Value of travel time changes: theory and simulation to understand the connection between random valuation and random utility methods

This paper identifies and illustrates the theoretical connection between the Random Valuation (RV) and Random Utility (RU) methods for Value of Travel Time Changes (VTTC) analysis. The RV method has become more and more popular recently, and has been found to lead to very different estimation results than conventional RU models. Previous studies have reported these differences but did not explain them, which limited the confidence in the RV model as a useful foundation for transport policy analysis. In this paper, we first analytically show in what way exactly the two models are different and why they may generate different estimation results. Based on this deeper understanding of the connection and difference between the two models, we formulate hypotheses regarding the conditions under which differences in estimation results are expected to be smaller or larger. Using synthetic data, we empirically test these expectations. Results provide strong support for our hypotheses, allowing us to derive a number of practical recommendations for analysts interested in using the RV and RU models in their VTTC-analysis.