Co-blogger Arnold Kling posted last week on health economist John Nyman’s 2004 article, “Is ‘Moral Hazard’ Inefficient? The Policy Implications of a New Theory,” Health Affairs, 23, No. 5: 194-199. I won’t repeat Nyman’s argument, which is brief. Instead, I want to point out a non sequitur.
Nyman argues that many health economists have overstated moral hazard in health insurance. Whether that’s right is an empirical issue, but let’s say he is right. What follows from that? In a section titled “Implications for Policy,” Nyman writes:
The new theory suggests that health insurance generally makes the consumer better off. Therefore, the subsidies that encourage consumers to purchase insurance voluntarily, or a national health insurance program for the entire U.S. population, would improve society’s welfare.
But from the fact that health insurance makes the consumer better off, it doesn’t follow that government should subsidize its purchase. To advocate subsidies on standard efficiency grounds, Nyman must argue that there’s a positive externality. In his next sentence, he suggests that there is, writing, “This is especially true because we as a society are altruistic and benefit when all who become ill have access to modern medical care.”
I think that reasoning is weak, but let’s say you disagree and think it’s strong. The point is that he shifted arguments in mid-stream. He went from arguing that there’s less moral hazard in health insurance than we thought to a conclusion that the government should subsidize health insurance, a non sequitur. Then to support his case, he brought in positive externalities in health care consumption. These may or may not exist, but they have nothing to do with his original argument. Absent externalities, his argument that we have overstated moral hazard simply does not justify subsidizing health insurance. His ultimate argument for subsidies that could make sense is based solely on the supposed positive externalities in health care, not on the “new theory” that he exposits in his article.
READER COMMENTS
jsalvatier
May 9 2011 at 8:24pm
I normally don’t really like your writing, but in the interest of encouraging what I *do* like: this is a great point.
Shangwen
May 10 2011 at 10:38am
I’m not going to tussle with the econometrics in the debate, but as a senior administrator in a health care system I think the policy and practical implications of what he suggests are nonsense. I work in Canada, which of course has a single-payer system for the 60% of health care that is government-purchased.
Neither I nor any of my colleagues would agree that “free” or richly insured care predominantly leads to optimum welfare gains. In his paper Nyman presents a caricatured version of moral hazard ambiguity, contrasting cosmetic surgery with life-saving liver transplant. Clearly those are different things and the former leads to more hazard if insured. However most insured procedures are in a vast grey zone in between the life-saving and the medically trivial. The clinical boundary at which maximum treatment gains are reached is rarely clear, even to experienced clinicians, and in my experience they are very reluctant to wind up care if the patient wants to continue. Sometimes the motivation is altruistic, but it can also be financial, or because the non-sick are easier to treat than the sick. One of the worst examples of this is psychotherapy, where 100% coverage and no rationing can result in many easy, non-sick individuals crowding out many except the suicidally ill.
PLW
May 10 2011 at 10:43am
Perhaps I’m reading Nyman incorrectly, but I believe his argument is not that there is less moral hazard than we thought. Instead, he’s arguing that moral hazard is a good thing — that is, it’s welfare increasing, not welfare decreasing.
From Nyman’s 2003 book on “The Theory of Demand for Health Insurance”:
“…the clinical literature paints a far different picture of moral hazard than economists do. In contrast to the image that economists evoke of additional care that is frivolous and discretionary (for example, designer sunglasses and cosmetic surgery), the image of moral hazard from the clinical literature is one of reasonable, effective, and often lifesaving care for people with common illnesses, care that appears to be well worth the resources spent in producing it.” (p. 22)
This leads Nyman to the conclusion that government should subsidize insurance. This raises another important question in how the DWL of taxation to pay for the subsidy compares to the Nyman’s welfare gain.
mark
May 10 2011 at 10:51am
Great dissection.
Also, “we as a society are altruistic” is absurdly overgeneralized ( “we”? “as a society”? what do these words mean and how does that meaning translate into subsidized health insurance?)and flies in the face of much of the past 50 years of evolutionary biology theory, such as kin selection, etc.
Also thanks to commenter Shangwen for an insightful comment.
David R. Henderson
May 10 2011 at 1:04pm
@jsalvatier,
Thanks, I think. 🙂
@Shangwen,
Nicely done. I checked out some of your posts and am impressed. If you care to reveal, what city in Canada are you in?
@PLW,
You stated his point more accurately than I did. His point, though, even if accepted, does not imply subsidies to health insurance as an efficiency-enhancing measure.
@mark,
Thanks.
Shangwen
May 10 2011 at 2:40pm
David, thanks for the kind words (and the plug). I keep my blog anonymous so that I can be frank about things, being a public servant.
PS, I do have a new post up….
Comments are closed.