Both the FASB and IASB have published their proposals on the recognition and measurement of expected credit losses for financial instruments. The good news is that they are available for comment – we all see this as a step forward. But there is bad news as well. The proposals are not only different, but arguably both depart from the underlying economics in favour of operational shortcuts.
Neither proposal revealed any surprises compared to what I expected last Autumn. I mentioned then that the FASB model might be simpler from an operational perspective, but one has to question the conceptual merits of recognising impairment losses when an entity originates or purchases a financial asset at fair value. A ‘day one’ loss does not reflect the economics of a market-based transaction. Others have also questioned whether being forced to book a full expected loss on day one will act as a further constraint on bank lending. The IASB model recognises a smaller day one loss than the FASB model and, while still not conceptually pure, at least tries to acknowledge that credit risk is priced into the interest rate on debt instruments, and that the income to cover expected losses is earned over the life of the loans.
So which proposal is best? Both models aim to address the often-articulated criticism of the current ‘incurred loss’ model: ‘too little, too late’. The FASB goes further down that road than the IASB. However, both proposals compromise on operationality and conceptuality. The question is whether to accept that a conceptually pure answer is too difficult and choose one of the compromises so we can finish the project.
Digging deeper, the next question is whether there is any correlation between the current level of a bank’s reserves/capitalisation (banks are the most impacted by the proposals) and the economy it operates in, and its preference for a model. It seems that there is. While some banks expect that, under the IASB model, their impairment loss allowance will double, others argue that they would end up releasing loss allowances. Setting aside the conceptual question of how that might happen when moving from an incurred loss model to any iteration of an expected loss model, I have to admit that the impact of the proposal is highly dependent on current reserve levels and the economic situation.
The original IASB proposals contained the concepts to do the job better. Identifying the credit risk element of interest pricing, and setting it aside as a future loss reserve, would reflect the true economics of a transaction. But the proposed model was extremely complex, and so banking industry complaints were listened to. Previously, preparers, users and regulators had considered that convergence was extremely important. The question is now whether respondents believe that the urgency of the need to move to an expected loss model outweighs the drawback of having two versions.
What will happen after the comment periods end? The boards have said that they will share their feedback and decide on the next steps. Although many institutions (including the Financial Stability Board and the Basel Committee) have expressed their disappointment on the lack of convergence, a converged model is unlikely. I don’t expect either board would move to accept the other’s proposal.
Potential compromises might include taking a period longer than 12 months, but shorter than lifetime, to measure expected credit losses on initial recognition. But, during previous discussions, the boards considered and rejected the 24-month alternative. Another possibility would be to try again with a conceptual solution (for example, consider the lifetime expected credit losses at inception but, rather than recognise immediately in the profit or loss account, spread the expense over the instrument’s life), but this would add a further significant delay.
Whether we like this or not, we have until 31 May (FASB) and 5 July (IASB) to finalise our responses. As ever, I welcome your thoughts.