As the COVID-19 pandemic shows early signs of coming under control across large parts of the world, the policy responses to its economic consequences are starting to unwind, so far in an orderly manner. In turn, the effects of the pandemic on banks' asset quality are becoming a little less uncertain and a little less negative compared to our February 2021 forecasts.

Across the 86 banking systems S&P Global Ratings covers, we now expect that credit losses will amount to around $1.6 trillion over the two years to end-2022 (see chart 1). This is around 8% lower than we had previously forecast (see chart 2, and for more details, see "Lower And Later: The Shifting Horizon For Bank Credit Losses," published Feb. 2, 2021). Moreover, the latest data for 2020 show that bank credit losses amounted to around $820 billion, some 8% lower than our previous forecast of $892 billion.

Still, our forecast for credit losses across the three years to end-2022 amounts to around $2.5 trillion in aggregate--more than 1.5 times the run rate of losses in 2019. We expect that 2019 marked the end of a multiyear period of benign credit losses for banks globally, even as economies continue to recover from the pandemic.

Chart 1

Chart 2

We use the term credit losses to refer to the income statement charge (in U.S. dollars) by which banks add to their balance sheet provisions or allowances for expected losses on domestic customer loans, plus any direct write-offs of domestic customer loans. Banks often describe credit losses as "provisions for" or "charges for expected credit losses," among other similar terms. Credit losses generally precede charge-offs, the actual write-down of loans that detract from the balance sheet allowances for credit losses. The cost of risk (or credit cost ratio) refers to credit losses as a proportion of customer loans.

Our forecasts are based on our analysts' estimates for nonperforming loans (NPLs) and related provisioning. Regulatory forbearance, jurisdiction-specific rules, or bank management decisions regarding slowing the pace of loss recognition, will affect the timing and recognition of credit losses in banks' financial reporting. In our view, and as the aftermath of the 2008-2009 global financial crisis (GFC) showed, delays in the recognition of credit losses by banks, or a lack of transparency in reporting such losses, could yet undermine investor confidence in banks and may delay the path to recovery for some countries.

The Global Credit Cost Ratio Rose Sharply In 2020, And Will Remain Elevated Over 2021-2022

We expect that on a global scale, banks' credit cost ratios averaged around 97 basis points (bps) in 2020, and will slip a little to around 85-90 bps in 2021 and 2022. This ratio is well above the 2019 average of around 70 bps (see chart 3). We estimate that the ratio was around 100-120 bps in the aftermath of the 2008-2009 GFC. That said, a comparison with credit losses in the GFC is not a like-for-like assessment. This is because of factors such as current accounting rules, which require a more timely recognition of credit losses than the rules in place during the GFC, as well as the composition of global lending, which is now more weighted toward developing-market economies (including China) that tend to have weaker asset quality. The nature of the support mechanisms is also very different from in the GFC, as governments have moved rapidly to support households and businesses rather than needing to directly support banks. The GFC also had a more limited effect on loan asset quality in some regions, including Asia-Pacific, for example, than we expect to be the case now.

Our projections for credit losses vary widely among regions (see charts 3 and 4). Globally, we expect that the total increase in credit losses from their level in 2019 will amount to around $870 billion over the three years to end-2022. (By increase, we mean the sum of the differences in projected credit losses for each of 2020, 2021, and 2022 compared to the actual credit losses for 2019.)

Chart 3

Chart 4

China alone accounts for over one-half of this increase. This largely reflects the sheer size of the Chinese banking system in a global context (see charts 5 and 6) and stringent provision-coverage regulations for NPLs. In terms of customer loans, the Chinese banking system is approximately the same size as the U.S., Japanese, German, and U.K. banking systems combined. Moreover, the banking system in China is much more important in supplying credit to its economy than in the U.S., where borrowers typically benefit from a deep, liquid, and mature bond market, a large nonbank financial institution sector, as well as access to funding via the banking system.

The COVID-19 pandemic is an important component in our estimates of asset quality for Chinese commercial banks, but other factors are also at play. We expect reported NPL ratios to continue to climb at a moderate pace, with certain firms failing to keep up with the fast pace of China's economic transition. Default rates will likely increase, including among state-owned enterprises, and we anticipate high formation rates of new NPLs, rising credit costs, and growth of loan write-offs exceeding the pace of loan growth over the next two-to-three years. Rigidity in provision-coverage requirements elevate credit losses, notwithstanding the fact that the health and economic impact of the pandemic is less in China than in many other jurisdictions. The Chinese economy maintained positive (albeit lower) economic growth in 2020 compared to 2019, whereas all other major economies contracted.

Western Europe accounts for a further $161 billion of the increase in credit losses, even as the region's economies rebound. This reflects the more gradual recognition of credit losses in the region relative to the U.S., as well as the emergence of problem loans as borrower-friendly support schemes unwind. In contrast, we now expect that the run rate of credit losses in North America over the three years to 2022 will only be around $40 billion higher in aggregate than losses in 2019. Indeed, we forecast zero credit losses for the U.S banking system in 2021. U.S. banks' first-quarter 2021 results showed provision reversals of $14.5 billion, and as the economic recovery strengthens, it's possible that these reversals will continue resulting in negative losses for the year.

Major Banks Are In A Good Position To Absorb Credit Losses From Earnings

We estimate that the top 200 rated banks represent about two-thirds of global bank lending. For these banks, we estimate that credit losses absorbed about 40% of their 2020 pre-provision earnings, and will absorb a little less, around 37%, in 2021 and 2022. The ratio was under 30% in 2019. We expect the earnings of the top 200 rated banks to recover further in 2021 and 2022, fueled in part by lending dynamics across much of the world. This is particularly the case in China, where we expect lending growth to average around 14% per year over this period (see charts 5 and 6).

However, in Europe, despite a general pick-up in mortgage activity and some recovery in consumer credit, we expect that overall loan growth will be more muted in the period to end-2022. We expect European loan growth to average around 2%, weighted more toward 2022, after very strong growth in 2020. This primarily reflects the higher level of household savings built up during the pandemic, as well corporates taking up loans on a precautionary basis in 2020, reducing their demand for credit now.

Chart 5

Chart 6

Still, the pandemic is not over yet, and of course the path to recovery could be worse or take longer than we assume in our base case (see table 1). Such setbacks would likely lead to higher credit losses as more borrowers become distressed, as well as lower pre-provision earnings from reduced economic activity, and potentially even lower interest rates. This combination of events would inevitably affect many banks across the world. Even if the spread of the virus were to end tomorrow, the effects could linger, especially if social distancing becomes the norm, or business and consumer spending doesn't bounce back as people remain cautious in the face of uncertainty.

Table 1

The Unwinding Of Fiscal Support And Moratoria Schemes May Reveal Fragilities In Banks' Asset Quality

The unprecedented level of fiscal support that many governments across the world deployed in response to the pandemic proved to be a key factor in supporting households and businesses, particularly during lockdowns. Many banks introduced payment moratoria or other forms of borrower forbearance as part of their response to the effects of the pandemic. Both types of support helped stabilize borrowers' creditworthiness during the pandemic, and prevented borrowers dealing with liquidity shortfalls from becoming insolvent and defaulting on repayments. As these measures unwind, any underlying but previously hidden fragilities in banks' asset quality will become more apparent. That said, so far, we have not seen any evidence of significant new weaknesses in asset quality, and our base case is that in aggregate, banks will be able to comfortably absorb the losses that will arise.

What's On The Horizon For Bank Credit Losses?

Below, we outline our credit loss forecasts for regional banking systems across the world to end-2022.

Asia-Pacific: Outside China, credit losses will decline in most jurisdictions over the next two years

We expect credit losses to be lower than our previous forecasts for most Asia-Pacific countries, as prospects for the region's economic rebound will support banks and their borrowers. Our revised forecast for credit losses for Asia-Pacific banking systems from 2020 to 2022 is $1.53 trillion, down slightly from our February 2021 forecast of $1.57 trillion. We also expect credit losses as a percentage of loans to decrease in nearly all Asia-Pacific countries over the next two years. Targeted assistance to stretched borrowers will likely continue in many places until pandemic-related challenges abate substantially. Given its size, China continues to account for more than one-half of the region's loans and more than three-quarters of its credit losses.

In China's case, we anticipate that credit losses will rise. The effect of COVID-19 on credit costs in the country may extend over several years. This is notwithstanding the fact that the health and economic impact of COVID-19 is less in China than in many other jurisdictions. The Chinese economy maintained positive (albeit lower) economic growth in 2020 compared to 2019, whereas all other major economies contracted.

While we expect the trajectory of credit losses in Asia-Pacific as a whole to be lower in 2021 and 2022 than in 2020, most banking jurisdictions will still find it challenging to lower credit losses to 2019 pre-pandemic levels by end-2022. We expect that China, South Korea, Singapore, Hong Kong, and Australia may be among the first major banking jurisdictions in Asia-Pacific to recover to 2019 financial strength, but this is not until the end of 2022. China's recovery is notwithstanding rigidity in provision-coverage requirements and associated credit losses compared to some other jurisdictions.

While the Japanese banking system's profitability is low by Asia-Pacific standards, it has contended well with COVID-19 challenges. Our most recent credit loss estimates for Japan to 2022 are much lower than we expected this time last year.

For India and Indonesia, the path to recovery from the pandemic will be more difficult. India has endured a steep recession, and while our expectation is for a strong economic rebound, India had very high NPLs leading into the pandemic. This contrasts with most other Asia-Pacific banking jurisdictions, where NPLs were at historic lows when the pandemic struck. Indonesia is among the countries hardest hit by the pandemic in Southeast Asia. Regulatory forbearance that allowed restructured loans to be classified as performing until end-March 2022 has mitigated damage to Indonesian banks' financials. We believe that the underlying deterioration in asset quality could become more apparent after this classification expires.

North America: Easing pressure on asset quality will likely allow for a sharp drop in credit losses and shrinking reserves in 2021

Pressure on asset quality has eased substantially in North America thanks to robust economic growth spurred in part by significant government stimulus measures and an accommodative monetary policy. This should allow banks in the U.S. to reduce the high reserves they built in 2020, likely through minimal or perhaps even negative provisions for credit losses. In Canada, we also expect provisions for credit losses to fall significantly compared to 2020 on improving conditions.

In the U.S., the 2020 implementation of the current expected credit losses accounting standard--which requires banks to set reserves for the expected lifetime losses on their entire loan portfolio--forced banks to build reserves last year faster than they would have under the prior standard. With reduced expectations for lifetime losses, banks in the U.S. are likely to reverse at least some of that build-up.

In fact, in the first quarter of 2021, U.S. banks reported provisions for credit losses of negative $14.5 billion, equating to negative 54 bps for average loans, and early second-quarter earnings reports have shown further negative provisions. The ratio of reserves to loans for the U.S. banking system fell to about 2.0% from 2.2%, but is still well above the pre-pandemic level. In our base case, we expect provisions for credit losses to be around zero for the full year, but see a reasonable possibility of negative credit losses for the year.

Western Europe: An orderly recovery means that credit losses will decline even as forbearance and fiscal support measures wind down, but will remain higher than before the pandemic

We have revised down our forecasts for European bank credit losses in 2021 and 2022 by around 6%. We estimate that their credit losses peaked at around $120 billion in 2020, and will decline by a little under 10% this year to around $110 billion. We expect losses to fall faster in 2022, declining by a further 20% to around $88 billion. This still represents an elevated amount of losses compared with pre-pandemic losses of just $53 billion in 2019. In part, this reflects the more gradual recognition of losses in European banks compared to other regions such as the U.S., as well as a historically low level of credit losses in 2019. That said, we expect that some of the region's banking systems, including those in Sweden, Norway, and Finland, will return closer to pre-pandemic loss levels by the end of next year.

While we have yet to see the full impact of the pandemic on European banks' asset quality, in our view, the proven effectiveness of government support measures and confidence in the strength of the economic rebound now underway have reduced the chances of severe asset-quality deterioration. Indeed, the gradual end to debt moratoria across European jurisdictions has not resulted in a meaningful hike in delinquencies so far. While some moratoria remain in place, most have expired, and we see most borrowers resuming normal repayments.

Though problem loans are yet to rise--particularly in banks' corporate loan books--better prospects should support borrowers' creditworthiness and result in a manageable, rather than significant, increase in nonperforming assets at most banks. Still, we expect that asset-quality performance will vary across banks and countries, with problems likely to be concentrated in the sectors hardest hit by social-distancing measures, namely, hospitality, lodging, transportation, entertainment, and retail.

We have seen some U.K. banks release provisions in the first quarter of this year, reflecting the fact that these banks had set aside higher provisions than other European banks in 2020. In contrast, most other European banks opted for a more gradual recognition of credit provisions in 2020, so we don't expect to see similar provision releases for them this year. In addition, the economic rebound in the U.K. is likely to prove stronger and come earlier than in Continental Europe, providing further support to asset quality.

Latin America: Credit losses should decline thanks to robust provisioning prior to the pandemic and higher provisions in 2020

We forecast credit losses of $102 billion to end-2022 for banking systems in Latin America, with the bulk of these in the region's largest economies, Brazil ($46 billion) and Mexico ($17 billion). We expect credit losses for the region to be around $54 billion in 2021, and to recover to 2019 levels by end-2022, mainly due to the Brazilian financial system's recovery. Additional factors that will contribute to this recovery include the region's higher provisioning coverage prior to the pandemic; a robust rise in additional provisions in 2020, helped by healthy margins on loans that ease the pain of earlier loss recognition; and the short-term nature of loans in the region. Moreover, the loan portfolio mix has become less risky over the past few years, and as a result, we expect loan losses as a percentage of loans to be lower than historical levels.

As economies continue to rebound, we expect businesses and individuals' debt-payment capacity to improve. On the other hand, although asset quality has been better than we expected so far, it's likely to worsen as loan-deferral programs and government-support measures are phased out. However, the sharp rise in provisioning last year puts banks in a good position to deal with the possible deterioration. A key risk for the region's banks remains the high levels of informality in the economy and fragile conditions for small and midsize enterprises (SMEs). The pandemic has hit the middle- and lower-income households the hardest, and increased social and political tensions will generate a high degree of uncertainty about government policy in the post-pandemic years. This could further dampen investment and lower long-term GDP growth.

Central and Eastern Europe, the Middle East, And Africa: Banks are still navigating their way through the pandemic

We forecast credit losses of around $150 billion to end-2022 for banking systems in Central and Eastern Europe and the Middle East and Africa, similar our February 2021 forecast.

Turkey.  We expect operating conditions for Turkish banks to remain weak, and lending growth to slow to an average of 15% in 2021, compared with 35% in 2020. In 2020, lending increased due to heavy stimulus to households and SMEs and the depreciation of the Turkish lira, which inflated borrowing in foreign currencies. We expect that asset-quality indicators will continue to worsen once the government's support and regulatory forbearance measures are gradually lifted. On Sept. 30, 2021, the Turkish regulator, BRSA, extended the forbearance measures regarding capital-ratio calculations and NPL and Stage 2 loan classifications. In addition, we estimate that about 7%-9% of loans still benefit from moratoria.

Banks with exposure to the construction and energy sectors, commercial real estate, SMEs, and more generally, to sectors badly affected by the pandemic, like tourism and transportation, will contribute to the deterioration in asset quality. Under our base-case scenario, we expect NPLs to reach around 11% of total loans by the end of 2022, and problematic loans (NPLs plus restructured loans) to exceed 20%. We also expect the cost of risk to rise to 320 bps on average in 2021 and 2022, from an already-high 290 bps on average in 2019-2020.

South Africa.  For South African banks, we forecast that the growth of credit to the private sector will be subdued in 2021. We estimate that credit losses are likely to start decreasing, to 1.7% in 2021 from 2.1% in 2020, while NPLs will likely increase to 5% of total loans over this period. Despite a sharp drop in earnings, the major banks remained profitable in 2020, and their capital ratios remained resilient. We expect this trend to continue in 2021. Although top-tier banks are exposed to wholesale short-term deposits, these largely stem from domestic nonbank financial institutions. Banks are therefore not exposed to large-scale refinancing risk or a reversal of investor sentiment because they do not rely on international funding.

Saudi Arabia.  The Saudi economy will recover in 2021-2022 from the shocks of 2020 as global demand for oil recovers and private consumption increases. Real GDP is likely to return to 2019 levels in 2021, while nominal GDP may take longer. We expect credit growth to stay strong in nominal terms in 2021-2022, but to slow down due to a high base effect. Corporate credit growth may pick up as Public Investment Fund programs generate business for contractors. Retail credit growth will stay strong due to a continued focus on mortgages, although the market will gradually become saturated. We expect the cost of risk to stay elevated in 2021, at about 120 bps. This reflects our view that the volatile global health situation and international travel restrictions still weigh on the economy. Additional provisions may be needed to offset the wind-down of SME support programs in 2021, including payment deferrals. Combined with very low interest rates, this will weigh on banks' profitability.

Russia.  We expect Russian GDP growth to recover to about 3.3% in 2021 and 2.2% on average in 2022-2024 after a contraction in 2020. The recovery in economic growth will be supported by internal consumption, as well as by the global economic recovery. We observed strong lending growth in the first quarter of 2021, largely driven by diminishing but still-high demand for mortgage lending under government-subsidized mortgage programs and some increase in corporate lending demand. We expect a gradual normalization of credit losses in 2021-2022 on the back of the anticipated economic recovery and accommodative monetary policy. In our base case, we project that the banking sector's cost of risk will drop to 1.5%-1.7% in 2021-2022 from about 2.1% in 2020. We think that NPLs will likely remain elevated, at about 9%-10% in 2021, reflecting the weakened macroeconomic environment, the absence of positive dynamics in disposable income, and potential downside risks to the economic recovery. Additionally, recognition of problem loans and provisioning might be delayed as some banks will try to manage the negative impact on their capitalization.

Estimations Of Expected Credit Losses In Banks' Financial Reporting

A key aspect of banks' financial reporting that is still relatively new--and broadly untested in a downturn--is the estimation of credit losses. Both International Financial Reporting Standards (IFRS, the financial reporting rules applicable across much of the world outside the U.S.) and U.S. Generally Accepted Accounting Principles (U.S. GAAP) require banks to take a more forward-looking approach when estimating credit losses on their loan portfolios. The new rules took effect in 2018 for banks reporting under IFRS and in 2020 for large banks reporting under U.S. GAAP.

The two sets of rules are markedly different in their approach. In broad terms:

• The U.S. GAAP model is based on an estimate of current expected credit losses, which requires banks to set reserves for expected lifetime losses on their entire loan portfolio.
• The IFRS model, expected credit loss (ECL, as set out in IFRS 9), requires a dual measurement approach under which a 12-month ECL allowance is established for performing loans and a lifetime ECL for underperforming and NPLs.

These rules mean that as banks increase their expectations for future credit losses, they may have to increase reserves markedly. And the difference between the two approaches is likely one reason why U.S. banks' credit losses in the early stages of the pandemic were markedly higher than some European banks', for example. We are now seeing U.S. banks reduce loan loss reserves, reflecting a combination of better-than-expected borrower performance to date and a more positive economic outlook.

For banks reporting under IFRS, a crucial question is whether loans under forbearance measures are experiencing a significant increase in credit risk since origination, because that is the trigger for the loan moving from a 12-month ECL provisioning requirement to a lifetime ECL. We expect to see increasing stage migration over this year--as economic uncertainty reduces and support mechanisms wind down, the underlying asset quality picture will emerge more clearly.

Related Research

Global:

• Ratings Component Scores For The Top 200 Banks Globally--July 2021, July 13, 2021
• Global Credit Conditions Q3 2021: Reopening, Reflation, Reset, June 30, 2021
• Global Economic Outlook Q3 2021: Picking Up Steam, Fueled By Vaccinations, June 30, 2021
• Lower And Later: The Shifting Horizon For Bank Credit Losses, Feb. 2, 2021

Asia-Pacific:

• Intervention Worked: Credit Losses Set To Decline For Most Asia-Pacific Banks, June 1, 2021

North America:

• Canadian Banks 2021 Second Half Outlook: Canadian Banks Are Poised For A Significant Recovery In 2021, July 12, 2021
• Rating Component Scores For U.S., Canadian, And Bermudian Banks (June 2021), June 30, 2021
• Fed Stress Test Results Give Banks Room To Release The Capital They Built Up During The Pandemic, June 28, 2021

Western Europe:

• As Near-Term Risks Ease, The Relentless Profitability Battle Lingers For European Banks, June 24, 2021

Latin America:

• Resilient Earnings and High Provision Coverage Position Latin American Banks Well To Face Tough Conditions, March 9, 2021
• Brazilian Public Banks Aim To Shed Government Debt Gradually, May 11, 2021