I provide novel evidence on the real costs of political interference in bank lending. Analyzing staggered state elections in India, I show that politically motivated increased bank lending to farmers before elections crowds out lending to manufacturing firms. These lending distortions are larger where farmers have more political weight and where incumbents have more influence over banks. Reduced bank credit forces manufacturing firms to cut production and operate at lower factor utilization. I also provide evidence suggesting politically motivated increased agricultural lending before state elections contributed towards excessive indebtedness of farmers and a subsequent costly bailout in 2008.

Using a novel dataset that combines information on customer-supplier trade relationships with information on firm-bank lending relationships, we show that common banks that lend to firms at both ends of a trade link grow and strengthen such trade relationships. To establish causality, we use bank mergers, which generate exogenous variations in the presence of common banks, and show that common bank relationships between customers and suppliers increase trade relationships by 49.6%. We find that the role of a common bank is greater when it is more informed and when supply chains suffer from larger information and holdup problems. We also document that suppliers with common banks face lower spillover risks from a distressed customer. Overall, our findings show the unique role of banks in driving inter-firm growth and investment by mitigating information and holdup problems, which arguably leads to greater economic growth.