COVID-19 and provisioning for expected losses

The provisioning standards in effect before and during the global crisis were based on the concept of incurred losses. They have been criticized for excessively delaying the recognition of credit losses by banks, and therefore for being a source of procyclicality (Financial Stability Forum 2009). In response, the International Accounting Standards Board and the Financial Accounting Standards Board in the United States introduced new rules based on the concept of expected loss. In the EU and much of the rest of the world, IFRS 9 came into effect in 2018, while in the US, the current approach to expected credit losses (CECL) is slated for 2021. With some differences , both standards call for the recognition of credit losses which, based on available information (including macroeconomic information), can be expected over time horizons ranging from one year to the life of a credit instrument. . The goals behind this more forward-looking approach to provisioning included inducing more cautious lending behavior in good times and encouraging earlier corrective action in bad times.

Many observers and supporters of the new standards have been intellectually enthralled by the idea that provisions based on “expected” rather than “incurred” losses should be less pro-cyclical. However, the complicated internal modeling required to estimate future credit losses under IFRS 9 and CECL would only be truly countercyclical if banking modelers were able to predict turning points in the cycle, or the arrival of disasters such as the COVID pandemic. -19 two or three years in advance. In such circumstances, they would be called upon to fund their provisions (loss buffers) at the right time: against the profits made in the two or three years before the cycle reversal or the arrival of the catastrophe. This would bring the banks into contraction with additional buffers.

Provisioning for expected losses can add procyclicality

However, as we show in Abad and Suarez (2018), in the absence of banks’ ability to anticipate unfavorable developments in aggregate conditions sufficiently in advance, the initial recognition of expected future losses may paradoxically imply that, from the onset of economic contractions, banks experience a sharper decline in profits and, therefore, capital than under the old loss paradigm. Intuitively, the expected loss approach implies having to provision, when the contraction begins, for certain anticipated credit losses that would only be provisioned as and when they appear in the incurred losses approach (see Figure 1).

Figure 1 Effects of the arrival of a contraction under different provisioning rules

To note: This figure is based on the results of Abad and Suarez (2018) which quantify, according to alternative provisioning standards, the impact of the onset of an average recession on a bank portfolio of loans to European companies and its implications on the income statement and capital (CET1) of the correspondent bank. The horizontal axis represents the years after the onset of a recession.

For banks subject to capital requirements and having a limited capacity to raise new capital, the loss of capital may imply a contraction of their credit supply, as has already been pointed out in the literature on the procyclicality of capital requirements. Basel own funds (Kashyap and Stein 2004, Repullo and Suarez 2013). This mechanism is particularly relevant in the context of the COVID-19 pandemic where an unforeseen shock caused a sharp contraction in economic activity. The least traumatic absorption will require channeling sufficient credit and liquidity to the most affected sectors, thus questioning the implications of provisioning for expected losses in such periods.

The quantitative results of our analysis indicate that, for a hypothetical bank wholly specializing in lending to European companies, the onset of an average recession would imply increases in the impact of IFRS 9 and CECL equivalent to around 100 basis points of CET1 (as a fraction of total assets), i.e. more than a third of the bank’s fully charged capital conservation buffer (CCB). This impact is approximately twice as large as that implied by the previous approach to losses incurred.

COVID-19 crisis calls for preventing procyclical effects

If the COVID-19 contraction were to have an impact on foreseeable defaults several times greater than an average recession, the initial impact on banks’ capital could easily exhaust voluntary and regulatory capital buffers (in addition to minimum requirements) with which banks enter. this crisis. Additionally, beyond its initial impact, uncertainty regarding the depth and duration of the COVID-19 crisis may cause banks to face greater variability in their provisioning needs (and greater volatility). of their available capital) within the framework of the new provisioning standards compared to the old ones. .

The sudden and unexpected nature of the pandemic, and the certainty that its implications will be less damaging to the economy if there is a lot of credit flowing to affected businesses and households, suggests the advisability of preventing the pro-cyclical damage that new provisioning standards may cause in this crisis.

In a recent column advising against any changes to current provisioning standards, Nicolas Véron asserts that “there is no perfect accounting thermometer for credit risk in banks’ loan portfolios, but breaking the current thermometer in the midst of a crisis would do a lot more harm than good. . Fortunately, however, there are ways to neutralize the impact on capital (and likely credit supply) of IFRS 9 (or its American companion CECL) without ‘breaking the thermometer’, that is. – say without abandoning the estimate of expected credit losses resulting from this crisis or any other in the immediate future, and their reflection in the banks’ income statements.

No need to break the thermometer: use the transitional arrangements

Such a solution is based on the authorization of capital adjustments of the same nature as those which have been applied in the EU in the form of “transitional arrangements” since the introduction of IFRS 9 in 2018. Such arrangements have been made. set up to smooth over time. the impact on capital and credit of the additional provisioning requirements implied by the new standards. Banks have had the option of using them or not.

In essence, the existing transitional arrangements allow banks to take back as available capital up to 95% of the absorption of additional losses implied by the new arrangements in 2018, 85% in 2019, 70% in 2020, 50% in 2021 and 25% in 2022. The transition was scheduled to end in 2023. Within the framework of these provisions, provisions are constituted according to the new paradigm of expected losses, so there is transparency on the estimates of losses expected by the companies. banks, but part of their impact on the bank’s available capital (CET1) is neutralized.

For us, ensuring that the new standards are not a source of additional procyclicality in the fight against the economic consequences of the COVID-19 pandemic calls, at the very least, to stop the clock of the transitional arrangements (which in 2020 would still allow banks to resume 70% of additional provisions). In fact, we think it would make sense to delay the clock to its 2018 position: allow banks to add 95% until further notice.

To avoid the temptation of banks to enter into games where they signal their strength by withdrawing from the transitional arrangements, we are arguing for making the adoption of the amended transitional arrangements compulsory. And to avoid “wasting” the capital released on distributions, we would expect banks and / or their supervisors to ensure the freezing of dividends and share buybacks during the stand-by period, as part of ” a “counterpart” agreement. Finally, our proposed measure would come with the announcement that once this exceptional period has passed, the path of transitional arrangements would be resumed, after appropriate rescheduling.

In the EU context, with the transitional provisions of IFRS 9 already in place, dictating the freeze or postponement and freeze of the transition clock would involve minor legislative adaptations. This would not call into question the verification of compliance with the rules relating to the status quo (in which such transitional provisions already exist). From the point of view of preserving the informative value of the accounting statements and the continuity of the costly procedures necessary for the adoption, internal validation and application of new standards, we believe that our proposal is superior to the pleading alternatives. for a simple deviation from new standards (Stein et al. 2020) or by applying them with additional flexibility (and ambiguously defined), as recently recommended by the Bank of England and the ECB.


To conclude, if we sympathize with the principle that we must avoid changing the rules of the game in the middle of a game, the nature and severity of the COVID-19 crisis and the expected importance of bank credit to deal with it would call for and fully justify the adoption of our easily applicable proposal by the authorities.

The references

Abad, J., and J. Suarez (2018), “The Procyclicality of Expected Credit Loss Provisions”, CEPR Discussion Paper 13135.

Financial Stability Forum (2009), “Financial Stability Forum Report on Combating Procyclicality in the Financial System”, April.

Kashyap, A., and J. Stein (2004), “Cyclical implications of the Basel II capital standards”, Federal Reserve Bank of Chicago Economic Outlook, 1st trimester, pp. 18-31.

Repullo, R., and J. Suarez (2013), “The procyclical effects of bank capital regulation”, Review of financial studies 26, p. 452-490.

Stein, JC, R. Greenwood, SG Hanson and A. Sunderam (2020), “How the Fed Can Use its Bank Prudential Authorities to Support the Economy and the Flow of Credit”, Harvard University, March.

End Notes

1 See

2 See

3 See and https: //

Comments are closed.