As Congress looks increasingly likely to force a default on U.S. government debt for no real reason, there are two possible approaches for a rational outside observer to take, which are (1) despair and (2) hare-brained scheming. Option 1 is probably correct, honestly, and you can get your fill of it elsewhere; Martin Wolf's despair is eloquent. But Option 2 -- coming up with creepy tricks to avoid reality, reality being the debt ceiling and a pointless default -- is more fun so let's talk about it here.

The creepy trick that has swept the nation* is the platinum coin option, in which Treasury mints a $1 trillion platinum coin, deposits it at the Fed, and suddenly has an extra $1 trillion of money to spend without incurring any debt (and, thus, without breaching the debt ceiling). This is a good trick as tricks go, and it's been extensively advocated by Josh Barro, Paul Krugman, Matt Yglesias, Joe Weisenthal, basically every economics blogger really. I am unaware of any good arguments that the platinum coin wouldn't work, but it does have the problem that it is really really really really obviously a trick. I mean, it's a trillion dollar coin, come on. So it's sort of sub-optimal symbolically, and would make people really mad. It's a crisis-enhancer, although with the benefit of avoiding immediate default.

My preferred creepy trick is creepier and less of a trick and it goes like this:

  1. The U.S. government takes in $277 billion in tax revenues each month, and spends $452 billion each month, for a monthly deficit of around $175 billion.**
  2. It also has, on average, call it $100 billion of Treasury notes coming due each month.***
  3. Instead of just rolling those Treasuries -- paying them off at 100 cents on the dollar by issuing new Treasuries at 100 cents on the dollar -- it should pay them off at 100 cents on the dollar by issuing new Treasuries at 275 cents on the dollar and using the extra money to pay its bills. The 10-year yield today is around 2.6 percent, so you could sell a 10-year with a 23 percent coupon for 275 cents on the dollar.**** The 30-year is about 3.9 percent, so a 14 percent coupon should get you there. Etc. Math here.
  4. That's it. You aren't adding debt, so you never hit the debt ceiling, but you keep getting more money.

This idea is not original to me -- several people on Twitter have championed it -- and I wrote a long post about it during the last debt ceiling crisis. You should read that post if you want the long/technical version of the argument for why this works, but the short version is: I think it works. The debt ceiling applies to the face amount of bonds, not the amount raised, so selling a $100 bond for $275 only counts $100 against the debt ceiling and gets you $175 in debt-ceiling-free money. There are rules against issuing premium Treasury bonds, but the rules are just Treasury rules and they can be changed unilaterally by Treasury with no notice and no Congressional approval. So Treasury could do an auction tomorrow seeking to sell $100 billion of 23 percent bonds for $275 billion and as far as I can tell no one could stop them.

Of course no one would have to buy them either, and you might be skeptical that anyone would. I don't think there's any reason to be too skeptical though. Yes, high-coupon bonds are weird.***** But the interest payments here, though high, would be full-faith-and-credit-of-the-United-States obligations. Those payments are no easier (or harder!) to default on than any other Treasury obligation. The Treasury market is already accustomed to separating out interest and principal payments into just a string of zero-coupon obligations, so it should be perfectly well equipped to handle this weirdness.******

Does this cost the government money? Meh, not really. The "real" cost of these bonds is not the sticker coupon; it's the yield, and the yield shouldn't be much more than the yield on regular-way Treasury bonds.******* They'd effectively be amortizing bonds: Instead of borrowing $100, paying a little interest each year, and then paying back the $100 all at once in ten years, Treasury would borrow $275, pay back $17.50 per year for ten years (plus interest), and then pay back the last $100 at the end.

I think this basically works and, as lunatic schemes go, it gives off less of a banana-republic-from-outer-space whiff than the platinum coin does. That's partly because it looks like a banana-republic-from-earth sort of scheme. Because it is! This happens! Countries really do enter into contracts that require high future payments but that create less formal "debt," in order to get around restrictions on their debt. Greece did exactly that, entering a swap with Goldman that required large future payments but that reduced its official debt-to-GDP ratio in order to formally comply with EU treaties.

That's what this is: Incur large future payments to reduce our official debt numbers in order to formally comply with the debt limit. Unlike the platinum coin, this scheme is practically normal, if by "normal" you mean "scandalous even in Greece." Which, I mean: really is pretty much normal for the U.S. government these days.

* The nation of Twitter, I mean, ugh.

** Source is Bloomberg's useful CLIF [go].

*** Source is eyeballing T [Govt] [go], sorting by maturity. There's around $93.7 billion in October, $127.9 billion in November, $98.6 billion in December, $99.6 billion in January, $130 billion in February, etc.:

This is not, like, a *great* visual aid.
This is not, like, a *great* visual aid.

**** Really like 22.6 percent, but my math adds a 25 basis point yield bump for general creepiness, which I think is fair-to-generous. See the footnote below -- there's an argument for that bump being zero or even negative.

***** Why? Well, mainly because a high-premium bond has bankruptcy risk: If you buy a $100 bond for $200 because its coupon is high, and the issuer goes bankrupt, you only have a $100 claim in bankruptcy. Does this matter where the issuer is the U.S. government? I mean, let's hope not. If the U.S. government actually does an actual restructuring of its debt, your par claim is the least of your problems.

This high-premium problem flows through various mechanisms of the bond market, e.g. dealers don't like to give as-good financing on high-premium corporate bonds as they do on par bonds, but again that rationally shouldn't matter here and I'm not aware of anywhere where that's true of Treasuries.

You can eyeball existing high-premium vs. par Treasuries. There are no Treasuries trading at 275 cents on the dollar but look here:

Red boxes highlight pairs of high-premium and par-ish bonds, guess which is which.
Red boxes highlight pairs of high-premium and par-ish bonds, guess which is which.

I've highlighted pairs/triplets of 7-to-10-year notes with the same maturity but different coupons. All of the high premium bonds trade at lower yields than the low-premium/par/discount bonds. [Update: Because, of course, they have a lower duration -- you get more of your money back quicker because the coupon is higher. This argument applies even more strongly to my creepy bonds, suggesting that my math is way way too generous. You could probably get away with a 22 percent coupon on the 10-year.]

****** Other flavors of weirdness, like moral-obligation Obama-bonds that are not authorized by statute and backed by the full faith and credit of the United States, are probably harder to handle. But these are just totally regular Treasuries, just with high coupons.

******* Incidentally tax law follows economic reality here: If you're a taxable investor and you buy these bonds, you don't treat your full 23 percent coupon or whatever as taxable interest; it's mostly basis recovery and your taxable interest income is effectively the economic yield. So that's not a reason not to buy them.