Most disaggregate models of household vehicle ownership are static models. Such models are not capable of predicting the time-path for the size and composition of the vehicle fleet. Predictions of this nature however are very important from a policy perspective, because they can give the rate of market penetration of low emission vehicles or vehicles with modern catalysers. Another modelling approach is to focus on vehicle transactions, such as replacement, disposal and acquisition. Vehicle transactions models can be developed as duration models. In the competing risks models in this paper, a state is defined as a combination of vehicles (possibly one) owned by the same household. The duration models explain the time duration models explain the time that elapses between the beginning of this state and the end of the state. The competing risks duration models are linked to models for the choice of make/model/age-of-car (conditional on replacement or acquisition), annual kilometrage and fuel efficiency. All models are estimated on data from three-waves (1992, 1993 and 1994) of a panel of Dutch car users. The paper also contains simulation results for the short- and medium-run impacts of policy measures (raising all fuel prices, raising prices of some of the fuels, raising the road tax, raising the road tax for large cars only) on composition of the fleet (e.g. by fuel type or engine size class), vehicle use and fuel consumption.
Abstract