Transport costs, trade, and geographic concentration: Evidence from Canada

Our objective is threefold. First, we explain how to estimate transport costs and the geographic concentration of industries using trucking microdata and geocoded plant-level data. Second, we document that transport costs explain between 25% to 57% of the observed relationship between trade and distance across Canada's economic regions. Last, we show that changes in transport costs have a substantial impact on geographic concentration patterns for vertically linked industries, depending on the strength of the links. A one standard deviation increase in transport costs leads to a 0.02 standard deviation decrease in geographic concentration for industry pairs at the bottom decile of the input-output coefficient distribution, whereas the corresponding effect at the top decile is a 0.02 standard deviation increase. This gap between weakly and strongly linked industries stands up to a wide range of specifications and is robust to instrumental variables estimations.