Making sense of the market in US CDS

By Felix Salmon
October 3, 2013

Matt Levine had an excellent post last week on the bizarre market in credit default swaps on the USA — a market which people only ever look at during times of crisis or potential crisis. The nihilists are out in force today, using this market to confirm their priors, but the problem is that it’s very, very hard to look at US CDS, or to look at the yield on short-dated Treasury bills, and draw anything much in the way of meaningful conclusions.

One reason why is that Treasury bills are unique in many ways, including the weirdest way of all: as worries about the creditworthiness of the US government increase, the price of Treasury securities tends to go up, rather than down. Even if the US hits the debt ceiling, that won’t hurt the price of US debt; instead, general nervousness will only cause investors to flow into Treasuries and out of riskier assets. Which is to say, out of everything else.

And although Levine has managed to piece together a scenario under which a temporary technical default on US debt could cause a real payout for holders of US CDS, I don’t think that scenario really explains the price action either. The problem with it is that the government would still need to miss an interest payment on its Treasury securities, and there’s no way that it’s ever going to do that, whatever happens to the debt ceiling.

Think about it this way: if I roll over my debts, then my total debt does not actually increase. So if a T-bill is coming due today, then the government can pay it off in full, and issue a new T-bill, without increasing its total indebtedness. It’s true that a failure to raise the debt ceiling would prevent the government from funding its expenditures with new borrowing — although John Carney makes a good case that the government could just issue Obama Bonds instead.

The government still receives substantial tax revenues every week. So although the government would have to live within its means, spending no more than it got in revenues, its revenues would still be far greater than the total amount of debt service. And with Jack Lew (or anybody else, really) as Treasury secretary, you can be sure that debt service payments would be priority number one. US government payroll — especially for the president and Congress — would probably be the first thing to get cut; the armed forces might be next, just to place maximum pressure on House Republicans. Then Medicare and Medicaid, maybe — the doctors and hospitals providing those services would just have to wait until the debt ceiling got raised before they received their checks. Failing to meet any of those obligations would not be considered a debt default, and would not trigger CDS.

So what’s going on in the odd corners of the financial markets which are suddenly receiving so much attention? The simple answer is that they’re trading markets. They don’t only go weird when the debt ceiling approaches: something similar happened back in January 2009. And here’s what I wrote back then:

Anybody who bought protection at, say, 25bp is now sitting on a very nice profit if they close out their position. Maybe this is just a form of black swan insurance: buying US government CDS is a way of making money when everything else plunges in value. You’re not really insuring against an actual default, you’re just betting that if the world starts to implode, the price of your CDS is going to rise even higher.

No one knows exactly how high CDS rates would go if we pierced the debt ceiling, but it’s a reasonable assumption that they would go higher than they are now, even if (as is almost certain) they never pay out a penny. The US CDS market is a speculative, greater-fool market: the trick is to buy at a low level, and then sell at a higher level. A bit like bitcoins, really. If you think that the debt ceiling is going to be hit, then it makes sense to buy CDS today, just because spreads are going up rather than down. The only trick then will be trying to time the perfect moment to sell.


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

“…you can be sure that debt service payments would be priority number one.”

There is neither legal basis nor technical ability for Treasury to prioritize payments.

Posted by ScottHavens | Report as abusive

You’ll have to educate me, Felix. I thought that Treasury didn’t have the ability to prioritize payments, therefore first-in = first-out.

Posted by GRRR | Report as abusive

There is a debate over this, but the I think view of the Treasury was that they cannot prioritise spending, other than within the day. By the end of the day, all bills had to be paid. (This was in a court submission in the 1980s, from what I recall.) In practice, this would imply a default given the lumpiness of Treasury payments. There are plenty of Republicans who disagreed with that interpretation, but I do not know what their justification is.

There was a technical default on some retail Treasury investor products in the late 1970s. I saw that in an article on Forbes. (Debt ceiling problem, back office mixup over New Year’s.)

In any event, if the U.S. really defaulted with a big rightdown, would your counterparties really be able to pay up? I do not think this a market to take too seriously.

Posted by BrianRomanchuk | Report as abusive

FS: “[W]ith Jack Lew (or anybody else, really) as Treasury secretary, you can be sure that debt service payments would be priority number one.”

Reuters’ Jason Lange: “Many analysts speculate that the Treasury would give preference to some bills over others in an attempt to keep [from missing a debt payment], but a senior Treasury official told journalists on Thursday it would be impossible to prioritize payments on debt, as some Republicans on Capitol Hill have proposed.” [ 03/us-usa-fiscal-treasury-idUSBRE9920KA2 0131003]

I’m not sure what to make of the senior official’s “impossible to prioritize” comment. I came up with a few plausible possibilities:

(1) Impossible = “not legally authorized” — i.e, Treasury believes it would be illegal (e.g., unconstitutional) for it to prioritize payments.
(2) Impossible = “not technologically operationalizable” — i.e. Treasury believes that, even if it wanted to prioritize and was so authorized, it could not technologically make that happen.
(3) Impossible = “over my dead body [will I prioritize]” — i.e. Even if Treasury could technologically prioritize and was legally authorized, it simply would choose not do it. I can imagine a few plausible reasons for an unwillingness to prioritize: (a) Treasury thinks it would be politically unwise; (b) Treasury finds it unethical; (c) Treasury would rather risk credibility with bondholders than set a bad precedent–lest a future Secretary Cruz use it as a pretext for not paying Obamacare debts.
(4) Impossible = “I want you to think its impossible.” — i.e. Treasury is willing, authorized, and able to prioritize when the time comes. It’s merely (though wisely) bluster.

Of course, these aren’t mutually exclusive. And I have no idea which is most probable. Felix, you seem to think it’s #4.

But whichever it is, I find #2 (technological impossibility) to be the scariest. It implies that the proximate, technical consequence of credit ceiling inaction is… unknown. Worse still, it’s at the whim of a mammoth payment infrastructure that very possibly no one individual actually understands or is capable of understanding. If so, the consequence is unknowable. Somewhere, a series of machine logic gates is predestined to behave a particular way when Payment Zero is owing and The Balance is insufficient. Somewhere, an unknown mechanical arbiter lurks, controlling the fate of who gets paid what… or at all. Maybe the system will get trapped in an infinite loop, with no one getting paid. It’s like the Y2K uncertainty all over again! Perhaps that should be comforting. After all, Y2K was a dud. Alas, it isn’t.

Posted by Sandrew | Report as abusive

I think selling pressure on the USDollar will precede Tbond selling, so watch the trade-weighted value of the USD, not CDS prices on Tbonds.

Posted by TBV | Report as abusive

“Think about it this way: if I roll over my debts, then my total debt does not actually increase.”
Yes it does. US Debt is accounted on a principal basis only, rolling over implies paying the last coupon and that does increase the debt level.

Posted by alea | Report as abusive

At this time the areas getting the most activity are shady areas. Trying to protect the casino bets. Buying delusional insurance to pay for a bet on which way a particular market will fall. Then create the situation you want to happen (or rigging) the market to collect on the insurance that is bought and paid for buy USA workers. It is like a cheater at a casino who knows how to count cards allowing him to rob the house. At this point, money is just binary symbols bouncing around a cow casino.

Posted by 2Borknot2B | Report as abusive