Abstract

We develop three propositions about the ex ante value of reduced risk. If there is a continuous outcome distribution and if self-protection influences outcome probability and severity, then: (1) unobservable utility terms cannot be eliminated from the ex ante value expressions; (2) knowledge of the convexity or the nonconvexity of dose-response functions is insufficient to sign changes in these expressions; and (3) self-protection expenditures need not be a lower bound measure of these expressions. Therefore, many restrictions applied in recent empirical work on the economic value of risk changes are not immediately transferable to settings where endogenous risks prevail.