Scott Sumner Agrees that MMT Policy Proposals Are Not Inflationary

By Scott Fullwiler
Scott Sumner sets out to debunk theories of the price level not based on a form of the quantity theory of money, and lumps MMT in with those approaches that “deny open market purchases are inflationary, because you are just exchanging one form of government debt for another.” While this is true, what’s interesting is that from within Sumner’s own paradigm, MMT-related proposals should not be inflationary. This is clear right off the bat when he lists his first “qualifier” or exception to the quantity theory:

  1. If the new base money is interest-bearing reserves, I fully agree that OMOs may not be inflationary.  That’s exchanging one type of debt for another.

And that is about all we need to hear. As we’ve said probably gazillions of times, you can’t have discretionary open market operations beyond that which is consistent with the Fed achieving its federal funds rate target unless an interest-bearing alternative to reserve balances is offered. Traditionally, this has been Treasury securities issued by the Treasury or sold by the Fed. The only way to leave all the reserve balances circulating and achieve a positive interest rate target at the same time would be to pay interest on reserve balances.
For instance, later, when Sumner writes, “Now suppose that in 2007 the US monetized the entire net debt, exchanging $6 trillion in non-interest bearing base money for T-securities,” hopefully he realizes that this is not operationally possible without paying interest at the target rate on the excess reserve balances created unless the Fed wanted to have a zero-rate target.


So, in the MMT proposals, whether for functional finance fiscal operations without bond sales or basic coin seigniorage, or in our critiques of QE, we’ve always recognized that these were not operationally possible with a positive interest rate target unless interest is paid on reserve balances at the Fed’s target rate. As such, we always propose that the rate paid on reserve balances and the target rate be equal.

And just to be clear, when Sumner says above that “OMOs maynot be inflationary” due to what he later describes as expectation effects, again MMT agrees that there can be such indirect effects as when expectations of QE2’s ultimate effect were likely behind rising commodity and equity prices. Cullen Roche pointed this out literally dozens of times.

Now, obviously, the MMT understanding of the effects of interest on reserve balances—namely, to achieve an overnight target while adding reserve balances in a discretionary manner, and virtually nothing else—is completely at odds with Sumner’s view and the view of many other neoclassical economists, where interest on reserve balances is akin to tighter monetary policy. But differences in how the two paradigms understand the monetary system or how monetary policy is transmitted to the rest of the economy are not my point here. The point is, from within Sumner’s own paradigm, policies proposed by MMT’ers aren’t inflationary.

Finally, just as an aside, Sumner concludes with, “So here’s my question: Are there any non-quantity theoretic models of the price level?” Of course, the price level itself can be anything depending on which year uses as a base year and the value at which the base year is set, so what’s really of interest is understanding changes in the price level instead of the level itself. Interestingly, MMT is also a quantity-theoretic model of changes in the price level. The differences are (1) net financial assets of the non-government sector, rather than traditional monetary aggregates, are the MMT’ers preferred measure of “money,” and (2) desired leveraging of the non-government sector is akin to what one might call “velocity.” In MMT, the two of those together (net financial assets of the non-government sector relative to leveraging of existing income) set aggregate demand and ultimately changes in the price level, at least the changes that are demand-driven.

71 Responses to Scott Sumner Agrees that MMT Policy Proposals Are Not Inflationary

  1. Scott,My understanding of price level is limited to MV=PY which I believe I got from Bill Mitchell's blog, but my memory is spotty. Are there different equations for what you and Sumner describe?Thanks,John Newman

  2. Sumner: "MarkS, Yeah he “crushed me.” He claimed you can’t do an OMP without IOR. And the reason; because it would change short term interest rates. And of course short term interest rates never change. Perhaps he doesn’t recall the Volcker era."Scott can you please respond? I don't follow his retort.

  3. I want to know the effect of an OMP if there is no interest rate target, or if you prefer, if the Fed is using the OMO to change the target.Your final paragraph confuses price levels and price indices. Obviously every country in the world can set their price index equal to 100 in the year 2005 if they want to, that doesn't change the fact that price levels around the world vary quite a bit. I want to know why.

  4. BTW, the sarcasm in the retort to MarkS, was aimed at him, not you. He's a frequent commenter at my blog.

  5. Forgive me if I am stating the obvious or missing the issue here…Scott Sumner: "Well at least you are entertaining. “You’ve taught me?” I must be a slow student. And as I mentioned in this post, the QTM doesn’t rely on banking at all. I assumed no banking sector in my Iceland example."…."Brito, That’s right, there aren’t even any banks in my model. This whole posts clearly abstracts from banks. Nobody seem to want to critique what I actually said, which speaks volumes."MMT views the Iceland non-banking sector helicopter drop example as a vertical injection of net financial assets, and yes, that kind of operation can affect price in the MMT paradigm. So I don't see a real disagreement between you guys here. This seems to be what Scott F is saying in the last paragraph- that is where the QTM is in MMT. But when you introduce the banking sector and you start talking about OMOs- this is where Sumner and MMT may depart. The OMO is not the same as the helicopter drop in the MMT view.So Scott Sumner, and maybe I misunderstood you, but you essentially said "that's great, I don't care, let's stick to my post" to someone trying to explain the position that there isn't a direct transmission mechanism from qty of reserves to price. That's not fair. Unless I am missing something, that's what this is all coming back down to….In any case, I am enjoying this Scott vs Scott debate. I wish the blogosphere had direct debates between school of thought "representatives" more often. Honestly, you guys could make a business out of this. I bet people would pay to see these things hashed out in real-time (broadcast videos or something). Something like this is much more likely to lead to constructive consensus building, or at the very least broaden the discourse within the community and improve economics education, than a seemingly segregated blogosphere and journal-sphere that rarely directly interacts. Perhaps that's me being ignorant and naive as a total outsider…

  6. Dear Scott,The Fed is the monopoly supplier of reserve balances. It can't not set a price at which banks can obtain reserve balances or earn interest (even if 0% for the latter). Not supplying the quantity desired to settle payments at the target rate sends the rate toward one of those two. Too few sends the rate to the central bank's lending rate; too many sends it to the rate paid (again, could be zero). Reserve requirements reduce the inelasticity on most days, but not all of them. This is all confirmed in numerous empirical studies, even neoclassical research like Borio, Bindseil, Whitesell, and so forth.Even during 1979-1982 they had a target "range" and would do OMO if the rate was going out of the range; Meulendyke confirmed this over 20 years ago in her discussions with Basil Moore.And, yes, I wasn't trying to suggest you didn't know how a price index works.Best,Scott Fullwiler

  7. "The Fed is the monopoly supplier of reserve balances."So these can be called central bank reserves and are an asset to the bank and a liability to the fed? Plus, they ONLY circulate between the fed and its member banks? Also, central bank reserves can be here today and gone tomorrow? Are those correct?In MV = PY, why shouldn't M be the medium of exchange supply because that is what circulates in the real economy with a velocity? M has units of dollars and velocity has units of transactions per time or just per time.

  8. Scott, I'd like to get your thoughts about NGDPer's idea that once "back on trend" NGDP should be about 4.5% per year and pay no attention to real GDP.http://www.themoneyillusion.com/?p=10109"It’s the Fed’s job to target NGDP, they should pay no attention to RGDP."

  9. This comment has been removed by the author.

  10. John Harvey does a good post on MV = PY here:http://blogs.forbes.com/johntharvey/2011/05/14/money-growth-does-not-cause-inflation/As ever its all down to the assumptions. M expands and contracts endogenously. M is like haircuts, you can't supply more M without a demand for it.V is variable even over the short run.Y The economy can and does come to rest at less-than-full employment.So it's an interesting equation, but like the government deficit it is information – not a means of control of the econom

  11. Scott, I don't follow your first point. Suppose the fed funds target is 2.0% and the discount rate is 3.0% (as you know , they often differ.) Assume IOR is 0%. Now suppose the Fed does and open market sale, which moves the fed funds rate up to 2.25%. Does that policy affect the price level? I say yes (if it is seen as permanent.) I don't see where your reply addresses that situation. Second, so what if a OMO causes the target rate to change, the Fed changes the target rate all the time. I don't see why we even need to discuss interest rates when considering the effects of OMOs. The Fed need not even have a ff target rate. If they have a discount rate, they need not actually lend any money via that route. As it it's it's often almost inactive. They could make it totally inactive. It's all arbitrary. The Fed has the legal authority to move the base around and completely ignore any side effects of rates. As you know, the Fed doesn't really peg rates, as that would leave the price level indeterminate. Instead they adjust the Fed funds target as needed to meet certain macro goals. How do they assure the actual fed funds rate is close to the target? Via OMOs (except obviously at the zero bound.)I know you weren't trying to claim I didn't know what a price index was, but I was trying to claim that you were confusing a price index and a price level. So I don't understand your response. Does MMT have a theory of the price level? Or not?.

  12. Scott Sumner,As Scott F has explained to me many times before, it does not work that way; you are using textbook mythology . The Fed does not change the FFR using OMOs as you suggest, and it cannot. They change the OMO by either announcing it (called the "announcement effect" in the literature) or by changing the interest rate paid on reserves. Otherwise, OMOs inconsistent with the target rate they set will send the FFR to the interest rate paid on reserves (could be 0%) if too much are supplied or to the discount rate if too little are supplied. As Scott F says, this is well documented in the literature coming from both neoclassical and heterodox economists.So the point is, it is the change in the interest rate that may affect the price level, but this is irrespective of the quantity of reserves.I'll let Scott F further expound. Hope I didn't spread misinformation here.

  13. "How do they assure the actual fed funds rate is close to the target? Via OMOs (except obviously at the zero bound.)"So to reiterate- no that's not how it happens, and it is not possible this way. The FFR is changed to the target via the announcement effect or IOR. OMOs are then used after the fact to defend the rate at that equilibrium. It can't be done the other way around because reserves in consistent with the target would send the FFR shooting down to the IOR (could be 0%) or shooting up (theoretically to infinity but just the discount rate if there is one). Again, well documented.

  14. Dear Scott,I need to do a post on this because there are too many details to cover.But quickly on a few points:1. The Fed can't just choose the level of reserve balances it wants to have circulate. Banks are settling $3T/day with reserve balances and you can't provide them with less than they need to do that without risking the payments system's stability. But banks can do nothing with any amount beyond that which they need for this purpose. So, as many others have found, the demand for reserve balances is almost perfectly inelasstic–providing too many sends the rate to the floor (IOR or 0) and too few sends them to the lending facility. This is de facto interest rate targeting. There's no such thing as the market setting the rate–as Marting and McAndrews at the NY Fed wrote a few years ago, "The costs of reserves, both intraday and overnight, are policy variables. Consequently a market for reserves does not play the traditional role of information aggregation and price discovery." RR only reduces the inelasticity a bit on some days, but the basic framework holds regardless.2. The Fed can't not lend at the discount window. Again, the payments system depends on it and they can't get reserve balances anywhere else but the Fed. And even when banks aren't borrowing much overnight, they are borrowing over $100B/day on an intraday basis at the peak of payment flows. Overnight holdings are simply a buffer to avoid not clearing these intraday overdrafts given the steep penalty the Fed puts on that.3. Given the inelasticity of the demand for reserve balances, it's widely reported that the Fed changes rates via an announcement effect. Even prior to 1994 announcements, they would signal via a repo or reverse repo that would necessarily reverse. 4. I wasn't suggesting the Fed never changed the rate. I was merely describing how they hit whatever rate they are targeting at the moment. More later, as the above is incomplete and a bit choppy.Best,Scott Fullwiler

  15. Forgot to say . . . my comments at 11:04 are based on the pre-Lehman procedures in which IOR (zero back then) is set below the target rate. Obviously, once IOR is set equal to the target rate, or the target rate is allowed to fall to zero if there are no IOR, then the Fed can hit whatever reserve balance target it wants to, as long as it's sufficient for banks to settle payments.

  16. Scott Sumner,I am really looking forward to your response because you have afforded us the opportunity to expose you to how monetary operations experts (neoclassical and heterodox alike!) understand central banking operations in a non-convertible fiat, floating rate regime. It was mind blowing for me when I finally wrapped my head around this. The loanable funds framework we are taught in school all our lives is intuitively appealing in its simplicity, but it’s simply not reflective of actual real world institutions and dynamics!The subtlety but profound difference in these frameworks really underscores what it means to be the “monopoly supplier of reserve balances.” This is not just fancy but vacuous semantics. It makes all the difference in the world for understanding real-world central banking operations in a monetary regime such as the U.S. Like Scott F said, the Fed can’t *not* set a price on reserves; this system necessarily leads to de facto interest rate targeting! If the Fed doesn’t explicitly name a price, like you are trying to suggest, the rate either goes to the IOR (could be 0%) or the discount rate (theoretically infinity if there is none). There is no in between! And so, in this vein, any price impacts come from changes in the rate, not changes in reserve quantities! Tell me that’s not cool!

  17. Sumner said:"Suppose the fed funds target is 2.0% and the discount rate is 3.0% (as you know , they often differ.) Assume IOR is 0%."That's not even possible. The IOR serves as the floor. The Fed can't pay 0% on IOR and then announce a 2% target rate. Even a first year would understand this. This is intolerable ignorance. Pardon me, but this is entering the realm of the totally absurd.

  18. ^ MarkS, you are wrong. That is how central banking has worked before the crisis. The Fed did not pay IOR then and it had positive interest rates. It does serve as a floor but the floor can be below the target

  19. Scott F., be aware that Scott S. uses the monetary base for M. I don't.Scott S. said: "Scott, I don't follow your first point. Suppose the fed funds target is 2.0% and the discount rate is 3.0% (as you know , they often differ.) Assume IOR is 0%. Now suppose the Fed does and open market sale, which moves the fed funds rate up to 2.25%. Does that policy affect the price level? I say yes (if it is seen as permanent.) I don't see where your reply addresses that situation."I'm of the opinion that M should be the medium of exchange supply. If that 2.0% to 2.25% move in the fed funds rate doesn't change the medium of exchange supply or its velocity, then nothing happens in the real economy.Is that correct, and can you address what M should be? Thanks!

  20. Sorry, I am not being clear enough. There are obviously excess reserves so if the Fed stopped paying IOR the rate would fall below the 2% rate. Is that not accurate? Wouldn't the Fed have to first remove the reserves before being able to set the rate at 2%?

  21. Scott, We aren't going to get anywhere if we keep talking about interest rates and RRs and bank reserves. I'm talking about changing the monetary base, and I'm talking about doing so in normal times, when rates are above zero. So let's ignore the banking system, and assume the Fed buys $100 billion in T-bonds from the private nonbank sector. They pay for the bonds with currency. That was my post. I'd like to know the effect of that operation. The fact that the Fed doesn't actually do things that way is of no concern to me. I'd like to know what the effect would be if they did do that.It is not correct that large increases in the money supply drive rates to zero, as we have observed numerous times. On average, the faster the rate of growth in the money supply, the higher the nominal level of interest rates. It's true that when rates do fall to zero, there is also often a large increase in M, as the opportunity cost of holding base money falls to zero. But that's a separate issue. Go back to 1980 when inflation was double digits and interest rates were 15%. Now have the Fed announce it's going to flood the economy with cash. Buy up huge amounts of the national debt, and pay for it with currency. What would happen? I'm quite sure rates wouldn't have gone to zero. But if they had I would have gladly snapped up lots of gold and real estate with near zero interest loans!!The Fed can shut the discount window, until recently it was hardly ever used. In the early 2000s I remember checking, and discount loans were like $200 million. That's million, not billion. In any case, I have no interest in what is legally allowed. Obviously it is technically possible to close the discount window. I don't think it would have bad side effects, but even if it did that has no bearing on whether the window could be closed.The whole point of my post was very clearly stated. I want to know if there are non-QTM theories of the price level. In your first reply you confused the price level with a price index. I asked you again. In this reply you didn't comment at all. Can I assume the MMTers have no theory of the price level? That's all I initially wanted to know about MMT, although I obviously appreciate your giving other info about the model.

  22. If the Fed exogenously sets reserves via discretionary purchases, then it MUST pay interest or the interbank rate will go to zero. If it's not zero, then somebody is leaving free money on the table by holding reserves they could lend out at a profit. This is not something debatable. You either understand it or not.

  23. Sumner,You're trying to move the goal posts. You can't just remove the banking system from the equation. Fed policy works THROUGH the banking system. The two are synonymous. Your attempts to remove it are due to your lack of understanding of monetary operations. It might make things easier for you to understand, but it doesn't help move the discussion forward because the examples you are using are not realistic. You should be very concerned with the way the Fed actually does things because that's all that matters. If you want to debate the merits of monetary policy on Mars then be our guest, but there's no point in creating mythical monetary systems that confirm your preconceived notions.

  24. "It is not correct that large increases in the money supply drive rates to zero, as we have observed numerous times."but in the absence of Fed rate setting, what would happen, scott?

  25. "Now have the Fed announce it's going to flood the economy with cash. Buy up huge amounts of the national debt, and pay for it with currency. What would happen? I'm quite sure rates wouldn't have gone to zero."the Fed sets that rate. so you're being quite sure that the rate wouldn't have gone to zero means that you're assuming the Fed is setting a rate above zero

  26. "Obviously it is technically possible to close the discount window. I don't think it would have bad side effects, but even if it did that has no bearing on whether the window could be closed."you have been a monetary economist for how long, scott? how have you not thought about this? this how setting interest rates works

  27. While I can't follow the all the nuances of the above debate, I was intimately, hands-on, involved in money markets throughout the 1990s, and am so, from a slightly higher perch today, so can illuminate at least one point. Required reserves were, and I believe still are to the extent there are still some required reserves, on basis of a two week window. Therefore, it is VERY easy for the open market desk of the NY Fed to first guess at what amount of reserves will clear the market at a rate, and then perform OMO to manipulate the FF rate (and closely related Repo Rate) towards its target both throughout the 2 week window, and on the last day. It would be an error to think that the reserves have to EXACTLY clear the market each day. Daylight "overdrafts" and two week long overshoots and undershoots are a part of the system. At the end of the two weeks greater volatility in the Fed Funds rate was observed to the extent the Fed miscalculated by a few hundred million the actual demand for reserves…

  28. "Your attempts to remove it are due to your lack of understanding of monetary operations. It might make things easier for you to understand, but it doesn't help move the discussion forward because the examples you are using are not realistic."Dude, he's just asking a question about the price level.

  29. "Now have the Fed announce it's going to flood the economy with cash. Buy up huge amounts of the national debt, and pay for it with currency. What would happen? "You'd take the person who instituted that idea in those circumstances outside and shoot them.You simply wouldn't do that in those situations if you were following MMT principles. It's a straw man argument. You haven't understood, or are wilfully mis-representing the way MMT looks to maximise the economy.It targets the unemployment rate and the inflation rate via fiscal means. It allows the currency to float as a buffer.Therefore in double digit inflation and interest rates at 15% the approach would be to tax away the excess flow of funds in the system to get things back to equilibrium.So let's say we finally get rid of the failed monetarists that have causes the unemployed so much suffering and have to deal with that sort of situation.The policy approach would be to introduce a draining taxation measure – say a land value tax for sake of argument – and an Employer of Last Resort programme to pick up the pieces during the adjustment.You would probably need to look at capital requirements of banks and minimum deposit arrangements to prevent the banks creating too much money. Clearly in this situation there is an excess of willing borrowers.Then you slowly wind up the land value tax to drain the excess flow of funds in the system until you can bring down inflation and the monetary interest rates down to something sensible. And the national debt will slowly shrink of its own accord, slowly, because bonds would no longer be issued. MMT based policy wouldn't try to run the economy with just one lever.

  30. Neil, I don't think you understand what's going on here. My post isn't about MMT principles, it about what determines the price level. I can set things up any way I wish. I am not trying to see how a MMT monetary system works, I'm trying to see what happens if the Fed swaps currency for bonds. That's my only interest. If MMTers have something interesting to say about that thought experiment, fine. If they think it's a stupid thought experiment, simply refrain from commenting. But unless you tell me what would happen to the price level under my conditions, I have no starting point for trying to figure out why MMTers think OMPs don't affect prices. Obviously there are lots of differences and we could have an interesting discussion figuring out what they are, and how important they are. But unless I know your view of the price level under my thought experiment, I have no way of understanding how you think about the most basic principles of monetary economics–which go back to Hume and before.

  31. "But unless I know your view of the price level under my thought experiment, I have no way of understanding how you think about the most basic principles of monetary economics–which go back to Hume and before."the most basic principle of modern monetarism is the quantity theory? the quantity theory where m is treated as an exogoneous variable? yikes, monetarism sucks even worse than i thought

  32. @ Scott S"So let's ignore the banking system, and assume the Fed buys $100 billion in T-bonds from the private nonbank sector. They pay for the bonds with currency. That was my post. I'd like to know the effect of that operation. The fact that the Fed doesn't actually do things that way is of no concern to me. I'd like to know what the effect would be if they did do that."I think the point is is that this cannot happen. You cannot simply circumvent the banking system and to do so theoretically is just an abstraction with no bearing on how things work in the real world.To the best of my knowledge, MMT and mainstream economists agree on the effects of money once it 'enters the economy', but the divergence is what must happen for it to enter the economy.Which brings us to:"But unless you tell me what would happen to the price level under my conditions, I have no starting point for trying to figure out why MMTers think OMPs don't affect prices."My understanding is that the interest rate would fall (unless there was a floor set [IOR]). So, what would happen to the price level if the interest-rate fell (I think this is the essence of your question)?Well, that all depends. If the banks have lots of customers with good credit-ratings (or they were being lax in their lending standards), inflationary pressures would build as aggregate demand increased from loans taken out by the public. However if — as is now happening — we have high unemployment, great uncertainty and few willing lenders and borrowers forming contracts, then very little inflationary pressure builds and aggregate demand remains low.This situation — which we now face and which was also faced in the 1930s — shows clearly the poverty of monetary policy. This is why MMTers advocate fiscal policy as a better approach.This is a very interesting debate. More of this sort of thing. Seeing how different the two approaches are is very interesting (and I think I'm being confirmed in my suspicions that mainstream theory is based on aimless abstractions…).

  33. The price level – is it explained by the QTM – or is it rather the other way around?Scott Summner insist that the QTM is a good explanation of the price level.He then creates a fictive economy with a fixed amount of currency and states:“It’s likely that NGDP will end up being roughly 15 to 50 times the value of the stock of currency.”Of course, 15 to 50 times are not a really precise figure, not even as “ballpark” – it’s a factor of 3.3 – so the same quantity of money can have a) stable prices, b) 300 % inflation, or c) serious deflation? – So what does it explain?He further states:“BTW, prices in Japan are 100 times higher than in the US, and Korean prices are 1000 times higher.”Are they? AFAIK, all these countries have their own currency, so there is no common denominator. Comparing “Price levels” without noting that they use different units of account is rather pointless – not all currencies are “created equal”. That’s why there are currency exchange rates. Only if you compare prices (in different currency) at their exchange rate can you compare price levels.The actual price levels (in local currency) are the result of economic history. Comparative price levels are usually explained by the rate of productivity in the tradable goods sector. If productivity in the tradable goods sector in country A) is high, those producers can easily compete on global markets, make a lot of profit and pay good wages. This should create a high demand for labor in this sector (pull workers out of less productive sectors), and (due to the high wages/profits earned in the TGS) more demand for non tradable domestic goods and services (like housing, haircuts, dining out etc.).So, higher demand for non-tradables, and wage pull from the tradable sector, will pull the wage level up, and with this the general price level. So country A) – due to its higher productivity in the TGS should have a higher price level than country B). (And of course, it has a fair chance to have a trade balance surplus, so money flows in to the country, which expands the quantity of money – but this is an effect – and not the cause).So – what else causes inflation (and deflation)?If demand rises, and supply is inelastic, or, if supply falls, and demand is inelastic, supply will fall short of demand, and hence, prices will rise. Now, how can this be, if the quantity of money is stable?Well, some call it velocity, others call it leverage. Anyhow, people, chasing the rare goods, either liquidate savings, or take credit, to keep buying.If supply later becomes adequate again, prices should drop somewhat (but not all the way), but the private sector will be more in debt (relative to GDP) than before.If supply stays inadequate for some time, there are two possibilities. If the net money supply doesn’t increase sufficiently, then credit expansion will come to an end, there will be a demand shock, and recession (and quite possibly, deflation) will follow. If, OTOH, government runs a persistent deficit, which adds to net financial assets of the private sector (that is – increases the net money supply), then the rise in the price level (inflation) might persist, and if supply remains insufficient for quite some time, both deficit and private credit expansion can finally result in hyperinflation.

  34. continued…Or, to put it in other words: an increase in the quantity of money might follow inflation, and an expansion in the net quantity of money (trough new gold mined and minted, a trade balance surplus, or –rather the norm – a government deficit) is a prerequisite for a persistent rise in the price level – but it’s not the cause, it might be an effect.If – say under a gold standard – the amount of money is fixed, the price level can’t rise persistently – but it will fluctuate wildly – supply shocks will result in higher prices, but those will soon be offset by demand shocks (or just normalization).Increasing the money supply trough central bank lending to private banks doesn’t increase the NET financial assets of the private sector (both assets and liabilities of the banking sector increase). An increase in net financial assets of the private sector can only derive from a government deficit or a trade balance surplus (or, if the central bank directly buys NON FINANCIAL ASSETS – like houses, roads, commodities, or labor..).Paying interest on reserves is nothing but a bank subsidy. [In a similar way, paying interest on treasury bonds is a subsidy to bondholders..] And QE is just a special form of subsidy for bondholders – if they can now sell bonds at a market price above parity to the FED, instead of waiting till the bonds are finally redeemed at par at maturity.

  35. Second try with part one..The price level – is it explained by the QTM – or is it rather the other way around?Scott Summner insist that the QTM is a good explanation of the price level.He then creates a fictive economy with a fixed amount of currency and states:“It’s likely that NGDP will end up being roughly 15 to 50 times the value of the stock of currency.”Of course, 15 to 50 times are not a really precise figure, not even as “ballpark” – it’s a factor of 3.3 – so the same quantity of money can have a) stable prices, b) 300 % inflation, or c) serious deflation? – So what does it explain?He further states:“BTW, prices in Japan are 100 times higher than in the US, and Korean prices are 1000 times higher.”Are they? AFAIK, all these countries have their own currency, so there is no common denominator. Comparing “Price levels” without noting that they use different units of account is rather pointless – not all currencies are “created equal”. That’s why there are currency exchange rates. Only if you compare prices (in different currency) at their exchange rate can you compare price levels.The actual price levels (in local currency) are the result of economic history. Comparative price levels are usually explained by the rate of productivity in the tradable goods sector. If productivity in the tradable goods sector in country A) is high, those producers can easily compete on global markets, make a lot of profit and pay good wages. This should create a high demand for labor in this sector (pull workers out of less productive sectors), and (due to the high wages/profits earned in the TGS) more demand for non tradable domestic goods and services (like housing, haircuts, dining out etc.).So, higher demand for non-tradables, and wage pull from the tradable sector, will pull the wage level up, and with this the general price level. So country A) – due to its higher productivity in the TGS should have a higher price level than country B). (And of course, it has a fair chance to have a trade balance surplus, so money flows in to the country, which expands the quantity of money – but this is an effect – and not the cause).So – what else causes inflation (and deflation)?If demand rises, and supply is inelastic, or, if supply falls, and demand is inelastic, supply will fall short of demand, and hence, prices will rise. Now, how can this be, if the quantity of money is stable?Well, some call it velocity, others call it leverage. Anyhow, people, chasing the rare goods, either liquidate savings, or take credit, to keep buying.If supply later becomes adequate again, prices should drop somewhat (but not all the way), but the private sector will be more in debt (relative to GDP) than before.If supply stays inadequate for some time, there are two possibilities. If the net money supply doesn’t increase sufficiently, then credit expansion will come to an end, there will be a demand shock, and recession (and quite possibly, deflation) will follow. If, OTOH, government runs a persistent deficit, which adds to net financial assets of the private sector (that is – increases the net money supply), then the rise in the price level (inflation) might persist, and if supply remains insufficient for quite some time, both deficit and private credit expansion can finally result in hyperinflation.

  36. As a non economist watching this, it looks like Scott S. is trying to force the real economy into a mathematical form that produces consistent results: If you do A then B. While the MMTers are committed to empirical reality and refuse to indulge the abstraction because the mathematicization abstracts out what is relevant for actual humans. If we lived in a world where with 300m Americans the Fed bought directly from mom and pop without intermediation by banks, it would probably have different physical properties and be inhabited by a different species as well.Again as a non economist watching the debate, I have to say the MMTers have a lot more to say about the world I inhabit which has in essence been an externality to the maths of the mainstream for most of my life.

  37. "…say under a gold standard – the amount of money is fixed…"No beef Good Habit, we're in total agreement. But to take this one step further: I don't believe this and I'd love to see some of the MMT 'top dogs' do something on it.Under a gold standard is the money supply really fixed? I can't believe it could be.We say that the 'loan window' is never in contact with the 'reserve dept' and I think this is a strong point. Are we really to assume that they were under the gold standard?I'd love to see a proper MMT post on this. Did the gold standard/Bretton Woods ever operate as it said on the tin? Or was it just hokum that was largely circumvented by banks?Comments much appreciated.

  38. Dear Scott,Regarding your first paragraph, bank reserve balances are part of the monetary base, and they are the part changed when the Fed buys securities. Treasuries exist only in book entry form and are settled only via Fedwire using reserve balances, so it’s simply not possible to buy them with currency in the real world. Not to mention the fact that primary dealers (with whom the Fed does operations in the real world) have no use for that much currency. Further, even if you did (which you can’t, but for the sake of argument), what would happen is primary dealers would want to hold the currency largely as deposits (since that’s how they transact), and so would put the money in the banks, who would then have too much vault cash and would go to the Fed to exchange the vault cash for reserve balances. Only, in your world there is no banking system. This is very strange—it’s as if you want to abstract and abstract until at some point you get to a model that matches your textbook view of monetary economics. I don’t find that useful if we’re really after understanding and making policy in the real world (perhaps you’re not).Regarding your second paragraph, I never said (and would never say) that “increases in the money supply drive rates to zero.” I said that increasing reserve balances beyond that amount desired by banks to settle payments and meet reserve requirements would send the federal funds rate to zero or the remuneration rate. Those are very, very different statements. In your 1980 scenario with rates at 15%, if the Fed buys all the securities, it buys them with reserves (or are we assuming there were no banks in 1980 and treasuries didn’t settle on Fedwire?) and the overnight rate falls to zero unless the Fed pays interest. Other rates do not necessarily fall to zero; they may not fall at all depending on a number of things.Regarding your third paragraph, I was describing intraday overdrafts, not overnight lending. The former was well over $100B/day at peak payment flows. The latter is what you are referring to. The Fed’s daily approach (pre-Lehman) is to take a large chunk of payment settlement onto its balance sheet during the day, then by the end of the day have banks mostly settle up with each other. So, yes, overnight lending by the Fed is and was largely unimportant on a daily basis for payment flows at least given the Fed is fairly efficient at offsetting changes to its own balance sheet (with a lot of help from the Treasury) and estimating reserve demand. The point is that about 20% of GDP is settled on average each business day via reserve accounts. “Closing the discount window,” particularly on an intraday basis, if it amounts to bouncing checks unless some other arrangements are made, would be a catastrophe and perhaps illegal. The Fed’s legal mission—perhaps even it’s primary one—is to maintain the stability of the payments system; bouncing even a small % of $3T/day in payments would hardly be consistent with that. The Lehman crisis would be a cakewalk in comparison, and people running the Fed would lose their jobs.Getting cut off . . . will continue below

  39. Continued . . . Regarding your final paragraph, as I explained in my post’s final paragraph, MMT’ers are quantity theorists—I said we have a quantity theory of AD, but I meant the price level in the language you’re more familiar with (the former is Post Keynesian language). At any rate, as I said there, we’re quantity theorists, but we just think monetarists are looking at the wrong aggregate—we prefer net financial assets of the non-government sector as our M, relative to desired net saving of the non-government sector as our V. Further, at a very basic, simplified, theoretical level, because the government sector is the monopoly supplier of net financial assets to the non-government sector, it is also the price setter (or, in more casual language, the monopoly issuer of the currency is the price setter). It’s sort of an MMT version of the fiscal theory of the price level.Finally, let me add that we don’t find a monetary base explanation of the price level useful. Aside from QE (which, again, necessitates or at least results in the overnight rate being equal the remuneration rate), the monetary base is 100% endogenous. Currency is added to circulation as the private sector desires to hold more of it via bank vault cash (which is replenished when banks purchase more from reserve accounts). Reserve balances are related to payment settlement needs and (where applicable) reserve requirements, and are provided endogenously based on demand for them. Thus, it is economic behavior that drives the monetary base, not vice versa. Certainly the desired monetary base can be affected by interest rates set by the central bank, but there again it is the effect of the interest rate on behavior that is of interest and the causative factor; the change to the monetary base in that case is largely a residual effect.Best,Scott Fullwiler

  40. ScottS,The problem is that your thought experiment is based on a number of underlying modelling assumptions that you have not enumerated and have no backing from real world data.For example the idea that you can borrow endlessly from a bank at 0% looks to me like a money multiplier/loanable funds model which AIUI is rejected by MMT.Under MMT influenced policy model you wouldn't be able to do that because the banks would be capital constrained to prevent excessive endogenous money creation.I've noticed that the way economists tend to debate is not to understand the other guys model, but to try and surround them with their own model constraints in a weird intellectual pincer movement.I don't see how that helps understanding.Within your mathematical model, the equations will no doubt be consistent. The question is whether that mathematical model has any relevance in the real world.And the real world has banks in them.

  41. Don't know why this blog has swallowed the longer part of my post for the second time (I did a short end starting with the words: "Continued". I had tried an explanation of the price level, explaining that usually, the Quantity of money would follow changes in the price level (an not vice versa). The full text can still be found on Scott Fullwillers blog, and on my discussion forum…

  42. Typo – or freudian mistake..the full text can still be found scott SUMNERS blog.. sorry..

  43. Great Scotts:MMT’ers are understandably resistant to “thought experiments” whose structural foundations contradict the actual facts of present day monetary operations – at least without setting out the premise for why that should be necessary or feasible or useful in the sense of some defined institutional evolution.The Fed doesn’t swap currency for bonds. Why would anybody expect to gain insight from an operation that is so unrealistic? That’s not how either the supply or the demand for currency comes about at all in today’s world.To attempt to draw inferences from that about the price level or anything else really isn’t very helpful to understanding monetary economics in the context of the monetary system we actually have.That sort of thought experiment presumes much larger changes in the structure and operation of the commercial banking system and its central bank. Fine – but construct that institutional evolution first as a prerequisite – before inserting an unrealistic transaction willy-nilly somewhere into the current reality.Much of the monetarism I see written on blogs strikes me as outbreaks of willy-nilly-ism, ignoring the common sense need for some connection to how the monetary system actually works in the present reality.It’s always puzzled me how people try to carve out strange lessons about monetary economics without understanding the monetary and banking system we have now.Just one example of a different approach is Scott F’s comment:“Currency is added to circulation as the private sector desires to hold more of it via bank vault cash (which is replenished when banks purchase more from reserve accounts). Reserve balances are related to payment settlement needs and (where applicable) reserve requirements, and are provided endogenously based on demand for them. Thus, it is economic behaviour that drives the monetary base, not vice versa. Certainly the desired monetary base can be affected by interest rates set by the central bank, but there again it is the effect of the interest rate on behaviour that is of interest and the causative factor; the change to the monetary base in that case is largely a residual effect.”This is a completely accurate description of the way the banking system actually works today, which isn’t a bad start for a discussion – quite different from a “thought experiment” whose implicit premise is contradictory to such a reference point – particularly when it hasn’t justified its reasoning for why such a departure from present reality is useful at all.I personally have problems with MMT. But one thing I pay attention to is at least its ATTEMPT to anchor the development of any sort of theory with some kind of balance between the system we have now and an alternative system that might be possible.Theory is vacuous if it is entirely severed from some connection to present institutional realities. If there is no connection, there can’t be a logical destination for whatever usefulness the theory purports to have.Thought experiments that are impossibly contradictory and severed from that standpoint are not very useful.

  44. Scott Sumner said…"Scott, We aren't going to get anywhere if we keep talking about interest rates and RRs and bank reserves. I'm talking about changing the monetary base, and I'm talking about doing so in normal times, when rates are above zero. So let's ignore the banking system, and assume the Fed buys $100 billion in T-bonds from the private nonbank sector."Are you serious? You need to assume away banks to talk monetary policy? Without banks there is only fiscal policy. So anyway, we can agree that you understanding of economics is relevant only in a world without banks, when banks are present you have nothing relevant to say because you don't want to "focus too much on the visible, the concrete, the accounting, the institutions" – meaning you don't want to know how the monetary system actually works.

  45. Comments were getting caught in the spam for some reason, including my own. I went and published all those that weren't actually spam and weren't repeated comments from those trying to get aroud the spam filter. We need to keep on top of that more regularly. Sorry about that.Scott Fullwiler

  46. Very good to see you here JKH! When you have the chance, I'm sure all of us would be interested in hearing your critique of MMT. If you do get the time, it might be best to post it at Mosler's where it could he and others could interact immediately (and won't get caught in a spam filter!). Best,Scott Fullwiler

  47. @ Scott FMaybe you'd let JKH actually write a piece criticising MMT? I'd love to see it. And then see responses from you guys.

  48. As an MMT learner, I thought the answer to ScottS's question was going to be simply that the Fed buying bonds with currency does not impact the price level, or that it is somewhat deflationary due to the loss of interest income, whereas I would expect ScottS to argue that this would be inflationary "money-printing". But judging by all the opposition to the question I suspect that I am missing something important.Anyone care to tell me what it is?

  49. yes JKH. Please JKH post your critique of MMT @ warens blog. Yourself, anon, and Ramman are great posters who have contributed thoughtfull counterpoints to MMT.

  50. @Anonymous at 11:21 on 7/23I would agree with your initial assessment. Sorry I wasn't more explicit in taking it that far. Best,Scott Fullwiler

  51. Scott F.,Oops.I think I may have overstated something, or at least stated something with the wrong emphasis.My “problems” with MMT may have something to do with the organization of how the existing banking and financial systems are described in the context of MMT’s main themes. Not that there are micro inaccuracies, but just the overall organization of the material. But I would likely say the same thing if I were looking at any book on banking, and thought I might present it better or differently. Easy to say, or allege. That said, my thoughts on the subject concerning MMT are probably too vague at this point to be very productive in post form or whatnot.Moreover, my only reason for even saying something like that here is really that I wanted to emphasize by contrast a particular point on how MMT approaches economics, where I’m in 100 per cent agreement, as opposed to how “mainstream” approaches it, and this is a very important point in my opinion. That is, as I said, the sort of intellectual process whereby one develops an argument that can go one of two directions – either it connects robustly to the actual monetary system that we have, while exploring options for change, or it strays off in completely disconnected fashion, never to reconnect. My impression is that the monetarist people of all stripes really do not understand the existing system – specifically the fundamental difference between liquidity (e.g. money) and capital (e.g. net saving). They do not understand how banks work in this regard, and I’m beginning to think they do not understand how households work in a roughly parallel fashion. It is not inconceivable to me that the proper argument about banks and reserves and capital, which MMT’ers clearly understand, could be extended in analogous form to the behaviour of households, in terms of their responses to fiscal policy (which generates capital in the form of net financial assets and household net worth) versus monetary policy (which generates, at best), “liquidity” in the form of asset swaps. This is a matter of understanding balance sheets, and where to find liquidity and capital on balance sheets. In my view, even the most unsophisticated consumer has an intuitive connection at least to the facts of his/her own personal balance sheet in this regard, including liquidity and capital.With respect to the problem of intellectual “disconnection” from the existing system, I’ll draw an analogy. Suppose you could describe the universe as a Venn diagram, with one circle. The universe then consists of everything inside the circle plus everything outside the circle. Somebody then attempts to develop a theory of the universe based on everything outside the circle, without KNOWING what is inside the circle. Substitute the existing banking system for “inside the circle” and that may be my point about monetarism. By contrast, MMT’ers anchor themselves with factual knowledge of what’s inside the circle, and work out from there. And in doing so, they develop theories about how the universe might be “re-engineered” for the better (in their view) by reconfiguring what’s inside the circle – e.g. “no bonds”, or “zero rates”.BTW, when I said MMT "ATTEMPTS", please know that I meant – “attempts” versus “not attempts” – not “attempts” as in “attempts and fails”.That’s all for now…

  52. Scott F,Then does Scott S have a valid point that we can answer the question without delving into details of reserve accounting and the banking system? It doesn't seem beyond the realm of possibility that the fed could literally deliver cash in exchange for bonds. It seems to me that MMT should be able to predict what would happen to prices if they did this, even if the same prediction did not hold for our actual, current banking system.@Anonymous at 11:21 on 7/23

  53. Brian,I don't see how the Fed could buy tsy's with currency unless you changed so many things as to be describing a caricature of our current system. That said, if we just want to talk about what happens if tsy's on the pvt sector's balance sheets are ultimately replaced with currency, MMT'ers have always argued this would not be inflationary. If my retirement portfolio were suddenly converted to cash, do I all of a sudden spend it? In fact, since now my interest income is lower, now I'll need to save more to enjoy the same retirement I would have enjoyed with my retirement portfolio in bonds. As such, there's a good case to be made that I would actually spend less.This goes to the larger issue of (mis)understanding the relationship between "money" and spending. Sure, everyone uses "money" to spend. But someone with wealth–say, holding tsy's–always has at least 20 dealers making a market and so there is no issue with whether or not they can spend. Virtually any financial asset is the same way; there might be a capital loss, but that just suggests you should hold deposits instead of tsy's or equities if you're worried about that (further demonstrating that deposits don't necessarily necessitating spending). The qty of "money" held at any point is not the relevant measure of what necessitates or constrains spending. Without turning this into a treatise (I could go into much more detail), the overarching point is that if you don't understand the details/accounting of the real world financial system, or if you abstract from it too much, you can easily come to the wrong conclusions about causation.

  54. JKH,No worries at all, though your elaboration was certainly welcome.If you ever have the desire to further the discussion on this, I'm all ears (eyes?):"My “problems” with MMT may have something to do with the organization of how the existing banking and financial systems are described in the context of MMT’s main themes. Not that there are micro inaccuracies, but just the overall organization of the material. "

  55. Scott FThat helps a lot thank you! Your retirement portfolio analogy brings home the point.brian

  56. "If my retirement portfolio were suddenly converted to cash, do I all of a sudden spend it? In fact, since now my interest income is lower, now I'll need to save more to enjoy the same retirement I would have enjoyed with my retirement portfolio in bonds. As such, there's a good case to be made that I would actually spend less."Do my Apple comments from other posts apply here?I don't believe Apple (the corporation) has any plans to retire. So what are/is they/it saving all that $76 billion for?Instead of spending less, would you try to find another financial asset with the same or higher expected return?

  57. Had Enough,Regarding Apple, I have no idea what they're doing with the cash. Microsoft had $50B several years ago and paid most of it out as a special dividend. Berkshire Hathaway's sitting on a similar sized pile of cash right now and mostly waiting for an opportunity to invest it (obviously, in their case). There's any number of things Apple could do with it and any number of reasons they could be holding it according to basic corporate finance theory.Regarding "instead of spending less . . . " Yes, absolutely, further demonstrating that "money" doesn't necessitate spending. Again, there are many, many directions we could go in from there, so I don't want to give the impression that this is all there is to the MMT understanding of "money."

  58. Scott, JKH, or anyone else:http://www.themoneyillusion.com/?p=10116"Suppose you dump 300,000 Europeans on an uninhabited island—call it Iceland. The ship also drops off some crates of Monopoly money, and they’re told to use it as currency. Assume no growth for simplicity. Also assume no government and no banking system."Next by MarkS, "You said;“Policies like QE work by altering bank reserves.’"Reply: "That’s only been true in recent years. Prior to 2008 90% of new base money went into currency, not reserves. And of course if the Fed wanted that to happen again, they’d just need negative IOR.MarkS, Mishkin’s textbook is the standard in money and banking. I think he has 9 transmission channels for QE. You might want to check it out. As an example, one is the exchange rate. QE lowers the dollar, which boosts net exports."Next, "Are there any differences between your Monopoly money currency and your “prior to 2008 90% of new base money went into currency, not reserves.” currency?"And, “Fed up, No, not much difference between the Iceland example and pre–2008 America.” (for those scanning not the real Iceland)Next, "From what I can tell, I believe that there is a BIG medium of exchange difference between the Monopoly money currency and the pre-2008 America currency."Scott's reply: "Fed Up, And what is that difference?"Anybody want to "take a shot" at answering that?

  59. Had Enough,How about a banking system that creates deposits out of thin air, a central bank that creates currency and reserves out of thin air, and a currency issuing govt that creates net financial assets for the non-govt sector out of thin air, for starters.

  60. Scott, that's a good start.Define "out of thin air".What can't the gov't use/rely on the other three(3)?

  61. Scott F.,I am not clear as to why MMT considers in its quantity approach to price change that in the equation M*V=P*Q, M represents only the net financial assets of the non government sector. Personaly, I tend to view "M" as representing total debt oustanding in the private economy (private and public debt, of course not counting liabilities from the banking sector). The MMT interpretation of M appears quite restrictive and seems to suggest that an increase in the right side of the equation (PQ) could only come about from government deficit spending (ie. injection of net financial assets into the private economy). I do not agree with this -but I could be convinced- since an increase in PQ could also result from an increase in private debt. To use a real life example, as much as subprime loans or super easy credit conditions were a total fiasco in the U.S., one has to ackowledge that while they lasted they contributed to increasing "M", thereby increasing P*Q.I also tend to view velocity as the propensity to spend. Thanks, Qc

  62. Hi QcAs I said at the end of the post, "desired leveraging of the non-government sector is akin to what one might call 'velocity.'" So, we have both govt debt and private sector debt on the MV side. And, yes, V as propensity to spend 'out of existing income' is the point–that's where the "debt" part comes in.Best,Scott Fullwiler

  63. Qc said: "I am not clear as to why MMT considers in its quantity approach to price change that in the equation M*V=P*Q, M represents only the net financial assets of the non government sector. Personaly, I tend to view "M" as representing total debt oustanding in the private economy (private and public debt, of course not counting liabilities from the banking sector)."I agree. I call that the medium of exchange supply. What do you consider "liabilities from the banking sector"?Scott said: "As I said at the end of the post, "desired leveraging of the non-government sector is akin to what one might call 'velocity.'"I don't call that velocity. Isn't leveraging just more debt?"To use a real life example, as much as subprime loans or super easy credit conditions were a total fiasco in the U.S., one has to ackowledge that while they lasted they contributed to increasing "M", thereby increasing P*Q."Yes, indeed. Did geekspeak need that increase in M to prevent price deflation from cheap labor (especially WTO entry for china) and positive productivity growth?"while they lasted"So what happened to M when mortgage defaults went above capital set aside for the losses and the collateral fell in value?"I also tend to view velocity as the propensity to spend."So if someone saves (doesn't spend or invest) then does the velocity of the amount of medium of exchange that is saved go to zero(0), thereby lowering the overall velocity of the medium of exchange supply?

  64. Scott said: "No worries at all, though your elaboration was certainly welcome.If you ever have the desire to further the discussion on this, I'm all ears (eyes?):If it is OK, I'm all ears (eyes)? too. I'd like to hear the discussion.

  65. Had Enough,"Velocity" in MV=PY is simply a ratio: V=PY/M. Don't take it literally, as it's never been about "the same money making more transactions." It's simply about how much "saving" there is in the form of M (whichever version of M you want). If there's more (as when there's a basic portfolio shift to liquid assets), then V falls for a given PY, and vice versa.So, "leveraging" is a desire to NOT save. V rises for a given PY.

  66. The two Scotts should have this debate without all of the side comments. Could you guys agree to hammer this out in threads with comments disabled? I was educated at U Chicago (MBA), but I've rejected the macro that I learned because it does not fit real-world observations. I am intrigued by MMT, but I've long struggled to understand how eliminating government debt issues and providing a job guarantee would affect price levels. So, please, keep talking Scotts.

  67. ^Agreed, except I would prefer things to be hammered out in real-time, in-person debate…

  68. Also agreed re S&S hashing it out. (I'd like to watch…) First step: agree on definitions of terms, at least for use in the discussion.

  69. Perhaps the two Scotts will weigh in.