Getting governance right on IFRS 9 Expected Credit Loss: accounting policy and implementation decisions: PwC In depth INT2015-16

Publication date: 19 Nov 2015

Governance processes and controls are an essential part of any bank’s control environment. They will be particularly critical for banks implementing IFRS 9 Expected Credit Loss (ECL) and making key decisions on accounting policies and how practically to implement the new impairment requirements. The importance of strong governance is further reinforced by the draft Basel Guidance, which emphasises the need for a robust and high quality implementation.

Thisadobe_pdf_file_icon_32x32In depth outlines some of the key governance challenges we have seen in practice when making IFRS 9 ECL accounting policy and implementation decisions, as well as how best to respond. Although primarily focused on banks, many of the areas discussed will also be relevant to other financial institutions implementing IFRS 9 ECL.

Although not covered in this publication, banks will also need to consider many other aspects of governance during their IFRS 9 implementation projects and beyond. These include data governance, model governance and governance & controls over the ongoing ‘business as usual’ IFRS 9 reporting process.


At a glance

Governance processes and controls are an essential part of any bank’s control environment. They will be particularly critical for banks implementing IFRS 9 Expected Credit Loss (ECL) and making key decisions on accounting policies and how practically to implement the new impairment requirements. The importance of strong governance is further reinforced by the draft Basel Guidance, which emphasises the need for a robust and high quality implementation.

This ‘In Depth’ outlines some of the key governance challenges we have seen in practice when making IFRS 9 ECL accounting policy and implementation decisions, as well as how best to respond.  Although primarily focused on banks, many of the areas discussed will also be relevant to other financial institutions implementing IFRS 9 ECL.

Although not covered in this publication, banks will also need to consider many other aspects of governance during their IFRS 9 implementation projects and beyond.  These include data governance, model governance and governance & controls over the ongoing ‘business as usual’ IFRS 9 reporting process.

Background

Publication date: 19 Nov 2015

Many banks are currently discussing their key IFRS 9 ECL accounting policy and implementation decisions, before then commencing model builds and IT implementation based on those decisions.  Governance failings that lead to initial IFRS 9 implementation decisions being changed later could be extremely costly and jeopardise project delivery.  Additional uncertainties around how the final guidance of Basel, the Enhanced Disclosure Task Force (‘EDTF’) and the IFRS Transition Resource Group (‘ITG’) will be applied in practice, as well as general emerging industry practice, add to the complexity in decision making.

Given the significant impact that IFRS 9 ECL will have on banks, there are a wide range of stakeholders with a strong interest in these decisions and how IFRS 9 is implemented. These include senior management, audit committees, regulators, shareholders, investors, analysts and auditors.  Getting governance right will therefore be key to making effective accounting policy and implementation decisions, that can be justified to internal and external stakeholders, both at implementation and in the future.

But this isn’t easy.

There are a range of challenges and issues we see that need to be considered in designing and operating an effective governance process over IFRS 9 ECL accounting policy and implementation decisions, which are summarised below.

Picture1

Each of these areas is discussed in more detail over the following pages, along with our perspectives on how best to address the individual challenges presented. However, the use of a Project Management Office (‘PMO’) to build these individual responses into an overall project plan and track progress, will be one way of ensuring an effective, co-ordinated response to the various challenges faced.

1. Right areas

Publication date: 19 Nov 2015

Many banks are currently considering and debating the key ECL accounting and implementation decisions in their internal governance forums. The most actively discussed include the definition of a significant increase in credit risk (‘SICR’), the use of forward looking information and certain practical expedients, particularly the use of changes in a 12 month Probability of Default (‘PD’) as an approximation to the changes in a lifetime PD when assessing SICR.

There are however a number of other accounting and implementation decisions which have had less industry focus to-date, but which we also see as highly judgemental, and which will require a similarly robust governance focus:

  • Derecognition: An area that banks are increasingly reconsidering is their asset derecognition policy under IFRS 9. An entity’s policy on derecognition impacts the ‘origination date’ for a loan and hence the date at which data is needed to assess if there has been a significant increase in credit risk since origination. This is particularly relevant for revolving credit facilities such as credit cards, where many consider that comparing current credit risk to the risk when a credit card was first issued (possibly 20 years earlier) does not represent the real change in risk.  But if a derecognition policy is to be changed, and the criteria of IAS 8 for such a change are met, other products also need to be considered, for example residential mortgages and corporate lending. This will ensure that the policy is consistently applied across all product types and that the impact of any proposed changes to the policy are fully understood, on ECL and more widely. For example, incremental losses may be recognised if more loans are derecognised, given the requirement for the new loans to be measured initially at fair value.  If a contract modification isn’t judged to result in derecognition, then the requirements of IFRS 9 paragraph 5.4.3 to recognise a modification gain / loss will need to be considered.
  • Quality of credit risk data at origination: Paragraph B7.2.2 of IFRS 9 states that on transition, an entity should seek to approximate the credit risk on initial recognition by considering all reasonable and supportable information that is available. If an entity is unable to make this determination without undue cost or effort, IFRS 9 para 7.2.20 applies and an entity should recognise a loss allowance at an amount equal to lifetime ECL at each reporting date until that financial instrument is derecognised. Judging whether data sufficiently 'approximates' origination credit risk, so that ECL does not default to a stage 2 lifetime expected loss at transition, is likely to prove very judgemental. Few banks in our experience have yet considered this area in any detail.
  • Dealing with regulatory-driven prudence: Paragraph 5.5.17 of IFRS 9 states that an entity shall measure the expected credit losses of a financial instrument in a way that is unbiased. In many areas banks are planning to leverage models and data already used for risk and capital purposes.  In contrast to this ‘unbiased’ principle of IFRS 9, current risk and capital practises often embed an element of in-built prudence within models and data. Completely identifying this prudence and assessing its impact on IFRS 9 implementation is likely to be challenging.

2. Right people

Publication date: 19 Nov 2015

One of the inherent challenges in implementing IFRS 9 is the range of people within a bank that will need to be involved.  Unlike some other accounting standards, Finance really cannot do it alone. Set out below are perspectives on who should be involved and how.

  • Finance vs Risk: Many banks have historically seen limited interaction between the Finance and Risk functions. However, IFRS 9 will require significant involvement from both. This can introduce complexities in deciding who has overall responsibility for the implementation project, though ultimately it will be the CFO who will need to ‘sign off’ on the ECL number as part of the financial statements. Their different backgrounds can also present a challenge to Finance staff in understanding ‘Risk-speak’ (and vice versa) and in ensuring there really is a common understanding of what is being explained and decided. This isn’t always the case in practice. Similarly, it can be challenging for Risk staff to understand the detailed requirements of IFRS 9 and appreciate its differences from common risk modelling practice. For example, the importance in IFRS 9 of relative changes in credit risk run counter to the absolute thresholds that typically underpin existing ‘Watchlist’ and other credit risk monitoring controls.  Promoting a ‘one team’ culture and continuous communication are key to overcoming these challenges.
  • Composition of IFRS 9 committees: Banks will need to consider who should be included in IFRS 9 committees and working groups, in particular the Steering Committee ultimately responsible for monitoring progress and approving key decisions. Risk, Finance, Regulatory Reporting, IT and PMO will typically be key members. But ECL will have a major impact on the economics of many businesses, so input from the business themselves should be considered. This will be particularly important if IFRS 9 may lead to product changes, for example changes in credit card terms to clarify origination dates, so that IFRS 9 projects anticipate these changes and respond accordingly. Investor relations will also play an important role in the messaging of the impacts of IFRS 9 adoption to the market.
  • Audit Committees: To allow audit committees to effectively challenge management on IFRS 9 judgements, briefing and education sessions will be key. Given the complexity of IFRS 9 ECL, the time needed shouldn’t be underestimated. Agreeing upfront the key dates at which the Audit Committee will be consulted will help ensure ‘no surprises’ and allow internal and external audit to share their views and insights at the appropriate times – 2018 will be too late.
  • People change: IFRS 9 projects will be lengthy. Most will need all the time available up until 1 January 2018, some banks are considering deferring non-critical elements beyond even this date and there is likely to be significant competition for skilled staff in the market place. As a result, there could be considerable people change over the course of the project. Banks should ensure that decisions and analysis are promptly documented, the knowledge of key individuals is actively shared and handover procedures are adhered to, in order to minimise the risk from departures.

3. Right now

Publication date: 19 Nov 2015

There are many complexities in implementing IFRS 9 ECL that pose challenges to effective governance. Some of the most pervasive that we see in practice are discussed below.

  • Theoretically perfect vs practically possible: Issues such as historical data gaps may mean a fully 'text-book' approach to IFRS 9 is considered impractical. But this requires assessment of the impact of the proposed simplifications or compromises, to either justify them on technical grounds or to demonstrate that the effect of applying practical workarounds is not material. This brings with it the inherent difficulty of demonstrating the impact of an alternative approach, without actually calculating the fully ‘text-book’ approach in the first place. In practice, this means that decision papers will not be ‘pure’ technical accounting papers. Instead, they will need to include data such as size of portfolio, historical losses, estimated stress case losses etc, to put the judgement in context and help those charged with governance come to an informed decision.
  • Practical impact of decisions: It is very difficult to decide between different possible options based on theory alone, without seeing the actual impacts of the different options. For example, what impact will different proposed approaches to determining SICR actually have on how many loans will be transferred to stage 2, the timing of transfer, the number that will quickly be transferred back to stage 1 etc?  Without this understanding, a lot of time can be spent debating issues that have little impact in practice. But the actual impact can often only be fully understood afterwards, once a model based on the agreed theoretical approach has been built. Simplified simulations of the proposed approaches, using either test data or samples of actual portfolio data, will help minimise this uncertainty. Where this is not possible, there needs to be clarity on when, and how, the practical impacts of the proposed approach will be reviewed by the governance process, to reconfirm their initial decision or change it if appropriate.
  • Managing uncertainty: Given the many possible uncertainties at the time of decision making, it is important that decision makers have clarity on any contingencies inherent in a proposed implementation decision. For example, if a planned modelling approach assumes that it can be demonstrated that changes in the 12m PD approximate changes in the Lifetime PD:
    • What evidence indicates this is likely to be true, so that it is a reasonable working assumption?
    • What specific work will be performed to provide the final, comprehensive analysis demonstrating the assumption is valid?
    • When will that analysis be completed and how will project timelines be impacted if the assumption doesn’t prove valid? 
  • Level of granularity of decisions: When deciding the decision points to be considered by the governance process, a bank needs to assess whether all the same considerations will apply for credit cards, mortgages (including repayment, interest-only), commercial loans etc, or whether different product types will need separate consideration by the governance process. Aggregating too many products risks overlooking product-specific issues, but too granular an approach will generate an unmanageable number of decision points and technical papers and risk inconsistent approaches. Steering Committees will need to satisfy themselves that the granularity of the approach taken is appropriate, as well as the appropriateness of risk ratings if (particularly for larger banks) these determine which level of the governance hierarchy will conclude on each decision, for example working group, sub-committee or main Steering Committee.

4. The Future

Publication date: 19 Nov 2015

Addressing the immediate concerns of IFRS 9 ECL within tight project timelines can be tough enough. But to make a real success of implementation, a number of more forward looking aspects shouldn’t be overlooked.

  • Iterative decision process: Implementation decisions will often be arrived at by an iterative process, as governance processes raise challenges, proposals are reworked, and updated proposals discussed again. This increases the risk that not all elements supporting the final overall decision are documented appropriately. This will be fundamental for evidencing an effective governance process to all relevant stakeholders and the importance will be heightened for Sarbanes Oxley (‘SOX’) reporters.  In practice, collating a single document shortly after final conclusion is reached, that pulls together the various challenges raised, analysis and sub-conclusions, will avoid the need to do this retrospectively. Experience from IFRS conversions in 2005 and more recently showed staff may well have moved on or minutes may turn out to be incomplete or unclear if this is done some time afterwards.
  • ‘Real time’ controls implementation: Banks have a number of inherent challenges in ensuring robust controls are in place over implementation decisions, that will stand up to independent scrutiny for 2018 reporting. There is typically far more enthusiasm for debating the technical issues than for the formalities of controls documentation. Many Risk staff will not be familiar with the rigour of SOX controls documentation, a significant behavioural change when Finance functions first had to comply with SOX, and there will be similar considerations even for non-SOX reporters.  And whilst there may be a collective view that “all the right things are being done”, formally defining the design and required evidence for implementation controls often identifies gaps or areas requiring clarification. For example, what (if any) independent testing should be performed over analysis supporting key judgements, what data / analysis needs to be retained as evidence and to provide an audit trail? To pre-empt such issues, implementation controls should be formally documented, agreed and communicated upfront, with ongoing monitoring to ensure they are being appropriately implemented, documented and (where necessary) remediated in ‘real time’, before people move and memories fade.
  • Materiality judgements: Where materiality judgements are made, for example on ‘immaterial’ portfolios, or where practical expedients are used, consideration should nevertheless be given to:
    • Are judgements truly 'future-proof'? For example, will they still hold with planned growth in strategically important portfolios, or in a higher interest rate/stressed economic environment when IFRS 9 ECL reporting will probably matter most?
    • Is there clear documentation of the key underlying assumptions behind the materiality judgements, to facilitate periodic reconsideration of those judgements in the future?
    • How is the aggregate impact of materiality judgements assessed, to ensure collectively these judgements won’t introduce material bias?
    • Do materiality judgements consider the different materiality levels applicable to consolidated group and solo legal entity reporting?
  • Future developments: Given time pressures, banks are having to make accounting policy and implementation decisions before the various uncertainties of IFRS 9 ECL are resolved, for example how the final Basel / EDTF / ITG guidance will be applied in practice and industry views on good practice. Governance processes therefore need to consider how best to track these developments, identify the areas potentially affected both directly and indirectly, and conclude on the appropriate response.
  • Disclosures: The current focus of banks is naturally on the key decisions impacting measurement of IFRS 9 ECL, rather than other aspects of financial reporting, such as disclosures. But disclosures will be fundamental to explaining transition decisions and their resulting impact to investors and other external stakeholders. So questions that should be asked as decisions are made should include “How would we explain this in our 2018 disclosures?”, “What questions would a third party ask if they read that explanation?”, “How are systems and processes being designed to generate the information needed to produce the related disclosures required by IFRS 9”?
 
close
  • Add Selected section
  • OR
  • Add Full document
  • Print Selected section
  • OR
  • Print Full document
  • Copy link
  • Add to my favourites
close

Warning, this action will download the whole document into PDF format.

Cancel
close

Warning, this action will add the whole document to my documents.

Cancel
Back to top