FTX tapped into customer accounts to fund risky bets, setting up its downfall

Correct. The first operation ("writing a number on a computer") can occur regardless of how much money the bank has. But if operation 2 is not possible, then the number that was created in operation 1 is de facto not money. Remember, we're arguing whether the bank needs to have physical banknotes and central bank reserves in order to "create money" when originating a loan. We're not arguing about whether the bank can type in random numbers on a computer - on that point we are already in agreement. The disagreement concerns whether/when those numbers can be considered to be "money".

> All "money" is simply an accounting entry.

Excluding physical banknotes, yes that is true, but you are implying the reverse of that statement to be true and it's not true at all: all accounting entries are not "money". If I open an excel sheet right now and type in "99999999", that is an accounting entry, but it is not money. Likewise, if a troubled bank has completely ran out of capital and is not supported by structures like the FDIC, and it proceeds to type in "9999999" as an accounting entry for the account balance of the chairman's wife, that is not "money". The chairman's wife will not be able to exchange it to goods and services - hence, it is not "money".

> Deposits created by a bank are good for all transactions within that bank. If all people were customers of the same bank (some sort of "central bank" if you will) and all currency were digital, then the bank could indeed create infinity dollars without liquidity risk, because all transactions would occur within their own accounting system with no external settlements.

Yes, central banks have the ability to create practically infinite amounts of real money from thin air. The argument doesn't concern the central bank's ability to create money, it concerns a commercial/retail bank's ability to create money in the process of originating loans.

> > Follow-up question: if you genuinely believe [a bank which has NO MONEY is able to (in a practical sense) create infinite money out of thin air] to be possible, then why isn't anybody doing that? Surely there are many people working at banks who would like to collude with their friends and family to create infinite money. If you believe that to be possible, why has it literally never happened?

> it does happen, fraud in loan origination isn't that rare. Here's a recent example of a banker originating $16m in loans to Paul Manafort in exchange for a shot at working with Trump:

Nope, that is not an example of a bank issuing an infinite amount of money while having literally zero money in reserves. That is an example of a bank which has >16M in reserves, then issuing loans for 16M. Nothing weird about that. Completely unrelated to what I was asking for. Show me a bank which has 0 reserves, then issues loans for billions of dollars, and then exchanges those billions to goods and services. You can't do that, because that has never happened, because it's not possible.

> But let's imagine that a bank did attempt to operate without any settlement balances. What would that look like? It would originate loans by adding $9,999,999,999 to someone's account, and since they're not the only bank, the person has the choice to spend that money wherever they please and some of that money will end up being withdrawn as notes/coins or transferred to another bank. The originator of the loan would then have to borrow settlement balances sufficient to cover withdrawals in notes/coins and cover interbank settlements. What happens? The central bank, which is the lender of last resort, lends money to all those banks and they have to pay interest on it, but in lending the money, the central bank has created reserve balances

No, the central bank would NOT provide a loan in this outrageous, obviously fraudulent instance. Again, you're claiming this to be possible, but it has never happened.

> But fundamentally, we could have a banking system that operates exactly as I described where ALL money creation occurred only through the origination of private credit and subsequent lending of reserve balances by the central bank. It would be a horrible economy with massive inequality and instability, but you could do it, if you so desired. The reason we regulate it is to remove some of that instability (removing the inequality, we're still working on!)

Now I'm confused. So the example you provided was not supposed to reflect reality? It was just a "we could in theory have a banking system like this"? Yes we could in theory, but in practice we don't. In practice normal banks need reserves in order to issue loans.

No bank starts out with $0 capitalisation and then makes up money along the way. This has never happened.

> Commercial banks all agree with each other that they accept each other's demand deposit accounts as a form of money.

This is not true either. If a bank is known to have no capital, other banks will refuse to accept transfers from it without immediate settlement.

A failed bank like Citi should have been handled like smaller banks at the same situation were handled - taken over by the government, shareholders wiped out, new honest management installed, and criminal referrals made as appropriate. In addition, they should've broken up the bank to smaller units that will not be too big to fail.

The way it was done, most of the management layer remained and had a vested interest in covering up problems and their own complicity. Worse, it made it clear to other institutions that the bigger they become, the less likely they are to be allowed to fail. The incentives are perverse.

Edit: fairness is also important. A large, politically connected, bank benefited from bailouts that smaller, less well-connected banks and regular people did not get.

Yes, except that the government has the powers (and had them in 2008) to handle the situation.

> btw, based on your reply I think our positions are probably much closer together than they are far apart.

Agreed.

> I want Glass-Steagall back.

Glass-Steagall was explicitly repealed to allow the merger of Citicorp and Travelers Group which created Citigroup.

Robert Rubin, Clinton's treasury secretary, who promoted for the repeal, joined Citi immediately after leaving government and served as chairmen. He made around $140 million until he had to resign because of the financial crisis.

Citi was too big, too connected, and too corrupt. As long as it so large, it will not be effectively regulated, and it will be bailed out again and again.

If there is no effective regulation, and no legal sanctions, management has a strong financial incentives to take actions that are likely to bankrupt the bank.

There's a great paper about it by two winners of the Nobel Memory Prize in Economic Science. It is based on the lessons of the Savings & Loan crisis.

George Akerlof and Paul Romer: Looting: The Economic Underworld of Bankruptcy for Profit

https://www.brookings.edu/wp-content/uploads/1993/06/1993b_b...