Progress on the monetary policy and banking debate

We seem to be moving forward with this discussion on monetary policy, banking, and reserves. John Carney does a good job of summarising some of the initial forays in this back and forth. I am going to try my hand at framing the discussion here using my own analysis of the comments iteratively, with the assistance of more comments of course. Where there are mistakes, I will fix them accordingly.

The last post I wrote and a follow on post by Tom Hickey at Mike Norman’s blog get at the heart of the debate and so I will try to characterise what is being said here.


We have been living in a world predominated by floating exchange rates and currency non-convertibility for forty years now. Nevertheless, most of economics world seems to take a fixed exchange rate, Bretton Woods, or gold standard view of money and banking. In that world, as Warren Mosler quipped, bank lending is reserve constrained with the interest rate an endogenous variable via bank competition for reserves.

I put it this way in December [emphasis added]:

In the old gold convertible system, the central bank had to jack up rates to prevent an outflow of gold. Interest rates were the release valve. But in those old days, only by adjusting the gold peg i.e. depreciating the currency, could countries under attack get away with low rates once the vigilantes were on to them. That’s what happened during the Great Depression. Once the conversion was broken, the currency depreciated and depression lessened immediately.

Today the release valve is always the currency because there is no gold tether. So the currency gives way, not interest rates.

Bond vigilantes and the currency relief valve

What this in effect means for the domestic banking system is that in a nonconvertible floating exchange rate system, lending is not reserve constrained as banks can create reserves by making a loan that creates a deposit. Any one institution can always borrow reserves from other banks or from the Fed itself if it finds itself short of reserves (See this BIS paper from 2010 for further discussion).

The US government, as monopoly issuer of its own sovereign currency, has given the Fed monopoly power in the market for base money. The Fed then exercises this monopoly power by targeting the overnight rate for money, the fed funds rate. That is to say, the Fed targets a rate or a price, not a quantity. Almost all modern central banks of today operate with explicit interest rate targets, allowing the overnight rate to fluctuate within a range. Any monopolist can only control either price or quantity, not both. Now, central banks could target something else like reserves to transmit monetary policy into the economy; and they have done in the past. The Fed targeted reserves from 1979-1982. What the Fed found was that it had only a controlling influence on base money because targeting the monetary base meant volatility in interest rates (see this 2004 ECB paper for further discussion). But, more importantly, because bank loans create deposits that actually need reserves to maintain the integrity of the payments system, the Fed is forced to supply them according to its legal mandate.

In short, reserves are about helping set interest rates, not about pyramiding money on a reserve base.

Under present institutional arrangements, the Fed Funds rate is dependent on the Fed’s supplying the required amount of reserves at any given reserve ratio to keep the interest rate at its target or within its target band.  The Fed can’t target a rate unless it supplies banks with all the reserves that the banks need to make loans at that rate. This means that central banks must be committed to supplying as many reserves as banks want/need in accordance with the lending that they do subject to their capital constraints. Failure to supply the reserves means failure to hit the interest rate target. So in practice, if a banking system as a whole is at the reserve limit, central banks always increase the level of reserves desired by the system in order to maintain the interest rate. Not doing so means at once that the Fed cannot hit its target or that transactions fail as the payments system breaks down.

In sum: In a nonconvertible. floating exchange rate system, the amount of credit in the system is determined by the risk-reward calculations of banks in granting loans and the demand for those loans. Banks are not reserve constrained. They are capital constrained. Financial institutions grant credit based on the capital they have to deal with losses associated with that activity.

I don’t think anything I wrote is particularly controversial for those with banking and money as their primary economic discipline or area of study. But if you read textbooks like the one I got in business school by Glenn Hubbard, you find sentences like "The monetary base sometimes is called high-powered money because a given amount of base allows creation of a multiple amount of money" (p. 420, Money, the Financial System and the Economy, Hubbard, 1995). This suggests that the banks in fact are pyramiding credit/money creation on the back of reserves when this is not the case. I checked my intro college economics textbook by Baumol and Blinder from 1985 and it’s exactly the same kind of stuff. The reality is that banks are not reserve-constrained because the Fed must supply reserves to back loans already granted. Only if and when the Fed decides to raise the fed funds rate to curtail credit growth will reserves be constrained. And they will be demand-constrained, not supply-constrained.

Central Bank Flexibility – tactics, strategy, and policy

Given that framing above, the question everyone is asking is whether any of that matters over the long-term. Here’s how I explained Nick Rowe’s objection to the concept of endogenous money:

I think the real difference between what Nick Rowe is saying and what people like Scott Fullwiler and Steve Keen are saying is that Nick believes over the medium-term, central bank interest rate policy is endogenous. What I think Nick means is that Scott Fullwiler’s view is reasonably clear and straightforward in his view that central monetary policy is exogenous but that it only matters over a short-term time horizon because central bank interest rate policy adjusts endogenously over the medium-term to commercial bank and other economic variables such that it is really endogenous rather than exogenous.

Further, I think Nick Rowe is saying that it creates an expectation of central bank interest rate policy merely by announcing its target rate and the market moves to accommodate that target, knowing the central bank is the monopoly supplier of reserves. In that sense the central bank has control. But what he seems to suggest is that the central bank policy rate cannot be determined independent of macroeconomic variables (like inflation specifically) and that central bank may be forced to change policy based on these, making it possible to treat the central bank policy rate as medium-term endogenous.

Nick Rowe says my view of his previous commentary is fairly accurate. Scott Fullwiler doesn’t like the terms short- and medium-term. He would rather see us talk about Fed tactics, strategy and policy.

Scott frames it this way (with minor edits for readability):

  1. Tactics – can the central bank directly target reserve balances, monetary base, etc?
  2. Strategy – what sort of rules/discretion balance does the central bank follow in adjusting the target it has set tactically. How often? How big of an adjustment each time? By what criteria?
  3. Policy – How does the macro economy work and what role can or should the central bank play in stabilizing it?

Scott goes on to say that:

The debate between Krugman/Keen once it got to issues related to the money multiplier and loanable funds was about tactics–can banks individually or collectively create loans without regard to deposits or reserve balances? This is closely linked to an understanding of what banks are/do and hence Krugman’s view that they didn’t need to be included since inserting them didn’t change how one should view the money multiplier or loanable funds models. This is where I jumped in, because Krugman in my view was completely wrong on these points.

But Krugman’s reply to me, and Rowe’s post, brought in strategy and policy–”the central bank must change the interest rate target by adjusting to events and expectations” which is about how the central bank should adjust its target (strategy) within the context of how the macroeconomy works and interacts with monetary policy (policy)…

The MMT view is that we need to understand how the tactics work to inform our strategy and even our understanding of how the economy works. Krugman tried to suggest understanding the tactics is irrelevant to these two. This is a very significant distinction between the approaches.

Further, in MMT, we keep these three (tactics, strategy, policy) separate when we discuss them. Neoclassicals generally don’t–so, when I say the central bank must set an interest rate target (tactics) but can move that target wherever it wants (the possibilities for strategy), Nick says no the cb must set a target that responds to the economy and thus must be endogenous (strategy in the context of view of macroeconomy). We end up talking past each other as I have not invoked yet at all how central banks “should” set strategy with regard to how the macroeconomy works. While we will disagree on the latter, in our view jumping to that without clarifying and setting a common language for tactics and strategy complicates the discussion unnecessarily.

This is progress.

Translation: we agree on the basics here but semantically there are differences. 

  • MMT’ers believe the central bank, as monopoly supplier of reserves has monopoly power and therefore full discretion to act as an exogenous actor.
  • Nick Rowe says a central bank must set a target that responds iteratively to the economic variables like inflation and thus must be endogenous as a overarching strategy in the context of a macroeconomy).

I think that’s where we stand.

My Conclusions

  • We have been living in a world of floating exchange rates and currency non-convertibility but the economics world very often – and wrongly – takes a Bretton Woods view of money and banking.
  • The Bretton Woods world is one in which bank lending is reserve constrained with the interest rate an endogenous variable via bank competition for reserves.
  • In a nonconvertible floating exchange rate system, lending is not reserve constrained (over the short-term) as banks can create reserves by making a loan that creates a deposit. Any one institution can always borrow reserves from other banks or from the Fed itself if it finds itself short of reserves. If a banking system as a whole is at the reserve limit, central banks always increase the level of reserves desired by the system in order to maintain the interest rate.
  • But questions remain about what a central bank can target and to what effect and as to the discretion a central bank has in adjusting any target it has set "tactically". Some say that over the long-term a central bank must respond iteratively to macro economic variables. Others believe the central bank has full discretion to set policy as an exogenous actor.
  • My question is whether the above suggests banks MUST be including in any realistic economic model for it to have predictive power even in more extreme economic circumstances like the ones that existed during the great credit bubble. The Great Financial Crisis would suggest yes. yet, many in the economics field resist this notion. Hopefully, we can get more answers on this question as a result of this post.
  1. battenburger says

    Nice post, Edward, amazingly managing to extract some actual light from the entertaining explosion of irrationality that the two Ks seem to have produced.

    Did I spot you on twitter pulling SK & Mark Thoma apart & taking them out of the ring ? :-)

  2. Dan Lynch says

    Thanks for being diplomatic and professional about this debate, and for steering the conversation back to policy rather than endless debate about semantics and classifications.

  3. danny says

    The current monetary system is so vague, unnecessarily complex and unaccountable. There is so much debate whether money is exogenous or endogenous…etc

    Whether banks create loans first then source reserves or not doesnt really matter IMO.
    If money is endogenous banks have too much discretion in creating money. They also have too much control over the economy because their lending creates higher demand and wages and then better lending standards to increase loans again. The banks drive the business cycles in an unstable manner.

    On the other hand if banks are reserve constrained as some argue then banks still have the final say on policy because it is the commercial banks that deal directly with the economy not the authority.

    We need to realise that either way the system is ineffective and we need to develop an economic system which isn’t dominated by bank lending but instead facilitated by bank lending.

  4. Dave Holden says

    As an aside, even the capital constraint is weakened during the bubble phase of credit expansion.

    Ron T pointed this out on the Krugman post

    “But when there is a bubble and making loans is wildly profitable (at least short term) more capital flows into banks and this restriction is weak too.”

  5. warren mosler says

    economic models are ‘for further purpose’ so, for example, if it’s deemed the output gap is too large, one could argue for tax cuts and/or govt spending increases, depending on one’s politics.

    But you can then ask, should banks and nuclear power plants be included in that discussion? Both can alter economic outcomes. But for that particular discussion ‘all else is held constant’.

    So yes, what banks do or don’t do, for the purpose of that discussion, would alter the magnitude of the cuts/increases, and maybe even some of the specifics.

    And in the old days, when I first started, savings rates were capped, so a rate hike would defund the savings banks and cut off the mtg market. So institutional structure, again, matters a lot! And because of the institutional structure of banks, the forecasting of policy outcomes need put that front and center.

  6. Zimmer says

    Really nice post Edward.

    I love the fact that you never once suggested that banks create “money.” You always described what the banks create as “credit.”

    Too often, both traditional and non-traditional economists use the terms money and credit interchangeably.

  7. nicholsong says

    Posts like this are why I like Edward Harrison’s analysis so much. Knowledge, perspective, and the ability to synthesize it into nuanced, readable analysis.

    Well done, and thank you.

  8. Nick Rowe says

    Edward: This is sensible. But let me draw your attention to one analogy:

    Let’s start with the gold convertible system, as you describe it above, where the central bank pegs the price of gold. Now replace the peg with a crawling peg, so that the currency depreciates at 2% a year against gold. Now allow a slight spread, say 1% either way, between the upper and lower price of gold, so you have the old “snake in the tunnel” semi-fixed exchange rate system. Now allow a little “base drift” to the target exchange rate. Now replace one commodity standard with another. Replace gold with the CPI basket of goods. What you’ve got is what modern central banks do: inflation targeting.

  9. Nick Rowe says

    In other words, if the monetary system was reserve-constrained under the old gold standard, would it be not equally constrained, mutatis mutandis, under a silver standard, or under the new CPI basket standard?

  10. wh10 says

    Nick, I am admittedly not at all sure about this, but what are your thoughts on this?

    under a convertible system, the ratio of qty of reserves to the thing being exchanged has a direct impact on the exchange rate, right?

    but is that the same thing as the ratio of qty of reserves to the price level under a nonconvertible system? i struggle to see why because, even if i go along with your policy endogenous interest rate conceptualization, it’s still a story about the price or interest rate on reserves directly affecting appetite for credit, as opposed to qty of reserves.

    I think that’s roughly what Scott was saying here:

    “I would argue that it might look that way but the causation is wrong. The CB is targeting inflation via an interest rate target rule, which affects bank lending (customer desires to borrow, obviously), and then the demand for reserve balances at the target rate at any point in time, which the CB accommodates.”

    Disagree? (Probably :).)

    1. Edward Harrison says


      we’ll need to start a different thread on this. Nick has already posted some thoughtful comments in another post on this topic. See here:

      I am out all day. But I hope to dig into this when I return.



      1. Nick Rowe says

        Edward: thanks for posting the link.

        I should add (and I did an update on my post to add this) that I myself had always thought “Yeah, dropping gold changed everything!”. But your paragraph above made me reflect on this, and I’m now not sure that it really did. Oh God! I’m now finding myself moving away from a consensus! The irony!

        Damn capcha trying to exclude non-mathematical economists like me!

  11. Anon1 says


    You write the following:
    “So in practice, if a banking system as a whole is at the reserve limit, central banks always increase the level of reserves desired by the system in order to maintain the interest rate”

    So the banking system as a whole is reserve constrained? And if so, then what happens if the Fed chooses not to supply the “level of reserves desired by the system”? Then clearly individual banks become reserved constrained too.

    1. Edward Harrison says


      I don’t approve anonymous comments. I will make an exception here. But unless you provide a real e-mail address, your comments will be deleted. Your comment is specious because the Fed MUST supply the reserves to maintain the payments system.

      See here:

      The Fed is legally mandated to ensure the payments system and thus must supply the reserves.

  12. srini says

    Implicit in Nick’s thinking (and his framing of the gold versus CPI basket) is some kind of long-run supply curve that is independent of monetary policy–in other word long-run neutrality of money. I don’t think most PK or MMT economists would ever agree to that.

    Here is why. In Nick’s view, “If the central bank held the interest rate fixed, the banks could all expand deposits and loans as much as they wanted. But if the central banks responds to their expansion by raising the interest rate sufficiently, they can’t do that. In fact, if the central bank targeted inflation perfectly, keeping output always at the NAIRU, we are back in a loanable funds theory of the rate of interest.”

    Note that several assumptions here that are slid in as if they are uncontested truth. One, there exists a NAIRU–equivalent to assuming a LRAS curve that is independent of monetary policy–that everyone is aware of. Two, if loans grow faster than what the Fed wants, then inflation will accelerate. I think all of these are highly dubious propositions.NAIRU, if at all it exists, is probably constantly evolving and, given hysteresis, path dependent. I could easily argue that ultra low interest rates will beget greater investment, greater increase in capacity and human capital, lower NAIRU, greater productivity gains, therefore lower inflation. In which case, the tactics will matter. In other words, the long run is a but a series of short runs.

    In any case, we are certainly not “back in the loanable funds world.” Observational equivalence is not the same as actual mechanism.

    1. Nick Rowe says

      srini: suppose it were true that lower interest rates caused lower inflation, and higher interest rates caused higher inflation. That would mean that the Bank of Canada has been turning the steering wheel the wrong way. Which would mean it is a sheer miracle that inflation has averaged almost exactly 2% over the last 20 years since the Bank of Canada decided to target 2% inflation. Unless they are extremely lucky, drivers who think that the steering wheel is connected up the wrong way never get to where they want to go.

  13. joebhed says

    One might think that with all the kerfluffle between Neo and Post Keynesians that this little divide has something to do with real people and the real economy.
    It does not.
    In a separate article Ed explained that all of this disagreement is about HOW the CB “transmits monetary policy to the real economy”.
    Here’s a clue to Ed, Nick, Paul, Steve and Scott.
    The NEED is not for transmitting monetary policy objectives to the real economy, the NEED is for transmitting MONEY to the real economy.

    Consider for a moment Milton Friedman’s writing in his Fiscal and Monetary Framework for Economic Stability.
    Look it up.
    Randy Wray has a short chapter on it in his MMT Primer.
    More important, in the 70s(?) MF wrote his Program for Monetary Stability – basically a recommendation that the GOVERNMENT increase the money supply by 3 percent a year, or so.
    Imagine THAT.
    Friedman recommending the government create the money.
    In proportion to economic growth potential.

    All these psycho-intellectual econ biz-school gyrations become completely unnecessary.
    The tool for transmitting monetary policy into the real economy would be the direct deposit.

    But we’re all way too smart to consider Friedman and real, public, permanent money.

    1. Nick Rowe says

      Of the people engaged in this debate, my perspective is closest to Milton Friedman’s. Friedman abandoned his 3% money supply growth proposal. Sometimes 3% is too little; sometimes it’s too much.

      1. joebhed says

        I wrote “in proportion to economic growth potential.”

        So, if our long-term economic growth potential over the next ten years would average three percent, everything would be fine.

        Were it four 4s and six 2s, then we would have created 2 percent too much money over a ten year period.
        Hopefully the lesson we learn will have us create the right amount or 2 percent less over the next ten years.

        Compare that to the money growth records of the past decade. Using any measure of money.

        That would be about a generation of extremely stable monetary growth, which, when understood truly implies extremely stable financial conditions and a generally stable national economy.

        Not a bad generational legacy.

        The Minsky identification in these new ways of looking at money should eventually show that the systemic and institutional causes of the inevitable financial instability that this endogenous money system hands us lies in the debt-based money system itself.

        Using loans to create money is just a bad idea.

        We need a more scientific and sustainable monetary system in order to have financial and economic stability.

        Sorry but more scientific-based translates into less free-market-based when it comes to issuing the circulating media for the national economy.

        The government is the obvious issuer of stable money.

        1. Tom Usher says

          Hi Joe,

          You wrote, “The government is the obvious issuer of stable money.” No one else responded in agreement. Nevertheless, you’re right.

          Provided the Monetary Authority (MA) under the NEED Act were to do its job correctly, United States Money (USM) or United States Notes (USN) would be vastly superior to what the Fed does.

          You and I disagree on the form the MA should take in order to accomplish the goal in the most efficient manner. You think the MA as structured in the Act is the best method. I say that the MA should be in real-time via computer.

          Anyway, I didn’t want you to feel lonely here regarding USM/USN.


  14. wh10 says

    either way, depends on how much impact we think interest rates have on the economy.

  15. wh10 says

    either way, your incredulity assumes monetary policy has as much impact on inflation as does a steering wheel on the direction of the car.

  16. srini says


    Are you trying to debate on a high school team or are you trying to engage in advancing the discussion.

    You never pin down your mechanism. So, let me ask you exactly, what happens when the Fed has a particular interest rate and bank lending expands and reserve demand expands? How long before inflation picks up? Under what conditions? How long before the Fed decides that it is not noise and that underlying inflation has picked up because we are close to NAIRU?

    I just sketched out a case where inflation could fall for a while. The hardly means inflation will never rise if unemployment keeps dropping. You seem to think that without Friedman’s insight Keynesians have no idea that inflation would pick up. I suggest you go and read Joan Robinson’s book on Unemployment and enlighten yourself.

Comments are closed.

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More