HEALTH POLICY DISCUSSION
Is Inequality Bad for Our Health?

Thursday, October 11, 2001

Transcript prepared from a tape recording.

Agenda:

9:00 a.m.	Registration and Continental Breakfast
9:30	Presentation:	Jeffrey Milyo, University of Chicago
	Respondents:	M. Gregg Bloche, Georgetown University
		Nicholas Eberstadt, AEI
		Michael McGinnis, Robert Wood Johnson Foundation
		Sally Satel, AEI
	Discussant:	Newt Gingrich, AEI
	Moderator:	Robert B. Helms, AEI
11:30	Adjournment

Proceedings:

MR. HELMS: ---cancelled last month, as lots of things in this town were. The calendar in the morning news reminds me that it's exactly a month since the events of September the 11th, and they were occurring at exactly this time. So a lot has changed since then. This program was planned before then, but we still are trying to push on and address a number of important health policy issues.

Today we're looking at the issue of--it's our title: "Is Inequality Bad for Our Health?" As the introduction says, this has generated quite a body of literature now, a number of studies and people giving different interpretations of this and so on.

One thing that this reminds me of, I think a lot of times those of us who are dealing with policy issues and academic research and so on get very frustrated with sort of the policy process here in Washington, sort of, Why don't those people ever listen to the evidence, and so on. But on the other hand, a lot of times maybe they listen too much. There's that side of it, too. I think a lot of this issue I think illustrates that a lot of academics are doing things and they have sort of no real idea about how they're going to be interpreted and used in the public policy process.

So that works both ways. It's sort of what's convenience for the policy process, I guess, or the political process that determines what gets taken seriously.

Last week I had the opportunity to attend a conference on addressing sort of the non-health determinants of health at Princeton University, and not being an expert in this field, I went because, one, it was a good way to, I thought, catch up and learn some things. But there were several very good people involved in the program, so I wanted to hear what they had to say. They had separate papers on problems of drugs, alcohol, tobacco, and so on. A lot of--mostly economists, not entirely.

But I guess I was somewhat--I didn't know exactly what to expect from this conference, but somewhat surprised about how skeptical everybody was about sort of the ability of researchers to even determine things in this whole area, even in areas having to do with tobacco and alcohol. And one of the things they pointed out, I think, that impressed me was a lot of times the statistics we have about health outcomes and the various determinants of these things, income even, and so on, just don't--people keep saying, you know, that the numbers we have, the available theories and so on, don't always mean what people think they mean. And so it's very difficult to do good research in this area.

One of the papers presented at this conference was by Angus Deaton, a person who's written a lot about this topic. And he did kind of an overview of the literature and so on, and I just thought from the conclusions there were a couple of things I thought would be interesting for this, as an introduction to this topic.

After going through a discussion of all the literature about the relationship, you know, what we know about what affects health, he said, "I find it hard to argue that an exclusive focus on health inequalities makes much sense, especially one that targets ratios of mortality rates across different groups."

Then he goes on: "What about more general health policies that refocus attention away from health care and health-related behaviors and toward education and income? This seems a much easier case to make, and it is not hard to believe that the current U.S. system pays too much attention to health care delivery and drugs and too little attention to the effects of health on the upstream social and economic arguments." And he thinks the case for education is stronger than that for income.

He said, "In poor countries, a policy of income provision to the poor may well be more effective than spending the same amount of public funds on a weak health care delivery system," and so on. But I just thought those were interesting conclusions.

We're very fortunate to have--if I can get my list here, and so on. As our usual custom, we're not going to spend a lot of time introducing people. You have the bios of people in your packet. We're going to start out with Jeff Milyo, if I can find--Jeff is the assistant professor in the Irving B. Harris School of Public Policy Studies at the University of Chicago. So, Jeff?

MR. MILYO: Thank you. So is inequality bad for our health? To the uninitiated, that might sound a bit odd, so let me just take a moment to explain exactly what's meant by that.

The inequality hypothesis states that inequality, and in particular income inequality, has a causal effect on population health, quite apart from other relevant factors, such as poverty, lifestyles access to health care, quality of health care, et cetera; so that even if you controlled for individual behavior and attributes of individuals, you'd still find that if a person happened to live in a community with a lot of inequality, that their health would be worse than if they happened to live somewhere else but all their other attributes were the same.

So think of inequality as you might an environmental pollutant which, quite apart from your own behavior, has an effect on your health. And just like an environmental pollutant, it's quite plausible that the effects would be strongest for the poor because they have other disadvantages that might compound these effects. But the idea is inequality is bad for everyone's health, but perhaps especially bad for the poor.

This inequality hypothesis is most definitely a hot topic in the public health literature and the medical sociology literature, or as the kids back on campus say, people are really jazzed about it.

There have been scores of studies in prominent journals, such as the American Journal of Public Health, Social Science and Medicine, the British Medical Journal. A few important books now are out. Probably the best known is by an English scholar, Richard Wilkinson, called "Unhealthy Societies: The Afflictions of Inequality." And in a way it's quite fitting that Wilkinson was one of the early proponents of this hypothesis in that a lot of the early research appeared in the British Medical Journal because this income inequality hypothesis is in keeping with a long tradition of British scholarship, which stretches from Piltdown Man to crop circles. That's a little foreshadowing.

As I said, this is a hot topic, and a number of--part of the reason it's hot is because of the policy implications, and a number of scholars have really jumped to the policy implications, skipping right over any cost/benefit analysis just to the benefits of redistribution and the health benefits of redistribution. I think we'll have some time to talk about that later.

The inequality hypothesis is observation driven. It wasn't the case that there was some theory out there that suggested inequalities would be bad for people's health. Instead, there were a number of observations, and people connected the dots and said, How do we explain this?, and came up with this inequality hypothesis.

So there are four key empirical claims, and the first is that socioeconomic status, no matter how it's measured--race, education, income--is correlated with health outcomes. The better your socioeconomic status, the better your health. And that correlation often holds up even when you control for other factors. In particular, income is strongly associated with health outcomes when you look across individuals. So that's the first observation.

The second observation is, despite this relationship between income and health across individuals, there isn't such a strong relationship across countries. If you look at national income and compare differences in population health across countries, measured by, say, infant mortality or life expectancy, at least when you're looking at the subset of developed countries, you don't see such a strong correlation between GDP and health outcomes. And so this is something of a puzzle. If income is associated with individual health, why is it not associated with population health when you look across countries? And this is, I think, suggestive that perhaps it's not so much the material advantages of income that matters for people's health, but your relative standing within a society that matters.

So that when we see this relationship between an individual's income or, more generally, socioeconomic status and health, the reason that there's a correlation is because we're getting information about where a person is in a social hierarchy. So those at the top do better regardless of whether they're in a relatively poor developed country or relatively rich developed country, that somehow relative positioning within society matters for one's health. As I said, this is just suggestive.

The third empirical claim is meant to support this, which is that specific studies of social hierarchies seem to support this notion. The most famous is the British Whitehall Study where 17,000 British civil servants were followed. They were surveyed about their behaviors, such as smoking, et cetera, and their health outcomes were followed. And what was found is even when you control for some other relevant factors--education, smoking, sedentary lifestyle--you find a lower life expectancy for civil servants in the lower ranks of the civil service.

Well, it's not random how people get assigned into the civil service, so rank in the civil service might be correlated with other things. But you can do experiments on lower primates, and so there are citations to some of the literature where people have taken groups of monkeys and removed a dominant monkey or added a dominant monkey to a group, and the claim is that this has health consequences for the monkeys.

These three empirical claims, then, are taken to be very suggestive that there's something about social hierarchy that affects our health, but the real kicker is the fourth empirical claim, and that is that income inequality, no matter how you measure it, is correlated with health outcomes, no matter how you measure them, across countries, states, metropolitan areas, and I've even seen the claim that the correlation exists across postal codes--although I have to say I haven't seen the paper. I've been trying to get it for a year, and I can't even get in touch with the author. So let's hope he didn't wander into a bad zip code or anything.

But from these four empirical claims, proponents of the income inequality hypothesis say that we sort of are perhaps reluctantly but inexorably forced to conclude that perhaps there really is something about relative positioning in a society that affects one's health. And so how might that happen? And that's where this income inequality hypothesis comes in.

There are two causal mechanisms that have been suggested. The first is that social hierarchies affect the stress levels of individuals and that being particularly in the lower rungs of a social hierarchy is stressful and that added stress can eventually lead to cardiovascular disease or perhaps leads people to seek releases in self-destructive behavior--smoking, drinking, driving too fast, et cetera. And so that might be a link to health.

Clearly, this would seem to affect the poor the most, but because a lot of these activities have externalities, it may affect others as well.

That's the more direct link between inequality to health. The more indirect link, it's been hypothesized that inequality or living in a society marked by inequality causes people to view their neighbors as some alien other, different from me. And so there's less concern for the well-being of these other people. If you like, you can say that inequality puts people at risk of voting Republican, sort of less social provision, less of a social safety net, and so that's the mechanism by which this is hypothesized to happen.

So what I have done with my co-author, Jennifer Miller, at William and Mary, is write a series of papers where we really tried to take a closer look at these arguments. And on those first three empirical claims--SES and health, GDP and health, and social hierarchies--I'm happy to talk about those more, but let me just summarize our view of those, which is that that evidence is actually a bit exaggerated, there's a bit of an academic game of telephone going on where, with each retelling, this evidence seems to be more and more impressive. But I can spend time talking about that later, if you'd like.

I want to focus on the fourth empirical claim, which is really the key, this incredibly robust association between inequality and health, and look at that a little bit more.

First of all, there is a conceptual problem with this evidence. It's not clear what the comparison group should be. When we say inequality is bad for one's health, inequality among whom? Should we be looking within a neighborhood and measuring inequality? Should we be looking at the city level, the county, the state, the country? Should we be looking at some peer group, people with the same education or in the same profession? It's not obvious. And how inequality is measured across what group actually makes quite a difference.

Secondly, there's something known as the ecological fallacy, and I don't want to get into anything too technical other than to say that relationships that exist at the individual level are not always manifested in aggregate data when you're looking at, say, inequality at the country level and population health and vice versa. Sometimes you can find correlations in aggregate data that are not reflective of relationships in individual data. So as researchers, we really where we can prefer to test hypotheses about effects on individuals using individual-level data. And while more recently people have been doing that with the inequality hypothesis, for a number of years people were only really looking at aggregate-level data.

A third problem is what we might call unobserved factors. Inequality isn't dropped on a community out of the sky. It's not a true treatment effect. There are things that cause inequality. Some of them are easy to measure, such as migration and the age composition of a population and how many single-parent households there are. But some of them might be hard to measure. There might be social norms about two parents working in a family, et cetera. So we need to account for a number of other things that might cause inequality to be spuriously correlated with health outcomes.

And, finally, as a conceptual problem, the causal pathways that have been hypothesized would seem to suggest some time lag or some exposure required. The idea that inequality causes psychosocial stress which leads to cardiovascular disease doesn't suggest to me that we should be looking at contemporaneous correlations between inequality and health outcomes but, rather, some lagged correlation or some changes in inequality over a long period of time, if that has any health effects.

So in a series of papers, Jen Miller and I have tried to address these conceptual problems. And we've done three types of studies. First, we have looked at the cross-country aggregate data types of studies, and let me tell you what we found there. When we looked at the previous literature, what we found was that most studies were using data from around 1970, and they used few, if any, control variables. Many were just looking at bivariate correlations between inequality and measures of population health, either life expectancy or infant mortality.

So what we did was we used--looking at several decades, 1950 through 1990, and see if we can still find these same correlations when we look over a longer time period or we look at a larger sample of countries or smaller sample of countries. And we actually found that these correlations are sensitive to both the time period you examine and the number of countries. But I think more importantly, if you include some relevant control variables, for example, education levels in the population, that completely swamps this correlation and significance disappears.

And perhaps most importantly, when we examine changes in inequality across countries, we actually find the opposite relationship. So I wouldn't want to make much of this evidence for a couple of reasons, because we don't think much of the methods that people were using in the first place to do this, so we don't want to overstate our findings here. But if you believe the evidence that other people found simply making these cross-country comparisons, what we find is if you look at 10- or 20-year changes in income inequality across countries, you'll find that there are actually--where there have been increases in inequality, there have been improvements in life expectancy and infant mortality. So take that for what you will.

Now, many of us are kind of shy about these cross-country results because we don't think that the income inequality data are very comparable across countries. So let's put that aside.

The second kind of study we did was to revisit these cross-state results. We also look across counties and metro areas and find very similar things, so let me just speak to looking across states. And most of these studies really look at data that's around 1990 and don't look at other times periods.

Let me just show you two pictures. You don't need to see them very clearly. If you can just distinguish which states are darker, those are states with higher age-adjusted mortality. And you'll see that folks in the South fare worse. And let me show you another picture, and it looks almost the same. This is the Gini coefficient for household income, a measure of inequality, and the darker the shading, the higher inequality is. And so there's clearly a correlation across states between inequality and age-adjusted mortality. It's also clearly region-based. There's something different about the South versus the Midwest or Northwest.

Actually, if you take any course work in health economics or health policy, one of the first things you'll learn is that there are strong regional variations in health outcomes, and it's thought to be attributable to many things. There are differences in physician behavior across regions but, more importantly, there are differences in lifestyle, diet, et cetera, across regions. And we think these things are very important for health outcomes. So it's quite possible that this state-level correlation between income inequality and health might be attributable to these other regional differences.

So what we did, again, we looked at decade-by-decade data, 1950 to 1990, over U.S. states, and we simply tried to replicate the kinds of things other people had done using just 1990 data. And we can do that. We can get similar results if we use no control variables. But once we put in some control variables, either for racial composition of the state population or education levels of the state population, this correlation is not significant anymore, often changes sine.

I should say as a dependent variable we're using all-cause mortality as well as specific causes of mortality, such as cardiovascular disease, suicide, et cetera. And I think, again, most important, what we find is when we look at 10-year and 20-year changes in income inequality in the states, those changes in income inequality are, again, correlated with health outcomes but, again, the wrong sine for the income inequality hypothesis. And we actually find that 10- and 20-year changes in income inequality, the more income inequality there is over time within a state, the fewer deaths there are from cardiovascular disease. And this is quite contrary to the income inequality hypothesis.

I don't want to assert a causal relationship there, but other people have asserted causal relationships on this kind of evidence. We're just showing that we can actually reverse the sine or the association quite easily.

One of the reasons I'm being so cautious about asserting causality is that the two types of studies I've mentioned are these ecological studies. They're using aggregate data. We prefer to use individual-level data when we can. And so the last class of study I want to mention, we have used individual-level data from the current population survey. The people are asked to report their health status, so it's a self-reported health status, subjective, but it has been shown to be correlated with health outcomes. We use that as our dependent variable.

And if we don't use any other control variables, we simply look at the relationship between self-reported health status and the inequality in the community where you live, whether it's at the state, county, or MSA level, we will find an association consistent with this income inequality hypothesis. Living in areas with greater inequality is associated with lower health status, and significantly so.

When we add in control variables, then, based on individual attributes, income and, importantly, a nonlinear effective income, allowing for diminishing marginal benefit, and education, martial status, health insurance coverage, et cetera, this relationship is greatly attenuated. The coefficient falls by as much as two-thirds. But depending on your measure of income inequality, sometimes it's still significantly significant. So adding individual attributes alone didn't quash this association, or at least not consistently so.

But then when we finally account for these regional differences, then this association disappears. How do we account for these regional differences? Well, we don't have data on, in this survey, sedentary lifestyles, diet, et cetera. So what we do is we try a couple of different methods. One is to simply include a dummy variable for what state do you live in. We have several years of observation. So in that way you're controlling for anything fixed but unobserved about the state. If you do that, you quash this association.

The second thing we do is to be a little bit more conservative about this. Instead of having a fixed state effect, we have a fixed region effect, and so I think we use nine census division dummies. So we're simply saying, Are you in the Southeast? There's something about the Southeast that's different. People eat a lot of beignets. Maybe they do it because of the inequality, but maybe they do it because it's tasty. So once we include these division effects, again, it completely quashes this association between inequality and health.

There have been some other studies using individual-level data that are not yet published or are forthcoming, and I should just briefly mention those. Angus Deaton, as was said in the opening, well-known labor and development economist at Princeton, has done a number of studies looking both across countries and states, and he is also skeptical of the causal relationship between inequality and health.

Ellen Mira (ph) at Harvard Medical School has used a data set of four million births and has very good control variables on the attributes of mothers and has found that there's no relationship between inequality and infant health outcomes. And what's nice about this study is, because she has such detailed data, she doesn't necessarily have to sort of resort to these state effects or division effects. Just including individual attributes, because she has so much information, seems to cause this apparent relationship between inequality and health to disappear.

And then, finally, a study by Graysons(ph) and Stern--one of the authors is here today--from Rand finds--they look at a number of different medical diagnoses and, similarly, they don't need to resort to state fixed effects or some sort of area fixed effects because they have a very rich data set which includes a lot of information on individuals. And I look forward--actually, that study is coming out in the British Medical Journal, and so I'm sure there's going to be some responses to it as well. Those will be fun to read.

So what do we make of this in the end? I would say that this inequality hypothesis has been something of an ongoing case study in junk social science, that it was an intriguing hypothesis but really by now we know enough that there's really strong reason to doubt any causal relationship between inequality and health. Nevertheless, there continue to be studies published, and people who are strong proponents of it, they're well funded, and so more power to them.

I have a few other ideas for where future research might head, but why don't I let some others speak, and we'll revisit those if you'd like to.

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

What we'll do now is, if you see on your program, we have respondents and one discussant. And so we're going to just go right down in alphabetical order with our respondents, and then we'll let Mr. Gingrich wrap it up for us.

So our first discussant we welcome back. He's participated in this series before. Gregg Bloche from Georgetown University Law Center, and you can speak from there or the podium. It's up to you.

MR. BLOCHE: I guess I'll wander up there.

MR. HELMS: Okay.

MR. BLOCHE: When Sally raised the possibility, when Sally Satel raised the possibility of my participating in this event today, I thought the best part of her invitation was her plea not to do a PowerPoint presentation. I am a fellow Luddite on this, and proud of it.

It's, of course, routine for physicians to do slides and PowerPoint presentations, but not for law professors. We law professors know that a picture is worth a thousand words. We just prefer the thousand words.

[Laughter.]

MR. BLOCHE: There's been lots of words, as Jeff points out, said and written recently in both scholarly and popular publications about the relations between health and the social and economic pecking order in America and abroad. The World Bank and even some corporate executives are taking these relations quite seriously.

There's the growing realization that health care, despite the fact that we spend in excess of 14 percent of our gross domestic product on it, there's the growing realization that health care has less influence on health than a variety of behavioral and environmental factors that are not well understood. And this all is making political conservatives quite nervous.

The leaflet for today's forum reflects this apprehension. There's a reference to the "far-reaching implications" of the hypotheses being discussed today, including "restructuring of..."--I'm throwing in an ellipsis--"our economic system."

Why this nervousness? one might ask. After all, conservative don't much fret over the unequal distribution of economic desserts in general. CEOs and tax lawyers and, once upon a time, Internet entrepreneurs earn many times more than the rest of us. They buy BMWs and huge houses, and the rest of American society accepts this. There's no worry among economic conservatives about the possible "restructuring of...our economic system" to ameliorate the unequal distribution of driving pleasure.

The answer, of course, is that health is different than driving pleasure. No country assures an automobile, let alone a BMW, to every citizen. But every industrialized country in the world, except for the United States, provides universal access to basic health care coverage, even though it would seem that health care coverage doesn't make much of an aggregate impact upon health.

And even in the United States, there's widespread sentiment that health and health care are goods that should not be distributed on the basis of the existing distribution of wealth; that unlike BMWs and the great Cabernets, health and health care ought to be distributed more equitably.

In economics terms, health is a merit good, a good that many believe should be disseminated apart from the existing unequal distribution of wealth.

So the thesis that health itself is a product of the existing distribution of wealth poses a political challenge to the current distribution of wealth and to the institutions that support it.

The incentives for conservatives to challenge this thesis are clear, and Jeff has done so, I think ably, indeed audaciously, but only partially. Jeff has done a beautiful job of unstuffing a straw--I can't say "man" in the PC place where I teach--so straw person, and he's persuaded some, including me, that inequality per se--and I underscore "per se," as Jim Haus (ph) does in his commentary on Jeff's wonderful paper in a recent issue, the June issue, of the Journal of Health Policy, Politics, and Law. He's persuaded me that inequality per se has not been proven to adversely affect population-wide health.

The studies on this topic, as Jeff's work points out, are equivocal. But let's be clear about what Jeff has not done, and let's be clear about what Jeff has not said.

He has not challenged the basis thesis that health status is a function of socioeconomic status, that health status is a function of education, income, and wealth.

One could draw--and I'll really be a Luddite here. I'll be a kind of human graph. One could draw basically a curve [inaudible] have income along the horizontal axis and health along the vertical axis. Income can be wealth. Income can [inaudible]. And the curve that I think most people [inaudible].

MR. : You need another arm?

MR. BLOCHE: Yes, I do.

[Laughter.]

MR. BLOCHE: So the curve is kind of like this, right? A diminishing marginal [inaudible] in terms of health, as income or wealth or whatever [inaudible] on the X axis. It's a long--it's a concave, downward curve.

And so redistribution from the well-off folks up here to the worst-off down here, redistribution from the well-off to the worst-off, without a change in average income--and there's a key--without a change in average income will increase population-wide health. And, in fact, there's powerful evidence published online just earlier this week, a study by Ann Case and her group at Princeton. South African blacks in the early 1990s suddenly became eligible for pensions that were previously paid only to whites. And so there was a sudden infusion of a large cash subsidy to the households of pensioners. And in subsequent years, her study shows, there is a big improvement in multiple measures, multiple indices of health status.

On the other hand, population-wide health is itself a function of average income, and public policies that reduce average income thus risk reducing population-wide health. And as conservatives love to point out, aggressive income redistribution risks eroding incentives that drive economic growth and that, therefore, increase average income.

So it's important from a health perspective that we exercise a measure of restraint on our well-meaning redistributive impulses. It's also important from a health perspective that we target the redistributive public spending that we engage in toward those programs for the poor that achieve the biggest bang for the buck.

And this points to a vital research agenda. There are a lot of serious questions, including the question of how does being at the bottom end--

MR. : [inaudible].

MR. BLOCHE: How does being at that bottom end reduce health? Is the problem psychosocial stress or education or a lack of health care? Is it a matter of difficulty affording life's material basics--decent food and housing? Or do both poor health and poor economic status flow from some third factors, as Jeff and others have suggested, maybe maladaptive character qualities or behavior? Or does poor health lead to low economic status? Is the causal relation reversed?

We need research that's aimed at identifying the pathways of this linkage with an eye toward shaping public policy intervention.

Now, this agenda for public health research and advocacy is different--and I think Sally Satel's comments are going to address this--from public health's more traditional focus on changing individuals' disease-causing lifestyles and behavior. And this traditional approach has achieved a lot, most famously, a dramatically decreased incidence of many infectious diseases during the early part of the last century, even before the antibiotic revolution, and within the last ten years, in recent years, a decreased incidence in AIDS; and also a decreased incidence in cardiovascular disease. And shouldn't personal health be a matter of personal responsibility if we're to avoid the nanny state? If brie eaters who are bungee jumpers get the fate they bargain for, why should that be our concern?

In the 1996 presidential campaign, Candidate Dole challenged the conventional wisdom that tobacco is addictive and argued that people choose to smoke. Dole took a hammering, quite a hammering, from addiction specialists and the press, and he quickly backed off this perceived campaign gaffe.

But Dole had a point. You can tell alternative, autonomy-regarding, and deterministic stories about a vast array of health-related behaviors and lifestyles. There's brie eating as a complex neurophysiologic reaction to the presence of brie, or there is brie eating as a personal choice. You take a long-term risk in pursuit of short-term pleasure. And much of American politics, from social welfare policy to criminal justice, is about the choice between alternative approaches to characterizing human behavior.

From a public health perspective and an economics perspective, there's a pragmatic way to frame this choice. Which narrative, deterministic or autonomy-regarding, offers the potential to achieve the biggest bang for the buck in making people's lives better?

If, just to take one possible example, longitudinal research studies find that public support for, say, universal preschool or other social programs achieves large improvements in various indices of population health status, that's convincing to me. Yes, there are other values at stake. There's a regard for personal autonomy and responsibility as an end in itself, what conservatives tend to stress, and regard for a measure of equity as an end in itself, what political liberals tend to stress.

But the basic structure of pragmatism that I'm talking about can, I think, frame a rational policy debate about the relation between inequality and health, and we shouldn't, I submit, get distracted by the straw person of whether inequality per se, the gradient alone, ignoring the reality of big differences in average income, education, et cetera, makes a separate difference in health status.

Thank you very much.

MR. HELMS: Thank you, Gregg.

We now turn to Nick Eberstadt, who is the Hinder- (?) scholar in political economy at AEI and a member of the Harvard University Center for Population and Development Studies. Nick? There's a lavaliere mike if you want to use it.

MR. EBERSTADT: Thank you. Let's try it that way. Can you hear me?

I'm going to try to be a little bit quantitative for the next couple of minutes and to examine the proposition that inequality is bad for our health. I'm not going to look at the proposition that inequality leads to lack of social capital and thus to poor health, just offer you a little comment about that. People who think that inequality leads to lack of social capital should tell it to the Marines. I have a brother-in-law who's a gunnery sergeant. The Marines are not exactly an egalitarian organization, but as for cohesion and trust I think at this point they are probably second to none.

I'd like to examine with you the proposition that income inequality leads to poor health outcomes, and this proposition actually did not originate in the public health literature. It originated in the economic development literature and the demographic literature back in the 1960s and 1970s with work by development econometricians like Irma Adelman and other people, who adduced data and results which seemed to confirm the common-sensical proposition that, other things being equal, a society with a more unequal distribution of income or consumption would have worse health outcomes, higher mortality.

I thought that I would revisit some of these data, some of these comparisons with the new and improved information that we get from the Hylaga (?) , the World Development Report, and from the World Bank's World Development Indicators. There's a CD-ROM which contains data on over 200 countries and places with respect to economic and social characteristics.

It is a well-known fact that income and health outcomes correlate both locally and internationally, and if we take a look at the latest data from the World Bank, we see a very strong international correspondence between per capita [inaudible] income, output, per capita output, and life expectancy at birth.

It's a fairly robust relationship [inaudible] 66, and if you didn't have an AIDS epidemic in Africa, you'd get a higher correlation there.

By the same token, if one compares per capita output and infant mortality, there's also a surprisingly tight relationship [inaudible]. I've done these graphics on sort of a semi-log basis, which is to say that [inaudible] for every additional 11 percent per capita output, you get about another year of life expectancy [inaudible] every 5 percent additional per capita output, you get a point knocked off of your mortality rate.

Well, if inequality affects health outcomes in the first way that Jeff talked about [inaudible] disputed--disputed examined as carefully as it should--we would expect that the societies with the most unequal distribution of economic benefits would be underperformers against their incomes; that against their incomes they would have lower levels of life expectancy and higher levels of infant mortality than would be predicted by income alone. And, conversely, we would expect that countries with more even measured distributions of income would outperform their level of per capita income with respect to life expectancy and infant mortality results.

So I went through the World Development Report and looked at their data on about 130 different countries' distribution of economic benefits. When you look for a measure of economic inequality, you have to be careful what you're exactly measuring. There are two very different types of data that are adduced and collected in the World Bank. One is data on inequality of consumption; the other is data on inequality of income. Very different. Don't compare apples and oranges.

But what are the most uneven and the most even countries listed in the World Bank [inaudible]? Taking the genie out of the bottle, pick up the Gini coefficient [inaudible], and you'll see them listed here by per capita consumption and [inaudible] per capita income. [inaudible] unequal countries, but per capita consumption measures, according to the World Bank [inaudible] Sierra Leone, Central African Republic, [inaudible] Zimbabwe, all countries [inaudible] lack of social capital in Sierra Leone [inaudible] countries with low income [inaudible] like Rwanda [inaudible]. Similarly, looking at per capita income [inaudible] most uneven distribution of per capita income being Guatemala, Paraguay, [inaudible] Mexico, [inaudible] Denmark, Sweden.

What do we find when we look at the predicted health results for these different countries? Well, let's look at predicted life expectancy and consumption [inaudible]. Underperformers would be countries where the actual life expectancy is less than the income-predicted life expectancy and with the most unequal consumption [inaudible]. The differences are pretty stark.

There are some underperformers, by the way, in the most equal societies [inaudible] but the expected [inaudible] -- [tape ends].
 -- consumption-based measures.

Let's look at per capita income data. [inaudible] problem. Guatemala, Paraguay, Colombia, Chile, and Mexico are the most unequal countries with respect to Gini coefficients measured on a per capita basis, but they also consistently outperform with respect to life expectancy the place where they would be predicted to be on the basis of income alone. And we've got another whole big problem down here in Europe. France, Denmark, Belgium, and Austria are all performing rather more poorly with respect to life expectancy than they would be predicted to on the basis of their income alone.

What about infant mortality? We get the same story down here. If you look at the consumption-based measure, the most unequal countries have a much higher infant mortality rate than they would be predicted to have on the basis of per capita income [inaudible]. And the most equal [inaudible] in terms of per capita consumption have lower infant mortality rates [inaudible] performers.

But if we get back to per capita income, [inaudible], Guatemala's doing better than it's supposed to, Paraguay's doing better than it's supposed to, Colombia is doing better than it's supposed to, Chile's doing better than it's supposed to. Mexico is not. Austria, Denmark, Belgium, and Sweden are all poor performers by this measure.

So I should conclude by telling you that using international economic data, I can prove to you that inequality is bad for your health. I can prove to you that inequality has no impact on your health. Or I can prove to you that inequality is positively good for your health, which suggests to me, in conclusion, that this is a doctrine that is continuing to be in search of data.

Thank you very much.

MR. HELMS: Okay. A few facts inserted into this. Thanks, Nick.

We now go to another old friend, Michael McGinnis, who is a long time in the Public Health Service, an expert on prevention and so on, but Michael is now a senior vice president and director of the Health Group at the Robert Wood Johnson Foundation.

DR. McGINNIS: Thank you, Bob, colleagues. It's a treat for me to join in this distinguished group, and also to share the panel with two of our RWJ alums, Jeffrey and Sally--Jeffrey from our investigators program and Sally from our health policy program here in Washington. So we're--

MR. : Three.

DR. McGINNIS: Three. Gregg, too. Any more out there.

[Laughter.]

DR. McGINNIS: That's terrific.

I should start with a caveat, and that is that I have neither particularly strong views on this issue nor am I particularly well informed. I suppose it's a good thing that since I'm not well informed, I don't have strong views. But--

[Laughter.]

DR. McGINNIS: But I'm happy to comment from my vantage point. Actually, I'll be commenting from my perspectives resident in three arenas: first, from my perspective as a fan of science interested in the question at hand, that is, is inequality bad for our health? I use the term "fan of science" quite carefully because I'm not a researcher in the arena, but I'm certainly interested in the application of evidence-based--evidence to policy and basing policy on the best information we have.

Secondly, from my perspective as a health professional, I am a physician by training, wondering if the question as a health professional is relevant to me.

And, thirdly, from my perspective as a funder, exploring the nature of what our interests might be as a health philanthropy in this set of issues.

On the first issue, the notion of evidence-based policy, obviously the question is: Does this particular issue meet the traditional standards of evidence, standards of proof that we set out for our various policy decisions or for our personal actions, for that matter?

Standards of proof are pretty well defined and have been referenced implicitly along the way, but let me just review those in seriatim and very short order, which we used when I was in government. As Bob indicated, I worked in government for quite a while, and the sorts of standards that we used in the development of various Surgeon General's reports, issuing findings and recommendations around important health policy issues.

The first is the consistency of the association. The second is the strength of the association. The third is the specificity of the association. The fourth is the degree of exposure. And the fifth is biological plausibility.

Without going exhaustively through these, consistency of association, that is the consistency from evidence category to evidence category, biochemical evidence, animal evidence, epidemiologic evidence, clinical evidence.

I gather from what I've heard today and what I read in Jeffrey's fine analysis that there are some intriguing hints, but the evidence is pretty mixed at this point.

With respect to the strength of the association, that is, do all individuals who are exposed find themselves with similar outcomes all of the time, the answer is clearly no.

With respect to the specificity of the association, that is, a single cause yielding a single outcome, for example, the exposure of tobacco and the result of lung cancer, this is a very messy set of analyses to undertake before one can draw any conclusions of that sort. Obviously the potential for confounding variables in this particular arena is substantial, to say the least.

With respect to degree of exposure, that is, the dose/response relationship, I gather there's some but rather limited evidence of that dose/response relationship. Interesting Whitehall information which shows that nice gradient, and I suppose that there are similar findings from other loci, but it's clearly a work in progress.

Then the notion of biological plausibility is obviously very much a work in process. There are likely interesting relationships to what we see in the hypothalamic-pituitary-adrenal response system; but, still, tantalizing hints in many ways as opposed to definitive evidence.

So I guess that my conclusion as a relatively uninitiated, uninformed individual would be that the answer to the question--Is inequality bad for your health?--and a pure guess is probably yes. But with very different potency for very different folks in very different circumstances. The question is how and for whom. I don't think we can rule this out by any means. But it's unlikely, given what we've seen now, to be universally applicable across the board, regardless of who we are and where we are.

With respect to the perspective of a health professional, wondering is the question relevant for me, health professionals are generally interested in two things; that is, knowledge and action--knowledge about what it is that determines health, and action related to the most effective pressure points to protect and promote health. So, clearly, this is at least of passing interest to me, if for no other reason than interest in having a deeper understanding about what it is that shapes the health of individuals and populations.

And there is no question that with respect--if you couch this in the broader issue at hand, that is, do social circumstances affect health prospects, it's quite likely that there are interesting and important elements that would be of interest and relevance to a health professional.

What are the social circumstances at work? Are there commonalities in the way they exert their influence? Are there ways in which the health sector can address those commonalities?

Social circumstances do matter to people's health. Education, housing, income, for example, all impact health status, directly and indirectly. In some cases, they act by putting people in harm's way, subjecting them to hazardous lifestyles or environment that carry threats to their safety. In other cases, they act through depriving people of the means to pay for medical care they need.

But apart from the issues of lack of medical care or physical risk, the stresses of social engagement, and social estrangement in particular, that is, the absence of contact with family or friends, of being cut off from supportive interactions, feeling that choices are limited or that no one cares, act independently to increase people's risk for illness and death. We have enough evidence to indicate that social estrangement in and of itself is an independent risk factor.

It's possible that income inequality could act to increase health risk in a fashion similar to that of other aspects of social vulnerability. The question of the commonalities is a particularly interesting one. Do poverty, income inequality, education, housing, isolation, and other life circumstances, in addition to their sort of clear and present danger elements, also act through some more subtle, longer-term common biological pathways? Are there commonalities in the way that locus of control, empowerment, self-efficacy work through these biological pathways--again, the hypothalamic-pituitary-adrenal response system--to decrease immune function and increase risk for various infectious and chronic diseases?

And if so, if those commonalities do exist, are there ways in which the health sector can work to better identify those who have particular vulnerabilities and engage them in a fashion that will enhance their longer-term health prospects?

Which brings me in some respects to the perspective of that of a funder interested in engaging health issues and in doing the business of philanthropy in a responsible manner. Essentially as a health philanthropy, we're interested in knowledge and action, the same thing that any other health professional is interested in. We're also interested as a philanthropy in our opportunities for gap-filling and taking risks. Those are not only our opportunities, they're our obligations. We need to fill gaps that others are not taking on, and we need to take risks.

In this respect, we are interested in the notion of social circumstances. There's no question about that. We want to test the ways in which we might engage the socially vulnerable, identify them and engage them in ways that will enhance their long-term health prospects. We've been involved in this arena at the foundation for some time. A good example, I think, is our work in the nurse home visiting project, which is now a two-decade-long project to identify low-income, pregnant teens, engage them with home visits--these are among the most socially estranged and isolated folks in our society--and stay with them for a two-and-a-half-year period.

And we're finding 15 years later that not only were we able to improve, we, society, able to improve the prospects for those babies and those mothers in the near term, but that 15 years down the line we're seeing reduced incidences of teen pregnancy, of substance abuse among the 15-year-olds, and of susceptibility to violence. So there may be some longer-term issues. We as philanthropists want to do what we can to explore the nature of those effects.

Secondly, we are interested in complexity. These are complex phenomena, and we have an obligation to understand better how complex phenomena play out in society. These various domains of influence for us all--and there are basically five domains which influence the health prospects of a population or individuals: genetic predispositions, social circumstances, behavioral choices, environmental exposures, and the medical care we receive. But these domains don't act in isolation. The real interesting issues are at the intersections of those domains. How do they interplay to affect health prospects? And that's one of the things that we're interested in building from the perspective of philanthropy, a deeper understanding of how various domains intersect to influence health prospects and a deeper understanding of how society can take advantage of those intersections in order to sponsor activities.

So I'm delighted to have the chance to hear about this very important set of issues and to engage to some extent in identifying hints of ways in which we might as philanthropy make a contribution.

MR. HELMS: Thank you, Michael.

We now turn to Sally Satel, who is a W.H. Brady Fellow at AEI and is a staff psychiatrist at the Oasis Clinic in Washington, D.C. Sally?

DR. SATEL: I don't have any pictures, so you'll have my words.

As Dr. Milyo points out, the evidence for a causal relationship between income inequality and health is questionable. But he's staked out his intellectual territory and provoked an academic debate, and that is just the way it should be. But it's an altogether different thing to proceed as though the income inequality hypothesis were established fact and to advance policy objectives--that is, social reform--on the basis of that. Yet that's what we're seeing.

A few months ago, a physician named Steven Bizrushka (ph), also on the faculty of the University of Washington School of Public Health, wrote one of those "My Turn" columns in Newsweek, one of those personal essays. And the doctor wrote this: "Research over the last decade has shown that the health of a group is not affected substantially by individual behavior such as diet, smoking, and exercise." He went on to say that relative wealth, by contrast, was a very important factor in health, and so that better prescriptions for a healthier society would be those that reinforced egalitarian principles, or at least, as he proposed, imposed consumption taxes, better transportation, or universal child care.

What Dr. B. was telling the readers of Newsweek is that in order for Americans to be physically healthier, some of the basic aspects of our economic system would need to be overhauled.

Putting aside now the question of whether or not Dr. B. was correct in his assumption about income dispersion and population health, another question come to mind that health professionals must consider, which is: What is at stake when they throw themselves into the business of prescribing far-reaching social change, whether those changes come from the right, the left, or libertarian corridors? What are the implications? What is at stake when they recommend these kind of changes in the name of health? That's what interests me today.

Pronouncements like Dr. B.'s don't regularly appear on newsstand publications like Newsweek, but they are commonplace, as you've heard, in the academy and in advocacy circles. Indeed, for about the past decade, public health experts have become increasingly eager to expand their professional agenda beyond health and into broader controversies.

But let me be specific. By "broader controversies," I am not talking about policy matters that have a direct impact on health or health care. Many doctors, of course, rightly so, are involved in the policy debates over Medicare, insurance systems, prescription drugs benefits, even tobacco taxes, and reasonably so. I personally was involved in reshaping disability legislation a few years ago. What I am saying is that when health experts put their energies into utopian visions like social justice, they risk blurring their professional focus, diluting their resources, and, most importantly, forfeiting opportunities for practical ways to help patients and populations in a reasonable period of time.

So I do see it differently from the former dean of the Harvard School of Public Health, Harvey Feinberg, who says, "A school of public health is like a school of justice"--meaning social justice.

Likewise, a slogan of the American Public Health Association meeting a few years ago, empowering the disadvantaged, social justice and public health. Even the modified Hippocratic oath sworn to at some medical school graduations enjoin the young doctors to say: I recognize that I have responsibilities to my community. That's okay. But then go on to say: to speak out against injustice.

Hippocrates did not write that in the original oath. I think as early as 500 years B.C., he no doubt realized that everybody must fight injustice, but doctors must fight disease.

At its annual meeting in two weeks, the American Public Health Association will discuss a revised code of ethics for its profession. According to its website, the code affirms the World Health Organization's rather utopian definition of health, to quote, "A state of complete physical, mental, and social well-being, not merely the absence of disease or infirmity."

The code also affirms that public health should "advocate for or work for the empowerment of disenfranchised community members."

Now, how can they do that? Well, one way in which public health researchers and advocates seek to empower is through research that we've heard about, research on what's called the social determinants of health, really more properly called the social correlates of health. And these are income, community organization, education, class. The work is very hard to do. There are, as he knows, so many variables, although it certainly can be done well, and as I said, it's perfectly--it's excellent material for academic debates.

But at the same time, the funders of this work--NIH, the CDC, and health philanthropies--must be aware that this kind of work is easily politicized, and dispassionate scholarship can take a back seat to the researcher's own agenda.

Let me give you one example here. The NIH awards grants to study the relationship between health and--these are the variables that are in their RFA: powerlessness, classicism, racism, and discrimination.

I guess I should withhold judgment, but I'd be very interested to see those research proposals, because it's hard for me to imagine how applicants can manage to operationalize these variables in ways that will produce valid studies.

Just last week, I saw a memo from the CDC. It has a prevention fellowship seminar series, and the memo organized--the series organizers wrote a memo saying: In view of the relationship between health and income, education, race, and social capital--I think that's a fair statement to make; there is some relationship--some relationships that are demonstrated. But then it went on to say: In light of these relationships, participants are invited to discuss how they can test interventions to change the relationships.

Changing the relationship between health and income seems to me a poor use of health
professionals' time and knowledge. Such a questionable task has also been incorporated into a new blueprint that I just saw last week, a blueprint for health in the State of Minnesota. It's entitled "A Call to Action: Advancing Health Through Social and Economic Change." It was issued by the Minnesota Health Improvement Partnership, and in its precis, it says, almost sheepishly, "The social and economic environment is a major determinant in population health, but it has not been a focus of most health improvement efforts in Minnesota."

If this report has any influence, that might change, because among the recommendations to the Minnesota Department of Health are to help people move out of poverty and to assure opportunities for quality education.

Now, don't get me wrong. Helping the impoverished and promotion quality education are vital social aims. But they're taken on by local governments, not in the name of health specifically, but in the name of public good.

In conclusion, none of this is to deny that social conditions, especially poverty, affect physical well-being and length of life. And public health practitioners do have a responsibility to design policies that reliably prevent disease, reduce contagion, and minimize injury. But they are sorely mistaken in thinking that they have special expertise in changing income distribution and defining social justice or in producing the instruments that can attain it.

To be sure, attempts to understand the ultimate non-medical sources of ill health have occupied scholars for decades. It's certainly a legitimate topic of inquiry. But there is a huge difference between explicating these factors and claiming scientific authority for political remedies. It risks taking physicians and epidemiologists away from their traditional missions or trivializing their responsibilities.

Misguided political activism can also be fairly demoralizing. Professor Ronald Baer (ph) at Columbia University School of Public Health has lamented in print that so many of his colleagues believe that "public health officials can do little or nothing to change the prevailing patterns of morbidity and mortality in the absence of social change." He calls this mentality "public health nihilism."

As Professor Milyo showed, there is much energetic advocacy surrounding a deeply uncertain claim about the connection between health and the degree of inequality. Even if the linkage between inequality and health were clearly established, I think health professionals have no particular expertise in reducing inequality and solving broader problems of social injustice. And expending efforts in that direction siphons off energies and resources from the vital issues that they have addressed so well in the past and are doing--still doing now, of course.

I believe I've given ample reason why health professionals should define their territory more narrowly in ways that help patients and populations directly. The specter of bioterrorism now I think is even more weight, as if it were needed, to the virtues of the traditional mission of disease and injury prevention, which is the classical mission of public health. Congress has just provided $1.6 billion, a five-fold increase, to fund public health defense programs. So now the profession has a front-line, high-profile role to play in protecting the public. I would think that would help with a bad case of professional nihilism.

Thank you very much.

MR. HELMS: Okay. Thank you, Sally.

Our next speaker is truly a person that needs no introduction. You see him on the news all the time, on TV and in the press. But we might ask the question: What is Newt Gingrich doing on a health policy discussion here at AEI?

Well, he is a senior fellow here at AEI, and it turns out that health care is one of his primary interests. And I would remind you that before he became a politician, he was a college history professor. So he does know something about the relationship between academic research and public policy. Newt?

MR. GINGRICH: I had asked earlier what the difference between "respondent" and "discussant" was. But then I listened to Jeff talk about the bringing in of upper-hierarchy primates and the taking out of upper-hierarchy primates.

[Laughter.]

MR. GINGRICH: Which is a topic I have some knowledge of. And I figured "discussant" was just a new primate title. By the way, any of you who want to pursue that, there is a fabulous book by Frans De Waal called "Chimpanzee Politics," which is a study of the chimpanzee colony at the Arnhem Zoo, and you will never quite see the Pentagon or the White House or the Congress or, for that matter, any academic institution in quite the same way once you understand the non-verbal communications of primates as they struggle for hierarchy.

You may wonder how I got to that point, but let me--

[Laughter.]

MR. GINGRICH: Let me tell you why I think viewing the system from without--that is, not so much getting involved in the specifics of inequality analysis, but starting with a line that Nick Eberstadt used, which is that what you have is a doctrine in search of data, which I think is absolutely true, and leads you to ask the question: Why is the doctrine so important that this many people are spending that much time trying to find the data?

And I would argue that there are two reasons. The first is that there's been a 180-year reaction to the rise of the modern system in which consistently, particularly in the European intellectual tradition, there has been a continuing effort to figure out what the reason was for redistribution, that is, the state should be more powerful than the free market, created wealth should be redistributed away from those who created it, and the argument--the reasons for that change with each passing decade, you know, whether it's an environmental reason for redistribution and state control, or it was 180 years ago a series of arguments that the whole rise of the modern socialist critique of the free market has a very long tradition. And all it does in change titles.

I recently was on a panel with Jim Watson, the co-discoverer of the double helix, who made the comment that when they first came out with recombinant DNA in the late '70s, there were all sorts of wildly Luddite arguments about the notion that they were going to create the Andromeda strain and a whole series of things he said that were scientifically clearly ludicrous. But there were a whole range of distinguished groups, including the Sierra Club, who promptly lined up on the other side. And it took three or four years of real fighting to finally break these people.

In fact, his comment was that he thought all his life as an Irish Catholic that he was a liberal Democrat, until he got in this particular fight with the control wing of the left. And his comment was, after basically Helmut Schmidt and Thatcher and Reagan had intellectually defeated the control wing of the left in the late '70s and early '80s, they went quiescent for 15 years, and he said, if you look now at the genetically modified food argument, it is literally the same people; that after 15 years of being quiet, they finally feel comfortable coming back out again, and they found their newest cause. So he said the science is exactly as bad today on this fight as it was in the late '70s, and all the evidence in the last 20 years is they were just wrong. But the wrongness was irrelevant because it's the doctrine of control and the doctrine of redistribution that matters, and the current argument over data is simply this decade's manifestation of what is now an almost 200-year-long fight.

But there's a second reason--so first I just want to argue there is an intellectual tradition that is understudies that is a rejection of the rise of the modern market system and a rejection of the modern technique of having people create wealth, and that cult, that historical tradition is now at least 180 years old and has deep roots and is essentially a psychological, cultural rejection of modernism rather than taking each decade's argument as though it was a free-standing case. And there's a fabulous intellectual history to be written about that.

But there's a second thing going on. I've asked Jeff's permission to use one of his--if I can figure out how to do this? Ah, this worked.

What I was struck by looking at this starts with one of my passions, which is diabetes. I got deeply involved in the diabetes issue when the Center for Disease Control came to see me when I was Speaker and said every seventh dollar in Medicare is a function of diabetes, and it's growing. And it's mostly controllable. The Center for Disease Control argues you could reduce 95 percent of the blindness and two-thirds of the heart disease, kidney disease, and amputation of limbs which are related to diabetes.

Now, those are fairly startling numbers. And so I got very interested, and, in fact, with Erskine Bowles, the President's chief of staff, we created both a program for Type I diabetes and a program for helping Native Americans, who were the group that have the largest amount of diabetes.

And the more I got interested, the more I noticed that the problems were non-rational. I want to draw a distinction. They weren't irrational. They were non-rational; that is, they were statements of realities which were defendable on non-rational grounds, which is essentially cultural.

Now, the reason I point this back up is two of the white states with the lowest age-adjusted mortality are the Dakotas. But if you were to break out the Indian reservations, you would have stunningly different numbers. And we're not allowed to talk about it. We're not allowed to talk about it because it is a cultural, political problem. It is not a problem of income. It is an objective fact that being trapped in a fossilized culture with no future is a debilitating experience which leads to alcoholism and suicide.

Now, that is a statistically verifiable fact which is totally outside the politically correct society, and so you can't say it.

Furthermore, it is a fact that the Indian Health Service is a bureaucratically incompetent, politically dominated, subservient system which desperately seeks to avoid the anger of the Tribal Councils, who are, in fact, consistently failing to adopt what any modern system would describe as the best health process for their people.

And so you can't get any kind of solution because any kind of solution threatens the political structure of the people who are on top of the disaster, and they value staying on top of the disaster more than solving the disaster.

Now, these are objectively true. You can go over and you study the history of the Indian Health Service, and you ask why they don't do other things. And then you go out and sit with five or ten Tribal Councils, and what you'll discover every time is that power is more important than health. And, therefore, you can't change it.

Furthermore, retaining the cultural identity, even at the cost of alcoholism and suicide, is a more acceptable future than undertaking a program which would threaten the cultural identity. And I'm not--this is what I said. This is a non-rational argument. It's not irrational. But you can't sit down and have a rational dialogue about it because the core values of the two dialogues are profoundly different. Public health is essentially a function of modernity. And you see this if you look at every great breakthrough starting with smallpox vaccinations, which were vehemently rejected by many people on the grounds that they were in an intrusion of the state and they were a dangerous, unnatural thing to do by definition. True, by the way. I've often thought of writing a history of smallpox as seen by smallpox as we wiped it out. I mean, for those of you who believe in preserving endangered species, it is fascinating--

[Laughter.]

MR. GINGRICH: Here is a virus which we consciously hunted down planet-wide. We sought to extinct it. And except for a handful of laboratories, we have.

So I wanted to put this up because the aggregate data masks the greatest single disgrace in the American health system, which is the Indian Health Service. And I say that--if you break out the data, there is no subgroup more disserved than Native Americans. And we cannot discuss what the problems are because they're culturally and politically not appropriate.

Now, let me extend that. It occurred to me looking at the data that Nick put up--and actually I want to ask one of our interns to work with you and see if we can put this together. As bad as the worst Latin American country is, my guess is in mortality terms it is better than the average African country. That's a very important thing to think about. This is my second totally politically incorrect point. Culture and politics matter. Politics matters because if you don't have a stable system, even a bad stable system--you could argue Guatemala at its worst may--Guatemala and Paraguay--I guess there are two or three contenders for this title. Pick whatever the worst stable system was in Latin America. It nonetheless sent signals that people could organize their lives around.

So you find, for example, in the last 60 or 70 years, at least since the great war between Paraguay, Brazil, and Argentina, you don't really find anything comparable to Rwanda or Sierra Leone or Somalia or Southern Sudan. I mean, the list goes on. Because the political structures have been coherent enough to minimize the level of damage--have still been horrible. I wouldn't argue in any defense of the Guatemalan colonels, to take the most recent example.

But as a level of stability, people could adapt to the horror, and the result is in pure mortality terms, far fewer people died, and for most people they could actually eke out a pretty decent living because most of the time stable structures don't bother most people if they don't get in its way. And people learn in dictatorships just to stay out of the way of the dictator, don't do anything to draw attention.

Now, I'm not saying those are good things, but I'm suggesting it's amazing, if you look at the hostility of the left, correctly, against the old dictatorships in Latin America, and the total refusal to discuss the political collapse of large parts of Sub-Saharan Africa.

Now, why is that important? It's important because you can't put up any mortality tables for Rwanda--I mean, Rwanda happened to show up, and if you looked at Rwandan mortality in the last 20 years and tried to explain it as a function of inequality, it would be a sign you were totally out of touch with data. It was a political, cultural civil war in which at least a half million people were killed, which has to have a fairly big effect on the mortality tables.

And I'm not saying this in any way to be humorous, but I'm only saying to say that we don't allow the data to lead us to places that frighten us so badly, so we block it off and say can we go back to something else to explain it. The problems in Africa are very straightforward, and they're twofold: you have in many places a tribal culture which organizes life in ways that makes it virtually impossible to have modernity; and if you don't have modernity, you don't have a real public health system. There's a World Bank estimate that 83 percent of the drugs that are given away in Africa are sold because the value of them is so massive that people--the corruption just takes them out of the public health system, and people go sell them. That's one of the things one couldn't say during the recent argument about AIDS treatments in Africa.

So first I would argue you've got to look at stability. How do you get stability? That is entirely a political issue, which none of the inequalitarians want to talk about because it's so politically incorrect.

The second is cultural. If you have the President of South Africa explaining that AIDS is not transmitted in any of the ways that we scientifically believe it's transmitted, how do you have an effective preventive health program when the President of the country is saying that is stupid, that's wrong, don't even pay attention to it?

And so, again, it's a politically incorrect thing because nobody--although he was taken on pretty intensely, even by the scientists of his own country, but unless you look at these cultural issues, you can't understand the underlying health implications.

Let me just give you two other examples, something that Nick Eberstadt's written about a lot. The cultural crisis of Russian psychology--and I say it in that term because the individual psychology in Russia is a manifestation of a national cultural crisis--has had an enormous impact on mortality, far beyond any plausible economic explanation. And it is a fact that when people despair and have no sense of a better future, the number of abortions goes up astronomically. It is a fact--and this directly parallels, by the way, my comment about Indian reservations. It is a fact that when Russian men see no future, the amount of alcoholism and, in parallel, the amount of suicide goes up astronomically.

And none of these have much to do with the kind of things the left most likes to look at, because it requires, again, that you think about so what's the politics and what's the culture. How do you change those factors? Which really are very big factors and how you get to it.

My last point will be very simple. Preventive health is almost entirely a function of two things. As was mentioned earlier, it's either a function of the great 19th and early 20th century model, which is we go out and we eliminate yellow fever, or we put concrete pipes so that human waste doesn't sit outside your door.

Now, these breakthroughs were infinitely--with due respect to the doctors who are here, these breakthroughs were infinitely more important than breakthroughs in personal medicine. The fact that we eliminated cholera for all practical purposes, we eliminated typhus and typhoid, we literally eliminated smallpox, we eliminated yellow fever, we have done fairly well in the wealthy world in eliminating malaria, and terribly in the poor world where it's still one of the biggest killers on the planet--which, again, I would argue is a function of organizational structure, much more than inequality.

So in that setting, there's a zone of public health which is heroic and massive and happens without regard to your personal organizational skills.

There is a second zone, which you see if you go back to the old Rockefeller efforts in the South to get rid of ringworm and other kinds of problems, and that is teaching people things like, you know, wash your hands after you go to the bathroom, which may seem common to most of you--I won't claim all of you, but maybe because we've had some--we've had some breakdown in this cultural transmission. But as a general rule, there are a whole set of common-sense things which really became middle-class behavior and which were then driven down to the poor by very large voluntary and public health efforts that were very explicitly educational and very directly in conflict with traditional patterns. That is, people traditionally--in fact, there's a good book called "In the Wake of the Plague," which points out that bathing died out in Britain during the bubonic plague because people thought being wet was a function of the plague. And for almost 250 years afterwards, people bathed less than they had prior to the plague. This is the 1349-1350 plague.

So, you know, when you look at those kind of patterns, people had to literally change their behavior, and that means, by the way, very often outside Yankees coming into the Deep South and teaching things which were profound personal changes in behavior.

Now, I just want to suggest to you, when you look at the greatest interest--I start from a simple premise. And what I'd say to the inegalitarians is simple. If we could find a way to have a profound revolution in public health so that no matter how poor you were, you had these two things happened, we were eliminating malaria, for example, which is a definable, doable project actually underway, although crippled by political problems, and we were actively teaching people the scientifically based knowledge of preventive care. If those two things happened, it is clear that mortality would go down, life span would go up, and that child survival would increase.

Neither of those is a function of inequality. Both are a function of social organization. Both have cultural and political implications, and both are currently too incorrect for the academic left to be very comfortable looking at them.

Thank you.

MR. HELMS: Okay. Thank you very much. That gives us a lot to talk about, and I do promise all the panel members that I will give you a chance to respond to these things.

But right now I want to go to the rest of the experts in the room, meaning those of you in the audience, and give you a chance to pose some questions. I'd ask you to keep your comments short and your questions to the point. And I'd also like to know for recording purposes who you are.

So anyone want to start? Right here, please? Please wait for the mike.

DR. GILFOYLE: I'm Dr. Jean Gilfoyle (ph), and I have a question for Dr. McGinnis.

Dr. McGinnis, you made a comment in your presentation on your interest in the complexities of these issues, and that also appeared to address social and cultural indications. The one complexity you referred to were pregenetic dispositions, and I'm particularly interested in that because I know the Robert Wood Johnson Foundation has been closely associated with funding of school-based clinics, which are frequently located in minority areas.

Could you expand upon your comment related to pregenetic dispositions? Thank you.

DR. McGINNIS: Let me just take on the issue of our interest in complexity. Essentially, just to underscore a couple of points I made, it's very clear that the way the health of populations evolves is a function of the way that these various domains of influence intersect, how genetic predispositions, for example, intersect with social circumstances or behavioral choices intersect with environmental factors. They are cross-influencing in the way they play out.

We have enough trouble understanding what's happening within domains. Clearly, there needs to be a greater effort to understand the dynamics within any given domains, but if we're really going to understand and affect what happens in society, we have to better characterize and better address those cross-domain influences.

So what does that mean for foundation work? It means that, first, we're interested in developing a cadre of individuals who are comfortable with dealing with these complex problems. So we're developing a health and society scholars program to parallel our clinical scholars program. We'll be mounting a program of research to explore what happens at the intersections of domains. And we don't expect any quick, easy answers, but we feel that we need to avoid being reductionist in the way we approach -- [tape ends].

MS. KRIMGOLD: I'm Barbara Krimgold [ph] with the Center for the Advancement of Health. And this touches on a number of comments the speakers have made. I think Bob Helms mentioned Angus Deaton's paper, as did Jeff Milyo, and also it relates to Professor Gingrich's comments.

I believe Angus Deaton's paper at least, the recent NBER paper, makes a case for if you plug in race, then the effect of income and equality is no longer visible, and I notice that that's been raised here, and it makes--I wonder what your reactions are to us and whether in the U.S. we collect data by race and that's a proxy for class or economic disadvantage, educational disadvantage or socioeconomic residential segregation, or for these different cultural values that were mentioned. And I wonder what that suggests about future research on improving health and reducing health disparities.

MR. HELMS: Okay. I think I'd like to ask Jeff to respond to that, please. Please put the mike as close as you can, please.

MR. MILYO: That's a correct characterization of some of Deaton's work, which is basically he argues that the reason we find a correlation between equality and health outcomes across cities is because then equality is closely correlated with the racial composition of cities. And he argues that it's racial composition, which is a much more convincing correlate, and he doesn't have a causal explanation of why that might be. It's just left hanging there.

There's an enormous literature on the association between race and health, as there is on the association between income and health, which Gregg Bloche talked about a bit, but I would say, despite the reams of paper that have been written, we haven't yet got to the point where we're able to sort of parse out how much of that is attributable to discrimination, how much is attributable to choices that are related to a person's income or education, which are going to be correlated with race, how much are related to choices that are more, say, related to a subculture, how much is related to the fact that since people live in--tend to live in racially homogeneous communities, that they're going to the same kinds of providers, and that there might be differences in the quality of providers across communities, and that's something that we're just beginning to look at.

So it's left hanging out there really, and don't have an answer yet.

MR. HELMS: Two other panelists want to respond. Nick?

MR. EBERSTADT: Just a quick comment. If you take a look at the Census Bureau's data on income distribution, you see that the Gini Coefficient for African-Americans is hugely higher than for ethnic Anglos. And I don't know what contribution that would be, but that's something which has to be recognized there.

MR. HELMS: Newt?

MR. GINGRICH: I was just going to say again, I think, to some extent at least you've got to come back to cultural patterns and the degree to which culture becomes synonymous with race when you're analyzing certain kinds of problems.

MR. HELMS: Yes?

MS. DOUGHERTY: Bernie Dougherty from the Kato Institute. Two things. I guess I would totally agree with you. And I'm a little disturbed when I hear people talking about the impact of inequality on health outcomes, because it seems to imply--the normative that it seems to imply is that the richer some people get, the more they hurt the lower--the poorer people, which--and I really don't see how that is connected, so that disturbs me.

And I've looked at a lot of studies on income and equality, and it seems that those studies show that really matters is when at one point you are in the lower level of income, whether at one point of your life you're going to then end up in the higher level of income. And a lot of studies show that actually in the U.S. that's what happens, that you don't stay in the lower level of income and your chances of getting in the higher rank are very important.

I haven't looked at any studies, you know, in health level, and I was wondering whether you had information of that, like if at one point of your life you're really at the lower level of health outcome, whether there are studies that show that then you move up?

MR. HELMS: Jeff?

MR. MILYO: I don't know of any studies that look at sort of a change in the individual's health status over time given some treatment effect change in inequality of community or something like that.

MR. HELMS: Okay. [Inaudible].

MR. BLOCHE: I want to address the first point that you made. I think it's really important to see that this is not a story about the rich hurting the poor, and I do want to reply to what--I guess I can call him Newt even though I don't know him because he's Newt--what he said about data and how data leads us or it doesn't. I think you're absolutely right. But the problem here I think is illustrated by the way the subject has been framed for this event.

The inequality hypothesis, that's understood solely as the gradient, per se, ignoring the reality that income, wealth, education, and yes, cultural and behavioral factors have a huge impact on it all. Then it's easy to pick off, and what I think Jeff has done is picked off a weak target and perhaps killed the rest by connotation.

The hazard here is that the press, the public, the folks in politics might be invited by this critique to pay less attention to the role of income and wealth and education, and, yes, the left is responsible for our insufficient attention to cultural factors, but just as that's the case, I think, the right can--ought to take some responsibility for an ideologically driven, reluctant to look the issues of income and wealth straight in the eye. And it's not about criticizing the rich for doing something wrong. What this is about is looking at the data that supports the conclusion that income matters, and then trying to figure out the pathways, figure out the complexity, doing the research that Michael McGinnis has talked about funding.

MR. HELMS: Sally.

DR. SATEL: You know, this is a very thoughtful panel, so I don't think people have made the kind of radical statements that you're referring to, but I have seen what you're referring to in print in some certain books on public health, and an interesting manifestation--I mean what you're basically saying is somehow the haves are making the have-nots sick. An interesting manifestation of that I've seen is response to the Whitehall study. You mentioned it, Jeff. That's the famous civil servant study demonstrating a hierarchy in health between the top, and--there is five social classes. And, of course, what was observed is that there was a gradient in I believe mortality and cardiovascular health, but the researchers were surprised to see that the second to top ranking still had three times the incidence of morbidity and mortality. And that surprised them. They thought it would be a much shallower difference between the two top.

So their interpretation, as everyone knows by now, is that there may be something about hierarchy that is bad for one's health. Possibly it relates to the learned helplessness model we know about in animals, the idea that if you have responsibility but no control, that is a very stress-producing situation and it could lead to cardiovascular disease.

Okay, that said, what a number of folks I've seen on I guess you'd call the public health left, the way they've interpreted this is to use it almost as ammunition against the notion of hierarchy, and to argue then for, again, egalitarian principles, and what my only agenda, it's not from the right or left, just to say, well, there's another way to think about it. What about promoting the kinds of mechanisms that would enhance economic mobility? I mean that might be another way to undo the oppressive effects of feeling stuck, of feeling as though you have no control, but just responsibility. So there's that bias that you often do see in the literature.

MR. : Can I comment?

MR. HELMS: Yes.

MR. GINGRICH: Let me start by going back to your initial question, because you're exactly on target. The original critique of the modern wealth-creating market system started with what was a pre-modern notion, and that is literally, when you had the closing of the commons, for example, the rich were taking from the poor. I mean if you have a zero sum game and somebody gets richer, somebody gets poorer. The creation, starting in the early 18th century, of a dramatic increase in wealth, unlocked that model but didn't intellectually affect its critics, who all operate as though there's a zero sum game. So by definition, the rich must be hurting the poor. It's very deep cultural language in the critique of the left.

The second, in the European model, there are classes. There are still classes today. When you talk to people about the problems in biotech and the problems in computing, they'll tell you that there are deep class structures built into the European psyche and it really does matter who you are in ways that then have I think effect.

Lastly, I agree with--if I can call you Gregg--with what Gregg said. I think the great challenge to American conservatism is to say, okay, we want to raise everybody to some minimum. Now, how do we creatively do that within our values as opposed to trying to defend a system that's not defensible? The challenge is, when you start to do that, you both take on politically-entrenched interests who don't want the change because it will change or threaten or weaken their power, and you inevitably get into cultural change in ways that are very powerful. I don't know the exact data, but my hunch is that well-educated African-Americans who are upper middle class are closer to whites who are upper middle class in their standard of living than they are to very poor African-Americans, and that very poor whites are closer to African-Americans norm in terms of their health outcomes than they are to very well-off upper income whites.

DR. SATEL: Look at the Appalachian--

MR. GINGRICH: Which would argue that it is a cultural problem much more--now, there are specific problems such as sickle cell anemia, but at its core these are culturally-driven patterns much more than they are, I think, ethnically driven.

MR. HELMS: I'd like to go to the audience. Any more? Well, Barbara--

MS. : I did say something already, but I did have a response to that.

MR. HELMS: Very quickly.

MS. : That the data from David Williams and others who've looked at the intertwining of race and income and economics data shows that there are two ladders. There's a ladder for white Americans, and the ladder for African-Americans, the health is worse at every income level. I mean it's also a gradient effect, but even for upper income African-Americans their health is less good than it ought to be, say, given their income and wealth.

MR. HELMS: Okay. Right behind you here.

MR. GRANN: George Grann [ph], Johns Hopkins. On the issue of diabetes, it's almost directly the consequence of obesity, and obesity is a characteristic of poor whites, poor blacks, the American Indian, and have very little to do with income. It has much more to do with social class, and as one of my teachers referred to the welfare system as "the black man's Indian reservation" that destroys people's initiative, and makes them accept their state without any hope for the future or for their children. Their children get just as obese as they do, and develop the complications of diabetes and heart disease.

MR. HELMS: Any response to that?

MR. GINGRICH: I think Gregg wanted to say something.

MR. BLOCHE: I do want to briefly respond to Barbara actually. Barbara, you're absolutely right, but David Williams drives home the point though that class, the class ladder in the aggregate is much, much more important than the race ladder, that, yes, they're different and they're in parallel, but in the projects at the Institute of Medicine and elsewhere that focus on race and don't talk about class, David points to the data showing that class matters a whole lot more. Whether that's about culture or what the underlying determinants are, we don't know.

MR. HELMS: Okay, Jeff?

MR. MILYO: Before I jump out of my chair here, there's so much I want to talk about, but that would be selfish, so--

MR. HELMS: Well, actually, you know, I was going to give you the last word, so if this is where you want to do the wrap-up, that's fine.

MR. MILYO: Just some responses really, and this colloquy between Gregg and Barbara is one place to start, which is people use the words like class is more important than race, but when you look at those studies, they're just looking at associations. They're not being very sophisticated about causality. And I hate to be sort of an academic imperialist here, but economists are very good among social scientists about thinking about causality, and in their training to try to get at causality.

There's another important difference between economists and other social scientists and public health researchers. It's no secret. Economists are much more conservative, but I don't think that's the result of bias. If you look at opinion polls, economists, their political breakdown is very similar to the overall population. When you look at other academic disciplines, they are wildly to the left. So I think there is a reason why, in this literature of social determinants in health, if you read the economics literature people are much more cautious, there's much more attention paid to really trying to find out about causality, higher standards of evidence, and I think that's a result of those two factors. And, you know, it's kind of implicit in some of this discussion, but I think it needs to be made explicit.

And then I just want to echo on some of these remarks and Newt's comments about modernity. You know, the current issues of Social Science and Medicine, there's a raging debate on the inequality hypothesis. However, the terms of the debate are the following: they take as given that inequality is bad for one's health, and the terms of the debate are whether the original studies sufficiently impugned free trade globalization and capitalism. That's the nature of the debate, so it really is, I think, a manifestation of this political bias in academia. It certainly does exist.

Secondly, Gregg Bloche talking about income and health and that association. I don't want you to leave here thinking that we know that there's a causal relationship between income and health. We don't. There's a very strong association, and 99.99 percent of the studies done on income and health, socioeconomic status and health, race and health, are just documenting associations, and they're not seriously trying to get at the causal pathways, although people make statements about them.

Just taking the example of income and health, I can think of five reasons off the top of my head why income might be associated with health. First, to paraphrase I think Bernadette Peters from "The Jerk", it's not the money, it's the stuff. When you're wealthier you might have more and better food, shelter, you may have health insurance, et cetera, and it's those things that help make you healthier. So if we provided people those things, that might make them healthier. If you provide them money, they might spend it on those things, they might not.

Secondly, it's not the money, it's the ability to earn money, things like perseverance and prudence and farsightedness help you earn money, they help you hold a job, they help you engage in behaviors that are also good for your health.

Third, having a higher lifetime income might change the way you've behaved. So take the study that was mentioned by Anne Case when South African blacks had the opportunity to have pensions in their old age, that may change the way they behave, because they're going to have some income in their old age, so you're going to make different tradeoffs in terms of risky behaviors.

Fourth, income might be a proxy for relative status, and sort of address that.

And finally, a pretty obvious one, is that your health, whether it's manifested in an actual diagnosis or not, is going to affect your ability to earn income, and there are a number of people really trying to get at these different factors, and you shouldn't leave here thinking that there's a clear relationship in the literature that says if you give money to the poor, that their health will improve. We don't know necessarily that treatment effect.

There are intriguing studies though. If you look at the health of seniors, that's improved dramatically over time, and I think there's good reason to think that has to do with programs that address those individuals, but so have behaviors changed dramatically over time. And so there's a lot more need for caution than we've heard and certainly than we read in the literature.

Secondly, Gregg Bloche has kind of dismissed the inequality hypothesis as this perverse side stream, and I'll agree with both of those things, but if you think there's a causal relationship between income and health, that implies that perhaps redistribution will improve the health of the poor, but economic growth as well might improve the health of the poor. If instead you think that in developed countries we've done all we can in terms of the traditional public health improvements, good sewage systems and people have good hygiene, et cetera, and now it's these psycho-social stresses that are the most important determinants of health. Giving people better food, shelter, giving them more money, has no effect on their health or very little effect. Instead, what you've got to do is reduce the disparities in classes, and you can do that by taking money from the rich and throwing it into the ocean. That is literally the implication and that's the argument.

So there are very different policy implications. And researchers in this area are arguing very forcefully, and some famous philosophers as well, writing pop pieces on how Rawlsian social justice is good for our health, and it's definitely something a lot of people are taking seriously.

That brings me to Michael McGinnis. And I guess I have to say, hearing you go down this list of how you evaluate evidence, I was quite surprised at your conclusion. And so I've thought about this a little bit, because again, there is this divergence in what people infer from studies, and you see it across disciplines, across--well, let's leave it as across disciplines.

And let me give you one more example of that. Let's take the Whitehall Study. There are a couple of things you can infer from the Whitehall Study, where you see lower-ranking civil servants have worse outcomes. You might say being in the low ranks of social hierarchy is bad for your health. You might say being in the high ranks of social hierarchy is good for your health. You could say that social hierarchy is good for your health, but being in the high ranks is particularly good for your health. There are a number of inferences, but only one is made.

 So I think this gets to, what do you take as given? And I always get a big laugh when I make academic presentations. It's a great throwaway line to say that I'm objectively pursuing the truth, and everybody laughs, because we know we're looking for attention and funding and publications.

[Laughter.]

MR. MILYO: And objectively pursuing the truth is sometimes consistent with those goals and sometimes not. And as a true social scientist, what I'd like to do is have a null hypothesis. So I approach the data with the idea that there is no relationship between social status and health, and then look to disprove that. If you do that, with all due respect, there is no way you can be convinced that there is a causal relationship.

If on the other hand, you approach the hand already believing there is a causal relationship between social status and health, and instead all you're looking for is some excuse that despite the evidence, there might be a way to explain how that relationship still exists, then certainly you can do that.

And so I think as a scientist, we want to take a null hypothesis. Perhaps as a policy maker we want to be very, very risk adverse and sort of say, well, if there's any chance social status is related to health, then we want to act on it. I don't know.

A final difference, I think, in the way people approach this is risk factors versus causality. The public health literature, it's quite common to just look at by variate correlations and talk about risks. And absolutely I agree, if you present me with two people and ask me to bet on which one has the worse health or which one's going to die first, and all you tell me is which one lives in an area with high income inequality, I'll bet that person. It's a risk factor. But that's very different from saying that it's a true causal relationship. And that's what I think as scientists we need to be interested in and as policy makers we need to be interested in, because that's how you change people's health, by changing the true causal factors.

Sorry that was so long.

MR. HELMS: Okay. But that's a great statement I think to end on. Let me say that I forgot at the beginning to tell you that there were some extra copies of some additional articles by Jeffrey and Jennifer Milyo that are available out at the registration desk.

With that, I'd like to thank all of our panel members and our speaker for I think a very interesting discussion, and for you, for your participation. Thank you.