Abstract
We use a rich data set on long-distance freight trains in Norway to estimate the effect of variation in train unreliability over time on transport demand within a fixed-effect framework. The train services that we study carry high-value consumer goods and compete heavily with trucks. We find that if the share of substantially delayed trains increases by 10 percentage points, freight volume goes down by about 3 per cent. In contrast with evidence from passenger transport, we do not find that the demand elasticity with respect to reliability is lower than the one implied by stated preference (SP) studies.