The prediction of users' response to changes in transport supply is a necessary input into the resource allocation procedures of any transport agency. The need for the predictions to be as accurate as possible results from the many competing alternative investments open to decision-makers. To meet this need, researchers in travel demand modelling have progressed from the aggregate level models of the 60's to the disaggregate models, based on the theory and methods of discrete choice analysis of the 70's. The effect of ignoring or mis-specifying choice sets, however, is especially apparent in modelling problems where it is expected that major changes in society, family structure, and economy will occur over time. In this paper the impact of ignoring or mis-specifying the effect of captivity (i.e. The state in which decision-makers' choice sets have but one alternative) on predicted users' response to transport supply changes, at both the disaggregate and aggregate levels is discussed. It is concluded from theoretical and empirical analyses that the biases introduced into parameter estimation and model forecasts by ignoring the existence of captivity, can be significant. The implication of this for transport analysts is that decisions based on demand estimates that do not consider captivity potentially result in misallocation or scarce resources. For the covering abstract of the conference see IRRD 290118.
Samenvatting