For quite some time, I have been curious about the answer to the question below:
Keynesians have, over all, had a very good track record in this economic crisis; one indication of just how good is the remarkable extent to which people on the other side have had attack people like me for supposed wrong predictions we never, in fact, made (you endorsed Obama’s economic projections! No, I didn’t).
One area where things haven’t worked out as expected, however, is on the deflation front. Inflation has stayed very subdued; but coming in to the crisis I certainly thought that actual Japanese-style deflation was a real possibility. That hasn’t materialized (and for that matter, even Japan never had more than very gradual deflation). Why?
Paul Krugman gives an explanation that is both convincing yet seems to be missing something:
One immediate thing to look at was to see whether what was happening to inflation in the United States was consistent with historical experience of deep slumps that we know involved the economy operating well below capacity for an extended period. And it turned out that the Japanese deflation (which has never been very fast in any case) is pretty much unique. The IMF looked at Protracted Large Output Gaps — PLOGs — and found that in general inflation gets squeezed toward, but not below, zero:
And our own history actually points in the same direction: the 1930s were marked by sharp deflation in the early years, but considerable inflation as the economy partially recovered, even though unemployment remained very high.
So inflation seems “sticky”. But why? One immediate thought was that we might be looking at the effects of downward nominal wage rigidity: employers are very reluctant to engage in actual wage cuts. Way back in 1996 Akerlof, Dickens and Perry suggested that this would make inflation stubborn at low rates, breaking the usual link between high unemployment and disinflation.
Nominal wage rigidity is clearly an economic reality, but I kept wondering if inflation’s “stickiness” is a result of sticky wages or if the causality runs the other way. In my last posting, I identified what I believe to be a flaw in identifying the vertical turn in the aggregate supply curve as a long-term phenomenon:
But Krugman, a definite Keynesian economist, includes the same depiction of aggregate supply in his texts:
In Econ 101 textbooks (mine included) it’s standard to present the distinction between the long run and the short run with a picture something like this:
The classic short-run/long-run picture.
Here AD is the aggregate demand curve; more on that in a second. SRAS is the short-run aggregate supply curve, which is assumed to be upward-sloping because some prices and/or wages are sticky; LRAS is the long-run aggregate supply curve, which is vertical because in the long run price stickiness is assumed to go away.
Anyone else see it? Here’s the massive flaw: we’re supposed to believe real GDP cannot rise or fall in the long run? Looking at the above supply-and-demand model, one is left with than inescapable conclusion—real GDP must be a set value in the long-term to mathematically draw LRAS that way. Rising aggregate demand under these assumptions cannot have an effect on increased output, only a rise in inflationary pressures. But these assumptions are both empirically and logically false—clearly real GDP has risen precipitously during the 230 years since the end of the Revolutionary War. Am I supposed to believe the long-run horizon is longer than two centuries?
I believe the model should look more like this:
One important caveat: LRAS does not appear. Aggregate supply rising almost vertically is short-run capacity constraints. This means the key points are whittled down to just one:
- The aggregate supply curve illustrates the quantity of goods and services firms sell at any price level.
As the price level takes on an element of preeminence in this model, the most important controlling factor becomes aggregate demand, as a change in AD immediately effects a rise in both real output and the price level. [(I also argue LRAS is a resource limitation, thus affects SRAS shifts but otherwise is not plotted on the supply and demand model shown above). Graph originated here.]
I focus on changes in aggregate demand for a very simple reason—AD is far more dynamic and relentless than the economics field gives it credit for.
Aggregate Demand Relentless Dynamism
Krugman has come right out and asked the question why don’t we have deflation—which I believe is best answered with data from the Census Bureau:
# Increase: 22,164,068
% Increase: 9.78% (0.98% per annum)
Min % Pop New Arrivals: 8.91%
# Increase: 32,712,033
% Increase: 13.15% (1.32% per annum)
Min % Pop New Arrivals: 11.62%
2010: 308,745,538 Pop Density 31.42 persons/km²
# Increase: 27,323,632
% Increase: 9.71% (0.97% per annum)
Min % Pop New Arrivals: 8.85%
Oct 2013: 316,935,870 (U.S. Census estimate October 24, 2013 0430Z)
# Increase: 8,190,332
% Increase: 2.65% (0.66% to 0.88% per annum)
Min % Pop New Arrivals: 2.58%
I’ve calculated the increase between each decennial census since 1790 (American and Japanese Population Differentials), and included 1980 to the present day. In the first decade of this millennium, the American population increased by an average of 7,482 individuals per day. Estimates since the 2010 census show a depressing corollary:
2010-2011 per annum growth rate: 1.23%
2011-2012 per annum growth rate: 0.73%
2012-2013 per annum growth rate: 0.73%
Jan 2013-Oct 2013 per annum growth rate: 0.64%
Since 2010, American population growth has decidedly dropped, a reduction to a per annum rate similar to the U.S. experience during the Great Depression:
# Increase: 17,064,226
% Increase: 16.14% (1.61% per annum)
Min % Pop New Arrivals: 13.90%
# Increase: 8,894,229
% Increase: 7.24% (0.72% per annum)
Min % Pop New Arrivals: 6.75%
While the effects of population growth are always accounted for in monthly job growth statistics (somewhere in the 90,000 to 150,000 range required to keep up), never do I read that population growth from 2000 to 2010 has added $273 billion to current American GDP at the bare minimum (per capita GDP runs at more than four times the $10,000 I am assuming is the minimum expense to remain alive in the U.S. in 2013).
Someone with a working knowledge of economics and mathematics would immediately complain this is a drop in the bucket compared to overall American GDP. On New Year’s Eve 2012, U.S. annual GDP was tabulated as $15,684,800,000,000; the $273 billion from the 2000s population growth represents 1.74% of that total. Considering GDP growth was $693.5 billion more than the previous year, at first demographics appears to be a sideshow.
However, population growth also has a stimulus effect, one much larger (and more robust) than any fiscal stimulus Congress has even permitted to be spent. I would posit that the aggregate demand curve, which usually appears static or representing a single demand shift like so…
…rather shifts outward continuously, perpendicular to the tangent of the aggregate demand curve (also known as slope or price elasticity of demand), eventually abutting the point where the SRAS curve begins to slope upwards noticeably (toward the vertical). Passing the potential output point, prices rise much faster than real output…
…which, rather than indicating runaway inflation, far more likely is the price signal for firms to expand, resulting in an increase in capacity limits (SRAS shifts outward, to the right). Price increases abate from the shift in AS, until AD increases restart the cycle again. Over time, firms may also anticipate the boom generated from AD increases and expand in anticipation of the price signal rise. Nevertheless, this makes aggregate demand the prime driver of the economy, as the continuous increase of the AD curve yield both higher real output and an increased price level simultaneously.
Returning to Krugman’s question, I would posit that the continuous rise in American aggregate demand makes persistent deflation very difficult to engineer without a much more powerful force pushing against the population-rise AD effect. The setback from a leftward (contraction) demand shock tends to be temporary…
Annual Dec- Avg- Year Jan. Feb. Mar. Apr. May June July Aug. Sep. Oct. Nov. Dec. Avg. Dec Avg 2008 211.080 211.693 213.528 214.823 216.632 218.815 219.964 219.086 218.783 216.573 212.425 210.228 215.303 0.1 3.8 2009 211.143 212.193 212.709 213.240 213.856 215.693 215.351 215.834 215.969 216.177 216.330 215.949 214.537 2.7 -0.4
…the brief American descent into deflation that began in August 2008 yielded to inflation by December 2008 as new population growth pushed the AD curve outward. Inflation indeed looks “sticky” at first glance, but the reality might be that the relentless march of population-derived American aggregate demand increases makes persistent deflation almost the exclusive province of ill-advised policies like the gold standard. Almost exclusive, because the Japanese experience with deflation is the first major case of a very dangerous new phenomenon.
The Japanese Exception
The Japanese experience with population-based aggregate demand shows that the forces that push AD outward can (and do) have unbalanced forces that are equally relentless working in the opposite direction:
# Increase: 9,109,280 (10 year) 3,988,527 (5 year)
% Increase: 8.14% (10 year, 0.81% per annum) 3.41% (5 year, 0.68% per annum)
Min % Pop New Arrivals: 7.53% (10 year) 3.29% (5 year)
# Increase: 6,550,771 (10 year) 2,562,244 (5 year)
% Increase: 5.60% (10 year, 0.56% per annum) 2.12% (5 year, 0.42% per annum)
Min % Pop New Arrivals: 5.30% (10 year) 2.07% (5 year)
# Increase: 4,521,323 (10 year) 1,959,079 (5 year)
% Increase: 3.74% (10 year, 0.37% per annum) 1.58% (5 year, 0.32% per annum)
Min % Pop New Arrivals: 3.60% (10 year) 1.56% (5 year)
# Increase: 3,314,676 (10 year) 1,355,597 (5 year)
% Increase: 2.68% (10 year, 0.27% per annum) 1.08% (5 year, 0.22% per annum)
Min % Pop New Arrivals: 2.61% (10 year) 1.07% (5 year)
# Increase: 2,197,748 (10 year) 842,151 (5 year)
% Increase: 1.75% (10 year, 0.18% per annum) 0.66% (5 year, 0.13% per annum)
Min % Pop New Arrivals: 1.72% (10 year) 0.66% (5 year)
2010: 128,057,352 Pop Density 338.83 persons per km²
# Increase: 1,131,509 (10 year) 289,358 (5 year)
% Increase: 0.89% (10 year, 0.09% per annum) 0.23% (5 year, 0.05% per annum)
Min % Pop New Arrivals: 0.88% (10 year) 0.23% (5 year)
Oct 2013: 127,253,075 (U.S. Census Bureau estimate)
# Increase: -804,277
% Increase: -0.628% (-0.157% to -0.209% per annum)
Min % Pop New Arrivals: 0.00%
The population story in Japan is actually worse than the Japanese census lets on:
Unsurprisingly (to me at least), deflation sets in as population growth drops to such a low level (before turning negative):
CPI Japan 2013 -0.12%
CPI Japan 2012 -0.03%
CPI Japan 2011 -0.28%
CPI Japan 2010 -0.72%
CPI Japan 2009 -1.34%
CPI Japan 2008 1.37%
CPI Japan 2007 0.06%
CPI Japan 2006 0.24%
CPI Japan 2005 -0.27%
CPI Japan 2004 -0.01%
CPI Japan 2003 -0.25%
CPI Japan 2002 -0.90%
CPI Japan 2001 -0.80%
CPI Japan 2000 -0.65%
CPI Japan 1999 -0.33%
Japan might already be in line for suffering greater population loss than the country experienced during the Second World War…
# Increase: 2,743,956 (10 year) -1,076,967 (5 year)
% Increase: 3.96% (10 year, 0.40% per annum) -1.47% (5 year, -0.29% per annum)
Min % Pop New Arrivals: 3.81% (10 year) 0.00% (5 year)
…with virtually no chance to recover with an effect similar to the surge after the war:
# Increase: 10,124,566 (10 year) 11,201,533 (5 year)
% Increase: 13.86% (10 year, 1.39% per annum) 15.56% (5 year, 3.11% per annum)
Min % Pop New Arrivals: 12.17% (10 year) 13.46% (5 year)
If I were a Japanese economist, I would fear that current -0.16% per annum population loss rate in my country might actually hinder the effects of Abenomics…
Chart – CPI inflation Japan 2013 (yearly basis)
january 2013 – january 2012 -0.30 %
february 2013 – february 2012 -0.60
march 2013 – march 2012 -0.90 %
april 2013 – april 2012 -0.70 %
may 2013 – may 2012 -0.30 %
june 2013 – june 2012 0.20 %
july 2013 – july 2012 0.70 %
august 2013 – august 2012 0.91 %
…when compared to American inflation that is assisted by the stimulus of a 0.66% per annum growth rate:
I am not decrying Abenomics by any means at this point (one could wish the Federal Reserve’s QE was even a quarter as successful), but I feel compelled to point out that the beneficial effects of population rise boosting American aggregate demand appear to be working in reverse in Japan. If Abenomics is shut off (“tapered” in Fed-speak), I would posit the Japanese economy would quickly slide back into deflation (unless Abenomics is able to reverse the rate of Japanese population decline).
If deflationistas like Dean Baker might read this posting (a highly dubious proposition), I am certain that the ranks of economists that recognize these types of threats would nevertheless balk at my theorems:
The decline in population is in fact a benefit in many respects for Japan. It is a very crowded island with expensive land prices. A falling population will reduce the pressure on land making housing more affordable. It will also reduce congestion in cities. In addition, the decline in population will make it easier for Japan to meet commitments for reducing greenhouse gas emissions, if countries are ever held responsible for the contributions to global warming.
Falling population at first glance also seems to be a boon to the young:
The situation will be even better insofar as more workers are pulled into the labor force. As this article notes, because of weak demand, many younger workers cannot find jobs. If Japan were facing a “demographic squeeze” then young workers would have no problem finding jobs since there would be a shortage of workers. Also, because of the longevity of relative good health of many older Japanese, it is likely that many people will opt to continue working past age 65.
As Baker is likewise irritated by the “well-funded effort in the United States to try to place demographics at the center of economic policy debates,” I can understand why demographic arguments are viewed with suspicion by Keynesians. Supply-side dogma has so overtaken the economics profession that the mere mention of population growth automatically leads economists with a working concept of aggregate demand to suspect that describing population contraction a problem is a straw man for introducing more labor supply to reduce wages. Often such suspicions are right on the money. But in recognizing that Japan doesn’t need more workers (it has plenty of unemployed and underemployed freeters), do the deflationistas ask if Japan has enough (or enough for its) consumers?