Are capital intensive firms the biggest exporters?

This paper starts out from the observation that the export shares of firms (export to sales ratio) vary greatly among firms, and tend to be systematically related to the firms' capital labour ratios. This observation cannot be explained by the standard heterogeneous firms and trade model by Melitz (2003), which predicts that all exporting firms have identical export shares. In our model, we relate the difference in export shares to firm level differences in transport costs. Two factors influence a firm's transport cost in our model. First, firm scale can affect transportation costs, making freight rates lower for large firms. Second, we allow for an association between the capital intensity of a firm and its transportation costs. In accordance with data, we assume this relationship to be sector specific. This implies that our model can generate the result that more productive and capital intensive firms have higher export shares due to scale economies in transportation, but the model can also generate the opposite pattern that more capital intensive firms have lower export shares due to a strong positive association between capital labour ratio and transportation costs. We use Japanese manufacturing firm level data to calibrate our model by matching firm level export shares to data sector by sector. Regressing the calibrated transportation costs on actual data then shows that the calibrated (calculated) numbers can explain about half of the variation in the data.