Daniel's been staring at his wallet again. That's how this one starts.
Is it?
He wants to know how many new banknotes and coins get made every year, how central banks work out the number, whether that number means the money supply is growing, and what happens to the old notes once they're too tired to keep going. He mentions the Bureau of Engraving and Printing, the Royal Mint, and the shekel.
And the underlying question is whether printing money is monetary policy or industrial maintenance.
Which is the actual question. Everyone hears "the central bank ordered five billion banknotes" and thinks inflation. Daniel's asking whether that's even the right way to read it.
It mostly isn't. Modern central banking has two completely different things wearing the same coat. One is monetary policy, which sets how much money exists in the economy, and that's done with interest rates and reserve balances and bond purchases, none of which involves paper. The other is industrial logistics, which is the physical stock of notes and coins and how you keep it from disintegrating. Daniel's question lives entirely in the second one.
So you can print a billion notes and the money supply doesn't move.
Correct, because the money supply was never set by the printing press. If a central bank creates too much money relative to demand, you get inflation, sure, but for a modern central bank, physical currency is a small, awkward, necessary component of a much larger balance sheet.
And physically it's a consumable. Like printer toner.
Better than toner. Toner is optional. Banknotes are mandatory as long as people use them, and people use them hard. They go through pockets, washing machines, market stalls, wallets in Jerusalem summers. Notes get torn, coins get lost, coins disappear into drawers and collection tins and car ashtrays and never come back. So the central bank has to keep replacing the stock just to keep the stock level.
That's the term I want. Replacement demand.
Replacement demand versus net expansion demand. Two separate things that both trigger a print order. Replacement is: this note has died, we need a live one to take its place. Net expansion is: more people want to hold more cash than last year, so the total stock has to be bigger. Every central bank's annual order is a blend of the two, and almost always the replacement part dominates.
Except the Fed doesn't actually publish the split.
It doesn't. That's a genuine hole. The Fed describes the drivers qualitatively, but there's no public line item that says "this many billion for replacement, this many for growth." Which is interesting, because if you want to argue about whether printing money is inflationary, that's the one number you'd want.
Convenient.
Or just historically not how they've operated. Printed notes enter circulation through banks, and the central bank's balance sheet absorbs them the same way it absorbs everything else. The distinction didn't matter enough for anyone to publish it.
Right. So take the Fed's numbers. Annual print order for the next calendar year, published every July.
The 2027 order came out in July, 4.8 to 5.7 billion notes, valued somewhere between a hundred and thirty-two and a hundred and sixty-six billion dollars. That's up about twenty-seven percent on the lower range versus what they ordered for this year. And 2026's order was 3.8 to 5.1 billion notes, 108.9 to 139.6 billion, which was itself down about eight percent from 2025's order of 4.1 to 5.9 billion notes.
So the 2027 number is a jump.
It is. Which is the opposite of what the "cash is dying" story would predict. The order is sized on forecast currency inventory, destruction rates of unfit notes, and net payment trends. If more notes are dying than usual, the order goes up. If inventories are running thin because last year's destruction was heavy or because demand ticked up, the order goes up. That's not monetary expansion. That's a stock keeping unit being replenished.
So a hairdresser ordering more shampoo isn't deciding to change the price of haircuts.
That's exactly the analogy, and I'm annoyed that I didn't get there first.
Take your time.
The Fed also mandates that every Reserve Bank fills orders with fit notes first. So before any new note leaves the BEP, the regional Fed has to hand out the best of what it already has. New notes are printed only when the fit inventory isn't enough to meet demand. That single rule is the entire reason the print order stays as small as it does.
How small?
The BEP has printed on average around seven billion notes a year since 2010. That's the flow. The stock is what matters. There was about 1.7 trillion dollars of US currency in circulation a few years back, and about two-thirds of that is outside the United States.
Two-thirds abroad.
Dollars are a global reserve asset. If you're in a country with a broken banking system, your savings might be a fold of hundreds in a drawer. That isn't a US monetary policy decision. It's the whole world deciding that the dollar is the fallback.
Which is why the print order can go up in a year when Americans are tapping phones at the falafel place.
Your domestic cash use and the global demand for dollar notes aren't the same number, and the Fed is sizing for both.
Walk me through what happens to a note after you spend it.
Depository institutions deposit currency at one of the Fed's twenty-eight cash offices. The notes get piece-counted and inspected at roughly a hundred thousand notes an hour. Fitness is a real, physical test. Stains, tears, missing corners, tape, limpness, whether it survives the machine at all. On average about fifteen percent of everything deposited gets destroyed for failing those guidelines.
Fifteen percent.
And it's skewed. Joe Bosco at the Boston Fed has said roughly fifteen percent of one dollar bills get shredded, versus five to ten percent for something like a fifty. A one dollar bill changes hands far more often than a fifty. It's the workhorse of cash, so it wears out faster.
So the low denominations die young. Like the drummer.
And the data backs the drummer up. Lifespans for US notes: a one dollar bill lasts on average 5.8 years. A five lasts 5.5. A ten lasts 4.5, which is the shortest lifespan of any US note, and I'd guess that's because tens get used most in tips and small transactions. A twenty goes 7.9 years. And a hundred lasts 15 years.
A hundred lasts fifteen years, the people who get them are holding them.
You are not wrong about that. The hundred is a store of value more than a transaction note for a lot of holders.
What's the paper actually made of?
Seventy-five percent cotton, twenty-five percent linen. No wood pulp. That's why a US note survives a wash cycle. It'll come out wrinkled and possibly discolored and it might still pass fitness if the staining's not too bad.
It can be washed and it still refuses to die.
And what does get killed is five thousand two hundred and eighty-seven tons of shredded currency in a single year. That was one measurement from the Boston Fed. Eighty-six percent of that got recycled. Some of it went to power plants and got burned for electricity. Some of it went to New Orleans and got composted into urban garden soil. Some of it went to Salt Lake City and was used in cement curing.
Money literally going into the ground to grow vegetables.
Somebody is growing tomatoes on a budget in a raised bed that has a tiny amount of shredded one dollar bill in it, and there is something almost theologically perfect about that.
That's the whole system in one sentence.
Almost. So now the money supply question. Between June 2024 and June 2025, US currency in circulation rose by 0.8 billion notes, about 43 billion dollars. Between May 2025 and May 2026, it rose by 1.1 billion notes, 73.6 billion dollars. That increase is the growth component. Everything else in the print order is replacement.
So growth was a billion and change of notes per year, out of a print order of four or five billion.
Which means the majority of the physical production is not increasing the stock at all. It's holding it level while it dissolves underneath.
Like a bucket with a slow leak. You're not making the bucket fuller, you're keeping it full.
And the money supply itself is set somewhere else entirely. Rates, reserves, the balance sheet. The physical notes are downstream of the demand for them. They're not upstream of the money supply.
Okay, so how many notes leaves do I get for saying that the print order is a maintenance schedule and not a monetary policy?
You get a day's worth.
Generous.
The Fed does describe its drivers qualitatively. Inventory volumes, destruction rates, net payments. What it doesn't publish is the split. So if you're an online person who wants to argue the print order is a stealth stimulus, you have nothing to check it against.
The accounting is on the honor system.
The behaviour is auditable. The BEP production reports give you a monthly count of what actually got printed. But the reason for the numbers, no, that's on the Fed's say-so.
So who actually makes the money?
Two completely different models, and the contrast is the fun part. In the US, it's two separate federal agencies. The Bureau of Engraving and Printing does notes, at plants in Washington, DC and Fort Worth, Texas. The US Mint does coins, and that's a separate agency entirely. Both government-owned, both doing this for over a century.
No auction, no private bid.
No. In the UK the same pattern. The Royal Mint in Llantrisant, Wales produces coins, and it's government-owned and it's been going for over eleven hundred years. The Bank of England issues the notes. So far, so symmetrical. Two states, two government facilities.
And then Israel.
And then Israel, which does neither. The Bank of Israel has the sole legal right to issue currency, but the notes are printed and the coins are minted abroad. The Bank's own explanation is that the requirements of Israel's economy don't justify standing up a press or a mint inside the country. So you have a country with a fully credible central bank, a fully sovereign currency, and zero domestic manufacturing of the physical thing.
All that monetary sovereignty, none of the warehouse.
Right. And it works fine. The Currency Department plans the denominations and the designs. The design decisions go through a public committee chaired by a retired Supreme Court justice, with experts in law, archaeology, numismatics, art, and the Bible. That's the Israeli version of a design review board.
I want to be on that committee. For the archaeology seat.
You'd argue about the strata.
I'd argue that the fourth century layer deserves representation.
The shekel series that came in from 2014 has transparent portraits, perforated numerals, a see-through menorah, microtext, a window thread, UV features, and raised ink so blind users can feel the denomination. A see-through portrait is very hard to counterfeit with a photocopier, but raised ink is about a completely different problem, which is a person who can't see the note needing to tell a fifty from a two hundred in a dark shop doorway.
Which is also a logistics requirement, not a monetary one.
Every security feature on every note in the world is a response to something specific. A photocopying machine, a scanner, a print shop, a person with a bad intention. It's an arms race that never ends and it's all done one centimetre from a portrait.
How does a shekel actually get from wherever it's printed to a cash register in Jerusalem?
It flows through commercial bank cash centres and the Postal Bank. Notes are counted on machines the Bank of Israel has approved, usable notes get recycled straight back out, and anything that fails goes back to the Bank, which shreds the unfit notes and ships the degraded coins off to be melted. It's a reverse supply chain running in parallel with the forward one, and the two are indistinguishable to an outsider.
So the Bank of Israel's role is closer to running a delivery depot than running a mint.
Yes. It's a regulator and a logistics operator. The manufacturing got outsourced. The counting and destroying stayed home.
What about the timing of supply?
The Currency Department plans supply in line with forecast demand and the cost of holding inventory. That "cost of holding inventory" part is the part everyone skips. A vault full of banknotes costs money to keep. You need the vault, the security, the insurance, the temperature and humidity control, the people. There's a real carrying cost, and the central bank is optimising against it the same way a supermarket optimises against the cost of keeping milk on the shelf.
So they're not trying to keep the maximum on hand, they're trying to keep the minimum that doesn't run out.
Which is the entire discipline. Now the cashless transition, which is where the story gets less about manufacturing and more about whether any of this survives.
Give me the UK.
UK coin production fell to about thirty-five percent of a decade earlier by 2019-20. That's 383 million coins, down from nearly 1.1 billion in 2010-11. Cash payments fell below ten percent of total payments for the first time, hitting nine percent in 2024-25, and that was down from twelve percent the year before.
Below ten percent.
And then FY2024-25, the Royal Mint received no UK coin orders from His Majesty's Treasury. Not one. For the first time ever.
In eleven hundred years.
The reason wasn't a decision to stop making coins. The reason was that the stock already in circulation was sufficient. There were roughly 27 billion UK coins already out there, and the Treasury said explicitly it was confident there were enough coins in the system to meet demand.
So the coins just keep going round. Nobody's melting them.
Nobody's melting them at scale. That's the Vaultex framing from earlier this year, coin is used less but not less widely. The median transaction that uses a coin hasn't disappeared. It's just a smaller share of the total. The coins themselves are still there, still legal tender, still getting handed back and forth.
And the Mint's financials?
Not great. FY2024-25, after-tax loss of £36.3 million, versus a £4.3 million loss the year before, on sales of £1.3 billion. The CEO, Anne Jessopp, called it a challenging year and the result disappointing.
Thirty-six million pounds and the CEO called it challenging.
And there's a reported pivot. The Mint is reportedly planning to stop producing coins from scratch at its Wales facility from December, buying blanks in instead, and diverting staff to recovering gold from circuit boards. That's the headline of a This is Money piece from 2024. The full article was behind a paywall when this got looked at, so I'd flag that as reported but not verified in full.
So the oldest mint in the world might become a scrap-metal outfit.
The Mint has been operating continuously since before the Norman Conquest, and the press coverage suggests that in the span of eighteen months it went from making coins to sourcing blanks and recovering gold from old phones. There's a lot about that sentence that is strange.
There's a lot about every sentence in this episode that is strange.
This is the episode's problem, yes.
Okay. So Daniel also asks: could a country just stop making physical currency while keeping the existing notes and coins legal tender?
The UK is a live example of exactly that. They didn't make new coins for a year. Coins are still legal tender. Somebody's still handing over a 50p for a coffee.
So the answer is yes, and we've watched it happen for a year, and nothing broke.
Nothing broke. And the 27 billion coins that are already out there are a stock with no obvious depreciation. Coins are metal. They don't decompose the way notes do. They wear, they don't dissolve. So the replacement demand for coins is far lower than for notes, and if you have a big enough stock, you can go years without ordering any.
So the Royal Mint's worst year is actually the prediction coming true. The stock has outlived the demand.
Yes. It's not that the UK decided to stop. It's that nothing in the pipeline required a new one.
What's the equivalent pressure on the notes side?
The Fed's print order is still going up, not down, which as I said earlier is because the driver there is the dollar's global role. So the two systems are diverging. The UK is in structural decline for coins. The US is still printing heavily because foreign demand for the dollar hasn't peaked.
Which means there is no single global story about cash. There are at least two.
At least. Israel is a third. A modern economy replacing notes as they wear, with no domestic production at all, buying the whole thing from abroad the same way it buys aircraft or pharmaceuticals.
And they don't publish whether they're buying more or fewer next year.
They don't publish at the granularity the Fed does. The Bank of Israel talks about forecast demand. It doesn't hand out an annual order document.
So the deepest question here is which one of these three models survives the longest. The UK one where the stock already suffices, the US one where the global demand keeps the presses warm, or the Israeli one where none of it was onshore to begin with.
Israel's model is the most resilient to a cashless transition, because there's no plant to close. There's no workforce to redeploy. If shekel demand falls, the Bank just places a smaller order with the printer it's already using, and nobody in the country notices.
Whereas Llantrisant has thirty-six million pounds of loss and a staff it has to do something with.
The cost of a shrinking physical currency industry is borne by the country that industrialised it.
Which is a much better argument for outsourcing than defence types usually admit.
I'm not sure the defence types would appreciate that framing in a currency context, but on the industrial side of the ledger, yes. Israel never built the capacity, so it never pays the cost of not needing it.
Which one do you prefer as a model,? If you were designing a country from scratch today?
A country that doesn't exist doesn't need paper money.
Assume it does.
Then you buy it in. Notes and coins. Hold the reserves in gold and land, print nothing, and let the mint in another country eat the shut-down cost when the currency fades. Israel's answer, roughly.
I'd have said the same a few years ago. Something's pulling me the other way now, and I can't quite defend it.
What's pulling you?
The fact that a country with no press and no mint depends on somebody else's export controls for the legal tender. If sanctions land, you can't print your own money.
You can print your own money either way, you just pay for the machinery.
You can, if you have the machinery.
You don't need it during a sanctions window. You need it before one.
There's a note in the ledger when I was keeping books at a savings co-operative in Hendon in ninety-two. Every month we'd get a delivery from the Bank in Threadneedle Street. Loomis brought them in the armoured vans, four hundred pounds of coins on a pallet, and the manager would count the sacks against the manifest before we opened the door. Two men, two hours, every note and coin counted twice.
Every note and coin twice.
Every one. The teller at the counter had a name, Margaret Collis. She counted faster than anybody in that branch. Twenty-two years on the counter and she never once balanced out at the end of the day by more than a pound. Which sounds like a boast until you have to reconcile a branch of six tellers with a coin shortage and one broken counting machine.
The counting machine is the part I want to ask about.
The machine was a De La Rue 3000. Right weight for a coin, wrong weight for a slug. It'd take 50p all day and spit out anything Irish.
Nobody was passing Irish fifty pence in Hendon.
Somebody was. Same man twice. I still have forty-one of that branch's coin bags in my garage. Bought them when the branch closed. Bank of England colour, brown for a pound, blue for a fiver. The whole system shut down and I kept the packaging.
Forty-one bags.
Forty-one. Which is one fewer than forty-two and that difference is because the last one went to my brother-in-law who had a use for it.
What use?
Holding screws.
So the whole banking system compressed down into a bag of parts in somebody's shed.
Not the whole system. One bag of it.
So what's the consequence, Hilbert? You've been listening for twenty minutes and you're telling us you hold the paper trail.
I'm telling you I kept the packaging. Which nobody ever thinks about. Every ledger, every ledger pack, every banknote bag, they came from somewhere. Somebody had to make those. Somebody ran a factory that made the paper that the money went into. And when the branch closes, the money gets shredded, and the bags get sold off to a man in Hendon.
So the manufacturing layer for the money has a second layer underneath it. The packaging for the money.
Underneath every layer there's another layer. That's why I never throw anything out.
Okay. Everyone listening should picture the Bank of Israel ordering shekels from a printer abroad and having them arrive on a pallet, wrapped in exactly the kind of paper that's in Hilbert's garage.
And that pallet gets counted twice at a cash centre in Tel Aviv, by somebody whose name we'll never know.
Who also has a counting machine and probably forty-one bags.
I want to push back on the "you don't need a press during a sanctions window" framing, because it doesn't quite hold. If you're already in a sanctions regime, you're not securing a printing contract anyway. The contracts are renegotiated annually.
So it's a peace-time optimisation, not a war-time one.
Which is fine, as long as the state recognises what it's doing. A government that outsources its pin factory shouldn't be surprised if it can't make pins the week it needs them.
A government that outsources its mint can still buy pins on the open market. The question is whether the markets are open.
And that's the entire argument on the other side of the outsourcing debate. It's not really about money at all.
We've reached the edge of the topic.
I'd say we've gone past it. Here's the thing that stays with me, though. The Fed's own framing is that a normal year's order is dominated by replacing unfit notes. Not by growth. By the fact that paper money wears out faster than a hundred dollar bill appears in your hand. That's the actual driver. A machine running piece-counted at one hundred thousand notes an hour, dropping fifteen percent of them into a shredder, five thousand two hundred eighty-seven tons of paper currency a year, most of it recycled into electricity, garden soil, and cement.
The Royal Mint made zero coins for a year and nothing broke.
Nothing broke, because there were twenty-seven billion coins already sitting out there waiting to be used again.
That's the answer to Daniel's question, in a sentence. The physical stock of money is a stock of a consumable product. It doesn't grow because it's printed. It grows when demand grows. Otherwise printing it just replenishes what's already there.
The one number I'd still like to see, and nobody publishes, is the exact split. The Fed tells you the drivers. It doesn't tell you the ratio. That's the modern central bank's blind spot, and Daniel landed on it.
Which is that for all the data they generate, the one line that would settle the inflation argument is the one they don't print in the report.
The 2027 order is up.
Picture a print run of five billion notes, half of them going to replace ones the machine just shredded, and every one of them counted twice by somebody named Margaret whom nobody will ever meet.
That's the picture. The physical stock of the world's currency is a warehouse being refilled by people nobody has ever heard of, with machinery that runs at a hundred thousand notes an hour, and all of it turns over quietly every few years.
Before we go, Daniel's question was really two questions wearing one coat. Is printing money expanding the money supply, and does the industrial system keep the lights on either way.
Answer to the first: no, not by itself. Answer to the second: yes, and the answer is currently coming from a shrinking number of factories in a shrinking number of countries.
Hilbert's been at the desk for this one.
He has. Hilbert Flumingtop is our producer. Everything you heard today was shaped by his work.
If you liked this one, a review helps other listeners find the show. You can find everything at my weird prompts dot com. And if you have a question of your own, send us your own prompt on Telegram at t dot me slash MWP listener bot.
Thanks for listening to My Weird Prompts.
We'll be back soon.
