Innovative firms spearhead technological advancements (Chava et al. 2017) but may find themselves underfinanced due to their inherently risky high-tech nature, especially when borrowing from banks (Greenwood et al. 2010). Banks often specialise in certain industries to gain domain expertise that aids in assessing borrowers and monitoring loans (e.g. Gopal 2021, Beck et al. 2022, Giometti and Pietrosanti 2022). Our research (Gao et al. 2023) adds a new dimension: banks also accumulate technology knowledge that transcends industry boundaries. This is crucial because firms in different industries can leverage similar technologies. For instance, conveyors are used in mail services, steelmaking, robotics, and more. A bank experienced in one of these industries could apply its understanding of conveyor technologies when lending to companies in the other sectors. Similar examples can be found with artificial intelligence usage across many different industry sectors.
Using comprehensive patent data from the United States Patent and Trademark Office (USPTO) and syndicated bank loan data from the DealScan database, we uncover that banks develop a nuanced understanding of specific technologies through their lending practices. This accumulated technology knowledge notably lowers the bank's operational expenses associated with screening and monitoring loans. Importantly, these cost savings are then passed onto future borrowers using similar technologies, reducing their loan costs. It creates a cycle of mutual benefit that enhances the overall lending process.
Bank technology expertise and its economic value
To quantify the technology expertise banks acquire about their borrowers, we employ a unique, time-varying measure that captures the similarity of patent classes between the current borrower and the bank's prior borrowers. Specifically, we collect the patent data from the USPTO for the period from 1985 to 2020. The technology class of each patent is based on the Cooperative Patent Classification (CPC), jointly developed by the USPTO and the European Patent Office (EPO), which has replaced the United States Patent Classification (USPC) in 2013. Similar to Jaffe (1986) and Bloom et al. (2013), we measure the pairwise technology similarity between the prospective borrower and each of its bank’s prior borrowers in recent years as the cosine similarity of their patent classes. The average pairwise technology similarity is then used to measure how familiar the bank is with the borrower’s technology profile. This objective framework captures the dynamic nature of technological evolution, and the use of patents makes it highly relevant to the lending process.
Figure 1 Loan spread by borrower technology similarity
Under this framework, we find robust evidence of substantial economic impact. As shown in Figure 1, higher technology similarity between a borrower and its bank’s recent borrowers is negatively associated with loan spread. Moreover, this negative relation is statistically significant in a multivariate regression controlling for a wide range of borrower characteristics, loan characteristics, relationship lending, as well as an array of fixed effects such as borrower industry-year fixed effects, bank fixed effects, loan type and purpose fixed effects. Specifically, in our baseline model, a slight increase in our measure—just a one standard deviation rise—equates to about a four basis point reduction in loan spreads. In practical terms, it amounts to an annual loan cost-saving of approximately $170,000 for a mean (median) loan of $425 (166) million dollars in our sample. This isn't just a marginal reduction; it's a tangible economic benefit that enhances borrowers' bottom lines. Simultaneously, it allows banks to more effectively attract borrowers who align with their technological expertise, ultimately strengthening the overall financial ecosystem.
Further, our results are robust to additionally controlling for the concentration of the bank’s loan portfolio, product market competition, the borrower’s technology value, and alternative measures of loan costs. The documented negative relation between borrower technology similarity and loan costs is also not driven by the concerns that the borrower is a major prior borrower, or the bank has few recent borrowers. We then take several steps to investigate the underlying economic mechanisms. We start by showing that our measure is informative about the borrower’s creditworthiness and use a structural matching model to show that banks’ accumulated technology knowledge contributes to lending decisions. Matching borrowers to banks more familiar with their technologies, due to prior lending activities, achieves value maximisation for both borrowers and banks. By exploiting the adoption of intellectual property protection laws and the consummation of bank mergers and acquisitions, we use difference-in-differences estimations to confirm that such effect is causal.
Policy implications and recommendations
Moreover, the role of a bank's accumulated knowledge becomes even more pivotal in today’s global environment where intellectual property holds paramount importance, which not only translates to product innovations (Argente et al. 2020) but also long-run productivity growth (Mezzanotti and Simcoe 2023). We find that with enhanced intellectual property protection, borrowers operating within the realm of the bank's technology expertise receive larger discounts when seeking loans. This alignment with the global emphasis on intellectual property reinforces the relevance and importance of our findings.
Furthermore, our findings carry significant policy implications, particularly in the context of financing innovative technologies. The inherent high information asymmetry in such ventures often makes bank lenders hesitant to finance technologies they haven't previously encountered, contributing to the aforementioned challenges faced by innovative firms in securing bank financing.
One potential solution lies in government support to encourage bank financing to support technology innovation. This can be achieved through measures such as subsidising initial loans to innovative firms. As banks accumulate experience and knowledge on these new technologies, they become better equipped to assess the future prospects and underlying risks more accurately. Consequently, loan rates are likely to decrease over time, allowing for the gradual phasing out of government subsidies.
This approach draws parallels with successful policies aimed at small business financing and could prove particularly beneficial for the adoption of new technologies. By fostering collaboration between banks and innovative firms, we can stimulate technological progress, ultimately benefiting both the economy and society.
Conclusion
In a financial landscape where high-tech firms often find themselves under-funded, our research offers a new perspective. Banks don't merely gain industry-specific expertise through lending; they accumulate technology knowledge that can apply in multiple sectors. This creates a win-win dynamic that offers financial rewards to both banks and borrowers. It's high time that we appreciate and leverage this mutual benefit to address the funding challenges facing innovative companies today.
References
Argente, D, S Baslandze, D Hanley and S Moreira (2020), “From patents to products: Product innovation and firm dynamics”, VoxEU.org, 28 May.
Beck, T, O De Jonghe and K Mulier (2022), “Bank sectoral concentration and risk: Evidence from a worldwide sample of banks”, Journal of Money, Credit and Banking 54(6): 1705–1739.
Bloom, N, M Schankerman and J Van Reenen (2013), “Identifying technology spillovers and product market rivalry”, Econometrica 81(4): 1347–1393.
Chava, S, V Nanda and S C Xiao (2017), “Lending to innovative firms”, The Review of Corporate Finance Studies 6(2): 234–289.
Gao, M, Y Huang, S Ongena and E Wu (2023), “Borrower technology similarity and bank loan contracting”, CEPR Discussion Paper No. 18624.
Giometti, M and S Pietrosanti (2022), “Bank specialization and the design of loan contracts”, FDIC Center for Financial Research Paper, 2022-14.
Gopal, M (2021), “How collateral affects small business lending: The role of lender specialization”, Working Papers 21-22, Center for Economic Studies, U.S. Census Bureau.
Greenwood, J, J M Sanchez and C Wang (2010), “Financing development: the role of information costs”, American Economic Review 100(4): 1875–1891.
Jaffe, A B (1986), “Technological opportunity and spillovers of R&D: Evidence from firms' patents, profits, and market value”, American Economic Review 76(5): 984-1001.
Mezzanotti, F and T S Simcoe (2023), “Innovation and appropriability: Revisiting the role of intellectual property”, VoxEU.org, 27 September.
