I am here in Toulouse, in the south of France (not far from Spain), at a conference on “Risk and Choice.” The conference is in honor of Louis Eeckhoudt, my good friend and beloved coauthor, who is also a good friend and beloved coauthor of many, many people who do research on the economics of risk and the economics of insurance. Louis is here with us, enjoying the conference.
The most provocative presentation so far has been one by Isaac Ehrlich, on “The Problem of the Uninsured—Implications for Health Insurance.” (His picture is below, along with a link to his website.) Isaac is legendary to me for an interesting reason. In “A More Personal Bio: My Early Tweets” at the sidebar, I tell about competing in high school forensics. I went to the National Forensics tournament in Seattle in 1977 as a state champion in Extemporaneous speaking. Unlike Larry Summers, who was a college champion in debate, I didn’t do as well at debate. But I did compete seriously at debate for a while. At debate camp in the Summer of 1976 at Georgetown University, I ran a case arguing for the abolition of the death penalty along with my partner Tom S. Hales (not the now famous mathematician, Thomas C. Hales, who was one of my closest friends and classmates, but another classmate with almost the same name). The biggest obstacle we faced in our quest for victory were quotations from Isaac Ehrlich’s paper finding an apparent deterrent effect of the death penalty.
At the “Risk and Choice” conference in honor of Louis Eeckhoudt, Isaac has presented “The Problem of the Uninsured: Implications for Health Insurance,” which is coauthored with Yong Yin. The point of this paper, indicated in the title of my post, reminds me of another paper quoted a lot by high school debaters in 1976-1977: Sam Peltzman’s argument that requiring people to wear seat belts would make them feel safer, and therefore lead them to drive faster. This is now called the Peltzman effect, and I will refer you to Greg Mankiw’s post on "The Peltzman Effect" for a discussion of it.
In his presentation, Isaac points out that there will be the same kind of effect if you require people to have health insurance. They will feel safer in relation to their health, and so will take more risks. For example, a young person who has to pay full price for the visit to the doctor for antiobiotics to treat an STD may be more careful to use or insist on a condom. He and his coauthor Yong Yin have a theoretical model showing that such effects can be substantial in size.
Notice that it should be possible to get good empirical evidence on such effects. This is a great dissertation project for a graduate student in labor economics, public finance or health economics who wants to do empirical work in the “natural experiment” tradition. All it will take is finding a time when the government changed health insurance coverage in a reasonably exogenous way, plus data on behaviors that might plausibly be changed by feeling more safe given health insurance. If you know of anyone has done this already, please write in a comment.
When we talked during the break, Isaac pointed out to me that people who have looked have often found very little relationship between having health insurance and good health in places where they expected to see a relationship. He argues that this suggests some negative effect of health insurance on health like the “take more risks” effect he was talking about to cancel out the simple direct positive effect of health insurance on health that one would think must be part of what is going on. (By the way, I am reporting everything Isaac said to me with his permission.)
That leads to my last thought. Suppose that Isaac is right, and health insurance causes some category of people to take enough more risks that their health is about the same as it was before, overall. In theory, there would still be a redistribution argument for subsidizing health insurance for poor individuals in this category, but the welfare benefits from the redistribution would not show up as better health, they would show up as those individuals being able to enjoy taking greater risks with their health. This is not a line of argument likely to be pushed by advocates for expansion of health insurance, particularly since taking risks with one’s health often has negative effects on other people besides the person choosing to take the risk.
The Isaac Ehrlich-Yong Yin paper “The Problem of the Uninsured—Implications for Health Insurance” is not yet available. Given its timeliness, I encouraged Isaac to do what he can to get his paper with Yong Yin out as a working paper as soon as possible.
Update: Here I am talking with Isaac. He says that at the time, investigating capital punishment was a way of trying to show there was some deterrent effect of the criminal justice system, which was in doubt at the time. Often in empirical work it is useful to look at the extreme—in this case, the extreme sanction of capital punishment—in order to most clearly identify effects. Although as I high school student reading quotations on note cards in debate boxes, I thought of his technique as “multiple regression,” in fact he was implementing one of the first simultaneous equations models ever estimated in the economics of crime.
As I type, Isaac has been emphasizing that the effects are different for people who didn’t have health insurance because they were free-riding on the health safety net (such as hospital emergency rooms, charitable organizations, the Hippocratic oath that obligates doctors to help people, etc.) than for people who didn’t have health insurance because they were being careful to do everything possible to stay healthy. The “take more risks” effect is coming primarily from the people who didn’t have health insurance because they are going to do everything possible to stay healthy. But most of the policy discussion has been about people free-riding on the health safety net.