Disaggregate Urban Mode Choice Models: A Review of British Evidence with Special Reference to Cross Elasticities.

As part of an EPSRC funded research project (GRK52522) entitled 'National Multi-Modal Travel Forecasts', a review of the demand forecasting literature has been conducted. The principal aim of this project is to develop a set of national and regional travel demand forecasts by land-based modes. Such models will contain a set of own and cross-elasticities which can be used in strategic demand forecasting. A review of the literature, alongside fresh empirical work, makes an important contribution to this. A review of aggregate models is contained in Clark (1996). Such models are based on collective behaviour such as market shares or travel volumes. In contrast, disaggregate models make the individual decision maker the unit of observation. Within this project, Wardman (1997a) has provided a review of disaggregate mode choice models developed in the inter-urban context in Great Britain whilst Whelan (1997) has provided a review of car ownership modelling and forecasting.

This paper provides a review of comparatively recent research involving disaggregate mode choice models which have been developed in Great Britain in the urban context. The emphasis of this research is on cross elasticities for three reasons:

Mode choice models are well suited to the estimation of cross-elasticities;

The own elasticities provided by disaggregate mode choice models are underestimates since they do not account for behavioural responses other than mode switching (Oum et al., 1992);

There has long been a view (Dodgson, 1991) that there is insufficient evidence regarding the degree of interaction between modes and this view remains (Acutt and Dodgson, 1995; Wardman et al., 1997)

In contrast, aggregate models are well suited to the analysis of own elasticities since they take into account changes in the total number of trips yet they are generally limited in the extent to which they examine inter-modal competition and hence generally provide little evidence on cross-elasticities.

In this paper, we have drawn upon studies made available to us as part of a review study conducted for the Department of Transport into the value of time (Wardman, 1997b). Much of this evidence was provided on the basis that the identity of the studies remains anonymous. We have therefore provided the key parameter estimates from 34 studies without revealing the identity of these