Valuing Life_ A Plea for Disaggregation

2004] VALUING LIFE 433
should be undertaken even if WTP numbers do not justify it. Part of
the problem with those numbers is that individual behavior will not
reflect the “catastrophe premium” or “extermination premium” that
would almost certainly emerge if it were possible to test for it. People
may be unwilling to pay anything to avoid a risk of 1/100,000,000 that
they themselves face; but if they were told that every person in the
nation faced this risk, they might come up with a significantly higher
figure. It would take the right question to produce the higher
numbers. Another the problem is that WTP is not an adequate
measure of social responses to catastrophes—in part because people
are not familiar with making choices about risks of that sort.
In my view, this is a sound objection to the use of a low or nearzero
VSL in the context of catastrophic risks, even if the WTP
calculation would produce that VSL. As Judge Richard Posner
shows, 173
this is an important point when government is considering
how to respond to small risks of catastrophic harm. But notice that
the objection has built-in limitations. It does not apply to the
overwhelming number of cases in which VSL is used. In those cases,
the risks in question are 1/10,000 to 1/100,000, and no large-scale
catastrophe is at issue. Here, then, is a limitation on the use of WTP,
but the domain of the objection is restricted.
f. Third-Party Effects. A final objection would point to effects
on third parties. If outsiders would be adversely affected by the
undervaluing of a particular risk, and if their welfare is not being
considered, then the WTP calculus is seriously incomplete. This point
demonstrates a general and badly neglected problem for WTP as it is
currently used: agencies consider people’s WTP to eliminate
statistical risks, without taking account of the fact that others—
especially family members and close friends—would also be willing to
pay something to eliminate those risks. John might be willing to pay
$25 to eliminate his own risk of 1/100,000, but his wife, Jane, might be
willing to pay $25 to eliminate John’s risk also. If regulators add the
WTP, on John’s behalf, of John’s friends and relatives, the total WTP
might soon exceed $100. This is a real problem for existing uses of
WTP.
But thus far the discussion has been assuming that there are no
third-party effects. The argument for using WTP, on the stated
assumptions, is that government should not force people to buy goods
173. See id. at 196–98.