Felix Salmon

Bill Gross’s Investment Outlook this month includes the normal mix of hyperbole and mixed metaphors ("securitized WMDs", "the pyramid begins to unravel"). But get past that, and you’ll find him saying that the CDS (credit default swap) market poses a very serious systemic risk. I think (but I’m not positive) that he’s wrong on this one.

First, let’s be clear about the argument that Gross is not making, or at least not making here – and that’s the counterparty risk argument which always gets brought up in discussions of derivatives. The way that a CDS works, there are two counterparties, the protection buyer and the protection seller. The protection buyer pays a small "insurance premium" to the protection seller every six months. In return, the protection seller promises to pay the whole face value of the insurance policy to the protection buyer in the event that a given credit defaults. As part of the deal, the protection buyer also has to fork over some bonds, which Gross assumes are worth roughly half of what the protection seller is paying.

Now the protection buyer thinks that he’s protecting himself against default. But what if the credit defaults and the protection seller has no money? That’s known as counterparty risk, and it’s a serious issue. But it’s not the issue that Gross raises. Gross is saying that losses alone on the part of protection sellers could be a very serious matter, not just outright bankruptcy and default. Most protection sellers are big banks and insurance companies, and while the present situation is serious, I don’t read him to be saying that those banks and insurance companies are likely to go bankrupt in 2008.

Here’s what he says:

Let’s go back to the $45 trillion BIS estimate of outstanding CDS for more insight. If total investment grade and junk bond defaults approach historical norms of 1.25% in 2008 (Moody’s and S&P forecast something close) then $500 billion of these default contracts will be triggered resulting in losses of $250 billion or more to the "protection selling" party once recoveries are inserted into the equation… Of course, "buyers of protection" will be on the other "winning" side, but the point is that as capital gains and capital losses slosh from one side of the shadow system’s boat to the other, casualties and shipwrecks are the inevitable consequence. Goldman Sachs wins? Fine, but the losers in many cases will not be back for a return match… You have a recipe for credit contraction, a run on the shadow banking system… The unfairly "Ben Stein pilloried" Jan Hatzius of Goldman Sachs estimates that mortgage related losses of $200-400 billion alone might lead to a pullback of $2 trillion of aggregate lending. Even if this occurs gradually, he writes, "The drag on economic activity could be substantial." Add to that my $250 billion loss estimate from CDS, as well as prospective losses in commercial real estate and credit cards in 2008 and you have a recipe for a contraction in credit leading to a recession.

First, of course, I should thank Gross for jumping onto the Ben Stein bandwagon.

But think about what Gross is saying here. He’s well aware, of course, that the CDS market, like any derivatives market, is a zero-sum game (if you ignore counterparty risk for the time being): that’s why he says that "of course" there will be a "winning side". But, like Floyd Norris, he’s worried about gross losses, not net losses.

Now in a system where everybody is perfectly hedged, gross losses = net losses = 0. All of the CDS contracts cancel each other out, and no one has any net exposure to any default at all, since everyone buys offsetting protection on every credit that he sells protection on.

So let’s consider the polar opposite of that system, where there’s a total distinction between protection buyers and protection sellers. As defaults rise, capital will slosh, as Gross puts it, from the sellers of protection to the buyers of protection. The sellers of protection will lose hundreds of billions of dollars in capital, and all manner of nasty consequences result.

But let’s take another look at exactly what direction the capital is sloshing in. Yes, the sellers of protection will be remitting $250 billion to the buyers of protection as a result of those 1.25% of credits which have defaulted. But at the same time, the buyers of protection will be remitting all of their insurance premiums to the sellers of protection on the 98.75% of credits which haven’t defaulted. And on $45 trillion of CDS, those premiums are likely to add up to a lot of money – more, I’ll hazard, than $250 billion.

In other words, if default rates stay relatively low – and a 1.25% default rate is relatively low – then, in aggregate, the sellers of protection are likely to continue to make money, not lose it.

Of course, if you only sold protection on subprime mortgage bonds or CDOs, or you’re unlucky enough to have sold protection only on those particular credits which end up defaulting, then you’re liable to lose a lot of money. But we’re not talking about anything close to $250 billion, here.

And so I’m far from convinced that, as Gross would have it, $250 billion of theoretical CDS losses should be added to $200 billion or $400 billion in genuine subprime-mortgage losses. In subprime, loans were made and defaulted on – that’s real money down the drain. The CDS market, by contrast, is a zero-sum money-go-round where some people win and some people lose, but which isn’t going to have anything near the same systemic consequences.

Now, how could I be wrong about this? If the implied default rates on CDS contracts written over the past few years were much lower than 1.25%, the net capital flows might well be from protection sellers to protection buyers, rather than the other way around. Does anybody have that figure?