Quote from c*f(x):
I have been trading credit derivatives for a number of years. The problem is real, although fairly easily fixable. The main problem is documentation, and errors in documentation can cause a problem with leverage. First, a little about credit derivatives (we will just utilize a typical CDS contract here - but any of the more esoteric instruments - FTDs, etc. - suffer the same problems).
A CDS contract works in much the same way as an interest rate or FX swap. Two counterparties agree on terms and money changes hand on terms agreed to in the contract. The primary difference is the custom language used. Typically firms agree on an ISDA and CSA terminology that is mutually beneficial and fairly easy to settle. The problem is, though there is a standard ISDA and no one uses it - everyone has custom terms. The other problem is that CDS has a specific reference obligation and, unlike most swaps, requires custom documentation for each transaction. This is an enormous pain in the a**.
Contrast this with a standard corporate bond. If I buy or sell XYZ 7.5s of 2010, I will confirm with the other trader/salesperson and then - LATER THAT DAY - my backoffice will confirm with their backoffice. Everyone goes home square. With a CDS however, the backoffice confirmation does not occur until many days later. Normally within 30, but still, that is a long time.
This causes two problems. The first is obvious. If there is a trade error, one firm or the other may fail on a payment. Obviously this is bad. The other is slightly less obvious, but potentially more dangerous. Each relationship with a counterparty requires a credit line. In order for a firm - say Deutsche Bank - to do business with ABC Hedge Fund (or any counterparty for that matter), DB has to give them a credit line. For this, DB needs some transparency as to what sort of shape ABC Hedge Fund is in financially. They do this by knowing what positions (w/i a backoffice/credit function) ABC Hedge Fund has on. But the two firms communicate backoffice to backoffice. And if the trade hasn't been confirmed, well, guess what - it really doesn't exist. Arguably an honest firm will go ahead and disclose their true positions. One who is in a corner, however, has a disincentive to do so. If you are down, banks might not continue your credit line and thus you can't dig your way out of the hole.
There is also a delay in what we call "assignments" or "give ups". When A buys protection from B and then B buys protection from C, typically what happens is that B will assign A to face C. This usually happens when a fund does a transaction with a bank, then offsets it with another transaction with another bank and tells the banks they face each other - most always with some cash payment made or rec'd by the fund.
Additionally, there have been problems with the way contracts are interpreted. This came up years ago when the default wave of 01 and 02 occurred, as well as misunderstandings recently, such as those with Kerr McGee's restructuring. Not necessary to go into all of this here, but there have been bumps on the road.
What would solve this? Faster settlement for one. More homogeneous contracts would help. An exchange (sacrilege for bond traders) would also be stellar. But my real guess is that the Fed will pressure banks to speed up settlement and confirmations in order to provide better visibility for credit officers.
Crazy stuff, eh?