to about 1 percent of GDP. Therefore, growth should be a little higher than the CBO assumption during this decade or during this period, but not a lot higher. The punch line is that you get the 16 percent or 17 percent of GDP by the end of the retirement of the Baby Boom generation.

Tax Bite. Therefore, if we continue to tax an average 18 percent or 19 percent of GDP--as we have since World War II--almost all of that federal revenue would have to go to funding Medicare, Medicaid, and Social Security. Put another way, we would have to eliminate the rest of the federal government as we know it, including the Defense Department.

Of course, we could try to borrow our way out of this shortfall, but it would mean essentially trying to borrow the equivalent of $1 trillion per year in current dollars. That amount would grow dramatically simply because of interest costs and would quickly outpace savings here and abroad.

We could obviously attempt to tax our way out of this problem, but it would take enormous increases, at least for this country. To protect programs for the elderly and preserve our current level of government spending on other programs would require a tax increase of roughly 10 percent of GDP--ten times the size of the Bush tax cut of 2001--or a payroll tax rate of 30 percent to 35 percent, for example. By the way, if you decided to fill the financing gap by raising payroll taxes to 35 percent, you would have a very dramatic effect on labor supply, which would negatively affect the outlook for the economy.

The Crisis Is Now. It is pretty clear that in addition to borrowing or taxing or cutting other spending, we need to spend a lot of time thinking about these programs for retirees and the disabled and how to reform them. We also need to think about reform regardless of the status of the trust funds. While there has been much discussion about Part A trust funds going belly-up or being insolvent in the next decade, the real truth is that this year is the first year in which payroll taxes will not cover spending for Part A, which means the economic and budgetary consequences begin this year.

The point is that we are going to have to supplement Part A payroll taxes with real dollars. Those transfers may be called "interest payments," but they are payments of general fund revenues, either borrowed or taxed from the public. Eventually, there will be bonds cashed in, which is a perfectly fine thing to do, but those bonds have to be converted to cash and that cash has to come from taxpayers in the form of general funds. Therefore, this year is when the real economic and budgetary consequences start, not some future date of "insolvency."

The same point, by the way, applies to Social Security. It looks like 2019 will be the year when Social Security becomes insolvent--certainly not 2052 or 2049 or whatever the projected date may be.

In the end, trust funds are irrelevant to our calculation of the immediate impact on taxpayers and the budget. The trust funds do not contain resources that can be tapped to reduce the burden on the working, investing, and taxpaying population. In the main, the working, productive population will have to finance the Baby Boomers' public retirement, whether through taxes, borrowing, or reduced spending on other programs.

I should note another consideration here, the question of whether you try to finance benefits by imposing more beneficiary charges or by increasing taxes on workers. There are differential effects on the economy, depending upon the level of intergenerational transfers. Obviously, the more you use beneficiary premiums, the less you rely upon current taxes and current taxpayers, and the less you will have an effect on the economy (but never zero).

A relative handful of the elderly expend the lion's share of Medicare resources because they have the most health care needs. The top 20 percent of Medicare spenders account for 84 percent of the total Medicare spending in a given year--25 percent of the beneficiaries account for 90 percent of the dollars.

Smart Card. With this distribution in mind, consider a new paradigm, a new Medicare policy. Let's first make a distinction between the 75 percent of Medicare beneficiaries who expend only 10 percent of the resources and rarely end up in the hospital, and the high-cost beneficiaries. Take that 30 million people, give them a budget, think of it as a dollar amount encoded on a smart card, and let them spend it as they see fit. They could go anywhere they wanted and see any provider to get their health care. They could fill prescriptions with their smart card. They could see doctors with their smart card. They could have lab tests or whatever services or procedures they needed or wanted for their health care. No matter how you see the program working, the simple truth is that these 30 million beneficiaries do not spend enough, do not need enough health care that we should follow them around with a rulebook or require extensive reporting by providers.

You could income-relate the amount of dollars loaded on each card. You could vary the amount for patients with chronic diseases. You could have co-pays for beneficiary incentives. You could do all those kinds of things and more, but the point is you could cut loose three-quarters of the Medicare beneficiaries and not worry about controlling their behavior because they do not spend enough.

Focusing on High-Cost Patients. That leaves us with the one-quarter who do spend most of the money. What would we do about it? It turns out we do not know much about these patients. The data that we have is largely a mass of separate transactions. When my father sees a doctor, it shows up as a Medicare record with payment and diagnosis and other ancillary information. A lab test produces another record. If he goes to the hospital, there is yet another record, and until recently no one in the federal government added up what he spends in a given year or over his lifetime. More important, no one tracks what he is being treated for or what the outcomes of the treatments are. It is simply a series of transactions. The bills get filed and they get paid.

Fortunately, with Joe Antos's help before he left CBO, and others since, we were able to assemble a first dataset of some beneficiary-level data, by incorporating these transaction records into monthly spending data for about 3 million beneficiaries over the course of 12 years or so. It is not a very satisfactory database. For one thing, it does not have any pharmaceutical data: There hasn't been a pharmaceutical benefit in Medicare. It is also difficult to use. The transaction codes, the diagnosis codes, and the ICD-9 codes may be inconsistent or mask several conditions. Therefore, it is not particularly good data, but it at least provides some insight into these patients.

My specific pitch today is that we ought to start a program of collecting much better data, especially on these beneficiaries who are sicker. That would mean going out with actual surveys and doing medical abstracting, histories, lots of things. Pick a sample of 10,000 of these patients and really do a lot of work determining their conditions, how they are being treated, and their medical outcomes.

Given what we know, or are beginning to learn, it is encouraging to think that we could do a much better job of taking care of them and, at the same time, probably save a lot of money in the process. If you look at these "big spenders" you will find, for example, most of them go to hospitals at least once a year--and often several times each year. If we could reduce the admission and re-admission rates, we would likely produce superior results at lower costs.

Think again about the 75 percent who do not spend much money--some 30 million--with their theoretical smart cards. That is their Medicare program until they require hospitalization. However, once they enter a hospital, we are going to take a much closer look, with much more scrutiny--intense management, case management, or whatever you choose to call it.

By the way, this is a case in which disease management as it is currently practiced, while promising in some quarters, may not yield much for this population. Many of the elderly and disabled patients suffer from three, four, or five chronic conditions--in particular, heart disease, diabetes, high blood pressure, and chronic obstructive pulmonary disease. Therefore, how we treat these specific diseases in the younger, commercial population, does not necessarily apply to the elderly. For example, the drug regimens for multiple conditions have not been developed and we end up sending a lot of these people to emergency rooms with drug interactions. We do not know how to treat these multiple conditions very well.

We do know some instructive things about this population, however. We know they end up in hospitals repeatedly. We know from looking at these data that many of them have 10 and 15 physicians--two physicians for each of their conditions, a general practitioner and an internist, for example. They fill over 50 prescriptions a year.

We know no one is responsible for coordinating this care. There is no payment for coordination. There is no requirement for coordination. In the best of circumstances, there is a good doctor out there somewhere trying, or a good son or daughter, or the patient themselves. At the same time, the more conditions these patients have, the more likely it is they are also impaired mentally. Therefore the more coordination they need, the more likely they are to be unable to do it themselves.

With apologies to Jim Pinkerton, to the extent I offer any kind of a "new paradigm" for Medicare reform, it is that we need to think a lot harder about these beneficiaries who need the most care. If you look at populations that have similar, often multiple conditions, like Medicaid in which there has been fairly intensive case management, hospitalization rates have been cut by 20 percent to 50 percent. If we can manage to keep people out of hospitals, we are likely to improve their health and substantially reduce Medicare expenditures. The short, simplistic version of the new paradigm: Keep Medicare beneficiaries out of hospitals.

There are other ancillary data to support the notion that costs could be substantially reduced without changing health outcomes. For example, Wennberg and Fisher found substantial differences in the cost for Medicare beneficiaries from Florida to Minnesota. While the practice patterns are different, there are no differences in health outcomes, suggesting that if Minnesota practice standards were used around the country, Medicare costs could be reduced by 25 percent.

I hasten to repeat that this is only suggestive. This is the basis for some hope about a new approach and a whole new system that separates the high-cost Medicare patients from the low-cost Medicare patients. Most Medicare beneficiaries do not need intensive management. Let them do their own thing--they do not cost enough. Once they are hospitalized they could enter another, more intensively managed program. Additionally, you would try to pick up the people who were going to be expensive before they actually got expensive, perhaps through some predictive modeling or screening.

Physicals conducted upon entering the Medicare program could also be a screening device. A physical would help find diabetics, for example, and people with certain "pre-existing conditions." Also, if we had appropriate modeling, we could pick up people with a combination of factors and predict that they are going to be severely ill. We could perhaps defer expenditures, if not eliminate them.

Revolving Door. There are pieces of this problem that could be addressed without having the complete system developed. One of the phenomena we noted in the CBO Medicare database is that we have a revolving door of sorts between nursing homes and hospitals. Elderly people in a nursing home get sick, are taken by ambulance to an emergency room, and checked into a hospital. Doctors find it better to treat them there. It is easier and they have the facilities they need. There is not a separate reimbursement for physicians to treat patients in nursing homes. Therefore, the whole system promotes moving them from a nursing home into a hospital. After some days of treatment, they go back to the nursing home.

It is not uncommon to see some of these people revolve in and out of nursing homes and hospitals a number of times in a given year. True, it is a small part of the high-cost population, but a manageable piece you could analyze and work on. Another piece of the puzzle is the treatment of people at the end of their lives. About 20 percent of Medicare expenditures every year are for people who do not live through that year. Examining how they are treated is another way of thinking about a subpopulation of these expensive patients. It is encouraging, although perhaps not surprising, that many of the elderly want to die not in a hospital, but at home. However, many still end up dying in a hospital. There may be a way to better facilitate their desires.

Finally, I want to mention one small silver lining to this Medicare cloud. If you accept the premise that we need to keep people out of hospitals, and could devise a strategy that would cut hospitalization by half, you might worry about the consequences for hospitals. Yet the coming retirement of my generation is going to double the Medicare population, leaving hospital utilization unaffected overall.