To date integrated models of car ownership and use have not been developed in the UK. This paper describes the derivation of such a model for the Department of Transport based on 1991-93 National Travel Survey Data. In theoretical terms the combined modelling of car ownership and use based on disaggregate household data has a number of desirable characteristics: (1) car ownership and use have a large number of shared determinant factors, such as income; (2) based on a rigorous micro-economic framework the interrelationships between error structures underlying ownership and use decisions can be exploited to the full; and (3) the wide degree of variation in dependent and explanatory variable embodied in a large cross-sectional dataset provides an excellent basis for statistical analysis. Comparison with past aggregate and disaggregate car ownership and car use analysis methods is discussed. The paper explores the empirical techniques used to estimate a joint car ownership and use model, and the complexities involved in the sophisticated econometric techniques required. A key issue in the study was the explanatory power of the model when compared with base sample properties, and the performance characteristics of the model in forecasting mode. Empirical findings are discussed in detail. The study concludes with a discussion of the role of a joint model of car ownership and use in future National Road Traffic Forecasts.
Abstract