Scott Sumner responds on ngdp

Here is Scott’s lengthy response, do read the whole thing.

While I definitely favor having an ngdp futures market, and think it might improve monetary policy, its existence would not change (at all) any of the basic issues in my previous post.  Closer to the central point I think is Scott’s claim: “Any “real shock” that reduces NGDP expectations because the Fed responded passively is also a monetary shock.”  For me that is a real shock with insufficient monetary accommodation, not a “real shock,” as Scott gave it quotation marks, and also not a monetary shock.  I prefer not to smush real and monetary shocks together in that fashion, and I think some ngdp theorists are trying to claim an explanatory victory by fiat by doing so.  I do not think this matters for policy, and as I’ve stated I am very sympathetic to market monetarist recommendations.  But in terms of explaining downturns, again, I think they are trying to claim some victories by fiat.


Does the view that "the Fed is never passive" help?

I think this is the exact sort of sophistry that Tyler is accusing the NGDP crowd of cutting caught up in:

Isn't it much simpler (and more accurate) to admit that there is an economy in the world in which people are buying and selling things with money, and also an institution called the Fed, that responds (or sometimes does not respond) to developments in that economy? And that the two are analytically separable?

Sumner's argument-by-counterintuitive-definition method sometimes reminds me of Ronald Dworkin's books on legal theory.

But the Fed and the real economy interact with one another. All.the.time. Business plans reflect expected Fed policy, which itself is anticipating the future path of the real economy. The back-and-forth never ends. The Fed is always acting.

I think the confusion is more like circularity than tautology.

Tightness or looseness of the money supply can be thought of as the difference between the fed funds rate and the natural Wicksellian interest rate (though NGDP expectations are better). That is the effective number and so it is the relevant number. Nominal interest rates in the abstract don't matter- you can't tell if monetary policy is loose or tight based on nominal interest rates alone.

So if this number (Wicksellian interest rate-Fed funds rate) is dropping, the Fed is effectively sharply tightening monetary policy. If it does not adequately accommodate the real shock by expanding the supply of money, the effect is that the Fed is allowing the money supply to tighten dramatically.

Let's say a person is driving and they announce the amount of accelerator pedal deflection they are going to use. Then they start going up a hill. This is a "real shock" that causes the car to start going slower than they want it to go. If the car stalls out because the accelerator pedal is not deflected to a greater angle, is that an act of omission where the driver failed to adequately accommodate a real shock to vehicle speed? Would you say the hill caused the stall and not the driver?

Why limit that to the Fed? If Congress fails to enact a tax cut or a tax increase, are they not, in this sense, "never passive"? I'm not convinced by that argument because it equally applies to every actor (or non-actor).

Yeah. Congress should cut taxes and increase spending today. Then, in six months' time, if the economy bounces back, they should raise taxes and cut spending. That sounds pretty workable.

But you weren't talking about the relative nimbleness of the institutions in practical terms. You were making a metaphysical point about the nature of the Fed, which would apply equally to Congress, as the resident Nabakov fan pointed out with his/her perfectly sound reductio argument.

I guess I'd say it this way: Congress is usually passive. It only moves a couple times a year. That's the way it's made. Not the Fed.

I wasn't going for metaphysical.

"I guess I’d say it this way: Congress is usually passive. It only moves a couple times a year. That’s the way it’s made. Not the Fed".

I guess that makes all the difference--really?. How many times did the Fed move last year?

How many times did Yellen open her mouth?

Not as many times as Congressmen and women and the President, I'm sure. How about a serious answer?

Besides, I thought the point was she was acting by keeping her mouth shut.

It's a matter of credibility. Politicians promise all kinds of things. All of these promises, adjusted for the likelihood they will be enacted, are also factored into people's views.

Presumably, the Fed can deliver on its promises, so they get more weight.

Also, yes. Choosing to do nothing is not passivity. If we're driving on a long straight road, and you notice that I haven't moved the steering wheel in a while, you might make the mistake of thinking I'm just sitting there passively.

So, everyone who drives a car is not passive. Yellen drives a car. Therefore, she is not passive. I like these stories. Tell me another one.

I think the difference between the Fed and other actors is that (1) changes in money demand itself don't matter, only mismatches between money demand and money supply, and (2) the Fed can put the money supply curve anywhere; it is not constrained by some marginal cost of money curve. The Fed's job is to match money supply to money demand, either by changing money supply or influencing money demand. Given that there is no marginal cost curve that sets money supply, there really is no distinction between saying the mismatch was caused by the Fed not putting money supply in the right place vs. the mismatch was caused by the money demand curve shifting and the Fed just didn't accommodate that.

To Scott Sumner, all nGDP is the same. The only thing that matters is the aggregate quantity, never its composition. In his model, agents can never lose faith in a system that is committed to achieving its level target, even if they observe banks lending people money to dig a hole and fill it up again. There is no point agruing with him, he is a fundamentalist.

He actually has specifically mentioned that he thinks the banks should have just been allowed to fail and that with stable NGDP there would not have been a macro instability downside. The current policy is certainly much more friendly to 'private gains socialized losses'.

Bingo, JVM.

This point was brought about in Yudkowsky's post on market monetarism as well.

I vaguely remember a financial institution's slogan "There's always a bull market somewhere". In a market where the CB is targeting NGDP, there is always income somewhere.

I can envisage the psychology of the investors being completely different. When a bank fails, then some other bank might succeed, or some other store might succeed or something.

Currently the thinking is
" Ok, bank X has failed. That is 200 million dollars of cash flow that is not going to its employees and shareholders. Who will suffer and how much? Will it cascade?"

In the NGDP targeting environment it will be more like
" Ok, bank X has failed. That is 200 million dollars of cash flow that is not going to its employees and shareholders. Who will suffer and how much? Who will gain this 200 million dollars of cash flow and how much will they gain? What will they do with it?"

By reducing monetary shocks to a minimum, the environment completely changes, in mostly a positive direction.

To my knowledge, NGDPLT is supposed to work with the banks intact as well. And Scott does not favor guaranteed minimum incomes, so I don't know what his justification for helicopter money could be.

Lots of semantic games here.

> I prefer not to smush real and monetary shocks together in that fashion, and I think some ngdp theorists are trying to claim an explanatory victory by fiat by doing so.

It seems like you are talking past each other. There are two things to be explained. The fall in RGDP and mass unemployment. Using 2008 as an example I think Scott's claim would be as follows:

If we had (a) a real shock (housing crash / financial destabilization) and (b) a nominal shock (falling NGDP), we would get mass unemployment and fall in RGDP as we observed.

If we had (a) a real shock (housing crash / financial destabilization) and (b) no nominal shock (falling NGDP), we get no mass unemployment and but RGDP would fall at least relative to trend.

If we had (a) no real shock (housing crash / financial destabilization) and (b) a nominal shock (falling NGDP), we get mass unemployment and and RGDP would fall as a result.

If we had (a) no real shock (housing crash / financial destabilization) and (b) no nominal shock (falling NGDP), we get stability.

If mass unemployment always follows nominal shocks and is indifferent to real shocks, then real shocks don't actually have explanatory power in predicting mass unemployment and by Occam's razor should be eliminated. Sure, a real shock might trigger a nominal shock by causing a failed accommodation, but this is an accident of the current regime, so it is foolish to look for the "real cause" of mass unemployment as the real shock; that's only a problem due to the existing policy climate.

Are the MMers trying to 'declare victory by fiat' in explaining recessions? MMers would certainly agree that there are other causes of real shocks, but for mass unemployment business cycle recessions, they really are claiming that these are caused by monetary policy alone. "Recessions are always and everywhere a monetary phenomenon", albeit defining 'recession' as mass unemployment/deleveraging recession rather than the more traditional "2 quarters of falling RGDP."

There's also a semantic game going on about monetary policy changes. To MMers, changes to NGDP growth path are monetary policy changes, whereas Tyler it sounds like you consider the fed "taking action" to be a policy change; so in 2008 for MMers the fed massively tightened, whereas for you they "failed to adequately accommodate". This is purely semantic, but to me it seems sort of reasonable to talk about the outcome rather than the mechanism.

@JVM: " I think Scott’s claim would be as follows" - BZZT! Wrong.

Let Sumner answer, not you. If you read his blog you'll find, like any fine economist, he's very inconsistent, making him hard to pin down. The fact that he can't understand a fellow economist like TC, who is steeped in the same theory, is evidence of that. Just the other day Sumner (who to his credit does not moderate his blog and answers everybody, even abusive posters of which I'm not) said to me: "Ray, The fact that NGDP collapsed does not prove that it caused a collapse in RGDP, I agree with that claim. But of course there is a mountain of evidence that NGDP causes RGDP, which you choose to ignore. Bernanke’s written entire books explaining how tight money caused the Great Depression." ( So, according to Sumner, a collapse in NGDP is not necessarily a collapse in RGDP. But logically, this can be due to deflation only, since NGDP = RGPD + Inflation, would imply negative inflation or deflation and RGDP growth while NGDP collapses negative, which is possible, as in the late 19th century. Now Sumner switches gears and cites a 'mountain' of evidence, specifically the Great Depression, and 'tight money' which is probably a reference to the Managed Gold Standard (MGS) (the US/UK economies improved when they went off the gold standard as managed by central banks, which would have just been a coincidence IMO as other economies, like Argentina, did not improve when they went off the gold standard, and in fact got worse, see graphic here:

In short, Sumner's hard to pin down and you cannot put words in his mouth.

Idiot, Sumner said that the mere fact that NGDP fell does not mean that CAUSED a collapse in RGDP. The quote is about causality but of course your tiny IQ 120 brain cannot understand even that simple statement.

Sumner is incredibly consistent, you just don't understand anything that you read.

If I understand Tyler, he is asking, "If the money demand curve shifts, and not because of anything the Fed said or did, and the Fed fails to shift the money supply curve, did the Fed cause the shock or just fail to accommodate it?" I take "real shock" to AD to mean shift of money demand curve or, equivalently change in money velocity, not due to Fed forward guidance or other signalling.

This seems largely a semantic question, but consider the following: If a road curves to the left, the driver keeps steering straight ahead, and the car runs off the right side of the road, did the driver drive the car off the road or merely fail to offset the road designers' misplacement of the road? The Fed has a monopoly on the (base) money supply and, therefore, it seems reasonable to hold the Fed accountable for putting the money supply curve in the right place. After all, if we adopted a fixed rule, such as the Taylor rule, and a recession or inflation resulted from some imperfection in the rule, wouldn't we blame the flawed rule rather than the "real shock" that exposed the rule's flaws? If we replace the rule with Fed discretion, why should we not similarly blame the Fed's flawed use of discretion?

Finally, the absence of an NGDP futures market does seem to matter. Without such a market, the Fed could have a legitimate excuse that it didn't realize that the money demand curve shifted. In the car analogy, without a clear windshield, the driver could say that he didn't realize that the road turned. Even in this case, however, we should still say that the Great Recession (or Great Depression or Great Inflation) was caused by the Fed's inability to see changes in money demand rather than the changes in money demand themselves. There is nothing that prevents the Fed from shifting money supply to match money demand and, importantly, no one else is allowed to change money supply --- the Fed owns that.

@BC -firstly, money is largely neutral (Bernanke's 2002 FAVAR paper) so all of what you said was moot, but if you read the Fed minutes just after the Sept 2008 collapse indeed, as you imply, the Fed was steering with the wrong information, as they believed inflation was a possibility (read the minutes for a complete description, there's sentence to that effect in the minutes) and did not cut rates, though they promised to temporarily increase liquidity in response to the crisis. Minutes are out in the public domain now. Sumner is right in that the Fed appears negligent after Lehman's collapse, but given money is neutral the entire debate is nonsense.

Bernanke does not think money is neutral and the paper you cite does not say that. Bernanke thinks monetary policy caused the great depression!

I think you are the one "smooshing things together" when you say it will be a risk based recession. If there is some risk that China's economy could nosedive due to structural factors, and also risk that central banks around the world are currently using too tight policy and in operating in policy regimes that will fail to contain (and perhaps exacerbate) the damage, then it is not smooshing things together to say that monetary policy is the big risk here.

"I prefer not to smush real and monetary shocks together in that fashion, and I think some ngdp theorists are trying to claim an explanatory victory by fiat by doing so."

^^ This Sumnerizes - er, I mean summarizes - market monetarism to me. But what do I know, I'm just the peanut gallery.

""what do I know, I’m just the peanut gallery."


FRBSL folks have pointed out...

No, monetary expansion need not EVER lead to... if... (see FRBSL link above).

"I prefer not to smush real and monetary shocks together in that fashion."

I am curious as to how Tyler would define a "monetary shock" if not relative to how well it offsets "real shocks" to AD or NGDP, given that "real shocks" happen all the time. For example, Scott describes interest rate hikes in the 70s and their failure to tame inflation here: []. Would Tyler say that the interest rate hikes were negative (recession-inducing) monetary shocks that happened to be (more than) offset by inflatonary "real shocks"? It seems like one would have to say that if one viewed "real shocks" and "monetary shocks" as separable, independent things. In fact, in the 1970s, people might very well have said that the Fed was not causing inflation, just failing to offset all the inflationary "real shocks". Today, I believe the consensus is that loose monetary policy caused the Great Inflation, i.e., the inflationary "monetary shock" that caused it was the *net effect* of the Fed's moves and the "real shocks" taken together.

@BC - now reconsider your response given money is neutral. Open your eyes to new ways of thinking, like Fisher Black did. 1970s oil crisis caused a supply shock, and panic buying made it worse (oil has a few percentage points difference in supply/demand to make a glut/shortage,and the economy back then was more oil reliant than now), inflation followed for a decade until the panic subsided; btw Volcker Fed actually cut rates and inflation *fell*, sic, and raised rates and inflation *rose*, sic, contrary to the myth that Volcker 'broke the back of inflation', read the data yourself and see. Economy is non-linear, stuff happens, and the men/women at the Fed controls have very little if any effect (3.2% to 13.2% out of 100% says Bernanke in his 2002 FAVAR econometrics paper).

Sometimes it is clearly nonsensical to say that a lack of response to a shock is not the cause of a problem. And the existence of prediction tools like the NGDP market makes a difference in the analysis.

For example, if I am driving and the road turns, if I fail to turn with it and crash as a result I think it is pretty fair to say that I caused the crash. However, if a meteor strikes my car and I fail to avoid it then it it may not be fair to say I am the cause. But If I know in advance where the meteor is going to land that changes the equation again.

"Scott’s claim: 'Any “real shock” that reduces NGDP expectations because the Fed responded passively is also a monetary shock.' For me that is a real shock with insufficient monetary accommodation, . . . not a monetary shock."

Do you not see that this is a purely verbal dispute? Everyone should agree that, if monetary policy suddenly, unexpectedly tightens, that is a *monetary shock*. So the issue is: In the scenario you and Scott are imagining, *did monetary policy tighten*?

You say 'no': monetary policy just didn't loosen as much as was desirable; the Fed did nothing, and so did not change monetary policy, and so did not *tighten*. Scott says 'yes': the stance of monetary policy is to be judged relative to the expected path of NGDP one or two years into the future; if the Fed allows or causes this to *fall*, it has *tightened monetary policy*.

Either way of defining 'loosening/tightening', or 'the stance of monetary policy', seems acceptable, but we should decide between them. If we don't, let's at least recognize the ambiguity, and not confuse ourselves by taking a merely verbal wrangle to be a substantive disagreement.

There is a further issue about *your* way of talking: your notion of *doing nothing* is not perfectly clear. The Fed conducts routine operations; every business day it is, literally, "doing something." Precisely which of these routine activities are you brushing aside as not *really* doing anything (for purposes of your conception)?

Tyler: "While I definitely favor having an ngdp futures market, and think it might improve monetary policy, its existence would not change (at all) any of the basic issues in my previous post."

I really don't understand this comment. Seemed to me that one of the main "basic issues" in his previous post was his challenge to MMs to state their preferred EX-ANTE measure of the stance of monetary policy (rather than ex-post measures or statements along the lines of "NGDP was down, therefore monetary policy was tight..." etc)

So surely having an NDGP futures market and using this as the ex-ante measure of the stance of monetary policy should completely address this key "basic issue".

(Not saying that there aren't other issues cited by Tyler that NGDP futires market wouldn't address)

Well in relation to the current situation, we are not talking about a "passive" Fed, watching blithely from the stands as the economy suffers some kind of real shock. The actually are an active Fed - but in the wrong direction. They just increased interest rates and told the market that they planned to increase some more. And what else is increasing interest rates but monetary tightening? And now we see the market reacting, and the risk of a recession increasing. What more proof do you want that the driver is the Fed for goodness sake. There was no "real" shocks out there unless you count the fall in oil prices, which is unambiguously good for the US.

I a beginning to think Tyler is trolling Scott - willfully misunderstanding him and engaging in semantics rather than looking at the substance of the argument.

"But in terms of explaining downturns, again, I think they are trying to claim some victories by fiat."

By the way, Sumner and the MMs are not the only ones to declare that AD and NGDP fluctuations are purely monetary, at least away from the zero-rate bound. Here is Krugman from 1997 []*: "[demand-side] recessions do happen. However, such slumps are essentially monetary...they can usually be cured by issuing more money--full stop, end of story." Here is Krugman again, one month later []: "if you want a simple model for predicting the unemployment rate in the United States over the next few years, here it is: It will be what Greenspan wants it to be, plus or minus a random error reflecting the fact that he is not quite God."

MM is not really new, just reminding people of what they seemed to have forgotten that they once believed (and haven't since repudiated, just ignored). The one new part is looking to markets for information about expectations, hence the *market* in market monetarism. Sumner's most convincing writings draw mainly from Mishkin's textbook, Bernanke circa 2003, New Keynesians before 2007, AS-AD graph, plus graph of money supply vs money demand.

*Aside: that first Krugman reference is my favorite Krugman article. Packed with lots of stuff including: robots "replacing" humans, "loss" of manufacturing jobs, and pure empiricism as accidental theory.

IOW, Tyler, It's a tautology!!

I do not drive a Fiat, but I do believe the Fed should print a lot of money that will be used to pay off US debt and to effectively finance a FICA tax holiday. Let it rip.

The US economy grew by 8.5% real in 1965.

The US economy was much more characterized by structural impediments in the 1960s, but we had a growth-oriented Fed.

The Fed should gun the money presses so hard that fancy restaurants are hiring fat waitresses and we are wiping our rear apertures with Benjamin Franklins.

After that perhaps we can listen to sanctimonious sermonettes about inflation by little boys in short pants.

Tyler -- I mainly agree, obviously it's hard to separate out monetary policy effects from "real" effects, so here's a simple market monetarist rule:

If NGDP was below trend, the CB made growth/employment worse by not bringing NGDP up to trend.

Since we agree the CB has the power to do this (unlike some of the more extreme Keynesians, or, God help us, MMTers), that's a pretty cut and dried rule by which to make predictions -- if US NGDP is going to be below trend in 2016 (Scott is guessing around 3%), then policy is too tight to the extent they miss trend.

Unfortunately I doubt we can ever quite parse out how much monetary policy was at fault for any given outcome to any great level of precision or agreement, but if we have a severe recession in the second half of 2016, I at least for one would not blame a lot of it on current Fed policy, especially if there are easily identifiable real shocks (although, of course, the Fed's future response might be inadequate to those future conditions, which is a different question).

Comments for this post are closed