Banks’ Technology Expertise and Cost of Debt: Evidence from Syndicated Loans
DescriptionInnovation is the engine of economic growth (Solow, 1957). However, the inherent information asymmetry problem with innovation activities can lead to credit rationing or increased cost of debt, resulting in insufficient innovation investment (Mann, 2018). As financial intermediaries specializing in collecting private information for delegate monitoring, banks play critical roles in reducing capital market friction, especially for opaque borrowers such as innovation-intensive firms. Acknowledging the critical role of the banking sector, a growing literature has examined the impact of bank activities on the financing of innovation. In this proposed project, we examine how a most relevant dimension of bank expertise, banks’ technology expertise, affects the pricing of bank loans. Theoretically, a bank’s higher technology expertise with a borrower can help reduce the information asymmetry and agency problems in bank loans and lead to a lower cost of debt. However, it may also suggest a bank’s excessive exposure to specific technologies and result in a higher cost of debt.We propose a measure relying on the technology distribution of collateralized patents in a bank’s portfolio to capture the bank’s technology expertise. In recent years, using intangible assets such as patents as collateral becomes a popular mechanism to mitigate the market friction for debt financing (Mann, 2018). We argue that banks can gain technology expertise through the process of valuation and monitoring of pledged patents. We will employ the USPTO Patent Assignment Database to identify patents pledged to a bank and to calculate the technology expertise. We then plan to use a bank M&A setting, which exogenously changes a bank’s technology expertise with a borrower, to establish causal links from banks’ technology expertise to loan pricing. The impacts of loan syndication structure and borrower innovation performance will also be investigated.
|Effective start/end date||1/10/22 → …|