Scott Sumner  

Inflation targeting: It's even worse than you thought

Partial Response to Yoram Baum... Dear Nationalism...

Wolfgang Munchau has a nice article in the Financial Times on inflation targeting. I agree with much of what he has to say, but will offer mild criticism of two points. Here's how Munchau begins:

Back in the 1990s inflation targeting was all the rage. I was a sceptic. I recall asking a senior central banker at the time what he would do if faced with stagflation - high inflation, low growth. Would he raise interest rates and force the economy into recession just to meet the target? He said the situation would never arise.

He was right. It did not. Inflation targeting became an improbable success. But it is failing now for reasons different from those I feared.

I believe Munchau is wrong in believing he was wrong. In fact, as far as I can tell his 1990s fears have proven accurate. Consider the following facts:

1. In July 2008 the ECB tightened monetary policy because inflation was running above target. This was done despite very sluggish growth in the eurozone.

2. In mid-2011 the ECB tightened monetary policy because inflation was slightly above target. This was done despite the fact that the inflation mostly represented VAT increases for austerity purposes, as well as rising prices of imported commodities due to fast growth in places like China. The prices received by European producers were not rising rapidly.

Surely Munchau is aware of these facts, which makes me wonder why he did not cite them. Perhaps when he worried about "stagflation" he was thinking about the very high rates of inflation experienced in the 1970s. Yes, that has not occurred. But unless I'm mistaken the policy errors of 2008 and 2011 are pretty clear examples of exactly the sort of scenario that inflation targeting (IT) foes feared. The first almost certainly worsened the 2008-09 recession, and the second helped abort a recovery, and led to a double dip recession.

Munchau goes on to provide a lucid explanation of why a policy of targeting the path of prices is better than an IT regime that lets "bygones be bygones."

My only other quibble with the article is this comment about NGDP targeting:

Another option would be to target nominal output growth - which is best thought of as the sum of real economic growth and inflation. The trouble is that nominal output growth is so slow that, if you started this regime today, hitting the target would entail a larger stimulus programme than anyone would have the nerve to implement. If you are looking for a new policy, targeting the price level is a better choice.
This is an argument I see all the time, but it doesn't really hold up. Although both price level and NGDP targeting are completely rules-based once established, there is inevitable discretion in setting up the original trend line. While some NGDP proponents have recommended going back to the pre-2008 trend line (indeed that was originally my view), so much time has now passed that it seems far more likely that any NGDP targeting regime would set a new trend line along a track considerably below the pre-2008 trend. Thus adoption of NGDP targeting need not involve excessive monetary stimulus and inflation, it entirely depends on where you set the trend line, and also which growth rate is chosen. For instance, Bill Woolsey has proposed a 3% NGDP target path, which would almost certainly lead to a very low rate of inflation in the long run, even accounting for the modest reduction in trend growth in recent years.

HT: Nicolas Goetzmann

Comments and Sharing

CATEGORIES: Monetary Policy

COMMENTS (11 to date)
Zachary Bartsch writes:

In class, my professor always said "and Sumner wants to use the OLD growth rate for NGDP...". It was an off-hand remark that was mentioned repeatedly and dismissively. So that's what I knew about you.

It did seem odd to me at the time (Fall 2012). It reminded me of the European countries post WWI quasi-attempting to go back to the previous gold exchange rates rather than adopt new exchange rates. Maybe you advocated it at a time previous to 2012 when a quick recovery was still possible in order to prevent lackluster recovery.

Thanks for not being obstinate and being willing to recognize the new trend, addressing it as a fact relevant to policy prescription.

Posts like this are one reason that I enjoy your blog.

Further, When would you say you changed your mind about the prescribed trend path of NGDP? 2011 maybe? (obviously, even an announced new trend would have been better than the unannounced new trend that we kinda have).

Andrew M writes:

What considerations go into the pick for a trend line?
Once adopted, what would it take to prompt the CB to change course?
What would be the consequences of changing your previously agreed upon trend line?

Yancey Ward writes:

I just don't see how NGDP targeting is any better than inflation targeting in regards to Munchau's original complaint- even with discretion to reset the trendline. Couldn't one just ask- what would the central bank do if real growth was low but nominal GDP was growing beyond the target? And you have yet to show that NGDP targeting rules out stagflation.

Edward writes:

"I just don't see how NGDP targeting is any better than inflation targeting in regards to Munchau's original complaint- even with discretion to reset the trendline. Couldn't one just ask- what would the central bank do if real growth was low but nominal GDP was growing beyond the target? And you have yet to show that NGDP targeting rules out stagflation."

It doesn't.

What it does is to limit the amount of stagflation that will occur. If the economy is growing at 5% NGDp per year, and 4% is from inflation, clearly there are supply side problems that need to be addressed.

Edward writes:

I find the title of your article quite humorous.
It reminds me of several articles of the type of newspaper (businessweek or Newsweek maybe)that goes like this: "Why X is worse than you think" or (better)
I consider myself to be an intelligent layman,
And many of the topics I've never even heard of, let alone thought of? Until I've actually read the article.
You've said the public doesn't understand inflation,
I tend to agree
That's why I find the title of your piece funny. :-)
Just sayin''
But munchau's views are tragic.

Yancey Ward writes:
What it does is to limit the amount of stagflation that will occur.

Well, so does inflation targeting, or it does if the central bank doesn't balk at stopping inflation before it runs away. Nothing special in NGDP targeting prevents NGDP from exceeding the target while leaving real growth low or negative. The central bank always faces the same dilemma Munchau lamented.

Scott Sumner writes:

Zachary, It's not so much my views that have changed, but rather the situation. In 2009 I favored going back to the old trend line. In 2009 I did not favor going back to the old trend line if there had been no movement in that direction between 2009 and 2014. In 2014 I still think that in 2009 we should have gone back to the old trend line, whereas in 2014 I notice that we have not made in progress going back to the old trend line in 5 years, and hence don't think it makes sense today to go back to the old trend line. In that sense my views have not changed.

The easiest way to imagine why is with a reductio ad absurdum example where no progress was made for 1000 years. Should we (in 3009) try to go back to the pre-2009 trend line. No one would say yes.

Andrew, The biggest factor is whether the government taxes capital. The lower the tax rate on capital (zero is optimal), the higher the optimal NGDP growth rate. It also depends on the rate of population growth (per capita NGDP is what matters.) It also depends on whether there is money illusion (defined as workers irrationally opposing nominal pay cuts.) The more money illusion, the higher the optimal NGDP growth rate.

My views on the proper trend line have gradually evolved over the past 5 years, there was no sharp break. That evolution is consistent with my overall model.

Once adopted I would not change the trend line, unless there was powerful theoretical evidence that the old policy regime was flawed. The recent recession would not justify a change in the trend line, if NGDP level targeting had been in effect in 2008 (obviously it wasn't, so that's a moot point.)

Yancey, You misunderstood me. I do not favor discretion to change the trend line under a NGDP targeting regime. Yes, stagflation is possible, but it does less damage under NGDP targeting than IT.

And yes, nothing prevents the central bank from not hitting the target, but that's true of all regimes, including even the gold standard (as we saw in 1933.)

Thanks Edward.

Philo writes:

Gavyn Davies, in FT (, seems to have a definite idea, borrowed from Goldman Sachs, of what constitutes monetary tightening or easing. He reproduces a Goldman Sachs chart of "financial conditions," based on "financial condition indicators" ("FCIs"). "These FCIs, which have been kindly provided by Goldman Sachs, include several indicators which reflect overall monetary conditions, including short rates, long rates, credit spreads, equity valuations, exchange rates and even house prices. They therefore capture most of the channels through which monetary policy, including quantitative easing, is expected to work."

How closely does Goldman Sachs's measure of "financial conditions" approximate your own measure of monetary ease/tightness, in terms of expected level of future NGDP? How many people would you estimate have in mind something like the GS measure when they think and talk about monetary easing/tightening? (I have been wondering whether most people have *any* clear concept in mind; perhaps that was too harsh.)

Philo writes:

{A supplement to my previous comment, which, however, has not yet appeared:} Davies adds: "There is no standard way of measuring monetary conditions accepted by everyone. I have therefore shown [below] alternative indicators calculated by Bloomberg . . . . They are not identical, but contain a similar basic message." So Bloomberg and Goldman Sachs disagree slightly but are almost in agreement about how to measure monetary conditions. How close are they to Scott Sumner's measure?

Don Geddis writes:

Philo: Most of the FCIs you (and the article) cite, relate to the bank lending channel. But the Hot Potato Effect, and changing expectations of future nominal growth (NGDP or inflation) are far more important. So the claim that these FCIs "capture most of the channels through which monetary policy ... is expected to work" is false. At least, those are not the channels that Sumner expects monetary policy to work through.

I also note the article's sentence right after the part you quoted: "A 1 percentage point easing in these variables is expected to lead to a 1-1.5 percentage point increase in real GDP growth over the subsequent year." But monetary policy (directly) affects nominal aggregates, not real aggregates. This sentence seems to imply a model where changes in monetary policy directly affect real GDP. That seems highly suspicious, and is likely an indicator that they're completely on the wrong track.

Sumner would talk about how monetary tightness or easing would be expected to affect future NGDP. It's a completely separate, and much more complicated, question, to ask how a growth path in NGDP might relate to predictions of future real GDP.

So it seems to me that the model in the article you cite is very, very different from the framework that Sumner uses.

Philo writes:

@ Don Geddis:

Thanks for your reply.

"So it seems to me that the model in the article you cite is very, very different from the framework that Sumner uses." Different and, no doubt, inferior. They think easier money would directly produce a more-than-proportional increase in RGDP: why didn't that work for Zimbabwe, Weimar Germany, etc.? (Or do they think these were cases of *tight* money?)

But I find it interesting that these authors do have some definite measure of monetary ease/tightness--a rather complicated and sophisticated one at that--rather than (as I suspect is true of most people) just using monetary terminology with no clear idea in mind.

Comments for this entry have been closed
Return to top