Monday, November 10, 2008

CDS: An IM Exchange

(Felix Salmon) Many thanks to Robert Waldmann, who allowed himself to be roped into an IM conversation with me about the CDS market, after he left a couple of skeptical comments on the subject here and at Kevin Drum's. Here, then, is my (partially successful) attempt to persuade him that it's not as bad/dangerous as all that. And, as a bonus, if you make it all the way through to the end, you'll be treated to an indelible image involving Larry Summers.

So, wonks away:

Robert Waldmann: Thanks for the invitation. I am flattered. However, I am also fairly ignorant about developments in financial markets in the past 20 years, so please bear with me.

Felix Salmon: Bilateral derivatives have been around for much longer than that, so there's nothing in the CDS market which is really rocket science
In fact, CDS are much easier to price than options

Robert Waldmann: I've been learning about them fairly quickly, but just creating gaps of knowledge in my ignorance. I won't bother you with further warnings.

Felix Salmon: My feeling is that they grew out of big problems in the corporate bond market.
It was always very illiquid
I remember when I started at Euromoney in the mid-90s, bond traders would make an absolute fortune, thanks to bid-offer spreads you could land a 747 in
You can just imagine what would have happened to credit markets at a time like this, then, if it wasn't for CDS.

Robert Waldmann: Well that gets you as far as bilateral trades instead of submitting huge market orders and driving prices against yourself.
The legal form of derivatives has to have (I think) a lot to do with capital requirements.

Felix Salmon: The clever thing about the CDS market is that precisely because it's unregulated and bilateral, it has developed its own form of capital requirements, in the form of the seller of protection having to post collateral when prices move against it
This has actually worked extremely well, and the only problems in the CDS market so far have been in that small handful of writers of credit protection who had AAA ratings and managed to exempt themselves from having to post collateral.

Robert Waldmann: OK but the variable collateral can cause, uhm, cash flow problems for sellers having to get their hands on tens of billions given how fast and much prices can move.

Felix Salmon: only if they're stubborn enough not to unwind their positions in the face of prices moving against them. The good thing about the CDS market is that it's liquid enough that that is actually possible.

Robert Waldmann: A question. I read somewhere that AIG has had to post collateral now that (credit ratings of insured credit declined and/or its credit rating has uhm changed a bit and/or market prices of CDS it wrote changed).

Felix Salmon: Yes, the minute you lose your AAA, you have to start posting collateral.
Hence the gazillions of dollars in collateral which AIG has to borrow from Treasury at Libor +850bp.

Robert Waldmann: OK so it was their personal corporate rating that did it. Also the money will go back to AIG if when the crisis is over they get back to AAA and the insured credit in question hasn't defaulted (I understand not much has yet so relatively little paid to settle CDS contracts).

Felix Salmon: The money certainly goes back to AIG when the contract expires. I don't know enough about the language in AIG's contracts to be able to tell you whether the collateral requirement goes back down to zero in the event AIG regains its AAA.
That said, Berkshire Hathaway says in its most recent quarterly report that it hasn't put up a penny of collateral against the CDS it has written, despite taking over a billion dollars of losses on CDS and equity puts in the third quarter alone.
I don't like that. I think everybody should put up collateral, and compete on an even playing field.

Robert Waldmann: Seems to me that the complexity of CDS due to complete flexibility in writing the contracts makes them very hard to price.

Felix Salmon: That was more of a problem in the early days of CDS, but nowadays substantially all of them conform to standard ISDA boilerplate.

Robert Waldmann: ah hah. Where can I find standard ISDA boilerplate
(googling ISDA at their web page)
they seem to have a good bit of traffic right now oh it's a pdf
LCDS protocol published July 24 ... damn I'm being paid too little (0) to read this stuff.

Felix Salmon: The important thing is that everyone's using the same protocols. Which means that so long as you've dealt with counterparty risk, you can close out a position with a perfect hedge.

Felix Salmon: And it's definitely possible to deal with most counterparty risk, through a combination of CCDS and collateral requirements, as explained here.

Robert Waldmann: Yeah thought came to me that a cds on the counterparty would do it (so my thought was wrong because I didn't think collateral was relevant)

Felix Salmon: It's definitely kludgier than a CDS exchange, but it has actually proved surprisingly robust.
But I'm interested in a comment you left on my blog:
"Now as to CDS abuse, tell me a non abuse involving way in which the notional value of CDS is so much greater than the value of all insured D? If something like that happened with another kind of insurance, wouldn't you consider it abuse?"
I think I can answer that quite easily: I enter into a position, and then I unwind it by entering into an equal and opposite position
And since I'm a trader, I do that many times a day.
Pretty soon, notional is through the roof, even with no net exposure at all.

Robert Waldmann: Yes that seems to be an answer (and indeed quite easy).

Felix Salmon: And the data which DTCC is finally giving out would seem to indicate that net exposures are only ever a tiny fraction of nominal notional exposures.
Where was the DTCC a year ago, that's what I want to know, it might have saved us a lot of worry.

Robert Waldmann: That's for sure.

Felix Salmon: It's like they waited for the moment of absolute maximum panic, following the Lehman default, before sauntering in a nonchalant manner onto the scene and saying "everyone chill the fuck out, I got this"

Robert Waldmann: Well so long as they got this, fine by me.

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.
Basically the banks thought they were moving all the credit risk off their books, but they weren't, because they kept those "super-senior" tranches.
They could easily have moved the super-senior credit risk off their books too, but that would have made their CDO deals unprofitable, so they didn't.
That's not a problem with CDS technology, it's a problem with credit-risk modelling.

Robert Waldmann: Yours is a plausible hypotheis. As far as I know (and I don't know much) another is that they placed huge bets against each other and don't know yet who won big and who is insolvent (that is each is still sure they won big and the others are insulvent).
My sense is that it has to be something which is very different for different firms or the general equity injection for preferred shares (and promise of more) would have pushed the TED spread back to normal.

Felix Salmon: Bank prop desks certainly had the ability to make big bets using CDS. But those are the kinds of bets you mark to market daily.

Robert Waldmann: Yeah but CDS markets can not move not move not move JUMP. Historical data can be very misleading if distributions have fat tails and if history so far is not very long.

Felix Salmon: There are basically three kinds of institutions which wrote a lot of CDS protection on a net basis.
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.
As I say, it's possible that someone who sold credit protection could find themselves losing a lot of money one day when the CDS spreads gapped out
But that would show up in a bank's daily P&L, and in its quarterly earnings.
And generally the CDS desks at banks were so profitable just as brokers, taking no or very small net positions, that it seems improbable they would have tried to reinvent themselves as insurance companies and write a lot of protection on a net basis

Robert Waldmann: Some people (hedge funds mostly I'd guess) decided to bet on CDS misspricing by taking large not quite matched long and short CDS positions. They would be having uhm cash flow problems, if it weren't for AIG's credit facility at the FED.

Felix Salmon: You're quite right that the failure of AIG could have been catastrophic for the CDS market -- with any luck, we'll never know.

Robert Waldmann: The consequences of those cash flow problems have to do with CDS trading and I would count what would have happened without public intervention.

Felix Salmon: My point is that AIG is a unique case, thanks to the way it abused its AAA rating.

Robert Waldmann: I have been arguing with you in my mind usually remembering not to talk to myself out loud (makes people nervous). My argument was that just because things are OK with $90 billion in 27 days for liquidity assistance (most could go back to AIG) doesn't mean that the market behaved as a market should be allowed to behave.
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.

Robert Waldmann: I know I remembered your BH complaint so I guessed you would say yes (didn't guess the 2nd and 3rd yes).

Felix Salmon: Interestingly the US govt, which is undoubtedly safer than BRK, is posting collateral on the protection written by AIG.

Robert Waldmann: Well the US government has a long tradition of not exploiting its credit rating to the max. Without it, the USA would be the Union of soviet socialist republics of America.

Felix Salmon: it's also interesting though that when the US govt does try to exploit its credit rating, by backstopping Frannie's debt, it doesn't seem to work!
The market reckons that if the govt can guarantee Frannie debt, it can unguarantee it too.

Robert Waldmann: Well that was a wink wink nudge nudge semi maybe guarantee made by people who have left office to be replaced by people who hate Fannie and Freddie so not really a shocker.

Felix Salmon: Yeah, didn't Summers at one point propose a plan which would bail out Frannie's senior bondholders but do a debt-for-equity swap when it came to the sub debt?

Robert Waldmann: I don't know

Felix Salmon: Found it
"The government should use its new receivership power to protect taxpayers and the financial system. In the process, payments to stock holders, holders of preferred stock and probably subordinated debt holders would be wiped out, conserving cash for the benefit of taxpayers."

Robert Waldmann: Ouch

Felix Salmon: This is the man who will probably be treasury secretary come Jan 20
At least according to Intrade

Robert Waldmann: Yeah but last time he was treasury secretary they managed to teach him to not shoot off his mouth. I don't know how (I think a cattle prod was involved).
It's dinner time over here (I live in Rome you know).

Felix Salmon: I think the vision of Michele Davis wielding a cattle prod is a great place to end this, then

Robert Waldmann: OK thanks for the chat. Real fun. Bye for now

No comments: