An empirical investigation of the distribution of travellers’ willingness-to-pay: A comparison between a parametric and nonparametric approach

Abstract

Understanding the distribution of the willingness-to-pay (WTP) to reduce travel time or other unpleasant travel attributes experienced by travellers is crucial to making well-informed choices about infrastructure investments and the pricing of services. Given multimodality, heavy tails, and the impact of averaging errors in the classical parametric distribution-based analysis of WTP, this study compares WTP distributions across six distinct datasets using a range of model specifications, including multinomial logit (MNL), mixed logit (MMNL) and nonparametric logit mixed logit (LML). The models are estimated directly in willingness-to-pay space, highlighting that WTP is not uniformly distributed with a high likelihood of having a multi-modal distribution across the population. The findings confirm the intuition on the presence of negative WTP for some undesirable travel attributes. This study raises awareness among transport researchers, modellers, and authorities about using the correct WTP distributions in practice. It adds some empirical evidence regarding the true underlying WTP distribution of various travel attributes to the transport literature.