EBA Note Compares Provisioning Practices in US and EU Amid Pandemic

Looking at data over the past 13 years, following an economic shock, loan-loss provisions of EU banks tend to be less volatile than those of US banks. In a similar vein, in the first two quarters of 2020, the cost of risk of US banks was much higher compared to EU banks. However, in the second half of 2020, the cost of risk of US banks fell more rapidly compared to their EU peers. The impact of the pandemic on macroeconomic variables helps explain some of the differences in the cost of risk. The US suffered a higher increase in unemployment in the early stages of the pandemic that might contribute to the sharper rise in the cost of risk compared to EU banks. Similarly, a faster economic recovery in the US might explain the faster fall in the second half of 2020.

A preliminary analysis also reveals a riskier loan portfolio composition of US banks. The share of the portfolios potentially more affected by social distancing and containment measures such as commercial real estate or consumer credit over total loans granted is higher in the US. This could be a further explanation for the higher cost of risk at the onset of the pandemic. Another reason for variation in the cost of risk could be due to the different accounting rules. Under current expected credit loss (CECL) model, banks recognize lifetime ECL for all financial assets whereas under IFRS 9 the 12-month ECL is recognized for Stage 1 loans. At the onset of a crisis, the IFRS 9 impairment model presumably resulted in a rise in the cost of risk because of loan migrations from Stage 1 to Stages 2 or 3, for which lifetime ECL were recognized. However, this effect seems to be less material than the impact of applying the CECL approach to all financial assets.

Supervisory guidance might have also played an important role in the observed variations. EBA guidance issued in the context of the pandemic might have softened potential cliff effects by highlighting that the application of general public or private moratoria meeting specific requirements should not be considered as an automatic trigger to conclude that a significant increase in credit risk has occurred. EBA also recommended that when determining if a loan should be classified as non‐performing, banks should distinguish between obligors whose credit standing would not be significantly affected by COVID-19 in the long term and those that would be unlikely to restore their creditworthiness. Against this backdrop, in the first half of 2020, Stage 1 loans of EU banks increased by only 0.4% while Stage 2 loans rose by 29.1%. In the same period, Stage 3 loans fell by 0.1%. Similarly, US authorities confirmed that short‐term modifications for borrowers whose loans were performing prior to any relief would not be considered troubled debt restructurings. The US authorities also stipulated that banks were not expected to designate loans with deferrals granted due to pandemic as past-due because of the deferral. Nonetheless, since CECL does not envisage different stages for loan-loss recognition, the overall impact of this guidance might have been more limited in US than in EU.

For the EU, a sample of 160 European banks (unconsolidated number of banks, including 30 subsidiaries) as of December 31, 2020 has been used. This sample is reviewed annually by competent authorities and adjusted accordingly. For the US, the data is taken from the Quarterly Trends for Consolidated U.S. Banking Organizations of the Federal Reserve of New York (NY FED), which uses consolidated financial data across all reporting US parent bank holding companies and intermediate holding companies, and individual banks not controlled by a bank holding company or whose parent bank holding company does not report on a consolidated basis. The data excludes savings bank holding companies and branches and agencies of foreign banks. To ensure comparability between EU and US data, the cost of risk figures provided in this note do not include the amounts written off directly to the statement of profit or loss. Thus, they could differ from the cost of risk figures provided in other EBA publications such as the Risk Dashboard or the Risk Assessment Report.

Related Links

• Press Release
• Thematic Note (PDF)

Keywords: Europe, Americas, EU, US, Banking, CECL, ECL, IFRS 9, COVID-19, Credit Risk, CRR, Loan Loss Provisions, Cost of Risk, EBA