That Darn Coin

A coin in decay.

They say that a bad penny always turns up. But when it comes to crises these days, it seems that what keeps turning up is a bad idea—namely, the idea of having the U.S. Mint strike one or more trillion-dollar platinum coins.

As I explained last March, the idea, which was first broached in 2009 and has since become very popular among Modern Monetary Theorists, gained prominence in January 2013, when they and several more orthodox economists latched onto it as a way around that month's debt ceiling crisis.

Paul Krugman, who was one of the idea's proponents, observed:

Should President Obama be willing to print a $1 trillion platinum coin if Republicans try to force America into default? Yes, absolutely. He will, after all, be faced with a choice between two alternatives: one that’s silly but benign, the other that’s equally silly but both vile and disastrous. The decision should be obvious. …[B]y minting a $1 trillion coin, then depositing it at the Fed, the Treasury could acquire enough cash to sidestep the debt ceiling—while doing no economic harm at all.

So why not?

It’s easy to make sententious remarks to the effect that we shouldn’t look for gimmicks, we should sit down like serious people and deal with our problems realistically. That may sound reasonable—if you’ve been living in a cave for the past four years. Given the realities of our political situation, and in particular the mixture of ruthlessness and craziness that now characterizes House Republicans, it’s just ridiculous—far more ridiculous than the notion of the coin.

In the event, the Treasury put the kibosh to the big coin plan, and Congress went on dickering about the Federal debt until October 17, 2013, when the crisis came to an end with the passing of the 2014 Continuing Appropriations Act.


Now the idea has come back again, this time as part of the "Automatic BOOST to Communities Act," sponsored by Congresswoman Rashida Tlaib (D-MI). The BOOST Act's goal is to "provide a U.S. Debit Card pre-loaded with $2000 to every person in America," where the cards could be "recharged with $1,000 monthly until one year after the end of the Coronavirus crisis."  Instead of having the Treasury finance the plan by selling securities, the bill calls for it to use "its legal authority to create money via coin seigniorage, which is a statutory delegation of Congress’s constitutional power of the purse."

Because of the sum involved, the plan would have the U.S. Mint strike not one but two platinum coins, each containing one ounce of platinum but having a face value of $1 trillion. It would also compel the Fed to purchase the coins for their face value, by crediting the Treasury General Account by that amount.

Finally, the bill says that the plan it proposes "would preserve the historical separation between fiscal and monetary policy and avoid financial entanglement between the Treasury and the Federal Reserve which would eventually undermine the independence of the Fed."

We shall see about that.

Why Not Borrow?

When the platinum coin idea was being pitched back in 2013, it was, as I've noted, as a way to get around the debt ceiling, which many viewed as a serious impediment to needed fiscal stimulus. Allowing that the stimulus was indeed desirable, the idea had the merit of solving a real problem, even if it did so in a somewhat devious way, and one that risked establishing a precedent that could lend itself to future abuse.

Today, in contrast, there's no debt-ceiling impediment to aggressive fiscal action: as part of last year's budget deal, the debt ceiling was suspended until July 2021. That leaves Congress and the Treasury free to spend as much as they want, without having to compel the Fed to pay $2 trillion for a couple platinum coins that, so far as it's concerned, wouldn't be worth as many plugged nickels.

If platinum coins aren't needed to replenish the Treasury's coffers, why should Congress bother with them? It's a good question, and especially so considering that rates on many Treasury securities are now at record low levels. What's more, rates on some shorter-term Treasury bills are actually below the meager 10 basis points the Fed now pays on bank reserves.

Taken together with the fact that we're entering what's likely to be a deep recession with no immediate risk of "crowding out," what all this means is that the Treasury might actually save money by selling T-bills instead of making the Fed buy platinum coins from it. That's partly because it costs money to make the coins, including about $1250 bucks worth of platinum. But that's the least of it. The real difference is that, once the funds created get disbursed by the Treasury, and thence spent by their recipients, the coin strategy will ultimately increase bank reserves by about $2 trillion. The Fed will then have to pay interest on those reserves at the IOER rate, and so will have to deduct that amount from its Treasury remittances. If instead the spending program were bond financed, bank reserves would stay unchanged.

In short, under the current, "floor" system of monetary control, coin financing is equivalent to having the Treasury borrow the face value of the coins from the nation's banks, while paying them interest at the IOER rate, and wasting perfectly good platinum to boot.

Coins and Copters

As Greg Ip has observed, the platinum coin plan resembles a Treasury-initiated version of "helicopter money":

Banks won’t want a $1 trillion platinum coin, so the Fed will only buy the coin if Treasury forces it to. The Treasury, in “depositing” its coin at the Fed, is in reality ordering the Fed to print money. And if Treasury doesn't take the coin back, the money stays printed.

The helicopter money nature of coin-financed spending might make it appear capable of achieving a greater stimulus effect than ordinary debt-financed spending. But here again, the fact that bank reserves bear interest throws a wrench into the works, as Narayana Kocherlakota observed several years ago. It means that coin financed spending confronts taxpayers with the same interest burden they would bear if the Treasury financed its expenditures by selling consols with coupons pegged to the floating IOER rate.  Biaggo Bossone makes the point especially succinctly, and in a way fans of Modern Monetary Theory should find especially persuasive. "A positive remuneration on excess reserves," he says, "transforms them into fiscal liabilities (analogous to debt financing) in the consolidated public sector balance sheet."

Exposing a Charade?

If the coin gambit may not allow the Treasury to borrow for less, and isn't likely to accomplish anything that plain-old deficit spending can't, what use is it?

I posed this question on Twitter to Clint Ballinger and Nathan Tankus, two very smart Modern Monetary Theorists. And both said, in essence, that the plan's real purpose is educational. According to Ballinger, it would shine a light on the sheer silliness of the Treasury's "self-imposed rule" against simply creating and spending its own "tax-credit token[s]." Tankus likewise sees the plan as a way of "contesting the idea that the Treasury is not a monetary institution." The coin gambit is not, in other words, a way of allowing the Treasury to create money. The Treasury already creates money. It just does it in a "Rube Goldberg" fashion (Ballinger) rather than openly. While Fed officials may consider monetary policy their responsibility, the Treasury really calls the shots.

But does it? That would be so were the "self imposed rules" that Modern Monetary Theorists dislike merely so much window dressing. But they aren't. That the Treasury has thus far not considered the striking of trillion-dollar coins a legitimate exercise of its authority, that it supplies other coins only to meet the Fed's requests for them, and that it is not allowed to overdraw its TGA account, all have real implications. They mean that the Treasury is incapable on its own of expanding the Fed's balance sheet. That doesn't mean that it can't alter the quantity of bank reserves: it can and does alter that quantity when it allows its TGA balance to change. But once it exhausts that balance, its power to add to bank reserves comes to an end.

Present procedures also assign to the FOMC the exclusive authority to set the Fed's interest rate targets, while granting to the Fed Board authority to set the Fed's interest rate on excess reserves. Finally—and most importantly—Congress has delegated to the Fed responsibility for achieving "the goals of maximum employment, stable prices, and moderate long-term interest rates.”

Together these things mean that, while the Treasury is indeed a monetary institution, as is most obvious from its role in minting coins and printing Federal Reserve notes, it is far from having as much influence on monetary policy as it might under different circumstances, including one in which it considered it appropriate to strike and compel the Fed to purchase trillion-dollar platinum coins. The MMT coin gambit is, for this reason, not just about "contesting an idea." It is about changing the delegation of the government's "power to coin money and regulate the value thereof."

In particular, as Tankus observes, the platinum coin plan would be more in keeping with "the original spirit of the mint" which, when it was established in 1791, was this nation's principal monetary authority. However, in those days the dollar was a commodity money unit, based on either gold or silver. Consequently, the scarcity of those metals alone sufficed to place meaningful limits on the U.S. Mint's, and hence the Treasury's, monetary powers. There was no need for any further measures to guard against the "fiscal dominance" of monetary policy. With today's fiat money regime, that's no longer the case. Instead, some public authority has to be both  responsible for and capable of regulating the value of money. So far at least, the Fed alone bears the responsibility. It therefore needs the capability.

A Bankrupt Fed?

Would relaxing the Treasury's "self imposed" rule against minting trillion-dollar coins leave us as safe, or safer, from abuse of the money power than we are now? To answer that question, one has to consider possible future consequences of the change. That means asking what might happen if the Treasury could mint not just one or two trillion-dollar coins, but as many such coins as it liked, and not just to combat a severe recession, but for any reason. Once the government disburses the funds so raised, the Fed ends up with $1 trillion of liabilities on which it owes interest for each coin it has to accept, but without any additional interest-earning assets. Should the government's spending boost demand and prices, as it's likely to do, the Fed will have to raise its IOER rate, driving a still larger wedge between its interest earnings and its interest expense.

Nathan Tankus suggests that the Fed could prevent the Treasury from adding to its interest-earning liabilities by selling securities, that is, by "sterilizing" the Treasury's platinum coin deposits. But that would leave it with an even lower ratio of interest-earning assets to interest-earning liabilities. Allowing the Fed to sell its own securities—another Tankus proposal—would have the same result.

How many platinum coins would it take to raise the Fed's expenses above its interest earnings? Actually, a couple coins might suffice, for then the Fed's non-interest earning assets would exceed its non-interest earning liabilities, consisting of $1.8 trillion or so in circulating Federal Reserve notes, by about $200 billion. Assuming that the Fed's security holdings yield more than its IOER rate, it might still cover its expenses. But it would be a close-run affair. A third coin would almost certainly put it in the black. Probably a quarter-trillion coin would suffice.

To make the point more concretely, let's think back to 2019 for a moment. President Trump is convinced that more QE will make the economy "go up like a rocket," but the Fed won't go along. A Treasury staffer recalls the platinum coin loophole, and after some logrolling Congress comes up with an omnibus bill authorizing $2 trillion in spending on border walls, infrastructure repair,  zero-emission energy projects, and a job-guarantee program, to be paid for by having the Treasury jam a couple platinum discs down the Federal Reserve System's throat.

Looking at the Fed's income and expense statement for that year, without being saddled with the coins, the Fed earned about $100 billion in interest on about $4 trillion in securities, while paying about $40 billion in interest on $1.5 trillion in reserves. That left it with $60 billion to remit to the Treasury. (I set aside the Fed's non-interest income and operating expenses, as these are relatively minor items.) By resorting to the coins, the Treasury would have saddled the Fed with another $53 billion in interest expense, leaving it with a surplus to remit to the Treasury of a scant $7 billion! Like I said, that would be cutting things awfully close.

And what's to keep the government from coming up with still more ambitious projects, requiring many more coins? Perhaps it will take-up failed 1992 Presidential candidate Bo Gritz's plan to use its coining powers to pay off the national debt, as it might do for the price of 23 ounces of platinum, and whatever it takes to run some coining equipment for a few more seconds.

The problem isn't simply that the Fed's earnings might temporarily fall short of its expenses. The Fed's accounting practices allow it to handle temporary losses. As Seth Carpenter and several coauthors explain, when that happens, the Fed creates a "deferred asset" recording the loss. It then stops remitting funds to the Treasury until its earnings become positive again, allowing it to make up for its deficits.

But platinum coin financing doesn't just risk exposing the Fed to temporary losses: it could risk having it operate in the red for so long that its deferred asset ends up being worth more than the present value of its anticipated future earnings. That is, it could render it "policy insolvent." As Carpenter et al. note,

Because there has never been a deferred asset of any meaningful size, there is little guidance as to the whether or not there is a limit to the potential size of the asset. It may be plausible to assume that it would not be allowed to exceed the value of all future earnings, possibly in present discounted terms, given the fact that it is paid down through future earnings.

Plausible indeed, since it's not usually possible for a policy-insolvent central bank to go on printing money to cover its losses while also meeting its inflation target. It would then need to be recapitalized by Congress in order to stick to its mandate.  In short, the Fed might be forced either to sacrifice the budgetary autonomy that's intended to bolster its independence by allowing it to avoid periodic bargaining with the government, or to toss aside its mandate. Responsible monetary control could suffer either way.

How likely is such an outcome? I don't know. I only know that it would be very unwise for Congress to cast-aside current limits on the Treasury's money-creating powers until a serious discussion has taken place concerning that step's possible consequences.

The Direct Draw Alternative

None of what I've said should be taken to suggest that there are no constraints on the Treasury that couldn't safely be relaxed at once. The rules preventing the Fed from purchasing securities directly from the Treasury, or from otherwise lending directly to it, come to mind. The ban on direct Fed purchases of Treasurys is of purely cosmetic significance. Despite appearances it doesn't stop the Fed from doing anything it can't do by purchasing securities on the secondary market. Restoring the Treasury's "Direct Draw Authority," which until 1981 allowed it to repo securities with the Fed to meet its short-term cash needs, also wouldn't do any harm, particularly if the authority were limited to emergencies. Before we favor the Treasury with a platinum token, let's try these less gaudy options. After all, it's the thought that counts.