Description:
The gravity literature has focused on distance, borders and contiguity to measure geography's impact on trade. We add value to this literature in terms of data, method and assessment of effects. First, we expand existing geographical databases by adding topographical features. We supply novel detailed primary data on the international European river network. We also construct a new indicator for the ruggedness of trade routes for more than a thousand European country pairs. Second, we introduce a new approach to differentiate between contemporaneous versus historical trade costs. Third, we assess the impact of topography on trade across Europe by applying two-stage structural gravity estimations, identifying bilateral trade costs on the basis of a worldwide panel of manufacturing trade including countries' domestic trade. We show that positive effects of rivers on trade are less important - and also less persistent over time - than the negative effects of mountains. While border effect estimates remain largely robust against variations in topography, much of the historical - and all of the contemporaneous - trade costs usually attributed to non-contiguity can be accounted for by topography. Finally, counterfactual simulations for western (along the river Rhine) versus southeastern (along the river Danube) European countries suggest that historically topography may have contributed to the marginalization of southeastern Europe in European trade.