European banks and COVID-19 – What do 2020 annual accounts say?

Ewa Bialkowska | 14 April 2021

In our previous blog posts, we reported on the COVID-19 pandemic’s impact on the expected credit losses (ECL) disclosed by a selection of large European banks and five Canadian banks¹ .

Here, we look at what those same European banks have disclosed in their 31 December 2020 (Q4) annual reports. Similar to our previous reporting, the level of detail released by banks varies considerably. This means that the number of banks included in each of the ratios below differs.

Firstly, for 11 banks we compared the total ECL charge in profit or loss² and the profit before tax for the nine months to 30 September 2020 to the corresponding period in 2019. The percentage change in the ECL charge and profit before tax for each of the banks is shown below.

The percentage increase in ECL charge varied considerably – from 32% to 790% – although the bank with the largest percentage appears to have had a very small ECL charge for the year to 31 December 2019. 

Interestingly, the bank with the largest increase in its ECL charge also reported an increase in its profit before tax for the same period. The two banks with the smallest increases in ECL reported some of the largest declines in their profit before tax – similar to what we observed at the half year and Q3 2020.

For all of the banks in our selection, we also calculated the average increases³ in ECL charge for Q1, Q2, Q3 and Q4 2020 from the respective quarters in 2019. This showed a significantly decelerating trend between Q1 and Q3 but a rise again in Q4, as follows:

The chart below shows the loss allowance ratios⁴ for loans carried at amortised cost of eight of the 11 European banks at 31 December 2019 (YE 19), 31 March 2020 (Q1), 30 June 2020 (Q2), 30 September 2020 (Q3) and 31 December 2020 (Q4).

The average loss allowance ratio increased gradually between YE 2019 and Q2 but showed a slightly reducing trend in the second half of 2020. However, with many major European economies entering a second lockdown towards the end of 2020 and into 2021 this reversal may not continue for Q1 2021. The average loss allowance ratio⁵ for loans carried at amortised cost at the end of each quarter was as follows:

Two UK banks also disclosed loss allowance ratios separately for their retail and wholesale loans, as shown below. 

The two banks show different trends in the respective ratios: Bank A has a higher ratio for wholesale than retail, while for Bank B the reverse is true. Bank A showed increases in both ratios between Q3 and Q4 and Bank B a reduction.

Seven of the banks in our selection disclosed information relating to the volume of lending that was subject to payment holidays or other COVID-19 borrower relief programmes as at 31 December 2020. For six of these banks the gross carrying amount of the loans subject to borrower relief programmes represented between 0.2% and 8% of the total gross carrying amount of loans measured at amortised cost. Disclosures made by the seventh bank do not appear to have been made on a comparable basis.

All seven banks disclosed that the number and amount of loans subject to payment holidays or other COVID-19 borrower relief programmes has reduced significantly from previous 2020 quarter ends. 

In our earlier blogs, we looked at an analysis of loans by stages disclosed by seven European banks that had provided this information. The graphs below show, for the end of each quarter, the proportions of loans in Stages 2 and 3 respectively. 

Most of the banks in our selection saw a reduction in the proportion of their loans in Stage 2 at the end of Q4, compared to Q3. The average share of loans in Stage 2 for the seven banks increased from 6.77% at 31 December 2019 to 8.28% at 31 March, to 11.31% at 30 June, but then reduced to 10.34% at 30 September and reduced again to 8.99% at 31 December 2020⁶. The proportion of loans in Stage 3 increased slightly for most banks over Q4.

Previously, we have looked at how banks have tackled the task of assessing the impact of the pandemic on their future economic scenarios when measuring ECL. So, what is the position at 31 December 2020?

Of the 11 European banks, all stated explicitly that they had updated their economic scenarios for COVID-19 impacts in Q4.

Two banks reported using an additional fourth economic scenario at 31 December 2020 compared to the three scenarios they used at 31 December 2019. Meanwhile, one bank attributed a 0% probability to their upside scenario throughout 2020, in effect reducing its number of scenarios to two.

In previous quarters, the majority of banks changed the probabilities attributed to their economic scenarios in response to the economic impacts of the pandemic. For Q4, five banks explicitly reported applying a higher weight to probabilities for their base and downside scenarios compared to Q3. Two banks disclosed that they have applied a higher weight to their upside scenarios; the other four banks remained silent on this matter.

Three banks disclosed the specific probability weightings applied to their macroeconomic scenarios at both Q3 and Q4. The graph below shows the detail:

The majority of the banks disclosed the key inputs into their macroeconomic scenarios relating to the markets in which they operate. However, as they operate across a variety of different markets, we could not compare those inputs across different banks.

Eight banks reported the use of management overlays on top of the amounts calculated by their ECL models to respond to the prevailing uncertainties linked to the economic impacts of the pandemic and government support measures. Out of these, six specified the amount of the adjustments, with some giving more numerical analysis – e.g. per book or geography. The granularity and nature of disclosure varies, which makes a meaningful comparison difficult.  

In our blogs, we aim to share with you the emerging information on how banks have approached estimating ECL in this unprecedented environment. We will continue to monitor this information as it become available.