Passing the Buck on Medicaid:
How the States Pledge and the Feds Pay

August 5, 2003

Transcript prepared from a tape recording

9:30 a.m.	Registration
10:00	Presenter:	Michael S. Greve, AEI
		Jinney Smith, Northwestern University
	Discussants:	Dennis G. Smith, Department of Health and Human Services
		Alan R. Weil, The Urban Institute
	Moderator:	Robert B. Helms, AEI
Noon	Luncheon
12:35 p.m.	The Looming Crisis in Medicaid: One State's Perspective
	Nelson J. Sabatini, secretary of health & mental hygiene, State of Maryland
1:30	Adjournment

Proceedings:
MR. HELMS: Good morning. My name is Bob Helms. I welcome you to our periodic health policy discussions. Today we're going to be taking up Medicaid.

I was once introduced to a person named Dave Cooper over in HHS who was the department's Medicaid expert. He threw up his hands and said, "I know enough about Medicaid to know I'm not an expert." It was his opinion that the more you know about Medicaid, the more you know that it's very hard for anybody to understand. I must say, though, that I think we do have some people who might qualify as experts here today, and so, we're looking forward to a good discussion.

The title of the session is "Passing the Buck on Medicaid, How the States Pledge and the Feds Pay."

Now, let me start with just a few facts about Medicaid taken from the Kaiser Family Foundation's recent fact sheet. The program was established in 1965 as Title 19 of the Social Security Act. It provided health and long-term care for approximately 47 million Americans in 2002. Fifty-one percent of these are children, but they account for only about 17 percent of the expenditures.

The elderly, blind, and disabled together make up about 27 percent of the population covered by Medicaid, but they account for about 71 percent of the expenditures, and the Federal expenditures in 2002 were $147.3 billion.

The states add another $111 billion to that in 2002.

It is a program jointly financed by the Federal Government and the states, and it's operated through what's called a matching system of financing, and that's the subject for our talk today. Medicaid, as all of you know, was in the news, I'd have to say, more last year than it has been this year, although there are hints that things are always bubbling in the Medicaid issues.

Of course, the administration had a proposal last year, which I'm sure Dennis is going to tell us some more about later, and the governors made what many would call a valiant effort to come to some agreement about reforming Medicaid, but that kind of seems to have fallen apart.

So, now we have this situation where the expenditures are growing. It's a big concern to a lot of states out there in terms of trying -- who are under mandated balanced budget restrictions, to try to pay for this, and so, there's a lot of discussion in the Congress about what to do about this.

Today, we're going to start out with a talk based on a draft paper by Michael Greve and Jinney Smith.

Michael is with the American Enterprise Institute and runs our Federalism program here. Jinney Smith is a grad student at Northwestern University.

So, with that, I'd like to start with Michael.

MR. GREVE: Jinney and I are not economists, certainly not health care economists, and we're most emphatically not Medicaid experts. There are a lot of people in this room and on this panel who know vastly more about Medicaid than, frankly, I care to know. So, how did we get into this? I was writing a paper on what Congress was considering a few weeks or months ago, a bail-out for the states, and I wondered is that a good idea and a bad idea.

The upshot in the paper I wrote, which I think is available here, is it's a bad idea, but in that context, I took a look at the big Federal funding programs, education, and of course, the biggest one of them all, Medicaid.

I then asked our actual health care economists here at AEI whether they had any insights into this, and they all told me what Bob Helms just told you. It's too complicated. That's "A." And "B," it's too depressing. You don't want to go near it.

So, we dared to take a look at this program and take a whack at it, and from a very sort of political sciencey perspective. I'm not here to tell you whether we're spending too much or too little on Medicaid. Frankly, I have no idea. I'm not here to tell you that what the states and the Federal Government have been doing over the past decade or so to and in and under Medicaid is a good idea or a bad idea. Frankly, I have no idea. What motivated us and what this paper is about is how do we explain what individual states have done and how they behave? And the way you do that is you study variations among state responses or among state policy responses to identical or similar policy inputs.

So, the perspective is not this is one big Medicaid program and now let's look at the aggregate figures over time and look at the global trends, nor is it Medicaid is really 50 individual programs and now let's look at each one of them and see what constitutes best practices and what works and what doesn't work. Instead, what we want to do -- or what we wanted to do, at any rate -- is to look at the variations among states and to see what explains those variations and what, if anything, explains a common pattern among those states, and that way, you get a firmer grasp -- or so we hope and so political scientists generally hope -- on what drives political decision-making.

Now, you start at the end or you start with the observation that Bob mentioned at the outset. Medicaid has grown stupendously and continues to do so. It's grown faster than Medicare and now exceeds -- total spending on Medicaid for the first time exceeded Medicare, and there's a spending crisis in the states, and what the states say is, look, due to the recession or the -- in any event, the lagging economy -- there are more people eligible for Medicaid.

There's also continued health care cost inflation, ever more expensive procedures, more expensive pharmaceutical products, more expensive long-term care at the end of people's lives, and there is, of course, less revenue in the states to cope with all of them, and you want to see whether -- the first thing you want to see is whether that is, in fact, an accurate representation of Medicaid and what caused the spending crisis.

And so, the first cut we took is you look at the aggregate figures for Medicaid and you don't look at it over the past three years but you look at it over the period of the boom, which is the boom in state revenues and in the economy in general, which lasted, roughly speaking, from 1993 to 2001.

We have to lop off the first year, 1993, and the last year, 2001, because we couldn't get continuous data series for those years.

So, the core, with the exception of those two years, is the picture for the boom years, 1994 to 2000, and by the way, if this is not readily readable or legible for television viewers, the paper and all these complicated charts and graphs are available at either AEI.org or federalismproject.org.

These numbers are inflation-adjusted. So, what happened over that boom period was that the GDP went up by 20 -- over 20 percent, on a per capita basis by 14 percent, state revenues exceeded that growth by a considerable margin, the poverty population in the United States fell by 18 percent, the poverty rate fell by almost 15 percent, at the same time Medicaid expenditures -- that is to say, for state and Federal spending combined -- grew by an astounding 30.8 percent; on a per capita, per United States citizen basis, by over 23 percent.

Of course, if there is any time at all that Medicaid expenditures should fall, it's during a boom time like that. People get jobs. Some of these jobs have come with health insurance, believe it or not. So, that ought to reduce the population. Other people simply get wealthier and therefore exceed the eligibility criteria in the states. So, by and large, unless somebody decides otherwise, Medicaid expenditures should fall, but they didn't. They grew stupendously in boom time.

Now, the next question is, is that due to some health care inflation, a very problematic term, but I'll skip the complications for the purposes at hand, and we'll look at the next chart that looks at various indicators of health expenses and their change over time, and what you see there is, again, the GDP figure, and you see that aggregate Medicaid spending exceeded the growth or grew faster than any other health indicator, national health care spending, or for that matter, Medicare spending. On a per-consumer basis, Medicaid spending grew much more slowly than anything else, and that's a very interesting finding.

You don't want to make too much or infer too much from those drastic figures. So, this very low percent there, 7.5 percent expenditure growth per Medicaid consumer, could be attributable to a number of things. It could be that the Medicaid system is so big that it has an easier time to keep expenditures under control. Also, this period, 1994 to -- the tail-end of the 1994 to 2000 period coincides with the creation of the SCHIP program, which got a lot of kids onto the Medicaid rolls, eventually, and children, healthy children, at any rate, are generally low-cost medical consumers.

So, a lot of things could be going on there, but what the chart is consistent with is a rapid eligibility or at least coverage expansion. So, the growth in Medicaid, the disproportionate growth in Medicaid, seems to be due more to a growing population rather than something that's inherent in sort of health care expenditures, general consumption trends or cost trends in the health industry. And those decisions, if the population grew in times of rapidly rising wealth and falling poverty, those must be the states' decisions. There were very few things, if any, that the Federal Government did over this period to increase the population.What the Federal Government did was issue expansive waivers for states to cover additional Medicaid population in those states.

Now, in order to get a better grasp of this, you have to, as I said at the outset, look beyond the aggregate numbers and look at state-by-state variations, and that's what we're going to do now, in the next chart. What we did was we plotted the change in total Medicaid expenditures in each given state -- that's the vertical axis -- against the change in the poverty rate in each state, and there are a number of things that strike you.

First, the slightly ascending line there is called a regression line. What it means is that, all things equal, higher poverty rates correlate with higher Medicaid expenditures. No surprise there. That's not the interesting finding. There are two things, I think, that merit comment in this particular chart. First is the enormous dispersion along both dimensions. So, you find, on the one hand, if this works -- there -- states like Missouri or Minnesota, where the poverty rate dropped by over 50 percent. That sort of thing makes you think that maybe the son of man had it wrong, the poor will not always be with us, because half of them disappeared in six years.

At the other end, you find a state like Governor Dean's Vermont, where, in a period where poverty fell in almost all states, it increased by over 50 percent in the State of Vermont. Along the vertical axis -- that is, the growth in -- or decline, as it were -- in Medicaid expenditures, you find a similar dispersion. You find states like New Mexico, Oregon, Tennessee, all of them with Medicaid growth rates of over 60 percent compared to states -- two states, New Hampshire and Louisiana -- which registered very substantial Medicaid expenditure reductions. In Louisiana, the decline there is most likely due to statutory changes. States had figured out a lot of ways of maximizing their Medicaid take, so to speak. Congress caught onto that. This was particularly pronounced in Louisiana, and after Congress made those changes, Louisiana simply got less Federal money and had to figure out a way to get its expenditures under control.

Louisiana is an outlier in a lot of ways. It has one of the highest per capita medical expenditures in the country and one of the lowest levels of health. The most likely reason is that, in Louisiana, political corruption is not tolerated, it's mandatory. But in New Hampshire, the reductions are very genuine. So, you have this huge dispersion along both axes -- very, very dramatic differences among the states.

The second thing is the heavy -- that's noteworthy, I think -- is the heavy concentration of states in this quadrant here, number one. That is to say -- that's the states where poverty rates and poverty populations fell, but Medicaid expenditures, nonetheless, increased. As I said earlier, if there is any period at all when you should expect Medicaid expenditures to fall, that's when incomes are rising dramatically and poverty is falling but in fact, there's only one state, New Hampshire, that unambiguously pocketed a declining poverty windfall, if you want to call it like that.

By the way -- and this sort of semi-facetious -- the chart is consistent with the observation that all Medicaid-eligibles or people who might be Medicaid-eligibles migrated from New Hampshire there, which is stingy, into Vermont, which is not. I can't prove that that is what happened, but it's consistent with the chart.

Look at the next chart very briefly. We just wanted to make sure that we hadn't fooled ourselves. It's the same chart, but with the absolute numbers. So, this is the change of Medicaid expenditures per capita in the state. This is the change in the poverty population, rather than the poverty rate. It's substantially the same picture and nothing new emerges.

Now, how does all of this happen? How is it that poverty rates decline and, at the same time, the Medicaid expenditures go up? Let's look at the next one. This looks totally confusing, but in fact, displays substantially the same information that you just saw in a slightly different form. What we did was we grouped the states into four categories, according to the poverty declines. So, here's a group of states, 10 states, where poverty fell more than 10 percent, more than 30 percent. Here is a group of 11 states where poverty fell more than 20 percent but less than 30. Here are the states with slightly declining poverty, and here are 15 states with increasing poverty rates. Okay?

Now, it's called a box plot. What the lines here -- the outer lines -- indicate is the entire range of responses, policy responses, the entire range of states in that category, and here, again, is the change in state Medicaid expenditures. There was one state where Medicaid expenditures increased 80 percent while poverty increased. There was one where it increased only about 8 percent or so, the same year. The little boxes here show the middle 50 percent, the middle 50 percent of the states in each category, and this line here, the fat line, is the average for the states in those categories.

Now, two things, again, about this graph. The first is that, in states where poverty declined only slightly or even increased, the range is very, very wide and these boxes are very, very big. In other words, what you get in those states with slightly -- only slightly declining poverty or increasing poverty is a wide range of policy responses, and what that says -- or a cautious interpretation, at least -- is that these states simply faced very different policy choices. They could either say, gee, I mean poverty is increasing and that means our Medicaid expenditures are increasing in any event and we ought to be circumspect here. That's the states at the bottom end of these ranges. Or they said, no, look, it's a general era of prosperity across the nation, we should do more for our poor and uninsured, and so, if the poverty rates then increase or the absolute numbers of the poor increase in your state, then you confront very, very dramatic expenditure increases. The range here in these states where poverty fell dramatically, or almost as dramatically, is much, much narrower. In other words, what that means is those states all responded in a fairly similar fashion to the decline in poverty in their states.

Now, second thing. Even though these states here where poverty fell so sharply responded fairly similarly, expenditures still rose, right? That is to say, no matter what happened to poverty in these states over that boom period, on average Medicaid expenditure growth refused to go under 20 percent, and in fact, for the states with the most dramatic decreases in poverty, the expenditures rose more sharply than for either of the next groups, right? So, at some point, it seems to be the case that further declines in poverty cease to produce commensurate reductions, so to speak -- this is a very loose translation -- in Medicaid spending. Somehow, where poverty falls most sharply, it's not the case that, all things equal, the expenditures go down in proportion.

Who are these states? Well, one interesting fact about them. With the exception of Connecticut, they're all contiguous. If you draw a line northward from Mississippi, you catch them all. They're mostly rural states, and politically, rather diverse. They range from Minnesota to Mississippi, and that's basically the entire range of the country. I don't know why that is, and we refuse to speculate about it. Instead, we asked ourselves the following question. Can one tell a story that's broadly consistent with this kind of evidence? And here is our story, and again, it's a plausible story, it's not a proven explanation.

It's reasonable to assume that state legislators and governors look at Medicaid spending or Medicaid programs as kind of a floor. So, let's say, in year one, you're a state legislator, and you say to yourself, gee, look, Medicaid is now 18 percent, 20 percent of our budget. By and large, we can live with that. Now, we predict a rising tide, an economic boom, rising revenues. That means, by and large, the eligible population will decline. That signals declines in Medicaid expenditures unless we do something about it. So, why don't we use the money that's being freed up by these declining populations to expand eligibility or to improve coverage to make sure that everybody who is eligible for Medicaid, in fact, gets onto the rolls and is covered? Let's keep that in an affordable range, perhaps in line with our revenues, which we, of course, also expect to expand.

So, in other words, if that's the story, if that's how legislators think, if they don't sit around at all times and see how can we squeeze Medicaid expenditures further but, rather, think of it as, well, we can live with what it is now, it's high, but we can live with it, so let's keep it on an even keel, there's a substitution effect. As the poverty population declines, states being to expand eligibility. That's consistent with this pattern and with one further piece of evidence, which I'll show you now, or will show you now, and that's the next chart.

This is a regression analysis, and I'll spare you all the razzmatazz, but what this number means here is that the simple regression explains roughly two-thirds of the observe variation among the states, and what we plugged into the regression is the FMAP [ph] -- that is the Federal matching formula -- the growth of that formula over the years, the poverty growth, the revenue growth, and Medicaid growth as a percent of state revenue during the early years, during the first three in-years of the 1994 to 2000 period. The theory here is we want to know how state legislators think about Medicaid spending in future terms, and the Medicaid expenditures have a lag effect. Decisions you make in year one have an effect in year two, three, four, and so on out, and in order to mimic that effect in real -- I mean to mimic that real world situation in the data, we looked at the revenue growth over the first three years, Medicaid growth as a percent of state revenue during the early years, and figured out, does this explain Medicaid expenditure growth over the entire period, and lo and behold, it turns out that the answer is yes.

Those four factors here -- FMAP growth, poverty growth, revenue growth, and Medicaid growth in the early years -- explain two-thirds of the variation among the states, and these two factors, revenue growth in the early years and Medicaid growth in the early years, explain an awful lot. In other words, if you're worried about Medicaid cost control, the absolute worst thing that can happen to you is a rising state budget when the state says, under those circumstances, we'll make a slight increase in our commitment to Medicaid. What's wrong with this? If we're right about that interpretation -- and again, it's a plausible interpretation, not a proven explanation -- what's wrong with that kind of state behavior, to say that, look, it's a rising tide, let's lift all boats, as the poor decline, we'll give more -- we'll expand eligibility? Well, there are two things wrong with it.

First, it's easier to overshoot than to undershoot, and that's the next figure we'll show you. Between those two lines here, there are a bunch of states where the share of Medicaid as percent of state revenue stayed more or less even over the period. There are only five states where the Medicaid share of the state budget declined over the period. That includes the Louisiana abnormality. It includes California, which had just stupendous revenue growth over this period; New Hampshire, as explained, and Arizona and Nebraska. A third of all states, 16 states, overshot that target. That is to say, in 16 states, even in this period of rapidly increasing state revenues, in 16 states, Medicaid continued to chew up an ever larger share of the state budget. So, over-commitment and overshooting is one risk of the substitution philosophy.

And the second thing is -- the second problem with it -- and mercifully, that was the last chart. I don't have a chart for the following observation. If you think about Medicaid in those terms that I just tried to explain, you get a ratchet effect. It's understandable for a state legislature to say a rising tide should lift all boats. It's equally understandable for a state legislature to say, hey, an increase in Medicaid coverage is cheaper than doing anything else in the state budget, because it costs us, at most, 45 cents on the dollar, or 50 cents, in some states, on the dollar. The trouble, of course, is that revenues can't grow forever. Every one in a while, there's a slow-down, and now you face those problems in reverse.

Nobody believes that a falling tide should sink all boats, and so, Medicaid expenditures -- there will be some cuts, as there are now, but by and large, there will be fewer cuts than there were increases during this boom period, and second of all, a cut in Medicaid, all of the sudden, looks like an extremely expensive dollar, because every cut you make loses you, quote/unquote, Federal funds, and that's a problem.

So, what did we conclude from all of this? The first thing we conclude from all of this is that state decisions, independent state decisions, more than any exogenous factor, have driven Medicaid spending over the period we examined. The second thing we can see is that Medicaid will grow in bad times and in good times, and it will continue to do that so long as the shared Federal-state arrangement remains in place. Third -- and this goes to my initial observation. I can't tell you whether this is good or bad. If you want health care socialization -- that is to say, a move to a single-payer system, eventually -- the incentives for the states are optimal here, and the states behave as intended. What this is, in fact, is Hillary care on Federally subsidized wheels. On the other hand, if you're worried about trends towards -- in that direction and if you're worried about the aggregate expenditures here, you ought to be very, very worried, and the absolute last thing you want to do in that situation is to bail the states out of their periodic funding crises, because that's the only discipline on the process that otherwise cuts at all times in the opposite direction.

MR. HELMS: Okay. Thank you very much, Michael.

We now go to our discussants, as I say, the real experts here. Michael said he wasn't an expert.

We welcome Dennis Smith. Dennis has actually been with the department for a long time, off and on, I guess, but now he's a director of the Center for Medicaid and State Operations at the Center for Medicare and Medicaid Services.

Now, what that means is he heads the office that runs the Medicaid -- most directly concerned with running the Medicaid program and the old HCFA, or the agency that runs both Medicare and Medicaid over at the Department of Health and Human Services.

Dennis?

MR. SMITH: I was very interested in the paper that has been written, and since we've been, the last few minutes, looking at the last six years, maybe I'll shift gears and talk a little bit about what we were looking at in the coming 10 years and serving as the rationale for the administration's proposal on Medicaid this past year.

As was mentioned, Medicaid is now spending more dollars, when you combine Federal and state resources, than the Medicare program. Medicaid is growing faster than it. A large part of what we've been trying to do is really emphasize that Medicaid really serves three very distinct populations and really is two very distinct programs. What I mean by that is, as Dr. Helms mentioned, if Medicaid were only the program that it was originally conceived as, which was to serve low-income families on welfare and giving them access to health insurance, it would be a very different program than it is today. It would be a very much smaller program than it is today.

Less than a third of the kids on Medicaid are receiving cash assistance, and as Bob mentioned, kids make up 51 percent of the Medicaid population but only account for about 15 percent of the expenditures. Then when you add in their family -- other family members, parents and adult caretakers, we're really talking about 25 percent of Medicaid expenditures going to what we consider to be health insurance for low-income individuals. Medicaid really serves that population, but it is an important program for low-income senior citizens, as well. Medicaid has given senior citizens access to insurance coverage that Medicare has not provided for them in the past few years, most importantly prescription drugs, and obviously, now, Washington is addressing the prescription drug issues for senior citizens, and that will be an important consideration, and whatever happens in Medicare will have an impact on the Medicaid population.

But you have about 4.3 million senior citizens who are relying on Medicaid for prescription drugs or home health or their long-term care nursing home stays. While you have moms and kids, the traditional Medicaid population generally receiving acute care, and senior citizens generally using long-term care, in the middle you have 7 1/2 million people with disabilities who are reliant on the Medicaid program for both acute care and long-term care, people with disabilities who would not have access to private health insurance even if they could afford it, in many respects. So, Medicaid is an important safety net program for this population, as well, and those who rely on Medicaid more and more and you look at a lot of the growth in the program, is in the area of home and community-based services, and when you look at the growth in the coming years, home and community-based services and long-term care are important drivers of the Medicaid program.

Really, what our proposal was about was to look at Medicaid as those distinct -- serving distinct populations and being distinct programs, and the proposal that we came up with really was designed to shine a big spotlight on the long-term care side of the program. As I said, long-term care is going to grow even -- is growing even faster than the acute care side of the program, but we have decision-makings about -- in terms of states holding down the rate of growth in the Medicaid program, presenting new models for delivering services in long-term care. We think that this is really a very important area that, as states develop new systems for certain -- develop new models for serving people in long-term care and pick up on the themes that we have seen through waivers in Florida and Arkansas, for example, on cash and counseling, and giving people with disabilities greater control and, really, a role that they have not had in decision-making about what services they are going to purchase and who are they going to purchase from. We believe that that is -- when you look at states like the State of Washington, for example, that has moved aggressively towards home and community-based services, when you compare their rate of growth in long-term care to be much lower than what it would have been if those individuals had been served in institutions.

So, we believe that putting a big emphasis on the future in terms of spotlighting what Medicaid does on the long-term care side and giving the states the -- really, the incentive to come up with new models, which we believe that lead to greater satisfaction for the individuals being served and will lower that rate of growth that we're seeing in the Medicaid program. Right now, long-term care, Medicaid is spending about $100 billion on long-term care. By 2013, we're going to spend $225 billion on long-term care. On the acute care side, we believe that the model for future Medicaid reform is already there, in the State Children's Health Insurance Program that was created by Congress in 1997, and if you ask people what is one of the most successful programs in recent history, I think SCHIP would be very high on everybody's list in how that program is working, how it's been accepted, and the things that are built into the SCHIP program, and as a model, it dealt with capped allotments to the states, a great deal of flexibility to the states, the ability of the states to really look -- understand that much of the growth in the Medicaid population is not just families on welfare but families who are in the workforce. They are still low-income families, but many of them do have access to employer-sponsored health insurance.

The Medicaid program, though, is still running under the same rules that it was designed for them, designed for serving a population that was still virtually all of the welfare population. So, things like retroactive eligibility, for example -- you would never find that in the private sector. You don't find it in Medicare. A lot of these things that are built into the Medicaid program that reflect serving a population that is the welfare population. But indeed, the mission of Medicaid has grown and changed over the years, and part of this was about recognizing that change and to build into the program greater flexibility so states can make adjustments. SCHIP had great flexibility in terms of benefit package design, and again, I think that when you -- whether -- whomever you ask about SCHIP, part of that flexibility leads to the success of the program. States, though, responded very favorably to that flexibility, and I think most people would agree that the benefit packages that states have designed under SCHIP are very good benefit packages and are very realistic benefit packages. So, the flexibility that has been given to the states -- we believe that the states themselves, in looking at SCHIP, will tell us a lot about what they would do with the flexibility in the future under our proposal.

I think that a lot of the things that has been mentioned here this morning -- obviously, the state decision-making does drive Medicaid spending. Medicaid is a matching program, and the program is reliant on the state putting their dollar up first for -- and then claim a matching dollar. Part of what was mentioned this morning was, when you do look over the timeframe that was being discussed, oftentimes states were not putting up their full dollar but, in fact, finding ways to shift the cost of their decision-making to the Federal Government. We think that is only fair to bring that back into balance, and you saw it towards -- Congress responded to that in 1997 in terms the Balanced Budget Act and hard-wired a disproportionate share of payments, for example. Then at the end of the Clinton administration, Congress dealt with the upper payment limit financing mechanism, and during this administration, we have tightened that down even further.

Again, I think that one of the things that we need to agree on, as long as Medicaid is a matching program and is reliant on the balance of having that match, that all parties need to agree that the match needs to be real then and should not simply be shifted to the Federal Government. Oftentimes, the state has served in a state Medicaid program and know the discussions that go on in terms of Medicaid, and oftentimes, the rationale is to say, look, we are providing this service or this benefit, and right now, we're doing it with 100-percent state dollars. Since we're putting up money anyway, can't we find some way to get the Federal Government to participate in the cost of delivering that service, whatever that might be. So, in many respects, it's a very rational decision for state legislators to make.

Medicaid is a vitally important program and one that I think Medicaid reform is a matter of when, not if. I think when you look at the growth rates that are being projected for the program and look at the growth rates for state revenues, compare the Medicaid growth rate to other important priorities, and it is an issue about decision-making. You can only spend a dollar once. Where will you spend it? Medicaid, I think, has done very well in terms of -- in competing for that public dollar, both in good times and times now when they are rougher, but it does take a bit of a balancing act. I think that, again, states are making decisions to say, okay, do we expand to a higher -- do we expand into a higher income level, or do we spend that dollar on problems in our foster care system?

Do we spend that dollar on our education system?

Do we spend that dollar someplace else?

So, again, I think part of it, part of what goes on, oftentimes, is part of the thinking that going to higher poverty levels or increasing reimbursement rates to providers or whatever, that decision is, well, whatever dollar goes into Medicaid, de facto, is of greater importance than anything else. I think states are -- again, what is important to remember is, putting that dollar anywhere, you are competing against the dollars for the child welfare system, for foster care, for child care, etcetera, and what states did is they shifted the cost over to the Federal Government. We're saying, well, the decision can be made different now, because the Federal Government is now putting up a greater share of the cost of the program. To put that back into balance, again, I think what we did was look at the successful models of SCHIP. We looked at the future demands on the program and built into it, I think, some very realistic trend rates of what we were offering to the states in maintaining that Federal commitment to the Medicaid program.

Under the President's budget for the baseline that was built for 2004, the Federal Government was going to spend $2.7 trillion on Medicaid over the next 10 years. Under the President's proposal, we were still going to spend $2.7 trillion. So, the President's proposal was about maintaining the Federal Government's commitment to the Medicaid program. We were protecting every dollar that is projected to be spent in the Medicaid program.

There are service models out there that will improve the way services are delivered to individuals and help lower the rate of cost in the program.

I think that, again, when you look at Medicaid, you -- one of the most important things we have done this year, and I think we will continue to do, is participate in forums like this and really help explain what is in the Medicaid program, how Medicaid is serving different populations, whether it's the dual-eligibles, or talking about things like the long-term growth in -- the growth in long-term care and coming up with more models. We think that the models are out there. We have seen success in home and community-based waivers. We've seen success in cash and counseling, in getting greater involvement of the recipient to participate in decision-making. All of those things are important to do, and it won't be one thing or just two things. It will be a whole series of things, but each one of them are very, very important to bring into the program.

Let me stop there and again just -- I think that what we are looking at, the cost -- the future cost of the program, Medicaid must be looked at, but there's great -- this is not a situation to where -- you know, again, I think the way we can address the rise in the cost of the program, though -- we can meet with some strategies that are helpful not only to the program but to the beneficiary itself and improve the way services are delivered. When you look at cash and counseling, for example, you have extremely high satisfaction rates for the people who are doing that, who are participating in cash and counseling, that have improved quality, improved quality in terms of expanding the choices that they have and expanding access to care.

So, sort of beware of kind of, when Medicaid is sort of presented in either/or situations, don't get caught up in sort of the false choices that sometimes get presented. There are ways to improve services and, in the long run, help lower the rate of growth in the program, as well.

MR. HELMS: Before I go to Alan, I'd like to -- I'm sure this Washington crowd understands every term you used, but there may be some people watching on CSPAN that -- I just would like to ask you to clarify two statements that you made. One was hard-wiring the disproportionate share of payments, just to -- not getting too technical, but just briefly, and the other one was about establishing the expenditure caps. Could you just give a brief explanation of those two things for our own --

MR. SMITH: On disproportionate share, these are payments that Medicaid makes to hospitals that serve a higher proportion of Medicaid and low-income recipients. This program had been used by the states to shift a greater share of the costs to the Federal Government by -- because there were not firm rules about how much money could be claimed by the state and by hospitals. What I meant by hard-wiring was the statute defined by state how much money each state could receive under disproportionate share. So, Congress actually assigned a dollar amount to each state. Now, part of that was criticized. Some states that said, well, gee, we hadn't gotten into the DSH game yet, and so, now you have locked us into lower amounts of funding, but hard-wiring dealt with payments to hospitals, which -- DSH payments is about $18 billion of the Medicaid program.

MR. HELMS: I have some figures here that, in 2001, the disproportionate share payments was 7.1 percent of the Medicaid payments, and just for comparison, in-patient care was 32 percent, pharmaceuticals were 9.1 percent.

And the expenditure caps, very briefly?

MR. SMITH: Yes, sir.

The expenditure caps -- as I said, we started out with, again, the commitment that the Federal Government -- we did not propose this to save the Federal Government money. We put in -- that $2.7 trillion -- into the program that we were going to spend under current law. Then we basically came up with a couple of different models that we were using, again recognizing that states have a great deal of flexibility under -- and have expanded their program well beyond what the Federal Government requires the states to spend.

Again, if Medicaid were only the program that the Federal Government requires, we would only be -- Medicaid would be spending only about a third of the current expenditures. Only about a third of Medicaid expenditures are there, because the Federal Government requires it, whether it's the population that is eligible or the services that are being required. The states have expanded eligibility beyond what the Federal Government requires. It has expanded services beyond what the Federal Government requires. So, again, we recognize that if the states -- we were trying to do a combination of things: one, preserve the gains that had been made in Medicaid.

We did not want to see simply that the response of saying the states, with their budget situations, of just starting to cut back on Medicaid eligibility and Medicaid benefits, which, indeed, we have seen happen over the past year. So, we were trying to stabilize the program itself. So, we offered to the states additional money above the Federal baseline to help stabilize the program, but then we looked at the annual growth rates that the states were looking at, again looking at the distinction between mandatory spending and optional spending, telling the states that we would continue to guarantee the mandatory side of the program, that the states would not be able to reduce eligibility for the mandatory population and would not reduce mandatory spending, but find greater flexibility on the optional side of the program and build in trend rates that we thought -- again, the first seven years of the administration's proposal, the Federal Government would have spent higher than what was in the baseline, and then we would have adjusted in the last year.

MR. HELMS: Our next discussant is Alan Weil, and he's the director of the New Federalism Project at the Urban Institute. As many of you know, there are numerous think tanks in Washington. One of the most active in health care is the Urban Institute. I've been involved with Alan in a sort of health policy discussion group over the last several years and have always found him to be, one, very knowledgeable, but I think he's going to take a different view of some of this. That's my hope, anyway. Let me add, he not only brings an academic interest to this but also has some very practical experience in running a Medicaid program in the State of Colorado.

MR. WEIL: I really appreciate the invitation to be here this morning. I particularly appreciate it because when I'm done, I have a feeling I won't be invited back. Generally, when I'm asked to come in on a paper, I try to sort of summarize the things I like about it and the things I don't, but to be candid, I find it very difficult to find anything that I like in this paper, and I realize that the authors have confessed to not being Medicaid experts and to not being health economists and the paper is marked "draft," but I think it would be inappropriate of me to not say that broad circulation of this paper in its current form borders on being irresponsible, because it makes such a misuse of the data.

So, what I am going to do is talk about what I consider to be the five most egregious analytic errors in the paper and leave it to you to judge whether or not there is anything to be done about those errors and then spend a few minutes talking about the conclusions that the authors reach.

Fortunately, I will say that, in the presentation today, some of these errors were acknowledged, although they were not in the draft that I read, and that gives me some comfort, but it still doesn't take away from my overall conclusion that there's really no analytic contribution provided by the data presented in this paper.

The first error -- and it is, in my mind, the most important error -- is that the paper uses as its unit of analysis total Medicaid spending at the state level.

Now, as Bob mentioned in the introduction three-quarters of spending in Medicaid is for the elderly and disabled; only a quarter for low-income families, adults and children.

Therefore, the fundamental premise of the paper, without which there would be no paper, which is that Medicaid spending should be cyclical, is simply wrong, and anyone who looks at the Medicaid data and looks at where the dollars in the program go would not actually begin with the premise that Medicaid spending should be cyclical.

Medicaid spending is driven far more by demography than it is by cyclicality, and therefore, the inability to explain state-level Medicaid spending through economic terms should actually come as no surprise at all, and as a corollary, the choice of using either total Medicaid spending or Medicaid spending per capita, as a unit that tells us, really, anything about the program, is erroneous, and so, the fundamental premise of the paper, it just seems to me, can't be supported, and therefore, it's hard to take any steps beyond that first point.

But I said I had five, so I guess I'd better go through the other four.

The second criticism is that there is a regression equation presented, and some time is spent on it in the paper, but if you look at the regression model, Medicaid spending is on one side of the equation and the share of the state's budget that is Medicaid spending is on the other side of the equation. You just can't do that in statistics. You can't put the same variable on both sides of the equation. You do get a good high R squared. It says that two-thirds of the variability is explained by this model. But you can't interpret the results. They have really fundamentally no meaning. And so, I would recommend that you take the regression results in the paper and just sort of put an "X" through them and understand that, for sort of basic statistical reasons, that model can't be used to explain anything.

Third, the authors acknowledge that they're examining a period of unprecedented economic growth, and I think that that's -- it is a happenstance, I think, of the data set that they used, but I would be very cautious about generalizing from that period, for two reasons. One is they don't look year to year. They just look 1994-2000, as if the whole six years is one undifferentiated period. 1994 is an odd starting year. For those of who may be unlike the authors who have the misfortune or pleasure, depending on your perspective, of working on Medicaid for a while, 1994 had a couple of oddities about it. One is that it was right in the middle of big debates over the disproportionate share hospital program, and so, state spending on DSH, which is included in the numbers the authors use, is somewhat unusual for that year.

In addition, the last debates, not the current ones that this administration has proposed, but the prior debates over Medicaid block grants proposed using 1994 as one of the possible base years for setting the states block grants. States had tremendous incentives to try to find accounting mechanisms to build dollars into their '94 base, and I don't assume they were all equally able or interested in doing so. So, it's a difficult year to start with, but more important, using only a period of economic growth, particularly after a period of some economic downturn, highly skews what you observe in the Medicaid program. Let me give you a simple example.

When the economy is bad and when state budgets are not in particularly good shape, trying to preserve eligibility in the Medicaid program, states often do what they find easiest, which is to cut back or at least not give inflationary adjustments to the providers who deliver services to the people in the Medicaid program, and so, coming into the period that's analyzed, states had, I would hypothesize, been holding down their reimbursement rates for providers because they were trying to deal with very high rates of cost inflation. When the economy strengthens, because Medicaid is a chronic under-payer for services, the first thing that the providers do is they go to the legislature and say, you know, we haven't even gotten an inflation adjustment in the last few years, can't you at least give us that, and it's a dynamic I've seen played out in quite a few states. That would lead states to have the Medicaid program grow faster during a period of economic growth than it would during a period of downturn, at least on a per capita payment rate. That, it seems to me, at least is a testable hypothesis, but most important, it just, in my mind, suggests that if you want to really look at the role that state choices play in Medicaid spending, you can't just examine a period of economic growth.

Fourth -- and this point, again, thankfully, was made in the verbal presentation, although it is not mentioned in the paper -- the heterogeneity of the Medicaid case load does yield some anomalies with respect to looking at changes in per capita costs, and again, this was noted in the verbal presentation, and as was noted, the period analyzed includes the advent of the State Children's Health Insurance Program, which many states used as a foundation for outreach that then brought a lot more Medicaid-eligible kids into their program. In the average state, an elderly enrollee in Medicaid costs about 10 times as much as a child on Medicaid.

So, if you are doing outreach and bringing children into your Medicaid program, your per capita costs for the program as a whole are going to go down, not because your per person costs have gone down, but you've simply brought cheaper people into the program at a faster rate than more expensive people. Now, this is a fairly basic algebraic phenomenon, but why it's important in thinking about the results of this paper is that the authors are using the changes in per capita spending as sort of the thing to look at to try to figure out what's going on here, and when you have changes in the composition of the enrollees, per capital expenditures simply capture too many different things, and they're going to be dominated in many states by changes in the composition of the case load, as opposed to the per-person costs of the case load. I would also note that the period -- because it's early in the development of the SCHIP program, states had very different rates in their ramp-up of this program, and so, you would expect to see very different changes across states in the composition of the case load, which would have very different effects on per capita costs.

Finally, I would just make a point about the general structure of the argument in the paper. The argument is we have a phenomenon. We want to try to understand it. We're going to look at a number of hypotheses that might explain it, and lo and behold, none of the hypotheses seem to explain it, and therefore, we're left only with the un-analyzed hypothesis, which is that state policy must be the reason for cost growth in the Medicaid program. This is sort of an analytic trick that I don't think you should accept. Using a fairly simple data set, testing a series of fairly simple hypotheses, rejecting them one after the other, and then saying the one that we can't really analyze must be the one that is the explanation for the phenomenon -- it's just not really an acceptable -- in my view, it's not really an acceptable approach to analysis.

Take the example of the role of health care inflation. The general statement, as I understand it, in the paper is that because state costs varied so much from state to state, it can't have been health care inflation, because you would expect health care inflation to be about the same from state to state. Aside from the fact that that hypothesis is really never tested, and actually, if you look at the health care literature, you do find that there is tremendous variation across states in the adoption of new technologies that are expensive, it is simply not okay, in my mind, to line up a series of hypotheses, reject them, and say that the one that was not tested therefore must be the explanation.

So, I want to be careful how I describe my sense of this paper. It is not that I want to sit here as a researcher and take a few detailed points and say, as can be said with any piece of research, you could have done better here, you could have done better there, why didn't you look at this. As I read the paper, it's very hard for me to feel like there is any way to take the analysis that was conducted and use it to reach any policy decisions. That said, let me just spend a moment on the conclusions the authors reached, which actually maybe somewhat surprisingly, I don't disagree with so much, because they don't really draw from the data in the paper, but I have a somewhat different take on them.

The three conclusions that were just noted is, first of all, that state decisions have driven Medicaid spending. At one level, this is clearly true, because as Dennis noted, a small share of the Medicaid program is mandatory eligibles being covered for mandatory services, and states are making choices all the time, but I don't consider something like health care inflation, which is a major driver of Medicaid spending, to be under state control, and so, I think it's somewhat over-simplified to say that state decisions are responsible. Now, clearly, if states wanted to go out and make major cuts in their Medicaid program, some of them could, because they're spending beyond the base of coverage, but we would be talking about taking, for example, an 85-year-old widow who has a pension income of, say, $10,000 a year and needs to be in a nursing home and taking that person out of a nursing home or eliminating the nursing home benefit. That is an optional population, but I don't see states going backwards on coverage like that.

Second of all is the conclusion that Medicaid will grow in good and bad times so long as funding is shared. I basically think this is correct. This is a basic principle of the Medicaid program that states do not have to bear the full marginal cost of increased spending and they don't collect the full marginal cost of cuts in the program, and therefore, yes, the funding structure of the program is inherently expansionary. I don't think that's a bad thing, although I think others do, but I also don't think it's very controversial. Clearly, if states had to bear the full cost of all of their decisions, they would spend less. What I find interesting is how many states go no further than the minimum they're required to and a few additional basic populations that are not mandatory but are fairly obvious. Most states actually don't go very far beyond that floor. So, the notion that Medicaid costs will grow in good and bad times, I think is probably accurate.

The final claim is that, if you want a single-payer system, the incentives of the current system are optimal. I can't possibly see evidence for that conclusion in the data of the paper or my understanding of the Medicaid program. If you wanted to move to a single-payer system, you wouldn't leave major coverage decisions up to the states when states are clearly not in as good a financial position to spend dollars as the Federal Government. Their revenue base is not as broad as the Federal Government. They worry about interstate competition. The comment about New Hampshire and Vermont was made. And there are a large number of reasons that if you really thought you were moving to a single-payer system, you would not design the Medicaid program at all like this.

So, let me just conclude with two points. The first is what I find most frustrating in the paper is that it would be possible to look at the role that state policy plays in the cost growth in the Medicaid program. You could look year by year, instead of using six-year summary data. You could focus on the populations that you would expect to be cyclical, moms and kids, and not look at the entire program. You could actually look at state policy choices with respect to eligibility and payment rates, as opposed to just aggregating into one number which captures everything. Then you actually would have a hypothesis you could test. But that is not done, and so, it's very frustrating to me as a reader to see the gap between what the authors purport to want to do and what they've actually done, and the data exists to do what they purport to want to do, so it's hard to know why that was not done.

The final point I would make is just a very simple, as one who has worked in the Medicaid program for a long time. I really won't engage with the administration proposal to convert Medicaid into a block grant. That's a discussion, I believe, for another day. But what I find missing from the analysis here is the notion that there are actually people who are served by the Medicaid program. There are actually millions, tens of millions of Americans who would otherwise be uninsured. There are millions of people with chronic mental illness, with HIV and AIDS, with developmental disabilities, with spinal cord injuries, with degenerative muscular conditions. There are frail elderly with no one in the community or at home to take care of them, and their needs are met by this program, and they are not met perfectly and they could be met better, but it really is a value choice about how many of those needs we are going to choose to meet as a society, and to treat the Medicaid issue as a money issue without looking at the reasons and the very, in my mind, legitimate reasons people believe this program should be funded is looking at only one-half of the equation. And so, the approach taken, although we do a lot of econometric work at the Urban Institute and I don't challenge the notion that econometric analysis is a useful tool in thinking about Medicaid, if you want to actually reach some policy implications about the program, I think you have to look at the benefits side, as well, and not just the cost.

MR. HELMS: Do you want to respond now, or should we go to the audience?

MR. GREVE: Jinney Smith has some responses. I have some, too.

MS. SMITH: I'd like to begin -- well, the one, in particular, that I'd like to respond to is the second point about the setting up of the regression equation and your problem with that being that there is a portion of the independent variable being -- the portion of the dependent variable in the independent variable portion of the equation. I don't know if this will help alleviate some of your concerns in that area, but first, let me assure you that we did inter-correlation checks to make sure that we weren't putting things in the regression analysis that would be too highly inter-correlated.

Second, the sort of theoretical reason for putting that variable in, the early years spending in Medicaid, was that part of the argument of this paper has to do with the budget choices state legislatures are making, and in particular, the choices they are making in this early boom period and how they're raising the floor that they are going to be working with in later years when they're setting their Medicaid budget.

So, even though -- I won't even say it's not ideal to use a portion of the dependent variable in the independent side, as long as you have a good reason for it. If you look at Table 4 and you go to the standardized coefficients, one way you can tell when you're using a variable that really shouldn't be there, that it's sort of hogging up too much of the explanatory power, is that a standardized coefficient will completely blow away the standardized coefficients of the other variables, and while, admittedly, the Medicaid growth in the early period is the largest one, the remaining variables added together contribute more explanatory power than that one alone, and that's pretty much my comment on that part.

MR. GREVE: I just want to say, as far as I'm concerned, Alan should be invited, re-invited all the time, because this is the American Enterprise Institute, and we really appreciate people who don't hold back and who tell it as -- if not as it is, at least as they think it is.

MR. WEIL: I'll accept that.

MR. GREVE: I just want to say three very quick things.

The first is this.

There is really, I think, a disjunction between the experts' knowledge and perception of Medicaid, on the one hand, and the public perception. It is, of course, true that Medicaid has long been disconnected from its roots as a, you know, welfare program. I mean that was a long, long time ago. We know that, and you're entirely right. If the criticism is this paper just shows that that is the case, that that happened, yeah, criticism accepted. I know that, I mean you know that, but it is still worth observing and going through the numbers at some -- I mean sort of comparing a hypothetical poverty-connected Medicaid program to the actual Medicaid program that's actually out there, because my sense of it is that most Americans would still think of Medicaid, oh, that's for the poor, and if this, you know, egregiously flawed paper does nothing other than to correct that misperception, it will have done a useful service.

The second thing is, of course, look, with respect to certain issues you mentioned, the pick of the base year and so forth -- and it's not an arbitrary choice, it was just dictated by the data -- we are, of course, aware that, yeah, this is an invitation to take a more careful look; it's not a final study in any sense of the word.

That said, I want to say two things.

First, we did try to look at year-to-year variations. There's just too much noise in the system. You don't get a very good sense, and it doesn't tell you very much at all. That is why we looked over the entire six-year time span. Could we have broken it down further by three years, yeah, and looked at sort of time periods to smooth out some of the noise instead of getting this blanket six-year look, yeah, we could have, except, you know, time is short and we're all busy. And the other thing is some of the factors you mentioned, the heterogeneity among the populations and across states, the different dispersion of new medical technology among states -- all of that is true.

Do I think that, at the end of the day, a closer look at those kinds of confounding variables and factors would explain systematically a lot of the variation that we here observe? I'd be surprised if it were true. I'm not opposed to anyone sort of taking the data set, going out there, and testing it. Be my guest. But I don't think we will find out that much more is explained.

And my final comment is this. And Alan, you mentioned at the end, look, if you really wanted to move to -- you didn't call it socialized health care, you called it single-payer system, I call it socialized health care -- you wouldn't leave it to the states, and my sense is no, sorry, you would, precisely, on the current conditions, leave it to the states, and the reason is that 55 percent financial responsibility and 45 percent financial responsibility do not add up to 100 percent responsibility, they add up to complete irresponsibility and an un-traceability of the political decisions, and if you want to move under the current political conditions in the United States to a fully socialized health care system, zero political responsibility is precisely what you need, because no politician would go out and actually advocate it and take responsibility for that position -- no eligible politician, I should say.

MR. HELMS: Well, that should give us something to discuss. Let me also say that we will certainly invite you back, Alan. The original founder of AEI, Bill Brutus, Sr. [ph], used to say that AEI was established under the principle of competition of ideas, and so, we definitely want to promote that kind of discussion.

I know there are lots of experts in the audience. You don't have to be an expert to ask a question, but -- and there's lots of issues that have been raised. So, I'd like to go to our audience. I'd like you to identify yourself and also wait for Sharon to bring the microphone, and we'll start with Carl.

MR. POLSER [ph]: Hi. I'm Carl Polser, independent health policy analyst.

This is another methodology question.

Michael, I enjoyed your paper, but you mapped out the relationship between the poverty rate and Medicaid expenditures, and you seem to assume that, gee, if the poverty rate is down, then expenditures ought to go down, as if these are independent variables, but to what degree does the causation flow this way?

Increased expenditures on Medicaid might actually help lower the poverty rate, because people have their health care costs taken care of in these various situations that Alan described. Perhaps if they're working, then the employer can pay higher wages.

So, there actually may be a dependent relationship.

MR. GREVE: It's conceivable that there are circumstances for some segment of the population that sustained Medicaid eligibility prevents people from falling into sort of the category that would otherwise qualify as poor. The question is how large is that population? I actually doubt that it's very large.

MR. HELMS: Okay.

I was going to ask Joe if he wanted to answer it, but -- yes, okay.

MR. MILLER: Hi. I'm Dick Miller with Federal Funds Information for States, and I must say it's refreshing, after two or three years of being criticized for states cutting taxes too much, we now have criticism of spending too much during the same period.

A couple of points on the paper.

You have a couple of outliers, Louisiana and New Hampshire, and you mentioned that Louisiana is heavily leveraged and that can explain a lot of the variance.

New Hampshire is even more heavily leveraged. It's the most heavily leveraged in the DSH program of all states.

The percentage of the program that is -- that flows through DSH -- I think it's close to a quarter for New Hampshire, and Louisiana, by comparison -- these are off the top of my head -- is, I think, 17 percent, and 12 percent is the statutory max that states are not supposed to go above. Just a piece of information.

The second point I wanted to make is that the fastest growing Federal health program is not Medicaid but Medicare Part B. It's growing by leaps and bounds, and unlike the current year, where Medicaid growth has been cut in half, Medicaid Part B spending has accelerated. I think a lot of this has to do with the changing structure of health care in the United States.

Medicare Part A is basically flat from year to year, you know, 2, 3, 4 percent growth from year to year, and that's because we're not giving health care through hospitals anymore.

This has secondary impacts in Medicare Part B and Medicaid, and the growth over time for things like drug therapies in Medicaid and other therapies in Medicare Part B has a lot to do -- and Bob, I would suggest -- I'm really interested in the kinds of stuff you've done on like Social Security and Medicare on one side and Medicaid on the other.

You might consider a Medicare/Medicaid piece together, because I think a lot of what is happening in one is reflective of the other.

And finally, the point in the paper on the kind of cyclical intent of the founders of Medicaid -- I've looked at the history. I don't really find that.

I think it is intent to provide more to states with worse fiscal conditions, and that, to some extent, could be explained as a kind of cyclical nature.

I find -- and I would agree that it would be nice if we had a structure that would have a higher match rate in a recession and a lower match rate in boom times, and that might be something that someone might be interested in working on, but I also find that the FMAP changes from year to year are basically pro-cyclical, and I did a paper for AEI 20 years ago, as a matter of fact, on this, and I've been doing little bits and piece ever since.

You wind up with the lag in the data, the person linking data, with two to four years, and you wind up getting increases in FMAP's to states that generally needed less and declines in the other direction, and I think that's something that is a budget-neutral sort of thing that you could fix. There are ways to fix it.

No questions, just a couple of points.

MR. HELMS: Okay. Thank you very much.

Any response to that? Okay. We'll go to the next question. John, in the middle there.

MR. GREEN: Hi. John Green from the National Association of Health Underwriters.

Alan, this is for you.

You say that inflation is really the big driver in Medicaid, and I'm wondering how that could be since prices are set, and how do you tease out volume increases to, you know, distinguish it from what's going on in inflation?

MR. WEIL: Well, colleagues of mine have done much more work than I have in trying dis-aggregate the basis of the increased cost in the program, and inflation is a major component, enrollment is a major component, and there are others, as well, but with respect to your comment that prices are set, I mean they're set in a world where the government has a lot of leverage but does not have complete control.

There are Federal standards about access to services.

Until not terribly long ago, and certainly in part of the period discussed here, there were Federal standards with respect to the big ticket items of nursing homes and hospitals, with respect to the cost-based reimbursement.

Prescription drugs have been a rapidly growing area, and I'm not an expert on financing of prescription drugs, but that seems to have been a major source of cost for states, where they do not, at least in the first instance, have control over the pricing, although they can do some things on the back end that can affect it.

But I mean in a capitalist society where capital dollars will flow from place to place, if you are the government and you are purchasing health care on behalf of poor people who otherwise might have nothing to pay, you do have leverage, because at least to some of those providers, if you don't pay for it, those folks' care will be uncompensated, and most providers would rather have something rather than nothing, but if you set your rates too low, the capital that was the foundation for that provider existing in the first place will flow elsewhere.

So, it's not as if, because prices are administratively set, that there is no boundaries or no bounds on what states can do. There are both economic and political bounds on how low Medicaid rates can be.

MR. HELMS: Okay.

I'd like to go to Jim over here.

MR. FROGE [ph]: Thank you very much for the presentation.

My name is Jim Froge. I'm with the American Legislative Exchange Council, and for those of you who aren't familiar with ALEC, we represent 2,400 state legislators around the country, which is the largest bipartisan group of state legislators around the country.

So, since so much of this was about state legislatures and state legislators, we have a couple of comments, and Bob, if you will provide me some leeway, I know President Bush is extremely prickly about multiple questions, but I have three. I promise to make them very quick.

MR. HELMS: Okay.

MR. FROGE: The first one is I wonder if Dennis -- and anyone can answer this, but Dennis, if you could give a few examples of recent -- or maybe even going back as far as a decade or two -- examples of states pulling off various shenanigans to maximize their match even though it's arguable whether or not what they're using the match for is actually health care-related, and the incentives are such that there's every reason in the world for states to do that, to be very creative with how they define Medicaid in order to get their match, so if you could just give some comments on that.

The second one is there was a great piece by John Aglehart [ph] in Health Affairs in the spring where he gave a big overview of Medicaid, and one thing he pointed out is the MEDPAK [ph] survey from 2000-2002 in which the number -- percentage of doctors who accept all Medicaid patients, all new Medicaid patients, dropped from 48 percent in 2000 to 39 percent in 2002.

Now, that is despite a massive increase in Medicaid spending during that time, so massive spending increase, access to doctors decreasing. I think that's a very important phenomenon that deserves some comment.

And third, Alan, you said that Medicaid is inherently expansionary and you think that's not a bad thing and it tends to -- well, it does expand whether times are good or bad.

You may think it's not a bad thing, but a lot of state legislators certainly do, and I wonder if you think it's sustainable, however, because I think I would argue that it's not sustainable; it grows disproportionately in good times and bad. Is that sustainable?

And lastly -- and this is kind of relating to the third question.

Alan, do you have any idea for reforming Medicaid, such as the cash and counseling model, that don't simply involve spending more money.

Thank you.

MR. HELMS: Okay.

A question, I think, for each of you.

Let me just make one correction. I believe that the John Aglehart Medicaid article was in the New England Journal of Medicine, but he is the editor of Health Affairs, and I wanted to say this at the last, but since it's up now and we're talking about John Aglehart, he made available copies of this issue, January-February, of Health Affairs, which is devoted to Medicaid, and so, there are several articles there if anybody wants to do some more reading on this topic.

So, Dennis, would you like to go -- a question for each of you, so Dennis, you can go first.

MR. SMITH: Thank you, Bob, and thanks for the question, Jim, and again, I think that the underlying premise that states are in charge of the decision-making about Medicaid, I think is -- I think it's reality.

I think it is -- again, they base their decisions and Federal dollars then follow those state decisions.

I think that is clearly -- I think that's how the program was designed to work, and I think that is the way the program does work.

In terms of the financing, that leads into that, and Jim's question is getting at that. Perhaps those decisions were based, at least in part, because the Federal Government would end up paying a higher share of the cost of the decision that the state made, and what Jim was referring to was decisions -- the upper payment limits in which the Federal Government was taking on a higher share of the costs of the program, because in Medicaid -- it was driven by two things.

One, Medicaid reimbursement policy is not the same in every state, and they are fairly general and broad.

I mean when you look at the statute on reimbursement, you see things like reasonable rates. You see Medicare cost principles, things like that.

There is a second part that goes into that which allows the state to share their cost of the program with local governments.

So, what we saw and what the previous administration saw, as well, was that states were saying, okay, we are going to increase our reimbursement rates to our providers, whether they're a hospital or a nursing home or a public hospital, etcetera, but because of the lack of specificity in Medicaid and because of the ability to do inter-governmental transfers between local government and state government, the states would say we're going to raise our rates, but then the public entity who is one of the providers, like a public nursing home, is then going to return a share of that increase back to the state.

The effect of that is the Federal Government has picked up a disproportionate share of the cost of that reimbursement rate.

I don't want to single out any individual state, because basically what we found was, when we scored the regulation that would have saved tens of billions of dollars over a 10-year period, the idea was that, once the states -- you know, it starts with a handful of states, and then, as soon as that is discovered, everybody else will follow.

So, I don't want to single out one state, because the savings came from preventing other states from doing the exact same thing, which had been allowed.

But Congress did step in and said no, we're going to put limits around this, because the effect is the Federal Government -- it has -- the match has now shifted.

But again, those decisions were often made by the states in saying if there is a way to shift the cost, we are interested in doing that.

Now, will the decision-making be the same if the cost of the program is the same? That is, if you maintain your specific share, will you do -- will the state be willing to put up its share? I think the answer obviously is yes.

States will continue to put up their share of the program and to meet needs that they see, because having worked in a state for three of the years of your study, again, those are -- those things are obvious.

You have constituency groups come in, you have provider groups come in, who all are there telling state legislators that there is a public good here, that there is an access problem there, etcetera, about saying Medicaid is a good investment and Medicaid should get priority over some other program or some other way to spend that dollar.

So, yes, Medicaid spending will continue to rise, because you have a whole group of beneficiaries that benefit from it. You have a whole group of providers who will say I will not serve an individual unless reimbursement rates are increased, etcetera.

So, it will all continue to move forward.

Will it move as fast, though, if the states are paying their statutory share of the program?

You know, I think that that is what we are seeing now.

As UPL has been -- put restrictions on it, I think that that probably does speak to at least part of the slow-down in the rate of growth, but clearly, the program is going to continue to grow. We're talking about growing how fast, is the issue.

MR. HELMS: Okay. And UPL, of course, is upper price limit.

And there was, I think, a question for -- I know there was one for Alan.

So, go ahead, Alan.

MR. WEIL: Okay.

Well, as to the issue of whether Medicaid growth is good or bad, I mean I don't know anyone who wouldn't prefer that --

[End of Tape 2, Side A

[Begin Tape 2, Side B.]

MR. WEIL: [In progress] -- legislators and taxpayers would be two groups that I'm sure would prefer that.

So, I'm not saying I have a preference for high rates of Medicaid growth, but I think, again, it's a matter of value choice. The question is what are you getting for that spending? Are you taking care of needs that politicians observe in their communities that there are people who would not have those needs met without the existence of the Medicaid program.

Is the growth sustainable?

Well, in the long term, by which I mean a long time, clearly not.

You can't have one sector of the economy grow faster than everything else indefinitely, but you could say the same for medical inflation, in general.

Is the current rate of medical inflation in the economy sustainable? No, it's not either.

Things will have to change, and they will. They will adjust.

The question is can we adjust them in better ways or worse ways, and that really brings me to the answer to your last question, which is are there ideas for reform that don't just involve spending more money, and here the answer is sort of no and yes.

The reason the answer has to be no is that, as long as Medicaid is paying for health care services which grow and have grown and, by all estimates, will continue to grow faster than the economy as a whole, Medicaid spending will grow faster than the economy as a whole, and whether we like it or not, it's just a truth, and so, there will be more money unless we fundamentally alter what the program does, but the answer also is yes, there definitely are reforms that don't simply involve spending more money.

I mean, first of all, if you took the long-term care system, which Medicaid has become the financier of largely by accident, and turned it into a more rational system, eliminating some of the -- or much of the bias towards institutional care, providing services more effectively to meet the needs of the population, particularly those with chronic diseases, and rationalized the financing, it might not involve -- it certainly would involve spending less more money than the current system will involve, which is a funny way of saying it doesn't involve spending more money.

When it comes to the acute care population, I would say the only way, really, that you reform Medicaid that doesn't involve spending more money is to take on the chaos and the inefficiencies in the current health care system, and I was in this room when former Speaker Newt Gingrich spoke about a real rethinking about where the dollars go into the system and orientation towards outcomes and moving away from some of the chaotic dollar flows and lack of accountability that plague the current system, and although I don't agree with all of his ideas about how you would get from here to there, a different vision of a health care system that is designed to be more efficient and to yield health outcomes rather than dollar flows to providers and insurers and others is -- definitely wouldn't have to cost more money.

So, the answer, I would say, is if we look at Medicaid as a system or as part of a larger health care and long-term care system, then I think we can come up with ideas that don't simply involve spending more money, but so long as we expect Medicaid to pay for health care for low-income families and to be the financier of the majority of long-term care in this country, it's pretty hard to get away from cost increases.

MR. HELMS: Okay.

Jinney?

MS. SMITH: I'd just like to comment on some of the points that have been raised in the last few comments.

One thing Michael stressed in, I guess, his first rebuttal was this issue of accountability, and I think that's one of the key things that, as political scientists, we are interested in looking at, because we're wondering where will the pressure come to alleviate some of these spending increases, because not only is there a sort of upwards price pressure in that, for every dollar you spend, you get a dollar or sometimes even more, depending on the particular state, but on the flip side, there's no pressure to cut cost, because of you try to undergo some innovative program and really cut costs, you don't really realize the fully savings as a state; you're only realizing half the saving as a state.

But this question of where -- you know, do these costs have to go up forever -- and one thing that started after the period we looked at in this paper is the formulary reform movement, and that started in 2001, where when states were hit with their first year of a revenue shortfall, that's where the accountability came on the states side, and you began -- in the states, there began this movement of really negotiating and looking at their prescription costs, which make up something like 10 percent of the overall Medicaid budget, and now we're seeing in states -- there are some two dozen states now in the process of already adopting or in the process of looking at these formulary reform programs where they're realizing costs of upwards of 30 to 40 percent in their overall prescription drug care.

Florida alone is going to save about half-a-billion dollars over a two-year period by offering or recommending and highlighting generics and really evaluating different pharmaceuticals on a case-by-case basis that's disease-specific.

So, I mean it seems like there could be some pressure on keeping these cost-controlled, but there has to be accountability, and that's something that Michael and I were particularly interested in, because when you look across these states and what they're doing in terms of the great years, you know, there it should have been the easiest to start saving money and planning for, you know, years like 2002 and 2003, and that pressure really wasn't there.

MR. GREVE: I just want to add one observation, because you asked about state legislatures, and I think, in years ahead, they'll be a very awkward place to be, and here's why.

Some 12 years ago, I think Medicaid was about 10 percent of the average state budget. It's now up to 20 percent in the average state. Obviously, this cannot continue, right? It's crowding out everything else.

At the end of the day, the state will still have some fire engines or funding to locals for fire engines. It will still build some roads, right?

The big things -- but all of that so-called discretionary stuff has been really squeezed, and it's been squeezed from two ways by Federally funded programs, one -- the biggest one, the elephant, is Medicaid. The other one, of course, is education, which is also Federally co-funded, albeit not -- on different terms and in different ways and not as generously.

What will happen if those two programs or program categories and those two constituencies go head to head in the states? And it's bound to happen.

If the Medicaid stuff continues to expand and the education programs continue to expand, as for independent reasons, I think they will -- that conflict among very potent, very powerful, and very aggressive constituencies is bound to get much, much sharper politically in years ahead, and the place where you don't want to be is the state legislature when that storm hits.

MR. HELMS: Okay.

We have several more people that want to ask questions, and given the time, I want to ask you to please keep them short, okay, and to the point.

Right here.

MR. COLTRITER [ph]: Thank you. I'm Bill Coltriter. I'm president of the Center for Alcohol and Drug Research and Education.

I'm astonished to hear Mr. Weil say there are no strategies to contain the growth of costs in Medicaid if it remains in its present state.

First of all, the GAO says between 15 and 25 percent of all dollars in Medicaid are lost to fraud and abuse.

Number two, we have modern information technology that allows us to do intensive case management, to change drug formularies, to change models of delivery to preventive care, home health care, etcetera, and there are also medical practice models that can change the incentives to shift some of the burden to managed care organizations, for example.

So, I would appreciate any of the panel's comments on (a) the existing and (b) the potential mechanisms to cut down on -- or to achieve cost containment.

They are clearly there. They are clearly proven methods. They're available today. And as information technology grows, they will become even more robust and fulsome for tomorrow.

And the third rail of the discussion is changing the asset disposition rules to avoid the cost shifting of long-term care from private families to Medicaid, and no one is talking about that.

If you look at how we can contain the cost, the political will to continue to fund the program is more likely to be in place, whereas if the costs continue to spiral out of control, that political will may begin to evaporate.

So, I would invite comment from any of the panelists.

MR. HELMS: Okay. It's a good point. But it also brings up the issue I wanted to ask about. Are any of these ideas really being worked into these demonstrations and so on that are out there now?

Any comments?

MR. WEIL: If you took my comments to say that there are no opportunities for cost containment in the current system, then I was not at all as clear as I intended to be.

I think the overall structure of the existing Medicaid program if we don't do fundamental -- either system-wide reform or Medicaid reform -- is one that's going to continue to have cost increases.

That's not to say that there aren't lots of opportunities for cost saving.

I would hate to think that my remarks would lead to the conclusion that sort of cost growth, willy-nilly, is inevitable.

I was talking more about the notion that there are broader opportunities for -- there are greater opportunities for savings if we look systemwide than if we just look within the Medicaid program.

MR. GREVE: I entirely agree with that analysis.

One of the plain things is -- look, there are two things, right?

One is that sort of the built-in growth incentives that simply stems from the basic financial copayment arrangements between the Feds and the states, okay?

And that beast you're not going to get under control until and unless you monkey with the program design in a very, very fundamental way and go beyond the many caps that HHS is now proposing and go to real caps.

And the other part of the -- so, that you can't get out of the -- and that will be the underlying trend no matter what you do.

Are there sort of more marginal things that you can do in the way of cost saving? Yeah, sure, and I mean that's a matter of degree and of policy detail that I'm not familiar with, but what is important to keep in mind is the underlying growth incentive.

MR. HELMS: I want to take a few more questions and continue this, but I promised Alan I'd let him go at 10 minutes till.

Thanks very much, Alan. I really appreciate your remarks.

Okay. Dennis?

MR. SMITH: I did want to agree with your point that, again, there are, we believe, lots of things that can be done in the system that will improve the quality of services and at a more reasonable cost to the taxpayers.

We've been testing some of these out. You mentioned cash and counseling and home and community-based services.

Again, we -- if you look at states, where they spend their long-term care dollars, most states are still spending a majority of their dollars in institutional care.

We've got a handful of states, I think about three states, that actually spend less than 50 percent of long-term care dollars on institutions, and when you look at those states, those are very high-quality in terms of client satisfaction and are saving money relative to institutional care.

On the other hand, we still have quite a few states that are spending 90 percent, 95 percent of their dollars on institutional care.

It will take time to rebalance that system, but the administration has a number of proposals to do precisely that, and in the long term, that will improve the quality of care.

On health insurance side of things, again, to some extent, we need to help state legislatures to understand they are sitting on top of probably the largest or at least one of the largest health insurance companies operating in their state, but we don't always treat it like health insurance. Medicaid has a lot of different provisions in it that you don't see in the private sector, that you don't see in the Medicare.

We tend to pay for -- again, on a per individual basis.

The administration has encouraged family coverage policies. That, again, we think, is better outcomes for the families that will lead to higher insurance rates but also less expensive rather than paying, you know, $150 a month per head. You can buy as good, if not better, policy on a family basis.

So, in terms of -- on the mental health side, Medicaid spending, at least in Virginia, was spending 55 percent of the public dollars on mental health.

So, Medicaid obviously is a very important source of funding for mental health, substance abuse, etcetera.

The President's commission on mental health just came out and said, you know, there are -- you know, we've got to bring the program up to keep pace with the advances that have been made out there, whether, again, it's on home and community-based waivers, where people are talking about not an issue of health care for them. It's an issue of independence for them.

So, I think there's a great deal that can be done to change the way we deliver services and lower the cost, as well.

MR. HELMS: I want to take one more question, back here, please.

SPEAKER: A question on the Medicaid waiver, states that have a waiver to pay for assisted living -- for instance, I'm aware of Florida.

Are you seeing any savings or any comment on that strategy?

MR. SMITH: Yes. The cash and counseling waivers -- we also have an evaluation component built into those, and they do deliver services at a lower rate of cost, generally because you're avoiding trips to the emergency room, you're avoiding higher-cost institutional care.

So, we do see that, in the long term, they do save dollars.

One of the things that we have done in waivers, as well, tell the states that Medicaid will pay for what are called transitional costs.

Oftentimes, in moving an individual from an institution back into their own home or someplace else, there are additional costs associated with that that Medicaid can be used to pay for those things that will then save costs in the long run.

But when you look at it over the long term, yes, they save dollars.

I believe the Lewin Group [ph] did the study on the State of Washington, on their home and community-based waivers, and they showed savings.

MR. GREVE: I'm glad you asked that, because it illustrates, in part, what -- I mean the fundamentally flawed incentives in the system.

It is, of course, true that, if you could go to assisted living rather than long-term institutional commitment, that would, technically speaking, save money, and that is how these waivers are being basically peddled and sold.

The glitch is that the system is extremely bad at telling who would or would not otherwise be in long-term care arrangements, such that an assisted living waiver then amounts, de facto -- I mean it starts out as let's save money and not horde everybody into long-term care, but it ends up as a de facto gargantuan program expansion, and the only safeguard you have against that is the good sense of the governor, and the true fact of the matter is that some governors have good sense and others don't and view this as a vehicle to expand the entire program.

And that gets me to one last thing, which is that -- nobody mentioned this today, but I'll mention it and I'll be impolite.

These programs have frightful substitution rates and push out private insurance markets like you wouldn't believe, truly frightening rates, and the more generous the program, the more frightening the substitution rates.

SCHIP is the flagship of them all, and nobody wants to mention it, and I think you could make a good case that that is sort of a surreptitious long-term cost expansion.

I don't even want to argue that, but I want to raise that issue. We ought to be very, very nervous about it, all of us.

MR. HELMS: I'd just mention, Michael, there's a well-known term which I first encountered in the government, and it's called the woodwork effect, in the sense that you establish a program under the guise that you're going to substitute for something else, and then new people come out of the woodwork and it ends up costing a lot more than anybody ever expected, a very common phenomenon in government programs. I see that our luncheon speak is here, Secretary Sabatini, and so, that's a good sign, and we invite those of you who can stay for lunch to go right into the next room. We'll give you time for a bite to eat, and then we'll hear from the State of Maryland. I want to thank all of our panelists, Alan included, for a very interesting discussion.