Scott Sumner  

Japan refutes old Keynesianism

AT&T, Time-Warner and the end ... What do criminal records tell ...

Over at TheMoneyIllusion I have a post lamenting the recent revival of old Keynesian ideas. It occurred to me that Japan is an almost perfect refutation of these theories. In particular, Japan shows that:

1. Inflation is not caused by a tight labor market.
2. There is no permanent tradeoff between inflation and unemployment.
3. Even massive fiscal stimulus has almost no impact on aggregate demand.

In Japan, the unemployment rate continues to fall and is now 3.0%, well below the levels prior to the Great Recession:

Screen Shot 2016-10-29 at 9.50.11 AM.png
And yet inflation is slightly negative:

Screen Shot 2016-10-29 at 9.51.21 AM.png
And even if one excludes food and energy, Japan's inflation rate is roughly zero.

Some might argue that the unemployment rate is not the right measure of labor market slack. Maybe lots of Japanese have become discouraged, and dropped out of the labor force. Not so, the employment to (15-64) population rate keeps soaring to new highs:

Screen Shot 2016-10-29 at 9.54.19 AM.png
(Unfortunately Fred data ends in early 2015---does anyone have more recent data?) Notice that this is almost a mirror image of what happened in the US. In Japan, the employment ratio rose from just below 70% in the mid-1990s, to about 73% today. In America it dropped from about 73% in the mid-1990s to just below 69% today.

Update: Commenter Michael Makdad informed me that the Japanese ratio has increased another 1.4 points over the past 19 months---to the highest rate since 1974. That's an extremely strong job market.

Screen Shot 2016-10-29 at 9.57.27 AM.png
And there is also lots of anecdotal evidence that Japan has a very tight labor market:

Send us your construction workers, your care givers, your store clerks -- but for a limited time only.

That's the message from Japan, where the number of foreign workers, though still relatively small, has nearly doubled over the past eight years, and Prime Minister Shinzo Abe's ruling party is considering policies to speed up arrivals. . . .

A 2015 Manpower Survey found that 83 percent of Japanese hiring managers had difficulty filling jobs, compared with a global average of 38 percent.

Low and falling unemployment, a record high employment population ratio, and lots of anecdotal evidence of an extremely tight labor market. And what's the evidence against my claim? And what's the Keynesian explanation for the lack of inflation in Japan?

Market monetarists don't believe inflation is caused by tight labor markets; they believe that NGDP growth causes inflation. A good example occurred in the US during 1933-34, when the WPI rose about 20%, and even consumer price indices increased significantly despite the highest unemployment rate in American history. The cause? Extremely rapid NGDP growth.

Japan has had almost no NGDP growth in the past 23 years. So the old Keynesian model has no explanation for the strong Japanese labor market. None.

You might wonder how the market monetarists explain this pattern. After all, don't we emphasize the link between NGDP growth and employment? Yes we do, but the Natural Rate Hypothesis (NRH) is also a part of market monetarism (and New Keynesianism.) Over time, workers' expectations adjust to slow NGDP growth, and the labor market gradually returns to the natural rate. According to the NRH, it is sudden unexpected changes in NGDP growth that destabilize labor markets.

In the US, a sudden fall in NGDP in 2008-09 pushed unemployment up to 10%, and then unemployment fell back to 5% as workers gradually adjusted to the unexpectedly low level of NGDP. The recovery took longer than usual, as workers are especially averse to nominal wage cuts. Cuts were needed because NGDP declined during the recession, whereas usually there is just a slowdown in the rate of growth. In addition growth in NGDP was unusually slow during the recovery.

In Japan, unemployment rose in the late 1990s and early 2000s, as NGDP fell unexpectedly. Then labor markets started to recover, as NGDP stopped falling and then rose slightly. Once again, unemployment rose sharply in 2008-09 when NGDP plunged, and recovered once NGDP stopped falling. After 2012, Abenomics caused NGDP to start rising, but even if it had been flat, the Japanese unemployment rate would have kept falling.

Screen Shot 2016-10-29 at 10.47.23 AM.png
(BTW, this does not mean Abenomics was useless; it's main purpose is (or should be) to reduce Japan's dangerously high debt/NGDP ratio, by boosting NGDP. The BOJ needs to adopt a more expansionary monetary policy.)

Speaking of the deficit, fiscal policy is probably the area where Keynesian theory has failed most spectacularly in Japan. Despite running massive budget deficits in the 1990s and 2000s, which pushed the debt/GDP ratio to well over 200%, Japan has had essentially no growth in aggregate demand since 1993. This isn't just a slight miss; it's one of the most expansionary fiscal policies in all of history coinciding with the worst performance of aggregate demand ever observed in a major economy, over an extended period of time.

And as if that's not bad enough, the recent pickup in Japan (since 2012) was accompanied by a modest tightening of fiscal policy, mostly due to a rise in the national sales tax from 5% to 8%. (The following shows the deficit/GDP ratio):

Screen Shot 2016-10-29 at 10.24.02 AM.png
I'm trying to imagine what it would be like to be an economics grad student in Japan, learning old Keynesian ideas. Each theory that was presented would predict almost the exact opposite of what actually happens in Japan. I don't doubt that the Japanese are very polite when visiting (western) Keynesians present their advice, but just imagine what they are thinking in private.

Comments and Sharing

COMMENTS (22 to date)
Jose Romeu Robazzi writes:

Great post, Congrats Prof. Sumner, this material is something that could be input to perhaps an academic paper someday?

This not only destroys the core of what was mainstream economics for the most part of last century, it also makes some popular economists look very foolish these days, to say the least ...

foosion writes:

Japan provides a nice antidote to "debt will kill us all" hysteria.

happyjuggler0 writes:


Greece provides a nice antidote against those who like to point to Japan and say that high government debt does not matter.

Also, what happens in the long run matters, despite what Keynes and Krugman have claimed. We haven't seen the endgame in Japan with regards to whether or not their high government debt will clobber them.

Unsustainable trends continue to happen right up until they stop happening.

Rajat writes:

I think I understand why market monetarists don't see inflation as caused by tight labour markets. But I am less clear on why market monetarists believe that if NGDP begins to grow rapidly during a period of high unemployment (eg 1933-4), demand-side inflation can occur. Given sticky wages, if AD begins to grow fast and if AS is very flat due to excess capacity, why does inflation rise at all, at least to begin with? Is it due to hiring/production lags or something else?

Scott Sumner writes:

Rajat, I'd say that in most markets a rise in demand is partly met by higher output, and partly by higher prices. It may vary from one market to the next, and also may depend on how fast NGDP rises. In 1933 it was very fast, and it was hard for output to rise any faster. Keep in mind that industrial production rose 57% between March and July 1933; that's three times faster than any other period of American history. So perhaps output just can't scale up much faster.

Rajat writes:

Thanks Scott, that makes sense.

James writes:


How many economists do you think would agree with you that "NGDP growth causes inflation."?

This seems to invert the more conventional understanding that NGDP growth is caused by growth in the price level (inflation) and growth in output.

Is there any empirical test you would propose to decide between these theories?

Andrew_FL writes:
(BTW, this does not mean Abenomics was useless; it's main purpose is (or should be) to reduce Japan's dangerously high debt/NGDP ratio, by boosting NGDP. The BOJ needs to adopt a more expansionary monetary policy.)

This is a pretty transparent call for inflationary finance.

George writes:


I could hardly believe my eyes at your 3 points. So, at least one economist has come to his senses. What are the odds that others will look at Japan's data and begin to rethink the orthodox views?

Michael Makdad writes:

to answer your question, the participation rate in Japan for ages 15-64 has continued to rise since early 2015. September 2016 was 1.4 percentage point higher than September 2014 (comparing September to September because seasonally adjusted data are not available by age group) -- 0.7ppt for men and 2.2ppt for women. It is the highest going back to 1974, including for men.

David R. Henderson writes:

Good post.
I wonder, though, about this statement: "they [market monetarists] believe that NGDP growth causes inflation.” Just as you can’t erase from a price change, you can’t reason from NGDP growth without knowing what caused NGDP growth.

Scott Sumner writes:

James, You said:

"This seems to invert the more conventional understanding that NGDP growth is caused by growth in the price level (inflation) and growth in output."

NGDP growth equals inflation plus real growth, so it can't be caused by those factors. The conventional view 10 years ago was that NGDP growth was caused by monetary policy.

George, Low.

Michael, Thanks for that data. I'll add an update.

David, I would say the key issue is not so much what causes NGDP growth (monetary policy in my view) as what is happening to real GDP---the supply side of the economy.

HL writes:

About Japan's employment ratio, your (and St. Louis Fed's) data are from OECD

For the group aged 15-64 (common working age population definition).

Aug 2016: 74.5%
Jan 2016: 74.2%
Jan 2015: 73.0%
Jan 2014: 72.3%
Jan 2013: 71.3%
Jan 2012: 70.4%
Jan 2011: 70.6%
Jan 2010: 70.3%
Jan 2009: 70.8%
Jan 2008: 70.8%
Jan 2007: 70.2%
Jan 2006: 69.7%
Jan 2005: 69.2%
Jan 2004: 68.7%
Jan 2003: 68.3%
Jan 2002: 68.4%
Jan 2001: 69.2%
Jan 2000: 68.9%
Jan 1990: 68.4%
Jan 1980: 67.0%

And so on. I am also including a link to the chart just in case (

HL writes:

Michael is actually talking about a different series (the series you mentioned above - employment to population - does not have September data point yet).

OECD refers to labour force participation rate (a different concept, obviously) as "activity rate" for an unknown reason. Anyway, the OECD's seasonally adjusted "population activity rate" for Japan rose by roughly 1.3% from Jan 2015 (75.772%) to Aug 2016 (77.065%).

But these don't change the overall picture, which is Japan's labour market remains strong at least in terms of demand!

HL writes:

strong labour market in terms of demand and supply. Not just demand. Anyway, these are minor points.

There is a bit of Keynesian pushback by Richard Koo and Adam Posen, which is that most of these fiscal stimulus packages were in name only ("mamizu / fresh water" argument). At least the econ PhD guys I know from Japan are somewhat sympathetic to the argument, which is that the only time Japan did serious fiscal stimulus in the 1990s was after the Hanshin earthquake. They also believe that Hashimoto's VAT hike on the eve of Asian financial crisis was supportive for the good Keynesian explanation.

Anyway, I think you are right on this, and thanks for another great post.

James writes:


You wrote "NGDP growth equals inflation plus real growth, so it can't be caused by those factors."

The part after the "so" doesn't follow from the part before in any logical sense. I can cite plenty of examples where A=X+Y and is also caused by them. A trivial one is the money in my piggy bank, caused by and also equal to deposits plus withdrawals.

I asked you for an empirical test and you provided none. Did you have any evidence for your claim that NGDP growth causes inflation?

Alexander Hamilton writes:

HL, The "no true fiscal stimulus" argument is weak when applied to the UK or US but I can't believe anyone would seriously apply it to Japan. If a country can run deficits for twenty years (increasing it's debt to GDP ratio up to 200%) and still not be doing fiscal stimulus then the phrase is totally meaningless.

James, An empirical test isn't required because the direction of causation you suggest doesn't make any sense. The standard story is: 1) Increase in money supply 2) increase in spending (estimated as NGDP) 3) More money chasing same amount of goods 4) Prices start to rise

Why would the price level increase before there has been in an increase in spending?

HL writes:

Alexander, I cited the arguments of Adam Posen and Richard Koo not because I strongly agree with them but because they seem to get garner some sympathy from Japanese econ PhDs. I mentioned it because Scott wondered how econ grad students would respond to traditional Keynesian types doing lectures in Japanese universities. But this is entirely anecdotal, of course.

Alexander Hamilton writes:

HL, Yes I understood that. I didn't mean to suggest you were endorsing that view.

Mr. Econotarian writes:

BTW much of Japanese government debt is being sucked up by the Bank of Japan:

While Japan’s estimated gross government debt is now over twice the size of the economy, according to Schulz’s calculations using BOJ data, the shuffle of holdings from private actors like banks and households to the central bank is having a big impact. It means debt in private hands will fall to about 100 percent of GDP in two to three years, from 177 percent just before Prime Minister Shinzo Abe took power in late 2012, he estimates.


Ricardo writes:

Professor Sumner
I would kindly like to disagree with your post, you can read more about it in my blog:
Thank you

Alan Reynolds writes:

Japan's efforts at "fiscal stimulus" clearly refute the whole concept. The country's quantitative easing (buying zero-rate bonds with zero-rate reserves) also seems unsuccessful rather than insufficient.

Fortunately, demand management is not the only tool in the economists' tool box. Microeconomic structural reforms could improve incentives for enterpreneurship and innovation.

To take an obvious example, Japan is nearly the only country in Asia that has not cut top income tax rates to 20-35% on individual income and no more than 25% on corporate profits.

Comments for this entry have been closed
Return to top