Estimating road transport costs between and within European Union regions

Transport costs are a crucial element of applied spatial models such as spatial macro-economic models or interaction models for e.g. trade, migration or commuting. However, good estimates of these costs are not available at lower levels of spatial aggregation. In applied work, the distance between two regions is often approximated by the distance between the largest city in each region. For costs within regions, researchers often resort to crude ad-hoc approximations relating internal distance to the area of the region assuming a uniform internal distribution. This paper improves on this by considering averages of transport costs calculated between extremely large random samples of centroids which are drawn from a population grid. This allows calculating distances, travel times and transport costs both between and within regions, while taking into account the unequal distribution of population within the regions. The use of a detailed road network and many auxiliary datasets allows performing policy analysis. We assess the impact on transport cost of an increase in fuel prices, and find that it has a relatively high effect on transport costs in Eastern Europe. We evaluate transport infrastructure investment of the European Cohesion Policy program 2014-2020. The largest decreases in transport costs are found in targeted regions in Eastern Europe while the effect is much smaller in targeted regions in Southern Europe, suggesting decreasing returns to further transport infrastructure in these regions. We find significant inter-regional spillovers to regions directly bordering the regions targeted by the policy in Eastern Europe, such as in Germany and Austria, but also Finland and Northern Italy. The positive spillovers to EU regions in Western Europe are quite small.