The message Ueno's customers saw on the third of November was not a glitch.
We cannot display the balance.
That sentence is the most honest thing a bank has ever said to its customers, because it describes the relationship accurately. What you call your balance is a number the bank keeps for you. When the bank stops keeping it, the balance does not become zero. It becomes unverifiable, and for every practical purpose those are the same thing.
The rest of this chapter is about how that number came to exist, what stands behind it, and why the answer is stranger than most people assume.
A deposit is a loan you made and forgot about When you put money in a bank, you are not storing it. You are lending it.
The legal description is unromantic: the bank becomes your debtor, and you become its creditor. You do not own any particular banknote sitting in any particular drawer. You own a claim on the bank, payable on demand, subject to the bank's ability to pay.
This is not a technicality invented by critics. It is the structure, and it explains a great deal that otherwise looks inexplicable — why deposits are insured at all, why a bank failure is a legal process rather than a technical fault, and why the word "withdrawal" exists. You are withdrawing a claim, not retrieving a thing.
Compare it with the arrangement it replaced. If you put a gold coin in a vault, the coin remains yours. The vault company is providing a service — a building, a door, a guard — and if the company fails, the coin is still in the building. The failure is an inconvenience. With a deposit, there is no coin. There is a promise, and a promise is only as good as the party making it.
Every bank balance in the world sits on that foundation, including yours.
How the promise multiplies The mechanism was described plainly by the Federal Reserve Bank of Chicago in a booklet called Modern Money Mechanics, published in 1961 and used for decades to train bank examiners.
"The actual process of money creation takes place primarily in banks."
And, on the part that surprises people:
"Of course, they do not really pay out loans from the money they receive as deposits. If they did this, no additional money would be created. What they do when they make loans is to accept promissory notes in exchange for credits to the borrowers' transaction accounts."
Read that second sentence again. The bank does not lend out your deposit. It creates a new deposit when it makes a loan, and the loan is the asset that balances it on the other side of the ledger. The booklet walks through the arithmetic with an example: deposit a thousand dollars, the bank keeps a hundred and lends nine hundred; that nine hundred is spent and deposited at the next bank, which keeps ninety and lends eight hundred and ten; and so on until the sequence runs out. A thousand dollars of original deposit becomes ten thousand dollars of deposits across the system.
Your thousand is still on the statement. Nine thousand more now exists beside it.
The booklet assumed a ten percent reserve requirement, which is where that example comes from. That assumption is history. On 26 March 2020 the Federal Reserve set the reserve requirement ratio to zero, and it has stayed there. The constraint that limited the sequence in 1961 no longer exists in the form the booklet described. Lending is still limited by capital rules, by the demand for loans, and by prudence — but not by a requirement to hold a fraction of deposits in reserve.
You can see the result in the aggregates. In 1971, the year the dollar's link to gold was severed, the United States money supply was about 686 billion dollars. It is now above 23 trillion. That is roughly thirty-four times as many dollars, against an economy that did not grow thirty-four times.
The consequence is not a secret and it is not a conspiracy. The Federal Reserve targets two percent inflation per year, publicly. At two percent, prices double in about thirty-five years, which means a dollar held without earning interest loses half its purchasing power within the working life of the person holding it. Measured from 1971, the dollar has lost about eighty-eight percent of what it could buy.
So: the supply of the thing in your account is a policy variable. It is set by people, on a schedule, with a stated target. That is the first half of the problem, and it is the half everyone argues about.
The second half is quieter, and it is the subject of the rest of this chapter.
What stands behind the claim When a bank cannot pay, the question becomes what happens next, and the answer is a process rather than an event. Deposits are insured up to a limit, the institution is resolved, and customers are paid out according to a schedule they do not control.
In the United States the insured limit is 250,000 dollars per depositor, per bank, per ownership category. That number is reassuring until you look at what sits beside it. Forty-three percent of United States bank deposits are above the insured limit, which means that for nearly half of the money in the banking system, the guarantee is partial and the remainder depends on the bank remaining solvent. The fund that pays the insurance held about 1 dollar 48 cents for every 100 dollars it insured as of June 2026.
Those are the published figures, not an interpretation.
Now the timeline. In Korea, the statutory period to resolve a failed bank is 150 days. Regulators there have proposed shortening it to three or four days, which tells you two things at once: that the current period is long enough to hurt, and that the people who run the system know it.
Ueno is the case in front of us, and the numbers are worth reading slowly.
Three million customers. About 8.9 percent of all deposits in the Paraguayan banking system. A guarantee capped at 228.3 million guaraníes per depositor, per bank — roughly thirty-eight thousand dollars. A first payout announced at about 10 million guaraníes, roughly sixteen hundred and seventy-five dollars, "as soon as possible." A resolution that began with a unanimous board vote on the second of November, was announced overnight, and took the app, the cards and the transfers with it. Eight executives brought before prosecutors. The president of the bank resigned on the fifth.
In March, that institution had reported a total capital ratio of fifteen percent, against a required minimum of twelve. By the time the central bank adjusted the books, solvency was below half of the required minimum. Whatever the ratio measured in March, it was not a forecast of November.
And here is the detail that matters more than any of the others: the money inside those accounts included pension funds. Public money, held by people who had done nothing wrong, in an institution that had reported healthy capital eight months earlier, in a country whose central bank acted exactly as designed.
The system worked. That is the point. It worked, and three million people still could not see their balance.
Access is a permission, not a property There is a second layer, and it operates every day rather than in a crisis.
Your ability to move money is granted by an institution and conditioned on its rules. Accounts are frozen, restricted and closed routinely — for sanctions matches, for anti-money- laundering flags, for patterns a model did not like, for documents that did not arrive. Most of the time the customer is guilty of nothing except being inconvenient to a compliance system optimized for its own risk rather than theirs.
This is not a scandal. It is the normal operation of a system that must, by law, know who you are and what your money is doing. The complaint is not that banks break their own rules. The complaint is that the rules make your access conditional, and the condition is not yours to satisfy once a decision has been made somewhere you cannot see.
Under stress, the permissions tighten. That is when people need their money most, and it is precisely when the system is least able to provide it — as Ueno's customers discovered at the moment the app stopped showing them a number.
One clear illustration from the same three weeks, outside banking entirely: on the 29th of October, two stablecoin issuers froze more than eight billion dollars of tokens, about 2.4 billion of it within hours, because addresses had been linked to an incident. The tokens were on a public ledger. The freeze did not require a court, a hearing, or the holder's consent. It required an administrative decision by the issuer, and it took an afternoon.
A public address is a fixed target. So is a bank account. So, in the end, is anything that requires someone else's cooperation to move.
What "yours" would actually require Strip away the institutions and ask what property in money would have to look like for the word yours to be true.
It would have to be bearer: held by whoever holds it, with no registry naming an owner.
It would have to be final: settled without a counterparty who can reverse it, freeze it, or fail.
It would have to be permissionless: movable without an institution deciding whether the movement is allowed.
And it would have to be private: not merely unrecorded in a database you cannot see, but unreadable by anyone who is not party to the transaction.
Physical cash satisfies the first three and fails the fourth in an age of serial numbers and cameras. It also fails at distance: you cannot send a banknote across an ocean in a minute, and carrying one across a border is a decision someone else gets to review.
Digital bearer money has been the design problem of the last forty years: how to make something that behaves like cash while existing as information. Bitcoin was the first serious answer, and it answered the issuance question completely — a fixed supply, no issuer, no permission to participate — while leaving two parts of the problem open, which is what the next chapter is about.
What this chapter is not claiming It is not claiming that banks are useless, or that the system is about to collapse, or that you should empty your account and bury the proceeds.
Banks do real work: they aggregate capital, they underwrite risk, they move money between people who never meet, and they do it at a scale and speed that has no historical equivalent. Most days the system works, and on the days it does not, most people are made whole.
The claim is narrower and harder to argue with. The money in your account is a claim on an institution, denominated in a supply that someone else controls, reachable only through permissions that someone else grants. That is true on the good days too. It is a description of the instrument, not a prediction about the bank.
Which brings us to the three questions this book keeps asking.
Store of value. Can more of it be created, and by whom?
Medium of exchange. Can you use it, at a cost that does not punish ordinary payments?
Sovereignty. Can anyone stop you holding it, moving it, or spending it?
The first question is a monetary policy question, and the answer for the dollar is: yes, by a committee, on a target, indefinitely.
The second is a settlement question, and the answer for the dollar is: yes, cheaply and universally, within the perimeter.
The third is a property question, and the answer for the dollar is: yes, at several layers, starting with the institution holding the number and ending with the state that charters it.
Bitcoin was built to answer the first question differently, and it did. It was not built to answer the third, and the last seventeen years have shown what happens when a transparent ledger meets institutions that would rather not see certain transactions.
That is where we go next.