Learning to import from neighbors

This study examines the impact of learning from neighboring firms on new importers' behaviors. We first develop a learning model in which firms update their beliefs concerning import prices in foreign markets. These beliefs are based on the number of neighboring firms that import from that market and the level and heterogeneity of import prices. The updating proceeds according to the Bayesian rule. The model predicts that a positive signal regarding import prices revealed by neighboring importers encourages importing entry from the same country. Such signals play a stronger role when revealed by a greater number of neighbors. In addition, the model predicts that a positive signal increases firm-level initial imports and survival rates, but decreases post-entry import growth. Using a transaction-level dataset of Chinese importers during 2000–2006, we find supporting evidence for the model's predictions. Furthermore, importer learning displays heterogeneous effects on different firms and exhibits a spatial decay structure. Our robustness checks include controlling for various fixed effects, employing an alternative entry definition, and using subsamples comprising ordinary trade firms and direct importers.