Testing for inertia effect when a new tram is implemented

Abstract

Ignoring the inertia effects on transport-mode choice behavior may lead to erroneous decisions in transport policy. Around changes in the transport system, the majority of studies on inertia have relied on combining Revealed Preferences (RP) and Stated Preferences (SP) obtained prior to the introduction of new transport modes and measuring inertia as the effect that the real choices have on the choices in the hypothetical new scenarios. In this study, we analyze the role of the inertia using a novel panel data from the same set of individuals composed of two waves. The first wave was gathered before a new tram came into service and consisted of a RP survey and a SP survey which included the new public tram as a hypothetical alternative. The second wave consisted of a RP survey conducted two years later, after the tram started operating. Using these two waves, we estimate panel mixed logit models and found a significant inertia effect only between the RP waves which, having accounted for changes in other factors, increases the probability of choosing the car after the tram implementation. However, we did not find inertia effect on SP, hence taking into account only the RP-SP outcomes before tram might have led to wrong conclusions about the effect of the transport intervention on the modal share. Furthermore, we compare models with and without inertia effect and conclude that the models with inertia provide better fit to data, smaller direct car elasticities and increasing asymmetric effects between the car and public transport.