[kgwgk]

What I’m saying is that if you create a bank and I make a $1m deposit [0] your bank has

  Liabilities : deposit $1m
  Assets : reserves $1m

Now you can lend $900k to someone, who takes the money and tranfers it to someone else’s account in another bank to buy bitcoins or whatever. Your bank has now

  Liabilities : deposit $1m
  Assets : reserves $100k, loan $900k

And you cannot make more loans until you increase your reserves which means getting new deposits (so we’re no longer talking about a bank with $1m in deposits).

Say you get an additional million in deposits, your bank has now

  Liabilities : deposits $2m
  Assets : reserves $1.1m, loan $900k

And it could lend and additional $900k.

Alternatively, it could get funding from another source like borrowing from banks of issuing bonds. For example

  Liabilities : deposit $1m, bond $1m
  Assets : reserves $1.1m, loan $900k

I don’t think any of this is controversial, let alone wrong. By the way, for simplicity in those balanced sheets the “reserves” is all the funding available which has not been lent, not just the required reserves ($100k for $1m in deposits, etc.)

> If the banking system as a whole is leveraged 9:1, that implies each individual bank is leveraged approximately 9:1.

One of the main points of the paper under discussion is that this doesn’t happen. Reserves are not a binding constraint on lending.

[0] I don’t know if it makes much sense to talk about whether it is “actual cash”. Maybe my employer took a loan to pay me, maybe not. Maybe it was in cash, maybe by cheque, maybe by bank transfer. Would that change anything?

[louloulou]

That's no different than what I said, except you're conflating the ledger money created by banks with their reserves at the Central Bank.

Only Central Bank reserves and cash are considered "money" in this system. And I'm saying the "money" is leveraged 9:1 in our toy example, whereas you seem to be saying assets must always be greater than liabilities - which obviously I agree with. Since if they weren't, the bank would be insolvent.

[kgwgk]

You said

"if a bank has $1 million in deposits (of actual cash that people gave to the bank to put in their checking accounts) the bank can make $10 million in loans"

What I say is that it can only make more that $900k in loans _if_ the deposits held at the bank grow above $1m. Which doesn’t normally happen when lending because the most likely outcome is that the borrower takes the money out of the bank.

The reserves of the bank go down in that case, don’t you agree? They are just $100k after the $900k loan is made (and transferred away). $100k are the minimal reserves required when a bank has $1m in deposits.

Again, an individual bank is not the same as the banking system as a whole.

And I don’t quite get your remark about reserves. What I called “reserves” could be entirely held at the central bank (or in the vaults!) if the bank wanted to. Do you have an issue with those balance sheets?

[louloulou]

I shouldn't have quoted $1M in deposits, because I meant $1M in reserves. Since deposits can only become reserves if they are the liabilities of other banks that get settled by transferring reserves to your bank's reserve account.

My only point was individual banks, can lend up to whatever the reserve requirement is. If it was 10%, $1M in reserves at the central bank means the bank can could make $10M worth of loans. Do we disagree on this point?

It seems like we're arguing about what is actually money.

[kgwgk]

> My only point was individual banks, can lend up to whatever the reserve requirement is. If it was 10%, $1M in reserves at the central bank means the bank can could make $10M worth of loans. Do we disagree on this point?

I agree that bank can lend up to whatever the reserve requirement is. But the next sentence is extremely misleading at best.

For that bank with $1m in deposits and $1m in reserves before any lending that 10% requirement means that it can not let its reserves go below 100k (10% times $1m in deposits) so it can only lend up to $900k out.

I can agree if you say "the bank can make $9m worth of loans provided that the recipients of the loans never get them of the bank [and it ends with $10m in deposits]". That is reasonable (even approximatively true) for the banking system as a whole but is a ridiculous implicit assumption for an individual bank.

I would also agree if you said "a bank with $10m in deposits needs to have at least $1m in reserves".

[ubercow13]

>the bank can make $9m worth of loans provided that the recipients of the loans never get them of the bank

The bank can make $9m worth of loans (actually the bank can make any amount of loans, maybe even more), and some proportion of that may be transferred to other banks as reserves, and some other reserves will be transferred onto the banks balance sheet from unrelated transactions the bank makes. Then at the end of the day if the bank needs more reserves, it borrows them. The likely amount the bank needs to borrow based on the loans it makes and the cost of that reserve borrowing determines how many loans it will make. If it wouldn’t be profitable to make more, it’ll stop.

At no point does the bank only make 900k of loans so that it is fully covered in case all its loans are transferred out. The whole thesis of the paper is that that way of thinking is backwards.

[kgwgk]

Then you could just as well say that it can make $9m in loans, $99m in loans or $999m in loans as long as enough reserves are transferred onto the banks balance sheet from unrelated transactions the bank makes (including borrowing if required).

The amount of reserves can (and will) go up and down for an individual bank as it operates depending on their strategy.

10x the initial reserves has no particular meaning for an individual bank, only for the whole system (and the whole thesis of the paper is that even then the 10x number doesn't really matter).