Scott Sumner  

Would NGDP targeting lead to excessive inflation?

MSNBC Host Admits We Didn't Ha... We're Number 44! We're Number ...

In talking to other economists I often hear concerns that NGDP targeting could occasionally lead to excessive inflation. Oddly, I don't hear the opposite complaint, although there would be just as many periods of below average inflation as above average inflation. That in itself is quite revealing.

In papers like this one I have discussed why NGDP stability is better than inflation stability. I'd also encourage people to look at George Selgin's work on why inflation is the wrong variable to stabilize.

In this post, however, I'd like to address these concerns from a slightly different perspective. I think that many economists overestimate the extent to which NGDP targeting would lead to inflation instability. The most commonly cited objection is that if the economy fell into recession, a 4% or 5% NGDP target would require a relatively high inflation rate. Inflation would be fairly low under NGDP targeting, on average, but could shoot up to 5% or more during brief periods of recession. That is the fear.

To explain what's wrong with this fear, I'd like to describe three possible recessionary outlooks. Each one will be designated with a mammal, for reasons I'll explain later:

Demand shock recession with procyclical inflation (horse)

Demand shock recession with countercyclical inflation (unicorn)

Supply shock recession with countercyclical inflation (zebra)

In America, demand side recessions with procyclical inflation are far and away the most common. The rate of inflation usually falls during recessions. Occasionally, as in 1974 and (perhaps) 1980, we have a recession where the rate of inflation rises. Those are supply shock recessions with countercyclical inflation. And there are no demand side recessions where inflation rises, because such a thing is impossible.

In America, horses are very common, zebra are pretty rare, and unicorns don't exist at all. The fear that NGDP targeting would lead to lots of recessions with countercyclical inflation is akin to the fear that NGDP targeting would lead to there suddenly being lots of zebras and unicorns all over the place. It's very unlikely to happen.

So what's wrong with the reasoning process of those who worry that we'd see lots of cases of recession with unacceptably high inflation? Without realizing it, they are making a logical error in their thought process, holding two incompatible views at the same time:

1. NGDP is no panacea; hence we'd still have the same old business cycle to deal with.

2. Under NGDP targeting, NGDP growth would be pretty stable.

The problem is that today the overwhelming majority of business cycles are associated with strong shocks to NGDP growth. So if you accept assumption #2, then you can no longer accept assumption #1.

Since most people are skeptical about panaceas, let's look at if from the opposite perspective. Let's assume #1, that we have almost as many business cycles as before, even under NGDP targeting. Why would that be the case?

The answer is simple. Recall that NGDP targeting doesn't cause supply shocks, I think everyone agrees on that point. Rather if we continue to have business cycles at almost the same pace, it must be because NGDP targeting would fail to stabilize NGDP growth, perhaps due to policy lags. We'd still have periods like 2008-09, when NGDP growth plunged sharply. But in those cases we do not see the countercyclical inflation that so worries opponents of NGDP targeting.

Oddly, (without their realizing it) their fears about countercyclical inflation are due to a weird mix of assuming NGDP targeting fails, and that it succeeds.

To conclude, either NGDP targeting fails to stabilize NGDP, in which case inflation behaves much as it does today. Or it succeeds in stabilizing NGDP, in which case we have only three or four recessions a century, when there is a huge supply shock.

In fact, even the "problem" they worry about is not a problem at all. In the cases where inflation rises sharply as growth falls sharply (say mid-2007 to mid-2008), there is no widespread outcry among economists that money is too easy and that the Fed should raise rates to reduce inflation. The profession is just as concerned about the slowing growth as the rising inflation, and thus a 12 month period of 0% RGDP growth and 4% inflation does not lead to handwringing that inflation is too high. Wages (which respond to NGDP, not inflation) don't get unanchored when unemployment is rising, even if gasoline prices are rising. It's certainly unfortunate that living standards take a hit, but few people (other than former ECB head Trichet) feel a need to raise interest rates to do anything about it. (And how'd that work out?)

Remember, we already have a dual mandate, and the dual mandate implies exactly the same sort of countercyclical inflation path as NGDP targeting. Fears of countercyclical inflation are groundless, that's what we should all be rooting for. It's a feature, not a bug. And if we get it, the business cycle will be far, far milder.

PS. I see a few signs that Greece may do a deal after all.

Comments and Sharing

COMMENTS (15 to date)
Jon Murphy writes:

Very interesting post. Thank you for sharing. You've given me a lot to think about.

I'm relatively new to the idea of NGDP targeting, so I always enjoy your posts on the topic.

E. Harding writes:

"Or it succeeds in stabilizing NGDP, in which case we have only three or four recessions a century, when there is a huge supply shock."
-There's a third option: every recession that would be resolved through a demand shock before NGDP targeting gets resolved through a supply shock (a la ABCT). That doesn't sound too plausible, so I think NGDP targeting would halve the incidence of recessions.

BTW, what was the recession of 1990? Zebra, horse, or unicorn?

You should also do a post on the NGDP shocks of 1962, 1967, and 1986.

Don Geddis writes:

I suspect your opponents aren't quite making the logical error of your #1 and #2. It may instead (as E. Harding alludes to) be a disagreement with your subsequent sentence: "the overwhelming majority of business cycles are associated with strong shocks to NGDP growth". (I guess the question is whether "associated" means a causal connection, or mere correlation.)

I won't pretend to understand the macro model, but I can imagine someone believing that business cycles come from "real" shocks, and that stabilizing NGDP growth will do nothing but add the additional cost of "more inflation", without having any impact at all on the core causes of the observed periodic recessions. Treating the symptoms, in effect, rather than the root causes.

If that were true, then NGDPLT would take all the existing economic damage of business cycle recessions, and add additional costs of higher inflation on top of them.

Scott Sumner writes:

Thanks Jon.

E. Harding, 1990-91 was an adverse demand shock. However, as is the case in many recessions, there was a small adverse supply shock right before the recession.

The three NGDP shocks you mention were too small to cause recessions.

Don, Good point. Ironically I often get the opposite complaint, that it's a tautology that stabilizing NGDP will end recessions. People confuse the two types of GDP. You make a much better point, but I'd say that there is overwhelming evidence that NGDP has a causal role in RGDP. If in 2007 you told economists that the Fed was about to produce the sharpest drop in NGDP since the 1930s, what sort of real GDP outcome would they have predicted?

Andrew_FL writes:

If I had to guess, I don't think people saying this think 2 is true at all. I'd say they think we'd have the same old business cycle, but that when a stable NGDP path didn't change that, that we'd just switch to a higher NGDP target-and that that wouldn't work either.

In other words I think they reject your entire theoretical apparatus.

E. Harding writes:

"The three NGDP shocks you mention were too small to cause recessions."

-But they were all almost exactly of the same size as that in 1970, when a recession did take place.

Michael Byrnes writes:

I think some critics seem to think that under NGDP targeting, inflation would crowd out real growth. Of course they have it backwards - under an NGDP target real growth (ie positive supply shock) would crowd out inflation, while poor growth (adverse supply shock) would not allow higher inflation.

ThomasH writes:

An odd framework. Why would anyone worry about
inflation of 4%-5% pa for a few years?

Rather than just assume NGDP strays from its target path, I'd rather see how the machine works with NGDP targeting and real shocks but with some lags between the real shock, it's recognition, reaction to it [what does the Fed actually DO to manage NGDP?] and the response of the economy all of which has the effect of pushing NGDP off its target path.

Shocks to examine:

. Government decides to implement "austerity" (raises the discount rate at which it spends according to an NPV rule above the borrowing rate).
. "Animal spirits" of investors revive and they become willing to borrow to invest more at the going real interest rate.
. "Animal spirits" of discouraged workers revive and they enter the labor force at the going vector of wages.
. The nationality of the President of the monetary authority of a large trading partner changes and the new President is warned in a dream by a devil that inflation is just around the corner.
. Investors in a large multi country currency area suddenly realize that zero currency risk on sovereign bonds is not the same as zero credit risk.
. NIMBY coefficient falls and NPVs of many more infrastructure projects turn positive (and government actually follows an NPV rule.)

And I'd like to see the comparison with PL targeting and the same set of shocks.

Dustin writes:

Could tight money be considered a supply shock, given that capital is an input to production, that leads to a demand shock?

J.V. Dubois writes:

I think the gist of the argument was to the end of your blog, where you talked about slowing RGDP growth and high inflation.

Now I will try to play devil's advocate: if we have a shock that would under normal circumstances affected demand, we will need higher inflation (let's say 4 or 5% with 0 RGDP growth).

Now imagine that this situation goes on for some time and for some reason this inflation rate becomes anchored. Now this may seem very unlikely - isn't the whole point of NGDP targeting to anchor NGDP and not inflation? But imagine that there would be such a counterplay that it would be impossible to for CB to "costlessly" lower the inflation from 4 or 5% to 0-2% if RGDP picks up? Whenever they try to lower the inflation (tighten) RGDP that was about to pick up would be stiffled making the whole tightening unnecessary in the first place.

Or in other word's these economists may think that inflation is much more sticky and changing it can be costly even if we have different monetary target. Now I do not think this to be true. Inflation moved around quite a lot during gold standard era for instance. But I think that for somebody that thinks in inflation terms all the time such a thing can be quite hard to think about.

Jose Romeu Robazzi writes:

@Prof Sumner
This post does a lot to explain why some inflation fears are exagerated, but I think it remains a question that is more of a technical detail and a "stress test", let's say: if you have in place a level target regime and a severe and prolonged negative supply shock happens, then, after a while, maybe there will be inflation that may be a little too high even if we relax current standards. I know, I have asked this question before, this hypothesis implies that initially the stabilization policy failed (despite stimulus NGDP remained below target for a certain number of periods) and them it worked (because stimulus was increased until necessary). I think it is important to assess this situation, however unlikely it may be, because that is the real downside to the model. It is important to notice also that the higher inflation would not be contemporary to the supply shock, but would apper in subsequent periods, perhaps even maybe a few years later ...

ThomasH writes:

I think there is a distinction to be made between how a policy is made and how it is described to the public. I agree it sound funny to explain that we are trying to keep the average inflation rate constant, but I don't recall anyone ever trying to explain lowering interest rates or buying long-term asses that way. It is of course possible that explanation constraints might incline policy makers to NGDP target over PL target.

But it would still be interesting to see the model with endogenous monetary policy which worked out explicitly for each target with the same (negative real) shock:
When does the target variable reveal the shock? How long does it take the Fed to react? Exactly what does the Fed do and when to maintain/return the target variable to the desired path
How does the economy respond to those actions?
With which target is real income higher?

BTW, I think that the following misjudges the resistance to QE in 2009-10

Instead the Fed ran into a firestorm of controversy, as the public was outraged to hear news reports that the Fed was trying to raise their cost of living

I do not remember the resistance coming from "the public" worried about their cost of living at all but from people -- I'm looking at you, Rick Perry -- who oppose "debasing the currency" on moral - ideological grounds. That resistance would have been there as long as the expected effect of Fed actions (whatever target they were following) was going to result in inflation.

Scott Sumner writes:

Andrew, I find that when people raise those sorts of objections it's a waste of time trying to convince them---they've made up their minds and are just looking for excuses.

E. Harding. That was a tiny recession with unemployment running about 6%. It was mostly a relapse after a very overheated economy in 1969, with unemployment far below the natural rate in 1969.

Dustin, It has some effects on supply, but it's primarily a demand shock.

JV, Yes, lots of economist think it's inflation that's sticky, whereas it's actually wage growth that is sticky. If NGDP growth is stable then wage growth will be stable.

Jose, That raises a lot of extra issues. Here I was just trying to look at what happens if you actually stabilize NGDP. There are an infinite number of ways that NGDP targeting could fail.

Thomas, You said:

"That resistance would have been there as long as the expected effect of Fed actions (whatever target they were following) was going to result in inflation."

I believe that if Bernanke had said that the Fed was trying to boost the incomes of average Americans it would have been harder to demagogue on talk radio.

JJ writes:


you mention NGDP a lot. Can you briefly explain what NGDP is so that a person with a basuc understanding of economics training can understand? What is NGDP targeting? How could a government/central bank go about NGDP targeting? Why would a government/Central bank want to do NGDP?

I think I understand the inflation targeting position of the monetarist school, but I'm not clear on NGDP targeting.

Scott Sumner writes:

JJ, It's the current dollar value of all gross final output in the economy, during a given year. Or total gross income (which is the same thing.) This article better explains the idea of NGDP targeting:

Comments for this entry have been closed
Return to top