Modeling the dynamics of passenger travel demand by using structural equations

Levels of demand over time are analyzed for five modes of passenger transportation. The data are for the modes: car driver, car passenger, train, bicycle, and public transit. These data are compiled from week-long travel diaries collected at six-month intervals from a nationwide panel in the Netherlands. Three types of empirical relationships are present in these panel data: (1) autocorrelative relationships, capturing temporal stability in demand for the same mode at different points in time; (2) contemporaneous relationships, capturing complementarity and competition among different modes at the same point in time, and (3) cross-lagged effects, potentially capturing systematic shifts in demand. Simultaneous equation systems are used to test the temporal stability of demand for each mode and the stationarity of the contemporaneous relationships among the modes. The dynamic structure both of trip rates and of travel times are modeled successfully according to several goodness-of-fit indices. The equation systems capture nonstationarity in the contemporaneous relationships, as well as important cross-lagged effects. These results quantify changes in the structure of demand over time in the Netherlands and are shown to be directly related to the event of a public transit fare increase.