Thursday, May 16, 2013

Post Keynesians vs. Keynesians on Unemployment

LK has this quote on the subject:
"It is hardly necessary to say that Post Keynesians reject the ‘old classical’, ‘Bastard Keynesian’ and ‘New Keynesian’ argument that unemployment is due to the existence of a real wage above the equilibrium or ‘market clearing’ level owing to the trade union or government interference in the operation of the free market for labour. They also dismiss the ‘New Classical’ notion that unemployment is the (voluntary) result of intertemporal income-leisure choices by individual workers. As was demonstrated above, neither claim is supported by micro foundations; and neither has any macro foundation whatsoever. A Post Keynesian theory of unemployment would instead start from the proposition that in aggregate the level of employment depends on the level of output, which is itself determined by aggregate demand and therefore heavily influenced by macroeconomic policy. Unemployment is simply the difference between the level of employment and the aggregate supply of labour, which may - as explained earlier - safely be regarded as invariant in the short run with respect to the real wage, but variable with respect to the number of job opportunities.” (King 2002: 84)."
I don't think I quite agree with King, at least on the New Keynesian part (or the Post Keynesian part, for that matter) - or at least I think it's misleading. I think the importance of price stickiness for New Keynesianism is often misunderstood. People have this image in their heads of a non-clearing labor market when they hear "sticky prices". The market's just broken for a little while and eventually it'll get back on track. Austrian types don't like the implication that the market's broken. Post Keynesian types don't like the implication that it's just a hiccup.

But actually the price stickiness point has little to do with labor market directly. As I've pointed out in the past, the principle function is to explain why the short run aggregate supply curve is upward sloping and not vertical (that's why it's called New Keynesian and not New Classical - emphasizing non-vertical aggregate supply is very non-Classical). You might know the short run aggregate supply curve in more modern models as the Phillips Curve. Just rearrange things, switch from prices to inflation, and tack on Okun's Law and you've got the Phillips Curve. So price stickiness gives us the Phillips Curve tradeoff which, with the IS curve, can pin down an output level below a full employment level of output. [This is related, btw, to why I think the substantive difference between New Keynesianism and Old Keynesianism (not to be confused with Post Keynesianism) is overblown... but that's another discussion].

This is a lot less neoliberalish than Post Keynesians often make it out to be. This sentence by King:  "A Post Keynesian theory of unemployment would instead start from the proposition that in aggregate the level of employment depends on the level of output, which is itself determined by aggregate demand" could just as accurately be said of unemployment in a New Keynesian model.

There are of course lingering questions about tendency to move to the natural rate of unemployment and the stability of the natural rate of unemployment. It is true that New Keynesians are more likely to suggest that there is a strong tendency towards a stable (albeit changing) natural rate of unemployment. And there's probably good reason to say that - outside of liquidity trap situations the economy has seemed to exhibit that sort of behavior (and we've got good liquidity trap models when it doesn't). There's great discussion by New Keynesians in good standing like Mankiw, Ball, Moffitt, Akerlof, etc. about less conventional NAIRUs too.

The lack of a (necessarily) stark difference between New Keynesians and Post Keynesians on all these points is made forcefully by Stockhammer (2011)*. Of course, not all Post Keynesians agree. LK, in his post, also cites Lavoie and Davidson. It just so happens that Lavoie and Davidson are two Post Keynesians that are NOT inclined to agree with Stockhammer on this sort of point. It would be a mistake, though, to conclude that they speak for all Post Keynesians. LK is just sharing one side of the Post Keynesian position.

OK, so I say there's no real difference between New Keynesians and Post Keynesians on this particular issue raised by King. But surely there are differences between them, so what are they?

Leaving aside the point that someone like Stockhammer might specify or derive the Phillips Curve/IS Curve/NAIRU/MP curve nexus differently than someone like Mankiw I would say that a principle difference is the other contraptions they have that are often completely neglected by New Keynesians. These include:

1. Mark up pricing
2. Dividing the population into classical classes (with rentiers replacing landowners)
3. Various conflict theories of inflation
4. Explicit modeling of capacity utilization for inclusion in the investment function (capacity utilization is of course not much different from an output gap - but the inclusion in the investment function introduces accelerator effects which is interesting and [I think?] dropped from most modern analyses)
5. Modeling it with wage shares front-and-center so we can talk about distributional issues

There's plenty of heterodoxy in all that, chiefly through the distributional facets and the departure from marginalism. That, it seems to me, distinguishes Post Keynesianism. It's less a matter of one side talking about inadequate demand and the other not. In other words, to the extent that I'm a Post Keynesian (and I do like a lot of what they do) I'm a Stockhammer and (I guess? I should be much better read:) Pollin/Blecker/Dutt sort of Post Keynesian. I am less of a Davidson/Lavoie sort.

* "The Macroeconomics of Unemployment" in A Modern Guide to Keynesian Macroeconomics and Economic Policies, ed. Eckhard Hein and Engelbert Stockhammer


  1. Regarding the issue of nominal wage rigidity...

    There is a book on labour economics, Daniel Kuehn, that I think you would enjoy reading. It was written by a mainstream Keynesian economist at Yale University by the name of Truman Bewley, and published by Harvard University Press about ten or so years ago.

    The book is important because the research was done via good old-fashioned field-work. Bewley went out and interviewed businesspeople working in firms on this microeconomic matter. In the end, what Bewley found was that J. M. Keynes's observations in Chapter 19 of The General Theory of Employment, Interest, and Money were correct, while the writings of scholars who stayed closer to the position of A. C. Pigou and other classical economists, were wrong.

    Also, regarding the Post-Keynesians, Daniel...surely you're aware that there are different strands of Post-Keynesianism, right? There are "Kaldorians", "Kaleckians", "Robinsonians", "Shackleans", and "Keynes/Post-Keynesians" (the last being the strand from Davidson).

    Unless I'm mistaken, isn't Stockhammer a "Kaleckian"?

    1. Sure there's all kinds of internecine groups as with most heterodoxies. Yes, Stockhammer is often called Kaleckian although there are lots of different names. The innovations I listed are generally "Kaleckian" ideas too.

      The Bewley book is a good one. I've never read it all the way through, but have spent some time in it. I wish more economists did empirical research like that.

  2. "the principle function is to explain why the short run aggregate supply curve is downward sloping and not vertical"

    Do you mean upward sloping ?

    1. Gah! The important thing is "sloping" ;-)

      I was reading through some Stockhammer/AD material and I think that stuck in my brain in writing that sentence.

      Will change. Thanks.

  3. Here's, a question I'm interested in....

    I'm writing something that discusses the period of time GDP is measured over, one year. I haven't seen many attempt to justify using one year. It's certainly convenient because it matches what accountants do.

    I remember reading a Keynesian once who said that it's a good idea because the average length of a job contract is one year. So, it matches the period across which wages are most sticky. He said that Keynes used annual GDP for this reason. The problem is I don't know who wrote this, I thought it was Paul Davidson but I can't find it anywhere in his books.

    Does anyone know who made this argument?

  4. In the model you describe I think that in the short-term it would be true to say that "unemployment is due to the existence of a real wage above the equilibrium or ‘market clearing’ level". If wages (and other prices) adjusted immediately then the SRAS would also be vertical. Output would always be at the optimal level. I agree though that the story is more complex than just a "non-clearing labor market" one.

    I'm trying to make sense of the Post-Keynsian position as described in the post. King says:

    "A Post Keynesian theory of unemployment would instead start from the proposition that in aggregate the level of employment depends on the level of output, which is itself determined by aggregate demand "

    This (combined with the rejection of non-market clearing wages being a factor and belief in cost+mark-up pricing) seems to imply that they think the AS curve is not vertical even in the long-run (and logically might even be horizontal ?) and the only solution to unemployment is macroeconomic policy to stabilize AD. Do you know if that is the case ?


All anonymous comments will be deleted. Consistent pseudonyms are fine.