Banking industry focus on IFRS 9 expected credit losses

A banking industry focus on IFRS 9 expected credit losses

COVID-19: Top 5 IFRS Accounting Issues for Banks

At a glance

The COVID-19 pandemic has had and will continue to have far-reaching implications. In many parts of the world, governments have brought in never-before-seen measures including mass quarantines, social distancing, border closures, shut-downs of non-essential services and considerable (in some cases, unlimited) commitments to provide financial support to affected businesses and individuals. Just as the medical implications are emerging and evolving at breakneck speed, so too are those related to the economic and credit environment.

COVID-19 will impact many areas of accounting and reporting for all industries, as outlined in our publication In depth: Accounting Implications of the Effects of Coronavirus. For banks, additional challenges are likely to arise. In this article we provide our insights into what we believe to be the Top 5 issues for banks. These are:

1. Measuring expected credit losses (ECLs)

2. Identifying significant increases in credit risk (SICR)

3. Modifications and forbearance

4. Interim reporting under IAS 34 and other disclosure considerations

5. Government relief programmes

While this article focuses on the Top 5 issues, many others are certain to arise. As the situation continues to evolve, so too will the consequential accounting issues. For these reasons, the following is not an exhaustive list of all relevant accounting considerations. And while issues have been grouped under 5 headings, they will in many cases be interrelated.

1. Measuring expected credit losses (ECLs)

While the uncertainties arising from COVID-19 are substantial and circumstances are sure to change, we do not expect this to preclude banks from estimating their expected credit losses (ECLs). Estimating ECLs is challenging, but that does not mean it is impossible to estimate an impact based on the reasonable and supportable information that is available. A few things that may be helpful to keep in mind are that:

• Significant judgement will need to be applied in assessing the range of potential outcomes so as to meet IFRS 9’s requirement that the ECL reflects an unbiased and probability-weighted amount that is determined by evaluating a range of possible outcomes. An unbiased estimate is one that is neither overly optimistic, nor overly pessimistic.
• Given the speed with which events are unfolding, measuring ECLs at Q1 2020 is likely to be particularly challenging. Banks will need to develop an estimate based on the best available data about past events, current conditions and forecasts of future economic conditions. To the extent it is not possible to reflect the impact of COVID-19 in an institution’s models (and this is likely to be the case for many institutions, at least at Q1 2020), post-model adjustments or overlays will need to be considered. Updated facts and circumstances should continue to be monitored for any new information relevant to assessing the conditions at the reporting date.
• In terms of the methodology used to estimate ECL, no one size will fit all and different approaches may work best depending on factors such as local conditions, portfolio exposures, available data and existing models. Certain businesses or individuals may receive government support in some countries, while not in others.
• For interim reporting, in particular for Q1 2020, many institutions may not be able to perform a comprehensive ‘bottom-up’ analysis using loan-level probabilities of default that fully reflect all potential risks. Rather, it might be more appropriate to use top-down approaches (e.g. collective assessments or overlays) that focus on those segments that are most vulnerable.
• There is little doubt that economic conditions have deteriorated and this should be reflected in the macroeconomic scenarios applied by an institution and their weightings. In some cases, the prior period downside scenario may be an appropriate starting point for the current base case. Estimates are likely to be refined as additional information becomes available that is relevant to assessing conditions at the reporting date.
• Under IFRS 9’s ECL model, an expected credit loss will arise even where full recovery is expected on a loan, if payment is delayed and interest does not accrue during the deferral period at the effective interest rate of the loan. This is because there is a loss in terms of the present value of the cash flows.
• Disclosures are a critical component of ECL reporting, in particular given the level of measurement uncertainty resulting from COVID-19 (see 4 below).

2. Identifying significant increases in credit risk (SICR)

A key element in determining ECL is the assessment of whether or not a significant increase in credit risk has occurred, and hence whether a lifetime, rather than 12-month, ECL is required. In many cases and in particular at Q1 2020, it is unlikely that banks will have sufficient timely data to update loan-level probabilities of default which are often a core element of assessing SICR. As a result, a more likely approach may be collective assessments of qualitative factors and overlays, focusing on vulnerable segments of the loan book. Other factors to consider include the following:

• Assuming either that all stage 1 exposures move to stage 2 or 3, or alternatively that no exposures move to stage 2 or 3, is unlikely to be appropriate in many cases. The extension of blanket financial support to all borrowers in a certain class (e.g. all household mortgages) does not automatically mean that all such borrowers have experienced a significant increase in credit risk. Nevertheless, and notwithstanding extensive government financial support, debt levels are expected to rise and this will typically affect credit risk assessments. Judgements therefore need to be made to distinguish between those exposures that are significantly affected and those which are affected to a lesser extent, including within individual segments or portfolios.
• SICR is based on the likelihood of a default arising, and not on the likelihood of losses. Hence, some government relief programmes may not impact SICR assessments. For example, those programmes that provide cash directly to borrowers quickly and thus mitigate the risk of default should be considered, whereas those which provide funding or guarantees to financial institutions and only mitigate the losses incurred by those institutions should not. This may mean that a SICR has arisen, even in cases where it is expected that any losses that arise will be fully recovered by the bank.
• Staging might have a smaller impact on the overall ECL than other judgements and estimates since COVID-19-related defaults might be expected to take place quickly. For instance, this would be the case if COVID-related defaults are expected to arise within the next 12 months and so are already captured within Stage 1 ECLs.

3. Modifications and forbearance 

To help borrowers cope with the financial consequences of COVID-19, many banks and governments have announced various types of relief programmes that involve payment holidays, such as:

• Blanket moratoriums on debt payments for all borrowers in a certain class (e.g. all mortgages); and
• Case-by-case relief to:

                a. Those most affected;

                b. Any who request relief; and/or

                c. Those considered to have a good propensity to pay absent COVID-19.

Typically, these programmes require continued accrual of interest during the period of the payment holiday (where that is not the case – i.e. where interest is forgiven – additional considerations will likely apply). Given the unique features of many of these programmes, when determining the extent to which they give rise to a SICR past practices for payment holidays may not be appropriate. In particular, blanket moratoriums are unlikely to indicate all the loans in the affected population have suffered a SICR. However, certain customers within that population would be expected to have suffered a SICR, and so alternative ways of identifying this group would need to be considered. For Q1 2020, a starting point may be to use pre-COVID-19 risk ratings to determine which exposures were previously ‘closest to the line’ and hence are more likely to have suffered a SICR.

4. Interim reporting under IAS 34 and other disclosure considerations

Many regulators around the world are revising timelines and requirements for interim reporting. When banks do issue interim reports under IAS 34, it will be important to keep in mind the overarching requirement to explain events and transactions since the end of the last annual reporting period that are significant to understanding changes in financial position and performance. Key considerations in meeting that requirement, and when preparing other forms of interim reports, are likely to include:

• Critical estimates – Clearly identifying and explaining the critical estimates used in determining ECL will be important. Whilst 31 December 2019 disclosures on critical estimates will in many cases constitute a good starting point, a simple roll-forward of these disclosures is unlikely to be appropriate. There are likely to be new aspects of accounting that have become critical due to the changes in the economic environment and in market dynamics. Hence, past disclosures on previously identified critical estimates may no longer be relevant. In addition, those banks which previously disclosed numerical sensitivities may be unable to ‘re-base’ them at Q1 2020 to reflect the current uncertainty in a meaningful way. Indeed, such numerical sensitivities might actually risk misleading users if they are likely to be quickly superseded, in which case temporarily replacing them with a more qualitative analysis might provide more relevant information to users.
• Telling the story – Disclosures should reflect factors that are specific to the bank rather than being boilerplate, and should tell the story of how the estimate was developed. Such disclosures would include describing how the credit and other risks that the bank is exposed to have been impacted by COVID-19, how the impacts of COVID-19 have been incorporated into the ECL estimate, and the extent to which there is uncertainty and hence estimates might change in the future.
• Credit risk concentrations – Given the different impacts across sectors, updating previously disclosed analysis of portfolios by industry or region will be important. As was evident during the 2008 financial crisis, the level of granularity demanded by users will likely increase. For example, in the past a bank may have disclosed its exposure to the transport sector without further disaggregation. This may now need to be sub-analysed to help users understand the different underlying exposures and risks, for example by analysing the exposures into airlines, state-backed train companies, and haulage and freight companies.
• Credit risk management practices – The ways that banks manage credit risk are very likely to change, particularly given the large scale programmes to grant payment holidays and other reliefs that are being offered or mandated in many territories. It will be important to ensure there is a clear explanation of these programmes and their effect on credit risk practices, as well as any expected or potential impacts on the bank’s financial reporting.
• Fair values – Significant changes in fair value are explicitly required to be disclosed under IAS 34, as are significant transfers between levels in the fair value hierarchy. Given recent decreases in asset prices and liquidity in many markets, banks should provide sufficient information for users to understand these changes and their impacts. Where there have not been significant impacts, disclosing this fact may also be material information, given the risk of what might otherwise be assumed in the current environment.

5. Government relief programmes

Many governments, central banks and other agencies are developing programmes to provide economic support. Where this intervention is made through the banking system (e.g. by providing funding or guarantees to banks at potentially advantageous rates or terms), a key accounting consideration is whether an element of the transaction is a government grant. This can impact the timing of recognition of the effects of the relief, the presentation of those effects and what disclosures may be required.

In order to determine the appropriate accounting treatment, it will be important to understand the exact details of each particular support arrangement. Some of the factors to consider when assessing the accounting treatment are:

• Whether the programme is on arms’ length terms based on past transactions or market pricing, or can be considered to be ‘on-market’ for transactions of this kind in the current environment (i.e. including transactions with a government or government agency).
• If a programme does contain a government grant, whether there is ‘reasonable assurance’ that the grant will be received as required by paragraph 7 of IAS 20, taking account of factors such as:

a. Which aspects of the support remain uncertain and how critical are they?

b. Which transactions, with which counterparties, will be eligible for relief under the programme and how will that relief or benefit be received by the bank?

c. Will the government be able to deliver the stated reliefs, considering practical challenges as well as its ability to pay?

d. Are any subsequent clarifications adjusting post balance sheet events for accounting purposes?

• If recognition of a government grant is determined to be appropriate, the timing of recognition of the benefit in profit or loss, and its presentation and disclosure.