The Allocation of Public Guaranteed Loans to Firms During Covid-19: Credit Risk and Relationship Lending

Abstract

Using loan-by-loan data matched with supervisory and borrower balance sheet data, we investigate whether public guarantees supported firms that were already weaker before the pandemic shock or, conversely, borrowers that were low risk but needed liquidity to weather economic uncertainty. We find that government-guaranteed loans were more likely to be granted to borrowers who were safer, more liquidity constrained and for whom the granting bank was a significant lender. The availability of soft information on the borrower was not an important driver of allocation, consistent with the purpose of guarantees, which is to mitigate asymmetric information problems. Evidence from ex-post default data one year later shows that borrowers with guaranteed loans were significantly less likely to have repayment problems than those with no guarantee, holding constant the observable ex-ante risk. An asymmetric information test based on Chiappori & Salanie (2000) rejects the hypothesis that the allocation of guarantees was affected by large-scale adverse selection. 

Keywords: relationship lending, bank credit, loan guarantees, COVID-19, credit risk, asymmetric information

JEL Classification: G18, E63, H12, H81, G21