Does the Money Supply Increase When the Fed Buys Bonds?

I keep getting this question over and over again so I think it’s time I address it more directly.  First, let’s begin by understanding what I mean when I say “money”.  When I refer to “money” I am referring to the most widely accepted media of exchange – primarily currency and bank deposits.  It’s true that almost anything in the world can serve as “money”, but as it pertains to our every day lives it is bank deposits and currency that function primarily as money for purposes of transaction.

I won’t get into what gives this “money” value, but I do think it’s important to distinctly differentiate the primary types of money that exist in our monetary system.  ”Inside money” is the dominant form of money.  It is created “inside” the private sector in the form of bank deposits through loan issuance.  ”Outside money” is a facilitating form of money.  It is created “outside” the private sector and exists in the form of cash, coins and bank reserves.

Inside money is created through the issuance of loans.  And banks create these loans without having to multiply their reserves or check with the government first.  Contrary to popular opinion, the money multiplier is a myth.  The way banks create money is quite simple.  When creditworthy customers walk in their doors they issue loans which create deposits and find bank reserves (as required) later.  Banks are not constrained by their reserve balances as many neoclassical economists believe.  In this form, banks rule the monetary roost by controlling the most widely distributed form of money in our monetary system.  In essence, the government has privatized the money supply in what is really a market based system controlled by an oligopoly of banks who compete for loans.

Outside money is created outside the private sector.  This includes cash, coins and bank reserves.  Cash and coin distribution is maintained by the Federal Reserve who acts as an intermediary for the US government.  The actual cash and coins are created by the US Bureau of Engraving and the US Mint, branches of the US Treasury.  Bank reserves are deposits held on reserve at the Federal Reserve banks.  All member banks in the Fed system are required to maintain a percentage of deposits on hold at the Fed.  This system allows the banks to settle interbank payments in a way that is overseen and organized by a third party (the Fed).  You can think of this as an exclusive banking system where only banks can transact.  You and I cannot use bank reserves.  This is the “money” that banks deal in when settling transactions among one another.  Think of it like a nationalized banking system (as if there is one bank) without actually having to nationalize the banks (in other words, the interbank system brings settlement between different banks into one place, but maintains the market based competitive system in the private banking system).

So what happens when the Fed uses “outside money” to purchase bonds?   Like any bank, the Fed can create money “from thin air”.  This is how it creates reserve balances to transact monetary policy.  It has always done this.  For instance, in 2006 reserve balances increased by $50B at member banks as the Fed implemented policy, but no one complained about “money printing” and “debt monetization” back then.  In other words, this is ALWAYS how the Fed implements policy.  But QE has caused a great deal of confusion, thanks in large part to the neoclassical confusion regarding the definition of the “money supply”.

When the Fed purchases bonds they are simply changing the composition of the bank balance sheets.  Lets look at a simple breakdown here showing the bank balance sheet before and after QE:

It’s important to note that the net financial assets of the bank are exactly the same after QE as they were before.  In other words, the bank has experienced no change in its balance sheet except the composition.  We know that banks don’t lend their reserves so this policy is unlikely to have a material impact on the primary form of money – inside money.  There is much debate about some of the more nuanced points here (like QE’s impact on interest rates and bank recapitalization), but I don’t have the time nor the space to get into those details here.  The message I am trying to convey clearly is that the transmission mechanism via QE to increase the money supply is much weaker than most presume.

If we want to get very technical we would have to add that outside money in the system has increased.  Ie, since bank reserves increase the money supply that neoclassicals focus on (such as M1) has been altered substantially.  Assuming non-banks have been divested of a deposit, the supply of inside money has increased, however, the amount of net financial assets remains unchanged.  Hence, the reason why monetary policy appears to be so broken.  If they use that deposit to pay down debt then money has been destroyed.  In an environment where demand for credit is weak, the Fed’s policies simply cannot effectively and directly increase the money supply that matters to the everyday economy.

* See here for more on the actual workings of our monetary system.  

Cullen Roche

Mr. Roche is the Founder of Orcam Financial Group, LLC. Orcam is a financial services firm offering research, private advisory, institutional consulting and educational services.

More Posts - Website

Follow Me:


  1. On second thoughts the relevant point is whether it was a depository institution or a non-bank entity who financed the deficit in the first place.

    Non bank entities (households, firms, pension funds) must buy the bond by debiting a bank deposit. Then, as the government deficit spends they will receive the same amount of deposits back – because they are the beneficiaries of said deficit spending. The net creation of money (deposits) will be zero but the non bank sector will have gained new assets in the form of Treasury securities.

    When it is a bank that buys the government security in the primary market (whether this is done by using pre-existing reserves or by extending credit to the government does not really matter for the point that concerns us) no deposit will be debited in the first place. Thus, when the government spends by crediting household or firm accounts in the banks a new deposit will be created – aka “money”.

    In this case, for the private sector (bank and non-bank) as a whole the NFAs will be in the form of government securities because the newly-created bank deposits are assets for households and firms but liabilities for the banks. The “net” assets of the private sector will thus be the Treasury bills or bonds – just like it happens when it is the non-bank sector that buys the bonds in the primary market.

    • Read your 10:15 & 10:31 comments at MNE. They are correct. Consistent with Brett, JKH and all of MR. Also confirmed by my NY Fed source.

      So, looks like you’re an MRist? :-)

      • I like the rivalry between MMT and MR because it helps us to improve our knowledge of the monetary system.

        No struggle, no gain :)

        • I think “rivalry” is the wrong term. It implies that I am enemies or have angst against someone. I am only seeking the truth. I don’t dislike Mosler, or Sumner or other schools. I just want the truth. As a market practitioner I am concerned with understanding the world for what it is. All of economics describes a world as they want. Austrian econ starts with a small govt ideology and conforms their understanding around it. MMT starts with a big govt view and conforms a view around it. I’m not interested in the policy so much so my view is purely descriptive. In my opinion, MR has become one of the only non-ideological accurate descriptions of the money system that presently exists….

          • I don’t think “rivalry” is pejorative. I see it in the positive sense of competition.

            Just like two persons competing for a gold medal can be friends while remanining rivals. Competition will guarantee they they’ll do the most to succeed in their objective.

            They both strive to be the best – and MMT and MR should also try to attain the best, most accurate description of the real workings of the economy in general and the monetary system in particular.

    • Also, your 10:15 is very important. MMT likes to blur the lines on what money is so they can include tbonds in their description. Tbonds are securities. They are a claim on zero duration money and have no use as a medium of exchange in the same sense of moneyness that a bank deposit or currency has. Take a tbond into Wal-Mart if you dont believe me. Also, this gets messy when describing other securities. Does this mean AAA pvt sector entities issue money as well?? You can see the tangled web MMT weaves here.

      • Money is currency in circulation plus deposits; Tbonds are not money.

        An interesting point made by Richard Werner is that only 3% of money (deposits) is created by the government – the rest being created by banks as a byproduct of new loans.

        I’d like to know how he arrived at this figure. For the UK in 2012 , I suppose he’d have to sum up all new currency and deposits resulting from deficit spending and then divide by the total amount of net deposits created in the banking system.

        • “An interesting point made by Richard Werner is that only 3% of money (deposits) is created by the government – the rest being created by banks as a byproduct of new loans.”

          Can you post a link for the above?


          • I don’t have a link. The quote is from the second edition of the book “Where does money come from?” co-authored by Werner. It was recently issued in the UK.

            The book is a great read even though the more technical, accounting details are not as clear as the tables provided in the Lavoie paper on neo-chartalism that I cited above.

            • Also, your friend “Y” is wrong. He seems to think reserves are the money he has in his bank account. Tell him to call the US Treasury and ask them two questions:

              1) Can the Treasury spend before it obtains credits via taxes or bond sales? (hint: the answer is no).

              2) Can the private sector spend reserves at Wal-Mart? (hint: the answer is no).

              The accounting is rather simple and the reserve voodoo that MMT plays through settlement obscures it all (JKH already explained how the “settlement” idea in MMT is totally misleading – if one actually explores this it’s rather easy to understand how it’s a misrepresentation). The govt procures funds via taxes and bond sales thereby obtaining private inside money which results in a credits to the Treasury’s account. The Fed will intervene to add reserves if necessary (TO ACHIEVE MONETARY POLICY, NOT FISCAL!). Today, with the banks awash in reserves there is no need to intervene so the MMT point on reserves is wrong to begin with. Reserves are issued to help the Fed implement monetary policy. MMT confuses this to obscure the lines between monetary and fiscal policy. It’s wrong. When the Treasury procures funds (as it is legally obligated to do) it can then credit private sector accounts. So, it’s private inside money credit money in and private inside money debited out. The accounting is simple. If “Y” can confirm with the US Treasury that they can spend out of their account without selling bonds or taxing then I am wrong. But I am not wrong and I have verified this with sources at the NY Fed who actually work in these operations on a daily basis.

              Of course, you can obscure this process by consolidating Fed and Tsy balance sheet or by misconstruing what the Fed actually does via reserves, but then you’re not actually describing the US system. You’re describing something else. That’s all well and good, but it doesn’t help anyone actually understand the system we have.

              Govt spending is a redistribution of inside money. It is not a spending of outside money and bank reserves. The reserve process is monetary policy, NOT fiscal policy. MMT confuses the two, plain and simple.

              This was all confirmed by the NY Fed comments. So, you can either trust the guys who do this for a living at the actual institutions implementing it or you can believe a bunch of anonymous commentators on a MMT website. Your choice….

              • The answer to both 1) and 2) is no – I agree.

                But a MMT sympathizer might add on 1) that, while no is the technically correct answer, the point is not really very important because in practice the Treasury always manages to sell each and every bond that it decides to issue in order to finance government deficit spending.

                The private sector is always eager for risk free government securities providing it with a nice amount of interest income by the end of each period – a form of corporate and upper class welfare that we know well here in Brazil (for decades, the Brazilian T bonds provided unbelievably huge amounts of interest to said beneficiaries, although that “tit” is now drying up, so to speak, in relative terms).

                As for government spending always being a distribution of inside money, we’ve already agreed that this is so only when the government sells its new securities to the non bank sector. If banks are doing the buying in the primary market then at the end of the process of deficit spending the household/firm sector will have new, net deposits as assets – that is, new money.

                This can be seen in extremely clear terms in the accounting tables provided in the Lavoie paper. If you simply scrap his first step (banks “lending” to the government by crediting new deposits to the Treasury) it’s still the case that the private sector (non bank) will have gained new deposits by the last step.

                • I agree with you, but as Brett has showed in his MMT critique, the banks on-sell their inventory. They hold, what, 10% of the bonds they buy now? It’s a marginal piece of the pie. As my Fed source explained, the “primary” way this is financed is via non-banks. So, MMT can latch onto these minor points and claim they’re right, but they’re obscuring the bigger picture in doing so and doing no one a service in understanding how the system actually works. To me, that’s the very definition of latching onto some ideology. I don’t understand why that’s necessary. It’s not helpful in any way.

                  You also said:

                  “the Treasury always manages to sell each and every bond that it decides to issue in order to finance government deficit spending.”

                  This is a dangerous assumption. If you study LatAm hyperinflations you’ll find that this is most certainly not true. And it’s not due to the usual MMT excuses like currency pegs or whatever. It’s due to the fact that a sovereign currency issuer is not immune to the realities of hyperinflation. And in a hyperinflation the sovereign will not be able to procure funds via taxes or sell its bonds. So, what will happen? The govt will spend via sourcing its central banks funds. That’s all well and good except that it adds to the money supply by ADDING on top of the supply of inside money that already exists. This is why govt spending in a hyperinflation exacerbates the currency collapse. The inability to procure (redistribute) existing money exacerbates the currency collapse. This is really an Achilles heel in MMT. A thorough researching of LatAm hyperinflations would put a wrecking ball through some of the generalizatons and misrepresentations in MMT.

                  • Yes, but hyperinflations are an exception – an extremely rare exception.

                    In normal conditions, the Treasury of a monetary sovereign will have no trouble selling its bonds at a rate of interest close to the interbank rate that is manipulated by the central bank.

                    This has been so even in a country recently plagued by hyperinflation, such as Brazil. For the last two post-inflationary decades the Treasury has generally managed to sell its issues denominated in Reais at rates equivalent to the Selic (interbank) rate.

                    So that MMT sympathizer interpretation might be true “most of the time”.

                    • Well, dying is a rare occurrence in the grand scheme of one’s life. A doctor who doesn’t understand the causes of death in the system they are supposedly an expert on would not be taken very seriously. The sad reality of hyperinflation is that it’s this 1% occurrence that destroys 100% of the economy. So, to misunderstand it is a rather glaring hole in my opinion. Still, none of this even scratches the surface on why MMT’s description of the govt “creating” money is wrong. The govt absolutely does not create money in our current design.

                    • Oh, and “Paul” at MNE is saying that cash is the real money. Of course MMT will say this. As if anyone pays their taxes in cash that gets shredded (per Mosler). Ha. No, cash is outside money. It facilitates inside money accounts. You can’t obtain cash unless someone has drawn down an account in inside money. Cash is merely one way the Fed (via the Tsy) facilitates the existence of a private money system designed entirely around inside money. We don’t go through life seeking cash. We go through life seeking inside money. And in fact, cash is going extinct in many modern economies. Sweden, for instance, is going cashless by 2020. Inside money rules the monetary roost. It is a system by the banks for the banks. MMT has it completely backwards!

                      This is all plain as day for the person who is willing to open their mind to the realities. There’s no need for the voodoo and misrepresentations.

                    • IMO, a social science theory that is proved correct 90% of the time is a pretty good theory.

                      Say econometric studies show that, in 90% of the instances and countries examined, deficit spending resulted in a statistically significant effect in reducing unemployment and raising economic growth. And that in the other 10% no effect was found.

                      I’d vote for deficit spending in a recession, hands down.


                    • Yes, but I think it’s important to understand why deficit spending helps. The primary way it helps is by creating a flow in the economy. As you know, an economy is just a system of flows. I describe this like the human body. When the flow stops, the body dies. The govt can just about always keep the flow going (a flow in INSIDE money). It can almost always procure funds to spend into the economy and keep this flow going. That’s a very powerful tool we have.

                      The other important fact about deficit spending is that it adds NFA (but not money). This can bolster pvt balance sheets in a big way.

                      What I don’t like is when people describe deficit spending as “money creation” or “money printing”. That misrepresents what actually happens. Deficit spending is an increase in the flow in the economy and an increase in NFA. If people want to criticize where this flow goes then be my guest. That’s a conversation worth having. But don’t misrepresent what it actually is. The operational realities matter.

                    • Your 2:49 PM comment at MNE is very important and something I hope MMTers start to understand. This is the point MR has been making for a long time now. Brett made it years ago (he was way ahead of the curve here).

                      I just looked up the holdings of banks. Depository institutions hold 3% of the outstanding t-bonds. So your #1 example is practically a moot point. Banks on-sell their inventory or hedge it out so their exposure is not direct. Again, this is all consistent with my NY Fed source which is why I am so confident that he’s right.

                      So, should we welcome you onboard the MR train? :-)

            • Now he says:

              “The treasury receives payment in state money and spends in state money. Bank credit is a promise to pay state money. “

              No, the Tsy receives payment in bank money. Again, the NY Fed official explained this explicitly. I don’t know what else you need to know. Sure, you can obscure via the reserve settlement process (which is monetary policy), but that doesn’t change the facts that the NY Fed officials explained to me.

              Also, bank credit is not a promise to pay state money. You don’t pay your taxes in state money. Private citizens pay their taxes in bank money. But again, MMT obscures the settlement process as if there’s something special occurring just because we have an interbank market managed by the Fed in order to help the govt procure funding.

              It’s all voodoo based on an attempt to rationalize policies based on the myth of the “money monopolist”. The MMT reserve settlement explanation is a complete misrepresentation of the actual flow of funds occurring in the monetary system.

        • Forgot to add a term in the end.

          The formula for the percentage of “money” that is issued as a result of the government deficit spending should be something like:

          Currency plus deposits created as a result of deficit spending

          Diivided by

          Currency plus total net deposits created in the banking system

          Times 100

  2. Banks are not constrained by their reserve balances as many neoclassical economists believe.

    Hi Cullen,

    I agree that fractional reserve banking is taught in a misleading way, but I don’t think neoclassical economists believe that. I think nearly all economists understand that at the individual level, the scarcity of a resource is manifested by its price, not by direct visibility over the total amount available. That’s true just as true for bank reserves as it is for bank reserves.

    MMT-ers love to ridicule the establishment by attacking the wrong straw man, but generally they quote journalists or Austrians.

    Maybe I’m wrong, but do you have any evidence to show that the establishment is really that clueless about the banking system?

  3. “What I don’t like is when people describe deficit spending as “money creation” or “money printing”

    Well, usually it’s the neoclassicals who obsess about the evil of money printing to finance the deficit – meaning, presumably, the government selling bonds directly ti the NCB and bypassing the private markets.

    In fact, this gets it all wrong. As we’ve seen in our examples probably the most inflationary form of deficit spending occurs when the non bank sector buys the bonds. Because in this case, said sector will have a highly liquid asset (though not “money”) that will provide interest at the end of the period. Whereas, when the government sells bonds to the banks (same would happen if it sold them to the Fed, Canadian-style) the non bank sector will end up with a non interest paying (or lower interest paying) deposit.

    Likely, the household/firm sector will spend more if it ends up holding a bond instead of a bank deposit.

    Thus, inflation risk is also higher in this case.

    • Yes. You should read this Bezemer/Werner paper. I think you’ll like it. I’ve started doing quite a bit of research on their concept of “disaggregation of credit”. It stresses the focus of endogenous money in credit based money systems and the impacts by which the extension of this credit has differing impacts. They argue that there are productive and non-productive credit uses. The housing bubble, for instance, is a clear example of the way that credit can be used in non-productive ways (for speculation rather than real productive uses). I think this is such an incredibly important point. It’s not only MR consistent, but breaks down the thinking into a much more granular view.

      • I’ll read it yes, thanks for the suggestion.

        I did like Werner’s Positive Money book even though I found the technical, accounting parts less clear than those written by Marc Lavoie.

        It would be nice to have a textbook that explained clearly all the accounting steps that are required for deficit spending, clearing and settlement of balances, the role of the central bank, etc. With a realistic description of what happens in reality – as opposed to the fancy models that one reads in the monetary policy chapters of neoclassical manuals.

        There is a very good text on this at the introductory level – the two chapters on money written by Lavoie and Seccareccia in the Canadian edition of the Baumol & Blinder Macro textbook.

        There is also a book by Bindseil on monetary policy in the euro area that has been praised by Fullwiler, but I hesitate to spend the required time on a what is for me a rather unknown product and author (the paper that Bindseil co-authored on TARGET2 is interesting but unnecessarily complex, IMO).

        Perhaps JKH would care to write an intermediate text on Money? :)

        • JKH and Brett have a lot of different things in the works. Unfortunately, MR is all a work in progress to a large degree so the ideas are literally evolving and being developed as we go.

          I’ll look into the books you mention. Thanks. Have a happy new year!

          • Thanks.

            I also wrote a paper on debt rollover by eurozone countries and TARGET2 – but it’s in Portuguese, unfortunately :)

            Happy New Year for you!

            • From the looks of it you could probably quickly translate it to English! I know the regulars here would be interested in reading it.