So banks just create money out of nothing, do they?

There\’s an alarmingly large number of people who think that banks just create money out of nothing. You pitch up for a loan and they just magic the cash out of thin air, hand it over then charge you.

The eurozone crisis has left UK banks unable to raise the funding they need to make loans to businesses, evoking the spectre of the crunch that followed the collapse of Lehman Brothers.

If banks really did just create money then they couldn\’t possibly have this problem, could they?

The banks can only issue loans to companies and home owners if they can find sufficient funding on the markets.

That would be impossible, wouldn\’t it?

That banks must gain funding in order to make loans means that quite obviously, banks do not create money. The two positions are entirely exclusive.


39 thoughts on “So banks just create money out of nothing, do they?”

  1. Have you been following the argument between Positive Money’s Ben Dyson and FTAlphaville’s Izabella Kaminska about this?

    Dyson’s article on CIF:

    Kaminska’s critique:

    Comments on this are excellent too, including a long description from Kaminska of how fractional reserve lending works

    Dyson’s reply to Kaminska, which is as silly has his original piece:

    The comments on all three pieces are if anything even more interesting – and vitriolic – than the blogs themselves. Great fun.

  2. I understand what you say. Banks borrow from their depositors and inter bank funds and lend accordingly.

    Do all banks borrow on the inter bank funds?

    What is leverage?

    Tim adds: Leverage is the ratio of the bank’s capital to it’s liabilities (ie, what it owes to other people). It does not refer to the relationaship between deposits and loans, but to capital and deposits.

  3. Peter,

    No, some small banks don’t use interbank borrowing – Kingdom Bank, for example, says it chooses to fund all its lending from deposits, though it places excess funds on the interbank markets. It’s highly unlikely that any big banks are entirely self-funding like that, though.

  4. Banks do not create money “out of thin air” just as 24 frames a second do not, in themselves “move”.

    1. I think the biggest problem with this is how some people think banks operate now like goldsmiths did in the past – issuing scrip for gold deposits – and keep on spreading this seductive pap. Banks give you statements not scrip. If you want a scrip, you must withdraw. A totally different mechanism at the core.

    2. Like Izabella, I am constantly learning. Sometimes I learn a bit from people like Frances, who is kind enough to endure my discourse and has the integrity to agree when actually realising through the fog of my words what I am sometimes getting at is actually correct “sort of” (and most of the time most people are at best that when speaking on this topic).

    3. What I will say is that it appears banks, while not creating money out of thin air, do, as one part of their operations, agree to nett off their respective IOUs at parity.

    If I owe Tim Worstall £10, Tim owes Frances and Frances owes me, do we “create money from thin air” by calling it quits? I struggle to agree with that.

  5. It depends which banks you’re talking about. The central bank does create money out of thin air, effectively. The lesser banks are kind of, branches of the central bank so, the system as a whole creates money out of thin air. Which is how the money supply keeps magically rising, even when they aren’t “quantitatvely easing”.

    The actual problem they’ve got at the moment is there’s no particular commodity that everyone foolishly thinks will be worth zillions more in the future (dot coms, your crappy tumbledown cottage) so there’s nothing to spunk junk loans away on, so the loan system isn’t working as fast as it does during a good bubble. There isn’t a lack of money; there’s just as much (M0) as there was before the crash. They just can’t find enough suckers to turn it into M3.

    Don’t worry, memories are short. The suckers will rise again.

  6. Or to put that another way, I think the question has to be answered, “if the banks don’t create money, then where else does it come from?”

  7. Sorry Tim, but I’m lost here.

    My understanding is as follows and I would be very grateful to be corrected where possible…

    Whilst FRB does not change M0 (cash/bank notes in issue), any bank WILL be creating additional M1 (cash in issue + demand deposits) as it issues loans – at least up until the point where it hits its reserve ratio.

    Once it has reached its reserve ratio, then no, a given bank cannot continue to create more M1, but it IS creating “money” up until that point.

    If I’m correct, I think that’s your point: since most – if not all – banks generally do operate at the limit of their reserve ratio, there is not now new lending without going to the market for funds.

    If I’m not, what have I missed?

    Tim adds: Well, if we decide that we’re going to call M1 (or M3, M4) “monewy”, and in our definition of M1 we’ve got demand deposits, then yes, of course demand deposits are money. But that’s one of the things that confuses in this debate. By calling credit money everyone starts to think that the banks create money not credit.

    As to the bank. What limits the lending is the amount of capital it has. You’ve got to (under Basel etc) have a certain amount of capital put aside each and every loan that is made.

    When the bank makes a loan, sure, that’s an increase in M3 or M4 or whatever. But to fund the loan the bank must go and borrow that money from someone else. Deposits in short (which might come from individual depositors, from bonds they’ve issued, from the interbank market etc).

    It is true that the actions of the banks create credit and that we include credit in some of our measures of money, so they create money in those definitions. But it isn’t true that banks just magic up a loan out of thin air. They must have a deposit to balance that loan: banks books do balance.

  8. Kevin Monk:

    In the second link (vid 4) it is a shame he did not go forward to explain that although M0 is 1000G and M1 is 2710G and multiplied, the sale of the products from the factory and the taxes to fund the irrigation come from the spending of the farmers’, workers and factory contractors deposits to pay off the loan of the irrigation project and the Factory project. At the end of those loans being paid off, we end up still with 1000G physical and 1000G deposited.

  9. Tim,

    “But it isn’t true that banks just magic up a loan out of thin air. They must have a deposit to balance that loan: banks books do balance.”

    the books always balance anyway: when the loan is created, the bank must also simultaneously pay out cash so that assets fall by the same amount or they credit a demand deposit account with that cash creating the corresponding liability.

    The problem is whether creating such a loan (and exposing the bank to additional calls on its cash/capital) would cause it to fall below the minimum capital requirements.

    Now… a brand spanking new bank having just raised its initial equity capital can quite merrily issue loans without having to go to the market for funds. In this respect the assertion that banks can magic up money is true.

    BUT…. it doesn’t hold in reality because:
    1) all existing banks run close to their capital requirement. It’s what they do (LoLR makes this inevitable etc). So existing banks cannot issue new loans by magic – they’ve already done as much of that as they can.
    2) NEW banks do raise capital in isolation. For every £1 that a new bank A raises to allow it to magic up new loans, that £1 has been withdrawn from some other bank B somewhere else, which means that bank B has had to reduce its lending activity by the same amount.

    Is that a fair summary?

    If it is, the only problem is that that is too difficult to explain to your average GROLIES. 🙂

  10. We had a huge debate about all this only about a week ago. Really it amounts to arguing about about which end you crack an egg.

    Yes, when a loan is granted a bank doesn’t have to have settlement funds available – but it does have to have them once the loan is drawn. Provided it can cover any cash call it doesn’t have to have additional physical money at the time of drawdown, but it must at the end of the day cover all deposit withdrawals INCLUDING loan drawdowns either from existing deposits or by borrowing. So the “unfunded loan commitment” that everyone gets so upset about is at most a daylight overdraft. Really, what on earth is the fuss about?

    Of course the excess deposits they borrow through the interbank markets to balance their books could have arisen from people spending their loans. After all, no-one borrows money to stuff it under the mattress. They spend it, and that spending becomes savings for the recipient, which are very likely to be placed into a deposit (or current) account until they are spent in turn. No-one ever said this wasn’t circular. But that is not the same as suggesting that banks can invent money to settle loans. They can’t.

  11. Frances,

    Two things (and bear in mind that my understanding of this stuff is very sketchy so please take with pinch of salt – will gratefully accept corrections, no matter how apparently patronising…)

    “Of course the excess deposits they borrow through the interbank markets to balance their books”

    They aren’t ever borrowing to balance the books: the books always balance or you’ve got fraud somewhere: it’s simply an accounting identity – they’re doing it to cover the capital requirement aren’t they?

    “But that is not the same as suggesting that banks can invent money to settle loans. They can’t.”

    Unless the GROLIES are being spectacularly ludicrously howlingly stupid, I don’t think that that is being suggested. They are simply suggesting that banks can magically increase the money supply to create loans which others must pay back to them (which would rather destroy the point of using to settle a loan…)

  12. Roger (11)

    Well that’s screwed with my head!

    I’m not sure I understand your point.

    If the loans were repaid then the bank would have to deleverage and the whole process would reverse. M1 would drop back to M0.

    As I understand it, that’s why central banks pump money back into the system – as an attempt to counter the effects of deleveraging and the destruction of M1.

    This wouldn’t be the case though if the factory and the irrigation ditches were good investments? The farmer wouldn’t sell his irrigation ditch for 900G because it is now worth more to him. I’m guessing that good investment will limit a bank run.

    You could run this whole scenario where the people weren’t investing the money at all and just simply lending it on to the next person. You could run this scenario a thousand times and it won’t make the slightest difference to the total wealth. Therefore, I’m not sure why you think that if you completely deleveraged you’d end up with 1000G and £1000 deposits. You’d just have 1000G.

    * The usual “I’m an interested amateur who is keen to learn more” caveats apply.

  13. The Pedant-General

    Not quite. They are borrowing to balance their BoE reserve accounts, actually.

    Typically when a loan is created there is a debit to the customer loan account and a corresponding credit to the customer’s current account. When that loan is drawn there is a debit to the customer’s current account but the loan amount remains the same. That’s the funding gap. It is in effect a daylight overdraft. Banks cannot run an overnight overdraft in their reserve accounts without BoE approval, and they get charged penalty rates for doing so – so they usually borrow the difference on the interbank markets. I hope this is clear?

    This is an entirely different matter from the creation of credit, which as Tim correctly points out is capital constrained. Unfortunately the “banks create money from thin air” goons can’t tell the difference. The accounting entries for committed lending inflate the bank’s balance sheet and therefore drain capital. I slightly take issue with Tim over credit creation: in committed lending, a real deposit is created which is indistinguishable from a customer deposit and is therefore included in M(n>0). Therefore the money supply does increase as a result of committed lending, whether or not that loan has yet been drawn and irrespective of the availability of funds for settlement. Yes, in uncommitted lending (such as overdrafts) credit doesn’t become “money” until it is drawn. But in committed lending money is created when the loan is granted.

  14. The point here for me is this; while there are some goons who have a very simplistic view that a bank can just invent money willy-nilly, I think the more considered view which is widespread is that “the banks” refers to the banking system as a whole which quite clearly does create money. Or more specifically, it is not FRB itself that creates money (though it creates an Mx which is some multiple of Mo) but the combination of FRB and a State sponsored central bank that creates money.

    Hence Tim’s original post is a bit straw-man inclined.

  15. Francis,

    I think we’re getting there. It’s all beginning to sound like the early 80s when the news discussed the various measures of money almost every night. 🙂

    “That’s the funding gap. ”
    Got it – I’m assuming that they’re inside their reserve ratio and hence have cash to be able to allow the loan to be drawn down.
    If they haven’t got that cash, surely they are outside the reserve ratio – they can’t meet demands for withdrawals?

    But either way, it’s understood.
    “Hence Tim’s original post is a bit straw-man inclined.”
    I think that’s my concern: he’s right, but not quite for the right reasons. Your GROLIES need only show you a newly capitalised bank and he can demonstrate a bank doing precisely this.

    Tim’s point is that, taking the system as a whole, it never happens because the newly capitalised bank has reduced the capital somewhere else by exactly the same amount.

  16. @TW
    Ian B’s question @ 9 has to be answered : if the banks don’t create money who does?Or put another way: if the banks’ books balance, how come the money supply keeps on increasing?
    You’re the expert ,or so you keep saying telling us. I would have thought where money comes from is pretty basic for an economist of your world-wide fame.You seem to have immense trouble mustering an explanation that rebuts the thin-airists.It sounds like you’ve only begun to think about it.

  17. @DBC,

    You’ve not been following. This is all dealt with in the comments.

    To summarise:
    Banks create money in the very short time when the FRB system is first created. This process stops quite quickly when they hit a prudent reserve ratio. Lend too much and you are risk of liquidity problems and a bank run.
    This is so long in the past that it is simply no longer relevant.

    Thereafter, or rather, once you have a central bank/printing press/LoLR, it’s the central bank that prints money and the thin airists are very specifically NOT talking about central banks.


  18. @PG
    This is just a skim of the comments.You bank defenders are shifting your ground to say banks create money when they start out but this is” so long in the past it is simply no longer relevant ”
    Now its the central bank printing press.Yes right:you sure its not the trade unions who were once held directly responsible for all the inflation?Please explain how the nationalised central bank is churning out money.QE when the Bof E creates money by issuing unsupported cheques?And what supports the loan cheques the banks issues to customers then ?
    I would like an answer from the man himself please.

  19. PG,

    No, still not quite right, sorry. Lending is not reserve constrained, it is capital constrained. Banks can lend up to their capital limits. Reserves are balanced after the fact and are not relevant to lending decisions. So instead of “reserve ratio”, please substitute “capital requirement”.

  20. DBC Reed

    Sorry, I’m not TW but I do know quite a bit about this so hopefully my comments will be helpful.

    Money creation occurs in committed lending as a consequence of double entry accounting. Each loan is balanced by an equal deposit, which at the point of loan commitment is not borrowed but created. That deposit is placed in a customer demand deposit account, where it is indistinguishable from customer-deposited funds and can be drawn as a normal deposit withdrawal – you know, to pay the plumber in cash, or write a cheque, or do a bank transfer…. Demand deposit accounts are included in M1 to 4. Therefore the creation of balancing deposits in committed lending inflates the money supply. The pair of accounting entries (loan and deposit) also inflates the bank’s balance sheet and drains capital equal to the risk weighted amount of the loan. Banks can only lend up to their capital limits. The bank’s ability to lend, and hence to “create money”, is therefore constrained by the availability of capital.

    In uncommitted lending (eg overdrafts) the balance sheet asset and liability are created when the facility is drawn – prior to that it is a contingent liability, so off balance sheet. But it still inflates the bank balance sheet when it is drawn, it is still capital constrained and it still has to be funded by borrowing. Personally I think it also inflates the money supply, because a new deposit is effectively created somewhere else when the loan is spent. But we can argue about that – funding proximity does muddy the waters.


    No, commercial bank leverage of central bank money didn’t happen “so long in the past that it’s not relevant now”. It is continually happening now, because lending precedes funding (as I have described) and created deposits are indistinguishable from customer deposits so can be used to fund drawdowns. It is not constrained by reserve availability, as central banks in practice always provide liquidity to commercial banks that have lent more than they have funds to settle. It is constrained by capital availability, as I have described.

    I hope all of this is clear? The bit that everyone misses is the inflation of savings due to deposit creation. Those inflated savings are used to fund drawdowns.

  21. So I think we’re back to this; the banking system in toto creates money (“out of thin air”). I think everyone here knows that my local NatWest aren’t allowed to arbitrarily create it, which is why I said this has an air of the straw man about it. It’s a bit like arguing about which particular part of a car engine makes the wheels go round, when the answer is “all of it”.

    The central complaint of “thin-airists” is about monetary expansion. How the system expands the money supply is an interesting (and complex) subject- as we see above- much like the precise engineering details of an engine, but none of it denies the central charge that the engine makes the wheels go around. So, ultimately the sentence

    “That banks must gain funding in order to make loans means that quite obviously, banks do not create money.”

    does not posit two things which are mutually exclusive after all. Particular banks must gain funding to make loans, but the banks as a system do indeed create money, which is what everyone, or at least “thin-airists” actually care about.

  22. @FC
    Many thanks but I already know most of this ,particularly the point directed at the its-all-so-long-ago PG .
    The distinction between committed and uncommitted landing is new to me ,(it’s still amounts of money in both cases surely) ,and I am not clear what you mean by separating capital and reserves.After all,the conflation “capital reserves” is often used.
    Still not a peep out of TW.Looks like he’s keeping his head down and leaving his supporters to do the work.

  23. Gah, I really shouldn’t write complex stuff in the middle of the night. Loan accounting drains capital proportionate to the risk weighted amount of the loan, not equal to…..that’s why we have capital ratios…..

  24. DBC Reed

    Capital and reserves are totally different things, but they are frequently conflated and often confused. I wrote about this here:

    In committed lending the loan is legally committed to you and the bank cannot withdraw it once the agreement is signed. The money advanced is therefore “real” even though you haven’t drawn it, which is why it is on balance sheet from the point of commitment.

    Uncommitted lending is a facility which you MAY use but the undrawn portion of which can be withdrawn without notice at any time. It is a contingent liability from the bank’s point of view so off balance sheet. Off balance sheet liabilities don’t inflate the money supply. Therefore the undrawn portion expands credit but not the money supply. Credit cards and overdrafts are uncommitted lending, whereas a bank loan is committed lending.

  25. FC,

    “No, commercial bank leverage of central bank money didn’t happen “so long in the past that it’s not relevant now”. ”

    That’s not what I said: the bit that is all so long in the past is the FRB bit PRIOR to the central banks.

    Ergo – and this is the important thing – without the central bank printing money out of thin air, the banks cannot increase lending beyond their capital ratios (which they will reach quickly then stop).

    My understanding of the thin airist complaint is that they think banks OTHER THAN the central bank create money out of thin air. My assertion is that this is not so: they merely act as a multiplier on the central bank printing. Without that central bank printing, there can be no additional money creation by non-central banks.

  26. IIRC Teh Banks (SUM(B1:Bn)) would enlarge M1 if more wealth is accumulated in banks or new banks are formed from capital that would otherwise be put elsewhere.

    To me, though, the issue is what is M1? If I have lent money to another (deposited it), then the money is not mine anymore. All I have is a loan. I have to ask for it back*. If vast amounts of M1 are actually just records of lending to a Bank, then I do not call that “money”. You cannot spend what someone else owes you and has already been handed over to a third (normally another bank to balance their requirements).

    It only becomes money once it is instantiated and this is why I say the point that things get interesting is that banks freely swap each other’s IOUs in netting off. Even then, they are basically meeting the needs of their depositors who are happy to accumulate their own bank’s IOUs when a third party settles a debt. The depositor who is getting transferred an amount is basically saying “Can you lend my bank £x and let me have that loan as settlement for the debt”.

    What happens, of course, is that in netting off, one depositor-lender no longer lends to Bank A but lends to Bank B at the same time as another depositor-lender no longer lends to Bank B but now lends to Bank A. The first sells the loan to Bank A to buy a loan to Bank B while the second sells a loan to Bank B to buy a loan to Bank A. The banks are intermediaries in that pair of transactions. Is money being created here? These bank loans could be CDs. Do you think people holding CDs of de facto pure liquidity delivered by market makers are inflating the money supply?

    * I think one of the problems is this idea of a “demand deposit” making people think they have their cash sitting around in a shoebox somewhere…

    p.s. Everyone withdraw 10% of their deposits and put them under the mattress and lets see how much money private banks can “create”…

  27. PG

    Maybe it’s me that’s not getting it. Why are you talking about central bank printing in relation to bank capital ratios? The two are not in any way related.

    Capital is not static. Banks can, and do, increase their capital, through rights issues, retaining earnings or by selling assets. Securitising assets is the primary means by which American lenders get loans off their balance sheets to free up capital so they can lend more. There are other ways too, such as using derivatives such as CDS to disperse risk, or improving the quality of lending: both of these have the effect of freeing up capital. That’s why banks can continue to lend even when close to their capital limits.

    It is correct in theory to distinguish between money created by the central bank and provided to commercial banks as reserves, and money leveraged by those commercial banks through credit creation. In practice the effect on M1 to 4 is the same, because as I said it isn’t possible to distinguish between different types of deposits. You can buy a pint of beer with a loan deposit just as easily as with your wages. Bank reserves are included in M0, I think, so central bank printing increases narrow money as well – that’s the difference.

  28. @FC Not much point in replying now as this is about to drop into the void.But I would have though uncommitted lending would increase money supply if everybody maxed out their credit cards.Is n’t this what’s happening now? Also the distinction you draw between reserves and capital aids the thin airists since it appears that no deposits are included in reserves. (They tend to think some deposits underlie the pyramid of loans.

  29. DBC Reed

    Last try, honest…..

    1) Maxed-out credit cards are fully drawn uncommitted facilities. Personally, as I said, I think drawdowns against uncommitted facilities do expand the money supply. But others may disagree with me.

    2) Deposits are included in reserves. They are not included in capital as they are senior debt.

  30. Roger

    If a facility such as a credit card can be withdrawn without notice then it would be regarded as uncommitted and only the drawn portion would be on balance sheet. However, capital has to be allocated against certain off-balance sheet exposures as well. I’m really not sure what the position is with credit cards regarding capital allocation – it all depends on the bank’s notional risk exposure, which in turn depends on how the legal contract is interpreted.

  31. Anybody [email protected]
    Slipped up in the above by saying that no deposits seem to be included in reserves when I meant capital using your nomenclature.Since they do’nt contribute to capital,this is more food and drink to the thin airists who say deposits are not even looked at when making more “loans”.In fact this is entirely their point.

  32. Consider this: A bank has 1M in capital and a current balance sheet of 99M. The max balance sheet based on 1M capital is 100M. With its last 1M of potential balance sheet increase the bank approves a loan of 1M to the CEO of the bank at 2% annual interest (adding 1M in bank assets and 1M in bank liability). The CEO now draws 1M from his deposit account and buys newly emitted shares in the bank earning 3% annual dividend. The bank has to seek other funding for that 1M drawn by its CEO (to keep liabilities equal to assets). The bank now has 2M in capital and can expand its balance sheet by another 100M to 200M.

    As I see it all FRB issued money (moneyprinting) will at some point end up as capital in the banking system. Thus it is inevitable that at some point in a future credit boom 3T extra FRB moneyprinting will lead to approx 300T extra credit/debt/money in the economy based on current basel 2 capital requirements. Of course BIS have plenty of control over not just central bank money but also the world wide credit multiplier. They are in full control of the money system. They really are the money masters creating booms and busts at their own will and pleasure.

Leave a Reply

Your email address will not be published. Required fields are marked *