The Killing Fields

Ezra Klein argues that Joe Lieberman is willing to "cause the deaths of hundreds of thousands of people" for no better reason than to settle some old political scores.

This has put Washington Post editor Charles Lane in quite a tizzy. According to Lane, Klein is "essentially accusing Lieberman of mass murder," and Klein's "venemous post" is "beyond the pale."

The underlying issue is Lieberman's sudden discovery that expanding Medicare to allow people over 50 or 55 to buy into the program would require such a sacrifice in the way of Freedom(tm) that including such a proposal in any health care reform bill before the Senate would cause him to filibuster it.

Klein points out that a recent Institute of Medicine study suggests that lack of health insurance is causing about 20,000 excess deaths per year in the United States. This in turn suggests that successfully blocking health care reform would cause a lot of people to die who otherwise wouldn't. Another way of putting this point is to say that blocking reform will kill people. Yet another way of putting it is to say that someone who has the power to block reform and chooses to exercise that power is killing people. (Or, in Lane's intentionally hysterical formulation, "murdering" them).

As Yglesias points out, Lane's sudden squeamishness about these sorts of rhetorical tactics might seem odd, given that he seems to have no problem with, for example, Charles Krauthammer and George Will publishing highly misleading columns on crucial public policy issues.

Nevertheless as Socrates or Miss Manners or possibly both pointed out, two wrongs don't make a right, so the fact that Krauthammer likes to claim that Iranian agents are even at this moment contaminating our precious bodily fluids doesn't mean it's OK to publish misleading claims to advance one's policy goals, no matter how noble. And the study Klein cites is a pretty weak one. It uses a very crude methodology -- in essence it assumes that the difference in relative risk for mortality between insured and uninsured adults is accounted for completely by this single factor. That's surely not the case, any more than, for example, the difference in relative mortality risk between high school dropouts and college graduates is wholly a product of different levels of education.

Update: As a commentator points out this is actually incorrect. I relied on the Urban Institute paper's summary of the IOM's study's methodology instead of looking at the study itself, which was obviously a mistake. The study actually relies on a 1993 study that used a 1.25 hazard ratio associated with uninsurance after adjusting for multiple confounders. This earlier study, however, finds a relative risk running from 1.00 to 1.50 when employing a 95% confidence interval, which means that, using that confidence interval, the number of annual excess deaths associated with uninsurance ran from about 40,000 on the high end to zero on the low end.

To be fair, the study's authors admit their estimate is a rough one, and that even if their method overstates the effects of lack of insurance on mortality by 50% that's still a lot of dead people. (They could have made an even stronger argument by at least mentioning that the effects of under-insurance, given its prevalence, might be even more significant than those of uninsurance).

Furthermore their own data indicates that nearly half the excess deaths associated with uninsurance are taking place in one ten-year cohort: people aged 55-64. This, of course, is precisely the group that would benefit from the modest reform Lieberman now suddenly opposes.

Update: A further point that ought to be considered is that the effects of uninsurance on mortality don't nececssarily show up when a person is uninsured. It's quite plausible, for example, that lots of people covered by Medicare at the time of death died earlier than they otherwise would have because of the effects of decades of previous uninsurance.

The more general issue implicated by all this is to what extent it's OK to accuse your political opponents of killing people when they advocate policies that produce excess deaths in comparison to the policies you prefer. As a pragmatic matter, the answer of course is "it's OK to the extent it advances your goals." As a matter of principle, the answer, I think, ought to turn to a significant extent on the degree to which the policies you're opposing are actually intended to kill people as a first-order effect. Thus I see no possible objection even in principle to pointing out that Joe Lieberman (and other supporters of the Iraq war) wanted to kill a lot of Iraqis because he (and they) thought killing lots of Iraqis by invading the country would on balance generate good results. To be in favor of a war of choice, after all, is to be in favor of killing people who would otherwise not die so soon, because you believe killing them is necessary to achieve some worthy goal. That's what it means to advocate invading another country, although you would never guess that from listening to high-flown speeches on the matter.

Things get more complicated when the deaths caused by your policy preferences are second-order effects of those policies. Saying that you're in favor of invading Iraq even though this means you will be killing a lot of people in the process isn't the same thing as saying you're in favor of, for example, not criminalizing cigarettte smoking, even though not criminalizing cigarette smoking probably results in a large number of otherwise preventable deaths.

Lieberman's position on health care is more like the latter than the former -- which isn't to say that I have any objection to using the kind of language Klein uses to condemn him. As Yglesias says, stark moralizing language works. And in politics what works must, to a point, be given preference over more complete and accurate descriptions of reality. What that point might be is needless to say often a difficult question. In this case it isn't.