Winner of the New Statesman SPERI Prize in Political Economy 2016

Sunday 23 March 2014

Bank says money multiplier is wrong - should we be shocked?

For teachers and students of economics

A post I wrote nearly two years ago had the emphatic title “Kill the money multiplier!” A recent article in the Bank of England’s Quarterly bulletin, by Michael McLeay, Amar Radia and Ryland Thomas, is a little more circumspect, but the message is essentially the same. One of their ‘headlines’ is: “Money creation in practice differs from some popular misconceptions — banks do not act simply as intermediaries, lending out deposits that savers place with them, and nor do they ‘multiply up’ central bank money to create new loans and deposits.”

The article has created quite a stir. (See this post from Frances Coppola.) Some have tried to suggest that it represents a fatal blow to mainstream theory, or current policy. David Graeber writes that the article has “effectively thrown the entire theoretical basis for austerity out of the window.” This is nonsense. What the article does is outline the understanding of most of those currently involved in monetary policy (including academics), and contrasts this with how monetary policy is taught in undergraduate textbooks. (I should add that the article does this rather well, and is well worth reading.)

So why is there this disconnect between current thought and the undergraduate textbooks? The textbook approach does have its supporters: see this post by Nick Rowe for example. I could try and portray this as a continuing battle between Wicksellians and Monetarists, and discuss whether Quantitative Easing is a win for either side. That would make a nice discussion. However I think using textbooks as a pretext would be wrong.

Think of another standard part of textbook macro besides the LM curve and money multiplier. One of the first things students also learn is the Keynesian multiplier, where changes in government spending can lead to much larger changes to output because the marginal propensity to consume is closer to one than zero. Again this does not correspond with how most macroeconomists today think about the real world. Is this disconnect because of a rearguard action by old fashioned Keynesians who insist that the 1960s way of doing macro must survive? Of course not. (Any new readers please note, I am not arguing against Keynesian economics or that the multiplier is zero - see here. I just think we would be better off with new undergrads assuming a multiplier of one, and focusing instead on why output was demand determined in the first place.)

In both cases, this disconnect between undergraduate textbook macro and current practice has a far simpler explanation - the textbooks are out of date. The core of what is taught to undergraduates has not changed in fifty years, whereas macroeconomic thinking has changed substantially. However we are not using fifty year old textbooks. You will actually find a great deal of the more modern stuff in the textbooks, but essentially in the form of add-ons. So first students are taught that central banks fix the money supply, and then they learn about Taylor rules. First they are taught about a Keynesian consumption function (with a large mpc), and then they learn about consumption smoothing.

This is silly. It is also dangerous, because the problem with add-ons is that they may not get added on. In particular, students who learn all about the money multiplier may never go on to be taught that if banks are not short of reserves or have easy access to them, they can simply create deposits by issuing loans. I suspect it is this lacuna which helped motivate the Bank’s authors to write their article. 

So how does this silly and dangerous situation persist? One clue is that the same gap between what is taught and current practice does not exist at masters level. Masters teaching is much less dependent on the textbook. So I think we really need to look at the production of textbooks to understand what is going on. Now what follows is a theory, and as I have never written a textbook or talked about this to those who have it is based on no empirical evidence - but my theory is microfounded!

Suppose a leading macroeconomists wants to write a textbook, and they want to throw away the LM curve, and the Keynesian consumption function - or at least not start off with these out of date bits of kit. The publisher will do some market research. The market will consist of two types. There are the young radicals, who are just starting out and are desperate to teach in a more modern way. There are also the older traditionalists, who have been teaching macro according to the existing textbooks for some time. They will tell the publisher that while they have no objection to this more modern stuff appearing somewhere - indeed they think it is a good idea - they really need a textbook that starts off in the traditional way so that they do not need to rewrite their whole course. The publisher then persuades the author that, to make any money, they need to start with the traditional stuff.

If this textbook writer is representative, then the radicals will only have traditional textbooks to choose from. They have to be very radical indeed to teach without a textbook, so they start teaching in the traditional way. This unfortunately means that the radical becomes over time a traditionalist, and the process continues. The once radical will tell themselves that assuming money is fixed, while clearly not literally true, is not so misleading. The marginal propensity to consume may in practice be nearer zero than one, but at least it gets the kids to do some simple algebra and think about system feedbacks. And hey, a reserve constraint on banks issuing money could exist in some situations.

As a result, some students end up believing that banks just lend out deposits and that the central bank controls the money supply via a multiplier. And a central bank feels it needs to write an article pointing out that this is not so. That I think is a bit shocking.  


  1. Your theory of the textbook market is undermined by the fact that Robert Barro published a thoroughly non-Keynesian textbook at a time when he and freshwater macro generally were riding high. It was a flop, at least by comparison with Mankiw and Krugman-Wells.

    Also, is there a better graduate textbook than Benassy's? He's not ashamed to put ye olde IS-LM in there early on (in Chapter 2).

  2. I just finished posting yet another defence of the money multiplier -- giving it an international macro interpretation -- and then I read your post!

    Agreed that the first year textbook market is weird. It is very hard to please everyone, so too much stuff gets kept in. Mankiw was a fairly radical break, because it drastically shortened the total length, which kept slowly growing over time (Samuelson's first edition was a very short book) as the authors kept adding stuff to please individual instructors ("you can always skip it if you don't like that bit"). Mankiw deleted the Keynesian Cross (though it does have a short aside on the multiplier). That was a radical break.

    Many first-year students think that banks cannot create money, because an individual bank can only lend out the money that people have deposited at the bank. The simple money multiplier shows that that is a fallacy of composition, and that the banking system as a whole can create more money than the central bank creates, even if each individual bank only lends out (a fraction of) what has been deposited.

    OK, confession time: have you ever taught first year economics? (I have taught it many many times.)

    I wish all graduate students actually understood intro economics.

    1. Dear Nick

      It is because I teach first year students - up close at Oxford - that I have such strong views about the MM! I have found that what students take away from the MM (after reading Mankiw but before I talk to them), is that (a) the central bank controls the money stock by controlling base money (b) banks are pretty passive, creating deposits only when they receive reserves. I suspect in both cases this is not so much because of the words in Mankiw, but because equations are powerful and students read them in a causal manner.

      Now my turn. I suspect the real difference between us is that I'm a Wicksellian at heart and you are a Monetarist at heart? That was the post I nearly wrote. I did not, because I do not think most macroeconomists are Monetarists at heart. But I could be wrong.

    2. Simon: I take my hat off to you, for teaching intro. Many good economists think it's beneath them. But it's the most important course of all.

      "I suspect in both cases this is not so much because of the words in Mankiw, but because equations are powerful and students read them in a causal manner."

      Spot on. First year students love a "linear" story, in which A causes B which causes C, etc. Despite our teaching them the "non-linear" supply and demand story in micro.

      Before 2008 I was a (Neo) Wicksellian. Because thinking of the Bank of Canada targeting inflation by setting a nominal interest rate, and ignoring M, seemed to work. Until it didn't. Then came the ZLB and "QE". Reading the BoC trying to explain how QE worked from a Wicksellian perspective was like reading atheists explaining the efficacy of prayer. Or materialists trying to summon the ghost they didn't believe in. I returned to the Old Religion of Monetarism, when the Wicksellian god failed.

      Yep, that's the difference between us. Woodford has been very influential (though the BoC was Wicksellian long before Woodford). Probably you are right, and most macroeconomists are Wicksellian. I'm pushing back against the Wicksellian hegemony. It's a tough job, but someone's got to do it! (I really wish that a strong Wicksellian macroeconomist like you would consider my complaints that the Neo-Wicksellian model just assumes an automatic tendency towards full-employment. e.g.

    3. Nick: I'll try and think about your full-employment post. But on QE, I count that as a clear win for the Wicksellian point of view. QE is a demonstration that money is not a 'hot potato'. If QE had had a huge effect, then sure, but it did not, even though the increase in base money was huge. It didn't have a huge effect because interest rates were at the ZLB. What effect it has had has not come (at least in the UK) because banks wanted to loan out their excess reserves. Instead you have to think about what the central bank was buying, and how interest rates in those markets get determined. Which sounds Wicksellian to me.

      What I do agree with is this your description of how central banks talked about QE. But the problem there was that central banks (well at least the Bank of England) wanted to pretend that nothing much had changed except the instrument, because they did not want to admit that they had lost a good deal of control. That was/is a big mistake.

    4. I get the impression that one reason why the money multiplier has appealed to teachers of undergraduates is that, via the story of iterated lending and re-depositing of base money, it provides the opportunity for a neat mathematical trick from series mathematics. I used to teach the money multiplier that way myself until I understood how loans can create deposits, albeit needing reserves to support those deposits (for which reason I still believe that a money multiplier is still a useful concept).

  3. Paul Krugman blog, January 17, 2013 'All Your Base Are Belong To Us: What Is the Question?'

    "It may, however, change the operational details; textbooks that describe monetary policy in terms of the money multiplier will have to be rewritten. But that was already true: serious applied monetary economists pretty much stopped talking about monetary aggregates as a measure of policy years ago, and these days it’s almost all in terms of target interest rates. My concern is that when saying that money and debt are the same thing, it’s way too easy to lose sight of the real distinctions between monetary and fiscal policy that remain."

  4. "The Money Multiplier is dead" say Wren-Lewis, Copolla and the Bank of England. True, as there is no mandatory reserve requirements in the Bank of England system, there is no Money Multiplier.The Money Multiplier is dead in England! But that is just one way to run a central banking system.

    As Rowe has rightly pointed out in a previous post, China's Central Bank has reserve requirements, in fact 20% of deposits in banks are mandatory reserves in the central bank.

    Now that completely changes the argument how money is created, as there have to be excess deposits to create loans. To create a loan of 80, a commercial bank has to have deposits of 100! Unless there are excess deposits (aslready deposited as reserves at the Central Bank), a loan cannot be created. Chinese Banks are in fact deposit constrained! Or reserve constrained.

    So under the British Central Banking system, a commercial bank can create a loan, and deposit of 80 in the same time. Both assets and liability side of balance sheet increase by 80 . The only constraint is demand and sufficient regulatory capital by the bank in Britain.

    In China it is different.

    A Chinese commercial bank will have to have excess reserves at the Chinese Central Bank of 20, first. Only then will it be able to increase its balance sheet by 80, by increasing the loan (asset) and the deposit (liability) side by 80 each.

    I think we should not ignore how the second biggest economy in the world, and a vastly more successful one, runs its central banking system. Interesting would be a discussion what that means in terms of safety of the banking system, steering the economy (by varying reserve requirements), and what system is better for the health of the economy in the long run!

    Or for example, what would happen if quantitative easing took place in the Chinese banking system. There the money multiplier would work straight way!

    1. Good point. Many of my first year students are from China. Canada had required reserves in the past. It may have them in future. First year university is supposed to be about more than just the here and now.

    2. I forgot to add: but you still get a money multiplier with zero desired reserves, It is the reciprocal of the desired currency ratio. The money multiplier is a story about market shares, when the beta banks peg their exchange rates against the alpha bank. When most countries pegged to the US dollar, the Fed was the alpha bank for most of the world, and had a global money multiplier.

    3. Generally a central bank is making loans to other banks and that money can be used for reserves. Banks can issue securities to each other in circular patterns. After converting them to government bonds they are good enough as collateral for the central bank. Likely all this works in China, too. Whatever stability and prudence there may be comes from the fact that the four biggest players are state owned and state controlled.

      The BoE article is very good. In loan issuance phase it's almost impossible to do double entry accounting so that the bank would not create temporary money up to the time of loan withdrawal. Then the same practice opens the road for more temporary account money. See my

  5. I think you're too pessimistic about the textbook market. When I wrote the first edition of my intermediate macro book, about 15 years ago, I got excellent support from my publisher. They understood exactly what I wanted to do, and what niche of the market I was shooting for, and pretty much let me do what I wanted to do. You'll notice that the current version of the book (5th edition U.S., 4th edition Canada) contains no LM curve, and no mention of the money multiplier, which I like to tell people is the most misleading standard element of mainstream money and banking texts, and often appears in macro books.

    That said, most of the publishing industry is hopelessly backward, and it doesn't serve students very well. The standard business model is geared to print, and revisions every 3 years to kill the market for used books. It would be preferable to have an online version of a "book" that could be revised continuously, integrated with a database to update charts as new data comes in.

    1. Of course you are right. In my simple model they will still be teaching the money multiplier in 100 years, and I do not believe that will happen. What will happen is that the niche for texts like yours (or like Carlin and Soskice which is used at Oxford) gradually gets bigger as more 'radicals' adopt it. But its a slow process - too slow. Online textbooks must be much better - and surely must have been tried for some subject somewhere?

  6. The late Peter Donaldson, who was an Economics Tutor at my alma mater, Ruskin College, wrote in his book ‘Guide to the British Economy’ p 43 : ‘The essence of banking is thus to conceal or minimize the difference between currency and bank deposits. So long as the public has sufficient trust in bank soundness, they will not bother to distinguish between cash and deposits. Banking turns out to be a highly sophisticated confidence trick.’

  7. The usual. You say the Keynesian consumption function is an out of date piece of kit. I certainly agree that essentially all academic macroeconomists start modelling with something completely different -- PIH with a representative consumer and additive separability. Then there are add ons (which as in undegraduate study are often not added on). However, I think you should have written "out of fashion" not out of date. I know of no evidence that suggests that the modern approach to modelling consumption adds anything which isn't plain incorrect to Keynes's thoughts on the subject. In contrast I think the evidence strongly supports his ex ante critique of work that came later.

    Did he write anything on consumption with which you disagree ? If so please quote it and explain why you disagree. I am 100% serious.My thoughts at length (with data)

    I admit that I have cheated moving from 1960s macro to Keynes (and also from Hicks 1937 to Keynes 1936). This is a dodge because Keynes wrote a story not a model. the introduction of precise equations (from Keynes 36 to Hicks 37 was a necessary addition and a huge hostage to fortune. OK so lets do 1960 macro vs current macro. Note 1960s academic macro was not as simple as the stuff in textbooks. You are convinced that macro theory has progressed towards understanding since then. I think it has moved away. You should be able to present examples in which current models work better than 1960s era models. I have been looking without success for such examples (recalling that no prominent 1960s era macroeconomist ever interpreted an expectations un-augmented Phillips curve as a structural relationship -- check with fellow Oxonian James Forder who alarmingly studies the history of thought during my lifetime even though I am really still fairly young really)

  8. I think I understand why they write "nor do they ‘multiply up’ central bank money to create new loans and deposits” however I think that claim is a bit too sweeping (or open to misinterpretation) because banks do ‘multiply up’ central bank money in the simple sense that lending and borrowing in a fractional reserve banking system results in a quantity of money in circulation that is greater than the quantity of central bank money. Even if the story of how that happens is more nuanced than the simple money multiplier parable told to undergrads (which, I maintain, is perfectly serviceable so long as it's confined to being no more than an accounting identity accompanied by a just-so story, and isn't used to make the claim that bank lending is always reserve constrained such that central banks can set M by setting H. As with most of economics, it is a mistake to take the story too literally.)

  9. I don't really see why you are so hostile to the MM framework. Given a definition of M which includes some measure of bank deposits, D, then at any time - as a matter of identities - M can be seen as M = [1+c]*H/[c+r] where c and r are respectively the ACTUAL ratios of cash to deposit holdings of the public and the banks, and H is the monetary base. [Note that even if r is zero there is still an MM].

    The question is whether it's a useful identity or framework. If the CB does indeed set H then it will be a useful framework provided the c and r ratios are "stable". How the banks create the deposits is neither here nor there provided the ratios are stable: the standard textbook case - where the banks hand out some of the extra cash deposited with them as a loan and this gets back into the system - is one story; you can tell another in which the bank creates the extra deposits directly. But the expression above must still hold under the stated assumptions and so it seems a reasonable way of teaching people something about the quantity of money.

    And you can go on to explore, using the same framework, where that mechanism fails. An obvious example is where the r ratio becomes unstable, e.g. if banks suddenly desire to protect themselves by holding more reserves, or if they cannot find anyone it is worth lending too.

    And if the CB does NOT set H then you can replace H with a function, for example
    H = g(i), where i is an interest rate target and let that take whatever form you like and work out the implications for the LM curve. And that then would lead you on to discuss the sort of arguments that I think Nick Rowe makes: that i itself is not a long-term or ultimate target; it's a staging post for achieving some else: like a money supply target or an inflation target. Insert these targets and you get different implications for the behaviour of M and the LM curve etc.

    So it seems to me to useful pedagogic framework: it allows you to illustrate how a variety of approaches to monetary policy might work - and not every country has the same approach; shows where they break down; and allows you to talk about the importance of institutional details within a simple framework.

  10. For me the most important part of all this is the ending of the notion that banks are simply passive intermediaries. Exactly how the creation of money works is of secondary importance. What really matters is that people understand that the quantity of money - in its broad sense - circulating in the economy depends as much on the risk appetite of banks as it does on the policy decisions of central banks, and that these two are interdependent. So, for example, imposing tougher micro- and macro-prudential regulations (such as higher capital requirements) discourages banks from lending, and should therefore be regarded as monetary tightening. We have largely ignored the effect of tighter bank regulation on money creation, and I'm afraid I think this is because economics teaching over-emphasises the role of reserves and largely ignores the importance of capital and risk in bank lending decisions.

    There is no mechanical process by which deposits are "automatically" multiplied up into loans. There is simply an ex post ratio - that's all the money multiplier is. And even when there is a reserve requirement, banks can still lend in advance of obtaining required reserves if the interbank market is sufficiently liquid and their standing in the market is good. For creditworthy banks, lack of reserves on their balance sheets is not a constraint on lending.

  11. The argument that loans create deposits is unsound because there can be no bank that can issue a single loan without having any deposits at al all to start with. Do these economists know of any bank in the world that have ever created a single loan without having any liabilities (deposits) in tis books to start with?

    A real case of wrong causality. Economics has reached an end and serves no real purpose. A call to action

  12. Vitor Constancio, vice president of the ECB (2011):

    ‘It is argued by some that financial institutions would be free to instantly transform their loans from the central bank into credit to the non-financial sector. This fits into the old theoretical view about the credit multiplier according to which the sequence of money creation goes from the primary liquidity created by central banks to total money supply created by banks via their credit decisions. In reality the sequence works more in the opposite direction with banks taking first their credit decisions and then looking for the necessary funding and reserves of central bank money.’

  13. post keynesians talked about the moneyDIVISOR for decades. Yeah, "the state of macro is good":)

  14. My name is monica roland. I live in florida USA and i am a happy woman today? and i told my self that any lender that rescue my family from our poor situation, i will refer any person that is looking for loan to her she gave happiness to me and my family, i was in need of a loan of $100,000 USD to start my life all over as i am a single mother with 3 kids I met the GOD fearing woman loan lender that help me with a loan of $100,000,she is a GOD fearing woman, if you are in need of loan and you will pay back the loan contact her tell her that is monica roland that refer you to her. contact Mrs lisa peters via

  15. What if increased debt actually increases the value of money?

    See: Debt gives value to money


Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time. In addition, I cannot publish comments with links to websites because it takes too much time to check whether these sites are legitimate.