Bruno and Marino examine how banks reallocate credit when hidden credit risk is revealed

How banks adjust credit allocation when hidden credit risk is revealed? This is the research question of the paper “Unveiling Risk on Bank Balance Sheets: From Risk Disclosure to Credit Reallocation”, written by Brunella Bruno (Bocconi University and Baffi Centre, in photo) and Immacolata Marino (University of Naples Federico II). 

Methodology

The paper exploits the European Central Bank's 2014 Asset Quality Review (AQR), a unique supervisory exercise that revealed previously hidden credit risk on banks' balance sheets. The AQR generated substantial cross-bank variation in supervisory adjustments while, importantly, these prudential adjustments did not necessarily coincide with accounting recognition in banks' financial statements. This institutional setting allows the authors to distinguish the effects of supervisory risk disclosure from those of balance-sheet adjustments. 

The authors construct a novel bank-level dataset by combining supervisory information from the ECB Comprehensive Assessment with detailed quarterly credit portfolio data from the EBA Transparency Exercises and financial statement information from ORBIS Bank Focus. Using a difference-in-differences approach, they compare banks for which the AQR revealed substantial hidden credit risk (large supervisory adjustments to non-performing exposures, provisions, or both) with reviewed banks for which the supervisory exercise uncovered little or no hidden risk.

Rather than focusing only on aggregate lending, the analysis examines three dimensions of banks' responses: total credit exposures, the allocation of credit across portfolios, and portfolio-level risk-weighted asset (RWA) densities. This framework allows the authors to distinguish between credit reallocation across portfolios and within portfolios, providing a richer picture of banks' de-risking strategies.

Main findings

Banks for which the AQR revealed more hidden credit risk significantly reduce the riskiness of their credit portfolios while leaving overall lending volumes broadly unchanged. Rather than cutting credit, they primarily reallocate lending toward less risky borrowers, especially within existing credit portfolios.

The response differs across regulatory approaches. Banks using Internal Ratings-Based (IRB) models mainly reduce risk through within-portfolio reallocation, particularly in corporate lending, whereas banks using the Standardized Approach (SA) rely more on changes in portfolio composition. Capital constraints strengthen these responses but do not fully explain them.

Finally, the paper identifies two distinct transmission channels. It shows that supervisory risk disclosure itself disciplines banks' risk-taking, independently of whether supervisory adjustments are formally recognized in their financial statements. This finding highlights that supervisory expectations can influence banks' lending behavior beyond the mechanical effects of accounting recognition.