Why Old Keynesianism is looking worse these days, and other thoughts

I am happy to report that I heard another Paul Krugman talk a few nights ago.  Krugman had just flown in from Moscow, switching in Houston to a flight to Austin, with little time to spare.  Presumably he was very tired, he had a bad cold, he had to speak in a loud barbecue joint and bar, and yet he was entirely lucid and he charmed and instructed the crowd.  He even had praise for MarginalRevolution.  It was better and better received than if he had given a more traditional lecture.

The same Krugman has a recent blog post, or two, in which he is unhappy with me.  I will put some thoughts on that under the fold…

Krugman attacks me for being an anti-Keynesian when in fact I very much prefer New Keynesianism over Old, so the entirety of his critique is boxing at shadows.  As an aside, my first published article was in the Journal of Post Keynesian Economics, and as long ago as 1989 I wrote an essay arguing that Keynesian economics explained the Great Depression better than did alternative views; some of my libertarian and Austrian acquaintances still hold that piece against me.  Some MR readers also will recall the 12-part symposium I did on Keynes’s General Theory, full of praise for the book, though of course with some criticisms too.  I don’t expect Krugman to be an expert in the history of thought of me, but a) he has linked to previous posts of mine making clear my affinity for sticky price reasoning and other new Keynesian ideas, b) it is a rather simple proof that he has rather drastically misread me, as has DeLong, and c) if Krugman doesn’t know my views perhaps he should not attack them in such strident language.  By the way, to the extent I like Old Keynesianism, it is for the Shackle-Lachmann-Minsky strand, not IS-LM and the liquidity trap.

Moving on to substance, let’s start with a few reasons why I think the course of the recovery discriminates against Old Keynesianism, though not against new Keynesianism (definitions and contrast here).

1. New Keynesianism has a more optimistic attitude toward mean reversion than does Old Keynesianism.  Things are looking a bit better than expected half a year ago, so New Keynesianism gains in status over the Old.  Q.E.D.  That’s actually enough to establish the entire point.

2. Brad DeLong suggested not long ago that the short-run model, rather than the long-run model, will apply for a minimum of five years, and possibly up to fifteen years.  That claim is looking weaker these days and I am happy to admit that I am revising my own previous views on speed of adjustment.  It’s one thing to predict we won’t get to four percent unemployment for a long time, but I am now much more skeptical that the short-run model will apply for so many years.  (Since the labor force is not a homogeneous aggregate, the long-run model still can be more relevant than the short-run model even when there is residual unemployment.)  That is the view I am arguing against and Brad is not made of straw.

3. Krugman’s own writings show he is less worried about the particular perverse consequences of one version of the liquidity trap argument than say a year or six months ago.  Remember all the talk of the upward-sloping aggregate demand curve?  That would imply more output and the expectation of further output gains can make the liquidity trap worse.  Yet how does Krugman respond to the good job market report?  He claims, correctly, that it is good news.  Full stop.  There is no response like “it’s great those people have jobs but I’m worried that the extra output will worsen the liquidity trap.”  Common sense rules, as it should, though now it’s time to admit we have moved onto a different and better model.

3. Frankly, it is a bit of an embarrassment for many commentators that the (admittedly weak) recovery is coming right after the end of the fiscal stimulus.  Of course this does not refute the standard account of fiscal policy, namely that it can work but is hard to pull off politically in a manner which contributes to sustainable growth.  The correct answer for the timing of recovery, relative to the end of stimulus, is “confounding factors,” but that is exactly the point.  The confounding factors are more important than we had thought, and the fiscal stimulus not quite as important as we had been led to believe.   That is another point against the Old Keynesian view.

4. The Old Keynesian focus on the paradox of savings may have been relevant during the most dire points of the crisis, but it does not seem to be panning out in the recent data.

5. Liquidity trap models of unemployment stand in tension with Mortensen-Pissarides search models, as Krugman himself has noted in the past.  The notion that employer (and worker) expected profit influences search behavior and thus employment, seems to be showing up in the data.  That’s a real business cycle mechanism.

6. Corporate profits are strong.  Quite possibly the prevalence of labor-saving innovations helps explain why labor market recovery has been slower.  That too is a real business cycle mechanism.  The persistence of long-term unemployment, even in light of an improving labor market, suggests those particular workers face structural unemployment.  That too is a real business cycle mechanism.

7. A lot of people get upset when one praises “real business cycle” models, perhaps having in mind some overly simplified one-person contraption from the early 1980s.  The reality is that most economic phenomena and mechanisms fall under this heading, and that the objectionable simple models have been left behind a long time ago in favor of synthetic treatments.  I’m the one in the mainstream here (which doesn’t mean I am right), but it does make it odder to be pilloried and insulted for holding those views, as if it were some inexplicable tomfoolery not worthy of professional economists.

8. The old Keynesian view really does have trouble explaining turning points, unless there is a major fiscal or monetary policy action to account for the change.  This was well-known as long ago as the 1930s.  Brad DeLong cites a passage from Keynes about the depreciation of capital (see my previous coverage here, I am very familiar with this argument), but of course Keynes thought that mechanism was too weak and unreliable and thus he favored a mix of planning and nationalized investment.  In any case that mechanism explains only one chunk of the current recovery and furthermore I call a rising MPK a real factor too.

9. Krugman sometimes writes as if new and old Keynesian approaches are simply more and less rigorous versions of the same thing, but they clash on a number of important issues, a few of which I’ve mentioned above.  You don’t have to agree with Stephen Williamson on everything, or approve of his tone, to notice that many of his blog posts illuminate significant differences between New and Old Keynesianism.  Read through his archives.  There’s a lot of very smart stuff in there, even though it has become unstylish to respond to him very much.  He knows a lot about macroeconomics, usually more than the people he is criticizing, even if I often disagree with him on methodology.

10. Finally, it’s worth going through Krugman’s citations of his own successful predictions:

We said that as long as the economy remained deeply depressed, even a huge rise in the monetary base would not be inflationary, and that even huge budget deficits would not send interest rates soaring. And we said that fiscal austerity would be contractionary, not expansionary.

Funny enough, I also got those first two predictions right too, and on the third I have argued we haven’t seen real austerity yet.  I also was never very proud of myself for getting those predictions right, because I thought it did not show any particular acumen on my part, nor did I think it elevated me to an especially prophetic stature.  I actually prefer to think more about all the predictions I got wrong.  Here is Scott Sumner on the predictions he got right, an excellent record.  Perhaps Krugman’s successful predictions have come from a broader framework and not from Old Keynesianism per se?

In sum, one constructive response to the new data, and my post, would have been: “The Old Keynesian fortress actually is looking weaker these days.”

Another constructive response would have been to try to patch up the growing holes in the Old Keynesian approach, as applied to 2011-12, and try to salvage it vis-a-vis New Keynesianism and/or market monetarism.  No need to waste so much time on views which aren’t even in the running.

We didn’t get either of those.  This unfortunate outcome relates back to my earlier, broader point about focusing on the stronger arguments of the other point of view.  If one spends too much time knocking down and polemicizing against weak arguments, sooner or later all opposing arguments start looking weak.

Addendum: Mark Thoma adds comment.  And Ezra Klein comments.   Here is Krugman’s response.  There is no Orwellian move on my part, I’ve long been fine with new Keynesian ideas and there is a long paper trail to that effect, stretching back many years; to cite “real business cycle theory” in 2012 is in fact to refer to a bunch of them as part of the theory.  When it comes to RBC, Krugman is still living in 1983, a point which Stephen Williamson has made a few times as well.


Comments for this post are closed