Felix Salmon

CDS demonization watch, Bloomberg edition

By Felix Salmon
October 31, 2011

Bloomberg View’s Mark Buchanan has been taking a long, hard look at a 2009 paper by Italian physicist Stefano Battiston, Joe Stiglitz, and others; he explains what it says quite clearly and accurately on his personal blog. Basically, the paper quantifies the concept of “too interconnected to fail”: when you have a financial system with lots of banks, all of which have exposure to each other, then the system itself becomes much more fragile.

Here’s the chart:


What you’re looking at here is a blue-dotted “baseline”, where the probability of failure falls steadily as the number of counterparties goes up, and a sold-red reality, where a few counterparties help, but a lot of counterparties only serve to reinforce trends, exacerbate downward credit spirals, and generally increase systemic fragility.

So far so uncontroversial. But equally, there’s absolutely nothing in this paper to justify Buchanan’s Bloomberg headline: “Credit-Default Swap Risk Bomb Is Wired to Explode”.

There’s a germ of an interesting point in Buchanan’s column. Because the CDS market is unregulated, no one knows the degree to which it makes this syndrome worse, and exacerbates the interconnectedness of the financial system as a whole. So there’s a strong case to be made for moving the CDS market onto exchanges, where it can be watched far more closely.

But the paper Buchanan’s talking about never mentions the CDS market at all. And although Buchanan talks a lot about credit default swaps in his column and in his headline, the fact is that all of his points apply to just about any kind of interconnectedness. The paper concentrates on credit — the interbank market, basically. Buchanan extends that to credit default swaps. But most interbank derivatives exposure is not CDS related, and most of Buchanan’s points about CDS also apply, mutatis mutandis, to derivatives exposure more generally.

And this is far from convincing:

What reduces risk for individual institutions in small quantities spells trouble for the larger banking system when pushed too far. This is especially worrying when you consider that the number of CDS contracts outstanding on European sovereign debt has doubled in only the past three years, even after the AIG catastrophe. We don’t know if similar dangers lurk in the network of CDS contracts that links European banks with one another, as well as with banks in the U.S. and elsewhere.

Firstly, the amount of CDS written on European sovereign debt is tiny. Yes, it has increased over the past few years, as the Euro crisis has gotten steadily worse. These are called credit default swaps, after all: you only care about them insofar as you care about creditworthiness, and up until a few years ago European sovereign creditworthiness was not really an issue on the radar. But the European sovereign CDS market is still very much a backwater as far as the CDS market as a whole is concerned. (You can see one reason why when you look at Greece, where it looks very much as though in the first instance there’s going to be a sovereign default without a credit event.)

And secondly, we do know, pretty accurately, how much CDS protection has been written on European banks. We know this for the same reason that we know how much CDS there is outstanding on Greece: the DTCC does a very good job of keeping track of such things.

One of the interesting lessons of Lehman’s bankruptcy was that its CDS settled without a hitch, despite a lot of apocalyptic forecasts on the subject. That doesn’t mean the CDS market is fully robust, of course: it’s just one datapoint. But in the wonky world of counterparty hedging, the kind of effects that Buchanan is talking about are the first and biggest worry facing any desk. They’re not some kind of forgotten detail; they’re the whole reason why counterparty hedgers are in such demand in the first place.

It’s fair to say that CDS are more problematic than other derivatives in terms of interconnectedness issues, because of the “jump risk” involved and the fact that moves in CDS prices can themselves cause worries about a bank being close to failure. But it’s not fair to say that there’s a “risk bomb wired to explode” here. Banks don’t, in point of fact, hedge their European sovereign credit risk in the CDS market — if they did, the European sovereign CDS market would be much bigger than it is. And sovereign-debt crises are always banking crises, regardless of whether CDS even exist or not.

So while it’s fair to say that a sovereign crisis in Europe could threaten the entire banking system, it’s not fair to blame that fact on CDS. It would be true even if CDS had never been invented.

5 comments so far | RSS Comments RSS

Then why are the euro banking authorities going to such great lengths to avoid the official recognition of a “credit event” that would cause the CDS to have to be settled? If the CDS protection the Greek debt isn’t significant, the surely it would be easier to just admit that the writedowns on the Greek debt are a default and the banks will just have to suck it up. Are they delusional, or are you?

Posted by Moopheus | Report as abusive

What happened to MF Global? Was their bankruptcy caused by a series of stupid bets or by the failure of CDS to trigger in the Greek default?

Reminds me of the maxim, don’t invest in anything you don’t understand. Clearly they didn’t properly understand whatever they were investing in.

Posted by TFF | Report as abusive

Theres nothing inherently wrong with CDS’s, except they get completely abused. That can be pretty much stopped by two simple changes, the first of which is if they’re going to allow naked CDSs, where anybody can bet on a loan defaulting, even if they are not a party to it, then those kinds of bets need to be brokered by professionals: the casinos. They’ve been dealing with things like this for decades, and will price the swaps accordingly so there is no domino effect when a default happens.

The second change is if they’re going to let insurance companies like AIG write insurance against defaults, they need to prove they have enough liquid reserves to cover losses without bringing down their counter-parties. If they need to use assets with no public market, only a small fraction of the asset’s value should be allowed to be used as collateral (like maybe 20%). Letting insurers use a lot of leverage to cover their bets is the mechanism that amplifies risk to the financial system. Putting limitations on how swaps are covered will increase the cost of them, and reduce the demand for them, limiting their impact on the overall financial system.

Posted by KenG_CA | Report as abusive

Generic CDS’s will never become broadly used, because the accountants won’t allow it. Unless cashflows and term structure matches, swaps won’t qualify for hedge accounting. Sure, generics fine for speculators–but the big money is in hedging.

Read FAS 133 and IAS 39. The accountants got together and said that unless a derivative is bespoke, all g/l has to go through the income statement, while the hedged item remains accounted for at cost.

Oh well. Sure glad those folks at FAS and IASB are thinking about the second and third-order effects of their decisions.

Posted by Publius | Report as abusive

One reason euro authorities may not want CDS trigger is that it would be harder to claim that Greek bonds at ECB should be held at cost instead of market. Or put differently, if you marked the ECBs balance sheet to market today, how would all these sov bond purchases look?

Posted by Nicostrata | Report as abusive

Post Your Comment

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 http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/