News

# Instead Of A Trillion-Dollar Platinum Coin, Treasury Should Mint A Trillion-Dollar \$500bn Bond

Here are four problems in the world:

If I told you we could solve all of them in one fell swoop, what would you say? Would you say “shut up about that fucking platinum coin“? Ha of course you would. But it’s not that! It’s this, from pseudonymous1 blogger Sonic Charmer:

Could the Treasury just get around the debt limit by selling super-high-coupon bonds? (Say, 30% …) So they auction at a huge premium and Treasury earns proceeds that are a multiple of face value. Face value = low = stay under debt ceiling since it’s based on ‘money borrowed’ which is face value (?), but proceeds = high = can pay the ‘bills we’ve racked up’ (spending plans). Tune sizes/coupons/durations as required.

So: tentatively endorsed. Prove him (and now me!) wrong and you win a prize, the prize being a catastrophic government default. Let’s rule out some easy objections:

1. Math. The math is fine, easy even. Treasury takes in \$277 billion in revenue a month, and pays out \$450 billion in expenses, leaving a \$173 billion shortfall. I am too dumb to figure out how much Treasury debt matures each month but some random eyeballing on Bloomberg (T <Govt> go) shows around \$118bn of notes/bonds maturing in February 2013, \$98bn in March, \$101bn in April, \$118bn in May, \$94bn in June, \$94bn in July, \$129bn in August, \$95bn in September … let’s just ballpark it, ignoring bills, that \$100bn of Treasuries roll each month. If you sold \$100bn face amount of Treasuries each month for oh say \$273bn you’d be fine. I see the 10-year at a 1.836% yield and the 30 at 3.026%; by my math you could sell 10-years with a 12.3% [haha what?] 24.5% coupon or 30-years with a 9.3% 18.75% coupon and clear enough proceeds to make this work.2

2. The debt limit. Naah that’s fine. Here is the statutory debt limit:

§3101. Public debt limit
(a) In this section, the current redemption value of an obligation issued on a discount basis and redeemable before maturity at the option of its holder is deemed to be the face amount of the obligation.
(b) The face amount of obligations issued under this chapter and the face amount of obligations whose principal and interest are guaranteed by the United States Government (except guaranteed obligations held by the Secretary of the Treasury) may not be more than \$14,294,000,000,000, outstanding at one time, subject to changes periodically made in that amount as provided by law through the congressional budget process described in Rule XLIX 1 of the Rules of the House of Representatives or as provided by section 3101A or otherwise.
(c) For purposes of this section, the face amount, for any month, of any obligation issued on a discount basis that is not redeemable before maturity at the option of the holder of the obligation is an amount equal to the sum of—
(1) the original issue price of the obligation, plus
(2) the portion of the discount on the obligation attributable to periods before the beginning of such month (as determined under the principles of section 1272(a) of the Internal Revenue Code of 1986 without regard to any exceptions contained in paragraph (2) of such section).

As a human and a pretend lawyer3 I read that to say:

• For bonds not issued at a discount, the debt limit applies to the face amount of the bond, not its issue price.
• For bonds issued “on a discount basis” the debt limit applies to the redemption price, if puttable, or the issue price plus accreted OID, if not puttable.

And as a human and a pretend lawyer I submit to you that (1) a bond issued at a premium is ipso facto not issued “on a discount basis”3A and (2) even if somehow it is (a negative discount!) if you just make it puttable (at par) you get to treat it as a par bond. Meaning that the statutory debt limit by its terms applies to the face amount of debt, not the issue price.

3. Other statutes. Seem okay. 31 U.S.C. §§ 3102 & 3103 cover issuance of Treasury bonds and notes. Section 3102 is particularly worth reading; the authorization is broad:

With the approval of the President, the Secretary of the Treasury may borrow on the credit of the United States Government amounts necessary for expenditures authorized by law and may issue bonds of the Government for the amounts borrowed and may buy, redeem, and make refunds under section 3111 of this title. The Secretary may issue bonds authorized by this section to the public and to Government accounts at any annual interest rate and prescribe conditions under section 3121 of this title.4

It’s also worth reading the history that follows the text, which includes striking out previous restrictions on (1) interest rates above 4.25% and (2) issuance at less than par. (There do not seem to have been restrictions on issuance at more than par.) The historical movement has been towards more, not less, flexibility for the Treasury Secretary to decide what sorts of bonds to issue.

4. Treasury issuance mechanics. Here we are on shakier ground, but not as bad as it looks. First, the bad: Treasury’s rules exclude above-par auctions of bonds with interest rates above 0.125%:

(b) Determining the interest rate for new note and bond issues. We set the interest rate at a 1⁄8 of one percent increment. If a Treasury note or bond auction results in a yield lower than 0.125 percent, the interest rate will be set at 1⁄8 of one percent, and successful bidders’ award prices will be calculated accordingly (see appendix B to this part for formulas).

(1) Single-price auctions. The interest rate we establish produces the price closest to, but not above, par when evaluated at the yield of awards to successful competitive bidders.

(2) Multiple-price auctions. The interest rate we establish produces the price closest to, but not above, par when evaluated at the weighted-average yield of awards to successful competitive bidders.

So Treasury’s rules mostly prohibit premium bonds. This, however, does not seem to actually be a problem, since these rules are just Treasury’s rules and it can change them. (By, for instance, executive order.) Here is the adopting release for that 1/8th point rule; note that (1) it cites no statutory requirement and (2) it notes that “Because this rule relates to public contracts and procedures for United States securities, the notice, public comment, and delayed effective date provisions of the Administrative Procedure Act are inapplicable.” That means: Treasury can just change this rule whenever it wants. Now would be a good time.

Also – as noted in that release – the rule does not apply to reopenings, and in fact Treasury has done above-par reopenings in the past; here’s one at 104.5+. A reopening – just selling more of an already-existing series of high-coupon Treasuries – would be the simplest way to do this scheme, since it would require no rule changes or departures from ordinary practice. And there are some high-dollar Treasuries to reopen.5

On the other hand, “high-dollar” means like 150s, not 270s: there are no really high-dollar Treasuries to reopen and solve the whole problem. And if you wanted to do weird tweaks like making them puttable (see point #2 above) or deferring interest (see fn. 2) then you’d probably want a new issue, which would require changing some (purely Treasury-written and instantly-changeable) rules. In any case, the fact that above-par reopenings are kosher would seem to undercut the idea that there’s any sort of statutory ban on above-par Treasury issuance.

5. Market reaction. I dunno. It’s a little goofy, no? Presumably the boring sorts who buy Treasuries won’t enjoy the banana-republic thrill ride of buying a \$100 Treasury with a 18% coupon and a \$273 price. On the other hand … I mean, there are high-dollar off-the-run Treasuries out there that seem to trade a bit, so someone can buy them. And, hey: 18% coupon! That looks sort of nice, right? Right?

* * * * *

So what do you think? Is the above analysis wrong? (Where?) Is this a dumb idea? Is it dumber than the platinum coin? Meh. One thing I like about it is that it doesn’t really usurp any powers or change any procedures outside of some internal Treasury mucking about: Treasury just keeps on doing exactly what it’s been doing, but now it takes in a bit more money for doing so. Unlike the platinum coin, it’s not really in-your-face; you could do it almost silently.

Of course there are disadvantages too. The platinum coin doesn’t pay interest; the premium bond pays very high interest. This isn’t that big a deal – the economic cost is the yield, not the coupon; the high coupon can be thought of as basically just amortization of the amount initially borrowed. And presumably the yield will be reasonably close to that of par Treasuries. (Premium bonds often trade at a bit of an extra yield, though this is partially a matter of recovery value on default, which shouldn’t matter for Treasuries … right?) But the accounting cost matters too, since this is an accounting dodge more than it is an economic trade, and the accounting difficulties caused by paying 20ish% a year (of face value – like 7% of money raised) makes this just a short-term solution. The PR isn’t great, either – the US paying 20% on its bonds? Feh. Selling small-premium bonds (in e.g. reopenings) to get a little headroom would look a lot better than actually going out and trying to sell bonds at three times their face value, but a little headroom doesn’t seem to be what’s needed.

Mostly, though, I like the symbolism. Unlike the platinum coin, which presumably would be stamped with the words THIS WAS KIND OF A JOKE, the premium Treasury comes with its own elegant economic logic. Perhaps the most absurd part of the absurd debt ceiling drama is that uncertainty about whether the U.S. government will just voluntarily default on its debts actually drives people to buy Treasuries: in a time of uncertainty, the thing that investors value most is the safety of U.S. government bonds.6 Might as well make some money off of that, no?

1. Be warned, here begins my most pointless footnote ever. First: “pseudonymous” or “anonymous” or other? Does “pseudonymous” require, like, a name that could be a human name?

Impoco’s departure was a surprise to some. It had been expected that he would be moving away from the digital side to concentrate full-time on the eponymous luxury magazine he launched for Reuters last year …

do you think “wait the magazine is called ‘Impoco’?” or is that just me?

2. Of course then you’d have to pay the interest! This would be a fairly short-term solution; after a year of doing this you’d be incurring an extra \$250+ billion in interest payments that you’d have to fund by issuing ever-higher-coupon Treasuries, leading I suppose to a death spiral. If I were doing this on a tabula rasa I’d probably bump the interest rate a bit at the cost of deferring payments for a year or two, making this a breathing-room rather than permanent solution, but obviously the further you get from regular-Treasury-bond structure the weirder this looks.

3. Btw a lot of the references here come from this extremely helpful “Law & Guidance” page at Treasury Direct. Here’s what they say:

Are you a lawyer, or a financial advisor, advising a client who owns Treasury securities, or a law student researching a paper? Or do you just want to read the laws and regulations on which our programs are based?

… Or are you an annoying blogger with an impractical unrealistic proposal? Or maybe you’re just a crackpot who thinks that Treasury bonds are a fraud because the Pope was illegitimately elected by the Illuminati? YOU’VE COME TO THE RIGHT PLACE.

3A. Actually I get the distinct sense from other bits of the law (e.g. § 3121) that “on a discount basis” means “with zero coupon,” like bills, but it’s probably unprofitable to go too far down that path.

4. Section 3121 is non-restrictive and particularly says that “the Secretary of the Treasury may prescribe (1) whether an obligation is to be issued on an interest-bearing basis, a discount basis, or an interest-bearing and discount basis; … [and] (3) the offering price and interest rate.”

5. Bloomberg shows me a 7.625 of 2025 (CUSIP 912810ET1) trading at 159, which I think is the champion.

6. Though conversely there’s also a nice message in “you won’t raise the debt ceiling? Fine. Now we have to issue Treasuries at a 20% coupon.”

### 32 Responses to “Instead Of A Trillion-Dollar Platinum Coin, Treasury Should Mint A Trillion-Dollar \$500bn Bond”

1. Guest says:

Isn't this pretty much standard procedure in the new issue muni market?

2. Dobo says:

Charles Ponzi would be proud

3. NameEmail says:

I just scanned the comments, and wow what a bunch of immature butt heads. I'm from the Vietnam Nam era and this post tells quit a few stories no one will ever hear.

How long did it take you to scan both of them? I'm guessing quite a while.

• PermaGuestII says:

You mean the VietNam Nam era era?

4. WvvZvvW says:

You've got to be kidding me! Who's gonna lick a stamp that has Bill Clinton on it? Let a lone buy a bond. Home of the free indeed.

5. Matt for SEC Chair says:

I guess i'd rather have Matt take care of this then any of our appointed officials, meh, would like some charts though, also funny pictags are nice, Bernake never seems to include those.

6. passing pedant says:

This does not in fact solve any problem of "quest for yield", since all the math to calculate the issue price is based on the assumption that the high-coupon bonds would have exactly the same yield as normal treasuries.

7. Yona says:

Why don't they mint a gold ring with "1,000,000,000,000 USD" printet on it! One ring to pay them all! … plus they may get some money as advertising campaign…

• Lord of the Bling says:

And call it "The Precious"

• guest says:

Because gold coins aren't commemorative…

8. IR quant says:

Not happy with paying par for 3% 30y that has negative real yield and is guaranteed to lose lots of money when rates rise?
This new thing is guaranteed to lose 10% of its value with every rate hike – for 30 years! Don't all rush at once

• Anon. says:

• PermaGuestII says:

A 30yr with an 18.75% coupon would have a duration of something like 14yrs– so you'd lose less money with every rate spike than you would buying the current on-the-run 30 with a duration of 20yrs.

-actual fixed income guy

• passing pedant says:

it is also not going to stay a 30 yr bond for 30 years. in ten years, it will be a 20yr bond.

9. nachocheese says:

can I use this coin on the garden state parkway?

10. Shortofaname says:

What about taxes? The buyer will have to pay ordinary income tax on the inflated interest payments, but have a long term capital loss when the bond matures or is sold after a few years.

• anonhfm says:

11. anonhfm says:

If we accept the possibility that the US can default, then one would demand a yield premium for a high price bonds due to higher percentage loss given default (that is, even in the sovereign context, people tend to assume flat recovery as percentage of face). Even if you don't buy the story of flat recovery as a percentage of face, you would only be able to deliver the high-coupon Treasury into CDS on the basis of its face amount. Obviously, since the probability of default of the US is very low, the yield premium demanded for the high-coupon bond would probably be quite small.

12. bond noob says:

For someone who isn't finance-savvy, I need some answers to Matt's math on his spreadsheet.

-Where is he deriving the coupon rates from?

-Secondly, why isn't he dividing the given treasury rates by 2? I am confused as to why he is using 20 periods (listed 10 yrs x 2 ) but not dividing the given rate? Is that because the treasury rate is already assuming the semi-annual pmts? Hence the crossed out initial coupons?

If posting a question of this noob nature doesn't get me banned from dealbreaker or eviscerated by the commentariat, I would like an answer

• guest says:

Mostly it will just get you ignored.

• 1,000 Crickets says:

Yup.

• Guest says:

Not Matt, but in (non fixed income) finance, so here's a shot at answers:

1) He's not deriving them, he's plugging and getting to the approximate number he wants in pv terms.

2) He kind of is. In the formula =-pv(C5,C3*2,C4*100/2,100), the C3*2 figure represents the semi-annual payments and the C4*100/2 represents the payment amounts. The key is the "/2" at the end, which is doing what you want to do by dividing the interest rate in half.

13. dandraka says:

WARNING : POINTLESS GREEK LANGUAGE COMMENT (yes, even more pointless than footnote #1)

"Does “pseudonymous” require, like, a name that could be a human name?"

Yes it does : pseudonymous (ψευδώνυμο) = ψευδός (i.e. false) + όνομα (name).

E.g. Diane Keaton is a pseudonym. 50 Cent or Kirk Douglas is not.

14. sohbet says:

ortant ways unknowable: you could list what the stuff was, if you had a big enough piece of paper, but it was hard to get a handle on what it wo

15. sohbet says:

oyment to talk to your publish. It is well-articulated, razor-sharp, along with goal. What beneficial instruction you may have learned that you put onto this kind of release. Searching to

16. lish. It is well-articulated, razor-sharp, along with goal. What beneficial instruction you may have learned that you put onto this kind of release. Searching to

17. chat says:

turkish numberone chat room's!

18. sohbet says:

turkishchat.net

19. chat says:

turkish chat

20. sohbet says:

officers to investigate Blackstone's fishy financial transactions, but they wound up giving them

21. A nice one there. …:)