Friday, April 18, 2014

Did Kalecki Accept the Labour Theory of Value?

I thought that Kalecki did not accept the labour theory of value, but Matias Vernengo raises a question about this in a comment here, and asks for the evidence.

I have to admit to relying on memories of some reading on Kalecki quite a while ago, but some further research yields this interesting statement from the Polish Marxist economist Włodzimierz Brus, which seems to be from his personal conversations with Kalecki (who had
returned to Poland in 1955):
“While recognizing the enormous significance of the Marxist approach and of many Marxian tools of economic analysis, Kalecki never felt that he had to accept every component of Marxian economics or to retain those parts of it which have become obsolete in view of new experience and of the progress of economic theory. He felt, for example, a strong distaste for the Marxian theory of value, which he considered metaphysical and (if I am not mistaken) never wanted to discuss.” (Brus 1977: 59).
Of course, Brus might be mistaken, but other writers on Kalecki do seem to agree that Kalecki pretty much avoided the subject of the labour theory of value in his writings, and his silence suggests that he did not support it (Sawyer 1985: 148; Toporowski 2004: 217; Chapple 1983: 551; King 2001: 609; Chapple 2013: 302: “He had no time for the labor theory of value”) or (at least) had no interest in defending it:
“Kalecki seems never to have declared himself in print as to where he stood in relation to the labour theory of value. He assumed that prices were fixed by firms according to their unit costs of production to which margins or mark-ups were added to secure certain levels of profit.” (Harcourt and Hamouda 2003 [1988]: 219).
Brus, Włodzimierz. 1977. “Kalecki’s Economics of Socialism,” Oxford Bulletin of Economics and Statistics 39.1 (February): 57–67.

Chapple, Simon. 1994. “Kalecki’s Theory of the Business Cycle and the General Theory,” in John Cunningham Wood (ed.), John Maynard Keynes: Critical Assessments. Second Series. Croom Helm, London. 538–559.

Chapple, Simon. 2013. “Kalecki, Michał,” in Thomas Cate (ed.), An Encyclopedia of Keynesian Economics (2nd edn.). Cheltenham, Edward Elgar Pub. 301–304.

Hamouda, O. F. and G. C. Harcourt. 2003 [1988]. “Post-Keynesianism: From Criticism to Coherence?,” in Claudio Sardoni (ed.), On Political Economists and Modern Political Economy: Selected Essays of G. C. Harcourt. Routledge, London. 209–232.

King, John E. 2001. “Kalecki, Michał,” in Phillip Anthony O’Hara (ed.), Encyclopedia of Political Economy. Volume 1. A–K. Routledge, London and New York. 607-610.

Sawyer, Malcolm C. 1985. The Economics of Michał Kalecki. Macmillan, Basingstoke.

Toporowski, Jan. 2004. “Kalecki’s Arguments for Socialism,” in Zdzislaw L. Sadowski and Adam Szeworski (eds.), Kalecki’s Economics Today. Routledge, London and New York. 215–221.


  1. Mind you, I also think that there are several metaphysical things in some versions of the LTV. Sraffa would have also thought the same, and I would argue Marx too. Sraffa emphasized the objective aspects, and produced a logical solution. Something missing in the marginalist supply and demand story. And even if Kalecki never said anything, which really makes it impossible to say he rejected it, the question is whether short term PK views on pricing are compatible with classical (and Marx's) long term prices. Both are based on the notion that prices are based on objective condition of production and exogenous distribution. Many Sraffians would suggest these are perfectly compatible.

    1. Thanks for this comment.

      Is perhaps possible to see cost-based pricing as a better version of, or successor to, the LTV?

    2. As long as we take care to note things Marx was actually purporting to show with his theory of value, I'm not sure a "better version" exists. The problem arises more when it's mistaken for a theory of short-term prices rather than the laws of motion of a capitalist economy more generally. But otherwise, the role of cost accounting has always been fundamental.

      The other problem is that there are quite a few interpretations of the theory vying for prominence. (See, e.g., this paper for a survey.)

      That said, if we're viewing the major divide in price theory through the lens of "neoclassical vs. accounting" traditions, the various LTV's definitely seem to share a side with the various PK traditions, so it's not unreasonable at all to see a line of succession there.

    3. I agree, the real distinction is between theories where prices are determined by the objective conditions of (re)production of a surplus (cost of production approach), and theories where prices are determined by subjective utility under scarcity conditions. Furthermore, the fact that an author does not discuss a given topic only gives more degrees of freedom to combine her/his perspective with that of others, provided they are not incompatible in other respects. Each authors focuses on different things, and the differences between them may often appear to be greater than they are due to the particular issues each was studying (effective demand, or the transformation problem). In Sraffa's 1960 book, input prices are equal to output prices, which fits well with theories of administrative prices in order to obtain a surplus, for example. And prices can be reduced to labour values even in those cases as Sraffa also shown.

  2. This is the quote I had in mind. I assume that, given the language, Kalecki's rejection was identical to Robinson's.

    I strongly disagree with the idea that because he never discussed it he therefore never rejected it. By that logic if he never discussed alchemy he never rejected it either. An historian of ideas should take from the fact that it does not appear in his work and there is anecdotal evidence of his rejection then he rejected it. Another piece of evidence is that the Cambridge Keynesians generally rejected it quite vocally and Kalecki ran in that milieu.

    1. No one said that “he never discussed it he therefore never rejected it” (that would not be a very intelligent thing to say). You must distinguish between the analysis of the thought of an author, and the synthesis of elements of that thought within another framework in a way that the author did not discuss. Knowing what the author thought on a given issue is absolutely essential for the analysis of her/his thought, but not for the synthesis of elements of her/his thought in a wider framework. And unless you believe that there is a given author or book who showed the whole truth and left nothing else to discuss, synthesis is always necessary. Furthermore, the fact that an author does not discuss in print a given topic, especially when it was so important, often means that the author was not so sure what to say about it. I remember some of Sraffa’s unpublished manuscripts where he is more hostile towards the labour theory of value (writing that labour is simply the name of the whole process of production), and others where he changed his mind. It is a mistake to assume that every author was very sure about everything. And to focus only on analysis/exegesis, while neglecting synthesis/reconstruction of economic theory, is a strategic error, a startegic error that partly explains why Post-Keynesianism never became dominant (it spent much of its time in internal fights between different camps and left the synthesis to the “neoclassicals”, who in turn had no concern for analysis/exegesis and distorted Keynes’ message).

    2. I really don't understand this obsession with trying to salvage the LTV. It is a completely outmoded doctrine that doesn't live up against even the most casual scrutiny. The vast majority of PK authors (and yes, I agree, Sraffa too) rejected it. But for some reason the present generation have insisted on digging up its bones. The only explanation is that this is about ideology.

  3. Sraffa thought we could reduce prices to dated quantities of labour.

    What casual scrutiny do you have in mind?

    1. What if all the labour were done by machines?

      How would the LTV work in such an instance?

    2. I've had this debate too many times so I'm going to anticipate something that you are probably going to do before you do it. In order to do this let us first clearly establish what you mean by the LTV. Do you agree with the following definition from the Wiki page:

      "The labor theories of value (LTV) are heterodox economic theories of value that argue the value of a commodity is only related to the labor needed to produce or obtain that commodity and not to other factors of production (except as those elements can be regarded as embodied labour.)"

      P.S. LK, the LTV crowd would say that those machines are all dated labour inputs.

    3. I find debates about the LTV tiresome.

      Out of interest, in this post, UnlearningEconomics defines the LTV as:

      value = “socially necessary” labour time required to produce a good + current necessary cost of the capital used up in production ("constant capital") + value added by raw commodities (which are used up completely and so add all of their value).
      UnlearningEconomics states that "Labour is generally paid less than the value it adds, and therefore is the sole source of profit."

      I assume all these variables are measured in money terms, but the fatal flaw appears to be that businesses do not charge mark-ups on costs and that prices are at costs of production.

      I cannot see how prices at the cost of production are in any sense a "natural" or normal element in real world capatalism or a realistic way of viewing prices.

      How can there be any long-run tendency towards cost of production prices, when the vast majority of businesses will charge a mark-up on costs, always.

    4. I guess they would say that the "mark up" is the source of the "surplus value".

      Note that assumption that those who receive "profits" add nothing to the production process. Anyone who believes that needs some serious and very immediate exposure to the real world, rather than remaining immersed in their socialistic fantasy world.

      The LTV is far worse than the marginalist theory of value. It is ten times as ideological.

  4. The fundamental contrast, as I said, is between theories that focus on the process of production when explaining value, and theories that focus on marginal utility. The problem that arises when you focus on the process of production concerns the standard of value you use to measure costs. You may use embodied labour as Ricardo, or socially necessary labour as Marx. Or you may reduce labour to the quantity of land that sustains the labourer, as Petty and Cantillon suggested, and focus on physical entities. Or you may not use an universal standard of value at all, and use instead a mixed commodity, as Sraffa did (which could be reduced to dated quantities of labour presupposed in the process of production). The definition you suggest from Wikipedia is too simplistic and does not take into account all the different ways in which these (and other) different classical authors addressed the problem. Whichever way you go (whatever standard you want to use to do your cost-accounting), all of these approaches follow a method (of focusing on the actual production process) which is much more solid than the marginalist method (which presupposes infinitesimal changes while assuming everything else constant, as Marshall noted).

    Marshall also noted that a common mistake when interpreting the classical authors is to forget that what they call the “cost of production” already includes the surplus, or the “mark-up”, and so the classical approach is entirely consistent with mark-up prices. In fact, when people like Fred Lee study mark-up prices, they are developing a classical theory: they are looking at the process of production and trying to find a way to explain prices in terms of the costs of production.

    Hope I was not too “tiresome”.

    1. And yet the fundamental problem comes already when you use the word "value".

      Mark-up pricing theory is describing and explaining price, not value.

      Failure to define and distinguish between "value" and "price" leads to intellectual catastrophe.

      "Value" has a legitimate and empirically defensible sense of the subjective worth or importance any person x places on a thing, and with respect to any commodity y the subjective utility/pleasure/satisfaction that a person derives from it.

      "Price" is simply the monetary measure of a good.

      Mark-up pricing is explaining price, not value.

      If "value" in the LTV is to be understood as "price", then the LTV seems straightforwardly false to me.

      When "value" in the LTV is to be understood not as price nor as "value" in the sense I have defined it above, then it seems to me it has collapsed it into a metaphysical sense, which is theoretically dubious and empirically falsified.

  5. Nuno puts it well. And indeed, where Marx speaks of "prices of production," he's already included in the figure an average (or "general") rate of profit over "cost price," mirroring what modern accountants would call standard or target costing.

    As far as Marx's particular theory is concerned, I would be happy to answer any questions about the much-maligned nuts and bolts of it. If everyone's already all sick and tired of debate before it begins, then I will offer nothing more than that; just answers to questions.

    If there's no interest in that, either, that's also fine; just figured I'd offer.

    1. Do you accept at least that the word "value" has the following legitimate and empirically defensible sense (amongst it various menacing):

      "the subjective worth or importance any person x places on a thing, and with respect to any commodity y the subjective utility/pleasure/satisfaction that a person derives from it."?

    2. If my usage of the word "value" confused you, you may replace it by "price", which is what Sraffa actually does in the 1960 book (in page 9 he explains why). There is no harm in doing so as a temporary working hypothesis.

      Metaphysics is someting you simply cannot avoid, it will appear at some point, as I am sure you know.

      I find definitions of value in terms of subjective utility highly ambiguous, since mental states are highly volatile.

      Hedlund, I just read the article by Rob Bryer you mention above and I liked it a lot. Could you send me any further information you have on the topic of Marx's exchanges with Engels on accounting to my e-mail, (unless other people are interested too and you prefer to discuss it here)

    3. Thanks for this answer, Nuno Martins.

      But even if we interpret "value" to mean "price" we have serious problems with the LTV.

      Do you accept this as a valid definition of the LTV?:

      value = “socially necessary” labour time required to produce a good + current necessary cost of the capital used up in production ("constant capital") + value added by raw commodities (which are used up completely and so add all of their value).
      First, it is not labour time per se but average unit cost of labour at a given quantity of output produced that is used to calculate price.

      Secondly, calculation of "constant capital" in production of any individual good x seems extremely difficult if not impossible to calculate, as far as I can see.

      Finally, what businesses actually set prices in this way? None, as far as I am aware. It is empirically irrelevant.

    4. If you replace "value" with "price" as a temporary working hypothesis, what you get is not what you write above, but rather something like Sraffa's system. I see no problem in using something like Sraffa's system while assuming administered prices. You can set prices to achieve a given target rate of return, just like Sraffa sets a rate of profits and sees what prices emerge. And you can assume that prices are fixed administratively since Sraffa is focusing on the prices that are consistent with a given reproduction process. You can also focus on average costs rather than variable costs by adding fixed capital, as Sraffa does. Surely something to think about, I guess.

      As for the theory of value, I have published a book last year where I discuss it, I didn't mention it earlier since I don't like using blog's comments to do self-promotion. But luckily, Amazon's webpage has put some pages free online which are exactly the pages with the preface and the first part of chapter 1, which is precisely on the theory of value. So if you want to have a look (and as I said, most of what is relevant for our discussion can be seen in the "look inside" option), here it is:

    5. Do you accept at least that the word "value" has the following legitimate and empirically defensible sense

      "the subjective worth or importance any person x places on a thing, and with respect to any commodity y the subjective utility/pleasure/satisfaction that a person derives from it."?

      I do accept that definition as valid for some considerations, but not others. In the case of any classical value theory, it is not. Basically, whenever I discuss conflicting theories of value, I find the most helpful thing to do is to just eliminate the v-word from the discussion altogether, since it becomes very easy for all parties to lapse into equivocation without realizing. Subjective value, I will call "utility," and value in a classical or objective sense is "socially necessary labor time" (or just "SNLT"). Hopefully, proceeding on these terms will spare some confusion, since "value" is pretty much the most contested word in economics since "rational."

      Anyway, the latter is the definition Marx uses. And SNLT = price is indeed a simplifying assumption made in Capital volume 1, though by the end he identifies three distinct but closely related categories: SNLT, "price of production," and market price. They relate roughly like this: a) price equals price of production where the commodity realizes the general rate of profit, and b) price of production equals SNLT where the ratio of means of production to labor equals the average such ratio (for a given rate of surplus SNLT, but I won't get into that just yet).

      Thus, the case in which price = price of production = SNLT is exceptional (though we can postulate something similar to Sraffa's "standard commodity" to illustrate it), since all of the above needs to hold. In reality, the three usually diverge from one another for a given commodity, and such divergences effectively represent flows of labor time between different capitals. Thus, prices above (or below) SNLT facilitate faster growth (or outright shrinking) of capitals, shifting the distribution of social labor from, e.g., one sector to another.

      Nuno: Thanks for sharing that book, it looks neat!

    6. Hedlund, I just read the article by Rob Bryer you mention above and I liked it a lot. Could you send me any further information you have on the topic of Marx's exchanges with Engels on accounting to my e-mail, (unless other people are interested too and you prefer to discuss it here)

      Sorry, I missed this. I don't have any further references off hand, but I've exchanged emails once with Prof. Bryer via the address given on the paper, and found him very approachable and helpful. I'm sure he'd be willing to chat at length on the subject.

  6. The LTV certainly helps this Marxist critique of Piketty's new book: