Modal Split Functions for a Swiss National Freight Transport Model
N Fries, J Wichser, ETH Zurich, CH; G Abay, G Moreni, Rapp Trans AG, Zurich, CH
This paper presents the methodology and results of a project on modal split functions for a Swiss national freight transport model. Stated preference experiments based on revealed preference data were conducted in the context of a shipper survey.
In recent years freight demand modelling has undergone a remarkable development. All around the world freight demand models have been developed and implemented on regional and national levels to provide a decision aid for logistics and transport planners. After the successful launch of a national passenger transport model for Switzerland in 2005 the Swiss Federal Office for Spatial Development (ARE) decided to develop a similar model for freight transport to provide a comprehensive tool for transport demand forecasting in Switzerland including passenger and freight transport. The creation of this new freight demand model has been divided into four sub-projects plus a preliminary study. This paper presents the methodology and results of the project focussing on the development of modal split functions that represent the shippers' demand elasticities.
The core part of this project consisted in preparing and executing a survey among shippers and freight forwarders in Switzerland and its neighbouring countries. Stated preference (SP) experiments based on revealed preference (RP) data were conducted within the framework of the survey to collect the data necessary for the following estimation of modal split functions for different commodity groups. As expected the analysis of Swiss transport statistics showed significant differences in transport mode shares on the one hand between commodity groups and on the other between internal and border crossing transport. In order to take these biases into consideration seven commodity groups were defined, each of them divided into internal, import/export, and transit transport. In order to guarantee an optimal survey design the development of the SP-experiments was closely followed by logistics and modelling experts. Based on the experts' experience it was decided to choose an adaptive design for the choice experiments based on revealed preference data collected during the first part of the interview. Between April and October 2007 about 100 telephone interviews were conducted supported by an online-questionnaire providing a sum of 180 SP-experiments. The survey covers the whole range of commodities shipped over medium and long distances in Switzerland thus representing the largest freight demand survey of its kind so far in Switzerland. Although in the end the sample size was found not to be sufficient to allow a fully detailed estimation of separate modal split functions for each subgroup of the single commodity groups, overall results are good and seem realistic in terms of comparability with existing literature data and Swiss transport statistics. Interestingly in several commodity groups the importance of shipments? on-time reliability was ranked equal or even higher than the one of transport cost. The impact of travel time was generally less important than the one of price and on-time reliability.
The paper sets off with an introduction to the particularities of the Swiss freight transport market, followed by a description of the methodology and assumptions leading to the experimental setup. After some comments on the survey execution phase part four deals with the results of the modal split functions estimation. The paper finishes with a summary and draws some conclusions including recommendations for further research in the field of freight demand modelling in Switzerland.
Association for European Transport