Monday, March 8, 2010

More Naked CDS

According to Rajiv Sethi the prohibition of naked credit default swaps can be supported on the grounds of Diamond and Dybvig's financial instability model:

"In this case, expectations of default can become self-fulfilling even when solvency would not be a concern if expectations were less pessimistic. What does this have to do with naked credit default swaps? As John Geanakoplos notes in his paper on The Leverage Cycle, such contracts allow pessimists to leverage (much more so than they could if they were to short bonds instead). The resulting increase in the cost of borrowing, which will rise in tandem with higher CDS spreads, can make the difference between solvency and insolvency. And recognition of this process can tempt those who are not otherwise pessimistic to bet on default, as long as they are confident that enough of their peers will also do so. This clearly creates an incentive for coordinated manipulation."
I don't understand this argument for (at least) two reasons. First, unlike bank runs there is no sequential service constraint with sovereign bonds. If people try to exit the price falls, which prevents runs. Bank runs happen because of the advantage to rush to the head of the line.

Second, and more important, Sethi seems to forget that with naked CDS the purchaser is out money, the seller is receiving income. It is negative carry for the buyer. Thus there is a bias already against shorting in this fashion. The naked CDS holders are putting their money where their mouth is. Hard to see why they would do this based on some expectation that others will also do it, thus raising the cost of their play. Most traders would want to keep quiet to buy the insurance at a lower cost, and hope that others do not follow.

See The Money Demand for more interesting critiques of the Sethy argument.

No comments: