Scott Sumner  

Intellectual regress

PRINT
Demography and Decency... If Politicians are Evil, What ...

Intellectual progress is much more common than regress, but I do believe that the latter problem occurs more often than we might assume. Mid-century Keynesianism may have represented progress in some dimensions, but it also reflected regress on our understanding of the role of monetary policy.

Jeff Holmes sent me a fascinating blog post on scurvy, which he noticed had an interesting connection with some of the ideas that I blog about.

The post is too long to quickly summarize, but the gist of it is that knowledge of how to avoid scurvy was more advanced in the early 1800s than in the early 1900s. Indeed scurvy caused major problems in Scott's 1911 ill-fated expedition to the South Pole. This is odd, because by 1800 the British navy knew that eating citrus fruits could prevent scurvy, but did not understand the role of vitamin C, or indeed even the importance of fresh food. As shipping speeds increased with an advent of steam power, scurvy became far more rare. Gradually fresh lemons were replaced by bottles of rather stale lime juice, which had very little vitamin C. The inadequacy of this replacement only became clear on the long Antarctic expeditions.

I think something like that happened in economics, especially with the Phillips Curve. The Phillips Curve model worked pretty well under the gold standard, and indeed its policy implications were very important in the early 1930s. If we had prevented the Great Deflation of 1929-33, then we would probably have also avoided the Great Depression.

But just as the key to stopping scurvy was fresh foods, not highly acidic stale lime juice, so the key to stopping depressions was stable NGDP growth, not stable prices. The Phillips Curve model was a reasonable approximation of the truth, until it no longer worked very well.

Several commenters sent me a recent WSJ article, where lots of well-known economists express puzzlement over the recent period of low inflation. This outcome seems to violate the Phillips curve model, as unemployment has now fallen to 5%.

Those of us in the market monetarist camp are not at all surprised, and indeed we predicted that inflation would fall short of the Fed's expectations. While it looks like low unemployment might cause higher inflation, that's a statistical illusion caused by the fact that highly inflationary policies sometime create low unemployment, and deflationary policies often create high unemployment. But in both cases you only get that result when NGDP moves in a certain way relative to wages. The actual cause of high inflation is not low unemployment, but rather monetary policies that cause NGDP to rise very fast relative to trend RGDP growth. But recent Fed policy has led to slower NGDP growth, and we in the market monetarist community noticed that the bond markets seemed to expect continued very slow growth in NGDP.

The profession wrongly assumed that monetary policy was "expansionary" when expectations for NGDP growth showed that the stance was actually highly contractionary. That, combined with the plausible but not quite accurate Phillips Curve model, led to false predictions about the likely path of inflation. And this is not the first time the Keynesians screwed up. As Lars Christensen recently reminded me, they also misused the Phillips Curve in the 1970s, expecting that high unemployment would bring down inflation. But inflation didn't fall, at least not until the early 1980s, when Volcker finally slowed the extremely rapid growth in NGDP.

The recent failures of central banks to hit their inflation targets have led to a crisis of confidence:

Central bankers "thought that it must be their own doing," said Jon Faust, the director of the Center for Financial Economics at Johns Hopkins University, who served two stints at the Fed during that period. "We thought we figured out macro policy, and we could deliver low, stable inflation and stable output and low unemployment and all things good."
They had "figured out macro policy" in one sense, but they didn't know how they had achieved those results. It wasn't Phillips Curve thinking, it was keeping NGDP growth close to 5%. It just so happened that both approaches yielded similar results during the Great Moderation. But when they diverged after 2008, the Fed followed the wrong model, the Keynesian Phillips Curve model rather than the market monetarist NGDPLT approach.

Adam Posen sees the problem:

"There's no way in hell anybody reasonably predicted, using the mainstream models, that you would end up with inflation this low," said Adam Posen, the president of the Peterson Institute for International Economics, a think tank with an international focus.

"Macroeconomics is in the era of Kepler and Copernicus and Tycho Brahe. We built Ptolemaic models and thought we were doing quantum mechanics," said Mr. Posen, who served on the Bank of England's monetary policy committee from 2009 to 2012 and pressed his central bank colleagues at home and around the world to launch more aggressive monetary action to revive stagnant economies and boost inflation rates.

Mr. Posen was on the winning side of those debates--major central banks doubled down on stimulus efforts--but expected inflation never materialized.


The key phrase here is "using mainstream models." Unfortunately, there are signs that the Phillips Curve model will be replaced with something even worse:

Former Fed Chairman Ben Bernanke, in an interview, pointed to Congress as one culprit for inflation's weak performance. As Fed chairman, he urged Congress in the aftermath of the recession to temporarily boost government spending and focus on longer-run measures to control debt. By raising demand, that spending would also boost inflation. After an $830 billion stimulus plan in 2009, however, Congress turned to shorter-run budget cutting.

Mr. Bernanke said in an interview that a central bank's ability to raise inflation would have to assume "at least reasonably cooperative" fiscal policy. In other words, the central bank can only do so much to stabilize the U.S. economy if lawmakers are working against Fed efforts.

A new theory about low inflation has emerged from former Bank of Japan governor Masaaki Shirakawa. While his former professor, University of Chicago economist Milton Friedman, said inflation could only be caused by a surge in the money supply, Mr. Shirakawa raised demographics as a cause.

Japan's aging population during the 1990s and 2000s seemed to unleash powerful deflationary forces, according to his theory, by lowering expectations of growth, straining the government's budget and putting a growing proportion of Japanese consumption in the hands of older people who draw on savings rather than younger wage earners.

The population of Japanese ages 15 to 64 has fallen from around 87 million in the mid-1990s to about 77 million in 2015, according to data from the Organization for Economic Cooperation and Development. That means fewer working-age people to buy existing homes or purchase products, putting a damper on the economy's ability to boost demand or bid up prices.

In a prescient 2012 speech, Mr. Shirakawa said the U.S. and Europe would face similar conditions: "I cannot entirely rule out the looming menace that may unveil itself into downward pressure on inflation rates."

In 2014, a trio of economists at the International Monetary Fund endorsed many of Mr. Shirakawa's hypotheses, arguing there are "substantial deflationary pressures from aging" and that "this applies not just to Japan, but also to other countries with aging or declining populations."

Debate over inflation and its causes has come to a boil in recent months, with other theories joining Mr. Shirakawa's ideas about aging countries and Mr. Bernanke's blame of stingy fiscal policy.

Economists challenged Fed officials about their understanding of inflation during two days of talks this summer at the Kansas City Fed's annual monetary conference in Jackson Hole, Wyoming,

Mr. Faust of Johns Hopkins University said the long-standing view that central banks could control inflation was a myth. "There's very little support for the view that inflation is simple and we had it figured out," he said.


Reading this makes me very demoralized; it seems that macroeconomics is regressing. When money is tight we get slow NGDP growth expectations, and that leads to low T-bond yields. It should be no surprise that inflation is low in that sort of policy environment. And this has nothing to do with the Fed being unable to hit its targets because of the zero rate bound. After all, they are planning on raising rates this week. This is all about the Fed and the ECB having the wrong target.

Just to be clear, I did not expect inflation to get this low, as neither the market nor I is able to predict sudden oil prices collapses. I'm talking about the sub-2% trend rate of core inflation in recent years, which was entirely predictable. The fact that mainstream economists didn't predict it should lead them to re-evaluate their models, not come up with bizarre theories of "demographics" causing deflation.

Note that the Japanese economist who used the demographics excuse presided over the last portion of a near two decade period where the GDP deflator fell at 1%/year. Since he was replaced in 2013, the Japanese deflator has trended upward at 2%/year, even as the demographic situation has worsened dramatically. The yen was at about 80 when Abenomics was first announced in late 2012, and now is at about 125 to the dollar. It's odd to see that one of the architects of Japan's deflation is still excusing his tight money policies by pointing to "demographics", after the dramatic success achieved by his successor. What would Milton Friedman think of his former student?

HT: Benny Lava, TravisV


Comments and Sharing






COMMENTS (10 to date)
Meets writes:

Although I find the demographic theories compelling, I think this is more correct.

So how to change the model economists use? NGDP target is on their radar but I don't think anyone is talking about making huge changes to their models.

How to make it happen?

E. Harding writes:

"While it looks like low unemployment might cause higher inflation, that's a statistical illusion caused by the fact that highly inflationary policies sometime create low unemployment, and deflationary policies often create high unemployment."

-An MR commenter pointed out that, given a constant rate of NGDP growth, less labor slack does, indeed, cause higher inflation due to falling RGDP growth as the recovery turns into a period of full employment. But rapidly rising employment (as in 2014) should then still be correlated with lower inflation, so this doesn't excuse the terrible WSJ article.

E. Harding writes:

Also, there are two easy rebuttals to the demographic theory of the price level: Russia and Belarus.

Kevin Erdmann writes:

Great post, Scott.

I fear that we have a similar problem with the Lehman Brothers failure paired with the disastrous Fed meeting in Sept. 2008. The lesson that is taken from Sept. 2008 is that we shouldn't have let Lehman fail, and I don't see how one would mount a convincing counterfactual to show that a Lehman bailout would have been a big jar of stale lime juice, while aggressive, generalized accommodation would have been the fresh citrus.

Steve writes:

Demographics produce lower Wicksellian interest rates.

Bad policy models translate lower Wicksellian interest rates into deflation.

And mass produced bottles lead to sugary placebos for the body (or mind).

marcus nunes writes:

Scott, They just "keep hanging on" to their misguided Phillips Curve" belief!
https://thefaintofheart.wordpress.com/2015/12/14/the-fed-insists-in-hanging-on-to-a-wrong-model/

Scott Sumner writes:

Meets, More blogging?

E. Harding, That's an odd way of looking at it. The Phillips curve model says faster growth is inflationary, whereas with stable NGDP growth it's deflationary.

Good point about Russia.

Kevin, Good analogy.

Steve, I do think that's why people wrongly think low population growth is deflationary.

Marcus, You always beat me to it.

E. Harding writes:

I thought the Phillips Curve said that lower levels of unemployment were inflationary, not faster growth.

pmsap writes:

Scott, quick question (or anyone else who wishes to answer).

According to the article, in the 70s, US had high unemployment and high NGDP growth. How do you related these two facts within the musical chairs model? Wouldn't a rapid NGDP growth devalue W as to bring unemployment down to "normal" levels?

Lorenzo from Oz writes:

Milton Friedman was right, monetary fallacies never die. Population growth (i.e. demographics) was used as an "explanation" for the Price Revolution aka Great Inflation from the late C15th to mid C17th. That far greater population growth in the C19th coincided with zero long run inflation does not seem to bother folk.

As Deirdre McCloskey pointed out, higher population is (ceteris paribus) deflationary, as you have more transactions for a given quantity of money in circulation. Similarly, falling population would tend to be inflationary on the same basis--fewer transactions for a given quantity of money in circulation. (Though, of course, why one would expect quantity of money in circulation to have any connection with population change is the question.)

An ageing population would tend to put downward pressure on interest rates, due to shifts in time preference. But that just takes us again back to Milton Friedman and not judging monetary policy from interest rates.

Comments for this entry have been closed
Return to top