Two-sided matching in the loan market

This paper investigates the matching between banks and firms in the loan market. We estimate a many-to-one two-sided matching model using the Fox (2010) matching maximum score estimator. Using data on the U.S. loan market from 2000 to 2003, we find evidence of positive assortative matching of sizes. Moreover, we show that banks and firms prefer partners that are geographically closer, giving support to the importance of physical proximity for information gathering and expertise sharing. We also show that banks and firms prefer partners with whom they had prior loans, indicating that prior loan relationship plays an important role in the selection of current partners.