Bank Learning

In this thesis I examine bank learning in the context of lending to large publicly listed firms. I hy-pothesize that banks learn valuable private information about their borrowers over the course of lending relationships, and that banks use this proprietary information to set subsequent loan terms. Additionally, I hypothesize that the opportunity for bank learning increases with the inten-sity of the bank-borrower relationship. Following a new empirical methodology recently intro-duced to finance, I construct a proxy for unobservable latent firm quality. By interacting this proxy with relationship time, I am able to examine how private bank learning impacts loan terms (inter-est rates and loan amount) over time. A significant relationship between the learning term and loan terms demonstrates that banks acquire proprietary information over the course of a bank-borrower relationship, which is then used to set subsequent loan terms. Examining a sample of 489 syndicated loans to 40 UK-based publicly listed companies made be-tween 1988-2004, I find evidence of bank learning and the incorporation of this learned infor-mation into subsequent loan prices. I find that loan interest rate spread is decreasing 0.15 basis points per loan interaction as a result of learning. Additionally, I find that the impact of learning on loan interest rate is more than three times greater in higher intensity relationships, as proxied by the presence of information-intensive acquisition advisory service interactions between the bank and borrower. I do not find loan amount to be adjusted as a result of bank learning, suggest-ing that the borrower rather than the bank determines loan size