Swiss Value of Travel Time Savings

Switzerland was the first German speaking country to initiate an official study of the value of travel time savings (VTTS). The context of the study is the development of an official cost-benefit guideline in the context of a wider sustainability analysis sponsored by the Swiss Highway Agency (Bundesamt fŸr Strassen) (see www.nistra.ch). The paper reports the work undertaken by the authors for this study with a special emphasis on those elements of the survey work and modelling experience which are often less well reported: sample development from the original RP sample to the respondents of the SP, selection of the choice contexts, final wording of the SP, stability of the VTTS as a function of the model specifications. The sampling frame of the study was an on-going survey by the Swiss Federal Railroads on the travel behaviour of the Swiss population. In the main study 51% of the sample of 2300 respondents replied (two reminders were issued). The pre-tests had shown that only the route and mode choice experiments produced reliable results, while a destination choice based experiment returned results inconsistent with reasonable expectations. The pre-tests also showed that the presentation of the different travel time elements had a strong impact on the responses. The main study sample indicated some over-representation of wealthier males with a higher than average usage and commitment to public transport. We will report the selectivity analysis and an analysis of the reply speeds. We will also report the results of removing answers of respondents who answered inconsistently. The extensive specification search indicated a substantial sensitivity of the VTTS estimates to the inclusion of variables describing the behavioural inertia of the respondents and to trip distance and respondent income. The models included random parameter formulations and non-linear elasticity formulations for distance and income. They were estimated using Biogeme, an advanced choice model estimator developed by M. Bierlaire. We will describe the sequence of estimation results for the different SP experiments and interpret them against the background of recent European experiences. The recommended mean VTTS values are comparatively low in terms of average hourly wage rates (about 30% across all purposes), but show a substantial variation by income and trip distance.