(Interstate) Banking and (interstate) trade: Does real integration follow financial integration?

We conjecture that banks present in two regions charge the appropriate risk premiums for trade-related projects between these markets, whereas higher rates are charged for projects involving shipments to markets where they are absent. These differences affect regional trade flows. US interstate banking deregulation serves as a natural experiment to test our model's implication with the Commodity Flow Survey data. Difference-in-differences estimates suggest that the trade share of state-pairs that allowed pairwise interstate entry increased by 14% over 10 years relative to non-integrated state-pairs. Instrumental variables estimates suggest that an actual increase in bank integration from zero to 2.28% (the mean) increases trade 17% to 25%.