An article by Michael McLeay, Amara Radia and Ryland Thomas (“Money Creation in the Modern Economy”) published in the Bank of England Quarterly Bulletin has gotten a lot of attention recently. JKH, who liked it a lot, highlighting it on his blog, and prompting critical responses from, among others, Nick Rowe and Scott Sumner.
Let’s look at the overview of the article provided by the authors.
In the modern economy, most money takes the form of bank deposits. But how those bank deposits are created is often misunderstood: the principal way is through commercial banks making loans. Whenever a bank makes a loan, it simultaneously creates a matching deposit in the borrower’s bank account, thereby creating new money.
The reality of how money is created today differs from the description found in some economics textbooks:
• Rather than banks receiving deposits when households save and then lending them out, bank lending creates deposits.
• In normal times, the central bank does not fix the amount of money in circulation, nor is central bank money ‘multiplied up’ into more loans and deposits.
I start with a small point. What the authors mean by a “modern economy” is unclear, but presumably when they speak about the money created in a modern economy they are referring to the fact that the money held by the non-bank public has increasingly been held in the form of deposits rather than currency or coins (either tokens or precious metals). Thus, Scott Sumner’s complaint that the authors’ usage of “modern” flies in the face of the huge increase in the ratio of base money to broad money is off-target. The relevant ratio is that between currency and the stock of some measure of broad money held by the public, which is not the same as the ratio of base money to the stock of broad money.
I agree that the reality of how money is created differs from the textbook money-multiplier description. See my book on free banking and various posts I have written about the money multiplier and endogenous money. There is no meaningful distinction between “normal times” and “exceptional circumstances” for purposes of understanding how money is created.
Although commercial banks create money through lending, they cannot do so freely without limit. Banks are limited in how much they can lend if they are to remain profitable in a competitive banking system. Prudential regulation also acts as a constraint on banks’ activities in order to maintain the resilience of the financial system. And the households and companies who receive the money created by new lending may take actions that affect the stock of money — they could quickly ‘destroy’ money by using it to repay their existing debt, for instance.
I agree that commercial banks cannot create money without limit. They are constrained by the willingness of the public to hold their liabilities. Not all monies are the same, despite being convertible into each other at par. The ability of a bank to lend is constrained by the willingness of the public to hold the deposits of that bank rather than currency or the deposits of another bank.
Monetary policy acts as the ultimate limit on money creation. The Bank of England aims to make sure the amount of money creation in the economy is consistent with low and stable inflation. In normal times, the Bank of England implements monetary policy by setting the interest rate on central bank reserves. This then influences a range of interest rates in the economy, including those on bank loans.
Monetary policy is certainly a constraint on money creation, but I don’t understand why it is somehow more important (the constraint of last resort?) than the demand of the public to hold money. Monetary policy, in the framework suggested by this article, affects the costs borne by banks in creating deposits. Adopting Marshallian terminology, we could speak of the two blades of a scissors. Which bade is the ultimate blade? I don’t think there is an ultimate blade. In this context, the term “normal times” refers to periods in which interest rates are above the effective zero lower bound (see the following paragraph). But the underlying confusion here is that the authors seem to think that the amount of money created by the banking system actually matters. In fact, it doesn’t matter, because (at least in the theoretical framework being described) the banks create no more and no less money that the amount that the public willingly holds. Thus the amount of bank money created has zero macroeconomic significance.
In exceptional circumstances, when interest rates are at their effective lower bound, money creation and spending in the economy may still be too low to be consistent with the central bank’s monetary policy objectives. One possible response is to undertake a series of asset purchases, or ‘quantitative easing’ (QE). QE is intended to boost the amount of money in the economy directly by purchasing assets, mainly from non-bank financial companies.
Again the underlying problem with this argument is the presumption that the amount of money created by banks – money convertible into the base money created by the central bank – is a magnitude with macroeconomic significance. In the framework being described, there is no macroeconomic significance to that magnitude, because the value of bank money is determined by its convertibility into central bank money and the banking system creates exactly as much money as is willingly held. If the central bank wants to affect the price level, it has to do so by creating an excess demand or excess supply of the money that it — the central bank — creates, not the money created by the banking system.
QE initially increases the amount of bank deposits those companies hold (in place of the assets they sell). Those companies will then wish to rebalance their portfolios of assets by buying higher-yielding assets, raising the price of those assets and stimulating spending in the economy.
If the amount of bank deposits in the economy is the amount that the public wants to hold, QE cannot affect anything by increasing the amount of bank deposits; any unwanted bank deposits are returned to the banking system. It is only an excess of central-bank money that can possibly affect spending.
As a by-product of QE, new central bank reserves are created. But these are not an important part of the transmission mechanism. This article explains how, just as in normal times, these reserves cannot be multiplied into more loans and deposits and how these reserves do not represent ‘free money’ for banks.
The problem with the creation of new central-bank reserves by QE at the zero lower bound is that, central-bank reserves earn a higher return than alternative assets that might be held by banks, so any and all reserves created by the central bank are held willingly by the banking system. The demand of the banking for central bank reserves is unbounded at the zero-lower bound when the central bank pays a higher rate of interest than the yield on the next best alternative asset the bank could hold. If the central bank wants to increase spending, it can only do so by creating reserves that are not willingly held. Thus, in the theortetical framework described by the authors, QE cannot possibly have any effect on any macroeconomic variable. Now that’s a problem.
Uneasy Money 
Well that’s funny because empirically QE appears to have an effect on short term (3-month) interest rates:
UK interest rates [graph at link]:
US interest rates [graph at link]:
These fit the following function to the 10-year and 3-month interest rates:
log r = c log ((1/κ) (NGDP/Mx))
with κ = 10.4 and c = 2.8 for the US and Mx being either the currency component (“M0”) or currency + reserves (MB). Use M0 for the long term rate and MB (includes the QE) for the short term rate.
LikeLike
David,
When discussing money demand, should one distinguish between the public demand for base money versus their demand for bank money? Are there, in some sense, two different money demand functions?
Similarly, should one distinguish between the public’s demand for base currency versus the demand for base reserves?
Also, why does your book have to be so darned expensive on Amazon!!!
Best regards.
LikeLike
QE impacts the macro indirectly if you consider that the gigantic reserves held by banks helps recapitalizing the banks, who then become more willing to lend.
LikeLike
David, I’m very happy you’re looking into this BoE paper!
O/T (although perhaps tenuously related), I asked both Nick & Scott about a very very simple hypothetical regarding the HPE (actually I just borrowed the hypothetical from Scott’s post explaining the HPE). Nick and Scott gave different answers:
http://banking-discussion.blogspot.com/2014/03/toms-epsilon-example.html
It’s a very short post. Perhaps you can answer my question there?
LikeLike
“The ability of a bank to lend is constrained by the willingness of the public to hold the deposits of that bank rather than currency or the deposits of another bank.”
Well you try to bring back the money multiplier from the backdoor by that discussion.
Credit created raises output and income and demand for bank money is also a function of income. So I don’t know what kind of constraint you really mean.
LikeLike
“In the framework being described, there is no macroeconomic significance to that magnitude, because the value of bank money is determined by its convertibility into central bank money and the banking system creates exactly as much money as is willingly held.”
The value of bank money is determined by its convertibility into underlying central bank money, but doesn’t the opposite apply as well? If bank money and central bank money are close substitutes, then an increased quantity of inside money might ‘displace’ central bank money from people’s portfolios, causing the value of central bank money to fall (and the value of inside money to fall too). Wouldn’t this effect on the price level be sufficient to give macroeconomic significance to bank money?
LikeLike
“The problem with the creation of new central-bank reserves by QE at the zero lower bound is that, central-bank reserves earn a higher return than alternative assets that might be held by banks, so any and all reserves created by the central bank are held willingly by the banking system”
I’m a bit confused by this statement as I can envisage a situation where CB pay 0% on reserves and yet these will still be in excess of what is needed for the optimum level of reserves at current lending levels and demand to hold money.
Suppose the interbank target is 0% that the CB is practicing QE. It buys $1trillion of risky assets (whose IR is still positive) from non-banks, which drives up the price and reduces the yield on these assets The non-banks sellers deposit the revenue from these sales in commercial banks.
The main effects of QE then is that interest rates on risky assets have fallen a bit. This will both lead to greater borrowing by firms (who can borrow cheaper for any given level of risk) and greater spending (as people have less incentive to save out of income). These changes will be multiplied up and positively affect NGDP. However even so it is quite possible (and likely) that as a result of this QE action the total reserves in the banking system will now further exceed the optimum reserves that would be needed for current lending levels. This will be true even if there was no incentive (IOR) for banks to hang on to reserves.,
I agree that the level of “broad money” is demand determined but at the zero bound (and when QE is used) the link is broken between base money and broad money in a way that can’t be explained by the fact the banks can get a better return from holding reserves than increasing lending.
LikeLike
JP, BTW, Sadowski needs a copy of Jürg Niehans’ book too. I suggested you two use your Vincent Cate proceeds to get one:
http://www.themoneyillusion.com/?p=26400&cpage=2#comment-324818
😀
LikeLike
David, just finished your post. A very interesting take on it! I await your response to JP Koning. Thanks.
LikeLike
In fairness though David, the BoE article is entitled “Money creation in the modern economy” not “The macroeconomic significance of money creation in the modern economy.”
LikeLike
“If the amount of bank deposits in the economy is the amount that the public wants to hold, QE cannot affect anything by increasing the amount of bank deposits; any unwanted bank deposits are returned to the banking system. It is only an excess of central-bank money that can possibly affect spending.”
QE counterparties to the fed that aren’t depository institutions will still hold the deposits. Much of these deposits become stagnant like excess reserves though. QE does have an effect but it is quite limited because the counterparties it deals with have a low MPC when compared to the general public.
The demand for money is suppressed because it is tied to the demand for loans. If money entered circulation debt free the supply of broad money would increase and it would positively affect ngdp.
LikeLike
” Whenever a bank makes a loan, it simultaneously creates a matching deposit in the borrower’s bank account, thereby creating new money.”
When I see the Sun coming out, I also see that mountain. Then, mountains are created by the Sun, according to some economists. And according to others, there is a limit to how many mountains Sun can create because of space limitations.
This is one more example why economic departments must be abolished asap. http://t.co/oBefwX4Yw8
The situation is more serious that I originally thought.
LikeLike
I think the flaw in your reasoning is that banks create no more or no less money than the public is willing to hold, hence you attribute no macro-significance to money. The fact of the matter is that money demand is also endogenous and more importantly, inversely related to money supply. Banks ability to create money is constrained by the demand for loans, which is inversely related to demand for asset money.
This post explains endogenous money demand tinyurl.com/k3kn3e9 and how it is the primary driver of the business cycle.
LikeLike
Money and deposits are created by loans, they say…
Then they say QE swelled reserves (deposits).
I am missing what?
LikeLike
Benjamin, deposits are created anytime the aggregated banks make a purchase of anything from non-banks, including loans (I like to think of borrowers as selling loans to banks… like they are selling personal bonds). Likewise deposits are eliminated anytime the banks make a payment to non-banks (e.g. pay their electric bill, buy donuts, pay salaries or dividends).
But that’s not the only way for deposits to swell… if the CB purchases assets for example. I don’t know the details, but imagine they paid with cash or a check… sent directly to a non-bank asset seller (I know, there are PD’s… but again, I don’t know all the details there). In the end though, somebody deposits a check or cash: the bank adds two things to it’s balance sheet:
1. reserve-assets (cash can count as reserves, or it can be electronic)
2. deposit-liability (making the bank a debtor to the asset seller)
These two should offset each other and the bank’s equity should remain unchanged. That’s how I view the process.
LikeLike
Jason, Wasn’t the point of QE to affect long-term rates? Why is it that QE affects short rates and M0 affects long rates?
Ilya, Good point. Only banks hold reserves. Banks and non-banks hold currency. If reserves and currency are non-interest-bearing, it is more plausible to aggregate the separate demands and think of an overall demand for base money; when reserves, but not currency, earn interest, I think that the two must certainly be kept separate, and it is only currency that is macroeconomically significant.
About my book, sorry, but I have no control (or input) over pricing. Like most authors, I would prefer a lower price than that set by the publisher. Amazon, I think, actually does discount it slightly.
Rafael, Interesting point. However, I would regard that as a second-order effect.
Tom, My initial reaction is that Nick seems right, because whatever the demand for reserves, it should be constant in real terms. But there may be issues about the transition from the initial to the final equilibrium that I haven’t thought through. If I have more to say I’ll respond on your blog. Welcome to the blogosphere.
JP, Thanks, that’s what I tried to do. Good to hear that from you.
Ramanan, There are often indirect feedback effects. At any instant, if banks create more money via the lending process than the public wants to hold, the excess money flows back to the banks. That’s the constraint.
JP, Excellent point. The issue, however, is how the banks induce the public to switch from holding currency to holding bank money. They can only do so by making their moneys more attractive to hold than currency. They can increase the interest that they pay on deposits and provide more ATM machines to make it easier to convert deposits into currency when desired. If there is a shift in demand from currency to deposits, then that would tend to raise the price level unless the monetary authority is committed to some nominal target and reduces the amount of currency accordingly. So I was being more categorical in my assertions about macroeconomic significance or insignificance than I should have been.
I will try to respond to the remaining comments later today or this evening.
LikeLike
David,
Thanks for taking the time to look at my comment. The initial intuition is essentially Sumner’s claim that the market thought QE was temporary, whereas actually printing money (M0) would be seen as more permanent: e.g. here http://www.themoneyillusion.com/?p=10888
Sumner was making an argument for why QE wasn’t inflationary, but I thought a similar argument would work for interest rates. The primary driver was the fact that the fit of a function r = r(M, NGDP) to the 10-year rate for M = M0 worked pretty well for the 3-month rate when M = MB. I don’t know the actual microeconomic reason it works 🙂 — maybe the short term interest rate markets look to the base MB while long term markets anchor to M0.
I believe “operation twist” was designed to increase the term of the assets on the Fed’s balance sheet; I thought QE was intended to affect short term rates like the overnight rate while being only slightly inflationary (i.e. not printing physical notes and coins). Maybe I am wrong? In the “model” I’ve been playing with, the sharp rise in the base (MB) isn’t inflationary at all because the price level is described by P = P(M0, NGDP): http://informationtransfereconomics.blogspot.com/2014/03/how-money-transfers-information.html
Regards,
Jason
LikeLike
Has no one noticed that every dollar here is created as a debt that is laid upon the poor and no one else?
Step back and take a look at the underlying financial strategy that is always held forth as tradition without portfolio.
This strategy, held, most irrationally, by what seems like every economist out there – that, despite being called a “fiat” currency, the “fiat” portion of its creation has been destroyed and no longer exists.
When this happened is for a historian to discover, but it hasn’t been this way forever, so it must have been within the past 250 years.
To restate this another way, when I see people talk about the creation of the currency itself, everyone insists that no money, despite the “fiat” label, could ever, ever, ever, ever be created without a “balance” of debt upon each dollar produced, that no dollar be allowed to exist unless it has also been stapled firmly to a form of debt that will and must, must, must be paid back by private citizens and no one else.
This is not “fiat” money at all, but a “loan-standard”-based currency, much like a gold-standard, and this is why the gold-bugs insist money creation always be matched by a corresponding debt upon someone else, not themselves.
I mean, this currency is created by a national treasury function (the US Mint in our case) by the nation, not a private individual.
To burden everyone who needs a loan with also being the only people who get the questionable nomination of being the sole supporters of our national currency, yet who are required to pay interest on the loan, is not a credible or sensible economic model.
So I wonder why, when, and how this silly burden came to be placed upon the poor, rather than the rich, as I very much doubt that it has always been this way.
Only the poor pay for the currency. Doesn’t that bother anyone?
Really…there is no reason to insist on currency being backstopped by loans upon the poor, or that those loans should be interest-bearing, since the central bank itself can set any interest rate it likes and sees no problem with lending out money to the wealthy at zero percent.
Why are private banks allowed to exist? They don’t add value but subtract it.
They truly are parasites, these private banks owned by the wealthy who can use them for personal gain.
Right?
LikeLike
Rob, You may be right. The demand for reserves may be unlimited even at a zero nominal rate if short-term real rates are sufficiently negative. However, I do not necessarily accept the premise that the central bank is able to raise the price of assets and drive down their yields. Assets are simply bundles of expected future cash flows. For the value of those assets to rise, either the expected cash flow have to rise or the discount rates at which the assets are evaluated have to fall. It is not clear to me by what mechanism the central bank purchases alter either expected cash flows or the discount rates underlying the market valuations of those expected future cash flows. I actually do have a story for how QE could affect expectations of future cash flows, but that story implies higher not lower discount rates. Alternatively, there is a story about constant expected real cash flows with increased inflation expectations implying reduced real rates, thereby increasing valuations.
I don’t know what you mean when you say that the total reserves in the banking system exceed the optimum reserves need for current lending levels. The optimum is defined in terms of marginal yield on reserves held for clearing payments and is not specifically related to lending.
Tom, I was extending (and correcting) their argument.
cmamonetary.org, At very low or negative real interest rates, bank lending is not profitable. That’s why an increasing share of the money stock has to be issued via open market operations.
Digital Cosmology, Sorry, but you are being way too subtle for me.
H. Publius, No, I am afraid you aren’t following me. I am saying that it is the creation of base money that determines the price level and nominal spending. Given a solvent competitive commercial banking system, the quantity of money can be left to take care of itself, so that there is no need for the monetary authority to concern itself with the total quantity of broad money in the system. If banking system becomes insolvent, it becomes necessary to take steps to restore solvency, which may entail central bank lending or more extreme measures.
I don’t understand what it means to say that money demand is inversely related to money supply. Perhaps you meant to say negatively related. To say that money demand is endogenous seems like a truism. Who disputes that? I agree that banks cannot function effectively in an environment in which real interest rates are too low for banks to earn a profit from their intermediary money creating activities.
Benjamin, Bank money is created by deposits without reserves. QE creates money with reserves.
LikeLike
David, quick follow-up question on your book (which I recently read and loved)–is there no chance of a Kindle release? I think it would be a shame if distribution were indefinitely limited to expensive, hard-to-find print copies.
LikeLike
The amount fiat money in circulation has tripled since 1996, to about $3,400 cash per resident in the United States, or $1.2 trillion. It is rising rapidly too.
Prices are up about 40 percent in same time frame.
Deduction 1: Radical increases in U.S. cash in circulation, even fiat paper money, are not associated with radical increases in U.S. prices. It just ain’t happening.
Deduction 2: And paper money is becoming much more important to the economy than anyone wants to admit.
A lot of people like to say, “Oh, about one-half of cash is circulating outside the USA, you know, underworld stuff mostly.” Nobody knows this for sure, but let that go.
Why would drug cartels absorb more, or less, of fiat money from one decade to the next?
So we cannot say “It is El Chapo!” unless we have evidence the drug world is growing much more rapidly than the economy.
Moreover, I do not think drug cartels are that big. Afghanistan is the globe’s dominant supplier of heroin (that happened under our tutelage, btw). Something like 80 percent to 90 percent of heroin comes from Afghanistan.
But the Afghanistan crop is estimated at $60 billion or so. Seems like druggies and crooks would need some cash, but not $600 billion….
Even if one says, “Yeah, I have seen movies and they have suitcases with cash, the bad guys do,” that leaves $1,700 in cash per US resident.
How many times in a year does a bill circulate? Once a month? Do the math…
Side note: A large briefcase will hold $1 million.
If one assumes there is $600 billion out there in underworld briefcases, that makes about 60,000 briefcases full of cash. No wonder I can’t find my bags at the airport carousel. All the effing drug money is clogging up the airports,
“60,000 Briefcases!” Maybe that will be the the title of my next crime novel….
LikeLike
David,
First , thank you for your detailed response to my attempt to understand this complex issue, its very much appreciated.
On:
“I do not necessarily accept the premise that the central bank is able to raise the price of assets and drive down their yields.”
I was thinking of it like this. Suppose a bond pays an annuity of $100pa with some risk of default. Different people will value this bond differently depending upon their evaluation and toleration of its risk and the discount they apply to the future stream of earnings. If the CB enters the market and starts to buy up this type of bond then this will cause the price to rise and only people who value the bond greater than its new price will choose to hold it. Those who sell their bonds will (as a result of the flattening yield curve) at the margin prefer present goods over future goods and spend some of their cash – leading to increased economic activity
On
“The optimum is defined in terms of marginal yield on reserves held for clearing payments and is not specifically related to lending.”
I agree that is the correct way to look at it. Banks do more than just lend (they could buy bonds directly for example). The reason they may sit on unused reserves even with 0% IOR is that this 0% nominal return is still (in their eyes) a better return than any other risk-adjusted return available to them
LikeLike
David, you write:
“Tom, My initial reaction is that Nick seems right, because whatever the demand for reserves, it should be constant in real terms”
But Nick’s nominal demand for money curve (Mdn) is proportional to the quantity of base money (in this example), thus the real demand for money (Mdr) is also proportional. It’s Scott that keeps Mdr fixed, whit that result that as MB -> 0, then P -> 0, which, IMO, seems strange *in this particular example*
LikeLike
John, Thanks for asking. Sorry, I have no idea. Sales are not that brisk, so I doubt that anyone out there thinks that there is a lucrative market opportunity waiting to be tapped. However, the publisher Cambridge U. Press does have an ebook version available on their site for slightly less than the paperback version ($50), but Amazon doesn’t seem offer it on its site.
Benjamin, I agree that the amount of currency out there is a big puzzle, but what about the amount of gold being held. Of course, the value of dollar currency holdings seems to be greater than the amount of gold holdings.
Rob, I get your argument, but why don’t the people with more optimistic expectations step in and buy more of the bond when the see the price falling?
Tom, OK, I will need to look at that more carefully when I get a chance. Sorry, but I haven’t been able to focus on your question so far.
LikeLike
” why don’t the people with more optimistic expectations step in and buy more of the bond when the see the price falling?”
Well, I suppose they do (they buy the bonds off the pessimists) but this only limits the amount the price falls when the recession starts. Then the QE program pushes the price back up again a bit.
LikeLike
David thanks so much. I had a similar question for Nick a couple of weeks back and he answered it in a similar way. In that case (another cashless society hypothetical) a CB with $1 in reserve-liabilities purchases the entire banking system (for a steal! Their equity was ~$0) and took over their operations. The banking system as a whole had $10 in deposits. Scott & Sadowski (and implicitly Nick) agreed that reserves become meaningless once all the banks are consolidated under the CB, and Scott & Mark (and Nick again) also agreed the bank deposits (now direct CB liabilities) would then become the MOA. Scott even said “base money.” But it was still Nick with the answer: prices would NOT go up by 10x in the long run because both supply and demand for MOA had increased “10 fold.” Neither Scott or Mark objected to Nick’s answer.
LikeLike
David-
Yes, gold is being held…but people do not pay for auto parts or groceries or rent or hair-dos in gold. I think that is a weak analogy.
Another fact: Since 1996 it has been difficult to assemble large amounts of cash. A bank withdrawal of more than $10,000 is recorded. So assembling a large block of cash—the briefcase with a million dollars—is actually difficult. It may be a fiction of the movies, I don’t know.
If so, that would suggest that since 1996 cash has played a larger and larger role in the non-drug US economy. Velocity for cash may be rising, we do not know.
I like to think it is the free market circumventing stupid Fed tightness policies. It is also an example that the USA economy is no longer inflation-prone.
LikeLike
David,
In an earlier comment you said: “However, I do not necessarily accept the premise that the central bank is able to raise the price of assets and drive down their yields. Assets are simply bundles of expected future cash flows. For the value of those assets to rise, either the expected cash flow have to rise or the discount rates at which the assets are evaluated have to fall.”
Keep in mind that future sale price is part of the bundle of cash flows that you refer to; the Fed’s bid price for a particular asset directly impacts expected future cash flows related to the asset. The Fed represents new demand, and new demand increases prices.
Do you disagree?
How well that translates to other assets not targeted by the Fed isn’t so clear.
LikeLike