Alea today posts the timeline for physical settlement of credit default swaps, once a credit event has been declared. He doesn’t say why he’s posting it, but the main thing to note is that it’s likely to take a couple of months between (a) the credit event being declared in Greece, and (b) the final settlement of all credit default swaps on Greece.
And that, in turn, reveals a significant weakness in the architecture of CDS documentation. It may or may not be a big deal, this time round. But market participants have already been spooked by the possibility that Greece might be able to default without triggering its CDS at all. Now they can add to that another worry: that Greece might be able to default in such a manner as to leave the ultimate value of the CDS largely a matter of luck.
The way that CDS auctions work, you start with a credit event. Then, using an auction mechanism, the market works out what the cheapest bond of the defaulting issuer is worth. If it’s worth, say, 25 cents on the dollar, then people who wrote credit protection end up paying 75 cents to the people who bought protection: that’s equivalent to the people who bought protection getting 100 cents on the dollar, and handing their bonds over in return.
With Greece, however, the bond exchange is going to complicate things — a lot. Remember that it has a natural deadline: March 20, when a €14 billion principal payment comes due. If Greece’s old bonds haven’t been exchanged for new bonds by that point, then things will get even uglier, and even more chaotic, than anybody’s expecting right now. So it’s very much in Greece’s interest, and Europe’s more generally, to have everything wrapped up by March 20. Bondholders too, truth be told — they hate uncertainty.
But then there’s the CDS holders. In the best-case scenario for Greece and Europe and bondholders, every €1,000 of old Greek bonds will get converted to new bonds with a face value of just €315. Those bonds will probably trade at about 30% of face value, which means the new-Greek-bond component of the exchange will be worth about 10 cents for every dollar in face value of old Greek bonds that you might currently hold. Add in another 15 cents of EFSF bonds, and the total value of the exchange will be about 25 cents on the dollar, which is why people are talking about a 75% “present value haircut”.
The important thing, here, is that Greece is issuing new bonds worth around 10 cents on the dollar, while the EFSF is issuing new bonds worth around 15 cents on the dollar. The structure of the new Greek bonds is secondary: these ones involve a nominal haircut of 68.5%, and a market price of about 30 cents on the dollar. But theoretically, Greece could have constructed bonds with a significantly higher coupon and a bigger nominal haircut — maybe the haircut would be 85%, with the bonds trading at 67 cents on the dollar. Bondholders would still receive about €100 worth of new Greek bonds for every €1,000 of old Greek bonds they hold. But instead of the new Greek bonds trading at 30 cents, they’d trade at 67 cents.
Why does it matter what the nominal price of the new Greek bonds is, so long as the total package, including EFSF bonds, is worth about 25 cents on the dollar? Economically speaking, it doesn’t. But for the purposes of the CDS auction, it matters a great deal.
The reason is that the key number in the auction is the nominal value of the cheapest-to-deliver Greek bond — that’s the price at which the auction clears. And here’s the rub: this auction is going to take place after March 20, after the old Greek bonds have been exchanged into new securities. Because Greece intends to use collective action clauses to change the terms of all its outstanding bonds, even if they’re not tendered into the exchange, there effectively won’t be any old bonds in existence by the time the CDS auction happens. The only outstanding reference securities will be new bonds.
In the auction, market participants will not be bidding on the value of the package that is being offered in return for every old bond. The new EFSF bonds are obligations of the EFSF, for instance: they’re not obligations of Greece, and they have no place in a Greek CDS auction.
The way that CDS auctions are meant to work is that once a borrower defaults on its debt, that defaulted debt continues to be traded in the market, and its value then determines the amount that credit default swaps need to pay out. But in this case, Greece’s defaulted debt might well not continue to be traded in the market. In which case, when traders need to find a cheapest-to-deliver bond to bid on in the CDS auction, they’re going to have to use one of the new bonds, rather than one of the old ones.
And now you can see why the nominal price of the new Greek bonds is so important. Right now, it seems that they’ll be trading at a nominal price of about 30 cents on the dollar, which is close (ish) to the current market price of the old Greek debt. But there’s no particular reason why that should be the case. If Greece had gone for an 85% nominal haircut rather than a 68.5% nominal haircut, then the nominal price of the new Greek bonds would be 67 cents on the dollar — and anybody who wrote credit protection on Greece would only have to pay out 33 cents on the dollar rather than 70 cents on the dollar.
In other words, Greece’s CDS really aren’t protecting holders of Greek bonds at all — or if they do, it’s more a matter of luck than of law. When they get paid out on their CDS holdings, people owning protection against a Greek default won’t get paid according to how much money they lost on their old bonds. Instead, they’ll get paid according to the nominal price of the new bonds.
What this means is that the CDS architecture is broken, and can’t cope with collective action clauses. And as a result, according to the hedge fund manager who tipped me off to the whole problem, “this Greece CDS imbroglio might be the final blow for sovereign CDS as a product.”
Now there is a possible solution here: ISDA could try to decree, somehow, that the total package bondholders receive in return for their old bonds will count as a deliverable security for the purposes of the CDS auction. Bundle up the new bonds, the EFSF bonds, the GDP warrants, everything — and that bundle can be bid on in the auction, to determine where the CDS pays out. That would be fair and right. But the problem is, it might not be legal. There’s really nothing in the ISDA CDS documentation which explicitly allows that to happen.
The whole point about credit default swaps is that they’re meant to behave in a predictable manner in the event of default; one thing we know for sure about Greece is that the behavior of its CDS is going to be anything but predictable. We don’t even know for sure whether they’ll be triggered, let alone what they’ll be worth if and when they are.
Now there are a lot of people, among them European policymakers, who would actually be quite happy if the Greek default killed off the sovereign CDS market as a side effect. But I actually believe that sovereign CDS, when they work, are rather useful things. It’s just that Greece is having the effect of showing that they don’t necessarily work. And if you can’t be sure that they’ll work when triggered, there’s really no point in buying them at all.