Learning and the Value of Relationships in International Trade

Abstract

We use U.S. import data to explore the characteristics of importer-exporter relationships. While many relationships split up after only one period, others are very persistent. Even when a relationship breakup occurs, importers overwhelmingly replace imports from that partner by buying from a supplier familiar to them. Furthermore, 43% of new product purchases come from firms that the importer interacted with in an earlier year. These results indicate the presence of large matching frictions in international trade. We develop a model to study and quantify the role of reputation and learning in explaining the observed patterns of importer-exporter relationships. Predictions of the model regarding the correlation between switching behavior and source country institutions, the number of export partners, and the number of products purchased are borne out in the data.