This Comment Letter was sent by BDO Global Coordination B.V. on behalf of BDO International, to the International Accounting Standards Board on 31 July 2009:
Dear Sir,
Exposure Draft (ED/2009/03): Derecognition – Proposed amendments to IAS 39 and IFRS 7
We are pleased to have the opportunity to comment on the above exposure draft issued by the International Accounting Standards Board (IASB), on behalf of BDO International1.
We are supportive of the IASB in the action it has taken as a result of the global financial crisis, and of the acceleration of key projects in response to changing conditions and the needs of constituents. We believe that the IASB has acted on a timely basis in addressing a number of urgent issues.
However, while we appreciate that concerns have been raised about derecognition (in particular that certain items may inappropriately have qualified for derecognition), we are not convinced that the requirements of IAS 39 are flawed to the extent that significant change is needed now. While we acknowledge that the existing standards are not based on a single conceptual approach, the accounting that arises from their application is generally appropriate. We note that the proposed guidance would be likely to increase the number of financial assets that qualify for derecognition, and not reduce that number.
Consideration of risks and rewards
A fundamental part of the existing guidance in IAS 39 is consideration of risks and rewards. In practice, we have found that this test has been effective in ensuring that financial assets that should remain on balance sheet are accounted for in that way.
As with our response to ED10 Consolidated Financial Statements, we believe that in addition to a control test, a complementary (second) risks and rewards test is required in order to arrive at an appropriate accounting outcome. Without this, as the Board has identified, certain transactions (in particular securities repurchases or borrowing/lending transactions involving financial assets that are readily obtainable on a market) would qualify for derecognition where the economic reality is that of a secured loan.
We believe that the analysis in the exposure draft significantly overstates the issues that arise in practice from the existing risks and rewards test which is, in fact, an extremely useful tool in preventing inappropriate off balance sheet accounting. While it may require a degree of judgement in its application, that is no different from a number of other aspects of IFRS, and it would be expected that some thought would be required in dealing with complex arrangements that are aimed at achieving derecognition.
The control model
Both ED10 and the derecognition proposals are based primarily on the notion of control, with risks and rewards having been given substantially less prominence. We are concerned that such a fundamental change in approach is being proposed without an opportunity having been given for a full analysis, including a discussion paper. We would encourage the Board to retain a substantial element of risk and reward analysis for both consolidation and derecognition at this stage (in particular because this test has been found to be highly effective in preventing bad practice and ensuring an appropriate accounting approach is followed), with a more considered review then being carried out of whether the appropriate model is control, risks and rewards, or a combination of the two.
Our suggested approach at this stage
We fully understand that the exposure draft has been developed on a fast track basis, in response to concerns that have been raised about off balance sheet accounting, and that the IASB has discussed the project with the FASB on a number of occasions. However, due to the concerns that we have set out in this comment letter and its appendix, we believe that the IASB should restrict any changes to be introduced to IAS 39 in the immediate term to enhanced disclosures, as set out in the exposure draft, which we believe would enhance the information available to users of financial statements.
Convergence of IFRS and US GAAP, and the alternative approach set out in the exposure draft
If new accounting requirements are to be introduced (in particular where, in our view, the existing model operates effectively which is the case for the derecognition requirements of IAS 39), we believe that it would be appropriate to converge the requirements of IFRS and US GAAP. We note that the proposed changes to IFRS would not achieve this objective.
We therefore suggest that, in addition to restricting changes to IFRS to disclosures at this stage, the IASB and FASB work on a joint project to identify an appropriate converged solution.
As noted in our response to the detailed questions raised in the exposure draft, we believe that there may be merit in the alternative views of certain Board members. It may be appropriate for the two Boards to explore this approach further as well as the two stage approach that we have outlined above in the section ‘Consideration of risks and rewards’.
Our detailed responses to the questions you have raised in the Exposure Draft are set out in the attached Appendix.
We hope that our comments and suggestions are helpful. Should you wish to discuss any of them further, please contact either Andrew Buchanan at +44 (0)20 7893 3300 or Tracey-Lee Massey at +32 2 778 0130.
Yours faithfully,
BDO Global Coordination B.V.
Appendix
Question 1 - Assessment of ‘the Asset’ and ‘continuing involvement at reporting entity level
Do you agree that the determination of the item (ie the Asset) to be evaluated for derecognition and the assessment of continuing involvement should be made at the level of the reporting entity (see paragraph 15A, AG37A and AG47A)? If not, why? What would you propose instead, and why?
Yes.
Question 2 - Determination of ‘the Asset’ to be assessed for derecognition
Do you agree with the criteria proposed in paragraph 16A for what qualifies as the item (ie the Asset) to be assessed for derecognition? If not, why? What criteria would you propose instead?
We agree. While the approach is rules based, it has been shown to be operational in practice and has not led to inappropriate accounting results.
We note the alternative views of certain Board members and believe that there may be merit in the suggested approach. As noted in our covering letter, we believe that an appropriate approach may be for the IASB and FASB jointly to explore this approach further, as part of a project aimed at achieving a converged approach. However, we believe that significant difficulties could arise in practice under the alternative view, in particular the valuation of a retained (disproportionate) interest.
While paragraph 16A is noted as bringing the existing requirements of IAS 39 into the proposed amendments, this does not appear to be the case. In particular, we note the newly introduced requirement, that ‘the performance of the part retained does not depend on the performance of the part transferred’ and believe that this may give rise to difficulties in practice. As an example, assume that an entity transfers the interest strip of a financial asset to a third party. If the debtor defaults, or is capable of prepaying the capital element of the financial asset, then there would appear to be interdependency between the capital and interest elements that would result in derecognition of the interest strip being precluded. We doubt that this is the intention of the Board, and suggest that either further guidance is included in any final amendments, or that this new language is deleted.
We note that where a transfer is of a group of financial assets, paragraph AG42A requires that the assets are evaluated for derecognition as a group. It is not entirely clear why a different approach should be applied to a group of financial assets; we consider that it is likely always to be appropriate to analyse derecognition on an individual asset basis. We note that where a transferor sells a portfolio of originated loans and provides a ‘first loss’ guarantee on a portfolio basis, the effect is that the transferor retains an interest in all of the assets within the portfolio and, in consequence, the risk of loss on each of them.
Question 3 - Definition of ‘transfer’
Do you agree with the definition of a transfer proposed in paragraph 9? If not, why? How would you propose to amend the definition instead, and why?
While we agree that it is appropriate to clarify the meaning of ‘transfer’, we believe that the proposed definition is too wide and that it may lead to derecognition in circumstances where this would not be an appropriate outcome.
Introductory paragraph IN10 notes that although many of the derecognition outcomes of the revisions would be similar to those arising under the current standard, there are exceptions which include repurchase agreements involving readily obtainable assets. The total return swap examples in paragraph AG52L(h) illustrate that the ability of a transferee to sell a readily obtainable asset is sufficient in itself to pass the transfer test, even where all of the economic effects of ownership of the asset are retained by the transferor. We believe that this is an inappropriate result, as the proposed amendments would appear to indicate that a legal form analysis is more important than actual economic exposure.
We do not believe that the ability of a transferee to sell a readily obtainable asset it acquires in a repurchase transaction should, in itself, permit derecognition by the transferor. Instead, we believe that lending and repurchase arrangements, where the transferor retains economic exposure to changes in the value of the asset(s) subject to the lending and repurchase arrangement, the analysis should result in the transfer conditions not being met.
In contrast, the example in paragraph AG52L(f) does focus on economic benefits arising from the 10% interest in entity B (and notes that there is no transfer because the issuer of the note is not obliged to remit all of the economic benefits of the investment in Entity B).
While the proposed revisions to IAS 39 deal with financial assets and liabilities, the Board should consider wider issues involving non financial assets. For example, it is common for a series of ‘single property SPEs’ to be set up by an investment property company, where each property is financed by a ring fenced loan which has recourse only to the property that it has financed. It may be that derecognition of the loan and the property would be precluded because the lender would have exposure to changes in the property value only it that value falls; if the value increases, the SPE is required only to repay the capital amount borrowed. However, what if the value of the property has fallen to the extent that it is considered remote that the value will ever recover to the extent needed to repay the loan in full? In such a situation it could be argued that the lender is exposed to all the risk and rewards that are expected to occur in practice.
Question 4 - Determination of ‘continuing involvement’
Do you agree with the ‘continuing involvement’ filter proposed in paragraph 17A(b) and also the exceptions made to ‘continuing involvement’ in paragraph 18A? If not, why? What would you propose instead, and why?
The effect of the filter as proposed would appear to be that a risks and rewards test in a transferred asset is (to a limited extent) retained. This is demonstrated, in particular, by the exclusion in paragraph 18A(c) for assets to be reacquired at fair value (we presume that this is intended to be fair value at the reacquisition date – this should be made clear in any final guidance).
As we are strongly of the view that a risks and rewards test is a necessary feature in an assessment for derecognition, we agree that such a filter should be included. However, if (as would seem to be the case), the Board intends that insignificant continuing involvement would be disregarded, this would seem little different from the existing test of whether substantially all the risks and rewards have been transferred. We therefore see no rationale for changing the approach from the risks and rewards test in the existing guidance which we have not found to cause difficulties in practice. In fact, the risks and rewards test has been effective in ensuring that arrangements are kept on balance sheet as appropriate.
Question 5 - ‘Practical ability to transfer for own benefit’ test
Do you agree with the proposed ‘practical ability to transfer’ derecognition test in paragraph 17A(c)? If not, why? What would you propose instead, and why?
(Note: Other that the ‘for the transferee’s own benefit’ supplement, the ‘practical ability to transfer’ test proposed in paragraph 17A(c) is the same as the control test in IAS 39.)
Do you agree with the ‘for the transferee’s own benefit’ test proposed as part of the ‘practical ability to transfer’ test in paragraph 17A(c)? If not, why? What would you propose instead, and why?
We do not agree with the proposal that the transferee’s practical ability to transfer an asset should be the primary determinant in considering whether control has been transferred. While the test in paragraph 17A(c) is very similar to the existing control test in IAS 39, that test assumes considerably greater importance in the proposed model due to the substantially reduced focus on risks and rewards.
We note, and agree with, the comments in paragraph BC20, that the economic benefits to be derived from a financial asset do not arise solely from its sale; they may also be derived from the contractual cash flows and it is the control over those economic benefits that is important. Consequently, we believe that if the ‘practical ability to transfer’ test is to be incorporated into a revised standard, then what is meant by ‘for the transferee’s own benefit’ needs to be analysed further, and widened.
Rather than the changes proposed in the exposure draft, we believe that a more fundamental debate is needed about the meaning of control, whether risks and rewards should be included in the assessment of control on the basis that the more exposure an entity has to gains and losses relating to a financial instrument, the more likely it is that the entity controls that financial instrument, and whether a focussed risks and reward approach would in fact result in a better representation of those rights and obligations that are retained, and those that are transferred, in a transaction that may qualify for derecognition.
We note that elements of a risks and reward approach are included in the Application Guidance. For example, the wording in brackets at the end of paragraph AG52A suggests that it is necessary to assess the extent to which the transferee is exposed to the economic consequences of holding the asset. However, as drafted the wording is not entirely clear, as the use of ‘full economic benefits’ suggests a broader test than just the proceeds from sale.
Question 6 - Accounting for retained interests
Do you agree with the proposed accounting (both recognition and measurement) for an interest retained in a financial asset or a group of financial assets in a transfer that qualifies for derecognition (for a retained interest in a financial asset or a group of financial assets, see paragraph 21A; for an interest in a financial asset or group of financial assets retained indirectly through an entity, see paragraph 22A)? If not, why? What would you propose instead, and why?
We do not agree with the proposals, as we believe that for paragraph 21A they are more complex than is necessary, and for paragraph 22A they could give rise to results that do not reflect the economics of the underlying transactions.
We note that paragraph 16A refers to the unit of account being part of a financial asset or group of financial assets that can specifically be identified. That being the case, we see no need for the complex calculation that paragraph 21A requires, and would instead simply derecognise the appropriate part of the carrying value of the entire asset. The gain or loss would then be a function of the proceeds received/receivable and the carrying value of the part of the financial asset(s) to be derecognised. If the cash flows of the portion being considered for derecognition cannot be identified in the way that paragraph 16A requires, then derecognition would be prohibited and there would be no need for any allocation.
We believe that the proposals in paragraph 22A would benefit from further analysis. While the proposed approach might be operational and appropriate in the case of an SPE set up for the purposes of a specific transaction for a single entity, or for the transfer of ring fenced debt instruments, the position may be different for other transactions.
As an example, Entity A might transfer a wholly owned subsidiary, B, to Entity C in return for a debt instrument issued by C. Entity C has a wide range of financial assets which generate income, in addition to the shares in Entity B. As a consequence, Entity C’s ability to service the debt instrument issued to Entity A bears little relation to the dividends (if any) that might be received from Entity B. In that case, it would seem inappropriate for Entity A to continue to recognise an investment in Entity B as a financial asset instead of the debt instrument that it actually has. We suggest that additional guidance might be added such that, where the original asset is to continue to be recognised, there would need to be a clear economic link between the asset(s) transferred and the income to be received in return by the transferor.
Question 7 - Approach to derecognition of financial assets
Having gone through the steps/tests of the proposed approach to derecognition of financial assets (Questions 1 – 6), do you agree that the proposed approach as a whole should be established as the new approach for determining the derecognition of financial assets? If not, why? Do you believe that the alternative approach set out in the alternative views should be established as the new derecognition approach instead, and, if so, why? What alternative approach would you propose instead, and why?
No, we do not agree that the proposed approach should be established as the new approach for determining the derecognition of financial assets.
In addition, we do not believe that the risks and rewards test in the existing guidance has given rise to significant difficulties in practice. We disagree with the Board’s conclusion in this respect. Instead, the risks and rewards test has been fundamentally important in providing an appropriate mechanism to keep on balance sheet those arrangements which belong there.
Further, we do not believe that the proposed approach would necessarily improve financial reporting and, from the perspective of securities repurchases or borrowing/lending transactions involving financial assets that are readily obtainable on a market, the quality of financial reporting might be viewed as deteriorating. We do not believe that those transactions are faithfully represented in financial statements by derecognising the financial assets transferred when, in reality, the transferee receives no more or less than a lender’s return on the funds (temporarily) advanced (ie a lender’s return on a loan).
As noted in our covering letter, we believe that it would be appropriate for the IASB and the FASB to work on a joint project to identify an appropriate converged solution for derecognition. The alternative approach set out in the alternative views to the exposure draft would merit further analysis, in addition to further consideration of our suggested two stage test which would combine control, and risks and rewards.
Question 8 - Interaction between consolidation and derecognition
In December 2008, the Board issued an exposure draft ED10 Consolidated Financial Statements. As noted in paragraph BC28 and BC29, the Board believes that its proposed approach to derecognition of financial assets in this exposure draft is similar to the approach proposed in ED10 (albeit derecognition is applied at the level of assets and liabilities, whereas consolidation is assessed at the entity level). Do you agree that the proposed derecognition and consolidation approaches are compatible? If not, why?
Should the Board consider any other aspects of the proposed approaches to derecognition and consolidation before it finalises the exposure drafts? If so, which ones and why? If the Board were to consider adopting the alternative approach, do you believe that approach would be compatible with the proposed consolidation approach?
We do not agree that the proposed approaches for derecognition and consolidation are fully compatible. Although in both cases the emphasis placed on an analysis of risks and rewards has been diluted in comparison with current literature, the extent to which this has been diluted appears different.
In ED10, both control and risks and rewards appeared to be required to be considered (although, due to certain inconsistencies in the exposure draft, the extent to which risks and rewards were in fact to be considered was not clear). Our response to that exposure draft noted that, while the Board had explicitly rejected the risks and rewards model as a basis for consolidation on the grounds that it was not conceptually robust (BC33), the exposure draft itself noted at paragraph 12 that ‘when assessing control, a reporting entity shall consider power and returns together, and how a reporting entity can use its power to affect the returns. This linked to paragraph BC53, in which it was noted that ‘….the Board’s assumption is that the entity that receives the greatest returns from another entity is likely to have the greatest power over that entity.’
While the derecognition exposure draft does, implicitly, include consideration of risks and rewards (through the test in paragraph 17A(c)), these are combined with the control test in a rather different way.
We believe that the existing requirements set out in the current IAS 39 and SIC-12 are compatible as both place significant emphasis on an analysis of risks and rewards, which are defined in the same way in both sets of guidance. We are concerned that the dilution of the importance of risks and rewards in the proposals for both consolidation and derecognition will result in that symmetry being reduced or lost. The extent to which that symmetry may be reduced or lost is not clear, because neither ED10 nor the derecognition proposals are clear in how the interaction of control, and risks and rewards, will in fact remain.
Question 9 - Derecognition of financial liabilities
Do you agree with the proposed amendments to the principle for derecognition of financial liabilities in paragraph 39A? If not, why? How would you propose to amend that principle instead, and why?
While we do not object to the proposed changes, we find it difficult to understand why the Board feels that there is a need to change guidance which is well understood, applied consistently in practice, is not viewed as being contentious, and results in an appropriate accounting result.
If the Board feels that it should change the guidance, we have the following comments:
- Paragraph 39A refers to ‘….a financial liability (or a part of it)……’. It would be helpful for this guidance to be clarified so that the identification process for financial liabilities to be derecognised (whether the entire instrument or a part of it) is consistent with that for financial assets.
- It would be helpful for paragraph 40A to include additional guidance to cover circumstances where one of the parties makes a cash payment as part of the debt renegotiation.
- Paragraph 42B notes that ‘…if an entity derecognises a financial liability as a result of an exchange of debt instruments or modification of terms, it includes any costs or fees incurred in the gain or loss recognised’. However, this does not identify which costs or fees – are these the unamortised costs and fees that were incurred in respect of the original debt, or is it both those unamortised amounts and any costs or fees incurred in respect of the new financial liability? If the latter, it would be helpful for the Board to explain how this links to the requirement for financial instruments measured at amortised cost to be measured initially at fair value, net of transaction costs.
Question 10 - Transition
Do you agree with the proposed amendments to the transition guidance in paragraph 106 and 107? If not, why? How would you propose to amend that guidance instead, and why?
We agree that it is appropriate to require prospective application, without a need for retrospective adjustments.
Where an entity considers early adoption, it would be helpful to clarify whether this is available only from the start of each earlier annual accounting period, or whether the guidance can be adopted from any date prior to the date of mandatory application.
Question 11 - Disclosures
Do you agree with the proposed amendments to IFRS 7? If not, why? How would you propose to amend those requirements instead, and why?
We support the disclosures which have been proposed, and consider that they achieve an appropriate balance in providing sufficient (but not excessive) information to users of financial statements to enable them to understand the effects of transfers of financial assets (whether derecognised or not).
1 BDO International is a world wide network of public accounting firms, called BDO Member Firms, serving international clients. Each BDO Member Firm is an independent legal entity in its own country. The network is coordinated by BDO Global Coordination B.V., incorporated in the Netherlands, with an office in Brussels, Belgium, where the International Executive Office is located.
Exposure Draft ED/2009/3 of April 2009: Derecognition - Proposed amendments to IAS 39 and IFRS 7, Comments due by 31 July 2009:
ED-2009-3 Derecognition.pdf