20 June 2013

FEE comments on financial instruments (expected credit losses)

FEE issued its comments on EFRAG's Draft Comment Letter on IASB ED Financial Instruments: Expected Credit Losses. FEE also provided its comments to Hans Hoogervorst, Chairman of the IASB.

FEE considers the proposed model as a good step forward in many respects, particularly the split of portfolios based on relative credit quality combined with the absolute threshold and the measurement approach to the “2nd and 3rd bucket”. FEE is seriously concerned regarding the measurement of the “bucket 1”, particularly due to the “day-one losses”.

FEE believes that this proposed expected loss model is more responsive to changes in credit conditions than the current IAS 39 incurred loss model. FEE acknowledges that the model lacks conceptual foundation in the bucket 1, particularly since it generates day-one losses and there is no conceptual basis in setting the 12-month expected loss period. However, in FEE's opinion the model introduces important practical solutions that are necessary to resolve some of the operational issues and implementation costs raised by many constituents, including FEE, in relation to the 2009 ED model and the subsequent Supplementary Document (SD).

In general, the model is considered to be an improvement to the 2009 ED model from a practicability point of view. However, FEE favours the development of an alternative model for “bucket 1” measurement, if possible, that would strike the right balance between:

a) a technically sound but almost impossible to implement model previously proposed in the 2009 ED, and

b) a model which provides a simple solution to operational challenges but lacks conceptual justification (pure pragmatic solution, which might be appreciated by preparers where the effects are less pervasive).

There should be appropriate consideration given to the costs and benefits model (as well as any other expected loss model) of a new basis for estimating credit losses. Although FEE is not proposing to keep the current incurred loss model, FEE wishes to highlight the fact that the proposed expected loss model is inherently more subjective and comparability among reporting entities will become an issue, particularly for long-term assets. An expected loss model required by this ED would be more subjective and very costly to implement for large and small banks and insurers, yet would not, in FEE´s view, provide better quality information for users due to the approach to measurement of bucket 1 assets.

FEE feels strongly that a new standard should be more principle-based, allowing preparers to have greater flexibility to implement the requirements in a cost efficient way. As FEE states above, the 2009 ED was more technically sound but the cost of implementation prohibitive. In light of the highly judgemental nature of impairment accounting, FEE believes there is a solution that has the technical soundness of the 2009 ED yet is scalable, more pragmatic and less costly to implement.

FEE shares those arguments expressed by Mr Cooper, where he disagrees with the establishment of the 12-month expected credit loss allowance, particularly for the period immediately succeeding the origination of a loan. The recognition of this loss allowance would fail to reflect the economics of lending activities and is likely to mislead users. The fact that an asset acquired at fair value is immediately mandatorily remeasured down to reflect losses that are compensated by contractually fixed and probable interest income accrued over the life of the instrument does not faithfully represent reality. When determining the consideration paid, it can be assumed that the creditor takes into account any knowledge he has about the debtor‘s credit risk. In fact, the model systematically reduces the asset returns in the first 12 month period and systematically overstates these returns in the last 12 months to maturity.

A balanced solution would be somewhere between the originally proposed models which produce meaningful profit and loss numbers and the current more practical solution which, however, would reflect the key general accounting principles contained in the Framework. Therefore, FEE would like to see an amended model that would deal with the day-one loss for “bucket 1” items and provide faithful representation of the income statement effects. FEE thinks that the BC25 examples combined with some practical simplifications, might achieve this. Once the income statement effects are solved, there may be more approaches allowed for preparers so that they can find the most relevant and cost effective way of implementation.

The results of the field test that EFRAG is currently conducting and the input received from the banking industry could be utilised in developing such a solution for “bucket 1” items.

Similar to EFRAG, FEE certainly does not support the FASB model, particularly since it produces significant day-one losses and front-loads losses before the related income is generated. This lacks any theoretical fundaments in a framework promoting faithful representation and neutrality. Furthermore, FEE is of the view that such a model would negatively and unfairly affect the access to finance by the European economy, which largely relies on debt and particularly bank funding.

Press release

Comment letter to EFRAG

Comment letter to IASB

 

© FEE