Beware the ripple effect of expected accounting losses
By Jane Fuller
FT, August 27 2009 03:00 | Last updated: August 27 2009
As preparers and users of accounts brace themselves for another regulatory onslaught this autumn, one thing they should not count on is an end to so-called "pro-cyclical" accounting.
Some believe the International Accounting Standards Board's proposed switch from an "incurred" loss model to an "expected" loss model will help smooth out peaks and troughs in bank profits over the economic cycle. But in responding to the IASB's "request for information" (deadline September 1) they should be careful what they wish for.
While factoring in expected losses from the start of a loan may top-slice profits in a boom, it will also condense the reporting of losses during the bust. That will intensify pressure on balance sheets.
It is worth remembering that, for all the fuss about "fair value" accounting, narrowly interpreted as "mark-to-market", it typically applies to less than half of bank assets. The incurred loss model for loans is a historic system that reports losses well after the cycle turns and we all start to "expect" that losses will mount.
This lag has been captured in exercises conducted by the International Monetary Fund and central banks estimating bank losses over the next couple of years. The question is not whether there is much more to come, but how many hundreds of billions of dollars. Witness, at last, the mounting loan loss provisions at German banks. The expected loss model would not wait until a loan has been stabbed, poisoned and shot, Rasputin-style, before a cut in value was recognised. But while some losses would have been anticipated, expectations are affected by the cycle. Changing sentiment at the onset of recession would cause them to be revised downwards and, hence, provisions to mount.
Resulting impairment charges would come hot on the heels of mark-to-market losses, which are leading indicators. This means that banks would have less time between waves of losses to recover and raise new capital from the private or public sector.
To cope with that, banks will have to carry higher capital buffers, as every regulator and commentator has suggested. Bank managements have bad form on this. In the EU, for instance, when IFRS was adopted in 2005, it was clear that balance sheets would get bigger and that results would be more volatile - with the inevitable multiplying effect on that bigger asset base. The obvious response should have been to build up a bigger equity cushion to absorb potential losses. On the contrary, profits were splashed out on pay, dividends and share buy-backs.
This is why prudential regulators are set to order additional capital buffers to cope with unexpected losses. US bank regulators have already done this under the Supervisory Capital Assessment Program. In the accounts, such buffers could be called "economic cycle reserves", an appropriation from after-tax profits that cannot be distributed to staff or shareholders.
But there is another way in which the expected loss model will continue the march against historical accounting. First, the trigger for loan impairments will be much more sensitive to current market sentiment. And second, the debate is being fuelled about what interest rate to use.
If the valuation employs the initial rate set, either fixed or variable according to a contractual formula, that is the one applied in the calculation of present value. But a fair value calculation would use the current market rate for that type of instrument. Such up-to-date valuations are shown in the notes under IFRS. The US standard-setter is inclined to put gains and losses in the "other comprehensive income" bucket, still excluded from earnings per share.
Controversial accounting changes have a habit of moving from the notes to OCI to the income statement. Those looking for accounts to be smoothed through the cycle (not this author) will find little comfort in the current proposals.
Jane Fuller is co-director of the CSFI and chairs the Accounting Advocacy Committee of CFA UK
