[milderworkacc]

> This is just completely wrong - banks only loan out money they have. If you go to a bank and get a loan the bank didn't just edit a database entry - they had that money. Banks loaning out money they don't have is extremely illegal.

I'm afraid to say that this is completely wrong. Commercial banks do in fact create money via lending! The 101 textbook explanation offered here is at best outdated and at worst misleadingly perpetuates a myth that simply must die.

The Bank of England's note on money creation in the modern economy [0] is the place to start - and more or less reflects the explanation in the article.

[0] https://www.bankofengland.co.uk/quarterly-bulletin/2014/q1/m...

[lesuorac]

Page 19 seems to show that the newly generated money has to be backed by somebody's deposit and not just minted though. Like for any situation that doesn't involve you taking a loan from a bank to use that money at the bank the BoE paper calls out that the bank needs to get deposits from where you're sending the loan.

So, you buy a house using say Chase and the seller has Wells Fargo then Chase "mints" say 1M the money for your loan but then solicits 1M of deposits from Wells Fargo. like why does it matter if the 1m for your loan came from Chase depositors or Wells Fargo depositors? The point is that it's backed 1:1 by cash that came from a person which BoE example shows.

[throwawaymaths]

> the newly generated money has to be backed by somebody's deposit

Ok. Suppose I am a bank, and pg deposits $10. Then I lend to you, lesourac $9. This $9 is "backed" by pg's deposit. But I, the bank, only hold $1, and you, lesourac, hold $9.

In your head, you hold $9. In pg's head, he has $10 of assets. There are $19 of imagined assets running around, even though the "real" assets are only $10.

It's all good until pg pulls his $10 sooner than I expected, or if you, lesourac declare bankruptcy and default on your loan, and unable to pay back those $9. This is why bankruptcies are deflationary.

We could have a safer banking system if loans were from individual to individual, possibly mediated by a bank, and the lender fully accepted the risk of default.

[lesuorac]

I think you've missed the argument.

Nobody is arguing that the Money Multiplier [1] doesn't exist. Nobody is arguing that a bank run won't cause loss of deposit (ignoring FDIC).

The argument is whether a bank takes in say $10 of deposits to then loan out $10 OR if a bank "generates" $10 at-will to make a loan of $10.

[1]: https://en.wikipedia.org/wiki/Money_multiplier

[fragmede]

Oh interesting. That seems academic given the reserve requirements by the Fed under Regulation D which basically says every bank must have $x dollars in reserve for every $10 loaned out. Thus, the bank already has to have that $10 deposit before it's able to make the loan, but as long as those reserve requirements are met, they can make all the loans they want. Given the size of a commercial bank with deposits and withdrawals going on constantly across many customers, they run with some amount of margin, but basically they'll be able to make a good number of loans before they're anywhere close to their reserve limit. So to answer your question... sorta?

https://www.federalreserve.gov/monetarypolicy/reservereq.htm

[lesuorac]

I think you've also missed the argument.

So, the side I don't particularly like goes like this. Alice deposits $10 into Bob's Bank. Bob's bank now has $10 of liabilities (Alice's account) and $10 of assets (Alice's ex-Money). Charlie _wants_ a loan of $10. Bob records an increase of assets by $10 so now the bank has $20 in assets and records a corresponding increase in liabilities (to balance out the minted money) of $10 so now the bank has $20 in liabilities. Then the bank gives Charlie this _new_ money. In essence, the bank is not giving out depositor's money for loan but instead new money. (See the Bank of England (BOE) paper for details showing this [1]).

My argument is that (1) the same BOE paper shows that whenever that minted money needs to leave the minting bank's computer systems an equal amount of money from somewhere (either that bank or the receiving bank) will be destroyed. (2) That money will commonly move between banks. Therefore the fact that money is minted is irreverent because it's subsequently quickly destroyed.

Side note, it makes total sense to me that a bank would rather mint money in a lump sum exactly equal to the amount needed for a loan than figure out what fractions of the loan should come from what depositor. Its just practical.

[1]: https://www.bankofengland.co.uk/quarterly-bulletin/2014/q1/m...

[Nursie]

This link sure does do the rounds! It's practically a meme in itself these days.

But it's only part of the picture, for instance, large banks have capital requirements - https://www.federalreserve.gov/supervisionreg/large-bank-cap...

[nerdponx]

> The argument is whether a bank takes in say $10 of deposits to then loan out $10 OR if a bank "generates" $10 at-will to make a loan of $10.

In some sense, the loans are real and backed by real money, but it's your balance that is generated from thin air when the loan is issued.

[throwawaymaths]

Do you consider the value in your yield bearing savings account to "not be money" in the sense that your stocks "aren't money"? If the median American does, then your argument holds water. I don't think this is the case, though.

[lesuorac]

The BoE considers savings accounts to be part of the 97% of money in an economy. Although I'm still not too sure if you're on the side of a bank using a depositor's money or minted money for a loan.

> Of the two types of broad money, bank deposits make up the vast majority — 97% of the amount currently in circulation.