The Circular Logic Behind Scott Sumner’s Claim That the Fed’s Policy is Contractionary

By L. Randall Wray

OK, I’m flabbergasted.

I came across, and commented on, a piece by Scott Sumner a few days ago. (DID SCOTT SUMNER FIND MMT’S ACHILLES’ HEEL? ) He claimed he had proof MMT is wrong: if the Fed doubles the base then ipso facto nominal GDP must double and ipso facto MMT is wrong. Well, the Fed tripled the base and nominal GDP didn’t budge. In any case, even if that had worked, it is not evidence against MMT. All Sumner did was to string together a series of non-sequiturs.

Sumner’s also behind an inane proposal that the Fed ought to use its demonstrated impotence to target nominal GDP. Right. I wish the Chairman would reduce the earth’s wobble instead.

A reader thought I’d been too hard on him. But in his latest piece, he’s removed all doubt. What puzzles me is that some people seem to take this seriously. I don’t get it.

He argues that with ZIRP and successive rounds of QE the Fed has not been running easy policy, but instead has subjected the economy to extremely tight policy. You see, zero interest rates, trillions of dollars of excess reserves, and goosing Wall Street by over-paying for all the toxic waste that big banks want to unload is NOT easy policy—it is contractionary.

Actually, monetary policy has been highly contractionary since 2008. Yes, interest rates are quite low, and a lot of money has been injected in the economy. But that was also true in the 1930s in the US and in the early 2000s in Japan. It does not indicate easy money…

How does Sumner know? Well, because GDP growth is low. If monetary policy were easy, GDP growth would be high. Indeed, the low interest rates are NOT due to easy monetary policy, but rather result from low GDP growth.

If one uses nominal GDP growth as the criterion, then since 2008 monetary policy has been tighter than any period since the early 1930s. Indeed this is even true if you average NGDP growth and inflation…. The highly contractionary monetary policy of the last five years has dramatically reduced nominal GDP growth, and this has caused much lower interest rates.

Yes, and black is white, and white is black, and down is up and up is down in Sumner’s world. You see, if you use nominal GDP growth as an indicator of the monetary policy stance since Wall Street destroyed the economy back in 2008, then it is tight.

Sumner’s argument utilizes a tight circular logic: the Fed must have caused the recession because GDP fell, and the Fed must have had tight policy since the recession because GDP growth has been sluggish. We will know when policy loosens because GDP growth will accelerate. Until then, blame the Fed.


Now, I’m a skeptic about the Fed’s ability to affect GDP—one way or the other—but at least I understand what tool the Fed has to try to affect it: the overnight interest rate. That rate is NOT determined by nominal GDP growth, although the Fed does react to growth. Low nominal GDP growth usually induces the Fed to lower the overnight rate. The GFC pushed the Fed to go all the way to ZIRP.

On conventional thinking, that’s all the Fed can do—it shot its wad. Chairman Bernanke (wrongly) thought he had more ammunition, hence, he adopted unconventional policy with QE injecting trillions of dollars of excess reserves into banks. As anyone who understands MMT or even endogenous money predicted, the effect of that would be close to nada: ZIRP+QE=Zilch.

(MMTers believe there is a slightly negative bias resulting from lower rates, as Treasury interest spending is somewhat lower than it would have been; further, substituting reserves on bank balance sheets for higher earning Treasuries acts like a small bank tax—which probably is not good for bank lending.)

But no one schooled in economics thinks that nominal GDP growth determines the policy interest rate in a causal sense.

Except Sumner, who apparently believes that low nominal GDP growth causes low interest rates. But even that claim is not quite enough for some new blogosphere branch of “economics” that is called Market Monetarism. (Note even its adherents seem to admit that its arguments have never been held up to any kind of scholarly scrutiny) Nay, not only has the Fed delayed recovery by adopting extremely tight policy since 2008, the Fed—not Wall Street—actually caused the crash in 2008.

It was not a Global Financial Crisis, but rather, it was a decline in—you betcha—nominal GDP growth that caused the crisis. And who should we blame? Why, the Fed, of course, which caused the economic downturn by pushing nominal GDP growth to negative territory. Since, of course, the Fed controls nominal GDP growth through its policy. Whatever that policy is—not interest rates since those are determined by—you betcha, again—nominal GDP growth.

A nice, neat, impenetrable circular argument.

They call it the “monetary disorder view”: The Fed’s tight policy caused the recession, and its tight policy prevents recovery. It’s the Fed coming and going. You see, it is not that the Fed is impotent, but rather that the Fed is super-duper potent.

And for some reason, the Fed is hell-bent on dragging the US economy through the mud. So, don’t you believe Uncle Ben when he says the Fed is doing all it can do to get the economy back on track. Actually, the Fed caused the recession and persists in ultra-tight policy in order to keep us in recession.

And, yes, the Fed keeps the planet wobbling, too.

I can remember when the main critique of the Fed was that it was always hell-bent on causing inflation by pumping too much money into the economy. Now the claim is that the Fed is biased toward running extremely tight policy.

Market Monetarism appears to be based on two Godfathers. The first is Uncle Milty Friedman, hence “Monetarist”. Friedman claimed that the Fed controls the money supply, which determines inflation. All “real” variables are determined by “real” things, leaving money to determine the “non-real” nominal stuff. Aside from short-term temporary fooling, all the Fed does is to determine the rate of inflation. And hence, inflation is always and everywhere the fault of excess money. Who do you blame? The Fed.

To get there, Friedman had to rely on magic: somehow the Fed controls the money supply through control over reserves. And somehow a rise of the money stock causes the flow of spending to rise above capacity, causing inflation. No one believes that magic any more. The so-called deposit multiplier is deader than door-nails and door-knobs. Trillions of reserves don’t generate trillions of deposits nor trillions of dollars of spending; hence, no inflation.

Everyone who understands anything about central banking now accepts that the Fed only controls the overnight interest rate—which is a very weak handle to try to control “the money supply” or spending and inflation—or the earth’s wobble or just about anything else that you might care about. As I’ve argued, we do not really even know if raising that rate will encourage or discourage lending, borrowing, and spending. At best, you’ve got impotence when you most need vigorous virility. And that’s why the Monetarist Bernanke turned to the QE fire hose to spray trillions of reserves at banks. To no avail.

So to resurrect Monetarism, the Market Monetarists adopted another Godfather, Knut Wicksell. He was an interesting fellow, but as I said, once you got Keynes you certainly don’t go back to Wicksell for monetary theory. The erroneous idea is that “real variables” grind out “real interest rates”. Supposedly, the interplay of thrift (from whence comes the supply of some scarce resource called loanable funds) and productivity (that generates a demand for scarce loanable funds) determines a real rate, quite apart from the nominal rate determined by the central bank. According to Wicksell, when the “market rate” exceeds the “real rate” you get recession, while a market rate below the real rate gives you inflationary expansion. All of these “real” variables are unobservables—hidden behind the magician’s cloak.

That notion of a divergence between the “real” rate and the “market rate” is what puts the Market into Market Monetarism. And it is that Wicksellian magic that is behind Sumner’s cryptic argument:

The highly contractionary monetary policy of the last five years has dramatically reduced nominal GDP growth, and this has caused much lower interest rates. In supply and demand terms, the investment schedule or the demand for loanable funds schedule has shifted to the left, depressing interest rates and also the quantity of money borrowed. If this had been caused by an easy money policy, then the low interest rates would be associated with higher levels of borrowing and investment.

The Fed caused investment to collapse—which lowered GDP as we moved into the Great Recession—and as the demand for loanable funds fell, they became less scarce, causing interest rates to fall. It wasn’t the Fed that lowered rates, it was the decline in the demand for loanable funds.

Why did that happen? Because monetary policy was too tight. How do we know? Well, we cannot look to either monetary aggregates (A.K.A. money supply or base—the Friedmanian Monetarist view) or to interest rates (everybody else). We must look at GDP. Faltering GDP growth is evidence that monetary policy is too tight. Sumner again:

As a general rule higher interest rates are associated with robust economic growth (nominal) and low rates reflect weak growth. It’s a mistake to evaluate the impact of interest rates on the economy. What economists need to do is evaluate the impact of the factors that cause interest rates to change. Especially nominal GDP growth.

As Ronnie Reagan says, “There you go again”.

If the Fed does not implement money policy by controlling the money supply (Friedman) or the overnight interest rate (Everybody else), then how does it do it? Well, by the Expectations Fairy. Here’s Sumner:

Monetary aggregates are neither good indicators of the stance of monetary policy, nor good policy targets. Rather than assume current changes in (the money supply) affect future (aggregate demand) with long and variable lags, I assume current changes in the expected future path of (the money supply) affect current (aggregate demand), with almost no lag at all.

It is the expectation that future money supply growth will be below future demand for money supply that causes an immediate collapse in aggregate demand. That’s how the “market” modifies old-style “monetarism” to give us the new whiz-bang Market Monetarism.

So, the crash in 2008 was caused by the market’s belief that the Fed was going to engineer a reduction in the future growth of the money supply. However, that reduction would not be accomplished by actually reducing money supply growth (which the Fed cannot accomplish) nor by raising interest rates (which are determined by GDP growth).

No, the Fed would lower future growth of the money supply by whispering into the Expectation Fairy’s ear that it would tighten policy, and that would then lower GDP growth today—which would lower rates now. Hence, the low rates are actually evidence that policy is tight, not loose.

It is evidence that that Expectation Fairy is flitting about, warning every one of the Fed’s intention to lower GDP growth. As Sumner likes to say, monetary policy operates with “long and variable leads” (that’s the Fairy’s whispers) and affects aggregate demand “with almost no lag at all”. If we could just get that Expectation Fairy to start whispering about future monetary ease, then GDP growth would rise and interest rates would rise—which would be evidence of easy money policy.

If you are following all of this, I feel your pain—as President Bubba used to say.

29 responses to “The Circular Logic Behind Scott Sumner’s Claim That the Fed’s Policy is Contractionary

  1. Face… Palm….

  2. Great piece Randy. Unfortunately, the MMers are just part of the problem. Their picture is not all that different from Krugman’s views on the need for the Fed to “credibly commit to be irresponsible.”

    Mainstream macroeconomists believe that there is some kind of short-term correlation between inflation and the rate of real growth – or maybe its just between changes in the rate of inflation and changes in the rate of real growth. They also believe that The Market behaves like an ideally rational agent who has internalized the macroeconomic theories of contemporary macroeconomists. So they think that if The Market comes to expect higher inflation it will also expect higher growth.

    They also believe The Markets form their expectations of the future path of inflation based on Fed statements about the future path of inflation. Putting it all together, they think the Fed has to signal that there will be higher inflation than previously predicted. As a result the The Market will expect higher inflation, and thereby come to expect higher growth. And if The Market expects higher growth, The Market will invest more in growth to bring its behavior in line with expectations .

    The only difference between the Market Monetarists and the rest of them is that the the Market Monetarists think the Fed should target a nominal GDP path, rather than an inflation path. Otherwise the basic picture is very similar.

    I’m not sure why, but in recent weeks mainstream macro seems to have lurched in a conservative direction. By “conservative” I mean that they seem to have doubled down on embracing the established theories and defending their own. There was an incredible surge of support for Eugene Fama and rational expectations from all corners when he won the Nobel. I think mainstream macro is running scared due to the increasing volume of critiques of mainstream macro as a pseudo-science. So they are circling the wagons.

    • >Krugman’s views on the need for the Fed to “credibly commit to be irresponsible.”

      In other words, his solution to global economic crisis is that Ben Bernanke should wear hawaiian t-shirt and play banjo. How ridicilous is that?

    • For some odd reason mainstream economists ignore wealth completely in their analysis.

      It is really quite simple: certain amounts of net wealth in the private sector generate certain amounts of nominal spending in the economy. How much depends on saving desires. That is how aggregate demand gets determined, and they do not get it.

  3. Here’s another where Sumner’s thinking goes cockeyed:

    He sez:

    1. Fiscal policy only works when the Fed’s incompetent.

    2. The Fed’s incompetent.

    So by his reasoning, fiscal policy works.

    But he says it doesn’t, because it’s always trumped by the Fed.

    But the Fed’s incompetent, so…

    Repeat loop.

  4. And I’m still waiting for Sumner to reply to his own compatriot, David Beckworth, who pretty resoundingly obliterates Sumner’s thinking:

  5. I’m impressed you made it that far through a Sumner post. I cant read the guy anymore, so I let Mike Sax read him and I just get the juicy parts from Mike . I suggested to Mike that we need a site called “Stupid Things Scott Sumner Says”. It would have lots of material for sure.

    You should have seen the discussion Scott had months ago with Mike Sankowski of MR (and it was revived recently by Mike) regarding a futures market for NGDP Futures. Mike is a trader and has actually derived some futures contracts. He looked at it and said it would just be a boondoggle with Hedgies just making hundreds of billions off the Fed with no upside to our real economy but Sumner was having none of it. He just KNOWS that his NGDP idea cant miss and sees these futures contracts as a way to realize this idea. He’s so pigheaded he wont listen to professional traders, the guys who would structure these contracts, when they tell him its a no go.

    To quote Bugs Bunny ” Geeeeez, what a maroooon!”

  6. If we put a train on a closed loop of track just as long as the train, and hitch the engine to the caboose (last car for the young set), we will find that the caboose is pushing the train which eventually gets around to pushing the engine (or maybe the caboose is pulling the engine). In either case, we could take the engine out of the loop altogether and let the caboose do its job, saving the cost of manning the train with an engineer. Anyone want to help start a RR?

  7. I don’t find Scott’s logic circular in the least. Instead, I feel you have mis-characterized it. Consider my interpretation of his logic:

    (1) Shock makes money tight + (2) Fed *can* offset but doesn’t => Fed implicitly allows NGDP to contract => monetary policy is too tight

    So frequently people describe monetary policy as ‘loose’ or ‘tight’ with respect to historical policy norms. This makes no sense. If the policy is designed to achieve a target (even simply the current dual-mandate) the policy should be defined by where it leads with respect to the target.

    In other words, if I am flying a kite straight to the sun, and a strong wind blows my kite to the left of my intended direction, then I must steer it to the right in order to maintain course. So I begin to steer the kite to the right, more to the right than I ever have done before, but alas the prevailing wind still has my kite off course to the left of center. Fact is, I am not steering my kite to the right ‘enough’, and in so doing I am accommodating the new errant direction of my kite.

    If one agrees that the Fed’s current dual-mandate targets are desirable, then they must also feel that the Fed should loosen policy until those targets are arrived at.

    Now, we can argue all day long whether the Fed is impotent (I don’t think it is), but that is separate from the fundamental line of reasoning. The fact remains that the Fed being under target is resounding and absolute proof that policy is not loose enough to accommodate the target; whether policy can be ‘loose enough’ has no bearing on this.

    • “The fact remains that the Fed being under target is resounding and absolute proof that policy is not loose enough to accommodate the target”

      You start with the assumption that the Fed can control nominal GDP absolutely and then conclude that therefore if nominal GDP isn’t where it should be this means that the Fed isn’t doing enough.

      Why not start with the assumption that rain-dances and animal sacrifices cause rain to fall, and so if there is a drought this is evidence that we are not doing enough rain-dances and sacrifices?

      • It seems that you are conflating my interpretation of Scott’s logic with my assertion: “The fact remains that the Fed being under target is resounding and absolute proof that policy is not loose enough to accommodate the target”

        – The Fed has an inflation target of 2%
        – Loosening monetary policy aims to increase inflation, and vice versa
        – Inflation has been lower than 2% since the beginning of 2012
        => Monetary policy is not loose enough to achieve its stated target (or unemployment for that matter)

        So if one believes the Fed’s targets are just, then so too must they believe that the Fed should loosen policy to arrive at those targets.

        Now, with regard to Scott’s logic you are correct that the inherent assumption is that the Fed can control NGDP. That reality of this assumption wasn’t the point of my post. I would argue that, yes, the Fed can influence NGDP, more or less- as to ‘absolutely’, no I would not make such a claim.

        • if you assume that NGDP (or inflation) below target is always evidence that the Fed isn’t doing enough, then you must assume that the Fed has absolute control over NGDP (or inflation). If you assume that the Fed has absolute control over NGDP (or inflation) then you can assume that NGDP (or inflation) below target is always evidence that the Fed isn’t doing enough.

          However, those that live in the real world don’t make such ridiculous assumptions.

          • Synonyms of ‘absolute’ per Merriam: complete, total, utter, out-and-out, outright, entire, perfect, pure, decided

            No, I do not believe the Fed can dial in inflation to a basis point on a minute-to-minute basis via perfect control. The fact should also be pointed out that such perfect control is not a necessary assumption for the stated logic. In reality, the core assumption must be that the Fed can exert general influence in the upward or downward movement of inflation…. and don’t you truly believe that is reasonable? Isn’t it true that you feel the Fed could drive up inflation if they really really really wanted to?

            It would be more defensible if you were to claim that the Fed lacks surgical precision and fears brute force tactics could push inflation too high.

    • L. Randall Wray

      Dustin: what does the following mean (if anything):
      (1) Shock makes money tight + (2) Fed *can* offset but doesn’t => Fed implicitly allows NGDP to contract => monetary policy is too tight

      Define “shock”; Define “money tight”; define “Fed can offset”; define Fed allows GDP to contract.

      Or is it just magic and circular reasoning? Is a shock a decline in GDP? what caused it? Is tight money a decline in GDP? (circular) or Fed raised rates? (really? When?) What caused the “shock”? What did the Fed do to allow GDP to contract?

      • Well please let me reiterate that I do not speak for Scott, only my understanding of his position, which, based upon said understanding, I largely agree with; wouldn’t it be delicious irony if he felt my interpretation was way off base? 🙂

        Now to the definitions you request – I would invoke the generally accepted and oft used definitions:

        – Shock: ‘Change in exogenous factors’ as per wiki, which is well-said
        – Tight money: This is a market condition where credit is difficult to secure and results from a negative shock. This is not a policy decision.
        – Fed can offset: Fed has the ability to slacken the tightness of money in the market through monetary policy
        – Fed allows NGDP to contract: This is a simple logical result of the above premise that the Fed can slacken tight money conditions but doesn’t due to an irrational fear of actually hitting the 2% inflation target.

        Even after the subprime mortgage crisis (shock), credit tightened (tight money), IMF predict mild recession ( Unfortunately, that ‘mild’ became ‘Great’ because not enough was done to ease tight money conditions. I don’t at all read Scott’s position as NGDPLT will mitigate the onset of recessions. Rather, NGDPLT would moderate recessions from devolving into Great Recessions.

        There is nothing remotely circular about this logic.

        • L. Randall Wray

          Sorry Dustin that is a handwave. First, “subprime shock” is not the Market Monetarism vector; second you have not told us how the Fed tightened credit–give us the mechanism; tell how how the Fed “let” NGDP fall; and how it could have responded differently to keep GDP from falling. Handwaves don’t cut it.

        • By what mechanism does the Fed “ease tight money conditions”? And how could the Fed possibly do more than it is already doing (which hasn’t done anything at all really besides inflate certain asset classes)? The Fed has a very limited tool-kit which has only an indirect influence on inflation and GDP.

          If you think about a “shock” like the GFC, caused in large part by sub-prime lending and over-indebtedness, why on earth would we want to jump-start credit growth to get the economy going again? If the banks want to lend, they will lend, it has nothing to do with the Fed. You’re attributing this massive power to the Fed that it simply does not have, our fate sadly lies with the other side of the government.

      • money tight: the supply of base money is insufficient to fulfill demand for money. This results in lower NGDP growth.
        shock: anything that moves either demand or supply. In this case, I assume that something caused demand for base money to soar. Possibly some misbehaving banks?
        Fed can offset: All the fed needs to do is to expand the money supply to match demand. And to make people expect that the money supply will match demand in the future, too.
        Fed allows NGDP to contract: By not offsetting the higher demand for base money, M*V decreases, thus bringing NGDP down.

        With the above definitions we can translate:

        (1) Something caused demand for base money to soar + (2) The Fed did not expand money supply enough, or signaled that future money supply would not expand enough => Unmet demand for money means M*V goes down => this is how we know monetary policy has been too tight.

        One important point is that (expected) NGDP growth is how we know that policy has been tight. Because NGDP plunged and didn’t recover, we know policy was extremely tight.

        The other important point is that what caused the shock is irrelevant. Real GDP might be impaired if there is a supply shock, but the massive unemployment seen need not happen.

        PS: The fed refused to cut rates after Lehman failed in sept 2008. Moreover, it implemented the IOER policy *after* Lehman failed, 3 weeks or so. It increased the rate paid in October, and then again in November. The Fed explicitly wanted to raise money demand in the middle of a massive crash in inflation expectations (as measured by the TIPS spread).

        • L. Randall Wray

          sorry felipe, more hand waving and circular argument. Low GDP growth DOES NOT show us that monetary policy was tight; it is your DEFINITION of tight monetary policy with no explanation of how the Fed did it.

          • Almost, but not quite. Low NGDP *is how we know* money has been tight.

            The point is that the Fed controls M. If V has crashed (and it doesn’t matter why), then the Fed can increase M as much as needed to prevent a fall in M*V.
            The Fed need not have caused the crash in NGDP, but it allowed it to happen. And that is why it is the Fed’s fault. The Fed failed to offset the increase in money demand with an increase in money supply. Not only that, but the Fed increased money demand by several measures (including the IOER policy).

  8. Sumner has repeatedly posted over the last couple of months that the “Keynesians” are all silent now that his NGDP targeting is working its magic and creating economic growth. In fact he’s made clear that he is pleased with current growth levels, yet suddenly monetary policy is too tight and keeping growth too low?

  9. This market monetarism never made any sence to me. If you start asking where is your transmission channel from reserves to real economy then they are kind of using Friedman’s magic terms, hot potatos and all that.

  10. Just one little mistake :

    You’re quoting Sumner’s article while lining to Brad De Long’s. here’s the correct hyperlink :

  11. Very interesting, thanks. I learned a lot from the Wicksell / “Natural Rate” bits. But this opponent is unworthy of your steel. Irrelevance and oblivion will wash over Mr Sumner soon enough.


  12. I read both sites and generally agree with the MMT view, and it seems clear to me, but maybe not to Sumner that conventional Fed actions do not have a transmission mechanism to higher NGDP but I think Scott might ask what if the Fed buys unconventional assets? They could buy any and all assets in the USA if they wanted to and would you still say that it wouldn’t have an affect on inflation or consumption? And if that is true…can’t you draw a logical walk to the Fed focusing on raising NGDP expectations without actually going through with (or maybe doing just enough unconventional purchases to make people believe) the Nuclear Option Purchasing? Unwinding the unconventional purchases I guess would be a different matter as you couldn’t just let them mature off the books…It still seems like some hocus pocus to me, and could the Fed ever really get people to believe it would follow through? I don’t think so, but they do, and they want to try it…

    • What’s the difference between the Fed buying unconventional assets and fiscal policy? The nuclear option appears to be a way to get some form of fiscal policy happening while maintaining a charade that MMers were right about the usefulness of monetary policy all along (by calling it unconventional monetary policy). But if there is no difference in accounting terms I don’t see why there would be any difference in the results.

      Of course anybody should be massively worried by the likely corruption which could easily come out of fiscal policy decisions being made by an undemocratically elected committee. I think we saw the sort of thing which can happen during the US bailout process, and subsequent Fed audit, and some of it looked pretty dodgy.

      • L. Randall Wray

        Nickzer: Treasury SPENDS, Fed LENDS. Fed purchases only change the form of your assets–yes, Fed normally increases your portfolio’s liquidity. Treasury SPENDING increases your income and wealth.

        Big difference.