The value of a statistical life: some clarifications and puzzles

Many people have wondered why the US government conducts cost-benefit analysis with close reference to the value of a statistical life (VSL). It is helpful to answer that question by reference to the “Easy Cases,” in which those who benefit from regulatory protection must pay for it. In such cases, WTP is usually the right foundation for VSL, because beneficiaries are hardly helped by being forced to pay for regulatory protection that they believe not to be in their interests. In the Easy Cases, arguments from both welfare and autonomy support the use of WTP and VSL (with potentially important qualifications involving imperfect information and behavioral market failures). The analysis is less straightforward in harder cases, in which beneficiaries do not pay for all of the cost of what they receive (and may pay little of that cost). In such cases, arguments from welfare and autonomy might not lead in any clear direction. In the harder cases, regulation might be justified on welfare grounds even if the cost-benefit analysis (based on VSL) suggests that it is not. In principle, a direct inquiry into welfare (the master concept) would be preferable to use of cost-benefit analysis. In the harder cases, distributional considerations might also count in favor of proceeding (as prioritarianism suggests). But at the current time, direct inquiries into welfare consequences and into distributional effects are challenging in practice, and hence regulators should generally rely on cost-benefit analysis, making welfarist adjustments, or adjustments based on distributional considerations, only in compelling cases.