Models of automobile ownership level and type choice are described by using a dynamic transactions model structure. The functional form of the model is two-stage nested logit: the higher level in the hierarchical decision process is a decision on the type of transaction in the car market. The lower-level decision is on type of car, which is conditional on the decision to buy a car. Automobile type alternatives are defined by make, model, vintage, and body type. The model was estimated with data from the Haifa urban area in Israel. The sample consisted of a choice-based (stratified) sample of 500 households that did not buy a car in 1978 and 800 households that bought a car during the same year. Each stratum was drawn at random from the respective population of the Haifa urbanized area. The models estimated in this paper are sensitive to attributes of the type of car, household characteristics, and accessibility by public transit and private car. The models take explicit account of the transaction costs that ae incurred when operating in the car market.
Abstract