Valuation of a statistical live saved: Experimental results from the Ticino

Abstract

afety, in particular traffic safety, is an issue very often only implicitly considered in many transport policies or projects. To enable an explicit evaluation of safety benefits related to a project or a policy, these benefits have to be expressed in monetary terms; in this way they can be introduced into cost-benefit analyses and weighted in the same way as all other costs and benefits. The problem now becomes: how to express these benefits in monetary terms? In the literature one solution proposed was to place an explicit value on the estimated reduction in fatality risk, known as statistical value of life (VOL), representing the economic value associated with one life saved. The thesis work will first approach the concept of VOL and then the discussion of the advantages and disadvantages concerning its use; subsequently different methods for its calculation will be introduced as well as some estimation procedures. The aim of this thesis, particularly, will be the application of two procedures from those described, the stated choice approach and the priority evaluator approach, with the aim of finding plausible estimates of the VOL. The experimental part concerns the development of a survey for a sample of 500 people selected randomly in the canton of Ticino (Switzerland). The survey comprises three sections related to the different approaches employed, and a final person questionnaire. All steps concerning the realisation of the survey, the sampling, as well as the pilot survey, will be accurately described. The results of the returned surveys (179 out of 500, with a response rate of 36%) are then described and analysed. For the VOL evaluation, firstly a descriptive statistic is used, to determine the reference values by the means and medians of the responses. Then, a discrete choice model, specifically the Multinomial Logit model, will be applied within each approach to estimate the VOL. The results of the experiment, which show consistent values despite its limited size, demonstrate also that the estimated value of life can depend on many factors: the approach used, the variables and their values and the alternatives presented. A brief comparison with other studies in the literature involving empirical estimates of VOL will also be presented.