Do Banks Create Money? Not From Your Savings

A bank lends a customer one thousand euros, and at that moment nobody's savings go down by a single cent. That is the worked example Germany's central bank published in April twenty seventeen, and in it nothing leaves anyone else's account. The bank types a new balance into the borrower's account. When a bank gives you a loan, it creates that money on the spot.

You have heard the opposite, that a bank takes in savers' money, keeps a slice in reserve and lends out the rest. So do banks lend out deposits? The answer three Bank of England economists gave, in a twenty fourteen article, is that the act of lending creates deposits, the reverse of the sequence typically described in textbooks. And the videos telling you banks conjure money from nothing don't match the documents either. Not from your savings, and not from nothing.

The number in your banking app is a record of what your bank owes you. So follow that one loan of a thousand euros with us, from the keystroke to the day it's paid back. Do banks create money, and if they do, where does the thousand come from? What happens when the borrower spends it at a different bank? If a bank can create money, how did Silicon Valley Bank run out? And where does the money go when the loan is repaid? One question we answer badly: does any of this change anything that matters?

First the case for the picture you were taught, because it's in print and it isn't stupid. A free university textbook from OpenStax, published in December twenty twenty-two, works through a bank holding ten million dollars in deposits. It keeps ten percent in reserve and, in the book's words, it will loan out the remaining nine million. That nine million is deposited somewhere else and lent again, round after round, until the book counts ninety million dollars of money generated. The Bank of England itself calls that story a useful way of introducing money and banking. Banks do create money in it, but only by passing your deposit along.

And it matches everything you can see from outside. You pay money in, the bank lends, and it charges the borrower more than it pays you. The Bank of England describes exactly that as a bank's business model. Banks also fight for your deposit by raising savings rates, and why would anyone pay for something they could make for free? There's a real answer to that, and part of the textbook picture survives it.

Then there's the exhibit a believer reaches for first. On the ninth of March twenty twenty-three, more than forty billion dollars of deposits left Silicon Valley Bank in a single day, by the Federal Reserve's count. Regulators closed it the next day. A company that can create money should not be able to run out of it. That objection is fair.

In the worked example from the Bundesbank, Germany's central bank, a bank checks that its customer is good for the money and grants a loan of one thousand euros. The bank simply writes one thousand into the customer's account as a deposit. Some economists, the Bank of England notes, call it fountain pen money, created at the stroke of a banker's pen. The central bank's books, the report says, remain unchanged. The bank's balance sheet, which is just the list of what it owns and what it owes, gets longer on both sides. It now owns a loan, the customer's promise to repay, and it owes that customer a deposit.

Now look for the saver whose money was lent. There are five sets of books in this example, and no account anywhere went down to make that loan. One thousand euros of spendable money exists that did not exist a minute earlier, created, in the Bundesbank's phrase, by an accounting entry. The report says a bank can grant loans without any prior inflows of customer deposits. The idea that a bank, at the time of lending, can only hand out what was deposited earlier has a name in that report: a popular misconception.

Three years earlier, three Bank of England economists had written the same thing: whenever a bank makes a loan, it simultaneously creates a matching deposit, thereby creating new money. They add that a deposit is a record of how much the bank owes you. That makes it, in their words, a liability of the bank, not an asset that could be lent out. Your savings can't be handed to a borrower, because they are the bank's debt to you.

The textbook story, where a bank keeps ten percent of your deposit in reserve and lends out the rest, needs a rule that sets the ten percent. In the United States that ratio was ten percent on the larger balances in checking-type accounts until the twenty-sixth of March twenty twenty, when the Federal Reserve cut it to zero for every bank. Its stated reason was the way it steers interest rates, not a statement about who creates money. In September twenty twenty-one, a Saint Louis Fed publication recommended that textbook authors and teachers drop the money multiplier. Its title carries three letters you normally see on a gravestone. A free university textbook published in December twenty twenty-two still has its example bank keep ten percent and lend out the remaining nine million. The ratio it is built on has been zero since March twenty twenty.

How much of the money around you was made this way? The Bank of England counted the money that people and companies in Britain held in December twenty thirteen, and ninety-seven percent of it was bank deposits, with notes and coins the other three. The same article says those deposits are created by commercial banks themselves, and the only word it gives for how many is mostly. So almost all the money in that count was some bank's promise, and most of those promises began with a bank. What happens to a promise when somebody spends it?

Nobody borrows a thousand euros to admire it. In the Bundesbank's example the borrower buys a machine, and the seller keeps an account at a different bank. The borrower's bank now has to hand that payment over, and in the example the two banks settle through their accounts at the central bank. It pays in reserves, balances that banks hold at the central bank, in a kind of account you and I can't open. After the payment the first bank still owns the loan, but it has lost the deposit and a thousand euros of reserves.

Notice that the thousand euros of money didn't vanish, it is now the second bank's debt to the seller. What moved underneath was a different kind of money altogether. A commercial bank can write its own IOU, a promise to pay, whenever a creditworthy customer signs a loan. It cannot write central bank money, and the Bank of England calls itself the monopoly provider. Payments to other banks are normally settled in that second kind of money, and every cash withdrawal comes out of it. So what happens to a bank that keeps paying it out?

One loan like that is nothing, but banks make many such loans every day. The Bank of England follows a mortgage to this exact point. A bank that financed all its new loans this way, it writes, would soon run out of reserves. So the bank has to pull funds back in, and one way is to offer you a better savings rate. That is why banks compete for your deposit. The Bundesbank says it flatly: deposits play a major role.

So was the textbook right after all? About the need for funding, yes, and that is the part we said survives. What runs in reverse, according to the Bank of England, is the order of events. The loan comes first, and the loan creates the deposit. The deposits a bank then chases are, in the Bank's words, newly created deposits, made by other banks' lending. Saving more, it adds, does not by itself give banks more to lend. That is why its article says banks do not act simply as intermediaries, and the word carrying the weight there is simply.

If banks create money, how did Silicon Valley Bank run out of it? On the eighth of March twenty twenty-three, its parent company said it had sold twenty-one billion dollars of securities, investments it had bought, at an after-tax loss of one point eight billion. The next day more than forty billion dollars of deposits left the bank, and management expected a hundred billion more the day after. For a large American bank that failed in two thousand eight, estimates put the outflow at about ten billion dollars over eight days. On the tenth of March, Silicon Valley Bank was closed. A bank can create its own IOU, and it cannot create the money it owes other banks.

That is our reading. Both central bank papers were written years before this bank failed, but each piece of the mechanism is in them. Customers wanted their deposits sent to other banks, those transfers had to be settled in reserves, and the bank's assets were securities that take years to pay back, and had lost value. We owe the believer a concession here, because this bank is not an example of money created by lending. The Fed's review says the deposits were invested primarily in securities with longer-term maturities, which is the saver's picture. It shows the limit on our claim, and it shows nothing about the creation.

For two days, anyone with more in that bank than deposit insurance covers did not know whether they would get the rest back. On Sunday the twelfth, the Treasury, the Fed and the deposit insurer said all depositors would be fully protected. They also said no losses would be borne by the taxpayer. That month the insurer estimated the cost to its own fund at about twenty billion dollars, to be recovered from banks. The rescue above the limit was a decision made that weekend, not a right.

The number in your account is an IOU. The Bank of England calls a bank deposit an IOU from a commercial bank, and says the notes in your wallet are mostly IOUs from the central bank. The state stands behind the bank's IOU only up to a line. In the United States the insurer's page, dated April twenty twenty-four, gives two hundred and fifty thousand dollars per depositor, per insured bank, per ownership category. In Britain the line has been one hundred and twenty thousand pounds since the first of December twenty twenty-five. Your balance is a bank's promise, and the guarantee has a ceiling.

If a keystroke makes money, what stops a bank typing all day? The Bank of England asks that on its own explainer page, can banks create as much money as they like, and answers no, they can't. Follow our loan and you meet the brakes in order. First, someone creditworthy has to want it, since the Bundesbank notes that loans normally start with the borrower. Second, the loan has to earn more than its funding costs. Third, regulators make the bank hold capital, its own resources, against the chance you don't pay.

And the fourth brake is the one the Bank of England calls ultimate, which is the interest rate the central bank sets. A higher rate makes the loan dearer, so fewer people want to borrow and fewer keystrokes get made. That is where the central bank reaches you, through the rate on your mortgage. Do the brakes on one bank hold when every bank moves at once? The same article says that if all banks decide to lend more at once, money growth may not be limited in quite the same way. The documents say limited. They do not say always.

Does the direction of the arrow change anything you could measure? In the multiplier picture, the stock of reserves is what sets how much money can exist. The Bundesbank reported in April twenty seventeen that euro-area bank reserves had grown more than seven-fold since the financial crisis began. Yet after the euro area's central banks started a bond-buying programme in March twenty fifteen, the broad money supply, the money households and firms hold, grew at around five percent a year. Reserves multiplied. The money, the Bundesbank wrote, did not grow in proportion.

When a bank gives you a loan, it creates that money on the spot, and when you pay the loan back, that money is destroyed. The Bank of England's own verb, in its twenty fourteen article, is destroys: the repayment of bank loans destroys money. Take a month of supermarket shopping on a credit card, which is its example. Every purchase created new deposits in the supermarket's account, and paying the bill in full destroys all of that newly created money. Germany's central bank uses the same word. The Bank of England's public explainer says the money is deleted and no longer exists. Banks create money, not wealth.

Which is how our thousand euros ends, with the loan repaid and the counter back at zero. It also answers the question people actually type, do banks create money from nothing? Neither central bank document we read says from nothing, or out of thin air. They describe an entry with two sides, a new deposit matched by a new debt of the same size, owed by a borrower the bank has vetted. Not from your savings, and not from nothing. Our words for it, not theirs: from a debt.

Could banks be turned into what most people already think they are? In nineteen thirty-three a group of Chicago economists proposed backing every deposit in full with central bank money. Under the version Germany's central bank describes, banks would simply be credit brokers between depositors and borrowers, and it adds that there is no evidence of how that would perform. Full backing turned up again in a United States law signed in July twenty twenty-five, this time for payment stablecoins, digital tokens meant to hold a steady value. The law's text says they must be backed at least one to one, and that they are not covered by deposit insurance. That backing may be a deposit at a bank, so the token sits on top of the system we just followed, not outside it.

Now the question we said we answer badly: does any of this change anything that matters? Two economists say banks are go-betweens after all. Paul Krugman wrote in twenty twelve that banks are just one channel linking lenders to borrowers. In twenty twenty-four George Selgin, of the Cato Institute, titled an essay Banks Are Intermediaries of Loanable Funds. He is not defending the textbook's multiplier, and he agrees that a bank makes a loan by crediting an account. What he denies is that this lets a bank lend at a profit without borrowing from someone else to make good on it. His target is the thin air story we also threw out, but he counts the Bank of England article we quoted as part of it. Of its line that banks are not simply intermediaries, he writes: surely this is splitting hairs. His verdict is that the intermediary picture comes much closer to how real banks work.

Against that we can put one model. It comes from a twenty fifteen Bank of England working paper, and with money-creating banks it predicts lending changes that are far larger and happen much faster than the same model with banks as go-betweens. It is labelled research in progress, not a measurement of the world. And the Bank's own article describes a path where newly created money is quickly destroyed and there need be no further effects on the economy. We can show you how the money is made, and that reserves do not set the limit. How much that changes the economy, we can't.

Your bank does not lend out your deposits, because when a bank gives you a loan, it creates that money on the spot. Most of the money in a modern economy is bank deposits, and most of those were created that way. The bank cannot do it without limit, since it creates its own IOU and not the central bank money it needs to honour it. It still has to fund the loan and hold capital against it. And how much gets created depends, in the end, on the interest rate the central bank sets. Created on the spot, and not without limit.

Your mortgage did not come out of a neighbour's savings. It was created when the bank credited the account, and each repayment of the amount you borrowed destroys that much money. Your savings were not lent out. They are a debt your bank owes you, guaranteed only up to a limit. What we could not show you is whether the economy would behave any differently if the textbook were right. So if every loan gets funded in the end, does it matter that the money was created first? Tell us which side you're on in the comments.