Expertise zum Projektbericht VALOR

Cost-benefit analysis for investment projects and for measures that impact on road safety require scientifically sound and valid figures for the willingness to pay of the society to increase road safety. Stated Chocie (SC) methods have increasingly been used to establish such values, both in the academic community and in practice.
These methods are based on SC surveys where individuals have to choose one out of different alternatives that are characterised by the relevant attributes for the choice task, with different attribute levels that are defined within an experimental design. Based on the assumption that the respondents chose the alternative with the highest utility, the individuals reveal their preferences and implicitly also their willingness to pay. Econometric modelling techniques applied to the data from the SC surveys enable the indirect estimation of the willingness to pay for non-market goods (ORTÚZAR and WILLUMSEN, 2011).
Meanwhile, there are two such studies for Germany available. These are the VALOR- study conducted by an international consortium for Belgium, France, the Netherlands and Germany with contribution of the BASt (Project VALOR–Value of Road Safety, cited as VALOR, 2021); and a pilot study for the willingness to pay for road safety in Germany (Project FE 82.0689/2017, cited as OBERMEYER et al., 2020). This report analyses the VALOR-study regarding the methodology and the estimation results against the international experience in the field and compares the results with those from OBERMEYER et al. (2020).
As the analysis shows, both studies are based on a sound scientific concept that meets the international state of the art in the academic community and in the transport policy area. However, the different aims of these studies (a methodology that enables country comparisons in VALOR, 2021, and the character of a pilot study with a smaller and biased sample in OBERMEYER et al., 2020) lead to considerable differences regarding the sample size, representativity, questionnaire design (in particular with respect to the used reference values), and the results. As a consequence of these differences, none of these studies meet the requirements of a comprehensive and representative willingness to pay study for Germany.
The results from both studies deliver valuable insights, however, the quantitative results are not suitable for a direct integration into the accident costs calculation framework of BASt and into the German cost benefit analysis for the Federal Masterplan on Transport Infrastructure Projects. Both from the scientific and practical point of view it is recommended to use the insights for the design of a comprehensive German willingness to pay studies for transport safety. The most important conclusion from the comparison of both studies is that the choice experiments should reflect as close as possible the reality faced by the surveyed individuals. The hypothetical road trip presented to respondents in VALOR (2021) with the predefined route length, the chosen reference values and the priors used to define the variation of attribute levels in the experiments do not adequately reflect the German situation. Therefore, the reference values should be based on answers of the respondents regarding their last trip.
Further analysis and justification is necessary regarding the presentation of accidental risks in the survey (absolut versus relative risk), the choice of roads type (experiment with a less safe road infrastructure, for example Landstraßen), the necessary differentiation of types of injuries and on the inclusion of a no-choice option in the experiment. Furthermore, both studies have shown that the treatment of apparently inconsistent choices and lexicographic answers is important. Finally, none of the studies have analysed the willingness to pay for risk reduction of non-car users. This holds also for externalities such as suffer and grief of relatives and friends, as well as for differences in the willingness to pay for risk reduction when travelling with passengers, possibly with children and grandchildren. For these aspects a general need for research – not only in Germany – does exist.