We combine high-frequency market price data and the road network to estimate the size of internal trade costs in Rwanda and evaluate whether better connectivity reduces these costs. We find that internal trade costs in Rwanda, defined as a function of distance are ten times higher than in the u .s ., three times higher than in Ethiopia, and two times higher than in Nigeria. Holding Rwanda pass-through rate at the u .s ., Ethiopian, and Nigeria levels would have reduced the size of internal trade costs by half. We also find that these costs are particularly high for manufactured products. We test two mechanisms related to the intensity of competition, and the accessibility to road infrastructure. We show that intermediaries are charging higher markups in locations where there is less competition. We also show that better road connectivity through access to feeder and national roads matter as connecting a location to a feeder road reduces trade costs by approximately two cents per mile.
We use the microdata underlying the Ethiopian CPI to examine the spatial dispersion in local prices and availability of 400 items across 110 cities. Remote cities face higher prices and have access to fewer products. Large cities also face higher individual prices but enjoy access to a wider set of products. To assess the welfare implications of these patterns, we build aggregate cost-of-living indexes. The cost of living is higher in remote areas, but not systematically related to population size. We then show spatial differences in the cost of living are a significant determinant of migration flows across Ethiopian regions.
The measurement of trade costs in Ethiopia
This paper compares trade costs estimated using spatial price differences with the actual trade costs incurred by the World Food Program in Ethiopia. Based on a model of intermediated intra-national trade, I estimate the trade costs elasticity with respect to distance for both non-food and food products. I find that the estimated trade costs elasticity is lower than the actual trade costs elasticity for non-food products. The type of products considered and the difference in the mode of transport explain this finding. For food products, I find a similarity between the elasticities of estimated and actual trade costs with respect to distance. This result suggests that inferring trade costs from spatial price dispersion is a good method to approximate actual trade costs.