That David Graeber was not happy with my previous blog did not surprise me. Yet, his reply gave me cause to pause and to reflect. Had I missed something? Had I misunderstood the point he made in his Guardian comment? Here is again Graeber’s response to my blog:
“I don’t see why anyone should take you seriously as an ecomomist (sic) if you haven’t even figured out the point of the essay was not a “discovery” of fractional reserve banking but exactly the opposite: that the fractional reserve / money multiplier explanation of money creation is really a popular myth. Seems to me that anyone who read the (briefly and clearly written) essay and didn’t even get that far has no business running an economics blog.”
Graeber believes that the Bank of England paper “Money Creation in the Modern Economy” refutes standard explanations of how fractional-reserve banking (FRB) works in a fiat money economy. These standard explanations are myths and this has meaningful consequences. Graeber seems to follow Steve Keen who drew similar conclusions here. Both, Keen and Graeber equate standard explanations of FRB with what they call the “money multiplier theory”. The Bank of England paper explains why the “money multiplier theory” does not offer an accurate description of how the Bank of England directs the money creation process today. What the Bank, Graeber and Keen understand the “money multiplier theory” to be, I come to in a minute. The Bank of England does not state that this refutes FRB explanations in general, as Graeber implies in the quote above. In fact, the Bank of England paper is full of fractional-reserve banking processes, and there is an obvious reason: this is still a fractional-reserve banking system.
The “money multiplier theory”, at least in the rigid form in which the Bank presents it, is not essential to FRB and not a necessary component of FRB explanations. I wrote extensively about FRB and I do not think I ever even used the term. Furthermore, the supposedly explosive conclusions that Graeber (and to some degree Keen) draws specifically from the Bank dismissing the “money multiplier theory”, such as that money is not a scarce resource, that money creation does not require an act saving first, that the banks do create deposits simply via the act of lending, these conclusions are in fact true for any system of FRB under fully elastic fiat money and central banking. Just by combining FRB with central banking and unconstrained fiat money you invariably get these consequences. You would even get them if the “money multiplier theory” were correct. For example, if the “money multiplier theory” were a correct description of modern money creation, it would still not require an initial act of saving for banks to increase lending but simply an act of reserve-creation by the central bank, money would still just be a piece of paper or an electronic book entry and just not inherently scarce, and the supply of money would still be entirely elastic. No-one who has grasped how FRB works under conditions of limitless fiat money from a central bank can be particularly shocked by the Bank distancing itself from the “money multiplier theory”, by anything else the Bank ‘reveals’ in its new paper, nor by Graeber’s list of “explosive” discoveries. My initial suspicion that Graeber must only now have fully grasped FRB under fiat money conditions is thus entirely understandable.
After re-reading the Bank of England paper and Graeber’s Guardian column I maintain the following: 1) The “money multiplier theory” is a detail, not a game changer. 2) There is nothing fundamentally new or surprising in the Bank of England paper (there are, however, some inaccuracies and outright mistakes in it). 3) Graeber’s list of amazing discoveries is broadly true but not surprising. Money is, of course, never scarce in a fiat money economy. 4) Graeber’s conclusions — why austerity?, what does it mean for the average mortgage borrower if he finds out how money is created? — are silly and confused, and they do not follow from 3). Keen does not appear to share these latter conclusions but he, too, is wrong, in my assessment, as to the importance of what the Bank of England has “revealed”.
What Is Meant by “Money Multiplier Theory” and How Does It Relate to FRB
According to the Bank of England paper, there is a theory out there by which many textbooks explain the money creation process in an FRB system. The components of this theory seem to be the following, according to the Bank of England:
1) In an FRB system the ability of banks to create more deposits through more lending is ultimately restricted by the available quantity of reserves. Banks can usually create more deposits by extending new loans (sure they can, this is what FRB means) but with more deposits the risk of outflows in the form of transfers to other banks or cash redemptions rises. To be able to meet those outflows, the banks need more reserves.
2) In a fiat money system, the central bank controls the quantity of available reserves. There is no inherent or “natural” limit to the quantity of reserves. This is the difference to a gold standard or a (still hypothetical) Bitcoin-economy. If the central bank wants to provide more reserves to the banking system, it can do so. It can literally create reserve assets out of thin air. The quantity of reserves is now simply the result of an administrative decision. Thus, acts of ‘saving’ or the act of putting reserve assets (such as, in the olden days, gold) on deposit with a bank, are no longer needed to boost the banks’ ability to lend and to thus create additional money in the process. The ultimate constraint on money creation is now political.
3) Some explanations of the money creation process assume that the central bank directly manages the quantity of reserves. When the central bank wants the banks to lend more, it gives them more reserves (for example by buying assets from them). As the banks are assumed to be “maxed out” on their present reserves, the extra reserves will quickly lead to more lending. The banks ‘multiple up’ central bank money.
4) Some theories even assume that there is a stable, linear relationship between the quantity of reserves and overall money creation. If reserves go up by Y, overall lending will go up by a x Y. According to this view, the present quantity of reserves tells us about the stance of monetary policy today and of the growth in lending and broad money aggregates tomorrow.
Here is the point: 1) and 2) are certainly correct. There is nothing in the Bank of England paper that refutes 1) and 2). They are beyond dispute. 1) is true for every FRB system. 2) is true for every fiat money system. Only 3) and 4) characterize the theory above as a “money multiplier theory” and only 3) and 4) are being contested.
As to 3), the Bank of England states that it does not “typically” fix reserve levels but instead fixes interest rates, which means the interest rates it charges the banks for borrowing reserve balances from the central bank and it pays the banks for putting reserve assets on deposit with the central bank. By varying interest rates the central bank affects the profitability of additional lending and thus encourages or discourages (depending on present policy objectives) overall bank lending and broad money creation. If lending rises in response to lower rates, the banks can borrow whatever reserves they need from the central bank at the fixed rate, or sell assets to the central bank to obtain reserves. But if the banks lend too much (create too much money) in the eyes of the central bank, the central bank can hike interest rates again, which, all else being equal, should tend to slow borrowing and lending (and money creation), and this will in turn dampen the demand for new reserves.
In a fiat money system the central bank has full control over the quantity and the cost of reserves. It “typically” conducts policy by setting interest rates, rather than fixing the quantity of reserves. Both are, of course, linked.
Here is how the Bank of England puts it (page 21):
“Central banks do not typically choose a quantity of reserves to bring about the desired short-term interest rate. […] Rather, they focus on prices — setting interest rates. […] The Bank of England controls interest rates by supplying and remunerating reserves at its chosen policy rate. The supply of both reserves and currency (which together make up base money) is determined by banks’ demand for reserves both for the settlement of payments and to meet demand for currency from their customers — demand that the central bank typically accommodates.”
It is, of course, conceivable, that the central bank could have established procedures by which it tried to affect bank lending through adjustments of reserves rather than interest rates. But as the central bank has chosen to “typically” use rates as a policy tool, the banks have learnt to take their cues from rates. Banks know that at the present policy rate they can always borrow more reserves but they also know that if the central bank should think that there was too much lending going on, the central bank could hike rates again and thus increase the cost of obtaining reserve balances.
As to 4), it strikes me as completely unnecessary and even silly to assume that changes in the quantity of reserves must somehow directly lead to equivalent changes in broad money. Of course, there is ultimately a powerful link between the two, and it may make sense for the purpose of simple illustration to assume that there even is a linear one. But neither is there such a direct relationship between lower policy rates and additional lending. The central bank can only ever hope to encourage additional lending. It can never force additional lending and borrowing onto the system. However, it can always constrain lending. (And it can use QE to circumvent a reluctant banking sector. And it is also conceivable that the central bank lends directly to the state, something that Graber seems to propose to avoid “austerity”. This process would then circumvent the private sector entirely.)
Reserves Still Matter
Because the Bank of England steers bank lending via rates rather than the quantity of reserves, today’s reserve levels are not a meaningful constraint on tomorrow’s lending and broad money creation. But this does not mean that the quantity of reserves does not matter. This is still a fractional-reserve banking system, and it can be clearly seen from the prominent role that reserves play in the diagrams the Bank of England paper uses to explain money creation.
To illustrate this, here is a quote from the paper (pages 18-19):
“In the first instance, the buyer’s bank settles with the seller’s bank by transferring reserves. But that would leave the buyer’s bank with fewer reserves and more loans relative to its deposits than before. This is likely to be problematic for the bank since it would increase the risk that it would not be able to meet all of its likely outflows. And, in practice, banks make many such loans every day. So if a given bank financed all of its new loans in this way, it would soon run out of reserves.”
Interestingly, the Bank of England then describes a key strategy for banks to avoid this, which is to try and attract extra deposits from its customers. This is the age-old strategy open to all fractional-reserve banks, even under gold standard conditions: offer higher rates on deposits so you get more reserves.
“But whether through deposits or other liabilities, the bank would need to make sure it was attracting and retaining some kind of funds in order to keep expanding lending.”
This is classical FRB stuff. To say that the standard fractional-reserve banking explanation of money creation is a “popular myth” is clearly wrong.
Back to Graeber’s column
Here are some of Graeber’s conclusions, which supposedly make the Bank of England paper “explosive”:
Graeber: “There’s really no limit on how much banks could create, provided they can find someone willing to borrow it.”
This is only true to the extent that the central bank does not impose a limit. In a fiat money economy the ultimate limit to money creation, at least in a technical sense, is monetary policy, not saving, although by creating ever more loans without savings the central bank will ultimately throw the economy into chaos. In any case, this follows from points 1) and 2) above, it does not follow from the Bank of England’s rejection of points 3) and 4) above (the “money multiplier theory”), as Graeber seems to believe.
Graeber: “What this means is that the real limit on the amount of money in circulation is not how much the central bank is willing to lend, but how much government, firms, and ordinary citizens, are willing to borrow.”
This is incorrect. The real limit is still the central bank. If the central bank thinks that too much lending, borrowing and broad money creation occurs, it will hike interest rates and make reserve balances more expensive. This certainly poses a limit to money creation. Here Graeber draws incorrect conclusions from the Bank of England’s rejection of 3) and 4) above.
Here is the Bank of England again (page 17):
“The ultimate constraint on money creation is monetary policy.”
There can be no money creation without the central at least sanctioning it.
Graeber: “So there’s no question of public spending ‘crowding out’ private investment.”
Of course there is. At least within the parameters for overall money creation set by the central bank, the government may crowd out private borrowing.
Graeber: “It’s this understanding (the old “money multiplier myth”, DS) that allows us to continue to talk about money as if it were a limited resource like bauxite or petroleum, to say ‘there’s just not enough money’ to fund social programmes, to speak of the immorality of government debt or of public spending “crowding out” the private sector.”
Money is certainly not a limited resource but that is because of 1) and 2) above being true, not because of 3) and 4) having been dismissed. I am not aware of anybody talking about money in a fiat money economy as if it were a strictly limited entity, and I am not aware of anybody proposing that the state should live within its means (its income from taxation), something that used to be the norm but is now called “austerity”, because our monetary system is unable to produce a sufficient quantity of monetary units.
We can, of course, quickly agree with Graeber that in a fiat money economy the central bank can use the printing press to fund the government. From the side of money creation, there are no technical reasons why the government could not spend what it deemed necessary and fund it by borrowing, and have the central bank make sure that enough fiat money was flowing towards the state. But the argument for balanced budgets — and ultimately for hard and apolitical money — are the economic consequences of inflationism: rising debt levels, misallocations of capital, distorted markets, capital consumption, and ultimately money destruction with grave consequences for society overall. The argument was never about technical feasibility. Yes, Professor Graeber, we are entirely capable of destroying our monetary system.
Inaccuracies in the Bank of England Paper
1) If the “money multiplier theory” makes it appear as if the quantity of reserves was all-important, the Bank of England paper is guilty of stressing the role of interest rates too much. The two are, of course, connected. If the US Federal Reserve wants to ensure that the targeted policy rate, in this case the Federal Funds rate, trades at a lower level, it may have to engage in open market operations with the banks, i.e. add additional quantities of reserves to the system, in order to get the rate to the new, lower target level.
2) The description of quantitative easing towards the end of the paper assumes that QE operations involve the buying of assets from non-banks, such as pension funds, and that banks are only intermediaries. However, many QE operations may involve the buying of assets from the banks. For example, QE1 in the US in the autumn of 2008 was certainly targeted at banks. The consequences for balance sheets are then different from what the Bank of England paper describes.
3) Two short videos accompany the Bank of England paper. One explains why we use money. In this video, it is claimed that a 20-pound banknote was an I.O.U.. This is certainly false. A fiat money banknote is a non-redeemable piece of paper. It has no maturity and it is not a claim on any other asset or any payment or delivery of any kind. If you bring your 20-pound banknote to the central bank you get a new note, or you get change. (Maybe you can call it an “I.O.U. Nothing”, as Doug Casey did.) This is certainly not an I.O.U. by any reasonable definition.
Money Proper Is Not an I.O.U.
The video on why we use money appears to be based on Graeber’s theory, as advanced in his Debt — The First 5,000 Years, that credit came before money and that money was first an I.O.U. In so doing the video highlights the inadequacy of Graeber’s theory as an explanation for the rise of the modern monetary economy.
A farmer wants to trade his berries for a fisherman’s catch of fish. This is an example of direct exchange, barter. The problem of “double coincidence of wants” presents itself immediately (the two parties can only engage if they want what the other has to offer) but is at first ignored by the narrator. Instead, the focus shifts to the problem of separating the times of delivery of the two parts to the transaction: selling berries now but buying fish later. The time lag can be bridged with the issuance of an I.O.U., and the assumption is made that this is how money came about.
But money as an I.O.U. cannot solve the problem of “double coincidence of wants” (it can only lessen it, the I.O.U. remains a claim on a specific good or service, it is not a general medium of exchange) and, as the narrator points out, it crucially depends on trust. As a recipient of an I.O.U. you have to trust that the issuer can deliver, will deliver, and will deliver in something that you later really want. This is why the really important break-through in the evolution of a monetary economy came when commodities were used as money. Gold and silver functioned as media of exchange for thousands of years precisely because they are not I.O.U.s but still valuable. Gold is an asset that is not simultaneously somebody else’s liability. Holding this form of money gives you purchasing power (as long as gold and silver are generally accepted) without tying you to any specific issuer, without, in banker parlance, imposing ‘counterparty-risk’ on you.
Gold and silver have been money for thousands of years and they are not I.O.U.s. Bitcoin could become money, and it is certainly not an I.O.U. Fiat money banknotes and coins are not I.O.U.s. Deposit money can be considered an I.O.U. It is a claim — usually instantly redeemable — against the bank for the delivery of money proper, that is, banknotes and coins. The only reason the public is happy to use deposit money to the large extent it does today is because it trusts that the central-bank –dominated, fiat-money-based infrastructure assures instant convertibility in notes and coins, that is, in forms of money that are not I.O.U.s.