The rate cut and Knightian uncertainty

Here's a paradox. On the day the Bank of England cuts base rates because "conditions in financial markets have deteriorated", the Royal Bank of Scotland announces that its losses on sub-prime mortgages have been a mere £1.25bn - loose change for a bank that size.
Which raises the question. Why have such trivial losses - Barclays were similar - caused such havoc in money markets; before today's cut, one-month Libor was a full percentage point above the base rate?
The answer is that banks' are reluctant to lend to each other not because they are potless, but because their precautionary demand for cash has soared. They fear further losses on mortgage derivatives, and are holding cash as a reserve against this.
The obvious source of further losses is simply the continued downturn in the US housing market. But there are two less obvious ones. 
One is a principal-agent problem within banks. Many CDOs are illiquid, which means that "mark to market" valuation can become "mark to fantasy." Traders have wiggle room in valuing them, depending on what copula you use (pdf). I suspect that quite a few have wiggled on the high side, hoping to grab one last Christmas bonus before jumping. Because banks' top bosses are less knowledgeable than their juniors about such intricacies, they cannot be sure how great their losses are.
The second problem is that valuations of such instruments are very largely a function of confidence. Take SIVs, which borrow money to buy asset-backed securities. If credit markets are buoyant, as they were in the spring, borrowing is cheap and SIVs do well. But if markets are nervous, borrowing costs are high and so SIVs lose a lot. But the problem is that market nervousness depends upon what they think SIVs are worth. This is circular. The threat banks face is that the downward spiral will continue - loss of confidence causes losses on SIVs which cause further losses of confidence.
The message here - which isn't appreciated sufficiently outside financial circles - is that banks' problems are not hard quantifiable ones, but rather those of genuine Knightian uncertainty. What Willem Buiter said of government might also be true of banks: "The possibility that the authorities could be both truthful and competent is barely worth considering." This isn't because they are fradulent, but because some things just can't be known.

Comments

[dearieme]

Known unknowns, then?

[FSK]

I wrote a good article about the Subprime Lending Scam a lot on my blog. You also might be interested in the Compound Interest Paradox, which highlights the fundamental structural flaw in debt-based money.

http://fskrealityguide.blogspot.com/2007/06/compound-interest-paradox.html

http://fskrealityguide.blogspot.com/2007/08/subprime-mortgage-lending-scam.html

[Matt Munro]

The basic problem is fractional reserve banking IMHO

[Meh]

All this implies something which I've heard banking types mention "off the record" which is that each bank has looked at how much of it's "mark to market" now looks like "mark to fantasy." Knowing that "the chaps over there in Bank X are pretty much like us" that means they guess Bank X has a similar level of "mark to fantasy" and so the everyone wants a bigger risk premium on loaned money.