The present article aims at providing better taxonomy and description of the different underlying factors hidden behind accident outcomes at country and regional level. The evidence is given on the existence of various structural factors on the example of two European countries, for which a Full Bayesian (FB) ecological regression model is applied to study different accident outcomes at regional level. This approach makes it possible to combine spatial and time variation within one modelling approach and allows to investigate existing spatial relationships. More precisely, the results of the FB ecological regression model in which a series of structural explanatory factors was tested one by one within road fatality risk model setting, using 2000-2006 data of 43 regions of Belgium, is presented. For example, population density or share of artificial surfaces exhibit a very strong elasticity of -0.20 (-0.34) meaning that an increase by 10% point will result in the decrease of road fatality risk by 2 and 3.4% respectively. Taking into consideration structural factors would allow isolating those factors which could be conveniently influenced by policy related measures, what in turn would result in more effective road safety management. Last, but not least, the role of regional and national politics on road safety performance is discussed. (Author/publisher)
Abstract