YET MORE ON THE CDS MARKET....I've been meaning to link to yet another Felix Salmon post about credit default swaps, since I know what a fascinating subject they are for everyone, but one thing led to another and I haven't done it yet. Basically, "one thing and another" means that I spent several hours yesterday trying to understand the whole CDS issue better, but I failed miserably. So instead of pretending otherwise, I'm just going to link. Salmon conducted an IM conversation with Robert Waldmann about the CDS market, and part of it went like this:
Felix Salmon: So, have I brought you around to the idea that CDS really aren't a major cause of the current crisis?
As you know, Kevin Drum calls me "disturbingly persuasive"
Robert Waldmann: Ah well that is ambitious. You have convinced me that there is a perfectly legitimate reason which can explain why face value is so huge. As to the cause of the crisis, I remain confused. Stupid CDS tricks could have done it. So could stupid CDO tricks and what all.
Felix Salmon: I will concede that there were indeed stupid CDS tricks
Robert Waldmann: I mean the situtation is I don't understand the new financial instruments and it sure looks like the trader types didn't understand them as well as they thought.
Felix Salmon: But the stupidity was in understanding credit risk, not in understanding CDS.
Well....sure, but this seems like a bit of a dodge. After all, pretty much all financial bubbles are based on mispricing risk in some way or another. It seems like we need to dig a little deeper and try to figure out if there were specific aspects of the way modern financial markets are regulated that encouraged even more risk mispricing than usual.
The question, then, isn't whether credit default swaps are useful instruments. They are. The question is whether there's something about the way they're managed in real life that makes them potentially more dangerous than, say, stocks or pork belly futures. And if so, what can we do to limit "stupid CDS tricks"? Here's one possibility:
Robert Waldmann: OK a reform proposal. CDS must come with collateral even if you find a sucker willing to buy one without collateral (this is a regulatory restriction).
Felix Salmon: Yes yes yes.
That's why I'm so astonished Berkshire Hathaway is STILL writing CDS without collateral requirements.
But a move to an exchange would have the same effect.
If I understand this right, the benefit of requiring collateral is twofold. First, it keeps the CDS market from going too crazy, since CDS sellers can only sell protection if they have collateral to back their positions. Second, it reduces counterparty risk, since even if the CDS seller goes bust there's collateral that's been posted to make good on the swap. The big problem with AIG, for example, which has been the most spectacular example of a financial firm losing its shirt due to CDS exposure, is that they were writing CDS willy nilly without having to post collateral (thanks to their AAA rating). When their rating started going south, and they had to post collateral to make up for it, the company went down the toilet. If they'd had to post collateral in the first place, that wouldn't have happened.
So I guess that's a good start: make CDS exchange traded and insist on collateral posting requirements for all writers of CDS.
But what I still don't have a handle on is the scale of the losses in the CDS market. Clearly, AIG and the monoline insurers lost a ton of money. But Salmon suggests that the broader banking industry didn't. Partly this is because only a small segment of the financial industry were net sellers of CDS protection:
Felix Salmon: There's AIG, there's the monolines, and there's the synthetic CDOs bought by institutional investors.
Given the zero-sum nature of any derivatives market, that means that everybody else, on net, was a buyer of credit protection.
So far, this seems to be right: net losses in the broader financial industry on CDS trades seem to be pretty modest. So far. Unfortunately, though, that still leaves the CDOs, which we don't know much about yet, and it also leaves everyone else, who might be choosing not to settle CDS contracts yet that have big losses associated with them. I have a feeling we might need to wait a while longer to know for sure if the broader CDS market is as benign as Salmon thinks.
But it might be. Obviously it's cheating a bit to single out the particular areas where CDS sales caused big problems and then say, "well, aside from those areas everything was fine" after all, it's always the case in every industry that aside from the problems there are no problems but still, if it turned out that the big abuses came from noncollateralized CDS sales and synthetic CDOs, that would make me happy. After all, I'm already on record as thinking that CDOs are the devil's spawn. Anything that heaps more abuse in their direction is fine with me.
Bottom line: I'm still confused. For one thing, an awful lot of smart people seem to disagree with Salmon. I'd like to see some of them engage with his arguments. For another, the CDS market is so opaque that we still don't really know how much exposure is out there and who has it. At the very least, that seems unacceptable. And finally, even if there were only three segments of the financial market that were net sellers of CDS protection, just how much is it going to cost us to bail them out?
One way or another, the losses in the financial markets appear to be far wider than just subprime loans. After all, the size of subprime losses in the U.S. seems to be about half a trillion dollars, but in the past year banks have raised something like $300-400 billion in new private capital and another $200 billion so far in government capital. So that means their overall capitalization levels should be OK. But apparently that's not the case. So where are all the rest of the losses coming from? CDS? CDOs? Currency forwards? What? Does anybody actually know? And if not, what will it take to find out?
POSTSCRIPT: And one more thing. It's really annoying that the plural of CDS is CDS. "CDS market" is fine, and CDS when referring to an individual swap is fine, but why not CDSes when referring to multiple swaps? As in, "Sellers of CDSes should be required to post adequate collateral for each CDS they sell"? What does Wall Street have against plural acronyms?