## Thursday, October 17, 2002

### On the Dollar Value of Human Life

In a recent post to the Volokh Conspiracy, Eric Jaffe asks for a means of putting a value on human life: "no matter how crass it may seem, we eventually need some means of valuing lives (or more likely, life-years) and of comparing that value to other disparate values. We do this in any event, and at some point it would aid clear thinking to bring it out into the open a bit more. Society will "spend" lives on lots of things, and it would be nice if we did so with some amount of introspection rather than just by bumbling along."

Exactly right. Some method, even if only a rough-and-ready one, is needed for valuing lives. Saying that a life has infinite value sounds awful nice, but clearly we don't believe it -- for if we did, we would never take even the slightest risk to our own lives. If your death has a value of negative infinity, and there is even the smallest probability of death from whatever activity you'd like to do (driving, riding a roller coaster, eating rare meat, whatever), the expected value of the activity is also negative infinity. Anything multiplied by infinity is infinity, and no finite benefit could possibly be large enough to outweigh an expected loss of infinite magnitude.

So I will humbly suggest the economists' metric for valuing human lives. For any given risk to human life, find out the minimum amount of money it would take to induce the average human being to accept the risk; call this value X. If P is the probability of death from this particular risk, then solve for the value of life (V) using the following equation: PV = X. For example, if the probability of death created by some risky activity is 5%, and it would take \$50,000 to persuade the average person to accept this risk, then the value of life is \$1,000,000. This value of human life would not work for all purposes, of course; it depends, among other things, on the size of the risk. The dollar value needed to induce the acceptance of risk most likely increases at a greater rate than the size of the risk; the person willing to accept \$50,000 for a 5% chance of death would probably need *more* than \$100,000 for a 10% chance of death. In effect, the value of life we use in our calculations would be situation-specific, but not arbitrary.

No, I'm not joking. We cannot avoid the problem of weighing lives. Virtually every human activity involves some risk of death to someone, and risk can never be completely eliminated. It can sometimes be reduced (by spending increasing amounts of resources on safety measures, by reducing activity levels, and so on), but eventually our risk-reduction efforts are subject to diminishing returns. We have to spend more and more to achieve smaller and smaller reductions of risk, and it would be infinitely costly to reduce risk to zero. So the question is not whether to have risk or not, but how much to have. Placing an infinite value on human life does nothing to address that question.

(This proposal is certainly not my own idea; see almost anything in the work of Kip Viscusi, among others, for further details on this approach.)