Abstract
This article develops a disaggregate discret-choice model for risky travel alternatives, considering travel time as a random variable. The theory of rank-dependent expected utility is used to describe traveler behavior. It disentangles attitude toward risk of travel time loss (perception of time distributions) from attitude toward wealth (perception of time outcomes). The key assumption is that decision makers formulate their own beliefs on probability distributions whenever information is public knowledge. The concept of travel-time reliability premium is then defined. It generalizes the notion of value of travel time savings by proposing a measure of the willingness-to-pay to get another time distribution. An application to air route choice is demonstrated by using a sample of regular business passengers who belong to an airline loyalty program. The empirical approach relies on estimating a Box-Cox Logit model.
Notes
*The rank-dependent expected utility theory appears as a special case of the CPT when outcomes are considered as either gains or losses. From a historical perspective, we point however out that the rank-dependent expected utility theory emerged before the CPT but nearly at the same time as the prospect theory (PT, CitationKahneman and Tversky(1979)). The CPT inspired then from the rank-dependent expected utility to answer some inherent limitations of the PT.
**The transformation they propose is defined in a different framework in which ordered outcomes are losses and gains.
†Unfortunately, because of privacy of data, I am not authorized to publish statistics on fares, trip frequency, and market shares of routes. Therefore, a summary table pertaining to the variables is not possible.