Tuesday, May 11, 2010

Darrell Duffie on Banning Naked CDS Transactions

Posted on Grasping Reality With Both Hands:

Among other things, [the amendment to the Senate financial reform bill that would ban naked CDS transactions] means a big limitation on the entire securitization market, since in many cases the loan default risk is transferred into the structured credit product through a CDS. (This is not to be confused with a synthetic CDO, which is much different, and not such a big loss in my view.) So, lots of loans to ordinary individual Americans (their credit card loans, home equity loans, mortgages, and so on) and American operating companies will get somewhat more expensive. (This effect is not huge. Credit card or home equity interest rates will not go up 4%, say, but they would be higher because of this, other things equal.) [Disclosure: I am a director of Moody's Corporation since late 2008. Moody's makes money rating these products. There would be fewer of them to rate if this passes.]

There seems to be a presumption that buying CDS protection is fine as long as you have lent money to the borrower. That's a bad presumption. Example: I lend you money. There are covenants on the loan that protect me by requiring you to run your business prudently, and to not borrow too much. Then I ask a CDS protection seller, X, to sell me protection on you. Now I don't care if you default or not, so I won't worry about monitoring or enforcing those covenants, because X will pay me if you go under. Bad news. Even worse, where it might be efficient to help you avoid default, so that I can eventually get my money back, I will pull the rug from under you by calling in the loan. You won't be able to pay me back, but I am covered by X. Bad news. CDS can be misused more easily by those buying protection when having lent to the borrower ("legitimate CDS" in this amendment), than by those who have not lent, and have no ability to affect the borrower.

Lost ability of Americans to reduce their risk. Another Example: This one was part of testimony to the House Financial Services last week, on the panel on which I sat. A congressman (Rep. Manzullo, I think) asked how John Deere, a tractor manufacturer in his district, could use derivatives in its business. I gave an example in which John Deere sells 1000 tractors to a Greek company. Another panelist, Bob Pickel, explained that the Greek buyer of tractors would probably not be available as a referenced name in the CDS market, but that John Deere could get a reasonable hedge against the default risk of Greek firms by buying CDS protection referencing Greek sovereign bonds. That is true. But, this amendment would rule that out. John Deere would be unable to hedge the default risk on the receivables of its tractor sales. This is just one of many examples in which CDS protection buyers who reference a proxy name to get a hedge would no longer be able to hedge. Also, the definition of a "valid credit instrument" will probably be too narrow to allow people to hedge against losses when a borrower defaults that are not losses on a valid credit instrument. For example, if Company X defaults, I will lose the opportunity to collect money owed to me on the foreign exchange derivatives I have with X. If Country Y defaults, the market value of my factories in Country Y would decline precipitously. I could no longer hedge that. Why would anyone want to prevent an American company from protecting itself from losses this way?

The reporting requirement is redundant. All CDS and all other OTC derivatives in the SEC's regulatory domain will be required to be reported to the SEC already under the "data repositories" provision of the bill.

The restriction of a maximum of 60 days for a dealer to be "short" without owning a credit instrument is not very elegant, to say the least.

If it passes, it will not be the end of the world, but it is a step backward. The cost-benefit analysis: Cost: moderate. Benefit: none that I can see...

No comments: