Classic and bayesian estimation of subjective value of time

Abstract

In this paper, we estimate different travel time components in an urban transport context in Bogotá, by using discrete choice models. For the analysis, we evaluate two different approaches for model estimation: the maximum likelihood estimation (classical) and the Bayesian estimation, allowing a comparison between both approaches when are used to estimate the subjective value of time (SVT) as the marginal rate of substitution between time and cost. We found that the average SVT when we used the classical estimation is US$0.14/min; meanwhile, the access time was US$0.21/min and the penalty for a transfer is US$0.74. On the other hand, the average values estimated when using Bayesian approach were US$0.15/min for the SVT in vehicle, US$0.21/min for the access time and US$0.83 for the transfer penalty.