Immigrant Networks and Remittances: Cheaper together?∗

We estimate the causal effects of immigrant networks on individuals’ remittance sending behavior for migrants from many different countries residing in Spain. Our methodology addresses typical issues that arise when estimating network effects: reverse causality, common unobserved factors, and self-selection. In particular, we instrument the size of networks by predicting the number of migrants in each location using the location’s accessibility by distinct methods of transportation and information about how migrants from each country arrived in Spain. Our findings show that immigrants from above-average remitting countries remit more if they live in larger networks. Testing for mechanisms of network effects, we also find that these migrants are more likely to send remittances via bank transfers, which suggests that large networks of individuals who remit a lot might be better at sharing information about cheaper remittance channels (bank transfers compared to money orders in post offices or agencies). In line with this hypothesis, we find that due to network effects migrants shy away from the most expensive remittance channels, potentially freeing resources for additional remittances. Furthermore, cost spreads between the most expensive and cheapest providers are lower for countries characterized by high remittances and stronger networks, suggesting that network effects might be competition-enhancing. JEL classification: F24, J61, F22, O15, A14