Increasing relative risk taking in a choice context with source-dependent travel time risks

We use data on travel decision making to estimate a structural model that elicits travellers’ risk attitudes, with a focus on the mode-specific nature of risky travel time. We find notable intra-individual heterogeneity in risk attitudes when a decision maker faces two modes with distinctive risks in terms of different travel time variability shown in the choice experiment. Our empirical results show that, increasing relative risk taking, characterised by a traveller’s stronger risk-taking behaviour towards a more-risky mode with greater travel time variability, plays a statistically and behaviourally significant role in explaining mode choice behaviours. In addition to intra-individual heterogeneity, we allow the individual-specific risk attitude parameters to vary across the sampled population both randomly and systematically with an important observable variable (i.e., trip purpose). The choice dataset is best explained when the model accounts for intra- and inter-individual differences simultaneously. The highly negative correlation between two risk attitude parameters further highlights that ignoring inconsistent behaviour within a decision maker may lead to biased findings, thus misleading policy implications.