James Galbraith on flim-flamming “New Keynesianism”

2 May, 2014 at 11:55 | Posted in Economics | 56 Comments

Dear Simon Wren-Lewis:

You ask for some help with Keynesian views. Here are my thoughts:

“What New Keynesian theory does need is that falls in real interest rates stimulate aggregate demand (i.e. some form of IS curve), and in the basic model this comes from changing the intertemporal pattern of consumption. Is that wrong?”

Yes, it’s wrong. Keynes’s theory is a theory of consumption, based on income. Saving is the residual. Interest rates do not affect aggregate saving; hence the intertemporal pattern of planned consumption plays no role.

flimflam-2“What explains cyclical unemployment is real interest rates being at the wrong level. Movements in wages and prices get us out of a recession because they lead the central bank to reduce real interest rates. At the zero bound they cannot do that, and in those circumstances wage and price flexibility could make things worse. Is that wrong?”

Yes. Keynes did not offer a theory of “cyclical unemployment.” He offered a theory of underemployment equilibrium. Interest rates do affect business investment, and in the modern world, household durable-goods consumption. But that effect, which bears on cost, is usually small compared to that of the “marginal efficiency of capital,” also known as animal spirits. A decline in wages and prices per se increases the real interest rate (measured ex post), so the action by the central bank that you describe merely offsets this increase, but in any event, the effect is small since the psychology of a slump is overwhelming. Downward wage and price flexibility always makes things worse, not merely at the zero bound, since it always increases the real burden of debt.

“Now it is true that the standard New Keynesian model assumes a labour market that clears, but a model that replaces this with labour market imperfect competition would not behave very differently. That is what I actually teach.”

Chapter Two of the General Theory abolished the supply curve of labor, hence the “labor market”, and transferred the determination of employment to the domain of effective demand. Businesses hire when they can sell the goods that their employees make, and not otherwise. Any model with a “labor market” is an anti-Keynesian model.

“Equally the basic New Keynesian model assumes rational expectations, but if we want to change this to a case where agents make predictable errors that is easy enough to do.”

Rational expectations does not belong in any model that calls itself Keynesian. For a Keynesian theory, you do need to assess expectations, but whether they are found to have errors, ex post, is not important. What matters is the state of belief at the time decisions are made.

“(I think what I teach is pretty close to how many central bankers think, if not the rest of ‘my tribe’!)”

Not in my experience, though I may have known a different generation of central bankers. But that’s a detail.

I hope this helps clarify some of the key issues.

James Galbraith

 

About these ads

56 Comments »

RSS feed for comments on this post. TrackBack URI

  1. Ouch, Galbraith ate him alive.

    As every lawyer says: during trial never ask a question you don’t know the answer to. Wren-Lewis asked 5 such questions.

  2. As a resident of Japan since the 80s, may I say how much better Professor Galbraith’s explanation accords with my experience of deflation than Professor Wren-Lewis’s superficial treatment of these questions? If you haven’t experienced an actual deflation you are at a disadvantage in dealing with the questions Keynes raised. That disadvantage can be overcome by proper theory, as Prof Galbraith shows, but usually is not, as Prof Wren-Lewis demonstrates. He can perhaps be forgiven, for in a recent post he states that he is not interested in discovering what Keynes actually meant. (One fears, however, for his students.)

    In my experience ( and analysis) of Japan, I cannot emphasise enough the importance of a Keynesian income-based model of consumption. To be honest, I am becoming a little tired of the orthodox who know nothing of our age’s great deflation (which may have just ended after 16 years), yet feel in a position to offer theories which have no grounding in or relevance to an actual deflation as it is experienced in a modern , credit dependent, production economy. And I greatly regret that so much good work done in Japanese fails to garner attention outside this country.

  3. Oh, I like the way Galbraith phrases that. “Any model with a “labor market” is an anti-Keynesian model.” I’m putting that into the bag to pull out for later.

    For the usual trolls who will inevitably say “OMG! No way! There HAS to be a market for labour”, I would reply: then why do wages literally never fall when unemployment climbs? As we can see in the following graph we occasionally see a SLOWING of wage growth when unemployment climbs. But we NEVER see wages fall.

    http://bit.ly/1iRJhJU

    The half-baked answer I’ll get is predictable: “Oh, well they are flexible in the long-run…”. But then when does this long-run show up in the data? And if it doesn’t then what is its ontological status?

    • “The half-baked answer I’ll get is predictable: “Oh, well they are flexible in the long-run…”. But then when does this long-run show up in the data? And if it doesn’t then what is its ontological status?”

      First of all, just like Einstein did not substitute Newtonian physics but amended it Keynes did not substitute but amend the classical theory as well (and I hope everybody here knows that Keynes labeled the GT after Einstein). So heterodox economists who claim that there is no short-run and long-run are actually as anti-Keynesian as the New Keynesians who consider price or wage rigidity as the reason for the underemployment equilibrium.

      About why wages are stick, well, it is excusable that you have not read or understood the GT. But just observing the data (yawn, we have know that wages are sticky for decades) without offering any explanation is simply not science. There are ample of mainstream explanations: efficiency wages due to asymmetric information (Stiglitz) or psychological respectively sociological issues (Akerlof), implicit contracts and so on. Once you can provided a heterodox mechanism you would actually be engaged in science. Until than it is just polemics (I am all for criticizing the mainstream, as I wrote above mainstream macro models, RBC plus sticky wages, is total nonsense. But I don’t make a posture out of it and I never bash a model unless there is a superior one.)

      • Wages (and prices) stick because they are the outcome of political power struggles, not Walrasian double-auctions.

        And everybody principally involved in those power struggles have much stronger motives to not permit wages and prices to decrease than they have to prevent their increase. Partly because both firms and labour have large fixed nominal costs – principally debt service, partly because of a range of, more or less rigorously documented, cognitive quirks and shortcuts humans use when determining whether to accept or reject an offer.

        All of this has been in the literature since at least the late 1940s (though not necessarily in the *economics* literature), and could have been in the literature since the early 1870s if people had been paying attention to the consequences of retiring the Greenbacks without offsetting debasement of the specie currency.

        - Jake

      • I’m not debating you as to what Keynes meant. But why not tackle my question: if wages NEVER move then why call them sticky? Wage stickiness implies that they are fixed in the short-run and flexible in the long-run. But what IS this mysterious long-run you speak of? You and I both know that the long-run has practical policy implications in New Keynesian macro (i.e. fiscal policy is ineffective in the long run and the government budget should balance over the cycle). But when DOES this long run take effect? When do wages adjust in this long run scenario? It sounds like metaphysics to me.

      • “[I]f wages NEVER move then why call them sticky?”

        In the data you linked to the growth rate of wages fluctuates quite a bit so I don’t follow you here.

        Obviously political economy has an impact upon wages and declining union power and the increasing GDP share of capital are serious problems. But productivity is still a factor that influences wages.

        Whether you like it or not, Keynes never meant to substitute classical theory. he wanted to amend it which is why his book as the same title as Einstein’s. Now if you don’t wanna discuss it this is fine with me but I would really appreciate it if Post-Keynesian economists (which I appreciate in general very much, they are perhaps closest to Keynes among all the XYZ-Keynesians) would refuse from claiming that they are the only true apostles and at the same time propose something which is antithetical to Keynes’ writings (obviously the New Keynesian claim that wage and price rigidity are the key/only market failure that cause underemployment equilibria to exist is a far worse lie as Keynes made it crystal clear in the GT that more wage flexibility would decrease and not increase employment).

      • As an addendum, I think there is some merit to the Post-Keynesian notion that economies are always demand-constrained due to imperfect competition in the output market. But claiming or implying that this is in snych with Keynes is simply factually wrong.

      • Keynes, like Fisher and Robinson, is one of those economists whose mention in historical discussions should always be followed by a vintage.

        The Keynes of 1933 did not hold the same views, nor did he propose the same theories, as the Keynes of 1937, or the Keynes of 1920.

        There is a consistent red thread there, and one may reasonably extrapolate that to some extent. But discussions of “what Keynes wrote” are IMO futile, unless it is specified which Keynes we’re talking about.

        - Jake

      • Since you are not willing (or able) to answer my question then I suppose I will answer your’s.

        In the General Theory Keynes sought to answer the mainstream on its own terms. The early drafts of the book are very much more so “heterodox” but Roy Harrod, among others, convinced him that he should more so engage with the dominant theory of the day. When various studies came out in the late 30s on how prices are set Keynes changed his views explicitly.

        HOWEVER, this makes no difference to the above discussion. In the General Theory there is no labour market in the sense that a fall in wages will not bring about a rise in employment. Keynes is quite explicit about this when he writes:
        _____________
        “But it would, I think, be more usual to agree that the reduction in money-wages may have some effect on aggregate demand through its reducing the purchasing power of some of the workers, but that the real demand of other factors, whose money incomes have not been reduced, will be stimulated by the fall in prices, and that the aggregate demand of the workers themselves will be very likely increased as a result of the increased volume of employment, unless the elasticity of demand for labour in response to changes in money-wages is less than unity. Thus in the new equilibrium there will be more employment than there would have been otherwise except, perhaps, in some unusual limiting case which has no reality in practice.

        It is from this type of analysis that I fundamentally differ; or rather from the analysis which seems to lie behind such observations as the above….

        If this is the groundwork of the argument (and, if it is not, I do not know what the groundwork is), surely it is fallacious. For the demand schedules for particular industries can only be constructed on some fixed assumption as to the nature of the demand and supply schedules of other industries and as to the amount of the aggregate effective demand. It is invalid, therefore, to transfer the argument to industry as a whole unless we also transfer our assumption that the aggregate effective demand is fixed. Yet this assumption reduces the argument to an ignoratio elenchi. For, whilst no one would wish to deny the proposition that a reduction in money-wages accompanied by the same aggregate effective demand as before will be associated with an increase in employment, the precise question at issue is whether the reduction in money-wages will or will not be accompanied by the same aggregate effective demand as before measured in money, or, at any rate, by an aggregate effective demand which is not reduced in full proportion to the reduction in money-wages (i.e. which is somewhat greater measured in wage-units). But if the classical theory is not allowed to extend by analogy its conclusions in respect of a particular industry to industry as a whole, it is wholly unable to answer the question what effect on employment a reduction in money-wages will have.” (GT, Chapter 17).

      • I already wrote “obviously the New Keynesian claim that wage and price rigidity are the key/only market failure that cause underemployment equilibria to exist is a far worse lie as Keynes made it crystal clear in the GT that more wage flexibility would decrease and not increase employment”.

        We two agree with Keynes that wage flexibility would be detrimental in the short run. But you are simply at odds with the data when you claim that wages don’t move and you do disagree not just with those nasty mainstream economists but also with Keynes himself if you claim that there is no long run and that Keynes’ thought that his theory holds in the a long-run theory.
        The only long-run issue Keynes wrote about in the GE is what he called low marginal efficiency of capital (which implies low investment) and what we nowadays call secular stagnation (connecting low investment, i.e. low demand of capital, with a large savings, i.e. a large supply).

        If we connect the dots between secular stagnation and conventional Keynesian short-run macro it becomes clear that secular stagnation makes monetary policy ineffective (the Wicksellian natural rate of interest is persistently very low or negative, i.e. we are permanently in a liquidity trap, so monetary policy loses traction) and deficit spending is the only medicine. And in such a situation an economy might very well be permanently demand-constrained and labour markets might not clear (your claim that there is no labour market is simply total nonsense, you cannot ignore what is happening in the labour market and all progressive economists should care a lot about full employment).

        So yeah, the Keynesian short-run can become a long-run in this particular case. Otherwise the long-run behaves according to classical laws (which does not imply that there are not numerous market failures, imperfect competition, externalities and incentive problems due to asymmetric information).

      • We all know what Keynes thought of the long run (we are all dead!). Still, if you want to engage on this point answer my question: when does the long-run show up in the data?

      • “Still, if you want to engage on this point answer my question: when does the long-run show up in the data?”

        For ordinary recessions it usually suffices to check GDP growth (or just look at your data, recessions are marked there) and when we are at the ZLB you better also take a look at labour market data (obviously we are right not still in deep sh*t).

        But you are not really asking anything, you are denying that wages move at all and that labour markets exist (which is obviously utterly crazy) in order to support your “theory” that there is no long run.
        This is, as I already pointed out, anti-Keynesian (it is called General Theory and not New Theory) and I have little interest in engaging with people who disagree with fundamental Keynesian points, be it heterodox people who think that it is always the short-run (good luck finding any explanation for the wage growth in your data without long-run analysis), New Classicals who think that it is always the long-run or New Keynesians who think that more wage and price flexibility would undo the very existence of underemployment equilibria.

      • “When” is a really odd question here, but I think the long-run kicks in at about 10 year frequencies.

      • You’re simply not answering a very simple question: if wages are flexible in the long-run then when does this flexibility show up in the data?

        If the assertion is a testable hypothesis we should be able to see the long run flexibility of wages (and prices) in the data. I want to know how I would go about testing this.

        Pontus, this is empirics. I think that you’re out of your depth. Better to stick to the metaphysics.

      • I’m not going to do research for you here, Phil. But if you would – which you obviously don’t – know the difference between the time and frequency domain, you would understand that I’m talking about the frequency domain. You can look at wages at a, say, 10-15 year frequency and study their comovement with output per hour. That will give you a good idea of wage flexibility in the long run.

      • Oh Pontus…

        And when do the full employment effects appear? When I die and go to heaven? Haha!

        Go on now, back to your blackboard squiggles and your regressions and leave the adults to discuss economic policy.

      • Touche Phil. You really nailed that one, didn’t you? Well done!

      • Pontus, co-movement around a 10- or 15-year frequency peak is not sufficient basis for inferring long-run convergence toward a stable equilibrium attractor.

        Once you have non-linear effects in your system (which you oh so very much do in economics), you need to rule out complex dynamics before you can infer a stable equilibrium attractor. If you have complex dynamics in your system, then boundedness and co-movement in the frequency spectrum are also consistent with a variety of other attractor configurations, most of which have very awkward dynamic properties.

        Hydrodynamics offers plenty of examples of systems which are bounded and have variables that co-move on characteristic frequencies, but for which it can be proven that no unique “long run” exists – the system is quite literally a series of short runs, and innovations are amplified, not extinguished, so no “long run” can ever set in.

        - Jake

      • Oh dear, here comes the technodox to save the day! Strange bedfellows these hetero- and technodox.

        And no Jake. There are no (zero, zip, nada) ways of establishing anything “sufficiently” in social science using observational data. Evidence will simply never be conclusive, but just more or less compelling. Of course this leaves rooms for the real postmodern relativists — like Syll — to “reject” the whole notion of statistical inference based on observational data.

        Although I find it very unlikely that the precise mathematical curiosities that you bring up are the reasons for this.

        But whatever.

      • Not that it matters anyway, JakeS, because in order for this long-run stuff to have policy implications — like it does in New Keynesian economics — then it must actually appear in the immediate data at some point. Otherwise the long run is, indeed, a series of short runs in a highly complex and open system and people like Pontus are sitting around playing with their mathematics waiting for Kingdom Come. When you approach them with this point, of course, they hide behind the latest statistical technique that they’ve read about to sound clever. But everyone else can see that they’re being vacuous. Which is why no one listens to them.

      • Yours truly — a die-hard critical realist — a “postmodern relativist”? That got to be some weird Cambridge joke …

      • Besides, he’s not even talking about the same thing. He’s saying if that we can find correlation between real output per hour and wages then we know that wages are flexible. But, of course, wage flexibility means far more than simply that it tracks productivity. It also mean — crucially for New Keynesian theory — that it clears the labour market.

        Pontus is one of those awful mathematical economists who thinks that empirics means proving a correlation between two variables using fancy (but actually not that difficult) techniques like transforming a time domain into a frequency domain rather than making inferences about the world using the intuitive insights of a model.

        Besides, even if we take him on his own terms, a cursory glance at the data he is talking about shows that in the period 1965-1981 hourly wages grew at a consistently higher average rate (6.6% y-on-y) than in the period 1981-2013 (3.2% y-on-y) while hourly output per worker had around the same average growth rate in 1965-1981 (1.9% y-on-y) as it did in 1981-2013 (2.0% y-on-y).

        Now if output per worker is growing far faster relative to hourly wages in one period than it does in the other what might this suggest about productivity and income distribution… hmmm… Nah! Best not to inquire. Best to find some other way to engage in evasion by playing with the data.

        http://bit.ly/1jziba7

      • Pontus, in which fictional alternative universe is a unique equilibrium attractor a more plausible prior than all other possible combinations of attractors? Unless you have a compelling theoretical reason to prefer a unique equilibrium attractor – which you don’t – then you are making a highly unparsimonious inference here.

        Besides, if you’re doing frequency spectrum analysis, then testing for non-linearity is a basic misspecification test that you should be doing anyway.

        - Jake

      • Um, Jake, he DOES have a compelling reason to choose a unique equilibrium prior as his specification: because it is in line with the formal properties of how he models the economy.

        The process goes like this: Imagine extremely dubious thought experiment in which a simplified and unrealistic model of the economy converges on a unique long-term equilibrium => Assume that this thought experiment reflects the real world because that means what you do for a living isn’t nonsense => Try to prove certain properties of this thought experiment by searching for unique long-term equilibrium in the data (using techniques that may be telling you something entirely different anyway).

        It’s what I like to call Matthew 7:7 methodology…

        “Ask and it will be given to you; seek and you will find; knock and the door will be opened to you.”

      • Oh Phil, you are sometimes just too damn cute. The graph you showed compared nominal wages with real output per hour. I thought any idiot would know that if inflation was, say, higher before 1981 (pre Volcker) than after, then it would look as if wage growth was higher simply because of the escalating price level. If you look at real wage growth instead there is no such pattern: http://research.stlouisfed.org/fred2/graph/?g=zZ4

        Oops.

        And no, only the garden variety new Keynesian model relies on market clearing in the labor market. Any model with a frictional labor market doesn’t. So this is just wrong.

        Jake, more technobabble from you. There are pretty strong reasons to believe in a cointegrating relationship between wages and productivity. That the stability shown in the graph above would be an artifact of “variety of other attractor configurations, most of which have very awkward dynamic properties” is a hard sell.

      • “Oh Phil, you are sometimes just too damn cute. The graph you showed compared nominal wages with real output per hour. I thought any idiot would know that if inflation was, say, higher before 1981 (pre Volcker) than after, then it would look as if wage growth was higher simply because of the escalating price level.”

        I wonder how he’ll rationalize his inability to distinguish between nominal and real. I am really interested in discussing with heterodox economists as some of them have great ideas … but talking with folks who make basic errors which would be embarrassing for an undergraduate is a total waste of time.

      • Pontus… New Keynesian models require labour market clearing IN THE LONG RUN. Frictions are, by definition, a short run phenomenon. Have you already forgotten what we’re talking about? Are we banging up against that communication barrier again?

        And even if we take the real data into account (which I find absurd theoretically but let’s play regardless). Well, then we have to take different periods to account for the real wage erosion in the 1970s inflation. Here’s what we get:

        Period I 1965-1973
        Avg. Real Wage Growth: 1.8% y-on-y
        Avg. Productivity Growth: 2.7% y-on-y

        Period II 1973-2013
        Avg. Real Wage Growth: -0.2% y-on-y
        Avg. Productivity Growth: 1,9% y-on-y

        That still doesn’t make any sense. In Period I productivity growth grows about 30% faster than in Period II and real wages growth chugs along nicely.

        But in Period II while productivity slows it does keep growing. Meanwhile real wage growth goes negative!

        No tidy correlations here folks! Back home we go!

      • Pontus, dude, if you’re going to bullshit with numbers, find a more credulous audience.

        There is absolutely nothing stable about the relationship in that graph, other than the fact that both variables are clearly bounded within roughly the same range. Which you will trivially see if you pull the actual data, compute the candidate cointegrating relationship and plot it.

        Or, you know, just make a scatterplot of one against the other. You don’t even need specification tests to call bullshit here.

        If you put in enough break dummies you’ll get a well-specified cointegrating relationship, of course, and all that costs you is any pretense at building a model of cause and effect, rather than a concise but contentless condensation of the data.

        Or you can join the 90 % of the economics literature which goes “specification testing? I don’t need no steenkin’ specification testing.” That naturally makes your life a lot easier, and all it costs you is a ticket to the pseudoscience crazy-train.

        As an aside, a cointegration relationship does not imply an equilibrium attractor, except under highly restrictive assumptions which are unlikely to be realized in any actual economy.

        - Jake

      • Hold on Phil, did you just change the years? Comparing an 8 year period with a 40 year period (65-73 with 73-13)? Why on earth would you do that? To move the goal post?

        Let’s check.

        Average growth in labor productivity according to your initial dates:

        1965-1981: 1.6%
        1981-2013: 2.0%

        Average growth in real wages during the same years:

        1965-1981: 0.08%
        1981-2013: 0.12%

        Yep, that’s the reason! Cherry-picking dates to get your desired results. Very ugly, Phil.

        Lastly, “New Keynesian models require labour market clearing IN THE LONG RUN. Frictions are, by definition, a short run phenomenon.” is simply wrong. New Keynesian models do not – per construction – require labor market clearing in the long run. And frictions are not, by definition, a short run phenomenon. I am starting to believe you might actually be a bit, eh, special?

      • I don’t follow the New Keynesian literature closely (why would I? do you follow the Sraffian literature month by month?). And I’m sure some “revolutionary” has made long-run New Keynesian models in which there is no long-run market clearing of the labour market. Good for them. But this is not how those models are generally understood by the discipline. Don’t be dishonest. You know this.

        Now, onto the empirics. I explicitly stated that I changed the dates, and I quote, “to account for the real wage erosion in the 1970s inflation”. What I am trying to show — as any competent empirical economist would realise — is that there have been changes in distribution over the past 50 years that cannot be accounted for by changes in productivity. If we use nominal wage data we will see these changes most clearly in the years 1981+, if we use real wage data we see it most clearly in the years 1973+.

        If you are uncomfortable with comparing an 8 year time period to a 40 year time period would you prefer that I break it down into 6 different 8 year time periods? I guarantee that I will get basically the same results. So, come on, no cheap rhetorical tricks. You said that we could check for wage flexibility using this data (I disagreed but agreed to play the game). Well, I don’t see it. Sorry. Institutional changes have instead effected different distributions. And if you read anything outside of your statistical handbooks (like a history book maybe) you will know why this is.

      • Mind you, a quick Google search for “long-run labor market frictions” gives rather disappointing results. Perhaps no one has informed Google about this wonderful discovery yet…

      • Do you seriously wanna argue that productivity has no influence at all upon wages? Do you seriously think that the low human capital job at the supermarket or in the factory which you did to sustain yourself while studying is more productive than your current job which is more human capital intense?

        I am a radical left-winger, I recently read some books about the French Revolution, I am for 70% top marginal income tax rates (like in the well-known recent Diamond peter) and I think that the decline of the labour share of income and the rise of the capital share of income is ENTIRELY political economical.
        But I will not lie and pretend that political economy is the single only factor that influences wages. Productivity is the other significant factor and any kind of good economic analysis takes both factors into account. If some left-wing or right-wing economists lie and claim that only one factor influence wages they are engaged in ideology and not economics.

      • Well, Funky, I’m NOT a radical left-winger and I think that allowing your politics to colour your economics too much is vulgar. Now that we’ve all established what club we’re members of I’ll answer your question.

        Yes, higher productivity jobs often have higher compensation than lower productivity jobs. But sometimes low productivity jobs have extremely high compensation (finance! hello! or celebrity hairstylists!). It all depends. Go out into the world and judge for yourself. A rule of thumb saying that “high productivity factory workers will likely receive more than low productivity people pushing trolleys” but for God’s sake don’t turn it into some ridiculous axiom.

        Teaching students that “The Market” (reification alert!) distributes income according to productivity is enormously misleading. And guess what: only those with “left-wing tendencies” will ever look outside the box and try to explain why the model isn’t working. If you could see beyond the tip of your own nose you would notice that the garbage theory you propagate is actually giving the “other side” all the ideological cover they need.

      • Phil,

        “If you are uncomfortable with comparing an 8 year time period to a 40 year time period would you prefer that I break it down into 6 different 8 year time periods? I guarantee that I will get basically the same results. So, come on, no cheap rhetorical tricks. You said that we could check for wage flexibility using this data (I disagreed but agreed to play the game). Well, I don’t see it. Sorry.”

        Nah, I can do it much better myself (it’s amazing that you still don’t understand what a tool you look like when you don’t know the difference between nominal and real variables):

        I calculated the 15 year moving average from 1965-2013 for both real wages and real productivity per hour. I then regressed wages on productivity. The estimated slope coefficient is 1.59 (with confidence bounds 1.50 and 1.68), meaning that one percentage point rise in labor productivity growth is associated with a roughly 1.6 percentage point rise in real wage growth. The R-square is 0.9. Beat that.

        So that guarantee was pretty much as solid as your knowledge of economics, which is, as we know, laughable.

        “I don’t follow the New Keynesian literature closely”. So maybe you should do what people of average intelligence do when they don’t know something: shut up instead of acting like a yapping little ignorant poodle making a fool out of yourself. It will definitely improve your image.

        “But this is not how those models are generally understood by the discipline. Don’t be dishonest. You know this.” Now we’re being special again Phil. I explicitly wrote “And no, only the garden variety new Keynesian model relies on market clearing in the labor market. Any model with a frictional labor market doesn’t.” So I’m not sure how I’m being dishonest here.

        “Mind you, a quick Google search for “long-run labor market frictions” gives rather disappointing results. Perhaps no one has informed Google about this wonderful discovery yet…”

        Sometime I get the feeling that being heterodox is one big competition of using the work “rather” as some false modesty.

        Here why you got “rather disappointing results”: all labor market frictions are of a long-run nature! You searched for the equivalent of “blue blueberries” and use the lack of webpages devoted to the topic as evidence against blueberries being blue! But if you’re not too busy, perhaps you can take a quick look at Dale Mortensen’s and Christopher Pissarides work on the topic. They did get the Nobel prize for this invention. It’s been around for 30 years now. Welcome to the modern world.

        Is Kingston college a place for special people? I’ve never heard of it before, I hope it’s not their best and brightest I’ve met. That would be a scary thought.

      • Re Mortensen & Pissarides, I think that says more of what kind of prize the “Nobel prize” is, than the true quality of the models …
        http://larspsyll.wordpress.com/2014/01/23/on-dsge-and-the-art-of-using-absolutely-ridiculous-modeling-assumptions/

      • “Do you seriously wanna argue that productivity has no influence at all upon wages?”

        No, that would obviously be stupid. Productivity determines the size of the cake which can be split, and does influence the bargaining power of the various factors.

        Now, staying in simplistic straw-man land, do you seriously wanna argue that the size of the cake is the main determinant of how it is partitioned, and not who holds the knife?

        “Do you seriously think that the low human capital job at the supermarket or in the factory which you did to sustain yourself while studying is more productive than your current job which is more human capital intense?”

        My particular job, no. But I could name “high human capital” jobs which are.

        “I calculated the 15 year moving average from 1965-2013 for both real wages and real productivity per hour. I then regressed wages on productivity. The estimated slope coefficient is 1.59 (with confidence bounds 1.50 and 1.68), meaning that one percentage point rise in labor productivity growth is associated with a roughly 1.6 percentage point rise in real wage growth. The R-square is 0.9. Beat that.”

        Please supply the results of your misspecification tests as well.

        You did perform specification testing, right?

        Or, failing that, provide your residuals for the benefit of those people in this thread who can actually do time series statistics to a level that wouldn’t flunk you out of a intermediate undergraduate course on applied statistics.

        Until and unless you do that, the numbers you provided are just blinking lights on my screen, because I have no way to verify that the estimators you are using are consistent – or even well-defined.

        P.S.: The Swedish Central Bank’s Memorial Prize is a bigger joke than the Peace Prize – the latter is at least a real Nobel prize.

        - Jake

      • Oh, here come the gadgets, right Pontus? This should be amusing. A few points.

        (1) Nominal and real wages was an error. We’re on the internet and I’m doing other things this morning, you know? Trying to score rhetorical points by pointing out an error outside of an academic paper that someone concedes to straight away says more about your character than my mistake.

        (2) Your regression results are spurious because you used too long a time period. Divide it up into six 8 year periods and you will get different results for each time period. This will suggest that you do not find a linear relationship and the relationship between wages and productivity is to be explained by some basket of confounding variable (to find out which ones, close Eviews and open a history book). [You can see this simply by breaking down the periods. See here: http://fixingtheeconomists.wordpress.com/2014/05/07/exploring-inequality-real-wages-and-productivity-growth/%5D

        (3) You're being dishonest insofar as you know that this is not what students are learning. You also know that it is not the sort of model that calls itself New Keynesian and bounces around policy circles.

        (4) Actually the Google result for "long-run labour market frictions" showed up lots of hits for "short-run labour market frictions" which is rather strange given that you have just written, and I quote, " all labor market frictions are of a long-run nature". Are you sure that YOU have been reading the New Keynesian literature, Pontus? Hahahaha!

        (5) I suppose I should insert some ad hominem attack in here about Cambridge but I only slag people off when they're being silly, stubborn or intransigent. I grew out of name-calling for the sake of name-calling -- especially with regards to highbrow insults about being a "retard" -- when I was about 16 year old.

        (6) Now that we've played your rather crude game how about we play mine? This data is going to be a little more sophisticated than FRED measures, so brace yourself. Here is a chart showing how various sectors in the US have contributed to inequality over 22 years [http://fixingtheeconomists.files.wordpress.com/2014/05/incomeinequality.jpg]. Now, tell me: do you think that the changes here are reflective of changes in the different levels of productivity over this time period? Do you think, for example, that the increasing share of the financial sector between 1995 and 2008 is due to their increased contribution to economic productivity? If you do you’re a moron and you should get another job.

        You see Pontus, it’s all very well to have a bit of back and forth where we try to prove to each other how much cleverer we are than the other person. But the reality is that income distribution is a tough topic. And it’s just not explained by the marginalist theory. People are coming to see this now that Piketty’s book has been released. So, your little stories are going to come increasingly under fire in the next few years. My only hope is that your pathetic need to outcompete with other people — in this case, me — will make you so stubborn that you won’t change your mind on this issue and you will sink into obscurity. That would, I think, make me feel as if I had done my job here.

      • “If you could see beyond the tip of your own nose you would notice that the garbage theory you propagate is actually giving the “other side” all the ideological cover they need.”

        It is pointless to argue with a braindead moron. How often do I have to point out that I as well as any other sensible person realizes that political economy and productivity influences wages whereas you and the Chicago retards think that only one of these factors matter?

        Of course productivity is not the only factor that determines wages as it merely influences supply. Labour demand also matters and those celebrity haircutters receive such a high income because they are in high demand. Did you sleep through your entire undergraduate studies or do I really have to teach you what demand and supply are?
        Classical musicians are still working more or less like in the 18th century but they earn far more due to demand. People in the financial industry do indeed, as you have pointed out, only earn so much for political economical reasons. The pay of somebody who works in industry is influences by union power, the pay of an ordinary haircutter who works for a small company is nearly totally influenced by market forces.

        Unlike you I am able to differentiate. If you wanna be a scientist and not just drive home an ideological point (wages are entirely influenced by political economy) you might start doing the same. But as you cannot distinguish between nominal and real variables and as you write utter nonsense like “labour markets do not exist” I don’t hesitate to calli you what you are: a postmodern scharlatan whose economics is as scientific as that of my grandma.

      • P.S. We’re talking past each other on the friction issue. I was originally talking about wage rigidities. Anyway, I don’t really want to discuss this because you have a tendency toward monologue on such points.

      • “No, that would obviously be stupid. Productivity determines the size of the cake which can be split, and does influence the bargaining power of the various factors.”

        I like how you do not differentiate at all between the numerous kind of jobs that exist. This one-dimensional thinking must make you a brilliant economist.

        Bargaining power matters for some jobs but not for self-employed or people who work in small companies.

      • (1) Really? It was an error? Was that why you wrote “And even if we take the real data into account (which I find absurd theoretically but let’s play regardless)”?

        You found it absurd a minute ago, now it’s an error. Come on.

        (2) Spurious? Huh, how do you know? Well, just for the hell of it I decided to play along in your little game. I split the data into 6 subperiods of 8 years each and re-ran the regression. Guess what? The slope coefficient is still 1.4 (with bounds 0.13 and 2.65). The R-square is still 0.7. This is an incredible fit using only six data points (the fact that the slope coefficient is still significant is quite remarkable and tells you something about how strong the relationship really is). But instead of taking this seriously you aim at obfuscating the data and present a bar chart (!). Why don’t you use a scatterplot and then you can actually see the pattern yourself.

        (3) This is ridiculous. You: “New Keynesian models require labour market clearing IN THE LONG RUN.” Me: “No they don’t. The garden variety does, but not in general”. You: “that’s dishonest because that’s not how students learn the model”.

        (4) “Actually the Google result for “long-run labour market frictions” showed up lots of hits for “short-run labour market frictions” which is rather strange”

        No it doesn’t. There are two sentences that match “short-run labor market frictions” on google first search page, and both are relating to the same paper published in the IER in 2011 with 5 (google scholar) citations. This is hardly seminal work. What is “rather” strange though is your impressive ability of not understanding that you shouldn’t lie about things that can be verified.

        (5) Give it your best shot. Say that Cambridge is a shitty university and that Kingston is much much better. It will definitely hurt.

        (6) Life is too short. I have a job to do.

      • Oh, Funky, you don’t wear insults well. At least your pal Pontus is good at it. You’re truly awful.

      • (1) I do find it absurd. But you wanted to play that way. So, I was in error on your terms. I did not argue with this (because I’m not stubborn).

        (2) Hang on… you split the data into 6 subperiods and ran the regression on these 6 subperiods? Oh, good God, Pontus! Are you serious? Did you really do this? Wow! Lars, I hope you’re listening to this. Here’s a nice confirmation of confusing running regressions with causal arguments. Save the tape. I think we should move on from this point, Pontus. I can’t believe you did that. What on earth did you think that would prove/disprove!?

        (3) Subjective how ridiculous this is, I guess.

        (4) Meh.

        (5) You’re really into this “clash of the egos” thing, aren’t you? Internet tough guy, I guess.

        (6) Yeah, I thought so. You might want to prepare yourself for this sort of data in the coming decade though. There’s a whole shitstorm of it coming. So, you’d better find some way of bending your models into pretzels to explain it. Otherwise you’re going to find yourself fading into obscurity. (Although I don’t see much evidence that your star has shone all that brightly anyway, so perhaps this is the wrong metaphor…).

      • “How often do I have to point out that I as well as any other sensible person realizes that political economy and productivity influences wages”

        How do you justify singling out supply, demand and productivity among all the varied and sundry factors going into the political power struggle?

        It sounds like you want to cordon off some subset of ways to obtain political leverage and call them “market forces” and pretend that they can be studied independently of the rest of the contest for power.

        That’s special pleading.

        ““No, that would obviously be stupid. Productivity determines the size of the cake which can be split, and does influence the bargaining power of the various factors.”
        I like how you do not differentiate at all between the numerous kind of jobs that exist. This one-dimensional thinking must make you a brilliant economist.”

        I like how you think that productivity is a property of each individual link in the value chain, rather than the value chain as a whole.

        “Bargaining power matters for some jobs but not for self-employed or people who work in small companies.”

        Heh.

        In my day job I do business with small companies. I guaran-fucking-tee you that bargaining power matters to people in small firms.

        “(2) Spurious? Huh, how do you know? Well, just for the hell of it I decided to play along in your little game. I split the data into 6 subperiods of 8 years each and re-ran the regression. Guess what? The slope coefficient is still 1.4 (with bounds 0.13 and 2.65). The R-square is still 0.7.”

        Are those naive or robust confidence intervals? What does the residual (partial) autocorrelation function(s) look like?

        Oh, and even if you don’t want to take the time to run a proper specification search, is it really too much to ask that you provide your goddamn regression residuals?

        - Jake

    • “You’re simply not answering a very simple question: if wages are flexible in the long-run then when does this flexibility show up in the data?”

      Sigh. As I wrote above this is totally pointless. I am not agreeing with the classicals that wages are ENTIRELY determined by productivity, in imperfectly competitive output markets a company has to distribute its monopoly rents somehow and how much capital and labout get depends on political power.
      I am also not agreeing with New Keynesians that wages flexibility would undo the underemployment equilibrium.

      And last but not least, I am not agreeing with heterodox people who are as anti-Keynesian as their Chicago brethren. Pontus is totally right, you guys have more in common with postmodern writers (there is no long-run, there is no labour market are is the most ridiculous nonsense I head in years) who pretend that truth is a very flexible notion than with economists who search for practical yet imperfect models and mechanisms to understand the truth.

      • Hey Funky, you write:

        “I am not agreeing with the classicals that wages are ENTIRELY determined by productivity, in imperfectly competitive output markets a company has to distribute its monopoly rents somehow and how much capital and labout get depends on political power.”

        Well tell me this now: do there ever exist economies with perfectly competitive markets in which what you call “monopoly rents” do not exist? If not, then why make this assumption to begin with? Why not just assume that the distribution of the social product is all about social/political/bargaining power?

      • “the distribution of the social product is all about social/political/bargaining power?”

        As I already wrote numerous time, you are just the mirror image of the New Classical who think that wages depends entirely on productivity.

      • New Classical = Bad Guy
        Me = Opposite But Same As New Classical
        Therefore: Me = Bad Guy

        Thanks Socrates for that illuminating exercise in logical deduction!

      • New Classical = Bad Guy
        Me = Opposite But Same As New Classical
        Therefore: Me = Bad Guy

        Thanks Socrates for that illuminating exercise in logical deduction!

        You still missed my point which was that you as well as the Chicago folks IGNORE all factors that influences wages (and as I pointed out above the influence of these factors depends upon which profession you analyze except for one. The only reason for this ignorance can be ideological. But even if it is not (it could just be stupidity) you are still ignoring all independent variables but one and the ensuing result is obvious total nonsense.

      • I don’t recall saying that wages were ONLY a matter of political. I just said that the marginal productivity theory was wrong. I can think of lots of factors that influence wages outside of class power etc.

      • No, Funky, we’re just subsuming all “market forces” (such as supply and demand) into the list of ways to gain political leverage, rather than according them a privileged position in which they are treated as independent from and co-equal with all other ways to obtain political leverage.

        If you want to understand what’s going on in the real world, you should start with the man with the gun and how he organizes production and tribute collection. Because trade piggy-backs on tribute, and production for trade piggy-backs upon production for tribute. Treating them as independent and coequal is an obscurantist nonsense.

        - Jake

  4. There’s a bit more to that thread, including from Simon, James, Nick Rowe, and Nick Edmonds, both above and below this comment.

  5. My research notes:

    Schumpeter says in History of Economic Analysis, “everybody is supposed to react to a particular kind of ‘real’ values, namely, to prices expressed in wage-units or prices divided by an average money wage per unit of labor, which is determined by bargains between employers and employees – a well-nigh desperate measure of simplification that makes results incomparable as between two different points of time unless wage rates are the same in both. But, there is an important exception to this postulate that people calculate in terms of real values in this sense: workman do so only insofar as they save and invest but not in their bargains about their labor; when they negotiate wage contracts they consider exclusively money wage rates.” p1175

    In the GT, Keynes does express his agreement with the classical theory that ‘central controls could succeed in establishing an aggregate volume of output corresponding to full employment. (Ch24) “It is in determining the volume, not the direction, of actual employment that the existing system has broken down.” He also states that “the enlargement of the functions of government, involved in the task of adjusting to one another the propensity to consume and the inducement to invest wold seem … to be an encroachment on individualism, I defend it, on the contrary, both as the only practicable means of avoiding the destruction of existing economic forms and in their entirety and as the conditions of the successful functioning of individual initiative.”

    Schumpeter states the classical theorists held that unemployment was merely ‘frictional,’ and that the neo-classical economists always assumed full employment as the norm, which he interpreted to mean that they understood employment to ‘tend’ towards full employment and no involuntary unemployment at all times. He then goes on, “it never occurred to the majority to look to business cycles for material with which to build the fundamental theory of capitalist reality.”

    In GT, ch22v, Keynes notes the trade cycle school but rejects it because he says that the majority would prefer increased income to increased leisure. And, in ch16, he addresses the division of labor. So, it does not seem to me that Keynes has rejected the notion of a labor market, but he is discouraged by its inefficiency.

  6. Wouldn’t wages be subject to the law of diminishing returns or falling rate of profits?


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Create a free website or blog at WordPress.com. | The Pool Theme.
Entries and comments feeds.