Saturday, May 25, 2013

Technology, distribution and the rate of profit in the US economy: understanding the current crisis

New paper by Deepankar Basu and Ramaa Vasudevan - published by CJE (see here - subscription required)

From the Abstract:
"This paper offers a synoptic account of the state of the debate among Marxist scholars regarding the current structural crisis of capitalism, identifies two broad streams within the literature dealing, in turn, with aggregate demand and profitability problems, and proceeds to concentrate on an analysis of issues surrounding the profitability problem in two steps. First, evidence on profitability trends for the non-farm non-financial corporate business, the non-financial corporate business and the corporate business sectors in post-war USA are summarised. A broad range of profit rate measures are covered and data from both the US Bureau of Economic Analysis (NIPA and Fixed Assets Tables) and the Federal Reserve (Flow of Funds Account) are used. Second, the underlying drivers of profitability, in terms of technology and distribution, are investigated. The profitability analysis is used to offer some hypotheses about the current structural crisis."


  1. It's a good effort, but they kind of dismiss the TSSI out of hand due to what appears to be a misunderstanding. For example, in footnote 7, they say: "we believe Marx was quite unambiguously in favour of using replacement cost valuation of the capital stock," as if to suggest that this is in contradistinction to the TSSI position. Because they don't recognize how historical valuation is employed, it must seem like utter nonsense to them, which I imagine is why they've dismissed the TSSI results of secularly declining profit rates as some kind of edge-case outlier — concluding that "the weight of evidence thus suggests clearly that the current crisis was not preceded by a prolonged period of declining profitability."

    In fact, the passage in note 7 does not pertain to a measurement of profit; it pertains to the transfer of value from inputs to commodities. The TSSI agrees that current cost accounting is absolutely essential there — more specifically, "pre-production reproduction costs," which is subtly distinguished from replacement costs (but that's a matter for a different discussion). The only place the TSSI calls for historical valuation of capital is when measuring profit rates, and the reason for this is not because the value of the capital stock doesn't change, but because the value of the total capital advanced doesn't change.

    It may seem like an odd thing to say, but bear with me!

    If you take a sum of money, and invest it in machinery, whether that machinery abruptly rises or falls in value, your ultimate measurement of profit is still in terms of the value of the money revenues versus the value of the original sum of money advanced, which cannot be revalued. Consider the circuit of capital: M-C-M'. Profit would be M'-M. And since causality only runs one way, the C is a determinant of the M' and not the M.

    To give a numerical example:
    M = money capital (USD, GBP, whatever); C = constant capital (machinery, raw materials) value; V = variable capital (labor) value. Rate of surplus value will be set to 100%.

    Say 1000M buys 800C + 200V. That M figure is now locked in as your invested value. Now, say the value of the constant capital suddenly halves, giving you a total productive capital of 400C + 200V. Once all's been used up, you now have a total output worth 400 + 200 + 200 (surplus) = 800. You sell it and realize a value of 800 as your M'.

    Comparing the value you started with and the value you ended with, you'd find that 800-1000 = -200; that is, that hit to your capital value wound up hurting you for this production period, and you took a loss. The rate of profit was -20%. If you were, on the other hand, measuring your profit in terms of current capital replacement costs, you'd note that it only costs $600 to replace what you put in, and conclude that your rate of profit for the period is 33%.

    Thing is, that second thing is not exactly measuring "profit" anymore, as a return on investment (M'-M). Indeed, if you produce for another cycle, you may recoup some of your losses (assuming capital value doesn't fall again mid-process), but there can be no question that a miser who just sat on the 1000M is presently better off than the capitalist who invested it and got back 800M. Show me a capitalist who pays $1000, gets $800 back, and concludes that he profited by 33%, and I'll show you someone who will not be a capitalist for much longer. Especially if he financed that $1000 through a bank that then expected a return of, say, $1050. Such an entity would no doubt take a very different view of his windfall.

    That is the sole use the TSSI has for historical measurements. It is unfortunate the authors were so quick to dismiss this, as it seems entirely in line with Marx's value accounting, and its results are quite compelling.

    1. @Hedlund,

      Before asking my question, I would like to make it clear that I have no definitive position in relation to the TSSI. In fact, I know next to nothing about it and, in general, my first-hand knowledge of these modern mathematical formulations of Marx is very limited.

      I also say this because the question I am going to ask may be really stupid (so, please, bear with me).

      What if, instead of having its constant capital losing value (as in your example), constant capital increased in value halfway during the production process. Say, it doubled in value: from 800 to 1600 (after you paid 800).

      The total output is now 1600 + 200 + 200 = 2000. If you can sell your stuff for that and you use historical cost, you evaluate your profit like so: 2000 - 1000 (your actual cost) = 1000. 100%! Not bad, eh?

      But, at least in this scenario, the "surplus" you get using historical cost, it seems to me, is misleading. After all, come next production period and you'll still need 1600 (the replacement cost!) + 200 to produce.

      This means that in reality, you only have 200 (the S!) left for you: 20%.

      Am I making sense, or just making a fool of myself? :-)

    2. The problem with TSSI is that in order to defend a particular view of what they think Marx said, they end up abandoning the surplus approach basis of Marx analysis. They tend to confuse Ricardo and Sraffa with marginalist economics, and in the process end up presenting something that resembles marginalism (because they don't understand the meaning of neoclassical economics), surprisingly with the modern version of intertemporal general equilibrium models a la Arrow and Debreu. The ultimate analysis of TSSI is given by Gary Mongiovi in his paper Vulgar Economics in Marxian Garb (here

    3. Magpie:

      You're not wrong at all! The point is that capital gains and losses of this nature have both immediate and longer-term effects. In the former case, shocks can temporarily raise or lower the returns in a single period. But, as in the latter, it effectively acts as an adjustment of the organic composition of capital. Assuming there are no substitutions to be made in your example, then yes, the long-term implication of that rise in capital value is a fall in the rate of profit. That is, your rate of profit will actually be LESS than 20% moving forward, since s/C does not evaluate to 200/1000, but rather 200/1800 — or about ~11%.

      Ultimately, though, the problem with replacement cost methods is that they're measuring profits with respect to the capital stock rather than profits with respect to the money advanced. A capitalist starts with money capital and ends with money capital, and the rate of profit concerns the rate of change between these quantities. I don't get why people think it's okay to throw away this consideration, seeing as this was never unclear in Marx nor in capitalist accounts.


      I've read Mongiovi, I've read Laibman, I've read Veneziani. I've also read the responses to them. I don't think there's anything to be gained from retreading this ground, since it often seems like folks are content to talk past one another.

      If it will help, let's pretend I don't fancy the TSSI. Let's say I fancy the Hedlundian interpretation: the "HI." I have explicated details of it in my first post, which made no claim whatsoever about neoclassical economics, Ricardo, Sraffa, et al. With that said: Do you object to any part of what I've said above? If so, why?

    4. Matías

      Like I said, I don't know much about the TSSI or, to be honest, any of these modern, fancy mathematical Marxist theories.

      It would be irresponsible from me to pretend otherwise and side with any party in this dispute.

      I'll check the article you linked to as a priority (hopefully, more on the discursive, as opposed to the highly mathematical, side :-)).

      If Hedlund is reading this, I'll invite him to provide a link to a hopefully similarly written defense of his views.

      In the meantime, it may sound corny or silly, but can't we all get along? I mean, aren't we all supposed to care about the same basic ideas?

    5. Hi Magpie. I have no problem with getting along and having a conversation with anyone including those that disagree with me. Not sure why you'd have a different impression. However, on the intellectual content of TSSI I must be also honest, as I'm with respect to marginalism (neoclassical economics), that it is logically problematic and ultimately incorrect. I would suggest that you also read Fabio Petri's paper on the recent developments of value theory here

  2. I am not in favor of the TSSI approach. From my assessment of the analyses within the paradigm, it seems as if data has been constantly manipulated (specifically by Kliman & Freeman) to fit a reductionist falling rate of profit hypothesis. Moreover, I think it is a bit unjust to accuse the authors of not knowing how the historical valuation of capital stock is employed - see, and (subscription required) Moreover, contrary to Kliman's fanaticism, I believe that not holding onto a dogmatic conception of the supposedly fundamental Marxian theorem is requisite for analytical implication - see (subscription required); this is not to suggest, however, that any 'proof' fulfills the objective of elucidation- there is one, see here:

    1. Mr. Fields:
      I wrote a response to this yesterday, but I don't see it anywhere. Did it not go through? Or might it still be pending approval? (Or could it have wound up in a "spam" folder or something? I don't know much about how Blogger works.)

    2. Dave Fields: "data has been constantly manipulated (specifically by Kliman & Freeman) to fit a reductionist falling rate of profit hypothesis".

      How, exactly? If you are going to accuse people of making up the facts to suit themselves you have to supply chapter and verse.

  3. Hi, Mr. Fields. Thanks for your reply. Before I respond, is there anything in particular I said above that you disagree with? If so, what and why?

    Fair warning: I can't see anything that requires a subscription. (Despite the warning in the post, though, the Basu & Vasudevan piece was actually free to download.)

    "From my assessment of the analyses within the paradigm, it seems as if data has been constantly manipulated (specifically by Kliman & Freeman) to fit a reductionist falling rate of profit hypothesis."

    Can you elaborate on this? How has the data been manipulated? What makes the hypothesis "reductionist"?

    "Moreover, I think it is a bit unjust to accuse the authors of not knowing how the historical valuation of capital stock is employed - see"

    Looking over this paper, the idea that the price index for capital goods is factored into historical cost measures is pretty much exactly what I had intimated above. The result is that if we're measuring profit in money terms — which is how capitalists measure profit — then capital depreciation will play a role, not by changing the amount of money transformed into capital, but by changing the value of the total sales receipts.

    I take issue with this second paper's lack of clarity on the definition of the rate of profit. In the introduction, Basu says: "The profit rate is measured as the ratio of profit income (over a period) and the capital advanced (to generate that profit income)." This definition is fine, but then in the conclusion he changes it to "the ratio of profit income (suitably defined) and the value of the capital stock." This is actually a different definition, since "capital advanced" refers to money capital, while the capital stock refers to what the money was spent on. Keeping the measurement profit in terms of money is the whole point of using historical costs. However, the word "money" makes no appearance anywhere in Basu's article. (Incidentally, the site hosting this paper is that of Michael Roberts, who also promotes a historical valuation rate of profit, though his book was unfortunately not included in Basu & Vasudevan's article.)

    You are free to consider my remarks unjust for suggesting that they misunderstand the TSSI, but in fairness, I have not been given cause to believe otherwise. I think "misunderstanding" is certainly more charitable than certain other accusations floating around in this discussion.

    "Moreover, contrary to Kliman's fanaticism, I believe that not holding onto a dogmatic conception of the supposedly fundamental Marxian theorem is requisite for analytical implication"

    I'm afraid I simply don't follow what you're saying, here, apart from "fanaticism" and "dogmatic" (which, unfortunately, are not the only points that correspond to my last remark). How does the FMT relate to what I've said above?

    Thanks again!

  4. By the way, are you Jeffrey D. Hedlund who wrote this paper:

    Alas, I am not. But that is another good Hedlund!

  5. you dont like tssi?

    Do you think value expands in the production process?

    1. TSSI is a terribly confusing explanation of short term prices, which has little if anything of relevance to say about long term prices (value) or accumulation. Remember than Marx and the classical authors (surplus approach) wanted to determine long term prices in order to establish the rate of profit and, hence, explain accumulation. By the way, in order to have a more relevant discussion of accumulation one needs to abandon Say's Law (as Marx did), but also include in the analytical framework effective demand (Keynes and Kalecki).

    2. So do you believe that value expands in the production process?

      Like when wood is transformed into a chair. The woods value expanded when it came into contact with labor.

    3. Output expands in the accumulation process and that goes hand in hand with a process of gravitation of short term market prices towards long term (normal) prices, which require a uniform rate of profit. The former process can be described by effective demand in the long run (super-multiplier), while the latter are explained by Sraffa's equations.

    4. No not output...not accumulation. I am talking specifically about the value of the raw materials...

      If I have wood as a raw material at a given value, and I put labor to work making a chair. Is the value of the wood now higher?

  6. No not output...not accumulation. I am talking specifically about the value of the raw materials...

    If I have wood as a raw material at a given value, and I put labor to work making a chair. Is the value of the wood now higher?


Inequality and Stagnation by Policy Design

By Thomas Palley (guest blogger) This paper argues the mainstream economics profession is threatened by theories of the financial crisis a...