Recent Posts

Wednesday, December 18, 2019

Bank Deposits And Funding: Quick Comment

There have been comments on my earlier article on bank deposits, and I just wanted to add some thoughts. I expect that I will eventual write a book (similar in concept to Abolish Money (From Economics)!) on Fractional Reserve Banking. (Abolish Money! has sections on banking, but not enough to cover everything that comes up.) In my view, overly-simplified primers on banking systems give a misleading picture about banks, leading to pointless mystical debates about bank "money creation."

The argument is straightforward: in the English-speaking world, we no longer live in a world where traditional banking is the main mechanism for allocating credit. Banks may be all over the financial system, but that mainly is because regulatory changes collapsed the traditional distinctions between the various types of financial firms.* The reality is that non-bank finance ("shadow banking") is a dominant factor in shaping outcomes.

Simple Example

Most bank primers focus on how the amount of deposits is unchanged despite the effect of most customer transactions, other than loans being granted("money creation" of mythology) or paid off.

But as soon as we get to "non-traditional" finance, this becomes murkier.

The easiest example to think of is the act of a bank selling a term deposit to a customer. For example, a customer withdraws $10,000 from a demand deposit account to purchase a $10,000 term deposit (e.g., CD in the US, GIC in Canada).

These term deposits are not included in narrow definitions of money ("M1") by most statistical agencies (?), but might show up in wider monetary aggregates. If we ignore those wide aggregates, the act of converting a demand deposit into a term deposit "destroys money."

Note that the bank is not deeply concerned by this; they are converting one source of funding (a demand deposit) into another (a term deposit). Under normal circumstances, the interest cost is greater, but that is traded off against locking in funding and probably reducing the duration risk of the bank. (Since the duration of assets will typically be longer than liabilities, terming out funding reduces the mismatch.) Banks offer term deposits for a reason.

Although there are obvious legal and operational differences between term deposits and a bond issued by a bank (e.g., eligibility for deposit insurance, put-ability, secondary market, funding cost), from a bank funding perspective, a pool of term deposits resembles a bond issue. This is a hint that bank bond issuance also "destroys money."

Slightly Modified Example

Imagine that a customer of Bank A buys a term deposit issued by Bank B (a situation that I think is more common in the United States).

What happens is that Bank A has an outflow in the payments system, and Bank B an inflow. If nothing else happened that day, Bank A would have to raise funds somehow. The easiest way to do this is to sell a financial asset -- and Bank B would be looking to buy a financial asset. In practice, we do not see such perfect matching, but rather something looking like equilibrium (!) with all buyers/sellers of liquidity matching up orders that clear by the end of the day.

The net result is that a demand deposit is destroyed ("money destroyed"), a "non-money" financial asset shows up on the non-bank aggregate balance sheet, and a financial instrument of some sort has moved from one bank to another.

(Thanks to cultural imperialism of American textbooks, this would be explained as a "reserve transfer." This is the worst possible way to explain what happens, and probably explains why there is considerable bewilderment about bank operations.)

Non-Bank Finance ("Shadow Banking")

We now leap to the dreaded "shadow banking." It is a historical accident that term deposits are not considered non-bank finance, but in practice, banks issue debt instruments that are not deposits.

Imagine that Bank A issues $100 million debenture that is conveniently bought solely by entities that held deposits at banks.
  • For deposits originally at Bank A, we get a destruction of deposits and issuance of a debt instrument that looks like the term deposit example.
  • For deposits elsewhere, there are outflows from other banks (and matching inflows at Bank A). The net result is those deposits are destroyed, and financial assets get re-shuffled from the rest of the banking system to Bank A. (This is the mechanism by which banks build their liquidity position: they steal it from the rest if the system.) This resembles the second example.
The mythology that revolves around bank money creation completely ignores this effect. However, the movement by non-financial firms to hold financial assets that are not deposits forces the banking system to issue securities that are not deposits (bank debentures, money market instruments, securitisations to shrink the balance sheet). Any individual bank will see an ongoing outflow from its customers moving into "shadow bank" assets, and will be forced to issue securities to rebuild liquidity - at the expense of other banks' liquidity positions. As stock-flow consistent models emphasise, the financial assets outstanding have to match up with what is demanded by holders of financial assets (equilibrium again pops up in post-Keynesian theory (!)).

The only way to stop this would be for banks to form a cartel and not issue non-deposit securities. This would force the "law of conservation of deposits" to hold. Nobody would like this, particularly regulators, which is why this does not happen in the real world.

De-Mystifying Deposit Loss

With this background out of the way, we can see why arguments that banks need to worry about deposits going to other banks in response to making loans is missing the point.

  • For the banking system as a whole, "deposit transfers" are zero sum, and since most banks grow loan books near the average rate (otherwise, regulators get cross), the losses that occur due to "transfers" are negligible.
  • There is a massive ongoing outflow of funds towards security holdings, which forces banks to issue debt instruments and securitisations. When housing markets boom in the United States and Canada, residential mortgage-backed security (RMBS) issuance goes through the roof. Any aggregate movement of funds by depositors towards fixed income securities will act as a drain on bank liquidity, which needs to be shored up.
The point is that banks cannot think about their liquidity position solely in terms of new loan origination; they need to react to the portfolio allocation decisions of customers, and shift their funding strategy to match.

Concluding Remarks

Any discussion of real-world banking has to take into account actual conditions, which is where non-banks mainly hold non-deposit fixed income assets. Once we do this, we realise that building a mythology around banks is silly.


* In Canada, this was the "four pillar" system: chartered banks, insurance companies, trust companies, and investment dealers. Now, the megabanks do almost all of these activities under the umbrella of a single holding company (bank "corporations" are in fact a jumble of thousands of corporations, some of which are operating companies). Although there is a wistfulness for the "good old days," there is no way we can go back to that type of system without a total sea change in how we regulate the economy. Currently, entire professions are built around evading the economic intent of the law. Unless that ethos changes, trying to rebuild the pillar structure would fail.

(c) Brian Romanchuk 2019


  1. "For the banking system as a whole, "deposit transfers" are zero sum"

    For the system a whole it is, but for the individual bank it is not. And the financial market is made up of many individuals institutions, and not a single big one. So it really matters to see a single bank as a single bank.

    "There is a massive ongoing outflow of funds towards security holdings, which forces banks to issue debt instruments and securitisations"

    It depends on the country, time period and many other factors.

    I don't think I could understand your point or what is the "money creation mythology".

    But the fact is that banks do need to hold base currency (bank reserves) in their balance sheets to face loan origination liquidity needs. If loan origination liquidity needs outpace the amount of base currency, the bank will face trouble. To cover that liquidity need, the bank will issue deposits or other kinds of liabilities, but just if the market accepts them and if it is willing to pay an adequate interest rate.

    So, I believe that "banks can issue deposits or security liabilities at will" is true but an incomplete (and misleading) story.

    1. But banks get corresponding inflows. An individual bank has to be originating loans faster than competitors for it to get net losses. Meanwhile, all kinds of other factors hit the liquidity position, and that has to be managed.

      The need for “reserves” is very contingent on domestic institutions. Canadian banks are run with a target of zero net deposits at the CB. The only base money they have is vault cash - for the branch and ATM network. Liquidity management is done in securities markets and funding markets.

      If a market cannot place securities in markets, it is pretty much an ex-bank. (Things might be different for the small banks in the United States or credit unions, but they are unusual.)

    2. "Canadian banks are run with a target of zero net deposits at the CB. The only base money they have is vault cash - for the branch and ATM network."

      I don't know how the Canadian financial system works, but I imagine that, if it is indeed the case, Canadian banks sleep holding zero reserves because they invest all excess reserves in the central bank's overnight repo operations (or whatever equivalent thing) to earn some yield.

      A bank cannot survive without excess liquidity in the form of bank reserves, government bonds, repo or other liquid assets. It cannot anticipate with high precision how much deposits will be withdrawn in the next days or weeks. And issuing securities and selling them to the market is a complex operation that takes weeks or months.

      As Warren Mosler puts it, bank reserves, gov bonds and repos are all similar. Bank reserves are like checking accounts, and gov bonds or repos are like savings accounts (they yield interest). However, they are 1) government money in one way or the other and 2) interchangable.

      Banks need necessarily to have some of it before lending. Yes, they don't need to previously have 100% of the amount they plan to lend (except if it is a small bank) but usually they need at least some 40% to 80% depending on their market share. It is a fact that a percentage of customers will withdraw or transfer the money as soon as the lent amount appears in their deposit accounts.

      Banks do need bank reserves to settle operations (including withdrawals, transfers and tax settlements) and I thought that this shouldn't be controversial...

    3. André, what do you mean that a bank cannot survive without excess liquidity? Are you saying a bank that has one million in demand deposit liabilities won't survive unless it holds more than a million in reserves or some equally liquid asset? Isn't it the business model of banking that a bank's assets will Always be less liquid than its liabilities? I'm not following you here. Even the old 'money multiplier model' recognizes that banks create money by lending, if only as a multiple of their required reserves ratios.

    4. "Are you saying a bank that has one million in demand deposit liabilities won't survive unless it holds more than a million in reserves or some equally liquid asset?"

      No, I'm saying that a bank that has one million in demand deposit liabilities won't survive unless it holds more than zero in reserves or some liquid asset.

      Banks are actively very worried about planning, calculating and keeping the level of liquidity they believe is necessary to serve their operations. Nowadays usually it means keeping some KPIs like LCR and NSFR above 100% or a higher threshold.

      Some MMTers downplays the relevance of a bank keeping some level of liquidity, as if the system would always automatically supply liquidity to banks so they don't have to worry about it.

    5. André - yes, they hold liquid securities, which is what I wrote. They also issue banker’s acceptances, which I believe are rediscounted loans, and are money market instruments that can be issued very quickly to manage liquidity.

      But stepping back, the key point to note that liquidity drains due to lending only really matter if a bank is growing faster than its peers. Nobody disputes that. But that does not imply there is a hard limit to bank lending from that source, unlike the stories people are telling. Otherwise, it would be impossible to have high nominal growth rates in an economy.

      People are beating up on a straw man - that there are no constraints on banks - which is not what anyone serious has ever said. This misdirection of the debate is just an attempt to save face by people who resisted the concept of endogenous money.

    6. This four page U.S. Federal Reserve article discusses the Swiss zero reserve policy circa 1989:

      I assume banks settle payment via reserves or exchange settlement funds even though the overnight system might net to zero where some banks are in overdraft and others hold interest bearing positions at the central bank or other settlement mechanism.

      Aggregate bank assets can be simplified to include reserves (exchange settlement funds), securities, and loans. The liquidity cushion is reserves plus securities.

      Aggregate bank liabilities can be simplified to include transaction accounts, other deposits, bonds issued, money market borrowing, and overdraft at the central bank.

      I think Brian is saying that there is no direct linkage between the mix of assets and the mix of liabilities on the books of a bank or aggregate bank sector. If the measure of money is transaction accounts and other deposits then ordinary efforts of bank customers to move funds into money markets or other investments only forces a bank or aggregate bank sector to issue fewer deposits and issue bonds or money market liabilities instead. Banks are good at doing this until there is a money market crisis which Hyman Minsky argues is the trigger for a banking crisis or credit crisis.

      Proponents of MMT argue that the central bank and/or central government must provide liquidity as lender of last resort to support the payment clearing and credit systems. Money and credit are endogenous meaning not under the strict control of the monetary authority which can only control the monetary policy interest rate. In my view a central bank can only raise the short term interest rates sharply, by withdrawing liquidity, and this would reduce a credit fueled inflation by making it difficult or impossible for some financial intermediaries to finance credit growth. A recession might occur as the solution to inflation because reduced credit reduces cash flow in the future and forces units in the economy to default on debt.

    7. During the financial crisis in the U.S. in 2007-2008 the aggregate bank sector did not have the ability to develop liabilities to repurchase bad loans that had been sold into off-balance sheet investment pools. The banks were originating loans, packaging into securities, and selling the loans off the books, to collect fees and boost profits, without net growth of assets and liabilities in the bank sector under originate-to-distribute activities. There were credit enhancements offered by the originators which means the bad loans had to be repurchased by the banks when the credit crisis struck. The money, credit, and bank equity markets did not want to finance new liabilities associated with repurchasing bad loans so it was not possible for the aggregate bank sector to create liabilities and force a flow of reserves or exchange settlement funds to purchase loans in high volumes from the aggregate nonbank sector.

    8. "But stepping back, the key point to note that liquidity drains due to lending only really matter if a bank is growing faster than its peers"

      Faster than its peers? Why?

      For small banks, no matter the rate of growth, they will always need to have enough reserves or liquid assets before lending (because the first thing the customer will do after receiving the deposit is transferring or withdrawing it).

      For big banks, no matter the rate of growth, they will always need to have part of the amount to be lent in reserves (or other liquid assets) before actually lending.

      Those are facts and they contradict part of what MMT claims about banks.

    9. Sometimes proponents of MMT say "Loans create deposits" and this is basically true in the aggregate bank sector. Hyman Minsky said banks develop reserve-economizing liabilities, which sweep bank liabilities away from the reservable categories (transaction accounts), and into categories that do not apply reserve requirements. During the banking crisis the aggregate bank could not rollover these sweep accounts, due to non-bank investors causing a "run" of large time deposits and other uninsured bank liabilities. This would cause a large spike in transaction accounts and a large demand for bank reserves which only the central bank could provide at that time. L. Randall Wray of MMT was a student of Minsky. MMT proponents also understand that the central bank must provide whatever level of reserves are necessary to "hit" the target monetary policy interest rate under a system with or without required reserves.

      One can understand almost everything about banking and central banking and federal deficit policies in the U.S. prior to quantitative easing by reading the these two papers:

      The Money Paradox:

      Interest Rates and Fiscal Sustainability:

      The Money Paradox explains how the central bank injects reserves and transaction accounts (checking deposits) into the aggregate bank sector via the purchase of securities from nonbanks. Then the new transactions accounts migrate to other bank liabilities and/or equity and the excess reserves either would earn zero interest or get interest from the central bank. If the policy is to keep banks from buying short term Treasury bills with reserves then the interest on reserves must be set above the rate paid on Treasury bills. This leaves Treasury bills in the liquidity cushion of nonbanks and excess reserves in the liquidity cushion of the aggregate bank.

    10. André - if a deposit is transferred out to another bank, it is transferred in to another. Banks will receive such inflows on average according to their deposit market share. They will only suffer net losses if they are growing faster than their peers.

      Anyway, this is completely irrelevant for macro analysis. The MMT/PK argument is that reserves are not a macro constraint on anything, and thus there is nothing like a money multiplier. And that is obviously correct. Attempting to tell micro stories about constraints is a pure distraction. Your story about loans triggering deposit losses is only one factor, and is probably insignificant in practice for most banks. The most important constraint that matters for loan growth is the amount of credit-worthy customers who want to borrow. Changing bank capital or liquidity rules is insignificant versus that.

    11. "The MMT/PK argument is that reserves are not a macro constraint on anything, and thus there is nothing like a money multiplier. And that is obviously correct."

      I agree with the claim that the mainstream economic theory (money multiplier and many others) does not correspond to how the real economy actually works.

      However, I believe that there is enough evidence to support the claim that reserves are one of many factors that do have important effects on the (macro) economy. Changing capital or liquidity rules do have expressive impact on the credit market. For me this is so obvious that I have trouble in understanding what do you mean by claiming it doesn't. How it doesn't?

      If a bank has credit-worthy customers but doesn't enjoy adequate access to funding sources (because it is new, small, has lower ratings, or whatever) it will impact its credit decisions and operations. If reserved where unimportant, this would not be true. Liquidity and capital constraints do define the pace of credit origination.

      “André - if a deposit is transferred out to another bank, it is transferred in to another. Banks will receive such inflows on average according to their deposit market share”

      Well, you do criticize the money multiplier but then this sentence is the credit multiplier theory.

      But even if you follow that money multiplier logic, the small bank has 1% of market share, for example. It means it needs 99% liquidity before granting a loan. It has to seek deposits to cover the liquidity need.

    12. It is receiving 1% of the loans granted by other banks, which represent 99% of deposits in the system. If it grants loans at exactly the same pace as other banks, net losses from outflows to other banks are zero. You are attempting to look at one loan in isolation, which makes no sense at either the macro nor micro level. No manager at a solvent bank has ever said “We will grant you this mortgage - wait, we can’t because we don’t have $100,000 in reserves!” Banks have a loan book, and they manage the liquidity of the entire firm based on their total liquidity flow position. For the banking system in aggregate, since banks tend to move as a pack, they can move the aggregate pretty much wherever they want over time. High nominal growth rates happened historically, and they happened in the presence of bank reserves.

      The “money multiplier” story is that deposits are a multiple of reserves, and thus the central bank can steer deposits outstanding via controlling the size of its balance sheets. That is definitely not implied by anything I wrote.

      Meanwhile, in the pre-2008 US system, excess reserves were effectively indistinguishable from zero, and outside the financial crisis, Canadian settlement balances were indistinguishable from zero. There is literally no information to be gained from those time series. How do they represent a macro constraint on anything?

  2. " .............., the act of converting a demand deposit into a term deposit "destroys money.""

    There is an alternate story, a 'mechanical' story, that brings banking actions and transactions into a more understandable (and I think 'accurate') framework. In the mechanical framework, we need to follow the chain of ownership that begins when money is created.

    Before building that framework, we need to admit that a central bank only makes purchases by using 'reserves'. We also need to admit that before the central bank makes a purchase with reserves, it must 'own' those reserves (even if it just now created them).

    Once we accept CB ownership of reserves, it becomes easy to follow successive owners as reserves flow through the financial system. This chain of ownership endures until the CB finally recovers ownership and all other claims of ownership are relinquished. At this point in time, each followed block disappears to whence it came--thin air.

    Nothing about CB creation of reserves precludes private banks from lending these CB created reserves. When private banks lend reserves, they place a second layer of ownership on each 'reserve' dollar. This is risky for the private lending bank because the new reserve dollar owner-for-a-while may transfer it away from the lending bank, leaving the lending bank unable to repay the original depositor (who is also the reserve owner of record).

    Now, back to your article at the point where the quote was found. In my opinion, your article begins to ring true if you point out that it is the destruction of private bank demand deposits that "destroys money". The reserves that underlay deposits are unchanged.

    Turn now to your example of a 'demand deposit-term deposit' exchange. The risk of reserve-transfer changes for the sponsoring bank, justifying a more expensive (than deposits) way of controlling reserves.

    You can see that I differ with you about the importance of reserves as part of bank operations. I hope you can see the importance I attach to a chain of ownership that is attached to reserves so long as they exist. Reserves behave as if they are 'mechanical'.

    1. If we look at countries like Canada, there are no reserves. In this case, your entire story becomes meaningless. We can understand a system with reserves as being a no-reserves system with an archaic bit of financial repression tacked on. American banks act like Canadian banks, they just have an extra hoop to jump through (required reserves).

      Given the widespread confusion about banks and money among people who have been educated with American textbooks that emphasise reserves, it is safe to say that is the wrong way to approach the topic. Banks use a wide variety of instruments to manage liquidity.

      But even if a system uses reserves, there is no 1:1 relationship between reserves and the deposits that are in M1. Deleting a deposit from M1 tells us very little about M0.

    2. You are right about 'reserves' and Canada. We need to call the pool of base money something other than 'reserves' if we are speaking about money in Canada.

      The use of the concept of reserves as being a base money pool in America is complicated by regulation wherein government bonds are considered 'reserves'. Government bonds do not have the same liquidity as money and are therefore not immediately useful as a means of transferring money between banks. As I understand things, this difference in liquidity is overcome with short term markets such as the REPO market, which facilitates better movement of the underlying base money pool.

      The concept of a base money pool becomes important when we logically think of managing the value of money. The size of the pool must be important when compared to less expandable resources such as land or even equities. If the size of the base money pool is expanded willy-nilly, participants in the economy have no stable reference upon which to judge relative values.

  3. To what extent does it take for these methodologies that you put without hesitation take to begin working? Attempting to translate these inquiries will know whether your systems are working. SEO service Pakisan

  4. Thanks for sharing the best information and suggestions, I love your content, and they are very nice and very useful to us. If you are looking for the best bridging loan hmo, then visit Tiger Financial Ltd. I appreciate the work you have put into this.

  5. I admire this article for the well-researched content and excellent wording. I got so involved in this material that I couldn’t stop reading. I am impressed with your work and skill. Thank you so much. Mycard ADP


Note: Posts may be moderated, and there may be a considerable delay before they appear.

Although I welcome people who disagree with me, please be civil.

Please note that my spam comment filter appears to dislike long "anonymous" posts.

Note: if you want to post comments from Apple devices (iPhone, iPad), you apparently need to turn off "prevent cross-site tracking" in Safari privacy settings. (The reason presumably is that another URL handles comments, and so the user session needs to be preserved when redirected to that site. I don't like this, but this is not enough to make me switch my hosting service.)