On Thursday I said that the US is not going to default on its bonded debt, even if the debt ceiling is reached: "with Jack Lew (or anybody else, really) as Treasury secretary, you can be sure that debt service payments would be priority number one".
This is not a popular view within the blogosphere, maybe because it's generally associated with Republicans trying to say that hitting the debt ceiling wouldn't be that bad. Both Cardiff Garcia and Dylan Matthews have come out with sterling attempts to answer the question of whether debt prioritization is even possible; Danny Vinik, for one, says that there's "pretty good evidence to demonstrate that prioritizing debt payments is not possible". The problem, however, as Garcia says, is that most of the primary sources you'd want to go to on a question like this "are vague and unhelpful".
It's worth stipulating up front that hitting the debt ceiling would be disastrous even with prioritization: Garcia calls it "breaking the economy's knees with a fiscal crowbar", while Paul Krugman says that it would be "a catastrophe". But it would be much better than the truly apocalyptic state of affairs that we would see in the event of a Treasury bond default. Deutsche Bank says that in that event, the S&P 500 would fall some 45% - and, boldly, puts a 0% probability on that actually happening.
It's also worth stipulating that before the debt ceiling is hit, a lot of very sensible politicians want to make prioritization seem as unlikely as possible, because that maximizes the incentive to avoid hitting the ceiling at all. On the other hand, after the debt ceiling is hit, the very same politicians should be willing to move heaven and earth to ensure those bond coupons get paid.
So, why is Matt Yglesias so convinced that prioritization is impossible? He gives four reasons.
The first is that prioritization is illegal: "Treasury is not authorized to unilaterally decide to pay certain bills and not others". This is true - but also a bit irrelevant. Treasury is under unambiguous Congressional orders to pay lots of bills - all of them, in fact. If it fails to pay those bills, it will be violating the law as laid down by Congress. Hence the 14th Amendment argument that the president should simply ignore the debt ceiling entirely, if it comes to that. But underneath it all, it's hard to credit any argument which says "Treasury isn't allowed to pay its own bonds". If that's what Treasury wants to do, then surely it can do so. Besides, who would even have standing to sue?
Yglesias's second reason is that prioritization is just not feasible: it can't be done in the real world. Both he and Matthews cite the Treasury inspector general, who does indeed say what they say he says:
Because Congress has never provided guidance to the contrary, Treasury's systems are designed to make each payment in the order it comes due.
It's worth reading the whole letter, however, because the inspector general says a lot more than that. And while the systems are designed to make payments in the order they come due, they have also been designed so as to effectively insulate bond repayments from all other payments. Bond repayments are made through a system called Fedwire, while all other payments are made through the standard banking ACH system. Logistically, it's entirely possible to keep up to date on all Fedwire payments without making any ACH payments at all.
And the inspector general was very careful to keep all options open:
Ultimately, the decision of how Treasury would have operated if the U.S. had exhausted its borrowing authority would have been made by the President in consultation with the Secretary of the Treasury.
What's more, the inspector general does rather fudge the central issue of prioritization, which is whether debt repayments can carry a higher priority than everything else. "Treasury officials determined that there is no fair or sensible way to pick and choose among the many bills that come due every day," he said - but that's false on its face: prioritizing debt repayments is very sensible, since defaulting on Treasury bonds would be much more harmful than simply paying all bills as they come due, whether they're a bond coupon or a fighter jet.
There is an argument from the left that prioritization constitutes "paying China first", and would "require the government to cut large checks to foreign countries, and major financial institutions, before paying off its obligations to Social Security beneficiaries and other citizens owed money by the Treasury". Well, yes. But I don't think anybody in Treasury is swayed by such arguments: they know that in the grand scheme of things, all Social Security beneficiaries would be much better off receiving their money in arrears than they would be if Treasury defaulted on US sovereign bonds.
Yglesias then rolls out the timing argument, which is further developed by Zero Hedge: debt repayments are lumpy things, and it would be hard to "save up" enough money before the big repayments were due, if you were paying any other bills at all. Zero Hedge improbably says that "Treasury will simply halt new Bill issuance" if the debt ceiling is reached, but I don't buy it: no one's requiring that the national debt go down. And investors generally want to be able to roll over their short term debt: failure to be able to do so would be better than default, but not much.
Could Treasury decide to prioritize Fedwire payments, and then turn on the ACH payments sporadically, only insofar as they didn't eat up enough cash to endanger bond repayments? I don't see why not. Treasury wouldn't like it, of course. And as Yglesias says in his final point, such a scheme might well be so messy that the markets would have to end up assigning some kind of credit risk to Treasury bonds anyway. Still, doing so would send a very clear message to markets, that Treasury cares about them more than it cares about the sick, the elderly, or any other recipients of government funds. And the markets, in return, would probably reward Treasury with lower interest rates on Treasury bonds. After all, in a crisis, money always flows into Treasuries - even when it's a Treasury-bond crisis.