Derivative Surprises Everyone By Accomplishing Purpose and Unwinding at Randomly Generated Market Value

Author:
Updated:
Original:

Yay, Greek CDS worked. But, as we talked about a bit, it almost didn't:

By happenstance, some of the new bonds Greece has issued in its restructuring have a market price close to the total value of the package creditors received — about 22 cents on the euro. Those bonds will help set the CDS payout, and trouble will be averted: CDS holders will receive about 78 cents, roughly equivalent to the loss bondholders suffered. ...

If the new Greek bonds had different terms — higher or lower interest payments for instance — their prices could be substantially different, changing the amount the default swaps would pay. Ben Heller, a portfolio manager at New York hedge fund Hutchin Hill Capital, which owns both Greek bonds and CDS, said that means the swaps aren't doing their job. He said that until the problem is fixed, he "will not use CDS as a hedge against credit exposures anymore."

In fact Heller told Felix Salmon:

When you think about it, it's a product that, on certain poorly defined credit events, offers a random payout. So if I want to do that, then I could play roulette at a casino.

So, first of all: yes! I think worry about the definition of credit events is a bit overblown, but the randomness of the payout is a real thing and bizarre and terrifying. It bears re-emphasizing that the method of calculating the Greek CDS payout bears no relation whatsoever to the default risk that it was supposedly hedging.

But, also: no! Greek CDS ultimately settled at 21.5. That randomly generated payout number is veeeeery close to the "right" number, i.e. the number that would have hedged that risk it was supposed to be hedging. (And, it's maybe worth mentioning, the payoff was too big, not too small.) Which is maybe a coincidence except Felix also tells the story of how the new Greek bonds' maturities got reshuffled so that the 2042 would lead to more or less the correct result in the CDS auction.

One thing to remember that CDS is just a contract between two parties, one or both of whom is generally a securities dealer. ISDA is just a third-party provider of infrastructure that those two parties often like to sign up to. If you don't like the thing that CDS is, go find someone to sell you a thing that is the thing that you want to have. Heller plans to; he's got his eye on bond futures.

Somehow this got me thinking of our friend Greg Smith. Here is some sensible stuff from one Jacki Zehner, who also left Goldman but not in a huff and in fact as a partner, a title she got at age 32, which she points out in passing, leaving it unmentioned that at that age Greg Smith was still a VP* so SUCK IT GREG, but anyway, where were we:

Though I left my trading position 12 years ago, I will tell you from personal experience that the vast majority of people I worked with cared deeply about our customers, and, if you were heard calling customers any of the things Mr. Smith mentioned, you would be in big trouble. BIG.

More relevant is that if you continuously ‘ripped your customers eyeballs out’, they would cease to be your customers. The behavior that Mr. Smith so graphically describes was, at the time, the exception and not the norm in my opinion. Were we supposed to make money? Yes. If you promoted that this came at the customer’s expense, that was a bad thing.

There was a period of time in the distant past where, by accident, my desk wrote contracts that allowed us to terminate them immediately.** This really was a pure accident - we'd been writing the contracts the same way for years, but market conditions had moved such that a trigger that you'd never expect to be hit was hit at inception of the trade. Terminating would be a good thing for us, since it would in theory let us take our entire profit on the trade without any more risk or balance-sheet usage. Our customers, who were sophisticated large companies advised by fancy law firms, had read the contracts and signed on for this, although of course they didn't know that - to be fair, neither did we, and we read a lot more of these than they did. Then we discovered the glitch, and ... well, we eventually fixed it in future documents. But for the contracts we'd already signed up, we had discovered a free termination right that was valuable to us. So we basked, briefly, in our newfound ability to terminate trades like assholes and generate millions of dollars in risk-free profits. Then we went back to work. Guess how many we terminated?

If your business is dealing CDS or equity derivatives or used cars or whatever, repeatedly ripping eyeballs out is ... well, bad business. Customers often use their eyeballs for things, like seeing, so they tend to notice when they've been ripped out and have hard feelings about the operation. They go elsewhere. So even if you've got their eyeballs in your grasp and could, with one quick tug, acquire those eyeballs and also probably money because who wants eyeballs, you might well forebear and let them see the next deal you've got cooking for them.

But even if you're the nicest guy in the world, it is nice to have optionality. If you're a derivatives dealer, you're particularly attuned to free options, even if all they're realistically good for is some gentle basking. If you're ISDA, or the dealer banks that make up the majority of its membership (and also maybe have some input into the Greek restructuring), a thing that you probably don't want to do is blow up the CDS market for ever and aye by just picking a wildly wrong random number. But you probably like having the (partial, constrained, of-uncertain-legality) option to do so - constructed with, say, ambiguous documentation or determinations committees that decide whether there's been a credit event and what obligations are eligible for auction. Because it turned out that banks' Greek CDS exposure was pretty small - but if it had been, I dunno, Italy, ripping out some client eyeballs might have been necessary for the banks to save their own.

ISDA is doing some entertaining throat-clearing about fixing the screwy CDS settlement provisions that were invoked here, and the Ben Hellers of the world are pushing them to do that. Maybe they will, especially since the next default could go against dealers. Or maybe, by accidentally or accidentally-on-purpose alighting on an attractive - reasonable but slightly protection-buyer-favoring - random number for the Greek auction, they'll avert calls for reform and get to keep a bit of optionality for the next, scarier default.

Greek Deal Highlights Flaws in Default Swaps [WSJ]
A top CDS trader quits the CDS market [Felix Salmon / Reuters]
[Greece CDS] Final auction results [FTAV]

* As was I! Ah, misspent youth.

** Look, I'm not really giving anything away here because these contracts are publicly filed. Go find them. If you can figure out the problem, then, well, good for you, go sell equity derivatives, I HEAR THEY'RE HIRING.

Also, like, this happens all the time everywhere. Who reads documents?

Related

Derivative Accomplishes Purpose And Unwinds At Market Price

You can read the Jamie Dimon "Don't gloat about how bad Goldman is. Did you hear me? Don't gloat about how BAD GOLDMAN IS. The fact that GOLDMAN is BAD is of no interest to our clients. Or the press. Don't leak this to the press!" memo two ways. One is, y'know, what it sounds like: Dimon gets to score some easy/meta points by spreading it around that his business practices are so superior that he doesn't even need to spread it around that his business practices are superior. The other is that making money off of clients isn't something invented at Goldman Sachs and anyone at JPMorgan who throws stones is likely to be clonked on the forehead by a ricochet. (Or possibly by a deranged fictional whistleblower!*) The latter interpretation is probably right for James Gorman's more full-throated defense of Goldman because whoops: