Friday, 9 February 2018

What is the Benefit to Banks from Money Creation?

In a response to a recent post by Brian Romanchuk, somebody made the following comment:

"If private banks are ..... allowed to create and lend out their own money, they can undercut the ..... free market rate of interest, and for the simple reason that printing money is cheaper than having to borrow it or earn it."

This seems to suggest a kind of model in which banks choose whether to finance themselves with someone else's money that they have to pay to borrow, or money they create for themselves for free.  I think the problem is that this confuses two distinct ideas: that there is a benefit from having monetary liabilities and that bank lending increases deposits.

The potential benefit that accrues to banks by virtue of their status as money issuers arises through a reduced rate of interest on monetary liabilities.  If a particular type of bank deposit, such as a positive current account balance, is readily available for making payments, then it typically carries a lower rate of interest than other deposits. 

Sometimes the rate of interest on such balances is zero, but it need not be.  The important point is that there is a benefit to the bank through a reduced funding cost.  Set against this is the cost to the bank of providing current account services in the form of the costs of premises, staff and equipment.

At this point it is worth noting that these costs and benefits are based on the level of the bank's outstanding monetary liabilities.  It is nothing to do with which bank makes the loan that creates the deposit.  It is quite possible to have banks making lots of loans, but having minimal liabilities in the form of immediately available deposits, because that bank relies on different funding techniques.  These banks would be creating new money, but not getting any of the potential benefit that arises from having monetary liabilities.

On the other hand, it would be possible to have a bank with very large current account liabilities but which never engaged in deposit creation.  This would happen if the bank was simply taking deposits through payments received in from other sources and making all of its loans in cash[1].  The potential benefit of operating current accounts would be very important to such a bank.  

The point here is that it makes no sense to say that it is cheaper for a bank to print money than borrow it.  What the bank does at the point of making a loan is irrelevant.  What matters is how it chooses to manage its liabilities going forward and in particular the extent to which it chooses to compete for current account deposits.

The extent of the benefit depends then on how competitive that market is.  Under perfect competition, banks would have to offer interest rates on current accounts that would simply leave them with normal profits.  However, it is likely that there is a degree of monopolistic competition in the provision of banking services, particularly at the retail level, and this means that there is some supernormal profit that accrues to banks as providers of monetary liabilities.

It is difficult to assess how profitable it is for banks to have monetary liabilities, largely because many of the costs are shared with other activities.  Even for the banks themselves, it is somewhat arbitrary how costs get allocated.  However, the point here is that any such profit is just regular monopolistic profit in the market for current account services and not something to do with money being created out of thin air.

[1] Making loans in cash does in fact "create money" in the sense of increasing the broad money supply, but it is not what people usually have in mind when they talk of banks "printing money".


  1. A bank with many customers (creditors) who transact with each other would not be forced to make as many inter-bank payments. This should give a profit advantage (economy of scale) to a larger bank compared to a smaller bank if its customers do business with each other using the bank's liabilities to clear payment and no other banks are involved in the transaction clearing.

    1. Maybe, but I think that the marginal cost of making inter-bank transactions is tiny compared with the other costs. The problem with evaluating this is that most of the costs are overheads, being things like staff and premises.

    2. How much staff is required to debit one account and credit another on the books of a single bank when clearing a transaction between two creditors? Not so much. There should be less cost involved compared to efforts to attract equity or debt investors from the aggregate bank sector.

    3. Well, I agree that the marginal cost of making a payment is small, even for those that involve cheque processing. It's all about the overheads. And these are considerable when you have lots of customers with very small balances. Automation has clearly reduced some costs, but at the same time other costs have risen, such as compliance.

  2. In my mechanical model of money, 'printing' money occurs whenever the creator-of-money spends and then never recovers the newly created money. This can occur when government borrows and never bothers to repay(eliminate) the loan.

    Only a Central Bank (with the cooperation of government) can accomplish this type of money creation.

    Continuing my mechanical model of money, private banks can only lend money that they have in-hand. Because they are privately held, banks expect repayment of any loans made. Hence, while a private bank loan may increase the amount of money as-measured-by-active-bank-deposits, the amount of base money available does not change. [So what is 'base money'? Base money is money created by government borrowing- never-repaid.]

    Interest rates do not matter in this money creation model. Interest rates are an imposed penality designed to encourage people to behave in a predictable way.

    1. This of course depends on how you want to define money and what you mean by printing money. On the whole, I try to avoid using the term "money" (unless in the context of a model with limited clearly defined assets) prefering to talk about deposits, cash or, maybe, monetary liabilities. I have used the term "money" here, but only because I am addressing someone else's usage.

  3. Ten years ago, when there was all of the news about Northern Rock collapsing, I think I remember people saying that Northern Rock was in a risky position because it had so much of its liabilities in the form of short term money market borrowing from institutional investors rather than in the form of retail customer deposits. The idea was that even though in principle customer deposits are the shortest term form of liability (ie instant maturity) in practice deposit accounts are typically kept at a stable level overall whilst money market lending gets withdrawn on a large scale much more often. So perhaps banks benefit from deposits as a stable source of funding and not just as a cheap source of funding? Of course if the time to maturity of banks' liabilities was matched to the time to maturity of their loans, then that would be the most stable arrangement of all but, as you say, that funding model would be more expensive for the banks.

    1. That is absolutely right. Retail deposits are seen as being more sticky that inter-bank funding. This is now formalised in the Liquidity Coverage Ratio and Net Stable Funding Ratio measures. That is why retail savings accounts can pay higher interest than the inter-bank rate, even those that are repayable on demand. Such accounts carry higher admin costs than wholesale funding, but much less than current accounts where there is frequent turnover. These are all factors to be taken into account when weighing up the costs and benefits of retail deposit taking. But you still come back to the fact that any excess of benefits over cost is just a function of the competitiveness of the market and nothing to do with loans creating deposits.

    2. The liabilities of each bank could be categorized in the first analysis as the equity (loss cushion), insured deposits, and uninsured deposits. In the United States the large depositors tend to be uninsured and are expected to have sufficient resources to monitor the asset quality and other stability factors of the bank. If the large depositors want something like deposit insurance then the bank would issue repurchase agreement liabilities in the overnight or short term money markets. One feature of the global financial crisis is that the repo haircuts increased which is like causing a bank run on the repo liabilities which want to revert to uninsured deposits. So the logic is based on who has an incentive to cause a "run" on bank funding under prevailing conditions and not whether deposits per se are more stable then short term money market borrowings. If bank assets are corrupted by the need for a write-off and such loss is not firmly allocated prior to the event there is an incentive to withdraw funding to the bank, and there is no incentive to put new paid-in equity into a bank, in anticipation of having a write-off of equity.

    3. I think you're commenting on stone's comment rather than my reply. I would say though that when BIS use the term "stable" in talking about net stable funding, they mean what you seem to mean in talking about susceptibility to runs. Also, on secured wholesale funding, this can indeed be treated as being more stable, depending on the collateral, but any assets used as collateral are then treated as encumbered on the other side of the ratio.

  4. Bank do benefit, in total profits, from being able to create liabilities in much larger quantities than CB clearing deposits, as it enables them to undertake more business than would be the case as the interest rate is lower.

    1. There's no question that banks make profits from providing monetary services - my point is that the level of those profits is determined by the usual issues about monopoly power and not by anything special about money creation.

    2. Considering the operational functions of the banking system a huge amount of the turnover is account for where liabilities are created and hence profit, because creating liabilities for the use of customers reduces the interest rate on loans , and so there are more customers, lending, turnover and profit.

    3. I'm not entirely sure whether you are agreeing or disagreeing with me, but "creating liabilities for the use of customers reduces the interest rate on loans" doesn't sound to me like something that is correct.

    4. This comment has been removed by the author.

    5. Hi

      I'm thinking that if every £ of loan to customers on banks balance sheets had to be pre sourced as assets of central bank money, then the interest rate would be higher, due to the demand for CB money. In a different system the CB could create more CB money. But as it is, with banks creating money supply over and above the supply of CB money, and the CB setting the CB money rate consistent with the knowledge of that process, banks are systemically benefiting from creating their own money in that system as it is as it enables the the banks to charge borrowers a lower interest rate.

    6. In general, it works like this. If Bank A expands its balance sheet by lending more, it increases the supply of Bank A liabilities (i.e. deposits with Bank A and the like) relative to everything else (disposable income, the total supply of financial assets, etc.). This means that Bank A has to induce people to hold a higher share of their assets in the form of Bank A deposits. To do this it has to pay a higher interest rate. How much more depends on how monopolistic the market is. The bank has a bigger balance sheet, but thinner spreads.

      Don't get fooled by the line that banks either borrow money at interest or create it for free. It's complete nonsense and totally misunderstands what's going on when private money is created.

    7. The monopolistic component of the banks ability to obtain reserves by attracting deposits affects the cost of obtaining CB money. That cost is the interest rate of using CB money.
      But, the lending rate that it CHARGES, is determined by the banks ability to create deposits over and above its use of CB money. It is lower because its cost of funding per £ lent is lower due to it only having to obtain a fraction of CB money when it lends.

      For example here is a quote from Federal Reserve Bank San Francisco

      "Increasing the (reserve requirement) ratios reduces the volume of deposits that can be supported by a given level of reserves and, in the absence of other actions, reduces the money stock and raises the cost of credit."

      And so, for increased demand at the lower price and the same profit profit margin per £ lent, total pounds lent increases at the lower price, so more bank lending and higher total profit.

    8. We have to be careful not to mix up regulatory systems here. You give a quote relating to reserve requirements, but also refer to £ lent. The UK has no reserve requirement rules like the US, so that quoted comment does not apply to the UK.

      If we consider a system that does have reserve requirements (and, critically does not pay interest on reserves) then the requirement to hold a quantity of non-interest bearing balances is an additional cost to banks (the lost interest). Increasing the requirement increases the cost, and banks will try to pass this on to depositors through lower interest rates on deposits. This is also likely to lead indirectly to higher rates being charged on loans.

      The fact that the rates on loans and deposits are being varied in response to changes in the cost imposed by the reserve requirement tells us that this cost is, at least partially, being passed on to bank customers. This is exactly the point I'm making.

    9. This comment has been removed by the author.

    10. Banks attract deposits because they want CB money for transacting with other banks. And the ratio of commercial deposits to CB money is about 10 to 1. If it were instead 1 to 1 the industry demand for CB money would be increased lets say 10 fold.
      I see your point that banks creating monetary liabilities is not cost free, but that cost is much much smaller, maybe 1 tenth of the cost, as a result of creating them without the acquisition of CB money, and not correspondingly bidding up the wholesale price of CB money in the process.
      This is maybe not, to be considered to be a "benefit" to the modern and mature banking industry as a whole because the CB sets its target rate for CB money in the industry in accordance with these operations. However banks on an individual basis can benefit from using ways to reduce there own particular use of CB money for transacting with other banks.

    11. If we are thinking about what happens if there is an increased reserve requirement, we have two different assumptions we can make. We can assume that the Fed maintains its target Fed Funds Rate, in which case it will have to increase the supply of reserves, or we can assume it keeps the supply of reserves constant and lets the Fed Funds Rate rise.

      It seems you are thinking of the latter. Letting the rate rise does indeed increase even further the cost of running deposit accounts and will lead to higher loan rates and lower deposit rates.

      But the key point here is financial institutions face the same cost regardless of the operational procedure of making loans. A bank that can "create money" by crediting deposits directly does not face a lower cost than one that can only increase its deposit base by taking in external payments.

  5. ... the abhorrent ("heterodox") simplicity of "loans create deposits"

  6. Hi Nick,

    I’m the “somebody” made the comment you referred to at Brian’s blog. Thanks for your response.

    You claim that banks’ funding costs are reduced by reason of letting customers withdraw money instantly from their accounts. True, but so what? That in no way impinges (contrary to your suggestions) on my point that acquiring money by printing the stuff is cheaper, all else equal, than borrowing or earning the stuff.

    The fact that a following wind reduces fuel consumption in a car in no way invalidates the point that cutting down on the car’s weight, all else equal ALSO reduces fuel consumption.

    Next, I don’t agree with your claim that “It is quite possible to have banks making lots of loans, but having minimal liabilities in the form of immediately available deposits, because that bank relies on different funding techniques. These banks would be creating new money…”
    Where a bank (or more accurately a “lending entity”) is funded just by equity, which is what is involved under full reserve, banks do not create money. (Funding purely via equity involves zero liabilities, at least in one sense of the word liability.) Reasons are as follows.

    Where a bank under the existing system lends on $X of deposits, the borrowers have access to $X, but the depositors still have access to $X as well. Magic! $X has been turned into $2X.

    In contrast, where loans are funded via equity, people buy $X of equity in a bank and $X is loaned on. Net result is that the borrowers have $X instead of the equity holder. No dollars have been created. (The “depositors”, if you want to call them that, own $X of equity, not $X.)

    As mentioned in my comment, I’m basically repeating a point made by Joseph Huber. George Selgin actually made the same point, namely that given a “base money only” system (i.e. full reserve banking) and assuming private banks are then allowed to create / print money, the latter banks will initially go wild and print like crazy, because they can undercut the existing system. That’s in Selgin’s “Capitalism Magazine” article “Is Fractional-Reserve Banking Inflationary.” See 2nd para onwards here:

    That’s not to suggest Selgin supports full reserve, but he does seem to be making a very similar point to Huber.

    1. Ralph,

      Thank you for expanding on your point.

      Nothing was really supposed to ride on the point you refer to under "You claim that..", nor was it intended to be contentious. I was simply trying to frame the issue.

      Technically, even if a bank is 100% equity funded, the loans it makes will create money. Consider Bank A which has £100 of equity and no debt. Its assets are £80 of customer loans, £10 of interbank loans and £10 of cash. If it makes a loan to a customer, it either gives up some of its cash or makes a transfer from its interbank balance. Either way increases the money supply, because holdings by one bank of instruments issued by another are not treated as money.

      Of course, you could define money differently or you could imagine a different system, but whichever way you do so, it will still be the case that any benefit that might accrue to then banks must be measured by looking at the amount of such liabilities that they have outstanding, not the rate they are created at. I'm pretty sure that this is what Joseph Huber would say. In "Creating New Money", Huber and his co-author talk about the benefit to banks as "Special Banking Profits" and they estimate this based on the outstanding stock of monetary liabilities.

      Where perhaps I disagree with Huber is on my second point which is what drives those profits. My argument is that these are just regular monopolistic profits and that if provision of monetary services was a perfectly competitive market, there would be no supernormal profit. Actually, it is not obvious to me that Huber would disagree with this, but I would say that if we are simply talking about monopoly profits, it would save a lot of confusion if we just said so.

      (Also, Huber seems to have forgotten costs when he calculates profit, so his figures really just represent net revenue.)

      I've had a brief look at The Selgin article. I would probably have some comments to make, but essentially it just seems to be an argument based on the money multiplier model and the quantity theory. I didn't notice anything that sounded like he was saying that "printing money is cheaper than having to borrow it".

      If we consider Selgin's transition from a 100% reserves regime to fractional reserves banking, we might compare two different ways this could happen: First, banks might make loans by simply crediting deposit accounts. This is what I guess you would call "printing money". Secondly, banks could take in deposits in the form of base money (Selgin's "warehoue bank certificates") and then lend these out. This, I suppose, is "borrowing" the money. However, in Selgin's set-up, it makes no difference, because you always end up at the same ratios.