Abstract

We consider the value of a statistical life (VSL) at risk in interurban road accidents for car users. We use a stated preference (SP) experiment involving three variables: travel time, toll charge and accident in a medium distance route choice context. Apart from the SP exercise, respondents answered an accident risk perception questionnaire loosely based on the work of Jones-Lee and colleagues. The data was subject to a thorough set of consistency tests, including departures from the linear compensatory hypothesis (i.e. lexicographic and non-linear compensatory individuals). The paper discusses the results of several classes of SP route choice models involving risk. These allow to infer subjective values of time (SVT) and willingness-to-pay (WTP) estimates for reductions in accident risk, which allow to derive VSL for the highway under consideration. The SVT values are compared with values obtained previously in the country in order to check respondents understanding of the SP experiment. Our results indicate that use of transferred values from the USA is not recommended as these are significantly different from our estimates.