Within-individual variation in preferences

The purpose of this research was to explore how the values of travel time (VTT) and preferences for different modes vary within individuals compared with the variation between observed trips. With 6-week revealed preference panel data and stated preference data from a mode choice context, both collected in Switzerland, a revealed-stated preference logit mode choice model was estimated. The model was applied to simulate how VTT and change in consumer surplus vary across trips within and between individuals over the 6 weeks in response to a hypothetical congestion-charging scheme. The variation in VTT arising from income differences was found to be substantially smaller than the variation in VTT between trips. Moreover, the variability in VTT averaged over all trips within each individual was considerably smaller than the variability in VTT for all observed trips. Therefore, the assumption that variation in VTT between observed trips reflects the variation in the average VTT between individuals, which is usually made in equity analyses, will overstate the between-individual variation. The results suggest that if intraindividual variation in preferences is not taken into account, the negative equity effects of congestion charges are likely to be overestimated.