Spatial Modelling of Car Ownership Data: A Case Study from the United Kingdom

Applied Spatial Analysis and Policy, Volume 3, Number 1 / March 2010

Stephen Clark and Andrew O. Finley

“In this paper a model is formulated to estimate the strength of the relationship between household car ownership and income using cross-sectional data. Whilst reports of such studies are not uncommon in the transport literature, this study is different in that it takes explicit account of the spatial distribution of the data. By incorporating this spatial element in the model formulation, the residual errors in the model are uncorrelated and hence allows for the estimation of parameters that are, in a statistical sense, the best available. These spatial models are fitted to a large data set provided by the United Kingdom Office for National Statistics, covering the area of England and Wales. The recommended model form is a Hierarchical Bayesian spatial regression model with the parameters in the model estimated using the technique of Markov Chain Monte Carlo (MCMC). A common feature of all the spatial models is that the estimate of the elasticity of car ownership with respect to income is seen to be larger than that from a non-spatial model.”