In my June 10 post on the penny, I wrote:
The U.S. government makes a pretty penny (pun intended) on seigniorage. It’s not as much as it used to be because more and more people use credit cards and even cryptocurrency to buy goods and services. Still, it’s a good amount.
The biggest gain from seigniorage is on the $100 bill. Printing one costs the federal government just 9.4 cents. So, when the feds spend this $100, they make a nice profit of $99.90. Not bad. Printing a $1 bill costs the feds 3.2 cents. So even on a $1 bill, the feds make 97 cents.
In the comments, Rob Rawlings wrote:
I’m a bit confused by the idea of the government earning seigniorage by printing new notes. Happy to be corrected if I’m wrong but don’t they earn seigniorage when they buy back their own bonds with newly created electronic money rather than when they print new paper notes? When they print these new notes (to match an increased demand to hold them rather than electronic money) then the costs of printing seems like it would be a real cost.
I agreed that the cost of printing would be a real cost but that that cost was small relative to the face value of a $100 bill and even of a $1 bill.
Rob responded:
If the newly printed note is provided to a bank in exchange for an equivalent amount of base money, then where is the “net seigniorage”? It seems the seigniorage occurred previously when the government created new base money by buying back bonds.
Somehow, in responding, I missed Rob’s second sentence above. I think that’s true. The bottom line is that there is seigniorage and that he identified where it happens.
I think I erred in even going in the direction of talking about “net seigniorage,” as you’ll see when I quote Jeff Hummel below.
I brought in my monetary theory and policy guru, Jeff Hummel, who sent me the following paragraph:
I think we just have a definitional difference here. If you want to look at net seigniorage as you define it, that is fine and sometimes informative. But what I think is the standard way to think about seigniorage is as a tax (a tax on real cash balances), analogous to explicit taxes and government borrowing, the other two main ways governments generate revenue. Both of those have their associated costs of collection that you could at least in theory net out. But no matter how a government creates a new dollar and puts it into circulation, whether with a coin, bill, or electronically as with non-interest-bearing bank reserves, the burden imposed on the government’s subjects (through an eventual reduction of real cash balances) is still ultimately a dollar.
I agree with Jeff. Well, almost. I’m going to be a little picky here and point out that the burden of a tax is never (except in the case of a lump-sum tax) the amount of revenue collected. It’s that amount plus the deadweight loss, in this case, from people economizing on their holdings of real cash balances.
It occurs to me now in retrospect that some readers might think I’m advocating that the federal government print more $100 bills. I’m not.
Instead, I’m making a more modest point, and it’s this. Let’s say that the Federal Reserve has chosen an optimal monetary policy, defined however. Scott Sumner will have one definition, John Taylor another, and so on. But let’s hypothesize that in choosing this optimal monetary policy, the Fed assumed that there would be no additional demand in other countries for U.S. currency. In other words, it assumed that whatever U.S. currency was currently being held abroad, there would be no additional demand.
But, it turns out, there is additional demand. Then the optimal monetary policy would not be the one the Fed chose. The optimal policy would be to print more $100 bills.
Note 1: Thanks to Rob Rawlings for raising good points and to Jeff Hummel for helping me think through it.
Note 2: I gave directions to ChatGPT to draw a picture of a $100 bill with, due to whimsy on my part, the size of Ben Franklin’s head exaggerated. At whatever size, I think Franklin’s expression makes him look a little like Jack Benny.
READER COMMENTS
Rob Rawlings
Jun 14 2025 at 9:18pm
Thanks again, David — I really appreciate your follow-up and Jeff Hummel’s clarification. I agree with both Jeff’s framing and your additional comments on it.
I still find it useful to distinguish between situations of money creation and money transformation — though I see that as more about describing the Fed’s operational and accounting mechanics than about the underlying economic theory.
One minor point where I might differ is with the statement: “Then the optimal monetary policy would not be the one the Fed chose. The optimal policy would be to print more $100 bills.”
If, say, the Fed were targeting 5% NGDP growth, and unexpected foreign demand for $100 bills pulled money out of circulation and caused NGDP to fall below target, the optimal policy wouldn’t itself change — it would still be to hit the 5% NGDP path. What would change is the quantity of money needed to achieve it, including a greater share going into currency. So I’d see the increase quantity of base money and the printing of more $100 bills as a neutral accommodation to uphold an unchanged policy goal, not as a change in policy.
That said, I’ve really enjoyed the exchange — it’s deepened my understanding of both the mechanics and the macro framing. Thanks again for engaging so generously!
David Henderson
Jun 15 2025 at 10:28pm
Thanks, Rob.
With your characterization of optimal policy, I agree with your point.
I’ve enjoyed the exchange too, and I’ve learned.
Scott Sumner
Jun 14 2025 at 11:23pm
I think this is basically right, but a small point to clarify the exchange of cash for bank reserves. If the reserves pay interest (as they do today), then the government earns a profit by avoiding having to pay interest on the reserves pulled out of circulation in exchange for currency. The present value of this saving should equal the value of the currency note put in circulation.
If the bank reserves don’t pay interest, then any injection of currency will usually cause the monetary base to rise by the same amount. (Assuming a stable inflation or NGDP target.)
David Henderson
Jun 15 2025 at 10:29pm
Thanks, Scott. Good point.
Rob Rawlings
Jun 15 2025 at 11:24pm
If I’m understanding Scott’s point correctly, the demand from banks to hold reserves is independent of the public’s demand to hold currency. So if the public increases its demand for cash while banks’ reserve preferences remain unchanged, the central bank will need to increase the monetary base to meet both demands. The net result — other things equal — is that when new notes are printed (albeit indirectly, and taking Jeff’s clarification into account), seigniorage follows as the central bank creates additional base money to meet total (public + bank) demand.
Jeff Hummel
Jun 16 2025 at 5:33pm
Here are some clarifications to answer your questions (I hope):
1. U.S. currency is printed by the Bureau of Printing and Engraving (within the Treasury Department) and then bought by the Fed at its production cost. The Fed decides how much currency to purchase. But it never issues it directly to the public but only to banks and primary dealers by reducing their reserves. And the Fed does not actually control how much of the reserves it creates becomes currency. It is the banks who request that their reserves be exchanged for currency in the form of vault cash. And it is the public that ultimately determines how much currency is in circulation by reducing their bank deposits in exchange for currency.
2. Prior to the Fed paying interest on reserves in October 2008, the total monetary base, consisting of currency (either in circulation or in the form of vault cash) and bank reserves, generated seigniorage for the government. But when the Fed started paying interest, it essentially converted bank reserves into another form of government debt. Bear in mind that the Fed purchases Treasury securities (and other assets) on which it earns interest. Now part of that interest was being passed along to the banks. Pure fiat money, in contrast, pays no interest and, therefore, allows the government to purchase real resources without incurring any future liability. As a result, now only currency in circulation and in bank vaults provides the government with seigniorage.
3. As of February 2025, out of the Fed’s total monetary base of $5.75 trillion, more than half, $3.43 trillion, was in the form of reserves. And except for a small amount of vault cash, $95.4 billion, those reserves were earning 4.4 percent interest.
Rob Rawlings
Jun 16 2025 at 6:54pm
Thanks to Jeff and Scott (and ChatGPT!) for addressing some (hopefully) final points of confusion I had. I know feel able to address my own initial comment quoted by David above — “I’m a bit confused by the idea of the government earning seigniorage by printing new notes” — my initial skepticism, while true at the surface level, turns out to be wrong once you dig into the mechanics more deeply.
In the pre-IOR world, when reserves didn’t earn interest, the entire monetary base — both reserves and currency — generated seigniorage If the public demanded more currency, banks would convert reserves into cash, but since they held minimal reserves, the Fed (to hit it targets) almost always had to expand the base to restore reserve levels. This expansion funded interest-earning assets, so adding to seigniorage. So, even though the notes were issued to meet demand and were a transformation of one form of base to another , the act still led — albeit indirectly — to an increase in seigniorage.
In the post-IOR world, banks hold large quantities of interest-bearing reserves. If the public demands more cash, banks can exchange reserves for currency without forcing a base expansion. But here too, seigniorage occurs — because the Fed now avoids paying interest on the reserves it converts into zero-interest currency. This savings on interest expense is seigniorage.
So in both regimes, printing new notes will lead to an increase in seigniorage — either through expanding the base or by shifting its composition. It’s a more subtle process than I initially appreciated, and I’ve learned a lot — and see now that my earlier assumptions didn’t fully hold up under closer scrutiny.
As I said then “Happy to be proved wrong!”