“Secular Stagnation! Larry Summers is wrong”

Roger Farmer posts: “Secular Stagnation! Larry Summers is right

Larry Summers has once again been advancing the secular stagnation hypothesis. David Andolfatto responds with this tweet which plots GDP per person in the United States since the late nineteenth century. I’m with Larry on this one.

Why does this matter and what does it have to do with secular stagnation? Those who would deny the secular stagnation hypothesis want you to believe that the economy has a very strong tendency to revert to a mean growth path which is independent of shocks. Leave the economy to itself, and the recuperative powers of the market will restore us to the social optimum. The secular stagnationists, and I am one of them, disagree.

We believe that, in the absence of corrective policies by the central bank or the treasury, the economy will never recover after a shock. The unemployment rate will not revert to its social optimum and, associated with that fact, the economy will never revert to its optimum growth path. After a shock, the data do not revert to the same trend that they followed before the shock hit.

GDP per person has a unit root. That is accepted by everyone who has studied these data. The interesting question is why?

If, as Robert Gordon believes, it is caused by random technology shifts then there is not much that monetary policy or macro prudential policies can do about it. If, as I believe, it is caused by random movements from one inefficient equilibrium to another, we should be thinking very hard about how to design a monetary/macro-prudential policy that keeps the economic train on the tracks.

I´m closer to Andolfatto. Furthermore, Farmer´s allegation that “a unit root in GDP per person is accepted by everyone is false! He links to a Cochrane 1988 paper. These two links (here and here) indicate “everyone” is not really everyone!

The chart below takes the post WWII years. I estimate the trend from 1950 to 1994 and project for the next 20 years. GDP per person remains on trend up to 2007, after which it drops AND stays down.

Farmer SS_1

During the late 1960s and throughout the 1970s, the economy was buffeted by significant demand (fiscal) and real (oil) shocks. Monetary policy was lousy (the reason behind the period getting named “Great Inflation”). Nevertheless, the chart indicates that real GDP per person always reverted to trend, contrary to what Farmer presumes.

During the “Great Moderation”, output per person hugged very closely to the trend, with little oscillation. That was broken in 2008 and for the past 7 years GDP per person has evolved far below trend. It´s not really a “Great Stagnation”, but a “depression” (even if it´s not “Great”).

I agree with Farmer that we should “thinking very hard about how to design a monetary policy that keeps the economic train on the tracks”. Market Monetarist´s suggestion is that monetary policy should strive to obtain nominal stability, in other words, keep NGDP evolving close to a level trend path.

The NGDP growth chart illustrates

Farmer SS_2

Update: Jérémie Cohen Setton at Bruegel has a take:

What’s at stake: The question of whether capitalist economies are self-correcting and will eventually revert to mean growth has received renewed interest given the underperformance of most economies six years after the onset of the Great Recessions. While the idea of persistent high unemployment was central to Keynes’ General Theory, it was quickly abandoned by the neoclassical synthesis.

Tyler Cowen writes that the most crucial issue is whether economies will return to normal conditions of steady growth, or whether we are witnessing a fundamental transformation, unveiled in bits and pieces. One relatively optimistic view is that observed deficiencies — like slow growth in real wages and the overall economy, persistently low interest rates and low levels of labor participation — are merely temporary.  Another commonly heard view is that we made the mistake of letting the last recession linger too long, allowing some of its features to become entrenched.



Ruled by a “phantom rate”

That is implied by this Gavyn Davies piece “Who is right about the equilibrium interest rate?”:

The equilibrium real interest rate continues to lie at the heart of discussions about economic policy in the US and elsewhereBen Bernanke has written that the equilibrium rate, and not the FOMC, is the ultimate determinant of interest rates in the economy, and claims that it is discussed at every Fed meeting. The recent debate about secular stagnation between Mr Bernanke and Lawrence Summers centres on a difference about the future path for the equilibrium rate. And Cleveland Fed President Loretta Mester says that it is “the issue policy makers are grappling with” at the FOMC.


If, for example, the Fed sets short rates higher than equilibrium, the economy will respond by generating a rise in unemployment and inflation will fall below target. And vice versa. Mr Bernanke believes that the equilibrium real rate is currently abnormally low, but that it will rise gradually in the next few years as economic “headwinds” abate. This justifies a large part of the rate increase shown in the dots chart, a view also explicitly stated by Ms Yellen in her speech on policy normalisation.

Regarding the “telltale” signs from inflation and unemployment to guide the “appropriateness” (relative to the “phantom rate”) of the FF rate, what was happening in 2007-08? And what is happening now?

Phantom Rate_0

In 2007-08, Bernanke and the FOMC only had ears for the “siren” of headline inflation, all due to the oil price shock, and couldn´t hear the “muted sound” of core inflation. Plus, unemployment was almost imperceptibly rising, at least initially.

Although unemployment was slowly rising, so was headline inflation. That doesn´t help gauge the “phantom rate” because if the FF rate is above the “phantom rate”, unemployment should be rising but inflation should be falling.

Since headline inflation was rising but unemployment was more or less stuck initially, probably the Fed thought that the “phantom rate” was above the FF rate, which made the Fed reduce the FF rate (after the Bank Paribas affair in August 2007) very parsimoniously. In fact, from April to September 2008 the FF rate remained at 2% (with a bias to increase!)

For the past three years inflation has been falling and is below target, but unemployment insists in falling, “messing up” the signals that inform the Fed about the “phantom rate”, resulting in a “stand-off”  that brings forth a lot of “noise” from policymakers.

If the Fed were to stop trying to figure the “phantom rate” and instead observe the very visible behavior of nominal aggregate spending (or NGDP), it could have tried to save the “plane from dropping too low” in 2008.

Phantom Rate_1

Now that the “plane has dropped too low”, the previous height may not be attainable. But all signals are that there exists an attainable intermediate height that the “captain” could aim for. But that height NGDP level is for the “tower” Fed to specify. Given that the plane´s engines economy has taken a beating, the previous “cruise speed” may have to be lowered. No matter, both labor and capital will “rejoice”.

Phantom Rate_2

Update: The “Phantom” seekers like:

Phantom Rate_3

The NGDP Targeting crowd prefers:

Phantom Rate_4


Another chapter on “What´s wrong with economics”

Wolfgang Munchau has an interesting piece on the FT: “Macroeconomists need new tools to challenge consensus”:

Macroeconomists are once again caught up in a discussion about the future of their profession. An example has been the recent debate between Lawrence Summers and Ben Bernanke about the deep causes of the economic slowdown. The former US Treasury secretary has defended the case of a “secular stagnation” while the former chairman of the Federal Reserve sees an excess of savings over investment.

It is an enlightened debate, but it also masks a much deeper problem within macroeconomics. Secular stagnation — a sharp fall in growth rates lasting a very long time — is not something that you can easily square with the current generation of macroeconomic theories and models.

Part of this debate reminds me of a discourse among mathematicians at the end of the 19th century. At the time, mathematicians — and physicists too — thought they had solved most problems, just as economists did until 2008.

The advent of chronic instability is the equivalent challenge for macroeconomics today. The present tools used by mainstream macroeconomists cannot deal with this adequately. New ones are needed. They exist in other disciplines, but to macroeconomists they look as weird today as the abstract stuff looked to mathematicians of the 19th century.

Some comments (with reference to the US):

1 I don´t think there is chronic instability. There was a spate of instability in 2008-09, that configured the crisis. For the past five years what we´ve had is a “great stability” at the “wrong level”.

2 The 1970s was a decade of “Great Instability” (“Great Inflation”) in real growth, unemployment and inflation; but that was followed by more than 20 years of “Great Stability” (“Great Moderation”) in real output, unemployment and inflation.

3 The “Great Disturbance” derived from the fact that policymakers, in particular the ones responsible for monetary policy, came to believe that the source of the “Great Stability” was having kept inflation low and stable (or on target). So when inflation got a direct hit from oil prices they “tightened the monetary screws” and reaped a “Great Instability” followed by an “Inadequate Stability”, also dubbed “Secular Stagnation”!

The necessary tools are certainly there. What´s needed is a revision of the mainstream views on the source of the “Great Stability”!

Nick Rowe has a discussion and suggestion:

Fluctuations in inflation are a noisy signal of monetary disequilibrium, because the firms that do change prices are not always representative of the firms that don’t. And by targeting inflation the central bank makes inflation stickier, and this reduces inflation’s signal/noise ratio. Fluctuations in output are also a noisy signal of monetary disequilibrium. NGDP targeting means targeting the sum of two noisy signals. NGDP targeting is unlikely to be exactly optimal, but may well be better than inflation targeting, which puts all the weight on one noisy signal and ignores the other.

The big puzzle of the recent recession is why the inflation guard dogs failed to bark, to warn central banks of recession. Even in those countries where inflation did fall, it only fell a little. In others it stayed on target, or even rose above target. The NGDP guard dogs barked loud and clear, giving a consistent and correct signal. That is what we need to model. And if we can model that, we may also have a model in which targeting NGDP can do better than targeting inflation.

Hiding behind “Secular Stagnation”

Scott Sumner comments on a post by Edward Hugh:

The focus of Hugh’s piece is Finland.  He points to the very weak recovery, and suggests that structural factors are involved. Perhaps so, but some of the data he cites point in exactly the opposite direction.  Finnish unemployment has been rising, and at 9.2% is at the highest rate in more than a decade.  Meanwhile Hugh’s post shows inflation in Finland falling to zero. Those data points suggest a lack of aggregate demand, not structural problems.

EH shows several charts to further his “secular stagnation” claims. One chart he doesn´t show is a comparative chart of nominal spending (NGDP) relative to trend. I do so below and add Austria and Spain as “evidence” of the degree of the AD shortfall in Finland.


Every country in the world could do with some structural reforms, some more, some less, but the big problem with most of those economies, as exemplified by the charts for the three countries above, where all are subject to the same monetary policy, is differing degrees of  aggregate demand shortfall.