Yuichi IkedaBack to index

  • Who buys what, where: Reconstruction of the international trade flows by commodity and industry


    We developed a model to reconstruct the international trade network by considering both commodities and industry sectors in order to study the effects of reduced trade costs. First, we estimated trade costs to reproduce WIOD and NBERUN data. Using these costs, we estimated the trade costs of sector specific trade by types of commodities. We successfully reconstructed sector-specific trade for each types of commodities by maximizing the configuration entropy with the estimated costs. In WIOD, trade is actively conducted between the same industry sectors. On the other hand, in NBER-UN, trade is actively conducted between neighboring countries. This seems like a contradiction. We conducted community analysis for the reconstructed sector-specific trade network by type of commodities. The community analysis showed that products are actively traded among same industry sectors in neighboring countries. Therefore the observed features of the community structure for WIOD and NBER-UN are complementary.


    In the era of economic globalization, most national economies are linked by international trade, which in turn consequently forms a complex global economic network. It is believed that greater economic growth can be achieved through free trade based on the establishment of Free Trade Agreements (FTAs) and Economic Partnership Agreements (EPAs). However, there is limitation to the resolution of the currently available trade data. For instance, NBER-UN records trade amounts between bilateral countries without industry sector information for each type of commodities [1], and the World Input-Output Database (WIOD) records sector-specific trade amount without commodities information [2]. This limited resolution makes it difficult to analyze community structures in detail and systematically assess the effects of reduced trade tariffs and trade barriers.