The EBA welcomes the joint effort by IASB and FASB in developing a common impairment model that provides a more forward-looking approach. In particular, the move towards a converged expected credit loss (EL) approach for impairment will improve the decision usefulness and relevance of financial reporting for users, including prudential regulators.

Like the original Exposure Draft (ED), the Supplementary Document (SD) appropriately continues to address one of the important weaknesses that has been identified with respect to the current impairment models under IFRSs and US GAAP, which is delayed recognition of credit losses associated with financial assets (i.e. the too little too late problem).

The EBA supports the time-proportional approach to recording expected credit losses for the reasons explained by the Board. The EBA recognises that the proposed approach incorporates a broader range of credit information than under IAS 39 and is more likely to draw from banks' risk and capital management (e.g. with the introduction of the distinction between good and bad loan books). The EBA also acknowledges that the proposed model, including decoupling of expected credit losses from the effective interest rate, is likely to reduce the operational difficulties of the original IASB proposal.

Further simplification could actually be reached. As indicated in EBA´s 30 June 2010 comment letter on the original ED, another source of complexity was that the proposed approach requires estimates of the amounts and timings of expected cash flows to be probability-weighted possible outcomes. To simplify this, entities should be enabled to use more widely average loss rates, in line with their risk management systems (so not only in the case indicated in Para B7).

The EBA notes that the proposed model does not satisfactorily address the issue of procyclicality raised in their 30 June 2010 comment letter on the original ED. As indicated in that letter, estimates may have to be revised frequently depending on management’s ability to forecast changes in credit risk factors accurately. This potentially gives rise to pro-cyclical effects and therefore remains a source of concern under the supplementary ED model. The EBA would appreciate having a full picture of the future impairment model and information on future steps before the finalisation of the standard.

The EBA has in general a preference for a single impairment model - based on the supplementary ED's general approach - for all financial instruments carried at amortised cost. Considerations for simplifications (e.g. for quoted debt securities) should be consistent with the principles as those for the general approach for open portfolios.

Finally, the EBA does not support the proposed free choice regarding discounting. The EBA would rather support a model where discounting is applied where relevant and likely to result in reliable information (i.e. essentially for the bad book). Regarding the discount rate, the EBA believes that using the effective interest rate of the asset is the only reasonable solution in an amortised cost model.

Full paper