An Overview of IFRS 9
Summary
IFRS 9 is a major upgrade to the accounting for financial instruments previously governed by IAS 39. Its key elements are;
1) New rules on the classification, valuation and revenue recognition of financial assets and liabilities with new restrictions on when amortised cost and accrual recognition may be used.
2) A major revision to the determination of impairment moving from an incurred loss model to an Expected Credit Loss model.
3) Moving the recognition of Own Credit Adjustment (OCA) changes from profit and loss to Other Comprehensive Income.
4) Improvements to the rules around Hedge Accounting, increasing both the range of instruments that can be hedged as well as the permitted hedging practices. The effectiveness test is more graduated and less "Pass/Fail" and should allow hedge accounting to better follow real commercial hedging.
The standard is applied to financial reporting periods starting on or after 1 January 2018, although firms can continue to apply the hedge accounting rules from IAS 39.
The new ECL impairment rules are receiving the most publicity because;
1) Every asset not held at fair value will have to have some level of ECL held against it and this is will reduce bank capital. An EBA survey predicted an average fall of 79bp in Tier 1 Capital.
2) Implementing ECL is requiring significant credit loss modelling development and as usual the industry has requested more time, which it is unlikely to get.
3) The mechanics of the categorisation and calculation may cause surprising volatility in impairment reserves, a significant increase in default risk will not only increase the ECL it may move the calculation from a twelve month horizon to a full life reserve magnifying the original increase's impact.
In practice any firm with sizeable financial assets which will require an ECL needs to be in full population test mode now to be ready for next year. Many firms are planning to run their ECLs either monthly or quarterly but will need to run them every time they publish a balance sheet.
Background
It's the curse of accounting
standard setters that they generally only get criticised for their work with
commentators unable to recall the problems with the previous rules. Banks entered the Credit Crunch with
extremely low impairment reserves as the old standard, in an effort to stop
firms from using overly subjective loss estimates pushed banks to only reserve
for impairment that had been observed i.e. through a debtor failing to
pay. Thus the banks entered a period of
significant credit losses with very low level of reserves as the previous
economic boom had reduced actual defaults.
Similarly, the standards included a valuation adjustment for the impact
on the fair value of an issued liability caused by a bank's own credit risk
existed but was generally not included on the grounds that it was immaterial
while banks' creditworthiness was not in doubt.
When the crisis hit and own credit became material it caused banks to
post large quarterly OCA profits which was totally counter-intuitive. Lehman Brothers would recognise hundreds of
millions of dollars of profits as its issued bonds collapsed in value, but
could never have realised any of the profit as it had zero cash and became
insolvent. OCA gains are excluded from
The new standard is trying to address all these points. It is bringing in forward looking credit loss estimation. OCA is now only an OCI movement much as say the revaluation of a company's property changes its capital base but does not have any contribution to its profits from its operations. Hedge accounting is more flexible although this is still work in progress as the standard setters try to devise workable rules that cover entities' hedging of open (i.e. dynamic) portfolios with a variety of hedging tools.
Classification Valuation and Revenue
Recognition
Revenue recognition is now driven by two factors;
Which business model is being used and the equity/debt nature of the product;
Financial Assets
|
|
Held for Trading? |
Only Principal and
Interest? |
Only Collect
Contractual Cashflows? |
Collect Cashflows
and also sell? |
Election of OCI? |
Initial Recognition |
Disposal Recognition |
|
Equity Instruments |
No |
|
|
|
Yes |
OCI |
OCI |
|
No |
Fair Value through P&L |
|
|||||
|
Yes |
|
|
|
|
|||
|
Debt Instruments |
Yes |
||||||
|
No |
No |
||||||
|
Yes |
No |
No |
|||||
|
Yes |
|
OCI |
P&L |
||||
|
Yes |
|
|
Amortised Cost, Accrual through P&L |
|
There is also a "FVTPL" option which allows any asset to be designated and fair value and reported through trading P&L if it reduces or eliminates an accounting mismatch.
Financial Liabilities
|
Held for Trading? |
Credit Risk
Component with FVTPL |
Embedded Derivative |
Separable Embedded
Derivative? |
Initial Recognition |
|
Yes |
|
|
|
Fair Value through P&L |
|
No |
No |
Yes |
Yes - Derivative |
|
|
Yes - Host |
Amortised Cost, Accrual through P&L |
|||
|
No |
|
|||
|
Yes |
|
|
Fair Value in OCI |
Specifically;
· Trading positions with be FVTPL
· Equity investments will go through OCI with no P&L impact on disposal or de-recognition.
· Only contractual collection of principal and interest can be recognised in P&L under the accrual method
· Debt with a separable embedded derivative must be bifurcated with the debt amortised while the derivative is fair valued.
· Non trading liabilities with fair valued credit risk must report this component in OCI (i.e. OCA).
All financial assets must have an
ECL computed. If there has been no significant
deterioration in the probability of default then a 12 month horizon may be
used. If there has been a significant
deterioration then a lifetime horizon must be used. The rules over expected credit losses have a
number of similarities to the
|
|
IFRS 9 |
|
|
|
|
No Significant Deterioration since initial recognition |
A Significant Deterioration in the risk of default. |
|
|
Risk horizon |
12 Months |
Lifetime |
12 Months |
|
Credit risk assessment |
Balance sheet point in time. |
Through the cycle |
|
|
Loss Given Default |
Balanced |
Downturn bias |
|
Only a change in the risk of default triggers the change in risk horizon, so firms are not required to consider other factors such as loss given default or exposure at default. This is somewhat counter-intuitive but can be justified in an attempt to keep the rules simpler. Obviously, limiting the assessment to only 12 months will be much easier to implement.
Although a conservative bias is not required simplistic methodologies such as flat percentage loss estimations and reserving of interest due are not permitted unless they are equal to formal computations of the ECL. The ECL should reflect the time value of money and also be probability weighted so a modelled solution at some level is inevitable. At the very least this will be the probability that a loss occurs and the probability that no loss occurs although it's very likely that much more sophisticated scenarios will be needed for the larger asset pools. A firm is only required to use information which available without undue cost or effort. The standard includes a rebuttable position that an amount overdue by thirty days or more is evidence of a significant increase in the default risk. If there is evidence of a significant reduction in the default risk a firm may move the asset's ECL calculation back to the 12 month horizon.
Banks are anticipating some volatility in impairment as an asset that moves to the Lifetime basis of calculation will have experienced an increase in its 12 month ECL and this will be compounded by the need to compute the loss over its full lifetime. There is some talk of firms adding a conservative bias to 12 month methodologies to try and reduce the impact of this but this is precisely the subjective gaming of the system that standard setters are trying to limit. I think the concerns of excessive ECL volatility need to be viewed in the context of the old rules where some banks reported impairment increases of more than $10Bn and 300% in a single year.
From the bank's perspective only real use that OCA has is to act as a governor (like on a steam engine) on profit and loss during a serious downturn where trading losses are counteracted by OCA profits as the banks' credit spreads widen and the value of their issued liabilities decreases ("debit issued liabilities, credit profit"). But this is illusory as such gains are extremely hard for banks to realise as they would need to be able to buy back their own debt during a period when their cash resources are likely to be under pressure. IFRS 9 resolves this by maintaining the requirement but only as a movement through OCI. Now that banks are much more strongly capitalised than they were going into the crisis their OCA changes have reduced from the hundreds of millions of dollars seen quarterly during the crisis. Some banks are describing the change as de minimis again.
Given the enormity of the new ECL regime I think many banks have de-prioritised making any changes to their hedge accounting during 2017 for 2018. Also until the rules for macro open hedging are finalised it is difficult for banks to assess the costs and benefits of extending beyond their current practices. That said there are a number of improvements which they can gain from IFRS 9.
Hedges still need to be formally identified documented at inception and there must be some economic basis to the hedge. The hedged item can be a solo instrument or a group but it must be defined and documented specifically. Similarly a specific component of the hedged item such as its interest rate risk can be designated the hedged item provided it can be reliably isolated even if the item is not a financial instrument which will assist corporate users of hedge accounting. The 80-125% test is no longer applied as a pass/fail threshold instead such tests identify the ineffective component which will be reported. The value changes cannot be dominated by the credit risk in the instrument. Hedging can be applied cumulatively, so that a core instrument can have say its FX risk hedged by one hedge and then the resultant aggregated position can be hedged by a second hedge such as an interest rate product. This was not possible with derivatives previously. There are three types of hedging under the standard;
· Fair value hedges where an element of the fair value is identified for hedging.
· Cashflow hedges where specific cashflows such as interest payments for the element to be hedged.
· Hedging the net FX risk in a foreign operation, so that the investment's OCI return can be FX neutral.
Derivatives can be designated as hedges and it is permissible to split the intrinsic value from the time value of an option to only designate the intrinsic value changes as the hedging component, similarly the spot impact of a forward FX contract can be split out. Otherwise the entire instrument must be designated although a percentage of the total can be apportioned as the hedge. Non derivatives such as FX positions can be designated as hedges however instruments whose credit risk related changes in fair value are reported through OCI (see above) cannot be designated as hedges.
Hedges can be rolled into new contracts. The gains or losses from the effective hedge is reported in profit and loss or OCI to follow the treatment of the item for which it is the hedge. Ineffective hedge gains and losses are reported in profit and loss.
A new term called "Rebalancing" is introduced which allows an entity to change the hedging ratio to ensure that a hedge continues to be effective. This was not permitted previously and will allow the accounting to follow commercial practice better.