Paradoxes of rationality in road safety policy

Rationality is an ideal for transport safety policy. As developed within normative welfare economics, rationality denotes the efficient use of safety measures based on cost–benefit analyses that include all relevant impacts of the measures. Efficiency in the technical sense of the term provides a perfectly clear and precise guideline for policy priorities. Nevertheless, some choices that are guided by cost–benefit analysis may strike us as paradoxical or counterintuitive. A paradox of rationality refers to any situation in which conflicting choices can both be defended as rational. This paper discusses a number of choices that may seem paradoxical. The first involves the choice between options that have identical impacts on safety, but in which these impacts are valued differently. The second deals with the tendency for preference reversals to occur when preferences for the provision of safety are aggregated. The third discusses the inability of conventional measures of willingness-to-pay to reflect the intensity of preferences. The fourth concerns the tendency for policy choice to favour the rich at the expense of the poor when willingness-to-pay is not adjusted for the marginal utility of money. A fifth situation refers to the fact that a policy option that looks attractive ex ante may fail an ex post compensation test because utility functions depend on health state. There is a potential conflict between individual and collective rationality with respect to the costs and benefits of some road safety measures. When developing a road safety programme, a set of road safety measures whose benefits exceed the costs when considered as stand-alone measures could have benefits smaller than cost when combined in a programme consisting of all the measures. Finally, there is a potential conflict between efficiency and negotiated consensus as mechanisms of resource allocation in the public sector. The sources of the paradoxes and ways of avoiding them are discussed. Some of the paradoxes can be avoided if changes in risk are valued in terms of a fixed price per unit of risk rather than according to a non-linear demand function.