In this paper we determine the effect of transport cost on commuting flows, on the basis of an analysis of home-to-work journeys between municipalities in Denmark. Special attention is given to a proper estimation method and the form of the distance-decay function. It appears that neither an exponential nor a power distance-decay function fits the data well. The specification of log trips as a (downwards) logistic function of log cost results in a better fit. We find that the cost elasticity of commuting reaches a value of –4 for distances around 24 km, while it is close to 0 for both very short and very long distances. Finally, we demonstrate that the choice of functional form for distance-decay can make an important difference for predictions concerning the effect of infrastructure improvements on commuting flows.
Spatial Interaction Distance-decay Function Commuting Denmark Estimation Heteroscedasticity