Comments from ACCA to the International Accounting Standards Board, September 2009.
ACCA is pleased to comment on the above exposure draft, which was considered by ACCA's Financial Reporting Committee.
We understand the pressure for rapid changes to IFRS in the wake of the credit crisis but we have reservations about this piecemeal approach to the consultation on the revision of accounting standards. In particular, we are concerned that this paper does not articulate an IASB view on how the exposure draft fits into
Our general approach to the measurement basis for liabilities is as follows:
Many liabilities derive directly from transactions such as payables, deferred income and borrowings and measurement should reflect the transaction value. For other liabilities where there is no immediate transaction from which the consideration received can be derived (including non-contractual liabilities), initial measurement should be at the net present value of expected future cash flows (NPV), examples being tax, pension obligations and provisions.
Most contractual liabilities should be at amortised transaction value including those released according to the performance of the contract, not including the effect of changes in own credit risk. Derivatives with negative values should be at fair value. Non-contractual liabilities and others should continue at discounted expected future cash flows.
In addition we would agree that exceptionally there might be certain sorts of liabilities, in addition to derivatives, that could optionally be at fair value to avoid accounting mismatches or to replace hedge accounting.
We would see for subsequent measurement therefore three bases for measuring liabilities
Our response to the questions below reflects these three categories.
QUESTION 1: When a liability is first recognised should its measurement: (a) always, (b) sometimes, or (c) never incorporate the price of credit risk inherent in the liability?
As a general principle, which perhaps might suitably be included in the conceptual framework, we support (a) other than in exceptional and tightly defined circumstances.
As an example of exceptions, we note that the appropriate discount rate to apply to defined benefit pension liabilities has been a contentious issue on which there have been a range of answers suggested. The corporate bond discount rate in the current standard might be seen as a pragmatic compromise in such circumstances. For liabilities such as tax, it is not clear how or whether a credit risk effect might be reasonably incorporated.
We support (a) on the grounds that the measurement of liabilities should, in our view, reflect the transactions from which they originate as far as possible. Where these are contractual, the transaction prices will inevitably reflect and incorporate the counterparty's assessment of the credit risk of the entity or the particular instrument. For other liabilities where there is no immediate transaction from which the consideration received can be derived (including non-contractual liabilities), initial measurement should be at the present value of expected future cash flows, examples being pension obligations and provisions. The discount rates should in principle include the effect of credit risk by its incorporation in an incremental borrowing rate (among other factors) on the rationale that the entity might have to borrow to pay the liability.
QUESTION 2: Should current measurements following initial recognition (a) always, (b) sometimes or (c) never incorporate the price of credit risk inherent in the liability?
Subsequent measures of liabilities should generally not include any effects of changes in the credit risk. Our preferred measure for most contractual liabilities is amortised transaction value and thus including the original credit risk but not including the effect of changes in own credit risk.
As explained above we do not support subsequent measurement of liabilities at fair value except in the case of derivatives and other restricted circumstances. Fair value should include the effects of changes in own credit risk as that is part of the value which might be realised through trading. We think for fair values it is all but impossible to isolate the credit risk element and measure it separately.
Non-contractual liabilities and others should continue to be measured at discounted expected future cash flows and avoiding changes in own credit risk.
We are persuaded by the widely expressed concerns about counter-intuitive results and the realisation of gains or losses from changes in an entity's own credit standing. In particular, entities suffering from deteriorating credit standing might have difficulty arranging replacement borrowings in many cases. If they could raise finance, they would be paying a higher rate of interest meaning that the immediate “gain” would be at the cost of higher interest charges in future. As the paper notes, there would seem little incentive for those with improved credit standing to realise the loss.
QUESTION 3: How should the amount of a change in market interest rates attributable to the price of credit risk inherent in the liability be determined?
In our model this issue could only apply to liabilities measured at the present values of future cash flows. The discount rates to be used here are synthetic rates which are built up from components – estimates of a risk free rate, risk associated with variability of the cash flows and a premium for credit risk. The credit risk element might be derived from the entity's incremental borrowing rates. However as noted above in answer to Q2 we think that NPV measures should generally exclude changes in own credit risk.
QUESTION 4: The paper describes three approaches to liability measurement and credit standing. Which do you prefer and why?
We support option (c) in paragraph 62 of the staff paper, for the reasons we have given in answer to Q2 and Q3 above. We do not think that economic reality is reflected in charging a risk free rate of interest on all liabilities and an initial lump sum premium as in option (a). Option (b) is a fudge which is doing the same but is designed to obscure the impact. Different entities can borrow at different rates and we think these should be reflected in their interest charges and results over the period when the borrowing takes place.
This option should only apply to liabilities measured at NPV. Those measured at amortised cost would not include any changes in own credit risk. Those few measured at fair value should include the changes in own credit risk in profit or loss – we would not support a valuation basis which was ‘FV except for changes in own credit risk'.
Last updated: 12 Apr 2012