A Quantitative Theory of Relationship Lending

41 Pages Posted: 27 Sep 2022

See all articles by Kyle Dempsey

Kyle Dempsey

Ohio State University

Miguel Faria-e-Castro

Federal Reserve Bank of St. Louis

Date Written: September, 2022

Abstract

Banks' loan pricing decisions reflect the fact that borrowers tend to have long-lasting relationships with lenders. Therefore, pricing decisions have an inherently dynamic component: high interest rates may yield higher static profits per loan, but in the long run they erode a banks' customer base and reduce future profitability. We study this tradeoff using a dynamic banking model which embeds lending relationships using deep habits (“customer capital”) and costs of adjusting loan portfolio composition. High customer capital raises the level and decreases the interest rate elasticity of loan demand. When faced with an adverse shock to net worth, banks with high customer capital recapitalize quickly by charging high interest rates and eroding customer capital in the short term, while banks with low customer capital face persistent financial distress. Using Call Report data to measure the franchise value of banks' loan portfolios, we find that this effect has strong implications for how individual banks and the financial sector as a whole recover from shocks.

Keywords: banks, Customer Capital, relationship lending, interest rates, financial crises

JEL Classification: E4, G2

Suggested Citation

Dempsey, Kyle and Faria-e-Castro, Miguel, A Quantitative Theory of Relationship Lending (September, 2022). FRB St. Louis Working Paper No. 2022-33, Available at SSRN: https://ssrn.com/abstract=4230688 or http://dx.doi.org/10.20955/wp.2022.033

Kyle Dempsey (Contact Author)

Ohio State University ( email )

United States

Miguel Faria-e-Castro

Federal Reserve Bank of St. Louis ( email )

411 Locust St
Saint Louis, MO 63011
United States

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
61
Abstract Views
214
Rank
637,858
PlumX Metrics