« Back to Results

Bank Lending

Paper Session

Saturday, Jan. 4, 2025 12:30 PM - 2:15 PM (PST)

San Francisco Marriott Marquis, Juniper
Hosted By: International Banking, Economics, and Finance Association
  • Chair: Diana Bonfim, Banco de Portugal and Católica Lisbon

Are Collateral and Lender Screening Efforts Substitutes?

Dong Beom Choi
,
Seoul National University
Jung-Eun Kim
,
Federal Reserve Bank of Richmond

Abstract

This study investigates whether lenders actively reduce their screening efforts when offered better collateral. Using loan-level data on underwriting timelines, we find that lenders approve mortgage applications more quickly when local house prices rise, indicating less thorough screening. This effect is more pronounced when lenders face greater frictions that collateral could help alleviate, such as limited access to soft information, fewer opportunities for loan sales, higher credit risk, and constrained operating capacities. Non-bank lenders, particularly fintech firms, demonstrate higher sensitivity than banks, due to their heavier (less) reliance on hard (soft) information. These results suggest an increased level of procyclicality in aggregate lending with the growing presence of non-bank fintech lenders.

Joining Forces: Why Banks Syndicate Credit

Steven Ongena
,
University of Zurich
Alex Osberghaus
,
University of Zurich
Glenn Schepens
,
European Central Bank

Abstract

Banks can grant loans to firms bilaterally or in syndicates. We study this choice by combining bilateral loan data with syndicated loan data. We show that loan size alone does not adequately explain syndication. Instead, banks' ability to manage risks and firm riskiness drive the choice to syndicate. Banks are more likely to syndicate loans if their risk-bearing capacity is low and if screening and monitoring come at a high cost. Syndicated loans are more expensive and more sensitive to loan risk than bilateral loans. Our findings contradict the hypothesis that reputable borrowers graduate to the syndicated loan market.

CovenantAI - New Insights into Covenant Violations

Vanessa S. Krockenberger
,
Frankfurt School of Finance and Management
Anthony Saunders
,
New York University-Stern
Sascha Steffen
,
Frankfurt School of Finance & Management
Paulina M. Verhoff
,
Frankfurt School of Finance & Management

Abstract

This paper introduces CovenantAI, a novel artificial intelligence (AI)-powered tool that tracks SEC-reported covenant violations with improved accuracy over existing text-search methods, covering data from 1996 to 2022. It accurately identifies amendments, waivers, and technical defaults, providing a detailed timeline of covenant breaches. We use a “quasi” regression discontinuity approach to analyze the effects of these violations on key firm outcomes such as investments, employment, or credit access, revealing complex patterns and pronounced effects during economic downturns such as the COVID-19 pandemic. The changing loan market, resembling bond markets with the rise of CLOs and secondary trades, has decreased covenant reliance and violations among non-investment grade firms.

Discussant(s)
Andrei Zlate
,
Federal Reserve Board
Margherita Bottero
,
Bank of Italy
Amanda Heitz
,
Tulane University
JEL Classifications
  • G2 - Financial Institutions and Services
  • G0 - General