Continued from Part I.
The idea that banks arbitraged accounting categories to avoid mark-to-market losses as credit markets seized up in 2008 and 2009, won’t exactly be surprising. Since 1996, financials have been allowed to choose whether to put assets in their banking book — held at amortised cost — or in their trading books, which is marked to market. The decision isn’t entirely up to them — they have to show an intent to actually sell or trade — but nevertheless arbitrage opportunities between accounting books exist.
FT Alphaville wrote back in April of a paper that found (US) banks used their accounting discretion, to lower mark-to-market losses as much as possible. Citi, for instance, reclassified about $60bn of debt securities from held-for-sale (fair value) to held–to-maturity in late 2008.
It’s nice then, to see the FSA capturing that issue:
Basing the boundary on ‘trading intent’ is flawed. In buoyant markets, firms demonstrated trading intent for a wide range of positions. In periods of market stress, the inability of the trading book framework to adequately capture the risk on these positions has been exposed – in particular, the ability to hedge and/or exit these positions within a short time horizon was undermined.
Indeed, the below chart from the paper shows the split of bank losses, by value, across regulatory books. About 22 per cent of losses were on banking book positions, and 77 per cent came from the trading book. However, the below table shows the extent to which loss-making positions moved between the (marked-to-market) trading book and the (held at amortised cost) banking book during the crisis. Some 26 per cent of all losses, by value, were in positions that switched from trading book to banking book during the period analysed, and there was ‘no evidence’ of any loss-making positions moving from the banking book to the trading book during the period.
According to the FSA:
The data highlights a disconnect between the fair value boundary in the accounting framework and the trading book boundary in the regulatory framework. Table 5.4 shows that around 30% of the total losses analysed were incurred in positions that were held at fair value but that were in the regulatory banking book. These positions lost value due to market movements which will directly affect capital resources, but these positions were not subject to a market risk capital charge – this represents a significant gap in the regulatory framework.
No kidding.
To fix this particular issue, the FSA is advocating a “consistent approach” across all credit positions, and an an “increased focus on valuing traded positions as an input into capital resources.”
But valuation is also a tricky subject — on then, to Part III.
Related links:
From Level I to Level III, the myth of fair value – FT Alphaville
Accounting for European stress – FT Alphaville