Thursday, March 6, 200

Session 2: American Health Care: The Problems of Access and Cost

Mark V. Pauly, Ph.D. The Wharton School University of Pennsylvania

CHAIRMAN PELLEGRINO : Our next speaker is Mark Pauly from the Wharton School at the University of Pennsylvania , and I'll point out again to Dr. Pauly , we do not give extended introductions with a curriculum vitae because they are so impressive that it would take all the time of the discussion. So I presume you would prefer to have the time for discussion. So we await your contribution. Thank you.

PROF. PAULY: Well, I approached this task with some trepidation. When Dan asked me if I'd come and make some remarks, I said, "I'm not an ethicist. I'm just a card-carrying economist." And he said, "That's okay. That's what we want you to come and represent." And so I'll try to do that to the best of my ability.

I am, in recent years, a more vigorous consumer of issues in bioethics since I've been appointed co-director of a new flagship undergrad program at Penn, the Vagelos Program in Life Sciences and Management, where the cream of our undergraduate crop study the combination of what I call "the discovery question" in both science and management, and in a way, "the discovery oxymoron," managing science discovery, is a little bit of an oxymoron.

Of course, the students want to know not only some of the economics that I can convey to them, but they're also very sensitive to ethical issues, both in terms of justice, in some concept of justice, or (in some ways we actually try to raise anxiety about this a bit) in terms of conflict of interest. So I've at least had more reason to worry about ethical issues in my Mr. Chipsian phase that I seem to have entered as opposed to when I mostly taught Wharton MBA 's, not that they're not ethical.

(Laughter)

So I thought in my remarks I would try not to punish people who read my two papers, which I reread and I still agree with myself, so I'm happy to report that, but instead make some remarks that are intended in a way to link them to the discussion du jour of health care reform and at least in my way of thinking what I think I was trying to say there actually seems to me to be fairly strongly linked, and so my remarks on the linkage may be of interest to you, and, at worst, they'll remind you of what I said in the article, whether you agree that I made the correction well or not.

So, okay. So my version of health reform breaks down to just two questions. There are a lot of questions that circle around these, but I think there are two main issues that people talk about when they talk about health reform and pretty obvious probably.

One is cost-containment, which — and I've tried to give the most lukewarm definition of each of these two objectives in hopes that at least we might be able to achieve that at long last, although there are stronger versions. But I guess literally "cost-containment" would mean stopping growth in health care. I don't think that's going to ever happen, and actually speaking of my MBA's, they would be very disappointed if it was because that would not be good prospects for their careers. But I do think that an issue of slowing the rate of growth in medical spending is an important part of the debate.

The other part is covering the uninsured. I'll put my heart on my sleeve from a policy point of view and say I'm in favor of universal health insurance but not universal health care, at least as I understand those two terms to be defined. But this is a more less-aggressive definition of the goal, but I believe actually one that may be within our grasp if we aren't too slow about it. It's already starting to slip away. But the idea would be to make some changes that would reduce the number of people in the US without health insurance, either public or private.

My policy prejudices are — and maybe this just common sense. I actually got this from Tim Russert , I think. But it's more likely that the government would be able to do something about covering the uninsured then cost-containment. And if you know from my paper, and I'll kind of hit it here as well, cost-containment is a much higher mountain to climb and a slipperier one, too, than covering the uninsured, and so I'll say that.

But the two can be independent — and here's where I'm not so sure I agree with myself — and yet they are interrelated. So I guess let me first define the position I don't agree with. Some people are taking the political posture, especially on the Republican side, that, well, we really can't do anything about the number of uninsured people unless we get health care costs under control. I don't agree with that. Having said that though and being an economist, I have to say on the other hand, I do believe there is a connection between what's happening to health care spending and what's happening to the number of uninsured. So the connection is not one-for-one, but there's definitely a linkage.

I'm not sure whether I need to apologize too much for welfare economics, but the normative framework that I carry around in my head is one based on welfare economics.

Richard Musgrave , a distinguished theorist of public finance and the person that I studied most intensively back in the day, generated what I thought was an interesting metaphor of how economists think about government and public policy. He imagined kind of applying Adam Smith 's vision of how to improve productivity in the pin factory to government so there would be an allocation branch that would be staffed by economists and it would have sort of full control over the term "efficient" and "inefficient." We wouldn't let anybody else borrow that, not even engineers, certainly not ethicists.

(Laughter)

But there would be another branch of government staffed, I imagine, by probably ethicists and lawyers and maybe clergymen, for all I know, who would talk about the distribution of real well-being.

Now one of the theorems actually that people made a fuss about at one point in economics was showing that the allocation and distribution can't really be done in separate water-tight compartments because what you do with improving efficiency affects obviously the distribution of well-being.

But I'm going to try to justify or at least link such normative conclusions as I come to in welfare economics and recognizing the most serious problem of welfare economics which is at the abstract level we can we think characterize states of the world as efficient or inefficient. There's a fancier definition. But the kind of plain garden-variety definition of efficiency or inefficiency in economics — I'll do it that way — is a situation where there is some change that could be made that would be to everyone's benefit, and we think we can characterize it even with that kind of wimpy criterion. It actually does allow you to step into the phone booth and come out as a caped crusader against inefficient things like protective tariffs and at least under some readings minimum wage. But the dilemma or the limitation of that is that it doesn't uniquely define the most desirable state. You actually need somebody to either take up the responsibility or have received the golden tablets on which are written the social welfare function that tells you how well off various subgroups of the population ought to be based on some judgment, not based on economics.

Okay. Well, so the most important things to know about cost-containment, I guess, one way to say it is, it's not like petroleum products, so we all are observing some cases fairly painfully. I haven't got my gas bill yet. Thank goodness for global warming actually. It's been kind of a warm winter in Philly. But total spending on petroleum products is rising, and, although some people call that inflation, that's really not the right economic term either. It's just a rise in relative prices. But the interpretation of the increase in spending growth on natural gas and gasoline and so forth is you're spending more to get the same thing, and that's been the growth over the last couple of years.

In contrast for as far back as we have data, which I think goes back to around 1946, as far as we call tell — and I'll say a little bit about why this isn't the strongest proof, but it's all we've got. But as far back as we can tell, the lion's share of whatever share that is of growth in health care spending is attributed to what's called technological change.

And the intuition behind that, actually the technically correct intuition is just using inputs and outputs differently, using inputs differently to produce output. But the usual interpretation we give of it is that there's been improved technology that, in large part, the spending growth has been accompanied by new products and what can also be called new technology, new uses of existing products — for example, the discovery that greater widespread use of statins that have been around for a couple of decades would be a new and good thing to do. And so the basic punch line is that in contrast to petroleum products or other sorts of things, it's not as if we're spending more for the same thing. We're spending more for something different and arguably at least for something better.

I do need to apologize here a little bit, and I guess show some of the dirty underwear. The measure of technology is a statistical residual in the analysis of the rate of growth of health care spending. So you take your data on how spending grew last year, you see what portion of it you can attribute to the aging of the population — that's usually not all the much actually, less than 1 percent, but it's something everybody can appreciate and me especially personally — and then you also adjust it for changes in input prices, and then what's left over you call "technology." So it's really the label of our ignorance, but on the other hand, it does seem to be correlated with observations on the changes in the way medical care is produced and used both, as I said, in terms of new products and new uses.

And so if you take that as okay that most of the growth has been coming from new technology, then the two kind of evaluative questions in welfare economics would be what I called here Question 1 and Question 2. They are really the same thing but stated slightly differently. So if you were a cost-benefit devotee, you could say, "Well, I would then judge whether the spending growth that we experienced in 2006, which is the last year for which we have data, 7-point-something percent, I believe, well, that provided some positive benefit. Is the benefit greater than the cost?" So you take a cost-benefit approach.

An alternative way of looking at what is happening, especially at the level of the economy as a whole which I think can be instructive is, this is basically invoking the concept of opportunity cost. Health care I will say, as I'm sure everybody knows, is a fairly labor-intensive product, in contrast maybe to petroleum products, which are not. It's a fairly labor-intensive product. So roughly speaking, what you can at least inquire about is whether the value of the outputs that the resources being devoted to health care are producing is greater or less than the value that those resources could produce if, say, they were used in the next-best place in the economy, whatever that might be.

So, I mean, I think it's kind of a useful question, mostly because then you get, I think, back the message, nobody knows the answer, but you ask, "Well, what if all the smart young people going to medical school or nursing school and so forth who are going into health care, what if they didn't go into health care but did something else? What else would they do and would we feel happier if they went fewer into health care and more into something else?" I guess my nihilistic position here is, I don't know the answer to that, and I'm pretty sure nobody else does either. So trying to stake out a hard and fast position of the desirability or undesirability of health care spending growth, I think is largely — maybe we can find something to lean on here — but is largely undercut by the absence of information on the relative value of health care output compared to other things.

I'm pretty sure if those smart young people went to law school, I probably — this will be my last slamming-a-lawyer comment — but that would not be the greatest thing, but if they went into engineering school and, you know, stopped the bridges from falling down over the interstates in Philadelphia, I would potentially regard that as a gain. But there have certainly been some amazing things that people do and can do in health care. So that's sort of the normative comment there.

And then this sort of peels it away a bit and elaborates on the idea of why it's hard to know. It's hard to value health either in money or compared to the value of labor for other uses.

I guess this is my cue actually to try to be a little more positive and to talk about at least some people who think they know some of the answer to this value question, and I will cite David Cutler at Harvard who, along with some others, mostly everybody who has looked at this question has come out with the same answer. Well, they decide to look at the low-hanging fruit, and the low-hanging fruit was the improvement in health related to cardiovascular disease, and David and some others have tried to determine both how much of the improvement in health related to cardiovascular disease was due to new medical technology — they guesstimate 50 percent. That's actually a fairly squishy number, but I think the language is something like "nobody could prove it was less than that," and on a somewhat more substantial level even very conservative methods in the philosophical, not the political, sense of attaching — or perhaps the methodological sense — of attaching money values to health tell you that in a way the investment that we have made in spending more money on the treatment of people with heart disease and stroke as possible outcomes has been worth it with a payback ratio of something like 4-to-1 to 7-to-1.

So this isn't all just hypothetical pie-in-the-sky and it isn't all nihilistic. At least for that particular development in new medical technology and for that particular and quite important disease category, it looks like we're ahead. Or, another way to say it is — and I've actually tried from time to time to get some of my MBA's to join me in offering health plans that are atypical — why not offer a health plan that says — or slogan — "Last year's technologies at last year's premiums"? That would control health care cost.

And then we could see how many people would want to buy that plan, and my suspicion is not very many either in the public or private sector.

But now the other side is arguing that — and I'm sure that Elliott Fisher will raise this as well this afternoon, so I have him kind of in my head — arguing that compared to not having improved technology at all, we're probably better off spending the money, at least collectively or in an aggregate, than using it for what else we might have used it for, that plasma TV or, if my wife ever lets me, the pickup truck I really want to buy. But instead of spending it on those things, spending it on health is probably desirable, while we can't really prove that we've done that in the most efficient fashion, perhaps 80 percent of the benefit could have come from 50 percent of the spending growth or whatever the fraction you might want to plug in there.

And some fraction of the spending growth, as I said, is also not attributable to new technology is attributable to — well, the way it's usually phrased in the article Uwe Reinhardt and Gerry Anderson write every year for Health Affairs , the subtitle for that article always is, "It's the Prices, Stupid," and so there has been particularly, both in terms of understanding why our health spending in the US is higher than it was last year, and most especially in trying to understand why our spending is higher than it is anywhere else in world, a major contributor especially to the latter but somewhat to the former is the level of wages and prices in the US. We do pay more, not probably as much more as you might think if you watch TV, but we do pay more for pharmaceuticals — that's prices. But, of course, hospital spending and spending on health professionals is a much higher fraction of total health care spending than drug spending. And there the incomes received by physicians, by nurses, by phlebotomists, even by orderlies and clerks in the health care sector has not only been growing more rapidly than wages in general, but it's also considerably higher than in other countries.

So actually once calculated, what would health care spending as a fraction of the GDP look like in the US compared to other countries if we did the simulation and let those other countries pay their health professionals like we do, and it turned out we were third from the bottom in OECD fraction. Only Turkey and Portugal were lower than us. So at least that's important, I think, as part of the anti-Paul Krugman protective shield perhaps or something.

Anyway, some technical reasons why it's uncertain, we think, and there's increasing evidence suggesting there are important cost offsets in health care meaning — this is the way it's usually described — spend more at least on some particular kind of prescription drug if you can get people to do it and comply like blood pressure medication or hyper-cholesterol medication, that probably can save costs somewhere else in the system over time.

There is a fervent desire, I think, on the part of policymakers to want to believe that you can save money by spending money across the board. My reading on this pretty skeptical, that the actual evidence we have that spending more even on things like colon cancer screening and such would actually reduce total spending is pretty weak. Only the old, cheap pediatric vaccines, at least in my reading of it, actually have a negative net cost. But it's certainly true that a lot of preventative measures have a net incremental cost after you take out cost offsets which are low enough so that it's a highly efficient thing to do. They're worth the money, but they still cost money.

The other issue has to do with the definition of efficiency and inefficiency in economics. Here I draw a parallel to at least what I remember is the definition of efficiency from physics which was there was a kind of theoretical efficiency of a machine and then there was the actual efficiency and they differed because of friction. So when Peter Orszag took over as Director of the Congressional Budget Office, he polled all the health economists asking them how much inefficiency did they think there was in the US health care system. I think the central tendency was around 30 percent. I was lower than that, but that's the guess. But at least when I've talked to people about how they answered, they were thinking of it in terms of the theoretical improvement in efficiency, but the dilemma is that for many of the practices that seem to be inefficient, at least so far, we haven't either found a reason why they exist, which would be the first step on trying to change behavior or a good way to change behavior.

The other sobering point I'd make about cost-containment is that, suppose that we've discovered a way to put in place some of the changes in the organization and production of health care that Elliott will talk to you about — and I actually have some of my own sure-fire, can't-fail interventions, too — suppose you put all of those into place for several years, that could slow the rate of growth for health care spending. So instead of being 2 percent more than the rate of growth of GDP, it could be less or maybe even zero. The main problem is that, as long as the enterprise of developing new medical technology continues to operate in the same way and at the same pace as it has, after you phase in your cost-reducing strategy — and we actually did that with managed care. I believe had we not done what we did with managed care, Denzel Washington to the contrary not withstanding, health care cost would be 15 percent higher in the private sector than they are now. But having said all that, the magic, it will appear, will eventually go away because we really don't know any interventions to slow the long-run growth rate of health care spending, and that's sort of the last rhetorical question.

There are some things, as I said, that people think could improve efficiency, so I thought I would pick on the one which there is most consensus among economists and most puzzlement on the part of normal human beings. So we think, at least an important cause of inefficiency in health care — and this is not the pontifical "we" here: Victor Fuchs actually did a survey of health economists and almost everybody agreed with this — that one of these, at least the easiest to target, the most defensible target for something that is inefficient and is causing health care spending to be higher than it ought to be is the tax subsidy to employment-based health insurance. So in the first dot line, I actually put myself in there. We have a wonderful benefits department at Penn, and so here's what they let me do as an upper middle-class person, you know, at Wharton not earning a too-shabby salary.

What they allow me to do is to shield about $8,000 of our family health insurance premium from taxation because that's technically paid by the employer. We have a cafeteria plan at Penn which allows my wife and me to shield about the $4,000 a year that we explicitly pay as a health insurance premium. And then we put the maximum of $4,000 a year into a flexible spending account, which is also tax shielded. And, fortunately, we both inherited the gene for myopia and so we never leave any money in that account. We always have the best collection of designer prescription sunglasses on our block. But the basic punch line here is that the laws, while I'm not breaking any laws here — I guess I'm not that proud of it either, but I do try to thank the Treasury every time I get a chance — the laws allow me to shield $16,000 of my compensation from all taxes, not federal, state, and even the dreaded Philadelphia wage tax. And at my marginal tax rate, that's probably worth something like $7,000, all told.

So that's a subsidy. It is, as I said, because it's tied to the taxable income. The value is greater the higher the income for a person, and because generally that means the tax rate will be higher. When people have tried to total up how much this tax subsidy amounts to, it's at least $200 billion and growing and, of course, it's disproportionately claimed by upper middle-class workers. That may sound inequitable to you. I hope it does. At least it sounds that way to me. But the efficiency side, if I remember where I'm coming from here, it's also criticizable from the efficiency side because we believe that induces upper middle-income people to choose levels and types of insurance that are not very cost-containing, essentially because, as I've already argued with my flex account, but it would work with everything, we're splitting the cost with the Treasury, and so that causes the level and the generosity of health insurance coverage chosen by middle-class and above workers to be higher than it would otherwise be, and there's another defect, which I think I have on the next slide.

So for one other problem with the arrangement in health insurance which seems to make it function differently than other kinds of markets, even differently than other kinds of insurance markets, is it appears that when the upper middle-class wants to enhance the generosity of their benefits or to have coverage of the latest new technology, even if it's very expensive and perhaps doesn't add so many quality-adjusted life years, there's a dynamic which I'm sure you can speculate on as well as I that suggests that that's going to be the standard of insurance for everybody.

And so I think what we've been seeing — and now this is the linkage in large part to the growth in the number of uninsured — what we've been seeing over perhaps the last 10 years is a continued growth in the generosity and costliness of insurance coverage chosen by middle-class people. Middle-class people up to this point are still able to "afford" it, meaning they're willing to continue to work at firms that provide that insurance coverage and, arguably at least, they prefer to spend their rising real incomes on that improved health care than on other things. But lower middle-income people are bailing out at a pretty rapid rate. So sometimes the metaphor is, it's as if we're trying to sell a new and improved Lexus every year to everybody which might be okay if you're well-off at least up to a point, but does seem to leave an important group behind. And, yet, the obvious solution, why not create a Kia version of health insurance runs into both regulatory, ethical, and I think even market-related complexities here. So that's sort of the story there.

So now that I'm talking about the uninsured, I might as well make my remarks on the uninsured and then we can have a discussion.

Health economists have not had a great agreement on how to conceptualize why there is social concern about health care and health insurance. I think that's fair to say. There actually are in the health economic journals a fair number of articles that talk about fairness and equity. They are mostly written by British and, for some reason, Norwegian health economists. I'm not sure I understand that part. And they are a little bit — and so I'll criticize a little. They do tend to get down to what seems to somebody across the pond — maybe I'm just not sensitive enough — to be kind of the angelic density of pinheads. And some part of them are explained using cricket analogies, which make no sense to me whatsoever. So I'm not sure about that.

(Laughter)

But the way I've tried to think about the reason why we are concerned about the uninsured, which seems to be an empirical observation on the human species, is to model that or conceptualize it as a kind of externality. Externality in economics is any time when one person does something if it affects somebody else's well-being in a way that's not immediately mediated by the market. And to get back from the gobbledygook and just sort of talk about common sense, my hypothesis here is, that because I think it's true of me and true of right-thinking people I talk to, that in addition to all the other things we might want to spend our income on, we might feel depressed or sad to note that other fellow human beings are not consuming effective medical care and we might want to think about making, not a transfer to them to make them happy, but a transfer to them to make them healthy by causing them to increase their consumption of medical care.

Now, of course, if we're talking about contagious disease here and you could catch something from somebody else, you don't even have to be unselfish. Even Scrooge McDuck would pay for that, I guess. But the great bulk of medical care is not related to contagious disease, at least that's treatable at the margin. The great bulk of medical care spending comes from other things, but I think —this is an hypothesis — but I think that the thought that I would be willing to pay something to cover the uninsured, I can assure you of that. I can also tell you it's not my total income, but there's some number in between there, and the general vision here would be to use government to kind of mobilize and assemble the willingnesses to pay of the reasonably well-off and healthy in society for things that would improve the health outcomes of others.

And actually you can't avoid it. When you subsidize a good for which demand is stochastic like medical care, it is insurance no matter what you call it. You could call it "welfare," but it's still insurance. And so we generally think of this as a case for subsidizing or maybe even furnishing some kind of health insurance to those portions of the population which, absent an intervention, would consume levels of medical care so low that these externalities in consumption would be there at the margin and would be bothersome.

And although this is obviously a kind of an ethereal concept at least in terms of the standard methods of economic measurement, I'm here to tell you we have had pretty good luck, and I'm kind of revisiting this. Using this concept to explain why some states provide more generous Medicaid plans than others, there is a price elasticity out there. If you change the matching rate, they suddenly get more generous. There is an income elasticity out there. If state income rises and the matching rate doesn't offset, higher income states are more generous. In the work we did some years ago, there was also the Southern dummy. States in the South just seemed less generally inclined through the public sector than others. It may be no surprise. I don't know whether that's still true. But that seems to be something important. So that's one reason, why covering the insured [varies].

A second reason that I've been trying to pursue in my own research with Jose Pagan at the University of Texas-Pan American — we think we're getting some interesting recent results, so I'm excited about them — basically say, if you are an insured high-income person, if you live in a town where there are a lot of low-income, uninsured people, the quality of medical care you get will probably be adversely affected, not just by the potentially larger charity-care burden in your local hospital, but also by the phenomenon that especially specialized high-tech kinds of medical care need a large enough market in order to exist. So Podunk doesn't have a symphony orchestra unless it's, of course, the vacation home of a lot of well-off people, just so in a community where most people are low-income and/or uninsured, the level of sophistication of medical care will not be what it is if the average income and insurance coverage were higher and that spillover will affect insured people.

Now in some markets for better or worse, and I think mostly for worse from the point of view of goodness here, markets have been able to segment themselves so that the low quality provided to the uninsured doesn't spillover into low quality for the insured, but we think it's at least still an important influence and, particularly in the smaller cities and towns where a lot of the uninsured live, a potentially important argument.

The third argument is actually the one made by the Institute of Medicine or at least an argument made by the Institute of Medicine that I find most persuasive but it's kind of third on my list in terms of persuasive power. It basically says, think of the uninsured people. Let's make an assumption about how much more medical care they might consume and how much healthier they would be if they were insured. Let's attach a value, a monetary value, like we sometimes do as economists — as Vic Fuchs says 'It's dirty work, but somebody has got to do it' — and so we attach a monetary value to the improved health of the uninsured. It looks like the value of the improved health is more than the cost of covering the uninsured, which kind of leaves you with the question of why don't the uninsured pay for it themselves, but you may be able to invoke something there. At least those are the three arguments.

So I will try to go through quickly some simple political economy ideas here and just kind of give the 25-words-or-less summary. I believe one of the problems, one the reasons why it's been hard to make the case for covering the uninsured is related to something that I was just talking about. We don't really have good and persuasive evidence of how much good insurance does. My reading and criticism to some extent of the IOM reports on this subject is that they're written to preach to the choir. All of us in health care feel health care is a good thing, but we're not really so persuasive in explaining to people who may value private goods or, for that matter, may value other public goods, like repairing the decaying infrastructure, why improved health care is the best way to spend money.

The strongest evidence we have on the impact of insurance on the use and outcomes of medical care are for poor people generated by Ronald Reagan when he was governor of California . As you'll know from looking at my paper, as you may know, most of the uninsured are no longer poor, if they ever were, and for that population we actually have much less evidence about how much good it would do.

Then there is what I call the "Altman conundrum" named after my friend, Stuart Altman , who is the perennial chair of any health economics task force in the Washington , DC area — Stu can do a whole lunch shtick on this one — explaining the reason why we don't have health reform. It's not because people at least inside the Beltway are not eager to have health reform and reduce the number of uninsured — they even are at the Cato Institute. It's that the next-best alternative from their point of view is to do nothing. And so compromise is what's needed. And I will say, playing a little bit against character here, I am optimistic these days because I think there is compromise in the air, although we shall see.

But the fundamental political economy question I think is whether the price tag of covering the uninsured which I estimate to be at least $1,000 per tax-paying household is something that people would be really willing to pay. We may want to talk about that.

The economics of crowdout. This is, of course, one of the points of debate, particularly about the extension of SCHIP. There is a concept called "target efficiency" which basically says, if you're spending more government money to cover more uninsured people, how much do you end up spending per uninsured person that's covered? And the problem or issue is that, as you move up the distribution of income, you find a larger and larger fraction of people who already have private coverage, which means that, if you do put in place a public coverage program — and you can't just say you're only eligible for this program if you're uninsured. That's not going to work — you end up drawing into the public program a large fraction of people who would otherwise have bought private coverage. The current SCHIP numbers are about 50 percent crowdout.

In one sense, since I'm not too keen on the government actually providing health insurance for people well above the poverty line, I kind of like that result. On the other hand, I have to tell you that from a pure efficiency point of view, that's not inefficient. If you are making a transfer to families who previously had made a sacrifice to take the lousy job so they could get health insurance coverage for their kids and now you offer a subsidy to them that evens up the burden, that's not inefficient. It's a transfer and perhaps a justifiable transfer.

The last dimension of crowdout I wanted to mention here which I think is important has to do with, in a way, the problem of plumbing. So here's the way the story goes: Suppose that there was universal health insurance coverage. That should mean that there would be fewer people receiving charity care in hospitals. Right? So what you're supposed to believe then if you're the middle-income insured tax-paying family, that will cause hospitals to reduce their prices because now they don't have to cover the cost of the uninsured. When hospitals reduce their prices, what you're supposed to assume is that will cause insurers to reduce their premiums because now they don't have to pay out so much in benefits. Then when insurer premiums fall, what's supposed to happen — I actually believe this may happen eventually, but it's a little bit of a tenuous point — your employer will give you a bigger raise because now they didn't have to spend so much for health insurance. And that's the way that you get the money back that you paid in initial taxes to cover the uninsured, and I guess — I'm not going to try to sell you a bridge here, but there's a lot of possibilities for leakage there.

So here is my normative comment on plans to reduce the uninsured. I think it should include the three M's and exclude the two E's, and I guess one reason I'm optimistic is that at least two of the three M's are present in the program proposals of all of the remaining viable candidates and maybe the last M we could deal with.

So one of the M's is markets. That's not shibboleth, that's — the general idea here is that, both on the Republican and the Democratic side, people are envisioning a setting in which people, consumers, could choose — subsidized ones as well as unsubsidized ones — could choose from a wide range of different health insurance plans. It wouldn't be a single payor. It would be universal health insurance but not universal health care. People could choose and, to some extent, the differences in cost between different plans would reflect the differences in quality between different plans like the breadth of their network, so a more market-like arrangement, more or less what's available to me as soon as I bite the bullet and go on Medicare. I can choose Medicare Advantage Plan as well as the government plan.

Means testing. So the idea of offering larger subsidies to lower-income people and tapering subsidies as incomes rise seems to be present on both the Republican and the Democratic side although you actually have to read John McCain 's plan carefully to see it there, but it is there. And then this is the part that's actually a splitting of the Democratic side and certainly splits them from Republicans but the possibility or the desirability of an individual mandate. So all of those might be possible and, I think, would probably be a good thing.

The two E's, the things that I think both are bad ideas in themselves and in my judgment make it harder to get consensus on the three M's, one is employer mandates. So if you're a true-believer economist, you believe in contrast to normal human beings that the boss doesn't really pay for your health insurance. Amy Gutmann , the President of Penn, is not shelling out of her purse money for my health insurance. That $8,000 is, I like to believe, money I'm worth to the University which they would have paid me in any case. The only difference is and the reason why it's there rather than in my take-home pay is because of the aforementioned tax breaks. So it's my money. And an implication is, for an employer that currently does not offer health insurance, if you make that employer pay for health insurance, the cost, at least the lion's share of it, will not end up falling on the owners of the firm, it will end up falling on the workers in the form of either lower raises or, if their wages are close enough to the minimum wage level, more layoffs, and that seems to be undesirable from both an economic point of view and from an equity point of view.

The other thing that I think ought to be excluded if you want a reasonable discussion here is the notion that, well, we're going to be able to cut the cost of covering the uninsured greatly by improving efficiency of the system — not that it's not a good thing to try to improve the efficiency of the system, but I guess I've already expressed my skepticism. I think that, as usual, politicians who propose to pay for something largely out of reduced waste, fraud, and abuse are probably deserving of some skepticism. So my ideal here — I'll just read it off because I want to time to talk — would be an employer-enforced subsidized individual mandate for income-conditioned minimum coverage.

And so this is sort of the picture, free comprehensive insurance for the poor. As household incomes rise, the subsidy — both the subsidy and the minimum required coverage fall. Households can choose among public and privately provided plans in a market setting like today's Medicare. I actually — it wouldn't be my first choice, but I actually would be quite comfortable with Medicare for all, if by "Medicare for all" you mean today's Medicare which is not a single-payor plan by any stretch of the imagination. And the reason why I don't like employer mandates is, for one thing, they're paid by workers. But, for another thing, employers think it's their money and so they fight like tooth and nail and fairly effectively to stop health reform where they really don't have a dog in the fight.

There are some hard tradeoffs. This is my comment on the mandate controversy between the two Democratic candidates, at least from the point of view of somebody who has tenure. What I can say is, well, you'd want to pay a subsidy to make household premiums affordable, wouldn't you? And once you've done that, then it's not obvious to me why you wouldn't be comfortable having a mandate. Or to reverse it, if you're uncomfortable with a mandate, it must mean you haven't yet ponied up enough money for adequate subsidies.

The next one should say to get near-universal coverage without a mandate, you probably need about a 90 percent subsidy. So the great virtue of — one virtue of a mandate — is it allows the subsidy to be somewhat less and, therefore, less collected through the tax system.

So I think I've said a few of these — this is the sermonette part of the talk — but we do need better evidence on the health benefits of coverage for the uninsured who are not poor. My version of the price tag for an adequate subsidy plan that didn't try to hide the costs by assuming efficiencies that have yet to be determined is that it's going to cost at least $100 billion and probably about $200 billion. Interestingly enough, that's pretty close to what's collected or what's foregone by the tax subsidy. So I actually think you could get the money there, or you could get the money out of less spending on armed forces, but it is within the realm of possibility, although I think the swooning of the economy is causing that door to close pretty quickly unfortunately.

And I guess that's the last point I wanted to make, so thank you very much.

(Applause)

CHAIRMAN PELLEGRINO : Thank you very much, Dr. Pauly. We've asked Dr. Nick Eberstadt to open the discussion. Nick ?

DR. EBERSTADT : Thank you, Ed . Well, you've all just seen why Mark Pauly is widely regarded as a force to be reckoned with in the field of health economics. And, in fact, he's been a presence in the field, although his youthful vigor belies this, for fully four decades since his Ph.D. dissertation was completed. He's been interested in moral issues in health economics from the very beginning. His initial work, in fact, was on moral hazard in health economics and, as you know, when economists talk of moral hazard, they are discussing the incentives that are built into systems that encourage perverse or unwarranted, unwanted behavior. And then in his work, initial work, on health insurance in particular, he noted that there was a moral hazard involved in the potential structuring of health insurance, not only with respect to avoiding health insurance, but also in using too much health insurance depending upon the sorts of perverse incentives one could find. Today in discussing the questions of health costs and uninsured, I think he gave us a kind of panoramic outline there.

There are a couple of points I'd like to press him on for my own curiosity. In talking about health costs, of course, David Cutler and his group at Harvard have done a lot of work about perhaps excessive health costs and David Cutler is a formidable economist. Of course, Kevin Murphy is no mean economist himself at the University of Chicago . And Kevin Murphy and Robert Topel published a fairly recent study in which they tried to estimate the value of the mortality improvements and the longevity improvements to the United States since 1970, from '70 to 2000. As a very rough ballpark figure, their conclusion was that, even after you factored out medical expenditures, the net benefit to American society was in the order of tens of trillions of dollars. I think $60 trillion. You know, what's a few trillion among friends?

But I'd like to ask Prof. Pauly if he thinks that that would be comparing apples to oranges, comparing those sorts of estimated benefits of health improvements to the sorts of expenditure numbers we're talking about containing.

On the question of the health insurance crisis in the United States , Prof. Pauly 's contribution to the briefing book and his presentation here touched upon the whole question of the nature of the uninsurance, the lack of insurance problem. And I would like to invite him to say a little bit more about that.

Who are the uninsured? We get very different numbers, a difference of 10 million people — but, again, what's 10 million among friends — if we look at the Census Bureau's estimates versus the National Center for Health statistics. What's the profile of the people who are uninsured? What are the health consequences of being uninsured? Maybe we have an idea, an understanding, of what the consequences are if you're uninsurable. But what are the consequences of being uninsured after we attempt to control for demographic differences, ethnic differences, income differences, educational differences,

socioeconomic-status differences? Can we say much about what the health consequences are? And, of course, this is right up Prof. Pauly 's alley: Why are the uninsured not insured?

In my own approach to this as I try to think about this, one analogy to the health insurance question might be a housing question. Should we think of uninsured people in the United States as homeless or should we think of them as renters? Or some of each? If we compare homeowners to renters — well, if we compare homeowners to homeless people, we've got an enormous moral and ethical crisis on our hands. If we compare homeowners to home renters, we know that home renters for the most part have somewhat smaller, less capacious houses. They don't get equity, economic equity, in the system building up. But it might imply a different sort of approach here or at least a different set of questions.

And, finally, there's a question, I think, for our Council here. Mark Pauly has outlined a number of very difficult policy questions for the United States at the moment. But are these bioethical questions? We're not a council on healthcare policy or are we even a council on medical adequacy. We're a council on bioethics. Do we have more, as council members, to grapple with from Prof. Pauly 's very valuable and meaty presentation than the initial issue that brought him on the scene, the question of moral hazard? Thank you.

PROF. PAULY: Let me try answers for a couple of them. So I guess the question is, one of the questions is, what more do we know about the connection between increased spending and improvement in quality? As I suggested, not to be pejorative, but we already knew that there was going to be a big improvement in health outcomes for cardiovascular disease and stroke. There have been efforts still ongoing to try to look at the connection between increased spending and outcomes in cancer. My interpretation of those so far, although we don't have anything definitive, is that probably the increased expense pays for itself, but not too much more than that. So the ratio of benefits to costs seems smaller for that kind of disease.

I guess in terms of thinking about whether current health spending growth is like past health spending growth — I mean, the best thing — the most truthful thing to say about health spending growth is it's always something. And so for, actually, for the period of the '70s, the lead in health spending growth was increased spending on inpatient care, hospital care, largely stimulated, we think, by Medicare.

The other periods that are interesting to characterize, from about 1998 — I don't know where you draw the line here — but around 2004, the lead in health care spending growth was very definitely pharmaceuticals. Drug spending was growing double digits; in some cases, up close to 20 percent a year. That has dropped dramatically, and actually my undergraduates, I talk about health spending growth and don't people think something ought to be done to slow it, and they all agree, and then I say, "Well, but the reason it's slowing in pharmaceuticals is because the rate of discovery of wonderful new products has tailed off a lot. Does that make you happy?" And so I only do that on a sunny day so I don't make them too miserable, but that's sort of the dilemma there.

The scuttlebutt is what's taken over for drugs is medical devices, although the medical device makers keeps saying, "Don't blame us." But I think — well, it's easy to be nihilistic here, but part of the reason — let me put it in terms. If we knew much more about the relationship between spending growth and health, I think we'd be able to be much more definitive and probably much more accurate in terms of what we might suggest as policy remedies.

And I guess that takes me to the what do we know and we don't we know about the connection between insurance coverage and health outcomes. As I mentioned in my remarks, the most definitive work on that was done actually by looking at the impact of cutbacks in a Medi-Cal program on the health of poor people. And I guess I haven't memorized either the physical numbers or some dollar value, just to say that the impact was pretty stark.

There is some evidence, pretty good, that indicates that when people who are uninsured go on to Medicare and viewing that as a more or less exogenous change in their fortunes, that seems to improve their health. So those are kind of on the side of the angels.

I have tried to look at what impact insurance makes on the health of lower middle-income young women. I found that they had a lot more spent on them when they had health insurance, or spent a lot more on themselves, but that it was very hard to find any improved health effects. Part of that is because we don't really have very sensitive measures of health. So I am — well, I can only repeat what I said in my remarks. I think if we had better information on that, we might be able to be more persuasive.

In terms of how do we measure the uninsured, my summary judgment here is virtually any measure is bound to be a little squishy. But I think what's probably more important than the absolute level is the trend, and the trend is definitely not to the good and it's not as if the sky is falling all at once. But there is more or less an exorable trend upward in the count of the number of uninsured. The only time it really went down was when the Census Bureau which collects the data that's most usually cited changed their form to add a question at the end asking the respondent, "Now, you said so-and-so and so-and-so in this household was uninsured. Is that right?" And we made a million people suddenly become uninsured by adding that question. So I guess that goes to some of the softness of the measures.

There is a good policy question though I think which is, "Well, if there numbers were lower than we think, how would that affect policy?" My personal opinion of why Massachusetts was able to be first maybe — it's sort of eroding a little, but in doing something serious about the uninsured — was because they had one of the smallest uninsured percentages in the country. And so a major improvement was in reach there; whereas, in places where a quarter of the people are uninsured like Texas or California , as we already saw for California and probably will never see for Texas , they weren't even able to get it off the ground.

We do know some of the reasons why people are uninsured. In fact, we know a fair amount about that. The main economic reason is that if people are in settings where health insurance costs more for the same premium — or for the same benefit, they are less likely to have it. So we mostly can drive that off of employment-based group health insurance. So if you're in a large group where the loading of insurance is also quite low, pretty much in the ballpark of what Medicaid is able to do incidentally and probably in the ballpark of what a true estimate from Medicare would be, you get a very high rate of take-up of health insurance and a very high rate of offering of health insurance, in the 90 percent-plus level. As you go down both the firm-size distribution and the income distribution, people fall away. So even relatively high-wage people working in very small firms are considerably less likely to have insurance even if they're middle-class than if they worked in large firms. And then when you go to the people who don't have access or haven't put themselves in a setting where they have access to employment-based group health insurance and so have to turn to the individual insurance market, I actually — I do want to say I believe the individual insurance market is not as bad as a lot of people say, but it's not all that good, partly because, unless you choose a high-deductible plan with a health savings account you're not going to get a subsidy and in a way partly because it's the dumping ground for all of the hard-to-insure people that have fallen out of the group insurance system. So it's not — I probably should say I'm on the board of a small nonprofit health insurer in Upstate New York, but so I maybe have less of a belief about the intrinsic evilness of people who run health insurance companies. But I think the individual market in a way is interesting to me these days in part because it's peculiar, but in part because it is the vehicle that might need to be relied on, however flimsy it is and however bendable it is, for actually making some steps forward if that's going to be done through the private health insurance system.

CHAIRMAN PELLEGRINO : Thank you very much. Dr. Carson ?

DR. CARSON : Thank you for that very spirited presentation.

Certainly it's an extraordinarily worthwhile thing for us as a society to try to cover those individuals who are not covered. But in addition to covering them, I wonder what your thoughts are about, in the inclusive of that coverage, trying to change their behavior? And by that, I mean, you know, a large number of uninsured frequent emergency rooms and places that really are quite expensive and inefficient and, perhaps, creating systems that might encourage them to seek coverage in a clinic, which is much less expensive, but more importantly might be able to address some of the long-term issues that in the long run would save us a lot more money than just patchwork every time they get a problem.

PROF. PAULY: Well, so as you know, I'm skeptic that you can save money by spending money. But I do believe that there are lot of interesting possibilities about insurance design that could improve things from where they are right now. The most obvious thing to say is that health interventions which have a very high cost-effectiveness ratio and which, for some reason, people are disinclined to use when the cost-sharing is set at the typical level might be candidates for lowering the level of cost-sharing.

So even in the design of most of the health savings account, high-deductible health plans, there's usually an exemption from the high deductible for certain identifiable preventative activities: Immunizations for children; for people who have been diagnosed with high cholesterol or high blood pressure, the appropriate pharmaceutical interventions there. So some sense is, I think, trickling into the system.

I guess the best way to formulate my own position on it is to say the design of an insurance plan in terms of — and you can offer incentives for people or disincentives depending on what you do with the cost-sharing. That sounds to me like a great question that a market could answer with different health insurers imagining or trying out different sorts of ways of doing it.

I'm much less skeptical or much more skeptical, I guess, on trying to have a consensus of experts agree on the ideal level of cost-sharing for everybody. But I do think there's a lot of possibility for creativity there, some of which we're already starting to see.

Having said that though, I probably would — I have to mention, although I'm not sure whether I'm proud of it or not. Out of the, sort of, most important scientific and intellectual breakthrough, not only in health economics but economics in general, which was the Health Insurance Experiment, the RAND Health Insurance Experiment, which basically can be interpreted in a lot of different ways, and how it's interpreted depends on whether you talked to the physicians who worked on it or the economists who worked on it. But the economist party line, at least my interpretation of it — I think they wouldn't disagree violently — is to say, "Beyond providing catastrophic coverage, which everybody in the RAND experiment had, it doesn't look like a further lowering of the level of cost-sharing produces much, if any, of an improvement in health for the, sort of, average population."

The only subpopulation where there was a big improvement were low-income people initially of poor health especially with high blood pressure. But for the average American, at least especially the average middle-class American, I lay awake at nights worrying they don't even have catastrophic coverage. But if we could get catastrophic coverage to them, I'd regard that as a major accomplishment and then actually believe we could wait and see whether it needs to be improved beyond that point.

To some extent, at least my amateur political theory here sort of goes along with Altman's conundrum. When you try to design a health insurance plan that would be offered to people and maybe even mandated as in Massachusetts, you tend to get a Christmas-tree effect where every group wants to make sure their benefit is loaded on there, and then that tends to make it so expensive that either the uninsured can't afford it or the taxpayers think they can't afford it.

But I would certainly endorse the idea of trying to use cost-sharing both to direct people to things that look highly effective that they may be somehow underestimating and may be dissuading them from things that are really not worth the money.

CHAIRMAN PELLEGRINO : Any other questions? Dr. Hurlbut and Dr. Dresser .

DR. HURLBUT : Just to give us a larger overview of what all this money is going for in health care spending — you mentioned drugs, for example?

PROF. PAULY: Uh-huh.

DR. HURLBUT : I don't want to exaggerate this to mention in medicine, but I brought up earlier the changing arena of medical care. And so, for example, when I was a medical student, male-patterned baldness was not considered a disease and now it's a multi-billion dollar pharmaceutical dimension. Right?

And can you give us some idea of what, assuming this might become a larger portion of medicine, at the present, how large a portion of medicine is — and I've got several different categories here I'd like you to comment on. First of all, how much of increased health care has come from what you might call lifestyle drugs or elective procedures like cosmetic procedures? They estimate there have been 2 million breast augmentations in the United States . That's quite a lot of money, I suppose — and lifestyle drugs. Even in a certain sense abortion is an elective procedure, not a therapeutic intervention exactly in the typical form, although it's hard, of course, to qualify that because the expense of abortion would have ended up as an expense of obstetrical care. But to what extent is that? What portion of medical expenditures is that?

And, second of all, what percent is experimental procedures? I mean, at the major medical centers, the university medical centers, it's obvious to anybody that works there they're very expensive things taking place that are in a way experiments or explorations for the future of medicine. I remember when I was a medical student at Stanford, Norm Shumway was pioneering heart transplant. It's become somewhat of a mainstream medical procedure, but it might not of. Can you give us some estimate of how much of medical, what's lumped into medical, is that?

And also, finally, what percent is inordinately expensive procedures that, if you had a certain theory of how distribution of care should take place you might cut those off if — you see what I'm getting at?

PROF. PAULY: Uh-huh, yeah. Well, I won't take very long to answer because I don't know the answer. I do know that when people attempt to kind of classify things you always get arguments about it. My own guess is that, as a fraction of the total, the lifestyle and giggle products are probably not all that large. Partly that's because the fraction of the total that goes for all prescription drugs including the ones that affect health is only about 10 percent. So that's not where most of the money is going. For the other ones, about all I can say is, it would be useful to know the answers to those. I don't.

One thing I do know the answer to, so I might as well mention it, is what the people sometimes ask, is what about the expenses for people in the last year of life? And my interpretation of the data there is that the proportion of the expenses that is going to that subset of the population has remained quite constant over time. It's not that we're spending a lot more on heroic measures there.

But, you know, I haven't looked at the other end of the distribution of things that almost everybody would agree we could do without unless, of course, you're the professional association representing the people who do that thing. I don't know the answer to that one.

DR. HURLBUT : Can I ask a followup question then? So one of the things we could perhaps in our Council contribute to is some anticipatory comment in the sense that seeing in some sense where medicine is heading and we're obviously not going to make a dramatic effect on the immediate situation, but we might help frame the discussion for the future.

Do you see trends in medicine that have a specific relationship to what a council on bioethics might comment on? I mean, I don't mean to overemphasize this, but one of them does strike me as those issues related to medicalization of natural life processes, and that's just an example. But do you see others? Where do you see medicine going, and how might an ethics council comment on that so we might shape where medicine goes, not just how it's paid for.

PROF. PAULY: Uh-huh. Well, let me try to offer some thoughts on that. I don't have a coherent-like view of it, but one point is for people who are not poor, at least from an economist's perspective, we wouldn't be that interested in making a value judgment. It's their money, and as long as they know what they're doing, setting aside the tax subsidy, of course — yeah, I guess I wouldn't want the Treasury to be paying 50 percent of some of things, which it probably is. But setting that aside, the truest economist answer would be, if people want to spend their money on those things, that would be okay with us, and we might resist an attempt to impose some kind of external judgment on it.

For people's whose care is mostly supported by the government, ultimately by the taxpayers, though, I guess those are the right kinds of questions to ask, although I think any kind of a differentiation proposal tends to raise a lot of opposition. That hasn't stopped, of course, Medicaid programs from paying such very low payment rates that the level of access to a wide range of providers for that population is much lower than for the average person.

So, I guess, those would be the main answers. I do think in my vision of how a market arrangement might work would be one in which — and to some extent we talk a lot about why this hasn't happened — this sort of Alain Enthoven's vision — there should be a variety of different health plans including the really buff and really good-looking health plan or the, you know, good-enough health plan, the pretty-good health plan, and let people choose amongst them.

As I mentioned in my remarks though, perhaps for good reason, what we see in the market is really not much product differentiation in the private sector, not a whole lot of choice about a lot of things that might matter, and then in principle people could make choices about — at least my take on how to address the kinds of questions you're raising for the bulk of the population would be to say, "I'll invoke the sainted market as a way to settle that as opposed to having it judged by some external standard." You can't, of course, pass the buck on people who are being paid for primarily by taxpayers.

CHAIRMAN PELLEGRINO : Thank you. I have a request from three Council members to speak: Dr. Dresser , Dr. Meilaender , and Dr. Lawler . I announce this at this point because we're under a rather stringent requirement for being back here at 2:00 . There's a special problem with our afternoon speaker, Elliott Fisher , which I'll explain at the end of this session. So I would ask for the usual brevity.

PROF. LAWLER : I'll pass. I'm all right. I'll be glad to pass.

CHAIRMAN PELLEGRINO : Rebecca ?

PROF. DRESSER : All right. Well, these are not a big deal, so you can be brief. On your third-to-last slide, you were talking about your ideas. The poor would get free comprehensive insurance. As household income rises, the subsidy and the minimum required coverage both fall. Why would the minimum required coverage fall?

And then my other question was, it's embarrassing to hear what a great subsidy most of us get, and I'm ashamed. Is targeting the subsidies to people like us part of the realistic discussion? Is that something that might really happen, or is that just idealism?

PROF. PAULY: Okay. Well, so the — what was the first question again?

PROF. DRESSER : In your plan, you would have —

PROF. PAULY: Oh, what does that mean? So one way would be to describe the other end of the spectrum. You've got the poor person with the comprehensive coverage. At the other end of the spectrum, you have Bill Gates who, in my view, wouldn't need a subsidy and if he wanted to buy a high-deductible health plan, that wouldn't bother me.

So not that people would necessarily choose to do this as income rises, but I guess to state the other way around, I am actually quite nervous about high-deductible health plans owned by lower middle-income people because at least it seems plausible to me there that the out-of-pocket payment may deter them from doing things that may be good for them and maybe the rest of us would care about. So you could be permitted to choose a much less comprehensive plan if you're an upper middle-income person and you would get a much smaller subsidy, if any. That would be sort of the data point there.

Absent the tax subsidy, we think that those people, despite what your intuition might tell you, would actually choose considerably less comprehensive plans. The general conjecture is that the average out-of-pocket payment in a private plan which is now about 20 cents on the dollar, it might go up to about 30 cents on the dollar, so a 50 percent increase though in the effective price of medical care.

In terms of whether people are interested in this idea or, if targeted, it is in two of the candidates' plans, and I'd be interested in seeing how long it survives. John McCain actually proposes to abolish it entirely and to convert that plus a bunch of other things like Medicaid into a system of just a flat $2,500 tax credit for an individual, $5,000 for a family. I haven't been part of that, so I'm a little worried if the numbers add up. My calculations are you'd need about $500 million to do that and you'd only get about $200 million out of the tax subsidy. You'd have to find the rest of the money somewhere else. And then Hillary Clinton in her plan has a proposal, I believe, to limit the amount of the tax exclusion in some dimensions. So there is some talk about it.

As long as I'm talking about these things, I believe, probably because one of my Ph.D. students was working there, the Council of Economic Advisors came up with a wonderful idea that, of course, has sunk without much a trace, the impediment to removing this inefficient and inequitable subsidy. Well, at one point, a group of us wrote a proposal more or less along the lines that I described where we calculated that by removing the tax exclusion we could actually pay the tax cost and the subsidies in the proposal, and we further calculated what fraction of the population would end up being better off when the dust cleared where their subsidy would be bigger than — the subsidy from the new plan would be bigger than the subsidy they were currently receiving, and I think it was like 60 percent of the population was better off. I came to Washington , and they told me that was the wrong 60 percent. So there's a political resistance to taking away tax breaks much beloved by the middle class.

What the Council proposed to do was to cap the exclusion at $15,000 when offered that exclusion of health insurance premiums from taxation for everybody who bought the basic minimum level of coverage so that there was no incentive at the margin. If you were willing to settle for a high deductible health plan to go further, you didn't get a bigger tax break. And it was scored actually as budget-neutral on that because the limit was going to increase at the economy-wide rate of inflation rather than the health care costs.

So I believe there are a lot of possibilities for creativity in trying to do something about that mechanism. And at least one potential benefit from doing something about it is that you wouldn't hear as much from health economists. You know, we'd be satisfied.

CHAIRMAN PELLEGRINO : I think we might be able to work Dr. Meilaender and Lawler in. I'll take a little bit of the Chairman's prerogative to cut into the luncheon time, which is not a popular thing to do. So, Gil , I didn't make it attractive. Did I?

PROF. MEILAENDER : No. It's an invitation to be hated actually.

(Laugher)

PROF. PAULY: There's a moral hazard here though I need to point out.

PROF. MEILAENDER : Well, I'll ask my question really quickly then. I suspect you can answer it easily.

Right at the start with your two parts of reform, the cost-containment, the covering the uninsured, and you had your preference, but you said the two were independent, though you qualified that in certain ways. And just a word about that? I mean, it seems counterintuitive to think of them as independent, not just because you'd be covering more people, but at least in my mind the more I get, the more I want in a sense. And so just a word about the independence of these two aspects?

PROF. PAULY: Well, I kind of said they're independent and they're not. The main way that they're dependent is that the rise in the price or the total expenditure of health care for the insured middle-class, as I mentioned, seems to be exacerbating the problem of the uninsured at the level of the lower middle-class. The positive argument as to why it's not as serious as it seems is to note that the real cost to the economy of covering the uninsured — they're 16 percent of the population — isn't 116 percent of what we currently spend on health care or the total expense wouldn't rise to that level. And the reason is that the uninsured currently receive a fairly large volume of services, not as much as the insured. So I think the politician who says when we cover the uninsured it will actually be cheaper, that can certainly be true in specific circumstances, but as far as know in research that's not true across the board.

Nevertheless, so here's the arithmetic I do in my head. If the uninsured use about three-quarters of the medical spending that they would use if they were insured and they're 16 percent of the population, the real cost to the economy of covering the uninsured would be 4 percent of medical care spending, which is within the measurement error.

But the story I told about how hard it is for the taxpayers to get the other 12 percent back is probably one of the impediments. I mean, I do think that, as I've hinted — well, I guess I don't want to water this down too much. But the price tag for at least making a substantial dent in the number of uninsured that exists right now is probably something that could feasibly be paid by both the health care system and by the tax system.

So in that sense, we needn't hold off covering the uninsured for some method of slowing the rate of growth of health spending. Of course, it is true — and I suppose this is what you're thinking of — that once we cover the uninsured entirely we in a sense put in the public sector the same problem we already have for covering all of senior citizens, that as their health care cost rises that gets put into the tax system and that raises some problems.

So to some extent you have to have a willingness to ignore the long run because the long run seems so impossible. But I think that might be left to its own time.

CHAIRMAN PELLEGRINO : It's your option?

PROF. LAWLER : I'll pass.

CHAIRMAN PELLEGRINO: Thank you. The reason we're wanting to be somewhat more strict on time than we usually are — and we try to be strict — is that our afternoon speaker, Dr. Elliott Fisher, has a personal problem. He has a daughter who is in the hospital at the present time, and I think very properly he is being with her rather than with us in presence. However, modern technology has made it possible for him to engage in a telephone conference with us, and we've agreed to that. And he will present his slides as well and have comment, so that he will not be there, but he must begin and we must begin at 2:00 o'clock .

So have a good lunch. I think you've only lost a few moments. Thank you, Peter, for your diffidence. We appreciate it, and have a good lunch.

(Luncheon Recess 12:31 p.m. - 2:10 p.m. )