The Greece basis trade: What could go wrong?

By Felix Salmon
November 22, 2011

Why did Gretchen Morgenson write that column on Sunday about Greek credit default swaps? The answer is that the irresistible lure of writing about CDS lured her into the very murky waters of the Greek basis trade — the trade where you own Greek bonds and then hedge them by buying credit protection on Greece. Now this trade is emphatically not a big deal even in the context of the Greek debt restructuring: it’s probably a couple of billion euros in total, and won’t make much difference either way. But the outcome of the trade is likely to set an important precedent for the sovereign CDS market more generally, so it’s worth looking in a bit of detail at exactly what’s going on here.

Basis trades belong to a set which is relatively common in financial markets: things which are meant to be very safe but which, in fact, aren’t. Merrill Lynch reportedly lost somewhere in the region of $15 billion on basis trades, and at the height of the crisis I proposed that the US government should step in and start buying bond-and-CDS packages as a way to make money and get a bit of price discovery and liquidity into the fixed-income markets.

In theory, basis trades are simple: you buy a bond, which either pays off in full or doesn’t. If it does, you’re golden. If it doesn’t, then any losses you make on the bond can be recouped by profits on the CDS. So long as you buy the bond at a higher spread than the cost of credit protection, you should be guaranteed a modest profit.

But the question is how do you get there from here. Because CDS are derivatives, they’re subject to margin calls, and if you can’t meet CDS margin calls, you might be forced to unwind your trade before maturity. And that can be very expensive, if the basis has widened and markets have moved against you. So while the US government can play the basis trade without worries, everybody else has to treat it with a certain amount of caution.

And all of that was true before various EU member governments started deciding, in a killing-the-messsenger kind of way, that there was something profoundly evil about the sovereign CDS market and that they wanted to start trying to ban it.

There are two main forces putting the Greek restructuring together: the Greeks, and the EU. The Greeks, frankly, don’t care about CDS; what they care about is their debt, and their debt-service payments, both now and in the future. The CDS market, like all derivatives markets, is a zero-sum game, and whether CDS get triggered or not makes little substantive difference as far as Greece is concerned. That’s why I was highly skeptical of the idea that BNP had a conflict of interest when it was working for the Greek government while sitting on a key CDS committee: the Greek government really doesn’t have a dog in this fight.

The EU, on the other hand, has quite a lot of politicians and technocrats who hate the sovereign CDS market and would love to see it die. Bonds can be illiquid; CDS are the market’s way of pricing sovereign debt as accurately as it can. And when sovereign spreads are spiking nastily, it’s easy to blame a relatively small number of CDS traders for scaring the market, exacerbating very real problems, and making it harder to persuade people that things aren’t actually all that bad.

So, if the EU wants to throw a wrench of some kind into the spokes of the CDS market, what could it do in Greece? One thing would be to simply encourage Greece to do a “voluntary” bond exchange which doesn’t trigger the CDS. If bondholders playing the basis trade end up taking a 50% haircut on their bonds but getting nothing from their CDS, then clearly the CDS hasn’t provided them with the protection they thought it would.

There’s been a significant drop in the net notional amount of Greek CDS outstanding, of late, and that might indeed be a function of people unwinding their CDS trades in the face of uncertainty as to how they’ll be treated. But here’s the good news, as far as the CDS market is concerned: as the unwind has been going on, Greek CDS have continued to trade at extremely elevated levels. On a mark-to-market basis, anybody in the basis trade is fine. If they bought protection quite cheaply, they can unwind that trade and sell it at a high price, which will give them profits to offset against any losses they are forced to take on their bonds. In that sense, it doesn’t really matter what happens after the exchange: the trade is over, and it’s done what it was designed to do. More or less.

Meanwhile, as the unwind has been going on, the size of the Greek CDS market has been shrinking, so the number of people affected by the final decision as to whether or not the CDS is triggered has also been shrinking. That could be seen as good news for the EU and for people wanting to kill off the CDS market — but there’s a case to be made that the basis traders have simply moved on to Italy instead, which isn’t really a net improvement.

And the EU doesn’t just want to shrink the sovereign CDS market, it would ideally like to harm it in some way, in the hope that, once burned, the CDS traders stay away altogether next time.

Which brings me to the threats that BNP’s Belle Yang reportedly made in conversations with bondholders. Gretchen Morgenson didn’t explain them very well, partly because — I’m told — Yang herself was a little bit incoherent in what she was saying. But one bondholder did explain to me what he took away from the meeting.

The thing to remember here is that if the CDS isn’t triggered in the bond exchange, it doesn’t just disappear in a puff of uselessness. It still exists, and it still protects bondholders from a payment default. If you hold the old bonds — if you haven’t tendered into the exchange — then in many ways your basis trade hasn’t changed. Either Greece continues to make the coupon payments on the old bonds, or else it doesn’t, at which point the CDS really should trigger and make you whole.

But there are two ways that the sovereign CDS market really could be damaged in the aftermath of an exchange. The first is if the untendered old bonds got impaired significantly while the CDS remained untriggered. For instance, let’s say that Greece, using its own domestic law as the instrument, changed the payment terms on the old bonds so that they were paying out only a fraction of what they were paying before the exchange. That would almost certainly be a credit event under the ISDA definitions, and would trigger the CDS. But under one reading of what Yang was saying, Greece and/or the EU might attempt to impair the old bonds and pressure ISDA to declare that the impairment doesn’t count as an event of default. I very much doubt that ISDA would ever make such a determination. But if it did, then that would be a serious blow to the sovereign CDS market.

The second possibility is that Greece continues paying the old bond coupons in full, alongside its new bond coupons. And the Greek CDS market continues trading without any credit event. Until, one day, Greece being Greece, the country defaults again — on both its old bonds and its new ones. At that point, the CDS triggers. And to collect your payment on your CDS, you need to give up your bonds. But here’s the cunning bit: Yang has been hinting, I’m told, in her meetings with bondholders, that the EU and/or Greece will find some kind of way to change the law so that the old bonds aren’t eligible to be deliverable into the CDS exchange — to get paid out on your CDS, you’ll need to pay for new bonds, and your old bonds will be worthless.

If this were a credible threat, it would certainly be a good reason, at the margin, to tender into the current exchange rather than hold out. I don’t think it is a credible threat, but it’s easy to see how some investors might not want to take the risk.

The thing is, CDS is a young market, which hasn’t been tested in lots of different circumstances. No one can know for sure how it’s going to play out in future. So far, the CDS market has held up pretty well — everybody prophesying doom in the wake of the Lehman bankruptcy, for instance, was proved wrong. And when CDS were triggered on Fannie and Freddie despite the fact that there was never any payment default, the market coped with that well, too. My guess is that CDS will do what they’re meant to do, in Greece. But BNP is trying to spread a certain amount of fear, uncertainty and doubt over whether that’s necessarily the case.

As a result, anybody in the Greece basis trade needs to have a good degree of self-confidence that they know what they’re doing. Which, frankly, they should have had all along. Using derivatives to arbitrage securities is always a little bit messy and fraught with legal risks. If you don’t have confidence in what you’re doing, you shouldn’t be doing it in the first place.


We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see

“So far, the CDS market has held up pretty well — everybody prophesying doom in the wake of the Lehman bankruptcy, for instance, was proved wrong.”

I’m not the most knowledgeable commenter here by a long-shot but isn’t this only because of the government’s nationalization of AIG? I can’t see how you can point to this as an example of the cds market holding up well.

Posted by spectre855 | Report as abusive

On a completely different note: Felix, got an opinion of Ron Wyden’s efforts to stop SOPA?

Posted by GRRR | Report as abusive

You would have to be completely nuts to do the Greece basis trade, the cheapest to deliver bonds is a very long maturity far exceeding that of the CDS and liquidity in that bond is very poor, plus the Greek CDS trades mostly in USD which complicates hedging further as the quanto adjustment is very volatile. This has to be a very small trade, a fraction of the net notional outstanding.

Posted by alea | Report as abusive

“Because CDS are derivatives, they’re subject to margin calls, and if you can’t meet CDS margin calls, you might be forced to unwind your trade before maturity. And that can be very expensive, if the basis has widened and markets have moved against you.”

Felix, you got it backwards again. The CDS side of the trade only demands cash if the credit improves. You’re BUYING insurance, remember? So if the bonds go down (as Greece has) you would be RECEIVING daily settlement cash. And if the credit improves, and you can’t meet the margin call, you can always sell the bond into a rising market to close out the trade.

Only one market will move against you. This is a hedge. Repeat that: this is a hedge.

The real risk comes via counterparty issues. If your counterparty was AIG, your CDS swap against the default of Aurora 06-5 M2 might not have worked out very well, except that the Fed stepped in and lent them the collateral they needed to fulfill their margin call.

The reason why total return swaps pay more than the cash instrument is because of all this complexity risk. With the Euro-bankers engineering “voluntary” writedowns that don’t trigger the CDS’s (according to the ISDA), I’d be worried that I’d get fooled again. That’s what basis traders get paid the big bucks for.

Posted by Publius | Report as abusive

“Bonds can be illiquid; CDS are the market’s way of pricing sovereign debt as accurately as it can.”

I’ll concede this one if the CDS market becomes as liquid as (or even fairly close to as liquid) the sovereign debt market. Right now, it’s just False.

alea has it right.

Posted by klhoughton | Report as abusive

@Publius — the basis trade blowups of 2008-9 were in fact margin calls on people who were buying protection and who had the market moving with them, believe it or not. If their own credit was deterorating, they were still subject to such calls.

@klhoughton — you try finding Greek bonds, and then Greek CDS, and then get back to me.

Posted by FelixSalmon | Report as abusive

@Publius, remember that bit about “modest profit”? Most basis trades need to be levered to be worthwhile, so you will be posting collateral on the long bond position. Collateral transfers on the CDS leg may be mtm settlement, but on the bond leg will be based on a haircut which may be reset based on perceived risk and not just mtm. Anyway, it’s entirely possible to lose on both legs of a trade; it’s so common that there is a special name for it: Texas hedge. Also, that volatile quanto adjustment that alea mentioned makes a big difference because it means that it would be expensive (in trading costs) to stay hedged. I’ll bet anyone who actually puts on this trade is punting some part of the currency exposure.

@klhoughton, sure alea has it right, but he’s not saying the same thing you are.

Posted by Greycap | Report as abusive