The Elusive Professor Krugman

Sometimes I become convinced that Paul Krugman made a mistake, and then when I go back and reread his post I find that he used his words very carefully, to artfully avoid being wrong while creating a misleading impression.  I was thinking about all his posts that claimed fiscal stimulus didn’t fail, because it wasn’t tried.  I recall that he kept pointing to graphs showing that declines in S&L government output roughly offset increases in federal output.  But what about taxes and transfers?  Aren’t they also fiscal stimulus?  When I went back and looked at one of the posts, here’s the conclusion I read:

But if they won’t say it, I will: if job-creating government spending has failed to bring down unemployment in the Obama era, it’s not because it doesn’t work; it’s because it wasn’t tried.

OK, I guess I can buy this.  After all, back in 2008 and 2009 Krugman had some posts pointing out that tax cuts were much less effective stimulants than government consumption and investment, at least according to some Keynesian models.  I don’t entirely agree with those models, but it’s a defensible argument.

But here’s my problem.  If taxes and transfers don’t count for much in 2009, then shouldn’t they also be disregarded in 1937?  I’m referring to the famous Keynesian explanation for the 1937-38 depression, one of the most severe in US history.  Keynesians often point to the tightening of fiscal policy that occurred in 1937 as the key factor that aborted the recovery.  But here’s the problem with that explanation.  Fiscal policy wasn’t all that contractionary in absolute terms, and more importantly the big changes in fiscal stimulus occurred in the area of taxes and transfers–the exact area that Krugman thinks are relatively unimportant.

The reduction in the budget deficit mostly reflected two factors.  First, the large one-time ”bonus” payments made to WWI vets in 1936 (an election year) was not repeated in 1937.  And a 2% payroll tax was instituted to pay for Social Security in January 1937.  As far as I know those are the major contractionary moves.  But it was always understood that the 1936 bonus payments were a one-time deal.  How could that have caused a severe stock and commodity crash in late 1937, as well as a sharp plunge in industrial production in late 1937?

I found the following data from Thayer Watkins at SJSU:

Year   RGDP   Cons.  Inv.   Gov.   Exp.  Imp.  Balance

1934  641.1  519.0  31.5  127.3  21.4  31.1  -9.7

1935  698.4  550.9  58.0  131.3  22.6  40.7  -18.1

1936  790.0  606.9  75.5  152.5  23.7  40.2  -16.5

1937  831.5  629.7  94.0  147.0  29.9  45.3  -15.4

1938  801.2  619.5  61.3  157.8  29.6  35.2  -5.6

I just don’t see it.  Government output does decline slightly in 1937, but is still far higher than 1934 and 1935.  If you apply a multiplier of 1.6, it should have reduced GDP by about 1%.  But RGDP rose more than 5% in 1937.  Then in 1938 government output rises by twice as much as it fell in 1937, and RGDP plunges by almost 4%.  (BTW, we should be using NGDP figures, but everyone else uses RGDP, and the qualitative results would be similar either way.)

The 1938 depression had two causes, or perhaps one proximate cause and two deeper causes.  The proximate cause was a sharp increase in real labor costs.  Nominal labor costs rose sharply in 1937, due to a powerful union drive after the Wagner act, which rapidly doubled union membership and led to a wave of major strikes.  The payroll tax also slightly boosted nominal labor costs (I believe by 1%, but am not certain.)  Because wholesale prices were pretty high in the spring and summer of 1937, at first the wage boost merely led to a sharp slowdown in growth.  But then prices plunged due to a worldwide bout of gold hoarding, which increased the purchasing power of gold.  This tightened US monetary policy and caused the WPI to fall about 9% between mid-1937 and mid-1938.  Now two factors were driving up real wages, higher nominal wages and lower prices.  A supply shock and a demand shock.  Output plunged.  Severe slumps almost always have multiple causes.

So which is it?  Is Krugman wrong about 2009?  Or is he wrong about 1937?  Keynesians can’t have it both ways.

BTW, I eventually did find the quotation I was looking for, in another Krugman post:

The Obama fiscal stimulus more or less evaporates when you look at it closely, and take state and local cutbacks into account; basically, all it did was to keep overall fiscal policy from being outright contractionary.  (italics added)

I feel like that poor hunter trying to nab the roadrunner.  He usually gets away, but once and a while I catch him.

About Scott Sumner 490 Articles

Affiliation: Bentley University

Scott Sumner has taught economics at Bentley University for the past 27 years.

He earned a BA in economics at Wisconsin and a PhD at University of Chicago.

Professor Sumner's current research topics include monetary policy targets and the Great Depression. His areas of interest are macroeconomics, monetary theory and policy, and history of economic thought.

Professor Sumner has published articles in the Journal of Political Economy, the Journal of Money, Credit and Banking, and the Bulletin of Economic Research.

Visit: TheMoneyIllusion

Be the first to comment

Leave a Reply

Your email address will not be published.