close
Share with your friends

Quarter 1 2020 reporting offers some insight on the impact for banks’ loan books

As most large European banks have now issued their Quarter 1 2020 reports – the first reports to consider the impact of COVID-19 – we can start to build a picture of the impact of the pandemic on the expected credit losses (ECL) in banks’ loan books.

When calculating their ECLs under IFRS 9 Financial Instruments, banks need to take into account their historic experience of losses, updated to reflect current conditions as well as forecasts of future economic conditions. Their Q1 reports offer an early insight into how incorporating forward-looking information about the economic impact of COVID-19 is affecting ECL on their loan books.

We’ve looked at the interim reports of a selection of large banks from eight European countries. The level of detail released by banks varies considerably, with some interim statements providing only a high-level summary. Here, we consider the ECL ratio, the staging of loans and the disclosure of forward-looking information.

The ECL ratio

For the eight banks in our selection that disclosed this information, the average ECL ratio for the loans carried at amortised cost (the ECL as a percentage of the total gross carrying amount) increased from 1.28% to 1.43% between 31 December 2019 and 31 March 20201.

Increase in ECL ratio from 31 December 2019 to 31 March 2020

Enlarge image

Two UK banks also disclosed the ECL ratio separately for retail and wholesale loans.

Increase in ECL ratio – Wholesale and Retail

Staging of loans

A key indicator of changes in the credit quality of a loan book is how much of it has been moved between stages2, as this indicates whether the loan book has undergone a significant increase in credit risk or has become credit-impaired. Seven banks in our selection disclosed the analysis of their loans by stages. The graph below shows how the proportion of loans in Stages 2 and 3 changed from 31 December 2019 to 31 March 2020.

Portions of loan portfolios in Stages 2 and 3

Enlarge image

So, what does this tell us about how the COVID-19 pandemic is affecting banks’ loan books? Most banks in our selection saw some increase in Stage 2 loans. The average share of loans being assigned to Stage 2 increased from 6.77% to 8.27%3. Four of the banks reported an increase in the proportion of Stage 2 loans of between 2.2 pp (percentage points) and 3.8 pp. However, three banks showed a small reduction of below 0.3 pp. There appears to be very little change for the loan book proportions classified as Stage 3, which could reflect the fact that lockdown measures implemented in many jurisdictions have been in place for only a short time.

Forward-looking information

As the IFRS 9 impairment model is forward-looking, banks are required to consider a range of possible future economic scenarios and their probability to calculate ECL. The COVID-19 pandemic has had severe economic impacts across many jurisdictions compared with 31 December 2019. Many governments, central banks and economists have been revising their economic forecasts to try to capture the likely impacts. This means that formulating future economic scenarios at 31 March 2020 has been a particularly challenging task for banks.

We looked at 11 European banks and found that the granularity of information disclosed varied greatly. Some disclosed a lot of detail about the scenarios used, respective probabilities and assumptions about economic variables such as GDP growth or unemployment rates. Three banks disclosed that their base line scenario assumed a rebound in economic activity in the second half of 2020, with one assuming recovery in 2021.

The banks we reviewed appeared to use the same number of economic scenarios at both 31 December 2019 and 31 March 2020. However, most of the banks reported that they have prepared specific COVID-19 scenarios and changed the respective probabilities of their base, upside and downside scenarios. Three of the banks disclosed how they changed the probabilities attached to the scenarios they had used as illustrated below.

Probabilities attached to economic scenarios - Q4 2019 vs Q1 2020

Enlarge image

Existing ECL models will use historical experience to derive links between changes in economic conditions and customer behaviour, and ECL parameters such as loss rates, probabilities of default and loss given default. Therefore, adjustments to model results, based on expert credit judgement, could be necessary to reflect the information available at the reporting date appropriately.

Four of the banks reported that they had used these management overlays on top of the amounts calculated by their ECL models to respond to the economic impacts of the COVID-19 pandemic and low oil prices, while the other seven were silent on this matter. These overlays represented between 4% and 17% of the total ECL balance.

What next?

The Q1 2020 interim reports offer an initial glimpse of the impacts of the COVID-19 pandemic on banks’ ECL at 31 March 2020. However, much has changed since that date and will continue to change, along with expectations about the potential impact of the pandemic and related government support measures. Our next article will look at disclosures that Canadian banks make in their half-year reports for the six months to 30 April 2020.

Director

KPMG IFRG Limited (UK)

Contact

[1] The average ECL ratio is calculated by adding the ECL ratios of all eight selected banks and dividing it by eight. This means that the average does not take into account the different sizes of bank loan portfolios – i.e. all banks are weighted equally.

[2]  Financial assets are grouped into three Stages – Stage 1 for financial assets that have not undergone a significant increase in credit risk (SICR) since initial recognition; Stage 2 when they have experienced a SICR but are not credit-impaired; Stage 3 when they are credit-impaired.

[3]  The average share of loans assigned to Stage 2 is calculated by adding the proportions in Stage 2 and dividing it by the number of banks assessed. This means that the average does not take into account the different sizes of bank loan portfolios – i.e. all banks are weighted equally.