Home > Uncategorized > Hey Joe, banks can’t lend out reserves

Hey Joe, banks can’t lend out reserves

from Steve Keen

I began another post critical of Joe Stiglitz’s analysis with the caveat that I like Joe. I’ll add to that that I respect his intellect too, both because he’s very bright—you don’t win a Nobel Prize (even in Economics!) without being very bright—and because compared to some other winners, he is very capable of thinking beyond the limitations of the mainstream.

But there are some mainstream concepts that are so deeply embedded in even highly intelligent, flexible thinkers like Joe, that they continue thinking in terms of them, when a bit of really serious thought would show that the concepts are in fact nonsense.

Some of these are so deeply embedded in the psyche of economists that they even permeate the alternate economic universe I work in—known as Post Keynesian Economics (even here I’m a bit of a maverick). I attended a presentation by a non-mainstream colleague recently, and while she was critical of the mainstream, she also seemed to agree with Joe on the same topic: that private banks can lend their excess reserves to the public.

  1. THEY. CAN’T.

Here’s Joe on this subject recently in Project Syndicate, in an article entitled “What’s Holding Back the World Economy?”:

As a result, excess reserves held at the Fed soared, from an average of $200 billion during 2000-2008 to $1.6 trillion during 2009-2015. Financial institutions chose to keep their money with the Fed instead of lending to the real economy, earning nearly $30 billion – completely risk-free – during the last five years.

As I’m about to explain, banks didn’t “choose to keep their money with the Fed instead of lending to the real economy”. They have no choice but to “keep their money with the Fed” and…

THEY. CANNOT. LEND. RESERVES. TO. THE. REAL. ECONOMY. But the fact that economists believe they can matters—really matters—to the real economy right now, because giving private banks excess reserves via QE is the main tool that Central Bankers are using to try to attempt to revive the economy. Since the economy is still pretty sick, 8 years after the global economic crisis of 2008, and because we in the public are in effect their patient, it matters that they give this patient the right medicine.


“Quantitative Easing”: it sounds like a bowel movement. Did you ever imagine you’d even hear such a term, let alone find yourself discussing it over a beer with friends, as many of you probably have? This is the “Magic Bullet” that Central Bankers worldwide are relying upon to restart the global economy. And because in most countries it’s been far less effective than they expected, they’re blaming bankers for not doing their job, and lending the damn stuff to the public.


For me, watching academic economists and Central Bankers (the vast majority of whom trained as economists) tell the banks to “lend your excess reserves to the public, dammit!”, is akin to watching some delusional person in a playground watching two kids playing on a see-saw, and criticising them because they weren’t both up in the air at the same time.

So why can’t banks do what the vast majority of economists believe they can and should do—lend the excess reserves that QE has created to the public? And why don’t economists realise that banks can’t actually lend reserves?

Here’s where I’m going to have to explain something to you that is, on the surface, really, really boring: accounting.

But it’s not, really. Yes OK, the day to day life of an accountant might be less exciting than, say, that of an airline pilot. But just like a pilot, they have some arcane knowledge which makes doing what they do both intellectually challenging, and very, very useful to the public. The useful stuff pilots know is beyond me (but I implicitly and happily rely on it every time I fly); the useful stuff accountants know is double-entry bookkeeping. Why don’t economists know this themselves? Today’s economists simply don’t study it—just like they don’t study history either (“Great Depression? Never heard of it.”). Economists of Joe’s generation often did learn accounting as undergraduates—it was often then required as part of an economics degree—but very few of them ever integrated accounting concepts with their economics.

There are good reasons for that in general: economics can’t be reduced to accounting, just as biology can’t be reduced to organic chemistry (if it could, we’d know how to create life). But just as you can make mistakes in biology if you have a theory that requires chemically impossible reactions to take place, you can make mistakes in economics if your economic theory relies on processes that defy the laws of accounting.

And there is a “Law of Accounting”: that “Assets equal Liabilities plus Capital”. And the belief that banks can lend out their reserves (excess or otherwise) violates the Law of Accounting.

I didn’t learn accounting at University; instead, I’ve learnt it the hard way as I designed an Open Source software package to simulate monetary flows that I named Minsky (in honor of the great non-mainstream economist Hyman Minsky; you can read about and download Minsky for free from here).

Minsky implements this Law, so it lets me show that banks can’t do what Central Bankers (and Joe Stiglitz) think they can do—lend reserves to the public. Most economists, on the other hand, believe in a model of money creation known as “the money multiplier”, which is an intimate part of the concept of “Fractional Reserve Banking”, in which lending reserves to the public is what banks actually do.

In this model, if the Central Bank creates say $1 trillion, and the “Required Reserve Ratio” is 10%, then that $1 trillion of new reserves in banks will create $10 trillion worth of money in the real economy—so that the value of the “money multiplier” is 10. This is the model that Obama’s economic advisors used to convince him that the best way to rescue the economy from the crisis in 2009 was not to give money directly to the public, but to give it to the banks, and for them to then lend to the public:

And although there are a lot of Americans who understandably think that government money would be better spent going directly to families and businesses instead of banks – “where’s our bailout?,” they ask – the truth is that a dollar of capital in a bank can actually result in eight or ten dollars of loans to families and businesses, a multiplier effect that can ultimately lead to a faster pace of economic growth. (“Obama’s Remarks on the Economy”, April 14 2009)

That is poppycock. The actual amount of money that banks can lend to “families and businesses” out of $1 trillion of new reserves is not $10 trillion, but $0. Zip. Nada. Nil. THE. MONEY. MULTIPLIER. IS. ZERO.

Therefore, the $1.4 trillion of excess reserves that QE has created in the USA alone  has added precisely $0 to the lending power of banks.

To understand why, you’re going to have to follow me down the arcane path of double-entry bookkeeping.

Double-entry bookkeeping is a set of conventions that mean that financial transactions are accurately recorded. One part of it, the so called Accounting Equation, is easy enough to follow: it simply says that if you subtract your Liabilities from your Assets, what’s left over is your Capital:

Assets minus Liabilities Equal Capital.

The hard bit to follow is the set of conventions that accountants have developed to make sure that each transaction they record is done so correctly. This involves the use of the terms DR (for Debit) and CR (for Credit), and rules about which term is used which vary depending on whether the account is an Asset, a Liability, or Capital.

I’ve implemented those in Minsky, but because I find them confusing—as many students of accounting do—I’ve developed another convention that is also quirky, but I think easier to follow. It has two rules:

  1. All transactions are shown as a positive entry for the originator, and a negative entry for the recipient; and
  2. Assets are shown as positive sums, while Liabilities and Capital (which I normally call Equity, since the word “Capital” has very different meanings in economics than in accounting) are shown as negative sums. Showing a Liability as a negative makes sense: from the point of view of the entity represented by the table, a liability is a negative. The counter-intuitive bit is showing equity as a negative too, but if you do that then you get the check on your logic that every row must sum to zero.

So a double-entry bookkeeping view of lending $100,000 by a bank, payment of $5,000 in interest, and repayment of $50,000 of the debt, looks like Table 1:

Table 1: A double entry view of lending, interest payments, and debt repayment

Action Assets (+ive) Liabilities  (-ive) Capital  (-ive) Row Sum
Loans Reserves Deposits Bank
Starting position 0 0 0 0 0
Lend money +100000 -100000 0
Pay interest +5000 -5000 0
Repay money -50000 +50000 0
Final position 50000 0 -45000 -5000 0

Note that every row sums to zero—as it should—and creating a debt increases the bank’s assets (and liabilities) while repaying the debt reduces them. Now what about the idea that banks can—and should—lend out the excess reserves that QE has created for them? How does that idea look in a double-entry bookkeeping table? In a word, it looks impossible.

The first stage is the Central Bank makes a loan to the Private Bank—say of $1 million. That is shown in Table 2, and at this point the accounting is accurate. QE itself is both an asset for the banks, and a liability: they get the reserves from the Central Bank, and they are liable to return them to the Central Bank if it asks for them. So that row—“QE from Central Bank”—sums to zero as it should.

Table 2: QE increases both the assets and the liabilities of the Private Banks

Action Assets (+ive) Liabilities  (-ive) Capital  (-ive) Row Sum
Loans Reserves Debt to CB Deposits Bank
Starting position 0 0 0 0 0
QE from Central Bank +$1mn -$1mn 0
Final position +$1mn -$1mn 0

But “lending from reserves?”. If a bank lends its reserves, its assets fall: it has to make a negative entry in its Reserves column. But to show that the money has been lent to the public, you need a negative entry in the Deposits column as well. So the row sum for that operation in Table 3 is not zero, as it should be. The bank simply can’t lend out its reserves.

Table 3: How “lending out excess reserves” looks in a double-entry bookkeeping table

Action Assets (+ive) Liabilities  (-ive) Capital  (-ive) Row Sum
Loans Reserves Debt to CB Deposits Bank
Starting position 0 0 0 0 0
QE from Central Bank +$1mn -$1mn 0
Lend out excess reserves -$1mn -$1mn -2000000
Final position 0 -$1mn -$1mn -2000000

Notice another problem. While the bank has—somehow—given money to its depositors, according to Table 2, it hasn’t recorded that it’s lent them the money. As it stands, as well as bad accounting, this is a free gift from the banks to the public (heaven forbid that that might happen!).

What if the bank tries to fix up this mess by adding an additional row to record the loan? Then you get the mess shown in Table 4. It shows that the impact of a bank attempting to “lend out its Reserves” results in the bank’s assets remaining constant (loans have risen by $1 million while Reserves fall by the same amount) and its liabilities rising by $2 million (the $1 million received from the Central Bank in QE and the $1 million it’s “lent” to the public).

Table 4: Attempting to fix the error simply creates a bigger mess

Action Assets (+ive) Liabilities  (-ive) Capital  (-ive) Row Sum
Loans Reserves Debt to CB Deposits Bank
Starting position 0 0 0 0 0
QE from Central Bank +$1mn -$1mn 0
Lend out excess reserves -$1mn -$1mn -2000000
Record Loan +$1mn -$1mn
Final position +$1mn -$1mn -$1mn -$1mn -2000000

What if the bank lends from its excess reserves liability to the Central Bank, rather than from its assets? That, at least is technically feasible, as shown in Table 5: the Law of Accounting is not violated. But just as with Table 4, so far the Bank has given money to its depositors but not recorded that it’s made loans to them: can’t have that!

Table 5: Lending from reserves as a liability

Action Assets (+ive) Liabilities  (-ive) Capital  (-ive) Row Sum
Loans Reserves Debt to CB Deposits Bank
Starting position 0 0 0 0
QE from Central Bank +$1mn -$1mn 0
Lend out excess reserves +$1mn -$1mn 0
Final position 0 +$1mn 0 -$1mn 0

Table 6 records the loan, by adding a $1 million asset in the Loans column. To balance the row, the bank has to record minus $1 million in the “Debt to CB” column—thereby undoing the effect of “lending from reserves” in the first place.

The final situation has Reserves increasing by $1 million (because of QE) and Loans increasing by $1 million as well (because of the loan to the public), but Reserves have now played no role in the lending: they are back to where they were after QE. Table 6: Bank records the loan and the lending from Reserves disappears

Action Assets (+ive) Liabilities  (-ive) Capital  (-ive) Row Sum
Loans Reserves Debt to CB Deposits Bank
Starting position 0 0 0 0 0
QE from Central Bank +$1mn -$1mn 0
Lend out excess reserves +$1mn -$1mn 0
Record Loan +$1mn -$1mn 0
Final position +$1mn +$1mn -$1mn -$1mn 0

This is, if you’ll pardon the pun, the bottom line: reserves play no role in lending at all, regardless of QE. The “Money Multiplier” model is a myth.

But that myth is the basis of the attempt by Central Banks to rescue the world from the economic crisis. It’s little wonder that that rescue attempt isn’t going as smoothly as planned.

  1. Michael Kowalik
    February 18, 2016 at 12:24 pm

    So, either the FED is incompetent, an no one ever goes to jail for being genuinely incompetent, or the FED only pretends to be incompetent and the true objective of QE is to help the banks to 100% (free) reserve banking at the expense of the real economy.

  2. Frank Peters
    February 18, 2016 at 12:48 pm

    I like Steve K. But this is just silly.

    At the same time as arguing banks CAN’T lend funds they’ve been credited by the central bank, he also argues elsewhere that banks neither need reserves nor deposits in order to lend – they can simply create money.

    Bank A can choose to make any loan it likes. At the individual transaction level the ‘money creation’ notion is correct. But banks are, or ought to be constrained by risk metrics such as reserve assets requirement, capital ratios AND loan-to-deposit ratios. So Bank A’s aggregate lending is, or should be, limited by these constraints.

    If there is a run on the banks’ deposits (either by private sector panic, or, as under the ECB, orchestrated to bring Governments to heel) then the bank is obliged to stop new lending, or call in existing loans. According to SK, deposits are immaterial. Just make another loan, create another deposit.

    In reality, the level of deposits is an absolute constraint on the total amount of lending. A prudent loan-to-deposit ratio is almost certainly 1 or <1. No, banks don't lend their reserves (which is poorly worded). But the level of reserves and the level of deposits are a limit on total loans.

    • February 18, 2016 at 5:56 pm

      I agree with this critique of endogenous money theory, with just one wrinkle. It lies in the accommodationist element of endogenous theory (Paley etc) which observes that central banks not only lend reserves but carry out asset purchase in normal times (Fed I think about $30bn pa ‘accommodation’, BoE more like only £1bn, which they call ‘stock of assistance’). This small-scale assistance leaves private banks with increased reserves and equally increased liabilities to the private sector. However, all of this is merely to cater for stochasticities in clearing, and has little impact on lending levels. There the absolute constraint is the capital ratio as you say, and which, as I point out below, gets considerably worse with the influx of large reserves/liabilities under QE.

    • February 18, 2016 at 8:13 pm

      Banks don’t have a loan to deposit ratio. They have a solvency constraint that assets >= liabilities, which means loans>= deposits. In the trivial sense that someone’s loans are someone else’s deposits these are in fact equal in the system.

      Banks may have a metric of how much of their balance sheet originated as deposits (someone came and deposited money) as opposed to how much originated as loans (someone took a mortgage). But in accounting terms they’re always balanced. If you deposit your money in a bank they buy bonds to invest it. Why? Because if they lent your money to someone else the (loan-deposit) so created would be neutral on their balance sheet and they’d still be stuck with your money.

      Banks have a % capital and % reserve requirement. These two are sometimes called liquidity constraints. These are regulatory, not accounting requirements. For example they must hold capital >=3% of loans and reserves >=10% of loans. Or thereabouts. Only the capital one matters as I explain in a later post. But they don’t lend out the capital or the reserves, they’re just required to hold it as collateral.

      • Michael Kowalik
        February 18, 2016 at 9:20 pm

        Deposits are not an asset but a liability. They cannot be lent out. This is clear in Steves accounting. Also, loans create deposits.

      • Frank Peters
        February 19, 2016 at 10:20 am

        Banks DO have loan-to-deposit ratios. It is a key risk metric used by the ECB and the Bundesbank

        https://www.ecb.europa.eu/pub/pdf/other/bankingstructuresreport201410.en.pdf (see commentary around chart 17).

        And, no, loans don’t equal deposits as then the ratio would always be 1 and meaningless. The chart shows the ratio falling from 125% in 2008 to 112% in 2014.

        The Fed only uses LTD to regulate inter-state bank takeovers, ie banking at a federal level. But the LTD is the metric used, and is central to primary legislation, the Riegle-Neal Act 1994.


        As assets are not solely comprised of loans then the LTD ratio cannot be the sole risk metric. But as loans usually comprise the bulk of all bank assets and deposits comprise the bulk of all liabilities LTD is, or ought to be, extremely important.

      • Michael Kowalik
        February 19, 2016 at 11:53 am

        @ Frank Peters. The excess of loans over deposits in the Euro-area may be accounted for by Euro denominated deposit accounts held in offshore and non-EU jurisdictions.

      • February 19, 2016 at 4:54 pm

        “If you deposit your money in a bank they buy bonds to invest it.”

        This language and way of thinking perpetuates the basic error.

        A bank deposit is a LIABILITY of the bank to the depositor. The bank cannot spend it to buy anything.

        If you deposit your “money” in a bank, it is used to offset a deposit leaving the bank so that the clearing with other banks is zero. That is how an incoming deposit is a positive for the bank.

        The bank can buy a bond simply by creating a new liability. BOND (asset) = Bank credit liability. It does need to get an equivalent liability (deposit) incoming to avoid a net liability that would force it to sell the bond.

        So, in that sense the statement quoted above is true but the way of thinking about it is misleading.

      • February 20, 2016 at 7:22 pm

        Here’s the difference between these two actions. If you “deposit” money into a bank and then the bank “buys bonds to invest it” it looks like this:

        +clearing balance, -deposit
        -clearing balance, +bonds

        End result: +bonds, -deposit

        If the bank went out and bought bonds just by itself:

        End result: +bonds, -clearing balance

        And all I can say is these are not identical. I don’t have the knowledge what prevents banks from running up infinite debts in the clearing system. If it’s rates, policy rules, etc.

      • February 20, 2016 at 7:37 pm

        @Frank Peters, yes sorry, I was oversimplifying and said something inaccurate. Thank you for providing detailed references.

        Of course banks can have things other than loans or deposits on their balance sheets, and as these get transferred between banks a specific bank ends up with a LTD ratio other than one. Indeed that’s a useful regulatory statistic.

        But as a first approximation, loans and deposits get created symmetrically as (+loan, -deposit) and should sum to zero in the whole system. I’m sure it gets more complicated as loans pay interest or get written off, so it’s not 1:1. I think we’re actually agreeing in basic terms.

  3. February 18, 2016 at 12:49 pm

    To all «non-marxist» mavericks out there that may be unable to overcome deep seated prejudices about reading Karl Marx, at least try to read John Stuart Mill… Particularly starting at http://www.econlib.org/library/Mill/mlP59.html#Bk.IV,Ch.IV

  4. Ricardo
    February 18, 2016 at 1:36 pm

    I support your view and think that economists desperately ought to have some basic accounting in their minds, specially when dealing monetary and central banking issues. Although, I am not sure why you say that bank’s reserves on the central bank are a liability to CB. Those reserves are a liability in the central bank balance sheet, and their asset entry in the QE case is all the toxic assets or bonds that they bought to the banks and private investors. Meanwhile, in the banks balance sheet, they sold those assets and increased their reserves in the central bank, which is other asset entry. The double accounting principle works between assets and liabilities, but also betwwen two kinds of assets or two kinds of liabilities. So what QE did is just to rearrange the assets compositoon in the balance sheet of banks, because they needed liquidity not toxic illiquid assets, and this does not increase they balance sheets. On the other hand, central banks did increase their balance sheets because they create those reserves out of nothing, just like banks creates deposits out of nothing when they give out loans.
    Would you agree with this?
    Another point I wanted to clear out is the lending out of reserves view. Banks can’t lend directly their reserves, but, having more reserves allow them to improve their liquidity ratios and other ratios alike, thus letting them continue sith their business as usual, create deposits giving loans. So in a way, more reserves does open the possibility to more lending.

  5. February 18, 2016 at 1:47 pm

    Keen makes an error in this passage from his post:

    “The first stage is the Central Bank makes a loan to the Private Bank—say of $1 million. That is shown in Table 2, and at this point the accounting is accurate. QE itself is both an asset for the banks, and a liability: they get the reserves from the Central Bank, and they are liable to return them to the Central Bank if it asks for them. So that row—“QE from Central Bank”—sums to zero as it should.”

    The mistake here lies in a confusion between QE, when the central bank purchases securities from the Non-Bank Financial Intermediaries (NBFIs in BoE parlance), and reserves averaging, when the central banks lends reserves to private banks (termed Monetary Financial Intermediaries or MFIs in BoE parlance).

    EITHER an MFI borrows reserves in which case, as Keen says, it has a liability to the central bank, OR the reserves of the MFI increase because the central bank credits them in order to purchase assets from the NBFI which banks with the MFI, in which case the private banks have a corresponding liability to their depositors. In the next stage of the QE process the NBFI deposits are transferred to the private holders of the assets sold to the central bank. Hence Keen is right that reserve increases through QE involve increased assets of the private banks matched by equal liabilities, but they are liabilities to the private sector not the central bank.

    The actual question Keen investigates should anyway be a non-question in the tradition of accommodationist endogenous money (as Frank Peters says in his comment above), because private banks do not lend ‘out of’ anything. They simply create a loan which makes for an equal asset and liability for the bank.

    Keen is right though that the massively increased reserves of the private banks after QE do not encourage them to increase lending one bit. That is because the increased reserves as assets, matched by increased liabilities to depositors, mean that the liabilities column looks weaker from this point of view: the capital ratio has got worse. It is the capital ratio that dictates banks willingness to lend, not reserves (which for many decades has been only a question of dealing with clearing stochasticity). Of course willingness to lend on the part of the banks has to be matched by willingness to borrow, which, as is well known, falls of a cliff in a recession.

  6. February 18, 2016 at 2:18 pm

    I read this article by Joe Stiglitz, a person I’ve read and traveled to hear speak.

    My take on the econ brain wash trap that has caught Joe is much more brief. It does not matter whether he is right or wrong about banking details (same total failure by Yanis Varufakis as a finance minister).

    Borowing money to grow faster at a time when the world economy is operating at 1.5 Earths is stark raving mad.

  7. February 18, 2016 at 7:04 pm

    Steve, your explanation is rock solid but the misconception in the public’s mind may be somewhere else. Let’s distinguish three theories:

    1. The fractional reserve theory you read in bad textbooks or Wikipedia where the bank magically gets an initial lump of money and lends 90% to A, the money gets deposited back, the bank lends 81% to B, 73% to C etc. is utter rubbish. This is simply not the mechanism at work. BTW why are teaching materials still wrong at this basic “Earth is flat” level? Banks like to be obscure about their workings but it’s not a secret, you can find bankers like the BoE who will tell you how it actually works.

    2. The credible, not complete rubbish version of the money multiplier theory is that banks create loans and deposits as they please, but every night they have to meet a % reserve requirement and a % capital requirement so they monitor these amounts and constrain their lending. For example if the reserve requirement is 10% and the bank has $1bn reserves, the bank can only hold $10bn loans on its balance sheet. It can’t make any more loans than that until it manages to get hold of more reserves.

    Everything about this theory is true except there’s no limit on the available reserves. When the bank is short on reserves they just ask the Fed, and the Fed always says yes. So the reserve requirement is not in practice a limit. The Fed rate is a disincentive, and that’s all.

    3. So the way it really works is banks find a willing borrower, they create double entry +$1m loan and -$1m deposit as per your explanation and then they measure their balance sheet and find they need more reserves. So they make another entry +$100k reserves -$100k debt to CB, and they’re good for a 10% reserve requirement. The Fed doesn’t say no. If they needed capital they’d have to raise it in a market and that would be a limit on their balance sheet, but capital requirements are still small.

    What can the Fed do to make it easier for banks to lend? Lower the reserve rate, which has been zero for years and thus not a constraint, or lower capital requirements that are already arguably too low. Banks are not in any way constrained from lending by absolute limits, as was your main point.

    What does QE actually do? It bids up bond prices reducing their yield. Private money flows out of bonds into stocks and therefore bids up stocks. Speculators take margin loans to buy stocks and amplify the effect. Big companies are happier when their share price goes up than when it goes down, and that’s going to make life somewhat more pleasant for their people. But mostly at the top. Compared to inflationary spending QE is a huge waste.

  8. Political Economist
    February 18, 2016 at 9:18 pm

    I think it might clarify things a little if you just noted that when a loan is made assets increase by the amount of the loan and no reserves are lost (“lent out”). So, instead of a decrease in assets though lost reserves, there is an increase, so the line zeros out.

  9. John Hermann
    February 19, 2016 at 1:48 am

    Steve Keen’s take on this is absolutely correct. What really needs to be emphasised is that we have a dual monetary system. The creditary deposits in commercial banks and banking reserves (exchange settlement funds) tag along with each other with every transaction, but – like oil and water – never actually mix. The demarcation line is never crossed. Once this reality is recognised, everything else falls into place.

    • February 19, 2016 at 3:26 pm

      Well put. The question that illustrates the point is when bank A transfers money to bank B on behalf of a customer what actually happens.

  10. Paul Schächterle
    February 19, 2016 at 2:51 am

    Steve Keen is somewhat right in that QE is not helping against the real problem of the crisis. However to say that reserves can’t be lent is wrong or at least mistakable.

    The devil is in the detail and therefore before using a rather theoretical model of payment flows we should examine how the relationships are structured in real life.

    What is “lending” as it is done by banks?

    In the first place a loan is a contract between a bank (B1) and a customer (C1). The bank promises to pay a sum of money instantly and the customer promises to repay that sum of money later, i.e. after the duration of the loan, and to additionally pay a fee (the interest). The terms of such a contract may vary.

    Can a bank MAKE such a promise out of thin air? Yes.
    In other words: Are reserves necessary to make such a promise? No.


    Can the bank FULFILL such a promise out of thin air? That depends, but basically no.
    In other words: Does the bank need reserves to fulfill such a promise? That depends, but basically yes.

    If the customer asks for cash, the bank has to deliver cash. That requires reserves.

    If the customer transfers the money from his account at B1 to another customer (C2) who has a checking account at different bank (B2), then the bank B1 has to make a payment to B2. That involves reserves. Alternatively B1 would have to get itself a credit from B2, which B2 may not be willing to give.

    Only if the customer transfers the money to a recipient who has a checking account at the same bank (B1), then the bank B1 can fulfill that without the involvement of reserves.

    Of course there may be lots of transfers back and forth that can be cleared, but in the end the bank may need reserves to fulfill a loan contract.

    But what about the balance sheet?

    In short a bank’s balance sheet is primarily a way to check the SOLVENCY of the bank, not its LIQUIDITY.

    The bank is solvent if its assets are greater than its liabilities. Both assets and liabilities can consist of totally different types of economic values. Particularly the time factor of such values may be different.

    If a bank lends out money (reserves) and gets in return the customers promise to repay with interest then that promise may be worth more than the lost reserves. But the time in which the bank can demand repayment is in the future.

    To fulfill any present obligations to pay, i.e. to remain liquid, the bank needs reserves.

    • Michael Kowalik
      February 19, 2016 at 4:43 am

      Some good points here, but it begs the question what he “loan contract” is really about. We ought to remember that in order to have a valid and enforceable legal contract there must be a transfer of “valuable consideration”: something valuable must be transferred from the lender to the borrower. Somewhere in the ‘definitions’ section of any loan contract made by the bank there is always a clause like this: “consideration may include credit”. Since a loan contract is actually ALL credit, the above clause may be translated as “we are lending you a loan”. This kind of absurdity results from the fact that banks don’t have the money they “lend” and so even calling such a “contract” a “loan” is a spectacular misnomer. Unless you would cash-out the whole amount of the loan, the valuable consideration is not actually transferred and so the contract is, in principle, void. But banks do get special treatment in that regard from the courts.

      But even if you wanted to prove the point and cash out the whole loan amount (not EFT but physical cash) there would still be millions of customers who could not possibly do the same, simply bacause the bank maintains only about 5% of its current liabilities in reserves. The valuable consideration stipulated in all the loan contracts is never transferred nor can it possibly be transferred because it does not exist beyond the 5% reserves.

      So where does your new buying power comes from if not from the banks? It comes from the real economy. It is effectively stolen by the banks from everyone else by inflating the broad money supply. We also ought to remember that this monetary inflation will not fully manifest as price inflation because it is offset by the price-deflationary effect of economic growth, that is, by those who create real value.

  11. John Hermann
    February 19, 2016 at 3:30 am

    “… If a bank lends out money (reserves) … “. Oops Paul. Did you really mean to say that?
    Banks lend out bank credit money to their retail borrowers, not reserves. Creditary deposits within commercial banks have come to be regarded, used and accounted as a form of money. It is true that banks effectively lend reserves to other banks, but other banks are not retail borrowers and in any case the reserves remain entirely within the banking system. The line is never crossed into the real economy.

    • Paul Schächterle
      February 19, 2016 at 5:44 am

      Yes, that was mistakable on my part. The bank makes a promise to pay money. If the customer demands it, then the bank has to deliver money (i.e. reserves). Normally a customer just make transfers from one checking account to another. So the bank does not need to turn out reserves in cash but only needs to be able to make payments to other banks or get credits in order to effectuate the transfer.

      My point is: the reserves have a certain role in credit contracts. The are the basis of the contractual duties. Banks need to manage their reserves to maintain their liquidity.

      And that is a task somewhat different from maintaining solvency. Balance sheets play a role in determining solvency, not liquidity.

      Of course there are various interdependencies between solvency and liquidity.

      The credit crunch was/is not only a problem of liquidity but also of solvency because some banks had/have bad loans in their balance sheet. Bad loans may need to be revaluated which may lead to a decrease in assets which may lead to insolvency, not only illiquidity.

      Fear of insolvency again may lead to illiquidity. Banks may refuse to give credit to other banks because they fear that those banks are insolvent and will not be able to pay the loan back.

      Also the burst of an asset price speculation bubble may lead to insolvency instead of just illiquidity because a) the borrower may get insolvent and b) the loan may no longer be adequately secured.

      On the other hand, reducing the interest payments for refinancing loans may reduce the liabilities of banks and prevent their insolvency. Hiking up the prices of speculative “assets” like stock shares may technically increase the value of their assets and therefore prevent insolvency.

      IHMO one big problem is that the policies to save the zombie banks i.e. the banks that made bad speculative deals and therefore are at risk to be insolvent if the assets are revaluated, are quite likely harmful for the rest of society. Also saving the banks may not be the right goal. Why not nationalize the banks (including the FED) and allow for a revaluation of assets?

      Why use the brute force of the national bank to keep asset prices elevated to speculation bubble level? Why not use the force of the national bank to promote infrastructure projects, health care, education, things that actually improve productivity and the living conditions of the masses?

      The bankers and the rich invested in speculative assets own the politicians, though, and prevent such policy changes. They also prevent regulation to avoid risky asset price speculation with public money and small savings because they profit from the central bank intervention. Asset prices and therefore rents and executive bonuses stay high. Wages and small savers’ interest payments stay low. Maybe that is exactly what some people prefer.

  12. John Hermann
    February 19, 2016 at 7:11 am

    ” The bank makes a promise to pay money. If the customer demands it, then the bank has to deliver money (i.e. reserves). ”

    Actually, this statement is technically incorrect.

    If a bank customer demands to be paid currency (coins or notes), what happens is that the customer’s creditary deposit is reduced accordingly, which means that the money supply (meaning, money accessible to and used by the general public, as measured by M1) remains unchanged. And at the same time, the bank’s currency reserves are depleted by the same amount. It is also quite likely that a little further down the track the bank will redress that depletion by exchanging with the central bank some of its exchange settlement funds for currency. And if the aggregate of such transactions leads to a sufficient depletion of creditary reserves across the entire banking system (as indicated by inflationary pressures) then the central bank will step in and inject new creditary reserves into the banking system via its open market operations.

    It is also important to recognise that – in the exchange of currency between the bank and the customer – the status of that currency was transformed at the instant that transaction occurred. Namely, it transformed from being part of the bank’s broader reserves into being part of the money supply. And the reverse process obviously occurs whenever a customer makes a deposit using currency. However this type of transaction does not represent what I have described as “crossing the line”, which could only occur if the money supply had changed in the process of withdrawing or depositing currency.

  13. February 19, 2016 at 3:44 pm

    Economists’ perennial trouble with accounting
    Comment on Steve Keen on ‘Hey Joe, banks can’t lend out reserves’

    You say: “… the useful stuff accountants know is double-entry bookkeeping. Why don’t economists know this themselves? Today’s economists simply don’t study it …. Economists of Joe’s generation often did learn accounting as undergraduates … but very few of them ever integrated accounting concepts with their economics.”

    Accounting is elementary mathematics and, true, it is regularly beyond the capacities of economists (2012). Unfortunately also of heterodox economists. This includes Steve Keen.

    The matrix is not the best tool to present the accounting interrelationships so I present my refutation of Keen’s argument in an alternative format here

    1 In the beginning there is only the central bank which creates overdrafts and deposits uno actu out of nothing. Overdrafts stand here for all forms of direct loans to the household or the business sector. The deposits of the central bank are money and used for transactions between the household and the business sector. Other forms of money are here kept out of the picture.

    2 The banking sector is now split between the central bank and commercial banks. The central bank creates 10 monetary units (million, billion, trillion, Euro, Dollar, Yuan) of overdrafts and deposits for the commercial banks only.

    3 The commercial banks start their lending business and create 100 monetary units overdrafts and deposits for the business sector. The deposits of the commercial banks are the transaction money used by the business sector to pay wages and by the household sector to buy consumption goods. The ratio of central bank deposits (= reserves) to business overdrafts is here 10 %, i.e. 10/100 units) and it is assumed that this is the maximal ratio. So, the commercial banks have here reached their limit of money creation. It is the central bank’s turn to act.

    4 In the course of quantitative easing, the central bank takes over 5 monetary units of business sector overdrafts (= loans) from the commercial banks. The ratio of central bank deposits to business overdrafts is now 15.8 %, i.e. 15/95. So the commercial banks have excess reserves. With regard to the 10 % limit they need 9.5 units of central bank deposits but have 15.

    5 The commercial banks now again take up their lending business and increase overdrafts to business by 55 units. Of course, the same increase takes place on the debit side (= business deposits +55). The ratio of central bank deposits to business overdrafts is now again 10 %, i.e. 15/150).

    In the strict sense it is misleading to say that commercial banks lend out reserves. In a pure credit economy the commercial banks create overdrafts and deposits uno actu out of nothing. The reserve ratio is not a practical but a legal limit.

    So, literally, is is right to say that commercial banks do not lend out reserves. But it is obvious that between step 4 and step 5 the banks have excess reserves and therefore are in the position to create money in the form of bank deposits for the business and the household sector. Between step 4 and 5 the credit multiplier is indeed greater than 0. Steve Keen’s conclusion “Therefore, the $1.4 trillion of excess reserves that QE has created in the USA alone has added precisely $0 to the lending power of banks.” is false.

    The lending power is there but of no use if the household and business sector prefer to deleverage (Koo, 2009).

    The real problem of QE is that the central bank takes toxic loans off the commercial/investment banks balance sheets and thus protects them from losses.

    Egmont Kakarot-Handtke

    Kakarot-Handtke, E. (2012). The Common Error of Common Sense: An Essential Rectification of the Accounting Approach. SSRN Working Paper Series, 2124415:
    1–23. URL http://ssrn.com/abstract=2124415.
    Koo, R. C. (2009). The Holy Grail of Macroeconomics. Lessons from Japan’s Great
    Recession. Singapore: Wiley.

    • February 20, 2016 at 3:47 pm

      Canada hasn’t had a legal reserve ratio since 1991. The UK also has no legal reserve ratio. Reportedly banks keep about 3% of their liabilities as vault cash by their own judgement. So the entire concept of the central bank limiting lending with a reserve ratio is not applicable, and in fact counterproductive to understanding that BORROWERS are the source of money. Period.

    • February 21, 2016 at 8:45 pm

      That would be true except in step 3 commercial banks just ask the central bank to borrow reserves as they need to meet the 10% requirement and the central bank says yes. Then they can move directly to step 4. Holding reserves is a liquidity requirement, not a limit on balance sheet size. I guess it could be used as a limit if the central bank refused to lend more reserves, but it does not.

      Quantitative easing is not when the central bank agrees to lend banks more reserves. It’s when the central bank creates reserves and uses them to buy assets (bonds, stocks, sometimes deliberately toxic securities) directly from the market. In so doing, reserves are credited to whichever financial institution previously held the assets where, as Steve points out, they sit paying interest to the bank and otherwise doing nothing for the economy.

      QE may help the economy somewhat, although not that much and mostly for rich people, by buying bonds and stocks. The part about reserves sitting in banks just helps the banks.

  14. John Hermann
    February 20, 2016 at 10:52 am

    Steve Keen said: “…. ‘the truth is that a dollar of capital in a bank can actually result in eight or ten dollars of loans to families and businesses, a multiplier effect that can ultimately lead to a faster pace of economic growth’ (“Obama’s Remarks on the Economy”, April 14 2009). That is poppycock. The actual amount of money that banks can lend to ‘families and businesses’ out of $1 trillion of new reserves is not $10 trillion, but $0. ”

    Steve seems to have become confused here between capital and reserves. Modern bank lending is constrained by bank capital, not bank reserves. And that capital is largely composed of financial assets (e.g. Treasury securities) which are unmatched by financial liabilities. Therefore Obama’s remarks are correct, and are not poppycock. Note also that Obama used the word “can”, leaving open the possibility that an increase in the capital multiplier (the inverse of the capital ratio) might actually be ineffective in stimulating economic activity — in situations where households and firms are engaged in deleveraging their existing debt or for any reason are disinclined to take on more debt.

  15. February 20, 2016 at 3:56 pm

    Not knowing much about accounting, i could not follow the explanation. BUT there is something strange about it. accounting identities are representations of some real world phenomena. It appears as if some accounting conventions are preventing banks from lending reserves given to them by the central bank. So the problem could be solved by changing accounting system. For example, what would happen if we stopped doing double entry bookkeeping and went to a single entry system instead? After all double entry dates back to a pre computer era where possibilty of bookkeeping mistakes was greater So if we change accounting conventions, would that enable banks to lend reserves? ALTERNATIVELY there are REAL explanations — like there are laws which prevent banks from taking money out of central bank and lending. OR in a heavily indebted economy agents are not interested in acquiring ore debt, OR SOMETHING That is a real explanation takes a different form from and accounting explanations. It is possible that steve is right annd the accounting identities represent some hard reality about real world finances, but the explanation is missing this half. I dont understand it

    • February 20, 2016 at 4:36 pm

      DEBT to a commercial bank equals money. BORROWERS create money by promising to pay it back to a bank. Money is a promise of someone’s future productivity. To think that the central bank can create money is nonsense. Central banks create central bank reserves, credit for cash when cash is called for by the depositors. Cash is a temporary physical form of COMMERCIAL BANK CREDIT.

      The only restraint on bank lending is the RISK of non-payment by the borrowers.

  16. John Hermann
    February 21, 2016 at 12:19 am

    There are several aspects to this issue Asad. As pointed out in an earlier comment, there is a fundamental reason why banks cannot transfer reserves to members of the public, and that is because we have a dual monetary system. Members of the public are not permitted to have deposits with the central bank, and it is this privilege which ultimately defines what a banking reserve is. Putting it another way, commercial banks are special entities in that they do not form a part of the real economy.

  17. February 21, 2016 at 1:01 pm

    Asad Zaman, John Hermann

    The two criteria of science are formal and material consistency (Klant, 1994, p. 31). The latter is established by empirical testing. How important this is for genuine scientists one may glean from the fact that physicists have built ‘the world’s most expensive and complex experimental facilities to date’ (CERN, Wikipedia*) in order to test a hypothesis that has been put forth in 1964 by six theoretical physicists.

    This vividly contrasts to the silly methodological motto of most economists, i.e. “it is better to be roughly right than precisely wrong!” (Davidson, 1984, p. 574)

    The analogon to the physicists’ fervor of measurement would be to install a giant facility which records every economic transaction in real time according to the principles of accounting. This facility then delivers the exact numbers (two digits) of total income per period, consumption expenditures, saving, and so on. And these numbers are the rock-solid foundation of empirical testing.

    Curiously, economics have never shown any ambition to build such a facility. Worse, economic theory is not even built upon concepts that correspond with what could be actually produced with such a gigantic bookkeeping machine. Just the contrary, economic theory has been built upon concepts like utility or equilibrium and it should have been evident from the very start that there is no testable correspondence to these green cheese concepts in the real world. Thus the scientific failure of economics has been methodologically pre-programed 140 years ago.

    What most economists have not realized to this day is that accounting is pivotal to their discipline. Their manifest incompetence consists in not understanding the elementary mathematics that underlies accounting (2012). This is the real mathiness problem.

    “Somewhere between the Political Arithmetician, alias the National Income Accountant, and the Financial Analyst, alias the Accountant, lies the task of the quantitative economist’s analytical role, and none of the theoretical or applied tasks of these two pragmatic and paradigmatic figures requires anything more than arithmetic, statistics and the rules of compound interest. These, in turn, require nothing more than an understanding of the conditions under which systems of equations can and cannot be solved. But what kind of quantities do these equations encapsulate as parameters, constants and variables? Surely, the kind of quantities that enter the equations of the Political Arithmetician and the Accountant cannot be other than rational or natural numbers — negative and non-negative? Eminent theorists, working in core areas of economic theory — price theory and monetary theory — have made this point in interesting ways over the past half a century.” (Velupillai, 2005, pp. 866-867)

    To be sure, accounting is not all of economics. But make no mistake, above the entrance to economics as science is inscribed the phrase: “Let None But Those Who Mastered The Elementary Mathematics of Accounting Enter Here.”

    Egmont Kakarot-Handtke

    Davidson, P. (1984). Reviving Keynes’s Revolution. Journal of Post Keynesian Economics, 6(4): 561–575. URL http://www.jstor.org/stable/4537848.
    Kakarot-Handtke, E. (2012). The Common Error of Common Sense: An Essential Rectification of the Accounting Approach. SSRN Working Paper Series, 2124415: 1–23. URL http://ssrn.com/abstract=2124415.
    Klant, J. J. (1994). The Nature of Economic Thought. Aldershot, Brookfield, VT: Edward Elgar.
    Velupillai, K. (2005). The Unreasonable Ineffectiveness of Mathematics in Economics.
    Cambridge Journal of Economics, 29: 849–872.

    * https://en.wikipedia.org/wiki/Higgs_boson

  18. February 24, 2016 at 6:33 pm

    The same discussion in German, translates OK. Points to note: the reserve requirement is very low. Zero in UK or Canada, 1% in the Eurozone according to Flassbeck.


  19. James Smarson
    March 6, 2017 at 6:40 pm

    I think people are misconstruing that only the Central Bank can reduce excess reserves for meaning banks can’t lend out excess reserves. Let’s do a simple theoretical example:

    Bank A wants to loan Steven Keen to buy a house
    Loan (Debited) 100K
    Reserves (Credited) 100K

    Bank A could easily sell its reserves for cash to another bank. Or it could repo its reserves for cash. Or it could got cash in exchange for reserves from the CB.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.