The article by Daniels, Kennedy, and Kawachi promises far more, I am afraid, than it delivers. This is all the more disappointing since I applaud the effort to join empirical analysis of health and health care to normative disputes about what public policies should be enacted and realistic discussions of what policies can be implemented. My reservations fall into two categories. First, I am concerned about the authors’ basic claim that inequality per se is “bad for our health.” Second, I am not convinced that, even assuming income inequalities cause significant health inequalities, justice requires more effort to reduce income differences than to make access to medical care more equal. In that connection, I raise questions about the general presumption that discovering causal pathways in social arrangements leads directly to what the authors call “policy implications.”
Before proceeding, however, two prefatory remarks. First, I want to declare an interest. As the co-editor of Why Some People are Healthy and Others Not: The Determinants of Population Health, I found the public health sections of this paper annoyingly self-referential: the main empirical research they cite is their own. Although I am not an epidemiologist, I spent five years meeting regularly with a number of Canadian social scientists who wrote the basic chapters of that book. Nothing in this article makes a large advance over the understanding that book communicated nearly six years ago. What’s more, there is considerable evidence that Kennedy and Kawachi did not carefully consider its findings. More generally, the authors’ whole discussion seems curiously disconnected from the extensive research that precedes their work. The Lalonde Report of 1975, the most widely distributed public document in Canadian history, argued that, at the levels of income in societies like Canada’s, conventional investments in medical care were unlikely to produce big improvements in population health. Not a word about that, let alone of the efforts of scholars like James House of the University of Michigan to chart the causal pathways by which “social determinants” work their health effects, appears in their essay. And there is only a throwaway line about what may be the most illuminating research on these questions–the Whitehall study that Michael Marmot has pioneered. In that study, the gradient in health outcomes, given equal exposure to risk, appeared to be connected to the amount of control over lives that differently situated workers have. Daniels, Kennedy, and Kawachi say little about the physical pathways that might account for these differences–even though Marmot, and others, have explored them.
Second, space considerations prevent my taking up the theoretical issues raised by the authors’ treatment of how justice claims and the social determinants of health are linked. This is a serious topic, done with considerable care in the article, and worthy of separate analysis. Instead, I want to focus on a couple of key points.
How Bad Is Inequality?
Have the authors made a compelling case for the claim that “justice is good for our health”? I think not, but one needs to begin by asking what precisely the authors are asserting. The most cautious presumption is that inequality appears to affect the health of populations. That, of course, is consistent with the correlations the authors present. But there is little in the way of a rigorous defense of this claim. One difficulty is that the precise meaning they attach to “inequality” is unclear: sometimes it seems to mean the Gini coefficient (a standard measure of income dispersion), sometimes (as with the section on the Rawlsian conception of justice) it seems to be linked to the distribution of stress and control at work.
But the most important problems, I suspect, are the technical arguments against using correlations–whether at the national, state, or local level–to support causal claims. Scholars such as Harold Pollack of Michigan, Jeff Milyo of Tufts, and Ingrid Ellen of New York University all contest the idea that, controlling for income, inequality itself necessarily matters. As Pollack put it recently at an academic meeting, “cross sectional regressions that use inequality measures such as Gini are virtually uninterpretable.” He goes on to say that “it is frustrating that uncritical use of these measures is so pervasive in public health analyses of [United States] and cross-national comparisons.” Pollack’s grounds, which I find plausible, are straightforward: “Money matters near the bottom of the distribution and may not matter at all for many outcomes when one exceeds the median. Controlling for the median income, then, any income dispersion measure is highly correlated with the percentage of the population under the poverty line.”1 So it is not dispersion itself that matters for health, but the proportion of the population that is poor.
The connection between inequality and health, then, is far from obvious to other analysts and this article does little to dispel the skepticism Pollack’s remarks exhibit. That does not mean there are no connections between inequality and health. Rather, the connections are less obvious than the article suggests, and their implications, for what counts as a just society or what policies such societies should pursue, are less compelling than presented.
I turn now to the topic with which I am most familiar: drawing policy implications from admittedly controversial empirical findings. Assume for the moment that the article’s central contention about inequality and health is correct. Assume further that we know how to reduce income inequalities, that the technology of redistribution is available and implementable. Does that mean we should turn to inequality-reducing policies as a matter of health policy, subordinating the claims on resources made by modern medicine?
The authors’ discussion of this issue is more nuanced than their treatment of the social determinants of health. Daniels, Kennedy, and Kawachi concede that identifiable illness commands our attention and utilitarian concerns about “net benefits” need not always trump our humane allocation of care for the ill. But two larger issues of public policy analysis remain. First, the discussion of the purposes of health care policy is terribly truncated. National health insurance finds justification not simply in efforts to improve on measures of a population’s health. We care about equality of access to medical care because suffering, pain, uncertainty, and the myriad other features of being ill or injured ought not, it is widely believed, vary primarily with one’s ability to pay. The fact that even in systems of universal health insurance there are pockets of unequal response to illness does not dispel the egalitarianaspiration–or the social cohesion and sense of fairness–that such efforts both reflect and symbolize.
Secondly, I am concerned that the authors understate the gains that fairness in allocating medical care has proven capable of generating, and that they overstate the likelihood that we can do very much about more basic determinants of social equality.
Put more generally, there is no reason to treat a theoretical possibility as a compelling policy option unless both the worth of that aim and its implementability dominate the alternatives. Nothing in contemporary America suggests that we are likely to move more quickly in reducing income inequality than we are to make health care more fairly available. The expected value of a policy option is, in short, its idealized results times the likelihood of achieving them. A one-in-ten chance of getting a dollar is a lot less valuable, looked at this way, than a one-in-two chance of getting fifty cents; indeed the latter is two-and-a-half times better.
It is worth remembering that when national health insurance was a more prominent option in the 1970s, one of the arguments against it was that medical care outlays were wasted, that more powerful tools for improving America’s health were available. A quarter of a century later, Americans enjoy longer lifespans, but with a distribution of health care that is shameful. Those nations with a reasonably fair distribution of income protection against the costs of illness might well gain by concentrating at the margin on health-improving policies outside of medical care. The United States, I contend, should address the do-able but difficult task of making medical care more fairly distributed before taking on the more utopian task that Daniels and colleagues suggest.
1 Delivered at the October 1999 meetings of the American Association of Programs in Policy Analysis and Management, in Washington, D.C.