Self-protection has been used to help define lower bounds on the value of statistical life (VSL). We show circumstances exist in which (1) the lower bounds are so low as to be more misleading than informative; and (2) the bound is an upper bound on the population's average VSL. The relationship between the bound and VSL depends on the degree and nature of individual heterogeneity, the fraction of the population buying self-protection, and the price and market setting for self-protection. Although some factors are observable, their impact is difficult to assess because they interact with unob-servable population characteristics.
展开▼