IFRS 9 – Ready or not?

IFRS 9 (the new accounting standard) is fast approaching with many organisations already in full swing in terms of development and with their chasing pack firmly in the planning stages for design and build.  But just how ready are you for the impending changes?

Here, we will explore some of the key changes at a high level, how to implement the changes, what to expect, how to get ready, key differences in approaches across organisations and the reasons for these.

This article will also cover in detail some of the more difficult concepts or those that are new for IFRS 9, such as:

  • Stage 2 Transfer Criteria

  • Disclosures

Why are the standard setters making a change?

The new standard is a distinct departure from the current International Accounting Standard 39 (IAS39) Financial Instruments: Recognition and Measurement, in that it focuses on accounting for expected credit losses as opposed to incurred losses. This is a response to criticism levied on the current standard, particularly in the wake of the financial crisis of 2007, 2008 and 2009 and the strain that this has placed on certain organisations and regions.

Criticisms levied at the current standard

There have of course been a number of criticisms levied at the current standard, namely: 

The final instalment will also include a summary of what we have learned and where the best sources of information from regulators and standard setters are.

  1. It was slow to respond to recognition of credit losses, specifically only recognising lifetime expected credit losses on exposures where there was objective evidence of impairment and an additional allowance (incurred but not reported) for exposures expected to become impaired imminently.

  2. That it was too reliant on backward looking information, particularly delinquency data and historical loss rates.

  3. That it only considered drawn loans, not commitments to lend. As a result many credit card and current account providers were not recognising sufficient loss allowances.

  4. That the information provided to users of financial statements is not transparent enough to be useful.

  5. It was too complicated to implement and maintain as there were multiple approaches depending on the asset.

What the new standard IFRS 9 aims to address

The new standard attempts to address these issues in a number of ways:

  1. IAS39 was slow to respond by design. It was a response to some organisations applying balance sheet management under UK GAAP to recognise losses that would not be incurred in good times so that they could be released to prop up the P&L in bad times. IFRS 9 addresses this by recognising lifetime expected credit losses for all exposures that have objective evidence of impairment, all exposures that have significantly increased in risk since initial recognition and exposures that are expected to default in the next 12 months.

  1. The standard demands that forward looking information is used and losses are accounted for now on the basis of what we believe will happen in the macro – economic environment in the future and, the specific impact that has on an organisation’s different portfolios.

  1. Under IFRS 9, the date of recognition is the date at which the commitment to lend is made. This means that for pipeline business and revolving credit products, losses must be recognised based on the expectation of draw down over the behavioural life of the exposure.

  1. The disclosure requirements are more granular, requiring split by risk grade, LTV, drawn and undrawn and also requiring disclosure of key assumptions (e.g. transfer criteria) to ensure users have the best chance of understanding the organisations underlying financial position.

  1. There is a single impairment model for assets measured under amortised cost and those measured under Fair Value Other Comprehensive Income (FVOCI).

Summary of Key Changes

There are a number of key changes, and whilst a summary is provided below, they will be discussed in more detail over the coming months.

Part One – here we cover the following elements:

Stage 2 Transfer Criteria

Under IAS39, there are only two states, Objective Evidence of Impairment (Impaired) and Incurred But Not Recognised (IBNR). Different organisations will have different definitions of Impaired, based on Days Past Due (DPD) or other evidence, for example Forbearance.

Under IFRS 9 Stage 3 aligns to Impaired (although some organisations are potentially revising their definition) but Stage 1 and Stage 2 have no direct comparison. The Standard clarifies that Stage 2 is where there has been a significant increase in risk, when comparing lifetime probability of default (PD) since origination, but does not define significant, more on that later.

Disclosures

IFRS7 is the Standard that describes the Disclosure requirements for financial assets, and specifically in this context the losses that result from the holding of those assets. IFRS9 has made amendments for those requirements, and in particularly the level of granularity, which is very much in line with IFRS objectives of providing more information to users of financial statements.

IFRS9 - How to make the changes

The changes are substantial and require significant thought and effort by individual organisations and their advisors to develop a compliant solution that is right for them. Some larger, more complex and systemically significant organisations have been working on this for a number of years and still don’t have all the answers, so it is important to determine appropriate timescales for your organisation.

Below is an indication of the various timelines organisations are working to:

IFRS9 Timeline of Required Changes Example

The process generally consists of four components:

  • A design phase, confirming scope, performing a gap analysis against current capability and the IFRS 9 standard. A methodology covering all the key elements of IFRS 9 and an appropriate approach for the organisation, considering the range of products and materiality and any supporting analysis. A design which utilises these findings and relates them back to actual data repositories within the organisation, together with proposed solutions.

  • A build phase, delivering to the design but sense checking along the way to ensure that components explained in the design perform as expected when developed in full. An engine build to tie all the elements together and perform checks and balances to ensure robust calculation.

  • A testing phase, developing scripts to test against, performing those checks on individual components and the solution overall and supporting UAT. This would be done just prior to parallel run.

  • A reporting phase based on all reports built based on design specifications and learnings from parallel run.

Why the different approaches?

Organisations are generally taking different approaches due to the complexity or size of the business or because of their reporting schedule.

As an example, a typical large high street banks, may well have a full service offering on their retail side, so have to devise solutions for Unsecured term loans, mortgages, overdrafts, credit cards and car leases. In addition to this, they may have different offerings for different customer segments, for example, sub-prime or High Net Worth (HNW) or legacy products as a result of an acquisition or merger. Whilst there may be an element of crossover, this all requires thought and analysis to determine the right level of segmentation and development. On the commercial side, there may be similar issues but these may be compounded by lack of data or lack of suitability to modelling, so they may have to define a more manual approach. Clearly the design and delivery is much simpler and requires less time for an organisation with only a retail mortgage offering for example.

The level of expectation is also different for different organisations. The recent guidance issued by Basel Committee on Banking Supervision (BCBS) 1 on Accounting for Expected Credit Losses refers to proportionality, and whilst it makes clear that this should not jeopardise a high quality implementation it should allow organisations to adopt approaches commensurate with the size and complexity of the organisation as well as the economic significance of it.

The reporting period also plays a part, IFRS 9 requires organisations to adopt in the reporting period after 1 January 2018, for organisations that are not due to report until year end until Q3 of 2018, some feel they have more time and therefore have a delayed start. Whilst this can be acceptable for a final deadline, it is certainly good practice to start the project sooner rather than later to ensure senior management have the right amount of time to consider all of the potential impacts.

What can I expect?

From experience there are a number of things common to every IFRS 9 project:

  • The solution you thought you would arrive with at the start is not the solution you end up with. This is a vastly complex challenge and whilst simplifications can be applied, they need to be relevant and justifiable for your organisation. This process takes a lot of thought and analysis and will sometimes prevent the organisation doing what they wanted to do initially or open up a whole new opportunity.

  • You will need some support. Even if you have enough people in your BAU team, it is unlikely they have all the requisite skills to get the right result for your organisation in a way that meets the standard. Whether that comes from specialist EL modelling experts or technical accountants, availability may be limited and needs to be planned in.

It’s an exciting challenge with lots of opportunities. Whilst this is a challenge it is a really exciting one, on every project I have been involved in there have always been a number of interesting debates that have lead to very interesting solutions

Transfer Criteria

Probably the fundamental change for IFRS 9 is the concept of ‘Significant Increase in Risk’ and I think it is fair to say it is the one area the industry is having the most difficult time in responding to. Credit Risk professionals tend to think in terms of absolute risk and managing that, rather than relative risk that the standard requires. One solution to this is to utilise internal current risk metrics and specific elements of the standard to develop tests to define significant movement.

First, to assess an initial range, one needs to be determined, this can either be done by judgement or analysis to show maximum movement from origination to default over a given time for individual segments. In any case, the actual range is not important but if there is a degree of confidence that the optimal value lies within the range it does make the process more efficient.

When these tests are applied to test data in isolation, it gives you a range to assess the movement to Stage 2 against.

Test Results Graph - Direction Changed Transfer Criteria

 

It is possible to see from the above that Test 5 works very well for all options, whereas for the rest of the tests there is some degree of differentiation depending on which factor is used as a potential stage allocator. To make these measures comparable it will be necessary to incorporate elements of judgement depending on how significant one test should be in comparison to another. For some of these tests, for example Test 1, it could be used as a limiter in the first iteration so for example, if internal judgement suggested that the result to this should be at least 65%, this would automatically rule out 1000% as a relative PD increase.

2nd Iteration TC Test Results (Direction Changed) Graph

 

The same tests are repeated but for a reduced range and their relative weighting reassessed, (based on judgement) to further refine the range. In this case, judgment and the results that this is based on have highlighted that Tests 2 and 4 are most significant in terms of optimisation and as such have been repeated in isolation.

Test Optimisation Graph - Weighted Optimal Test vs Marginal Improvement

The optimisation tests transform the original test into ‘marginal benefit’ tests. This stage isn’t critical and can be omitted if the organisation has a preference for a less technical approach.

As the two optimisation curves have been normalised, both normalised optimal rates can be plotted, and combined mathematically. In the example below, equal weight (of 10) has been given to both tests:

Weighted Optimal Function = (10 x Marginal Test 2 Benefit) x (10 x Marginal Test 4 Benefit).

This shows that the optimal multiplying factor for this segment is between c300% and c450%, these tests can be repeated for different segments and risk grades and further refined.

Disclosures

Whilst many of the requirements for Disclosures haven’t changed from IFRS 7, the introduction of different stages and the movement of these over the reporting period does, increase the level of granularity required.

An example below of the movement analysis required below is a good one, under IAS39 only the increase or release from impairment (collective and specific) would need to be considered whereas for IFRS 9, the detail required is much more granular across each asset.

Disclosures - Example of Required Movement Analysis

 

The tables in this section uses the following key

  • EL - Expected Loss

  • LEL – Lifetime Expected Loss

  • NCI Non-Credit Impaired

  • CI – Credit Impaired

  • SA – Simplified Approach

  • POCI – Purchased or Originated Credit Impaired

In addition to this, you have to show the drivers, for example balance movement across segments to align the two, some of which would have effected increases and decreases to provision across the portfolio.

Drivers for Change - Example Table

How do I get ready?

The next stage is really to consider whether you and your organisation have thought through all of the above and whether your plans fit the expectations of your auditors and are aligned to your peers. If the answer is no, the first thing to do is to formulate a plan, either internally or with the help of advisors.

Suggested Project Plan to Implement IFRS9

The process generally consists of four components:

  • A design phase, confirming scope, performing a gap analysis against current capability and the IFRS 9 standard. A methodology covering all the key elements of IFRS 9 and an appropriate approach for the organisation, considering the range of products and materiality and any supporting analysis. A design which utilises these findings and relates them back to actual data repositories within the organisation, together with proposed solutions.
  • A build phase, delivering to the design but sense checking along the way to ensure that components explained in the design perform as expected when developed in full. An engine build to tie all the elements together and perform checks and balances to ensure robust calculation.
  • A testing phase, developing scripts to test against, performing those checks on individual components and the solution overall and supporting UAT. This would be done just prior to parallel run.
  • A reporting phase based on all reports built based on design specifications and learnings from parallel run.

Why the different approaches?

Organisations are generally taking different approaches due to the complexity or size of the business or because of their reporting schedule.

As an example, a typical large high street banks, may well have a full service offering on their retail side, so have to devise solutions for Unsecured term loans, mortgages, overdrafts, credit cards and car leases. In addition to this, they may have different offerings for different customer segments, for example, sub-prime or High Net Worth (HNW) or legacy products as a result of an acquisition or merger. Whilst there may be an element of crossover, this all requires thought and analysis to determine the right level of segmentation and development. On the commercial side, there may be similar issues but these may be compounded by lack of data or lack of suitability to modelling, so they may have to define a more manual approach. Clearly the design and delivery is much simpler and requires less time for an organisation with only a retail mortgage offering for example.

The level of expectation is also different for different organisations. The recent guidance issued by Basel Committee on Banking Supervision (BCBS) 1 on Accounting for Expected Credit Losses refers to proportionality, and whilst it makes clear that this should not jeopardise a high quality implementation it should allow organisations to adopt approaches commensurate with the size and complexity of the organisation as well as the economic significance of it.

The reporting period also plays a part, IFRS 9 requires organisations to adopt in the reporting period after 1 January 2018, for organisations that are not due to report until year end until Q3 of 2018, some feel they have more time and therefore have a delayed start. Whilst this can be acceptable for a final deadline, it is certainly good practice to start the project sooner rather than later to ensure senior management have the right amount of time to consider all of the potential impacts.

What can I expect?

From experience there are a number of things common to every IFRS 9 project:

  • The solution you thought you would arrive with at the start is not the solution you end up with. This is a vastly complex challenge and whilst simplifications can be applied, they need to be relevant and justifiable for your organisation. This process takes a lot of thought and analysis and will sometimes prevent the organisation doing what they wanted to do initially or open up a whole new opportunity.
  • You will need some support. Even if you have enough people in your BAU team, it is unlikely they have all the requisite skills to get the right result for your organisation in a way that meets the standard. Whether that comes from specialist EL modelling experts or technical accountants, availability may be limited and needs to be planned in.

It’s an exciting challenge with lots of opportunities. Whilst this is a challenge it is a really exciting one, on every project I have been involved in there have always been a number of interesting debates that have lead to very interesting solutions

Transfer Criteria

Probably the fundamental change for IFRS 9 is the concept of ‘Significant Increase in Risk’ and I think it is fair to say it is the one area the industry is having the most difficult time in responding to. Credit Risk professionals tend to think in terms of absolute risk and managing that, rather than relative risk that the standard requires. One solution to this is to utilise internal current risk metrics and specific elements of the standard to develop tests to define significant movement.

First, to assess an initial range, one needs to be determined, this can either be done by judgement or analysis to show maximum movement from origination to default over a given time for individual segments. In any case, the actual range is not important but if there is a degree of confidence that the optimal value lies within the range it does make the process more efficient.

When these tests are applied to test data in isolation, it gives you a range to assess the movement to Stage 2 against.