Much is still unknown about the impact of the COVID-19 pandemic on European banks' asset quality. Governments have provided a significant amount of financial resources to counteract the economic impact of the pandemic. In addition, debt payment moratoria and other borrower-friendly measures have helped lessen the impact of lockdowns and other measures on household and corporate borrowers. These measures therefore also support loan asset quality. Therefore, it's no surprise that banks' half-year reports do not yet reveal the full picture, and their third-quarter reports are unlikely to do so either.

S&P Global Ratings believes that, as support and forbearance measures are unwound, the true impact on asset quality will emerge. Such measures, while aimed at helping borrowers through a period of liquidity constraints, may also mask underlying borrower distress. For example, a relaxation of corporate insolvency laws in some countries, such as Germany and the U.K., to allow companies to continue trading through pandemic-related uncertainties is likely to--at least temporarily--limit the number of company insolvencies that would otherwise occur.

The Extent Of Fiscal Stimulus Will Influence The Outcome For Banks

We expect that differences in the size of authorities' fiscal response to the pandemic (see chart 1), and how quickly it is withdrawn, will lead to wide disparities in the amount of banks' credit losses by country. For example, Germany's fiscal package of over €1 trillion, and the recent extension of its furlough scheme to the end of 2021, will limit domestic credit losses for the banking system, but also potentially delay the emergence of some losses.

Chart 1

Across much of Europe, national moratoria schemes were initially introduced for a three-month period, with many extended for a further three months. Now, the situation is more varied, with some countries, such as the U.K., having closed their schemes and others considering extensions. For example, in Portugal, the government has recently extended--for the second time--its public moratoria scheme by six months to September 2021. In Austria, the authorities are considering an extension to the end of this year. Differences such as these in moratoria schemes across Europe will also mean disparate outcomes for asset quality.

What's becoming clearer is that the pandemic will hit some corporate sectors harder than others. In Europe, we see the greatest impact on corporate entities we rate in sectors such as commercial aerospace, non-essential retail, leisure, and restaurants, where we expect a recovery of credit metrics only after 2023. In contrast, we foresee sectors such as essential retail and pharmaceuticals remaining fairly resilient to the pandemic-related downturn (see "The Shape Of Recovery: Uneven, Unequal, Uncharted," published July 1, 2020, on RatingsDirect).

To the extent that banks have significant corporate exposures on their loan books, their credit losses will vary depending on the share of borrowers in more troubled sectors versus stronger sectors. We also expect that small and midsize enterprises will be relatively worse affected than larger corporates, and retail lending--particularly mortgages--to be the least affected, although we still expect an uptick in losses across all exposures.

Asset Quality Metrics Vary Widely

We estimate that gross customer loans rose by slightly less than 3% in the first half of 2020, as subdued loan demand from households was offset by substantially increased demand from corporates. The migration of loans across IFRS 9 stages varied noticeably across banks, in particular for stage 2 loans (see charts 2 and 3). In our view, owing to the extent of judgement in banks' assessment of future loan performance, not all of this variability in stage migration is attributable to underlying asset quality or differences in the composition of banks' loan portfolios.

Chart 2

Chart 3

To date, there has been limited evidence of significantly higher delinquencies on European banks' loan books, since fiscal support programs and debt payment moratoria are helping borrowers bridge short-term liquidity shortfalls. As such, few loans have moved to stage 3 (see chart 4). That said, we consider that delinquencies will rise over the rest of this year and next because some corporate borrowers may not survive the economic slowdown and some industries will recover slower than others.

Chart 4

We expect European banking systems will see a spike in the cost of risk for domestic loan exposures to 63 basis points (bps) in 2020 from just 28 bps in 2019, before it declines somewhat to 51 bps in 2021. What's more, we see significant uncertainty regarding the amount and timing of these base-case estimates, particularly if the economic recovery we forecast for Europe in 2021 is delayed or weaker than expected. Using data from the first-half financial reports of the top 50 European banks, we estimate that the average cost of risk for customer loans rose to 107 bps (on an annualized basis) in the first six months of this year, from 37 bps at year-end 2019, including a large number of deviations (see chart 5). We may see some narrowing of that wide range of outcomes in banks' full-year results, particularly if uncertainty about Europe's macroeconomic prospects reduces.

Chart 5

In most cases, we expect these higher costs will be absorbed by preprovision earnings and thus we don't expect regulatory capital levels will be significantly eroded. Still, the sharp rise in credit losses, and more muted earnings resulting from slower economic growth and lower interest rates, leave banks' limited earnings headroom to fully absorb further losses.

The Pandemic Is The First Test Of IFRS 9's Expected Credit Loss Approach

IFRS 9, which came into effect in 2018, introduced a more forward-looking expected credit loss (ECL) model for provisioning. The greater influence of future expectations has inevitably required increased precision and judgment from banks' management teams in estimating credit losses (see chart 6).

For European banks, we see four key factors in IFRS 9's ECL model affecting their provisions:

Macroeconomic scenarios and weighting:  

• The application of multiple economic scenarios, and the weightings applied to them, can have a significant influence at key turning points of the economic cycle, particularly when the change is sharp and unexpected. Few banks have explained their non-base-case scenarios or the weightings applied, so it is difficult to make comparisons. Still, in our view there has likely been an abrupt shift in macroeconomic scenarios that most, if not all, banks use, given the shock resulting from the pandemic. We expect this is a key factor in their provisioning estimates this year. Barring further shocks, we expect macroeconomic forecasts will be less volatile and less of an influence on ECL movements from next year. It's also possible that banks using more negative macroeconomic forecasts in their models, or greater weights, may see reversals in their ECL estimates in future reporting periods.

Stage transfers:  

• Management judgments or model assumptions about when a loan moves from stage 1 (which requires a 12-month expected loss provision) to stage 2 (which requires a lifetime expected loss provision) depend on assessments of whether there has been a significant increase in credit risk (SICR) since the origination of the loan. While IFRS 9 rules set out some indicators for management to consider SICR, no quantitative metrics are mandated as the standard. We believe this is likely leading to inconsistent application of the rules in practice. For example, for NatWest, the share of stage 2 loans jumped to 26.3% of total loans as of June 30, 2020, from 8.1% at year-end 2019 (see chart 2). This was primarily because NatWest assesses a 10-bp increase in the probability of default for its wholesale exposures as an SICR indicator. While this might indicate a more conservative approach than peers, it's difficult to confirm due to gaps in comparable SICR disclosure across banks. Nevertheless, it is helpful that many banks apply a 30-days-past-due backstop for assessing SICR; in other words, loans more than 30 days overdue are moved to stage 2 from stage 1. We expect stage transfers will be the main driver of ECL movements through 2021. In particular, as fiscal support measures start to wind down, we expect that increased clarity about which borrowers are having financial difficulties will result in a larger transfer of loans to stage 3 than has occurred so far this year (see chart 4). We expect provisions for stage 2 loans may rise for similar reasons.

Model assumptions and management overlays:  

• ECL models incorporate banks' historical experience and data to estimate parameters such as probability of default, loss-given default, and loss rates. The sudden and unprecedented nature of the pandemic-related economic stop across much of Europe likely meant, in many cases, the application of management overlays as an add-on to the ECL estimate derived from ECL models. Overlays may also be applied to reduce the model-derived ECL estimate, for example to incorporate the effect of furlough schemes or other fiscal support measures. A lack of disclosure means that it is not always clear which banks have applied management overlays, their effect (that is, the amount of the ECL top-up or deduction), or factors considered in estimating the overlay amounts.

Regulatory guidance and treatment of loan forbearance:  

• Regulators have encouraged banks to ensure that their ECL models incorporate the extent of fiscal support from governments and longer-term economic forecasts for periods beyond the abrupt sharp downturn arising from the pandemic. Both the IFRS standard setter and regulatory authorities across Europe have highlighted the importance of not automatically treating pandemic-related forbearance as an indication of asset-quality deterioration that requires the transfer of loans to stage 2 or 3 from stage 1, particularly where the forbearance is available to all borrowers on request. We see this as an appropriate approach that can avoid undue volatility in reported ECLs, once the assumption of a temporary shock and rapid recovery materializes. We think it's unlikely that loans subject to extended forbearance measures--such as ongoing or additional support to individual borrowers still in difficulty after the debt moratorium ends--would remain classified as stage 1 loans. That said, the greater the extent of forbearance applied to a bank's loan book, the greater the uncertainty about the loan book's future performance.

Banks Should Learn From The Last Financial Crisis

The 2008 global financial crisis serves as a reminder to banks and regulators that transparency about banks' balance sheets and performance is critical to addressing investor concerns about viability. Disclosure initiatives such as the Enhanced Disclosure Task Force (EDTF) helped address some of the shortfalls in transparency after that crisis, particularly where national regulators proactively encourage banks to apply the EDTF's disclosure recommendations. Enhancements in the timeliness and comparability of banks' Pillar 3 disclosures have also enhanced transparency.

While the current pandemic has not led to a full-blown financial crisis yet, debt and equity investors recognize asset quality to be a critical element in their assessment of banks. Gaps in disclosure could further undermine investors' interest in banks, which are by nature confidence-sensitive institutions. The European Banking Authority has set out disclosure templates for banks, intended to show the extent of loans under forbearance and new loans granted under guarantee schemes, which should aid investors' understanding of the risk.

Still, because banks' asset quality metrics lack comparability and management's assumptions are critical to provisioning, we assume that bank management teams eager to retain investors' support will seek to go well beyond the minimum. Timely and informative disclosures will help the market understand the basis for reported metrics, ideally including also sensitivity analysis under a reasonable adverse case.

Uncertainty Still Clouds The Asset Quality Landscape

More than six months after the new coronavirus started spreading across Europe, it is still unclear how the economic fallout will ultimately affect banks' asset quality. We expect further insights through banks' annual and quarterly reporting, particularly as the cyclical economic effects become clearer. A further complication, although positive, relates to loans granted by banks under government guarantee schemes established in response to the pandemic. Although the umbrella programs are quite large, the sum of loans granted to date has been fairly small, though not insignificant (see chart 7). Ultimately, guarantees will only limit some of the loan losses banks will face, and not cover all of them.

Chart 7

The longer-term implications of the pandemic are likely to arise from secular changes in addition to cyclical ones. Consumer behavior and preference, corporate investment decisions, savings propensity, and many other factors may be permanently altered. We anticipate that banks' previous stance about the relative riskiness of certain lending exposures may also change accordingly. Yet how the pandemic will evolve is unknown, and thus the consequences for government policy, economic activity, and asset quality could change in unexpected ways.

Related Research

• Managing Through The Crisis, Europe’s Banks Look To The Future, Sept. 28, 2020
• Global Banking: Recovery Will Stretch To 2023 And Beyond, Sept. 23, 2020
• The $2 Trillion Question: What’s On The Horizon For Bank Credit Losses, July 9, 2020
• The Shape Of Recovery: Uneven, Unequal, Uncharted, July 1, 2020
• The Adoption Of IFRS 9 And Bank Ratings, Feb. 19, 2018