Expected Losses, Unexpected Costs? Evidence from SME Credit Access under IFRS 9

Abstract

This paper examines lending effects of banks switching from an incurred credit loss model to an expected credit loss (ECL) model. I find evidence that ECL transition deteriorates the credit landscape for SMEs as risky, opaque, and bank-dependent borrowers. Post ECL, Affected banks reduce SME lending by 16–20 percent. Banks’ financial reporting concerns and implementation difficulties appear to explain these findings, while regulatory capital adequacy seems less relevant and lenders’ learning does not play a major role. Additional tests performed at the borrower and loan-contract levels indicate rising interest rates and collateral requirements and reducing loan amounts and maturities for SMEs that do business with affected banks. Furthermore, such borrowers could not make up for the lost bank credit from other sources, suggesting real effects. Despite these findings, my inferences do not imply that the ECL paradigm is undesirable for bank shareholders or the overall economy.

Keywords: banking, loan loss provisions, expected losses, IFRS 9, real effects, SMEs, regulation

JEL Classification: G21, G28, G38, M41