Scott Sumner  

Is a falling population contractionary?

Ricardo Reis on price level ta... Is Immigration a Basic Human R...

One often comes across articles that suggest Japan faces "headwinds" of a falling population. This is supposed to contribute to falling aggregate demand and deflation. That may be true, but if so it's almost certainly not for the reason that many people assume.

One of the best examples of a falling population occurred in Europe between 1345 and 1400, when the population may have fallen in half (I'm not sure how accurate the data is.) There is some evidence that the effect of the Black Death was actually inflationary, especially for wages:

But the Black Death had other economic effects. Although the rise in real wages after the plague was not as dramatic as many assume, it was persistent, as this graph from Gregory Clark (who looks at England) illustrates:
This is exactly what I would have expected. A plague does not reduce "M" in the famous equation of exchange. It might reduce velocity, but over a period of many decades I doubt it would have much impact on the speed at which medieval people spent money. If MV doesn't decline, then that means there is no impact on aggregate demand. Since the Black Death would clearly reduce real output (Y) we can infer that it probably raised prices.

Many people get confused on this point, because they wrongly visualize "aggregate demand" in terms of "lots of people going shopping". That is, they think of AD as a real concept, when it is in fact a nominal concept. The most explosive growth in aggregate demand in recent decades occurred in Zimbabwe during 2008, yet there weren't all that many Zimbabweans at the shopping malls---it was mostly inflation.

Screen Shot 2017-03-18 at 2.20.41 PM.png
Japan is in the midst of an episode of falling population that may eventually rival the Black Death (fortunately with much longer life expectancy). Here is the Wall Street Journal discussing the Japanese labor market:

TOKYO--A labor shortage is helping Japan's temporary and part-time workers win greater pay increases than its full-timers, a reversal that adds to signs of gains by lower-wage workers globally. . . .

In January, part-time hourly wages rose 2.6% from the previous year, compared with a 0.4% increase for full-time workers' base wages.

On b-style, a listing site for part-time jobs, the average position was offering ¥1,087 an hour in February, up by more than 7% in two months.

"There is a labor crunch overall due to the falling birthrate. If you don't raise hourly wages, you can't secure people," said Keitaro Kawakami, head of b-style's research unit.

I'm not sure what to make of that, as those wage increases are not all that impressive. But the basic point is correct. Other things equal, a falling population would lead to rising wages. Why then does Japan usually have such weak wage increases? Because other things are not equal---Japan has an extremely tight monetary policy, leading to slow NGDP growth.

Another mistake is to assume that a falling population leads to weak AD because there is less "need" for goods. In fact, needs are limitless. Here's a FT article discussing Japan's housing market:

You could pity the Japanese house. Its average lifespan is only 26 years. Or, you could revel in a culture of exuberant renewal that has made the Japanese house the crucible of contemporary architectural experimentation.

It is no accident that Japan has the highest number of architects per capita in the world -- about five times as many as the UK and more than seven times as many as the US. Their country needs them. In 2015, Japan built almost a million new housing units.

That's almost as many as the 1.108 million homes built in America that year. If I had not read that FT article, I might have guessed that the US built 10 times as many homes as in Japan. After all, our population is still growing at a decent clip while theirs is falling rapidly. So on a per capita basis I'd expect far more home construction US as compared to Japan. Add in the fact that we have almost three times as many people, and you'd expect dramatically higher home construction here than in Japan, perhaps an order of magnitude higher.

The high level of Japanese housing construction makes sense if you think in terms of wants being unlimited. Japan has no problem coming up with lifestyles that people aspire to; it's just a question of whether the central bank will produce enough NGDP growth to make those dreams come true.

Screen Shot 2017-03-18 at 2.09.39 PM.png
PS. There is one sense in which a falling population could reduce AD. A falling population might reduce the Wicksellian equilibrium interest rate. If the central bank is foolish enough to target interest rates instead of NGDP expectations, then a lower equilibrium interest rate could lead to tighter money and reduced AD. But that's due to dumb monetary policy, not a falling population.

Comments and Sharing

COMMENTS (15 to date)
John Hall writes:

Japan also has much more sensible zoning laws.

Jerry Brown writes:

Is the picture of houses an example of those knocked down in Japan? Cause I would rebuild them too. Hopefully, it is not an example of the new houses built. If it is, they don't need more architects- just much better architects.

blink writes:

Scott, I'm confused by the claim that velocity is constant. It seems that, with more individuals active, money moves more frequently as there are simply many more transactions. In an extreme case, what if the population decreases from two to one. There is positive velocity before, but zero velocity afterward. Why does this intuition not extend?

E. Harding writes:

The one country you can always count on to have a declining (or at least flat) population and galloping Aggregate Demand is Belarus'. Compare that country with Japan.

An aging population is probably contractionary in real terms and in terms of the natural interest rate, but not necessarily in nominal terms.

rjs writes:

it depends on what is meant by contractionary. i would say falling output of real goods and services, inflation notwithstanding, is contractionary. by that definition, falling populations are usually contractionary...

Scott Sumner writes:

John, Yes.

Blink, I am not assuming that velocity is absolutely constant. But in the long run it seems unlikely that a falling population would have any dramatic effect on velocity.

Harding, Thanks, I did a post at MoneyIllusion.

rjs, Here I am discussing the impact on AD and prices, not output. I agree that it is contractionary for output.

Andrew_FL writes:

To put it simply, falling population is a supply problem, not a demand problem.

pyroseed13 writes:

"Other things equal, a falling population would lead to rising wages."

How exactly, Scott, do you square this with your view on immigration?

Hazel Meade writes:

First World Problems.

Somehow I imagine that Japan will survive the horrors of falling population, slow NGDP growth and booming housing construction.

Thaomas writes:

I guess it depend a lot on what is meant by "contractionary." If it means leading to lower p.c. consumption, not necessarily. Taken as a very slow moving supply shock, it would lead to a period in which investment ratcheted downward which require the monetary authority to adjust its ngdp target down (assuming the previous target had incorporated the optimal inflation component).

@Andrew FL,
Why is falling population a supply side "problem" if the issue is p.c. income?

RPLong writes:

Quite obviously, large populations are a sign that contraception has been loose, while small populations are a sign that contraception has been tight...

Andrew_FL writes:

@Thaomas-I said a supply problem, not a supply side problem.

Scott Sumner writes:

pyroseed, It's pretty much unrelated to my views on immigration. This post is looking at nominal wages, whereas real wages are what matter for living standards. I don't believe that immigration reduces real wages.

Lorenzo from Oz writes:

"I don't believe that immigration reduces real wages."
So the labour market is an undifferentiated mass and no amount of increase in labour supply of any form will decrease wages for any group because demand will magically counteract any supply shift for any group?


Lorenzo from Oz writes:

It will not even (given wages are sticky downwards) affect the path of wages? For any group?

So a massive increase in labour without a commensurate increase in capital will not affect the relative returns to labour and capital?

When, naive me, I thought the increase in capital (due to innovation and changing technology) relative to labour might have something to do with rising living standards since the 1820s.


Return to top