Does valuation research provide a credible basis for cost-benefit analysis of safety measures?

Many studies have been made to obtain a monetary valuation of reduced risk of death, usually given as the value of a statistical life (VSL), which is a reduction in risk corresponding to the prevention of one fatality. This paper asks whether valuation research provides a credible basis for cost-benefit analysis of safety measures. A cost-benefit analysis is credible, if its results cannot be criticised by reference to the valuation studies forming its basis. It is argued that a credible basis for cost-benefit analysis in this sense does not exist. The monetary valuations of a statistical life vary enormously. The enormous diversity in values is increasingly accepted by researchers working in the field as inevitable and consistent with individual utility maximisation and thus not necessarily anomalous. Some recent contributions argue that the value of a statistical life ought to vary depending on, for example, income. Such reformulations of the theory underlying valuation studies mean that the choice of a particular value of a statistical life within the huge range of such values is not necessarily more justified than the choice of a different value.