by Andrew Kliman
Here is the thirteenth installment of “All Value-Form, No Value-Substance,” the series of comments I’m writing on Fred Moseley’s new book, Money and Totality: A Macro-Monetary Interpretation of Marx’s Logic in Capital and the End of the “Transformation Problem.” It responds to part of Moseley’s reply to the twelfth installment.
Please see the “Miscellaneous” section on the homepage of With Sober Senses for links to previous installments.
The Incoherence of “Transitional Society” as a Marxian Concept by Andrew Kliman Kliman gave this talk to the Workers’ and Punks’ University May Day School in Llubljana, Slovenia, on April 27, 2013. The text […]
The Unmaking of Marx’s Capital Heinrich’s Attempt to Eliminate Marx’s Crisis Theory by Andrew Kliman, Alan Freeman, Nick Potts, Alexey Gusev, and Brendan Cooney Michael Heinrich’s recent Monthly Review article claims that the law of […]
A major dispute over the causes of the latest capitalist economic crisis, and how to respond to it, has erupted in the Committee for a Workers’ International (CWI), an orthodox Trotskyist organization. Its largest section, […]
Copyright © 2011 - 2018 Marxist-Humanist Initiative
This comment will briefly present my interpretation of Marx’s prices of production as long-run equilibrium prices. And a few questions of clarification for Kliman at the end. Later comments will respond to Kliman’s Part 13.
Marx’s prices of production are long-run equilibrium prices
I argue that Marx’s prices of production in Vol. 3 are (or were intended to be) long-run equilibrium prices with the following properties:
equal rates of profit across industries
change only if the productivity of labor (or the wage) changes
function as long-run center of gravity of market prices
Equilibrium does not mean no change, but means changes only for these specific fundamental causes. So if the productivity of labor and the wage remain constant, then Marx’s prices of production do not change.
Market prices also depend on the accidental and temporary causes of S and D and thus often change every period. Marx’s theory abstracts from the fluctuations of market prices and focuses instead in prices of production as the center of gravity of these fluctuations which is determined by the “immanent laws” of capital.
I think the textual evidence to support this interpretation of Marx’s prices of production as long-run equilibrium prices is very strong, which I have documented in a 1999 paper on academia:
https://www.academia.edu/27678884/Marxs_Concept_of_Prices_of_Production_Long-Run_Center_of_Gravity_Prices
and summarized in my book (pp. 289-96 and 333-37).
A few examples:
“The real inner laws of capitalist production clearly cannot be explained in terms of the interaction of demand and supply …, since these laws are realized in their PURE FORM only when demand and supply cease to operate, i.e. when they coincide. In actual fact, demand and supply never coincide, or, if they do so, it is only by chance and not to be taken into account for scientific purposes; it should be considered as not having happened. Why then does political economy assume that they do coincide? In order to treat the phenomena it deals with in their LAW-LIKE FORM, the form that corresponds to their concept, i.e. to consider them independently of the appearance produced by the movement of demand and supply.” (C.III, p. 291)
“If supply and demand coincide, the market price of the commodity corresponds to its price of production, i.e. its price is then governed by the INNER LAWS of capitalist production, independent of competition, since fluctuations in supply and demand explain nothing by divergences between market prices and prices of production – divergences which are mutually compensatory, so that over certain longer periods the average market prices are equal to the prices of production. As soon as they coincide, these forces cease to have any effect, they cancel each other out, and the general law of price determination corresponds to price of production in its immediate existence and not only as an average of all price movements, and the price of production, for its part, is governed by the IMMANENT LAWS of the mode of production.” (C.III, pp. 477-78)
[T]he actual movement of competition lies outside of our plan, and we are only out to present the internal organization of the capitalist mode of production, its IDEAL AVERAGE, as it were.” (C.III, pp. 969-70)
Marx also said many times that his prices of production are similar to Smith’s and Ricardo’s “natural prices”, which were long-run equilibrium prices or “center of gravity” prices, around which market prices fluctuate over multiple periods of time. Marx’s critique of Smith and Ricardo was NOT that the prices in their theories should not be long-run equilibrium prices, but that they were UNABLE TO EXPLAIN NATURAL PRICES with equal rates of profit and could not explain why and how natural prices differ from values. For example:
“The price of production includes the average profit. And what we call price of production is in fact the SAME THING that Adam Smith calls “natural price, Ricardo “price of production” of “cost of production”, and the Physiocrats “prix necessaire”, though none of these people explained the difference between price of production and value.” (C.III. p. 300)
I hope that readers will read the paper cited above and see what you think about the textual evidence to support the interpretation of Marx’s prices of production as long-run equilibrium prices with the three properties listed above.
The main purpose of Marx’s theory of prices of production was to *answer the main criticism of Ricardo’s labor theory of value* (especially by Malthus and also Torrens) – that the labor theory of value was contradicted by equal rates of profit and was unable to explain long-run equilibrium prices with equal rates of profit. Marx answered this main criticism of the labor theory of value on its own terms. Marx did not argue that long-run equilibrium prices were not important, but he showed how long-run equilibrium prices COULD BE EXPLAINED on the basis of the labor theory of value.
The long debate over the transformation has generally assumed (correctly in my view) that Marx’s prices of production are long-run equilibrium prices. And the modern critics of Marx’s theory have argued that Marx was NOT ABLE to successfully explain prices of production as long-run equilibrium prices on the basis of the LTV, similar to Malthus’ critique of Ricardo.
My book is a response to this modern critique of Marx’s theory of prices of production. I argue that, if Marx’s logical method is correctly understood – the determination of the total surplus-value prior to its division into individual parts and the circuit of money capital (M-C …) as the logical framework of Marx’s theory which implies that the initial M is the starting point of the theory and is taken as given – then Marx did SUCCESSFULLY EXPLAIN LONG-RUN EQUILIBRIUM PRICES of the basis of the labor theory of value, and thus there is no transformation problem in Marx’s theory. I rebut the long-standing critique on its own terms, as Marx rebutted Malthus’ criticism on the same terms.
I think this is a significant victory for your side.
QUESTIONS OF CLARIFICATION FOR KLIMAN:
1. What is the equation for the “static equilibrium” prices of production in Section III of Part 13?
2. What is the equation for the “actual” prices of production?
3. Would you please make available your Excel worksheet so we can have some Phun?
Thanks
Hi Fred,
I’m already having phun.
By popular demand, the spreadsheet is now here:
http://marxisthumanistinitiative.org/wp-content/uploads/2017/02/Centers-of-Nothing.xlsx
See the Equations! sheet for answers to your 1st 2 questions.
(The Equations! sheet corrects a couple of errors in the text of that section of Part 13; the computations were right and are not affected by the corrections.)
Hi Fred,
You wrote,
Don’t forget that, in Part 2 of “All Value-Form, No Value-Substance,” I showed that your prices of production–both sectoral aggregates prices and per-unit prices–do not have the second of these properties. They change even if the “productivity of labor” (or the wage) do not change.
I have been busy with other things, but I would like to get back to the discussion of Part 13.
But the link given for the Excel spreadsheet doesn’t work. Please fix that. Or send to me directly. thanks.
Also, I also have two questions about the equations on p. 4 of Part 13:
1. In the last equation, total price for the period (t+1) = total value for period (t). What is the rationale for that unusual assumption. Usually total price is equated with total value in the same period.
2. How are r(j) and r(k) determined in the first period, since there is p(t+1) in the numerator?
Thanks.
Below is my previous comment on Kliman’s Part 2.
II. Prices of production as long-run center of gravity prices and causes of changes
This is a reply to Kliman’s second post on my recent book Money and Totality: A Macro-Monetary Interpretation of Marx’s Logic in Capital and the End of the ‘Transformation Problem’. This reply focuses on whether Marx’s concept of prices of production are long-run center of gravity prices around which market prices fluctuate over multiple periods of time and which change only if productivity or the real wage changes (my interpretation) or are short-run prices that continue to change over multiple periods even though productivity and the real wage remain constant and thus cannot function as long-run center of gravity prices (TSSI)
I argued in my book that Marx’s concept of price of production are long-run center-of-gravity prices, around which actual market prices fluctuate (“gravitate”) from period to period. These long run center of gravity prices are in the classical tradition of Smith and Ricardo, which have three key characteristics: (1) they equalize the rate of profit across industries; (2) they are “centers of gravity” around which actual market prices fluctuate over extended periods of time; and (3) they change if and only if either the productivity of labor changes (due to changes in the technology of production) or (secondarily) if the real wage changes.
I argued that the TSSI prices of production have the first characteristic, but do not have the other two key characteristics and thus is a misinterpretation of Marx’s concept of prices of production. According to the TSSI, the transformation of values into prices of production in as ongoing process that takes place over multiple periods, even though productivity and the real wage remains the same in all these periods. And since TSSI prices of production change every period, they cannot be “centers of gravity” around which market prices fluctuate over longer periods of time.
My book presented substantial textual evidence to support my interpretation of both of these other two characteristic of prices of production. For the characteristic of long-run center of gravity prices, Marx repeated a number of times in Theories of Surplus-Value and in Volume 3 of Capital and in letters to Engels that is prices of production were essentially the same as Smith’s and Ricardo’s “natural prices” which were long-run center of gravity prices around which market prices fluctuate (pp. 334-37).
For the characteristic of “change only if …”, Marx argued in a number of passages, especially in Part 2 of Volume 3, that since prices of production are determined by the equation:
PPi = (Ci + Vi) + R (Ci + Vi)
changes in prices of production could be due to a change in Ci or Vi or R, or some combination of these. Marx argued further in these passages (reviewed in my book, pp. 289-96) that changes in Ci or Vi are caused by changes in the productivity of labor, either in final goods industries, or in industries that produce the means of production for these final goods industries. A change of Vi could also be due to a change in the real wage. Marx also argued that a change in R is also caused either by a change in the productivity of labor somewhere in the economy which changes either the composition of capital or the rate of surplus-value. A change in the rate of surplus-value could also be due to a change in the real wage. These discussions of the causes of changes in prices of production seem to imply the conclusion that, if the productivity of labor and the real wage remain constant, then prices of production would also remain constant. Marx does not mention in these passages any other possible cause of changes in prices of production, besides changes in the productivity of labor and/or the real wage. He certainly does not ever mention that Ci and Vi and prices of production might continue to change in successive periods as a result of the ongoing equalization of profit rates and the transformation of values into prices of production, even though productivity and the real wage remain constant (as in the TSSI).
Kliman doesn’t say anything in his post about the second characteristic of prices of production as long-run center of gravity prices, and thus does not dispute my argument and textual evidence on this important point. Instead, he focuses on the third characteristic of “changes only if …”. He argues that, since I define prices of production as gross annual industry revenue (not unit prices), another possible cause of changes of prices of production defined in this way that was not mentioned in these passages by Marx is simply an increase in the scale of production, since that would increase gross annual industry revenue even though productivity and the real wage remain constant. And he infers from this very slim basis that yet another possible cause of changes in the prices of production not mentioned by Marx is the ongoing multi-period transformation of values into prices of production, even though productivity and the real wage remain constant (as in the TSSI).
Kliman is correct that my definition of prices of production as “gross annual industry revenue” implies that an increase in the scale of production would increase prices of production defined in this way, even though productivity and the real wage remain constant. However, I argue that this fact does not bolster Kliman’s case that yet another cause of changes of prices of production is the ongoing transformation of values into prices of production.
I continue to think that “gross annual industry revenue” is the correct definition of prices of production in a general sense, but I now realize more clearly that in Part 2 of Volume 3 Marx analyzed prices of production in a restricted sense, as prices of production per capital of 100. All the industries in Marx’s tables and illustrations in Part 2 have a total capital of 100, with unequal compositions of capital (ratios of constant capital to variable capital). Marx did this in order to emphasize the effect of unequal compositions of capital across industries on the value and surplus-value produced in each industry (Volume 3, pp. 261-62). Therefore, it seems reasonable to assume that in the passages in Part 2 that I reviewed in my book and that discuss the two causes of changes in prices of production, Marx had in mind this restricted sense of prices of production per capital of 100. This restricted definition of prices of production rules out an increase in the scale of production as a cause of changes in restricted prices of production. In this context, it made sense for Marx to state repeatedly that there are only two causes of changes in prices of production – changes in productivity and changes in the real wage – and not to mention an increase in the scale of production (which is not theoretically interesting or important anyway) as a cause of changes in these restricted prices of production.
Kliman said in concluding his post:
Hence, if the TSSI misinterprets Marx because it implies that prices of production can change even when technology and the real wage do not, then Moseley misinterprets Marx in the same way.
I don’t think I misinterpreted Marx’s prices of production fundamentally, but I agree that I did not fully appreciate the significance of Marx’s restricted sense of prices of production (per capital of 100) in Part 2 of Volume 3 and the connection between this restricted sense of prices of production and Marx’s discussions of the two causes of changes in prices of production in Part 2. And I will gladly acknowledge that in increase in the scale of production is another cause of a change in prices of production in the general sense of gross annual industry revenue.
However, this additional cause of changes in prices of production in the general sense does not contradict Marx’s discussions of only two possible causes in his restricted sense. And it provides no basis for inferring that another cause of changes in prices of production (general or restricted) is the ongoing transformation of values into prices of production, as in the TSSI. There is no hint whatsoever in all of Marx’s writings on the transformation and prices of production that the ongoing transformation is another possible cause of changes in prices of production. No textual evidence is presented in this post or in previous writings to support the TSS interpretation of prices of production as short-run prices that continue to change over multiple periods (even though productivity and the real wage remain constant) and thus cannot function as “centers of gravity” of market prices.
The most reasonable conclusion seems to be that the TSS interpretation of short-run prices of production is a misinterpretation of Marx’s long-run prices of production.
CONTINUATION Feb. 28
I should have included this clarification in my comment on Part 13. In any case, I think my conclusion still stands: there is no textual basis for inferring that another cause of changes in prices of production (general or restricted) is the ongoing equalization (over multiple periods) of the profit rate and transformation of values into prices of production, even if productivity and wages remain constant, as in the TSSI.
A reply to Fred Moseley’s comment of Tue, 28th Feb 2017 9:12 am.
The link is now working. The equations page of the spreadsheet explains the computations in detail, and it fixes minor problems (that don’t affect the conclusions) in equations in Part 13.
Output prices of period t carry the subscript t + 1. (Input prices carry the subscript t.) With that in mind, and noting that the corrected equations fix errors in the subscripts for outputs, it should be clear that both sides of the total price = total value equation pertain to period t.
Noting again that the corrected equations fix errors in the subscripts for outputs, period 0’s outputs are data. They and the top equation determine the output-price ratio in period 0. The output-price ratio, the other data, and the total price = total value equation then determine the absolute output prices of period 0. The absolute output prices, the other data, and the rate of profit equations then determine the sectoral rates of profit of period 0.
In Section III of his Part 13, Kliman attempts to demonstrate that market prices do not fluctuate around static equilibrium prices (that he mistakenly identifies with my interpretation of Marx’s prices of production), but instead fluctuate around TSSI prices of production.
But, there are two problems with Kliman’s argument. In the first place, Kliman’s static equilibrium (SE) rate of profit that he is uses to calculate his SE prices of production is *not the same as the rate of profit in my interpretation of Marx’s theory*. His SE rate of profit is determined by equating the SE rates of profit for the two sectors and the SE rates of profit for the two sectors do not depend on the labor theory of value in any way. Instead, labor is only a cost in these equations and is not a producer of value. Indeed his SE rates of profit vary inversely with the quantity of labor (Ɩ is in the denominator of the equations for the SE rate of profit).
In Kliman’s model, the real wage (b) increases at the same rate as the quantity of labor (Ɩ) decreases (4% each period), so that the product bƖ remains the same and the SE rate of profit remains the same. However, if b were to remain constant and Ɩ decrease 4% a year, then the SE rate of profit would steadily increase because labor is only a cost, contrary to Marx’s theory.
Therefore, whatever conclusions Kliman derives from his model about the SE rate of profit and SE prices of production do not apply to my interpretation of Marx’s theory.
And even more that that: the second problem with Kliman’s argument is that the TSSI rate of profit that he uses to calculate his TSSI prices of production also does not depend on the labor theory of value in any way; labor (again) is also only a cost in the TSSI equation for the rate of profit and is not a producer of value, contrary (again) to Marx’s theory. The TSSI rate of profit is the average of the two sector rates of profit, and this average rate of profit also varies inversely with quantity of labor (Ɩ is in the denominator of the equation for the average rate of profit).
So the TSSI rate of profit is similar to the static equilibrium rate of profit – both assume that the rate of profit varies inversely with the quantity of labor employed – and neither has anything to do with Marx’s labor theory of value and surplus-value and the rate of profit.
I leave tomorrow for a two week trip to Buenos Aires (conference) and Montevideo (talks) and will rejoin the discussion when I return.
Fred, this claim of yours is false:
If you look at the code for p2 in the static equilibrium system, you’ll see that it is
p2 =
m(l1x1 + l2x2)
—————————————————
(p1/p2)x1 + x2 – (p1/p2)(a1x1 + a2x2)
Cross-multiplying and rearranging, we get:
p1x1 + p2x2 = p1(a1x1 + a2x2) + m(l1x1 + l2x2)
which says that total price equals total value, and the final RHS term is precisely the new value added by living labor.
So the static equilibrium system does apply to your interpretation, and the above equation is your own total value = total price equation.
This is all explained at the bottom of the Equations! page of the spreadsheet.
The following claim of yours is also false:
The temporally determined prices employ the temporal analogue to the above equation; it’s the 4th equation in the system (see the Equations! page).
Compute the monetary variables in both systems and you’ll see that everything checks out exactly.
And the sectoral rates of profit fluctuate around the average temporal rate of profit, not around your average rate of profit; the market prices fluctuate around the temporal prices of production, not around your prices of production.
I decided to compute the monetary variables myself, so that we can move on. As I said, “everything checks out exactly.” The enhanced Centers of Nothing spreadsheet is here:
http://marxisthumanistinitiative.org/wp-content/uploads/2017/03/Centers-of-Nothing-enhanced-3.9.17.xlsx
Andrew,
I’m trying to understand “Centers-of-Nothing-enhanced-3.9.17.xlsx”. In the sheet you put temporalist calculation and p in=out calculation into opposition. In Kliman/McGlone article
period 14 (and also 15, 16 etc., I presume) p in=out seems to be compatible with TSSI. So, why is that or what are the conditions for p in=out being allowed and when not?
When reading this article again I stumbled across this sentence: “(Solely in order to facilitate comparison with `transformation problem”solutions’, we begin
without any ‘error[s] in the past ; i.e . initial values are equal to the values of means of production and labour-power .)” Does it mean, in period 1 you could have started with cost prices that already diverge from value of bought commodities?
A reply to Herbert Panzer’s comment of Sat, 11th Mar 2017 12:10 pm:
“So … what are the conditions for p in=out being allowed and when not?” It’s allowed when the actual data and temporal value relations generate that result (a zero-probability event). Otherwise, it’s not. In other words, input and output prices may happen to be equal, but we don’t force them to be equal contrary to the facts–or force them to be unequal contrary to the facts.
We like facts. Facts are our friends.
“Does it mean, in period 1 you could have started with cost prices that already diverge from value of bought commodities?” Yes. The cost prices are always determined by the actual prices (market prices, monopoly prices, regulated prices, etc., etc.) that currently need to be paid to acquire the inputs.
Thanks for helpful reply*. Other point. In Vol3 p.264 Marx writes “How these capitals function after the average rate of profit is established, on the assumption of one turnover in the year, …”. This does not sound like average profit rate is changing every period. So, do you not think, your Excel-Sheet
(20,3%,19,6%) might be a bit out of specification?
It is also hard to imagine that Marx, while assuming smoothed out average profit rate, does have in mind continually changing determinants of this rate, like labor-saving technological change or wage rate.
*small side information: in natural science, majority of our work in the lab is about “zero probability events” and facts are our friends, too.
A reply to Herbert Panzer’s comment of Sun, 12th Mar 2017 6:53 am.
In the spreadsheet example, the main source of the fluctuations in the general rate of profit is the fluctuations in sectoral output levels. They change in every period–which I think is quite reasonable–and therefore the general rate of profit changes in every period.
This is completely consistent with Marx’s concept of the general rate of profit. According to his theory, the rate of profit has to change when output levels change because the composition of the total social capital changes. And nothing in his text says that the general rate of profit is stationary. The passage you cite does not, and 51 pages after it, there’s a whole 60-page Part on the tendential fall in the general rate of profit!
For this and other reasons, it is not hard for me to imagine that Marx “does have in mind continually changing determinants of this rate [of profit], like labor-saving technological change or wage rate.” Not hard at all.
But that’s really not the issue. The issue is: given Marx’s concept of the general rate of profit, does that rate change in response to continually changing determinants like labor-saving technological change and the wage rate? The answer is “yes.”
Herbert, in re my comment that “In the spreadsheet example, the main source of the fluctuations in the general rate of profit is the fluctuations in sectoral output levels,” I just checked. After removing the transients at the start, we have x1/x2 = 1.014, avg. r = 20.3% in every odd-numbered period, and x1/x2 = 0.986, avg. r = 19.6% in every even-numbered period. For periods 20-199, the regression line is avg. r = -0.04974 + 0.24965(x1/x2) and the R² = 0.99999997.
Andrew, I think it’s not so simple. Consider “For all the great changes that constantly occur in the actual rates of profit in particular spheres of production (as we shall later show), a genuine change in the general rate of profit, one not simply brought about by exceptional economic events, is the final
outcome of a whole series of protracted oscillations, which required a good deal of time before they are consolidated and balanced out to produce a change in the general rate. In all periods shorter than this, therefore, and even then leaving aside fluctuations in market prices, a change in prices of production is always to be explained prima facie by an actual change in commodity values, i.e. by a change in the total sum of labour-time needed to produce the commodities.“ [Vol3 p.266]
As Marx is assuming one turnover in the year (synchronized/clocked I’d like to add), this is our time unit.
And change in general rate of profit requires a good deal of these units. As oscillations sometimes compensate each other, sometimes add up, the movement of gen. profit rate may change
for, let’s say, 13 periods, then stay stationary (approximately, of course, i.e. zero-probability as you say) for 6 periods, then change again for 7 periods, then stay stationary again (but on a lower level, as a consequence of tendential fall) etc. .
Related one thing you are right: change of determinants can be on a higher frequency then gen. profit rate change. But even such shorter periods are systematically not just one time unit.
So, overall, this dynamic pattern does not look like your Excel sheet.
Herbert Panzer,
You’re assuming that “protracted oscillations” refers to oscillations of the actual rates of profit, not oscillations of the general rate of profit. Why assume that?
The text of the passage is fully compatible with my interpretation, in which “protracted oscillations” refers to oscillations of the general rate of profit, and “genuine change in the general rate of profit” refers to a rise or fall in the level of the general rate that isn’t counterbalanced by a subsequent oscillatory fall or rise.
This interpretation also makes Marx’s text make more sense than your interpretation does, since, as I’ve noted, “According to his theory, the rate of profit has to change when output levels change because the composition of the total social capital changes.”
“But even such shorter periods are systematically not just one time unit.
“So, overall, this dynamic pattern does not look like your Excel sheet.”
Right. If I wanted to produce a realistic model, this would be a problem. But I don’t, so it isn’t.
The simulation is not a model but an illustration. Its purpose is to illustrate the fact that, when prices trend downward (or upward), a static-equilibrium rate of profit, such as Moseley’s, will not be the center around which actual rates of profit fluctuate. Instead, they will fluctuate around their weighted average.
Why assume that? Because, if a change of something-A consolidates and balances into a change of something-B then *-A and *-B are not the same
I do not know many Marxist and/or economists that are able to differentiate between model and illustration.
Typically, in the lab you use a model and for presenting conclusions found there you use an illustration. Mixing this is a frequent reason for erroneous conclusions. So, I allege you can maintain your conclusions also related a model, in case being asked.
Moseley: I’m not that far. But as a potential outlook: have you taken into account the possibility that Moseley is considering not only static-equilibrium rate of profit, but static-equilibrium system state with no change
of relevant system parameters (like prices) at this state while changes happen in a non-equilibrium state and both states are in sequence and alternating?
POSTED AT THE REQUEST OF ANDREW KLIMAN:
Herbert Panzer,
The following still doesn’t explain why you assume that Marx’s reference to “protracted oscillations” refers to oscillations of the actual rates of profit, not oscillations of the general rate of profit:
Marx wrote,
My point is that the text is entirely consistent with my interpretation of it, according to which the protracted oscillations in the general rate of profit are consolidated and balanced out to produce a (genuine) change in the general rate of profit (a change not counterbalanced by subsequent offsetting fluctuations). Here, *-A is the protracted oscillations in the general rate of profit, and *-B is the (genuine) change in the general rate of profit—and they are not the same.
So, the question remains: why do you assume that “protracted oscillations” refers to oscillations of the actual rates of profit, not oscillations of the general rate of profit?
I have no idea of what this means:
In any case, the illustration in question isn’t a model, because it is not intended to depict the actual dynamics of an actual economy. It is simply intended to illustrate the fact that, when prices trend downward (or upward), a static-equilibrium rate of profit, such as Moseley’s, will not be the center around which actual rates of profit fluctuate. Instead, they will fluctuate around their weighted average. The illustration illustrates this, and I think it does so very successfully, even though it’s not a realistic depiction of the actual dynamics of an actual economy.
It is pointless to criticize an illustration for not doing what it isn’t intended to do and which no one claims that it does. It’s like criticizing a butter knife for being a bad economic model, even though it’s purpose is instead to spread butter on bread and the maker of the butter knife doesn’t claim that it’s an economic model.
I don’t understand what “in sequence and alternating” means. In sequence with what? Alternating with what? In any case, the magnitude of the static-equilibrium rate of profit is invariant to the absolute magnitudes of the prices. Only a change in the relative price (p1/p2) will affect the static-equilibrium rate of profit (see Equations page of the spreadsheet). But there isn’t any change in the relative price; it always equals 1 in this example. So I could easily hold the absolute magnitudes of the prices constant (by making one of the prices the numeraire, for instance) and get the exact same results for the static-equilibrium rate of profit. It would still be the center of nothing.
But Moseley is definitely not considering a stationary-price system (only), since he contends, incorrectly, that his rate of profit is the center around which actual rates of profit fluctuate—not only in a stationary state, but also when there is labor-saving technical change. And when there’s labor-saving technical change, his prices aren’t stationary.
Reply to Kliman (March 9)
Kliman quoted me:
“labor is only a cost in these [static equilibrium] equations and is not a producer of value.”
And he argued that this claim is false.
But that was not my specific claim. My specific claim was that labor is only a cost in the calculation of the “static equilibrium” (SE) RATE OF PROFIT and that the SE RATE OF PROFIT varies inversely with the quantity of labor employed (L) and has nothing to do with the labor theory of value. I said:
“Kliman’s static equilibrium SE RATE OF PROFIT that he uses to calculate his SE prices of production is *not the same as the RATE OF PROFIT in my interpretation of Marx’s theory*. His SE RATE OF PROFIT is determined by equating the SE RATES OF PROFIT for the two sectors and the SE RATES OF PROFIT for the two sectors do not depend on the labor theory of value in any way. Instead, labor is only a cost in these equations and is not a producer of value. Indeed his SE RATES OF PROFIT vary inversely with the quantity of labor (Ɩ is in the denominator of the equations for the SE RATE OF PROFIT).” (emphasis added)
Kliman did not contest this specific claim about the irrelevance of the labor theory of value for the determination of the SE rate of profit. Indeed, Kliman confirmed my claim when he says (on the equations page of his spreadsheet) that the absolute levels of SE prices (that do include a term for the new-value determined by current labor) “are irrelevant to the determination of the static equilibrium rate of profit”.
The aggregate equality total price = total value (that includes a term for the new-value determined by current labor) is only a normalization condition that affects only the levels of absolute prices; this aggregate equality does not affect the rate of profit. According to Kliman’s equations, any arbitrary normalization condition could be assumed and the rate of profit would be the same and the rate of profit would vary inversely with L. If the decrease of L assumed by Kliman were not offset by the assumed increase of b, then Kliman’s SE rate of profit would increase, contrary to my interpretation of Marx’s theory.
Further comment on Section III of Kliman’s Part 13
Kliman’s example assumes that the productivity of labor increases CONTINUALLY in both sectors IN EVERY PERIOD. Generalized, this would mean that productivity increases in all industries IN EVERY PERIOD. This is a very unrealistic assumption, as Herbert has pointed out.
If we consider a single industry, technological change that increases productivity usually does not occur every period, but instead occurs infrequently and less frequently than changes in market prices which also depend on supply and demand. Market prices fluctuate around relatively stable prices of production, not around prices of production that change every period. In Kliman and McGlone’s original 1988 paper on the transformation problem, the transformation from values to prices of production takes place over 14 periods, which are assumed to correspond to real historical periods, and technology is assumed to remain constant in all 14 periods!
When technological change does occur in an industry, the long-run center of gravity price of production of the good produced in that industry (calculated according to current cost of the inputs) also changes. But that does not mean that the actual price of the output in that period is equal to the new price of production. Rather it means that in future periods market prices will tend to gravitate toward the new price of production.
When technological change occurs in an industry (and everything else remains constant), then according to the TSSI, the price of that good as an input at time t is greater than the price of production of that good as an output at time t+1 (i.e. p(t) > p(t+1)), and as a result of that difference the historical cost rate of profit is less than the current cost rate of profit in that industry.
Then this good produced in period t becomes an input in period t+1 and is purchased at p(t+1). If there is *no technological change in period t+1*, then the price of this good will remain the same and the price of this good as an output of period t+1 at time t+2 will also be p(t+1). Thus in period t+1 the historical cost of this good as an input is equal to its current cost and the price of this good as an input is equal to its price as an output. There is only a temporary divergence between historical cost and current cost of this good and it takes only one period for the historical cost to “catch up” to the new current cost. And since the difference between historical costs and current costs has been eliminated in the second period, so has the difference between the historical cost rate of profit and the current cost rate of profit.
Also, the TSSI price of production calculated according to historical costs in period t+1 is different from the TSSI price of production in period t, and thus the TSSI price of production CHANGES, *even though there is no technological change* in period t+1. The historical cost price of production is only temporary and thus cannot be a long-run center of gravity for market prices. The long-run center of gravity price for market prices is the current cost price of production to which the historical cost price of production converges and which changes ONLY if there is technological change and thus is more stable and more appropriate for the long-run center of gravity price.
And even if there were two or three or more periods of continual technological change in an industry, as soon as technological change did not occur in a period, the input price of this good would be equal to its output price and the historical cost price of production of this good would converge to its current cost price of production. The historical cost price of production is always “catching up” to the current cost price of production and it catches up as soon as there is no technological change in a given industry.
Marx assumed that technological change in a given industry would occur infrequently and less frequently than changes in market prices. As Herbert put it, it is hard to imagine that Marx thought that technological change would occur in every period in every industry.
For example, in Marx’s many statements that “prices of production change only if productivity (or wages) changes” the implication seems to be that such changes in productivity (due to technological change) would occur infrequently.
In addition, there are a number of passages in which Marx stated explicitly or implied that prices of production were usually stable over “long” periods of time. In Theories of Surplus-Value, Marx equated his concept of “cost price” (an early term for price of production) with Smith’s and Ricardo’s concept of “natural price” and he stated:
“The natural price of the commodity is not the market-price but the AVERAGE MARKET-PRICE over a LONG PERIOD, or the central point towards which the market-price gravitates.” (TSV.II. 319; emphasis added)
Thus, Marx’s price of production is “the average market price over a LONG PERIOD”.
In Volume 3 (Chapter 21), Marx made a similar statement, this time explicitly in terms of prices of production.
“If supply and demand coincide, the market price of the commodity corresponds to its price of production, i.e. its price is then governed by the inner laws of capitalist production, independent of competition, since fluctuations in supply and demand explain nothing but divergences between market prices and prices of production – divergences which are mutually compensatory, so that OVER CERTAIN LONGER PERIODS THE AVERAGE MARKET PRICES ARE EQUAL TO THE PRICES OF PRODUCTION.” (pp. 477-78; emphasis added)
And in Chapter 50 of Volume 3, Marx suggested that in prices of production are stable “over a prolonged period”:
“Market prices rise above these GOVERNING PRODUCTION PRICES or fall below them, but these fluctuations balance each other out. If one compares price lists OVER A PROLONGED PERIOD, and ignores those cases in which the actual value of a commodity alters as a result of a change in labor productivity, as well as cases in which the production process is disturbed by natural or social disasters, it is surprising both how narrow the limits of these divergences are and how regularly they are balanced out.” (pp. 999-1000; emphasis added)
There is also an important footnote at the end of Chapter 5 of Volume 1 which states that the average price of oscillating market prices is the “guiding light” of merchants and manufacturers in their long-term investment decisions:
“The reader will see from the foregoing discussion that the meaning of this statement is only as follows: the formation of capital must be possible even though the price and the value of a commodity be the same, for it cannot be explained by referring to any divergence between price and value. The continual oscillations in prices, their rise and fall, compensate each other, cancel each other out, and carry out their own reduction to an AVERAGE PRICE which is their internal regulator. This AVERAGE PRICE IS THE GUIDING LIGHT of the merchant or the manufacturer in every undertaking of a lengthy nature.” (C.I. 269; emphasis added)
The average price cannot be the “guiding light” of manufacturers if it changes in every period, even if there is no technological change. Later in this footnote, Marx clarified that these long-run average prices “do not directly coincide with the value of commodities, as Smith, Ricardo, and others believe.” And Marx later explained in Volume 3 that these long-run average prices are equal to prices of production with equal rates of profit and current costs.
Finally, I recently came across the following passage from early in the Grundrisse that refers to the average price of coffee of an “epoch” and “over 25 years”, and Marx commented that this long-term average price is the “driving force and moving principle” of the oscillations of market prices (i.e. the “guiding light” of manufacturers):
“The value of commodities as determined by labour time is only their average value. This average appears as an external abstraction if it is calculated out as the average figure of an EPOCH, e.g. 1 lb. of coffee = 1s. if the average price of coffee is taken OVER 25 YEARS; but it is very real if it is at the same time recognized as the DRIVING FORCE AND THE MOVING PRINCIPLE of the oscillations which commodity prices run through during a given EPOCH.” (G. 137; emphasis added)
A reply to Fred Moseley’s comments of Mon, 3rd Apr 2017 7:26 am and 7:41 am
He wrote:
I do contest it—because it is false. It contains a glaring confusion between and conflation of calculation and determination. This new version of my spreadsheet– http://marxisthumanistinitiative.org/wp-content/uploads/2017/04/Centers-of-Nothing-periodic-tech-change.xlsx — calculates the static-equilibrium rate of profit in two different ways. In one calculation, the static-equilibrium rates of profit are equated, etc. But in the other calculation (columns AC-AH of the Data sheet), the static-equilibrium rate of profit is calculated in the exact way that Moseley likes! Profit is the difference between the new value added by living labor and variable capital, and the rate of profit equals profit divided by the sum of constant and variable capital.
And the two sets of calculations produce the exact same result, always, in every period. So “Kliman’s [sic] static equilibrium SE RATE OF PROFIT that he uses to calculate his [sic] SE prices of production is *[EXACTLY] the same as the RATE OF PROFIT in [Moseley’s] interpretation of Marx’s theory*.”
Thus, if Moseley wants to complain that the static-equilibrium rate of profit “do[es] not depend on the labor theory of value in any way”—as he should—he needs to join with me in complaining about his bogus interpretation of Marx. That’s because the static-equilibrium rate of profit that does not depend on the labor theory of value in any way is Moseley’s own static-equilibrium rate of profit!
He also wrote,
False, false, false! The “constant-b example” page of my new spreadsheet assumes that living labor decreases, but it assumes that the real wage rate is constant. Once again, the two different ways of calculating the static-equilibrium rate of profit produce the exact same result, always, in every period. The static-equilibrium rate of profit that does not depend on the labor theory of value in any way is, once again, Moseley’s own static-equilibrium rate of profit!
He also wrote,
This line of defense gets him absolutely nowhere. It makes no difference whether the technical changes occur in every period or less frequently. The new version of my spreadsheet assumes that technical change takes place only in every third period (starting with period 3 in Sector 2 and period 4 in Sector 1). (The real wage rate is also modified, such that, in each sector, the real wage per unit of output remains constant.)
Nevertheless, Moseley’s static-equilibrium “prices of production” and rate of profit are, once again, centers of nothing. Actual prices and rate of profit don’t fluctuate around them. Once the technical changes commence, his static-equilibrium rate of profit is consistently a good deal higher than both the sectoral rates of profit and the actual average rate of profit; and the actual market prices are consistently a good deal higher than his static-equilibrium “prices of production.”
He can make the technical changes even less frequent, and the conclusion remains the same. Go ahead, Fred, try it.
His quotes from Marx don’t change matters one iota. They simply indicate either that Marx was wrong or, as I believe, that Moseley is misinterpreting him. If labor-saving technical change occurs and the MELT is constant, then it follows inescapably from Marx’s value theory that prices of production and the average rate of profit have to change, and that Moseley’s static-equilibrium “prices of production” and rate of profit are, once again, centers of nothing.
Four points:
1. Andrew, thanks for the revised spreadsheet that includes periods in which there is no change in productivity (nor Δb); e.g. the first three periods and every third period after that. We can see that in these periods the TSSI prices of production CONTINUE TO CHANGE, even though there is no productivity change, which is contrary to Marx’s prices of production, as I have demonstrated with substantial textual evidence (e.g. my academia paper).
This was my main criticism of the original Kliman-McGlone 1988 paper on the transformation problem – that their prices of production CONTINUE TO CHANGE for 14 periods in their example even though productivity was assumed to remain constant in all 14 periods. And in the 14th period, Kliman’s ever-changing prices of production finally converge to (guess what?) the Bortkiewicz-Sweezy equilibrium prices of production! (as K-M acknowledged, pp. 75-76)
The reason that the TSSI prices of production continue to change in his current spreadsheet example is that the quantities of output in the two sectors change every period due to unequal rates of profit in the two sectors and the process of adjustment to equal rates of profit. (The continuing equalization of the profit rate is also the reason that prices of production kept changing in the K-M 1988 paper.) But Marx’s prices of production presumes that the long-run equalization of the rate of profit has already taken place and these long-run equilibrium prices of production will not change unless there is a change of productivity (or the wage).
Therefore, Kliman’s spreadsheet example of ever-changing prices of production is a misinterpretation of Marx’s prices of production which are relatively stable long-run equilibrium prices that change only if productivity (or the wage) changes.
2. In my last comment, I considered the effect of technological change in a single industry on the TSSI price of production in that industry, without the unnecessary and distracting complications of two goods oscillating around each other. I argued (and Kliman did not respond):
“When technological change occurs in an industry (and everything else remains constant), then according to the TSSI, the price of that good as an input at time t is greater than the price of production of that good as an output at time t+1 (i.e. p(t) > p(t+1)), and as a result of that difference the historical cost rate of profit is less than the current cost rate of profit in that industry.
“Then this good produced in period t becomes an input in period t+1 and is purchased at p(t+1). If there is *no technological change in period t+1*, then the price of this good will remain the same and the price of this good as an output of period t+1 at time t+2 will also be p(t+1). Thus in period t+1 the historical cost of this good as an input is equal to its current cost and the price of this good as an input is equal to its price as an output. There is only a temporary divergence between historical cost and current cost of this good and it takes only one period for the historical cost to “catch up” to the new current cost. And since the difference between historical costs and current costs has been eliminated in the second period, so has the difference between the historical cost rate of profit and the current cost rate of profit.
“Also, the TSSI price of production calculated according to historical costs in period t+1 is different from the TSSI price of production in period t, and thus the TSSI price of production CHANGES, *even though there is no technological change* in period t+1. The historical cost price of production is only temporary and thus cannot be a long-run center of gravity for market prices. The long-run center of gravity price for market prices is the current cost price of production to which the historical cost price of production converges and which changes ONLY if there is technological change and thus is more stable and more appropriate for the long-run center of gravity price.”
Andrew, do you disagree with this argument? If so, why?
3. Kliman said: “the static-equilibrium rate of profit is calculated in the exact way that Moseley likes! Profit is the difference between the new value added by living labor and variable capital, and the rate of profit equals profit divided by the sum of constant and variable capital.”
But Kliman misunderstands my interpretation of VARIABLE CAPITAL, which is NOT derived from a given real wage (i.e. b in Kliman’s spreadsheet example) (multiplied by the unit price of the wage good), but rather variable capital depends on the balance of power between capitalists and workers in the class conflict over wages and is taken as given in the determination of surplus-value and the rate of profit (as I discussed in my comments on Part 12).
There is no such thing as a “given real wage” (i.e. a “wage bundle”) determined in advance and the same goods consumed by all workers. Workers are paid a money wage and then decide which goods to buy.
Therefore, if the price of wage goods declines due to technological change, the money wage (or variable capital) *might not decline proportional to the price of wage goods*. In this case, variable capital will be greater than calculated by Kliman, surplus-value will be less than calculated by Kliman, and the rate of profit will be lower, both because surplus-value is less in the numerator and because a greater variable capital is added to the total capital in the denominator.
Thus the rate of profit in my interpretation is not the same as the static equilibrium rate of profit in Kliman’s calculations.
4. There is also a very significant difference of interpretation of CONSTANT CAPITAL. Kliman’s model assumes that there is NO FIXED CAPITAL (i.e. no machines). But how can there be a 4% increase of labor productivity in every period (!) if there are no machines and no new machines? Magic? Miracles?
The absence of fixed capital is especially important in an analysis of the effect of increase of productivity on the rate of profit, because an increase of productivity usually requires an increase of fixed capital, which has a negative effect on the rate of profit. But this negative effect on the rate of profit is entirely missing in Kliman’s model because there is no fixed capital. Also, an increase of productivity usually requires a new type of machines so that the quantities and unit prices of the two different types of machines are not comparable.
I argue that Marx’s theory takes as given the increase of fixed capital invested (as a quantity of money capital) and includes the negative effect of the increase of fixed capital on the rate of profit. But fixed capital is missing in Kliman’s model.
Due to time constraints, I’ll have to deal with Fred Moseley’s comment of Sat, 8th Apr 2017, 9:00 am, bit by bit.
The most important bit, BY FAR, is this statement of his near the end: “an increase of productivity usually requires a new type of machines [sic] so that the quantities and unit prices of the two different types of machines are not comparable.”
Absolutely right. So it makes absolutely no sense to insist that input prices have to equal output prices, or to define prices of production in terms of input price-output price equality. Some of the inputs are not produced as outputs, so they do not have an output price. And some of the outputs are new–they didn’t exist at the time of input–so they don’t have an input price.
An idiot-monster like Jefferies may not understand this, but surely you do, Fred.
I have answered Moseley’s point 1. There is nothing more to say. I can only repeat:
A. He is misinterpreting Marx on prices of production not changing unless “productivity” or the real wage rate change, as I noted in Reclaiming Marx’s “Capital”.
B. As I have demonstrated conclusively in a previous comment here (Part 2, if I recall), and have repeated since then, Moseley’s own prices of production ALSO change even when “productivity” (input-output coefficients) and the real wage rate do not!!! Not only do the industry aggregate prices of production change. The per-unit prices of production–the prices of a commodity–also change. The passage of Marx’s in question is about “the price of production of a commodity” (emph. added). https://www.marxists.org/archive/marx/works/1894-c3/ch12.htm
More on Fred Moseley’s comment of Sat, 8th Apr 2017, 9:00 am.
At the end of his point 2, he asked, “Andrew, do you disagree with this argument? If so, why?”
Answer: Yes. Because all of it is wrong.
Elaboration:
1. The terms “current cost” and “historical cost” are incorrect. As I have explained in detail in section 6.4 of Reclaiming Marx’s “Capital” and elsewhere, the amount of value transferred from used-up means of production to products depends on the current pre-production reproduction cost of the means of production (according to Marx’s theory as understood by the TSSI). The current pre-production reproduction cost of the means of production is not the same thing as their historical cost. And since the pre-production reproduction cost of the means of production is current, we need to call the simultaneist alternative “the (current) post-production replacement cost” to avoid confusion.
2. Moseley in effect assumes a one-off technical change in a one-sector economy: “… the effect of technological change in a single industry on the TSSI price of production in that industry, without the unnecessary and distracting complications of two goods oscillating around each other.” (If there is more than one sector, the technical change will affect other sectors and there will be feedback effects on the innovating sector. The only way to abstract from that is to assume a one-sector economy.)
3. In this case, the commodity’s price of production equals its value, so the temporally determined price of production is
P(t+1) = P(t)*a + n
and the simultaneously-determined (or “long-run equilibrium”) price of production is
Ps(t+1) = Ps(t+1)*a + n,
where a is the physical input of the commodity and n is the new value added by living labor, both per unit of output.
4. Assume that, initially, P(0) = Ps(0) = n(0)/[1 – a(0)] > 0. Then, after the one-off technical change, we have different values for a and n, which we’ll call a’ and n’. Assume that a’ > 0, n’ > 0.
5. Moseley claims that, given a one-off technical change, the pre-production reproduction cost will differ from the post-production replacement cost for one period only: “it takes only one period for the [pre-production reproduction cost] to “catch up” to the [post-production replacement cost].”
This claim is absolutely false. (Fred, why do you persist in making false claim after false claim, backed up by nothing other than your intuition? Haven’t you realized that your intuition is not an adequate substitute for proof?)
6. There are 3 cases to consider.
(a) a’ = 1
In this case, we have
Ps(t+1) = Ps(t+1)*a’ + n’
or
Ps(t+1) = Ps(t+1) + n’
or
0 = n’ > 0
which is self-contradictory. Thus, the simultaneist version of “Marx’s” value theory is exposed as the fraud it is. The allegedly “long-run equilibrium” price of production does not exist, not even in the long run!
In contrast,
P(t+1) = P(t)*a’ + n’
or
P(t+1) = P(t) + n’
Since (by assumption, above) the input price in the period in which the technical change is introduced is P(t) = n(0)/[1 – a(0)] > 0, the temporally-determined output price of that period is
P(t+1) = n(0)/[1 – a(0)] + n’ > 0
In every subsequent period (given no further technical change), the output price exceeds the previous period’s output price by the amount n’. (Since P(t+1) = P(t) + n’, it follows that P(t+1) – P(t) = n’.)
So the pre-production reproduction cost never “catches up” to the non-existent post-production replacement cost. Thankfully!
(b) a’ > 1
In this case, we have, in the period in which the technical change is introduced, and in all subsequent periods (given no further technical change),
Ps(t+1) = Ps(t+1)*a’ + n’
or
Ps(t+1) = – n’/(a’ – 1)
which is negative. The allegedly “long-run equilibrium” price of production is nonsensical. (And if the amount of living labor needed to produce the commodity increases, causing an increase from n’ to n”, the value = price the commodity falls, which means that living labor subtracts value. That is completely antithetical to Marx’s theory.) Thus, once again, the simultaneist version of “Marx’s” value theory is exposed as the fraud it is.
In contrast,
P(t+1) = P(t)*a’ + n’
and since (by assumption, above), the input price in the period in which the technical change is introduced is P(t) = n(0)/[1 – a(0)] > 0, the temporally-determined output price of that period is
P(t+1) = {n(0)/[1 – a(0)]}*a’ + n’
which is positive.
And in each subsequent period (given no further technical change), the output price exceeds the previous period’s output price, by an ever-increasing amount. (It is easy to show that the increase in the price from one period to the next is a’ times the immediately preceding increase.)
So the pre-production reproduction cost never “catches up” to the negative post-production replacement cost, but diverges from it to an ever-increasing extent. Thankfully!
(c) a’ < 1
In this case, in the period in which the technical change is first introduced, we have
Ps(t+1) = Ps(t+1)*a' + n' = n'/(1 – a')
which is also its value in all subsequent periods if there is no further technical change. So it is a constant.
In contrast, the temporally-determined output price is
P(t+1) = P(t)*a’ + n’
Thus
P(t+1) – Ps(t+1) =
(P(t)*a’ + n’) – (Ps(t+1)*a’ + n’) =
a’*(P(t) – Ps(t+1))
which means that difference between the temporally-determined and simultaneously-determined output price in any period is a’ times as large as the difference in the immediately preceding period. (Recall that Ps(t+1) is constant.)
So, in this (final) case, too, it is just wrong to say that it takes only one period for the pre-production reproduction cost to “catch up” to the post-production replacement cost. Even if we assume that there is no further technical change, ever, it would take forever for the two costs to equalize.
For example, if a = a’ = 0.5, n = 64, and n’ = 32, the pre-technical change price of the commodity (assuming that P(0) = Ps(0)) is 128. After the technical change, the simultaneist output price immediately falls from 128 to 64, and remains 64 forever, while the temporalist output price falls from 128 to 96, then to 80, 72, 68, 66, 65, ….
7. Moseley is therefore also wrong when he suggests that persistent changes in temporally-determined prices of production are caused by “[t]he continuing equalization of the profit rate,” but not by a change in “productivity” (input-output coefficients). In the one-sector example above, there was no issue of equalization of profit rates! Yet the temporally-determined prices of production were shown to change period after period, even though there were no further technical changes after the initial one.
So why did the temporally-determined prices of production change period after period? The cause–the sole cause–was the one-off technical change. A change in X at time t can cause changes in Y at times t + 1, t + 2, t + 3, etc., etc., even when there are no subsequent changes in X. The absence of subsequent changes in X is no warrant for concluding that the ongoing changes in Y are not caused, or not caused solely, by the change in X.
A rainstorm can cause your roof to leak, and continue to leak and eventually collapse, long after the rainstorm ends. The fact that it dry outside when your roof is leaking and when it collapses is no warrant for concluding that these problems weren’t caused, or weren’t caused solely, by the rainstorm.
8. Finally, Moseley ends his point 2 by claiming that “The [pre-production reproduction] cost price of production is only temporary and thus cannot be a long-run center of gravity for market prices. The long-run center of gravity price for market prices is the [post-production replacement] cost price of production to which the [pre-production reproduction] cost price of production converges and which changes ONLY if there is technological change and thus is more stable and more appropriate for the long-run center of gravity price.”
This is utter nonsense. Firstly, and most importantly, Part 13 of my series of comments (see main article above) demonstrates that the “[pre-production reproduction cost] price of production”–the static-equilibrium price–is the center of nothing when market prices trend downward (or upward).
(I will deal in a subsequent comment with Moseley’s quibbles about fixed capital and wages. I will show that, when fixed capital is added in and wages are computed in his preferred manner, it is still the case that the static-equilibrium price is the center of nothing.)
Secondly, Part 13 of my series of comments also demonstrates that the “[pre-production reproduction] cost price of production”–the actual, temporally-determined, price of production can indeed be “a long-run center of gravity for market prices.”
Thirdly, the present comment has shown that the actual, temporally-determined, price of production is not “only temporary.” It does not converge to the nonexistent static-equilibrium “price of production” in the a’ = 1 case, or to the negative static-equilibrium “price of production” in the a’ > 1 case. And while it converges in the a’ < 1 case, it takes forever for the two prices to be equal. A difference that lasts forever is not temporary!
Fourthly, there is a more important reason why the actual, temporally-determined, price of production does not converge to the static-equilibrium "price of production" (as Part 13 makes clear): when market prices trend downward (or upward), as a result of continuing technical change (for instance), the relative difference between the static-equilibrium price and the actual price of production can remain stable and even increase.
Once again on Fred Moseley’s comment of Sat, 8th Apr 2017, 9:00 am.
He has a persistent and extremely annoying habit of trying to evade the implications of my demonstrations by complaining about my “assumptions”–even though the “assumptions” in question are for the sake of convenience and have nothing to do with the real issue.
Remember when he complained about my demonstration that his rate of profit is quantitatively identical to the standard physicalist rate? The first complaint was that the demonstration had only one sector. Then he complained that the demonstration, which I had graciously modified for his benefit, had only one capital good and one wage good. When I again graciously modified the demonstration for his benefit, he threw up other roadblocks, like the bizarre and meaningless claim that the physical quantities that I have found to underlie his monetary magnitudes aren’t “actual.”
Well, now he’s back to the same trick of throwing up such irrelevant roadblocks. I’ve shown that static-equilibrium “prices of production” and the static-equilibrium “rate of profit” are CENTERS OF NOTHING. Market prices and the associated rates of profit do not fluctuate around these static-equilibrium monstrosities, but around actual, temporally-determined prices of production and the actual, temporally-determined, average rate of profit.
Instead of thanking me for pointing this out, Moseley complained that my demonstration “assumes that the productivity of labor increases CONTINUALLY in both sectors IN EVERY PERIOD.” When I graciously modified the demonstration for his benefit, and showed that the static-equilibrium monstrosities are still CENTERS OF NOTHING, was I then thanked?
No, Marxian economics means never having to say you’re sorry. Instead of thanking me, Moseley has ignored the modified demonstration and thrown up a new roadblock. He now complains (in points 3 and 4 of his comment) that my demonstration assumes a given real wage and lacks fixed capital.
So I have now graciously modified the demonstration once more for his benefit. Here’s the modified demonstration–“STILL Centers of Nothing, no matter how many roadblocks one throws up”–which includes fixed capital, has no given real wage, and doesn’t have technical change in every sector in every period:
http://marxisthumanistinitiative.org/wp-content/uploads/2017/04/STILL-Centers-of-Nothing-no-matter-how-many-roadblocks-one-throws-up-.xlsx
Shall I hold my breath, waiting in vain for acknowledgement that the simultaneist version of “Marx” is bogus, until I turn blue and die?
Here’s a nouvelle version of “STILL Centers of Nothing, no matter how many roadblocks one throws up.” It computes static-equilibrium prices exactly (analytically) instead of approximating them numerically, and this allowed me to make the demonstration interactive. You can now alter all of the physical quantities as well as the rate and timing of technical changes, the MELT, the rate of surplus-value, etc.
http://marxisthumanistinitiative.org/wp-content/uploads/2017/04/STILL-Centers-of-Nothing-no-matter-how-many-roadblocks-one-throws-up-nouvelle-version.xlsx
Reply to Kliman, April 9, 10:30 am
My main criticism of the TSSI interpretation of Marx’s theory of prices of production has been that the TSSI prices of production *change every period even if there is no change of productivity* (leave aside wages for now and focus on productivity), which is clearly contrary to Marx’s prices of production, as I have demonstrated with substantial textual evidence in my paper on academia.
Kliman had two responses to this criticism in his comment on April 9.
The first response has to do with the definition of “productivity”:
“A. He [FM] is misinterpreting Marx on prices of production not changing unless “productivity” or the real wage rate change, as I noted in Reclaiming Marx’s “Capital”.”
I argue that Marx very clearly defined “productivity” as a *purely physical concept* – as the quantity of output produced by a given quantity of labor or by a unit of labor.
For example, in Section 2 of Chapter 1:
“By ‘productivity’, of course, we always mean the productivity of concrete useful labour… Useful labour becomes, therefore, a *more or less abundant source of products* in direct proportion as its productivity rises or falls.” (p. 137)
And in Chapter 12 of Volume 1:
“By an increase of productivity, we mean *an alteration in the labour process* of such a kind as to shorten the labour-time socially necessary for the production of a commodity, and to endow a *given quantity of labour* with the power of *producing a greater quantity of use-value*.” (p. 431)
Kliman argues that Marx defined “change of productivity” more broadly to include not only a change of physical productivity, but also a change in the prices of inputs without any change of physical productivity. So that (Kliman argues) when Marx said many times that prices of production change only if productivity changes, he implicitly included a change input prices (with constant physical productivity) in his meaning of “productivity changes”.
I find this interpretation ludicrous and I argue that there is no textual evidence to support Kliman’s highly speculative and non-intuitive expanded definition of productivity. Productivity is always clearly and explicitly defined and discussed by Marx as the physical quantity of output produced per unit of labor employed. Nowhere did Marx state that a change in productivity could be the result of a change of input prices, without a change in the physical productivity.
In his book, Kliman briefly presented his expanded definition of “productivity” in footnote 2 on p. 102:
“[Moseley’s] argument does not succeed unless Marx and Moseley mean the same thing by “productivity”, which Moseley fails to show. In the passage to which he refers, Marx (Marx 1991a: 307-08) used “change … in productivity” and “change in value” synonymously. And as Moseley himself agrees, Marx held that a commodity’s value depends in part of the prices of the inputs needed to produce it (see e.g. Marx 1991a: 264-65). It thus seems reasonable to conclude that a “change … in productivity” in the above sense can result from a change of input prices. On this interpretation, temporalist prices of production do not change for reasons other than changes in productivity and real wages.”
The passage that Kliman cites is the following from Chapter 12 of Volume 3:
“The price of production of a commodity can vary for only two reasons:
(1) A change in the general rate of profit…
(2) The general rate of profit remains unaltered. In this case the production price of a commodity can change only because its value has altered; because MORE OR LESS LABOR is required for its actual reproduction, whether because of a CHANGE IN THE PRODUCTIVITY of labor that produces the commodity in its final form, or in that of the labor producing those commodities that go towards producing it. The price of production of cotton yarn may fall either because raw cotton is produced more cheaply, or because the work of spinning has become more productive as a result of better machinery.”
We can see that “value has altered” in this passage because “more or less labor is required”, which is a “change of productivity” either in the production of the final commodity or in its means of production. Nowhere is there a suggestion that a change of productivity could result from a change of input prices with physical productivity constant, as in Kliman’s interpretation.
It is true that a change of value can be due to a change of input prices. But a change of input prices is itself due to a change of physical productivity in the production of inputs, e.g. raw cotton in the passage quoted above. There is no mention in this passage or anywhere else of a change of “productivity” that is due to a change of input prices without a change of physical productivity,
Therefore, I conclude that Kliman’s interpretation of Marx’s prices of production, which change every period, even though physical productivity remains the same, is a misinterpretation.
Kliman’s second response was:
“B. As I have demonstrated conclusively in a previous comment here (Part 2, if I recall), and have repeated since then, Moseley’s own prices of production ALSO change even when “productivity” (input-output coefficients) and the real wage rate do not!!!
I have already replied to this argument twice (once in a comment on Part 2 and again in an earlier comment on Part 13). But I will repeat it again and hope for a reply this time.
Kliman is correct that my definition of prices of production as “gross annual industry revenue” implies that an increase in the scale of production would increase prices of production defined in this way, even though productivity and wages remain constant. However, I argue that this fact does not bolster Kliman’s case that another cause of changes of prices of production is a change of input prices without a change of physical productivity.
I continue to think that “gross annual industry revenue” is the correct definition of prices of production in a general sense, but I now realize more clearly that in Part 2 of Volume 3 Marx analyzed prices of production in a *restricted sense*, as prices of production *per capital of 100*, i.e. of a *given scale of production*. All the industries in Marx’s tables and illustrations in Part 2 have a total capital of 100, with unequal compositions of capital (ratios of constant capital to variable capital). Marx did this in order to emphasize the effect of unequal compositions of capital across industries on the value and surplus-value produced in each industry (Volume 3, pp. 261-62). Therefore, it seems reasonable to assume that in the passages in Part 2 that I reviewed in my book and that discuss the two causes of changes in prices of production, Marx had in mind this restricted sense of prices of production *per capital of 100 and a given scale of production*. This restricted definition of prices of production rules out an increase in the scale of production as a cause of changes in restricted prices of production. Within this context, it made sense for Marx to state repeatedly that there are only two causes of changes in prices of production – changes in productivity and wages – and not to mention an increase in the scale of production (which is not theoretically interesting or important anyway) as a cause of changes in these restricted prices of production.
Kliman said in concluding his post:
“Hence, if the TSSI misinterprets Marx because it implies that prices of production can change even when technology and the real wage do not, then Moseley misinterprets Marx in the same way.”
I don’t think I misinterpreted Marx’s prices of production fundamentally, but I agree that I did not fully appreciate the significance of Marx’s restricted sense of prices of production (per capital of 100) in Part 2 of Volume 3 and the connection between this restricted sense of prices of production and Marx’s discussions of the two causes of changes in prices of production in Part 2. And I will gladly acknowledge that an increase in the scale of production is another cause of a change in prices of production in the general sense of gross annual industry revenue.
However, this additional cause of changes in prices of production in the general sense does not contradict Marx’s discussions of only two possible causes in his restricted sense. And it provides no basis for inferring that another cause of changes in prices of production (general or restricted) is changes in input prices without a change of physical productivity (and without a change in the scale of production), as in the TSSI.
The most reasonable conclusion is that the TSSI interpretation of short-run prices of production that change every period even if there is no change of productivity or wages is a misinterpretation of Marx’s long-run prices of production.
Reply to Kliman April 10 2:48 am
Kliman interprets my previous comment on April 8, Point 2, as being about a one-sector economy. But that is not what I had in mind. Rather, I had in mind a *single sector in a multi-sector economy* and in which the good produced in this single sector is an input in other sectors (not an input to itself), and in which prices are prices of production, not values. Therefore, Kliman’s entire comment (“three cases”) is not relevant to my argument.
However, I realize that my previous comment was not as clear about this distinction as it should have been. I distinguished between “this good as an output” and “this good as an input”, but I should have added to the latter: “… in the production of other goods”. So I have revised my comment below with clarifications in brackets.
REVISED PREVIOUS COMMENT.
When technological change occurs in a [single] industry [in a multi-industry economy] (and everything else remains constant), then according to the TSSI, the price of that good as an input [to the production of other goods] at time t is greater than the price of production of that good as an output at time t+1 (i.e. p(t) > p(t+1)).
Then this good produced in period t becomes an input [to the production of other goods] in period t+1 and is purchased at p(t+1). If there is *no technological change in period t+1*, then the price of this good [as an output] will remain the same and the price of this good as an output of period t+1 at time t+2 will also be p(t+1). And in period t+1 the historical cost of this good as an input [to the production of other goods] is equal to its current cost and the price of this good as an input [to the production of other goods] is equal to its price as an output. There is only a temporary divergence between historical cost and current cost of this good as in input [to the production of other goods] and it takes only one period for the historical cost to “catch up” to the new current cost.
Also, the TSSI prices of production [of other goods produced with this good as an input] calculated according to historical costs in period t+1 are different from the TSSI prices of production [of these other goods] in period t, and thus the TSSI prices of production [of these other goods] CHANGE, *even though there is no technological change* in period t+1. The historical cost prices of production are only temporary and thus cannot be long-run centers of gravity for market prices. The long-run center of gravity price for market prices is the current cost price of production to which the historical cost price of production converges and which changes ONLY if there is technological change and thus is more stable and more appropriate for the long-run center of gravity price.
A reply to Fred Moseley’s comment of Fri, 14th Apr 2017, 8:01 am.
He quotes a bit of what I wrote above (Sun., 9th Apr 2017 10:30 am):
and then seems to respond to it–with his usual stuff about “prices of production as ‘gross annual industry revenue’” and “prices of production *per capital of 100 and a given scale of production*.”
But he ignores what immediately followed the bit he quoted. Here is the full passage:
I first called attention to this fact 11 months ago, on May 12, 2016, in Part 2 of this series of comments:
Although Moseley now complains that “I have already replied to this argument twice,” he has NEVER addressed this point about the causes of changes in the price of production of a commodity. Not once, in 11 months. And his latest reiteration of the usual stuff about “prices of production as ‘gross annual industry revenue’” and “prices of production *per capital of 100 and a given scale of production*” doesn’t address it.
In other words, the issue is not what happens to the total price of output of the car industry or coal industry; and the issue is not happens to the price of output of the car industry or coal industry per $100 of advanced capital. The issue IS: what happens to the price of production of A car or A pound of coal–one car or one pound of coal? I’m at a loss as to how I can make this any clearer.
The rest of his comment is about some quotes from Marx on causes of changes in prices of production. I have already noted (Tues., 4th Apr, 2017, 11:24 am) that
Moseley has not addressed this. Unless and until he does, I have nothing to add.
A reply to Fred Moseley’s comment of Fri., 14th Apr. 2017, 8:08 am.
You’re still wrong.
1. One reason you’re wrong is a reason I have already pointed out:
2. A second reason you’re wrong is that it is still absolutely false that, given a one-off technical change, the pre-production reproduction cost will differ from the post-production replacement cost for one period only, as you claim (“it takes only one period for the [pre-production reproduction cost] to “catch up” to the [post-production replacement cost]”).
A complete description of the possible trajectories of prices of production is beyond my analytical capabilities when there’s more than one sector. Instead, I’ve considered 1 initial case and 2 changes to that case.
There are 2 sectors. Sector 1 produces a means of production used by both sectors, and Sector 2 produces articles of consumption consumed by workers in both sectors.
The physical quantities in the initial case are: a1 = 0.5, a2 = 1/3, l1 = 0.5, l2 = 2/3, b = 0.4, x1 = 10, x2 = 15. Assuming that a static equilibrium prevails, that total price = total value, and that the MELT = 1, the simultaneist and temporalist prices of production, equal to each other, are p1 = 1.2 and p2 = 1.
The 2 changes are both changes to a1 and nothing else.
(a) If the new a1 > 1.0691996…, then the new simultaneist (static-equilibrium, “long-run”) prices of production are nonsensical. One or both of them are negative. When the new a1 = 1.0691997, these “prices of production” are either p1 = -888693354.013 and p2 = -337229325.997 or p1 = -0.210 and p2 = 0.920. Take your pick. (Please bear in mind that I’m talking about per-unit prices of production, not “gross annual industry revenue” or “prices of production *per capital of 100 and a given scale of production*”–though I guess those aren’t strictly non-negative either.)
In contrast, the actual, temporally-determined, prices of production remain positive and keep increasing indefinitely. When the new a1 = 1.0691997, the prices of production are
p1 = 2.021 and p2 = 0.908 immediately following the one-off technical change. 100 periods later, they are p1 = 688.217 and p2 = 252.970.
(b) If the new a1 < 1.0691996..., and if there is no further technical change, AND no change in the real wage rate, AND no change in levels of output, EVER, then the actual, new, temporally-determined prices of production "converge on" the new simultaneist (static-equilibrium, "long-run") prices of production. But it would take forever for the differences between these two sets of prices to be fully eliminated. So, in this case, too, it is just wrong to say that it takes only one period for the pre-production reproduction cost to “catch up” to the post-production replacement cost. It takes an eternity! 3. A third reason you're wrong is that it remains utter nonsense to claim, as you do, that the "[pre-production reproduction] cost prices of production are only temporary and thus cannot be long-run centers of gravity for market prices. The long-run center of gravity price for market prices is the [post-production replacement] cost price of production to which the [pre-production reproduction] cost price of production converges and which changes ONLY if there is technological change and thus is more stable and more appropriate for the long-run center of gravity price." Firstly, and most importantly, Part 13 of my series of comments (see main article above) demonstrates that the “[post-production replacement cost] price of production”–the static-equilibrium price–is the center of nothing when market prices trend downward (or upward). As I have now shown, that's the case even when technical changes are intermittent, and there is fixed capital, and wages are computed in your preferred manner;--the static-equilibrium price is still the center of nothing. Secondly, Part 13 of my series of comments also demonstrates that the “[pre-production reproduction] cost price of production”--the actual, temporally-determined, price of production--can indeed be “a long-run center of gravity for market prices.” Thirdly, the present comment has shown that the actual, temporally-determined, prices of production are not “only temporary.” In my example, they do not converge to the nonsensical static-equilibrium “prices of production” in the case in which the new a1 > 1.0691996…. And while they do converge in the other case, it takes forever for the temporalist and simultaneist prices of production to be equal. A difference that lasts forever is not temporary!
Fourthly, there is a more important reason why the actual, temporally-determined, price of productions do not converge to the static-equilibrium “prices of production” (as Part 13 makes clear): when market prices trend downward (or upward), as a result of continuing technical change (for instance), the relative difference between the static-equilibrium price and the actual price of production can remain stable and even increase.
In conclusion: Fred, why do you persist in making false claim after false claim, backed up by nothing other than your intuition? Haven’t you realized that your intuition is not an adequate substitute for proof?
I notice that Fred Moseley has responded to some of my replies to his comment of Sat, 8th Apr 2017, 9:00 am, but not others. I hope he will do so. In particular, I’d like him to respond to the first one:
“Due to time constraints, I’ll have to deal with Fred Moseley’s comment of Sat, 8th Apr 2017, 9:00 am, bit by bit.
“The most important bit, BY FAR, is this statement of his near the end: ‘an increase of productivity usually requires a new type of machines [sic] so that the quantities and unit prices of the two different types of machines are not comparable.’
“Absolutely right. So it makes absolutely no sense to insist that input prices have to equal output prices, or to define prices of production in terms of input price-output price equality. Some of the inputs are not produced as outputs, so they do not have an output price. And some of the outputs are new–they didn’t exist at the time of input–so they don’t have an input price.
“An idiot-monster like Jefferies may not understand this, but surely you do, Fred.”
I just realized that there’s a general proof–independent of the number of sectors, specific input-output relations, and everything else– that Fred Moseley’s claim that “it takes only one period for the [pre-production reproduction-cost prices of production] to ‘catch up’ to the new [post-production replacement-cost prices of production]” is false.
The proof is exceedingly simple. If we assume that the claim is true, we end up in self-contradiction.
Define
P1 = set of temporally-determined prices immediately after a one-off technical change
P2 = set of temporally-determined prices in the subsequent period
Pe = static-equilibrium prices immediately after a one-off technical change (and thereafter)
Moseley’s “it takes only one period to ‘catch up'” claim implies:
(a) P1 ≠ Pe (immediately after a one-off technical change, the temporally-determined prices haven’t caught up)
(b) P2 = P1 (since the temporally-determined prices have adjusted to a static equilibrium state –“caught up”–by the next period, such that the input prices (P1) equal the output prices (P2))
(c) P2 = Pe (the temporally-determined prices are now the same as the static-equilibrium prices)
But
if (a) and (b) are true, then P2 ≠ Pe, which contradicts (c);
if (a) and (c) are true, then P2 ≠ P1, which contradicts (b);
and
if (b) and (c) are true, then P1 = Pe, which contradicts (a).
Q.E.D.
And now I’ve realized that the proof above plus the following constitutes a proof by induction (or something similar, since the proof moves backward instead of forward in time) that the pre-production reproduction-cost prices of production NEVER “catch up” to the new post-production replacement-cost prices of production.
“Catching up” implies
(d) Pt+1 = Pt and
(e) Pt+1 = Pe,
which, taken together, imply that (e’) Pt = Pe.
But if (e’) Pt = Pe, then (d’) Pt = Pt-1
and (d’) and (e’), taken together, imply (e”) Pt-1 = Pe.
And so on, until we reach
(e”’…’) P1 = Pe, which contradicts (a).
So “catching up” doesn’t occur. I.e., the set of temporally-determined prices differs from the set of static-equilibrium prices FOREVER.
Q.E.D.
I am replying to Kliman’s latest comments individually as time permits. I am busy with other things these days.
Reply to Kliman, April 21, 8:01 am:
I don’t insist that input prices = output prices in a period in which there is technological change that changes the types of inputs or outputs. There would be a period of adjustment to the new long-run equilibrium.
My book is about the transformation problem, with the assumption of constant technology, so this issue doesn’t come up. The whole debate over the transformation problem has assumed constant technology, including Kliman-McGlone (1988 and 1996).
Fixed constant capital is included in my interpretation and is taken as given, as quantities of money capital advanced in the beginning of the circuit of money capital to purchase long-lasting means of production, both in the macro theory of the total surplus-value and in the micro-industry theory of prices of production. Since the economy is assumed to be in long-run equilibrium (in order to explain prices of production as long-run equilibrium prices, input prices = output prices.
Input prices = output prices is *not a necessary condition* in my interpretation. Input prices = output prices is a *consequence* of long-run equilibrium prices.
Reply to Kliman April 14, 2:58
Andrew, you are right that a change in the relative proportions of capitals with unequal compositions of capital can change prices of production (as industry total) and also can change unit prices of production.
Marx discussed the effect of the distribution of capital across industries on the rate of profit (and hence on prices of production) in Chapter 9 (pp. 262-63). For example, an increase in the relative proportion of industries with a lower composition of capital will increase the rate of profit (and hence increase the prices of production of all commodities).
However, in Marx’s many discussions of the “causes of changes in prices of production”, he never mentioned the distribution of capital as one of the factors that affect the rate of profit and hence affect prices of production. So I did not mention this factor in my summary of these discussions.
But this additional factor does not support your interpretation of prices of production that change every period even if productivity and wages and the scale of production and the distribution of capital across industries all remain constant. All these causes of changes in prices of production are conditions of production, and affect the production of value and surplus-value. Your prices of production change every period, not because of changes in the conditions of production of value and surplus-value, but because input prices ≠ output prices, and thus prices of production must change every period in order to equalize the rate of profit. Marx never suggested that prices of production change every period because input prices ≠ output prices, even if all the conditions of the production of value and surplus-value remain the same. That would not make sense because Marx’s prices of production are long-run equilibrium prices which implies that input prices = output prices.
Marx concluded his discussion of “The Causes of a Change in the Price of Production” in Section 1 of Chapter 12 of Volume 3 as follows:
“All changes in the price of production of a commodity can be ultimately reduced to A CHANGE OF VALUE …” (p. 308; emphasis added). But your prices of production change because input price ≠ output prices without a change of value.
Also, there is a problem with Table 1 in your Part 2. The relative proportion of capital in Branch 2 with lower composition of capital increases, but the rate of profit does not increase, which is contrary to Marx’s theory. How do you determine the rate of profit in this table? And also the rate of surplus-value decreases in both branches. Why does an increase in the relative proportion of capital in Branch 2 affect the rate of surplus-value in both branches?
Reply to Kliman April 14, 5:05 pm
Kliman’s comment does not engage my comment at all. My comment was about a single sector in a multi-sector economy and Kliman’s reply was again in terms of his two-sector economy.
My main point was this: if there is no further technological change in period t+1 in the industry that produces the input (and everything else remains constant), then the price of that input commodity at time t+2 will be equal to the price of that commodity as inputs in other industries at time t+1 (i.e. if no further technological change, then no further change of price of the input commodity). From this it follows that:
1. In period t+1, the pre-production cost of goods produced with this input = post-production cost of these goods.
2. In period t+1, the price of production of goods produced with this input valued at pre-production cost = price of production of these goods produced with this input valued at post-production cost, and thus the pre-production cost price of production converges to the post-production price of production after one period.
Reply to Kliman April 15, 6:35 pm
The flaw in Kliman’s first follow-up comment is step (b). We need to distinguish between the price of the good used as an input in the production of other goods (call it P(A)) and the price of goods produced with good A as an input (call it P(B)).
So Kliman’s (b) is P1(A) = P2(B). But this is not true. P1(A) is only one component of P2(B). That is:
P2(B) = P1(A) + X.
P1(B) = P0(A) + X.
And since P1(A) ≠ P0(A), then P2(B) ≠ P1(B).
That is, P(B) changes in period 2 even though there is no technological change in period 2, contrary to Marx’s prices of production, and P2(B) catches up and converges to Pe(B) in the second period.
Reply to Kliman April 15, 6:53 pm
The flaw in Kliman’s second follow-up comment is both (d) and (e). In the example I presented, “catching up” does not imply Pt+1(B) = Pt(B). Time t is the beginning of period t (the purchase of the inputs) before the technological change; so Pt(B) (the price of the output of period t) is not defined at time t. Pt+1(B) is defined (at the end of period t), but it is ≠ Pe(B), because P(A) is ≠ post-production cost (instead, Pt+2(B) = Pe). So obviously Pt(B) ≠ Pe(B), and nothing else follows.
Reply to Kliman April 3
In the above comment, Kliman altered his spreadsheet to include periods in which there was *no productivity change* (this was prior to his most recent spreadsheet with fixed capital). But I think that the revision he made was too limited. Only one period out of three had productivity constant in both sectors (not enough time to converge to long-run equilibrium prices of production) and when productivity did change, it was three times as large as in the original model (12% rather than 4%).
My hypothesis is that if productivity remained constant for a few more periods, then the average rate of profit would converge to the SE rate of profit and the TSSI prices of production would converge to the SE prices of production, as the TSSI prices of production converge to the Bortkiewicz-Sweezy long-run equilibrium prices of production in Kliman-McGlone (1988).
So I altered the equations in the spreadsheet to hold l1, l2, and b constant for periods 8-16.
The results are shown at http://www.mtholyoke.edu/lits/ris/convergencof%20TSSI%20to%20SE.xlsx
I hope the link works!
And the results confirm my hypothesis. The average rate of profit converges to the SE rate of profit in 4 periods and the TSSI prices of production also converge to the SE prices of production in 4 periods.
Therefore, market prices may fluctuate around TSSI prices of production in the short-run with productivity changes (although these fluctuations are due largely to Kliman’s unrealistic model with only two commodities in the economy which oscillate around each other), but in the long-run with constant productivity in an industry, the TSSI prices of production converge to the SE prices of production and market prices fluctuate around the SE prices of production.
The relatively stable SE prices of production are more appropriate for long-run center of gravity prices than the TSSI prices of production that change every period even if there is no change of productivity (or other conditions of production discussed in recent comments) until the TSSI prices of production converge to the SE prices of production.
The reason for this convergence (as in all the TSSI examples) is the TSSI “iterative” interpretation of prices of production. The output price of production of one period becomes the input price of the next period, along with the “iteration rule” of equal rates of profit (and a normalization condition). In these iterations, the input price becomes closer and closer to the output price and both CONVERGE to the long-run equilibrium price with input prices = output prices. This is the nature of the iterative method.
On Sun., 9th Apr 2017, 10:16 am, I wrote
I have since repeated this comment, and Moseley has now replied to it (Sun., 16th Apr., 2017, 10:41 am). I thank him for doing so. He writes,
But when has there ever been–under capitalism–a period without either “technological change that changes the types of inputs or outputs” or “adjustment to [sic] the new long-run equilibrium” (which occurs whenever there is tecnological change of any sort, and for a variety of other reasons)? I.e., when has there ever been a period in which the economy is actually in this alleged long-run equilibrium?
NEVER.
You know that as well as I, Fred. Everyone knows it, including Marx:
So your theory of “prices of production,” which refers ONLY to the alleged “long-run equilibrium,” therefore refers to something that has never existed, does not exist, and never will exist. In contrast to Marx’s theory of prices of production, it is not a theory of prices of production within capitalism. It is a theory of a nonexistent Sraffian mirage (or neoclassical mirage).
As I noted in Part 13 (main text above),
To anticipate my response to your spreadsheet, Fred, I note that convergence on something (given implausibly long lags in technical change and other kinds of change) isn’t remotely the same as fluctuations around that something.
(And that’s leaving aside the fact that the something being converged on isn’t your prices of production, since you supposedly reject the use of physical quantities and simultaneous valuation to determine your static-equilibrium prices!)
A reality check regarding Fred Moseley’s spreadsheet.
He claims that actual prices fluctuate around his nonexistent static-equilibrium “prices of production,” based on a simulation in which there’s no increase in productivity anywhere in the economy for eight straight periods. And that’s supposed to tell us something about what happens in the long run?!
It doesn’t. In reality, productivity continually increases– it is always increasing somewhere in the economy and therefore it continually increases in the aggregate economy as well. See http://marxisthumanistinitiative.org/wp-content/uploads/2017/04/No-they-do-not.jpg .
So, as I noted above, in contrast to Marx’s theory of prices of production, Moseley’s theory is not a theory of prices of production within capitalism. It is a theory of a nonexistent Sraffian mirage (or neoclassical mirage).
Reply to Kliman April 17
The reason I assumed no productivity change in both sectors is that “no productivity change” means both in a final goods sector and in sectors which produce inputs for the final goods. And in your model, both goods are inputs for the other.
And since in your model there are only two goods this looks like “no productivity change anywhere in the economy”. If there were more goods, productivity changes in other sectors that do not produce inputs for these two sectors would not affect the prices of production in these two sectors.
In any case, my main point is that in periods in which there is no productivity change in an industry or in the production of its inputs, the TSSI iterative interpretation of prices of production *will converge to the SE prices of production* (as in Kliman-McGlone 1988) because that is the nature of the iterative method.
Andrew, do you dispute this convergence of your iterative method?
I will be traveling for the next week, but I will return to the discussion when I get back.
Fred,
Stop trying to change the subject.
Your stuff about an economy without technical change, in which prices don’t trend downward (or upward), is just a smokescreen that diverts from the real issue here.
So let me refresh your memory as to what the real issue here is.
The “Centers of Nothing” example in Part 13 is about the fact that static-equilibrium prices are not centers of gravity of market prices in actual economies that experience technological change.
What I wrote, when introducing the example, was: “The notion that market prices fluctuate around, or converge upon, static-equilibrium prices is … a bizarre fantasy when it is applied to actual economies that experience technological change, as the following example illustrates” (emphasis added).
In a comment of March 14, I wrote: “Its purpose is to illustrate the fact that, when prices trend downward (or upward), a static-equilibrium rate of profit, such as Moseley’s, will not be the center around which actual rates of profit fluctuate. Instead, they will fluctuate around their weighted average” (emphasis added).
And I reiterated that point in a comment of March 16.
In response to an April 8 comment of yours that falsely alleged that “The long-run center of gravity price for market prices is the current cost price of production,” I wrote on April 10 that “Part 13 of my series of comments (see main article above) demonstrates that the ‘[pre-production reproduction cost] price of production’–the static-equilibrium price–is the center of nothing when market prices trend downward (or upward)” (emphasis added).
In the same comment, I also wrote, “Fourthly, there is a more important reason why the actual, temporally-determined, price of production does not converge to the static-equilibrium ‘price of production’ (as Part 13 makes clear): when market prices trend downward (or upward), as a result of continuing technical change (for instance), the relative difference between the static-equilibrium price and the actual price of production can remain stable and even increase” (emphasis added).
Please deal with the real issue here. Please turn your attention to whether static-equilibrium prices are centers of gravity of market prices IN ACTUAL ECONOMIES IN WHICH PRICES TREND UPWARD OR DOWNWARD BECAUSE OF TECHNOLOGICAL CHANGE (and other reasons).
Reply to Kliman April 19
Kliman complained that my recent comments “changed the subject” and that I should stick to the subject of the results of his Excel model, that with continuous technological change in his two-good model, market prices fluctuate around the TSSI prices of production and not the SE prices of production (which he mistakenly identifies with my interpretation).
But my recent comments did not change the subject. Rather they criticized the very unrealistic assumptions of his model and argued that the *conclusions of his model do not apply to the real capitalist economy*. I argued first that the assumption of continuous technological change in both industries (which amounts to the whole economy in his model) is very unrealistic. For any given industry, productivity increases are likely to be discontinuous and infrequent. And I revised his model and held productivity constant in both industries for 8 periods (in both industries because each good was an input to the other good) (and kept all the other unrealistic assumptions; more on this below) and showed that the TSSI prices of production converged to the SE prices of production in 4 periods.
And I argued that convergence to the SE prices of production is a property of Kliman’s iterative interpretation of prices of production, so that as soon as there are periods without productivity change in an industry (or in industries that produce its inputs), the TSSI price of production in that industry will always converge to its SE price of production (as in Kliman-McGlone 1988). I asked Kliman if he disputed this convergence property of the iterative method and he has not replied.
But this argument still assumed all the other unrealistic features of Kliman’s model which distort the results of the model and make them inapplicable to the real capitalist economy. These unrealistic features include:
1. Only two goods in the total economy.
2. Since there are only two goods, a productivity increase in one industry has a significant effect on the general rate of profit for the “total economy” and thus has a significant effect on the price of other good through the rate of profit.
But in the real capitalist economy with 1000s of industries, a productivity increase in one industry has a negligible effect on the general rate of profit
3. Furthermore, each of the two goods is an input both to its own production (that again!) and also each good is an input to the other good! These interrelations magnify and distort the effects of a productivity increase in one industry on both industries.
Again, there is nothing like these “twin goods” in the real capitalist economy.
4. In addition, these two goods are assumed to be locked into a reciprocal physical relation with each other such that the quantities produced of the two goods fluctuate around each other (between 101 and 99) (see the reaction function for x; the second of Kliman’s four basic equations). This locked in physical relation further distorts the effects of a productivity increase in one industry on both industries. The effects of an increase of productivity in one industry reverberate back and forth for a number of periods that slow down the convergence of the TSSI prices of production to the SE prices of production.
Again, there is nothing like this reciprocal physical relation between two goods in the real capitalist economy.
5. Finally, the two market prices are determined in an entirely ad hoc and unrealistic way (pi = total value / 2xi) (see the first and fourth equations of Kliman’s four basic equations). Market prices do not fluctuate randomly according to the accidental causes of S and D as in reality and in Marx’s theory, but instead are locked into relations with each other and with the relative physical quantities of the two goods (p1 = p2 (x2/x1)). So Kliman’s “market prices” are not actual market prices, but are instead artifacts of his very unrealistic algebraic model. Kliman’s artificial “market prices” fluctuate around his TSSI prices of production because of the above unrealistic assumptions of his model (as long as productivity increases continue), but this proves nothing about the fluctuations of market prices in the actual capitalist economy.
Therefore, the *conclusions of this extremely unrealistic model do not apply to the actual capitalist economy*. This is not an appropriate way to analyze the effect of a productivity increase on prices of production and the fluctuations of market prices around long run center of gravity prices.
I also argued that the effects of productivity increase in an industry on its price of production should instead be analyzed as described below.
In the following, I will use the following abbreviations:
HCPP for prices of production based on historical cost
CCPP for prices of production based on current cost
(Kliman has complained that my use of the terms historical cost and current cost is inaccurate and the terms should be pre-production current cost and post-production current cost. The difference between historical costs and pre-production current cost is that historical cost is the price of inputs when the inputs are purchased and pre-production current cost is the price of inputs when they enter production. So if there is a change of the price of inputs between the purchase of inputs and their entry into production, then historical cost will not be equal to pre-production current cost. However, there is no such change in the price of inputs in Kliman’s Excel model. The price of inputs in his model is the price at which the inputs are purchased (i.e. their historical costs). So for simplicity, I will refer to the price of inputs at historical cost (HC) and current cost (CC).
1. A productivity increase in an industry will lower the price of production of that good (call it Good A and its price p(A)). If good A is a final good (i.e. not an input to other goods), then the reduction of p(A) will have no effect on the price of all other goods. And that is the end of the story. Since there was no productivity increase in the production of the inputs of Good A, the HC of its inputs = their CC and thus HCPP(A) = CCPP(A). Market prices will fluctuate around the new PP according to the accidental causes of S and D.
2. If Good A is an input to the production of other goods (call them collectively Good X), then (according to the TSSI) in the first period the HC of the inputs to Good X will be > the CC of the inputs to Good X and HCPP will be > CCPP.
3. However, *if there is no further increase of productivity* in the production of Good A, then (according to the TSSI) in the next period the HC of the inputs to Good X will be = to the CC of the inputs to Good X and HCPP will be = to CCPP. Thus the HCPP of Good X converges to its CCPP after one period.
4. If productivity increases again in the production of Good A in the second period, then the second period would be a repetition of the first period, with both HCPP(X) and CCPP(X) decreasing further and continuing to be ≠.
However, as soon as the increase of productivity in the production of Good A stops, then (according to the TSSI) the HC of the inputs to Good X will be = to the CC of the inputs to Good X and HCPP(X) will be = to CCPP(X). The convergence is within one period.
5. While the productivity increases are happening, market prices MIGHT fluctuate around the HCPP if the perception of capitalists is slow to catch up to the new reality, but the HCPP itself converges to the CCHP as soon as there is no productivity increase, so that at that point market prices are fluctuating around CCPP, and the CCPP is the true long-run center of gravity prices, which will remain in effect until there is a new productivity increase.
A reply to Fred Moseley’s comment of Mon, 1st May 2017 8:19 am.
I.
When he says that he did not try to change the subject, he is either lying or doesn’t understand what Fred Moseley has written.
As I reiterated on a couple of weeks ago, the subject under discussion is whether static-equilibrium prices are centers of gravity of market prices in actual economies that experience technological change:
Moseley now claims that “my recent comments did not change the subject. Rather they criticized the very unrealistic assumptions of his model and argued that the *conclusions of his model do not apply to the real capitalist economy*.”
However, the evidence is clear; he did indeed try to divert the discussion onto what happens in economies without technical change:
But it gets even worse. Even in his latest comment—the one in which he claims that he didn’t change the subject away from whether static-equilibrium prices are centers of gravity of market prices in actual economies that experience technological change—he writes,
However, I did reply. My reply was that his stuff about an economy without technical change, in which prices don’t trend downward (or upward), is just a smokescreen that diverts from the real issue here, and that he should stop trying to change the subject.
II.
Moseley has now climbed back on his old hobby-horse—complaining about the “unrealistic features of Kliman’s model [sic].” As I noted on Mon, 10th Apr 2017 12:21 pm,
I have demonstrated that–in actual economies that experience technological change–static-equilibrium “prices of production” and the static-equilibrium “rate of profit” are centers of nothing. Moseley first tried to evade the implications of my demonstration by complaining that it “assumes that the productivity of labor increases CONTINUALLY in both sectors IN EVERY PERIOD.” I graciously modified the demonstration for his benefit, and showed that the static-equilibrium monstrosities are still centers of nothing.
Next—ignoring this modification—he tried to evade the implications of my demonstration by complaining that it assumes a given real wage and lacks fixed capital. Once again, I graciously modified the demonstration for his benefit, and showed that the static-equilibrium monstrosities are still centers of nothing.
Now–ignoring those modifications as well– he tries to evade the implications of my demonstration by complaining about the following “unrealistic features”:
Of course, I have neither the ability nor the computing power to produce an example spanning tens of periods in which there are “1000s of industries,” none of which use their own products as inputs, prices are determined in a manner that Moseley can’t find a way to pick at, etc. I am sure that he knows this. That is why I strongly suspect that the purpose of Moseley’s irrelevant complaints is just to prevent me from convicting him of error (as Keynes put it)—i.e., just to teach me the lesson that Marxian economics means never having to say you’re sorry.
But fortunately, I don’t need to produce such an example. All of Moseley’s latest quibbles—like the earlier ones about productivity changing in each industry in each period, a given real wages, and the lack of fixed capital—are irrelevant because, as I have already said, “the ‘assumptions’ in question are for the sake of convenience and have nothing to do with the real issue.” That is why it is an example, not a model. It is therefore possible to address the real issue without recourse to the evil “assumptions” of my “model.”
What makes my demonstration—and every modification of the demonstration—work are not its “assumptions,” but aggregate invariances that hold true whatever the assumptions may be concerning individual industries and physical input-output relations.
Consider a capitalist economy consisting of “1000s of industries,” in which productivity doesn’t change in each industry in each period, no industry uses its own product as an input, the real wage isn’t given, etc. Nonetheless, it is an economy in which prices tend to trend downward because of technological change. (“Tend to” means that this would occur in the absence of other changes, such as inflation.) Because there is technological change somewhere in the economy in (almost) every period—as I’ve shown to be the case in the US–the falling tendency of prices is (almost) continuous.
This has an important consequence. The prices that prevail at the start of next period are, on average, lower than the prices that prevail at the start of this period. And since the start of next period is the end of this period, the prices that prevail at the end of this period are, on average, lower than the prices that prevail at the start of this period. But since
(a) the static-equilibrium rate of profit is based on the “realistic feature” that end-of-period prices are not lower than start-of-period prices,
and
(b) the lower end-of-period prices are in relation to start-of-period prices, the lower the general rate of profit will be,
it follows that
(c) the general rate of profit must (almost) always be lower than the static-equilibrium rate of profit.
There are also consequences pertaining to individual industries’ prices and rates of profit that follow from these aggregate invariances.
First, since the general rate of profit must (almost) always be lower than the static-equilibrium rate of profit, and the general rate of profit is a weighted average of the actual rates of profit, it must also (almost) always be the case that individual industries’ rates of profit are lower, on average, than the static-equilibrium rate of profit. The individual industries’ rates of profit therefore can’t fluctuate around the static-equilibrium rate of profit.
Second, the individual industries’ output prices can’t fluctuate around static-equilibrium “prices of production.” If they did so, then the individual industries’ rates of profit would fluctuate around the static-equilibrium rate of profit. But individual industries’ rates of profit can’t fluctuate around the static-equilibrium rate of profit. The individual industries’ output prices therefore can’t fluctuate around static-equilibrium “prices of production.”
The following spreadsheet illustrates the fact that what makes my demonstration work are aggregate invariances that hold true whatever the assumptions may be concerning individual industries and physical input-output relations:
http://marxisthumanistinitiative.org/wp-content/uploads/2017/05/Centers-of-Nothing-aggregate-invariance-properties-5.3.171.xlsb
There are 10,000 industries! I make no assumptions about input-output relations or the real wage rate.
Because I have neither the ability nor the computing power to produce a 10,000-industry example that spans tens of periods and determines prices in a manner that Moseley can’t find a way to pick at, I simply consider a single period. (But one can easily generate additional periods by pressing “F9” and thereby generating a whole new set of capital advances.)
I set the static-equilibrium rate of profit at 25%. But I assume that there is labor-saving technical change that causes output prices to be 4% lower than input prices, on average. It follows that the actual, temporally-determined general rate of profit is only 20%.
The capital advances are randomly generated from a uniform distribution. The actual industry-level output prices satisfy the condition that (a) the actual aggregate price of output is 4% less than the aggregate static-equilibrium price of output, due to the 4% fall in output prices relative to input prices, and (b) actual, temporally determined rates of profit are normally distributed.
The results are quite clear. For reasons that have nothing to do with the nonsense that Moseley has been yammering about, but have everything to do with aggregate invariances and their elementary statistical implications, the example shows once again that static-equilibrium prices and the static-equilibrium rate of profit are centers of nothing in actual economies in which prices tend to fall as a result of labor-saving technological change.
* About 80% of the industries’ rates of profit are less than the static-equilibrium rate of profit, while about half are above and half are below the temporally-determined general rate of profit.
* About 80% of the industries’ output prices are less than the corresponding static-equilibrium prices, while about half are above and half are below the temporally-determined prices of production.
* The weighted average of the industry-level rates of profit turns out to be exactly equal to the temporally-determined general rate of profit, but 5 percentage points less than the static-equilibrium rate of profit.
* The temporally-determined aggregate price of output is exactly equal to the temporally-determined aggregate price of production, but 4% below the aggregate static-equilibrium price.
None of these results are compatible with the idea that static-equilibrium prices and the static-equilibrium rate of profit are centers of gravity. All of them are compatible with the idea that the actual, temporally-determined, prices of production and general rate of profit are centers of gravity.
I also tried different distributions of rates of profit–a normal distribution with a larger standard deviation, and a uniform distribution. The numerical details differed, but the basic results were the same. And I’ve tried different values for the static-equilibrium rate of profit and actual, temporally-determined general rate of profit. Unless the two rates are equal, the basic results are, once again, the same.
Note that these results depend on no assumptions other than the all-important one: technical change–somewhere in the economy–has caused end-of-period prices to fall below start-of-period prices, on average, which in turn causes the actual general rate of profit to fall below the static-equilibrium rate of profit.
Although these results seem to pertain to only one period, they actually pertain to (almost) every period since there is technical change—somewhere in the economy—in every period.
I hope the above has put a stop to the persistent and extremely annoying practice of trying to evade the implications of my demonstrations by complaining about my “assumptions”–even though the “assumptions” in question are for the sake of convenience and have nothing to do with the real issue. But I’m a realist. I understand that Marxian economics means never having to say you’re sorry. It’s part of the post-truth, alternative-facts world of Trumpism.
A reply to Fred Moseley’s comment of Sun, 16th Apr 2017 11:25 am.
He’s wrong. There is nothing wrong with my proof that pre-production reproduction-cost prices of production don’t “catch up” to the post-production replacement-cost prices of production after one period. And there is nothing wrong with my proof that they never “catch up.”
His criticisms make no sense, in part because he didn’t bother to understand my notation.
I wrote
Note that the P’s are SETS of prices and that the “subscripts” 1, 2, are PERIODS.
Moseley turns the “subscripts” into industry designators and thus turns the P’s into individual goods’ prices.
He thus begins, “Kliman’s comment does not engage my comment at all. My comment was about a single sector in a multi-sector economy and Kliman’s reply was again in terms of his two-sector economy.”
And it goes downhill from there.
A reply to Fred Moseley’s comment of Sun, 16th Apr 2017 10:51 am.
In that comment, he finally acknowledged that “a change in the relative proportions of capitals with unequal compositions of capital can change prices of production (as industry total) and also can change unit prices of production.” This is a welcome admission. It directly contradicts his claim—which he has repeated again and again and again—that the only two causes of changes in prices of production, according to Marx’s theory, are changes in “productivity” (i.e., input-output coefficients) and changes in the real wage rate.
This welcome admission comes 11 months after I showed, in Part 2 of this series of comments that Moseley’s “prices of production” change even if input-output coefficients and the real wage rate do not. Faced with a choice between his theory of how prices of production are determined and “what Marx actually wrote” about the causes of changes in prices of production, Moseley has opted to rescue his theory and throw overboard “what Marx actually wrote” (in marked contrast to his ostensible hermeneutical principle of long standing!).
In any case, the debate about whether changes in input-output coefficients and the real wage rate are the sole causes of changes in prices of production has come to an end. Moseley has acknowledged that he was wrong to claim that these are the only causes.
He now concedes that “the distribution of capital [i]s one of the factors that affect the rate of profit and hence affect prices of production.” He concedes this even though “in Marx’s many discussions of the ‘causes of changes in prices of production’, he never mentioned the distribution of capital as one of the factors that affect the rate of profit and hence affect prices of production.” Note that this implies that, in Marx’s theory, there can be factors that affect prices of production even though Marx “never mentioned” that they affect prices of production when he discussed why prices of production change.
But (with Moseley, there’s always a “but”) he continues to deny the validity of the temporal single-system interpretation of prices of production. Why? Because “Marx never suggested that prices of production change every period because input prices ≠ output prices, even if all the conditions of the production of value and surplus-value remain the same.”
Does anyone else see the glaring internal inconsistency in Moseley’s interpretive practice? He would have us believe that his interpretation of prices of production is fine and dandy even though Marx never mentioned one of the factors that affect Moseley’s prices of production when discussing why prices of production change. But he would also have us believe that the temporal single-system interpretation of prices of production was produced in the Antichrist’s workshop because Marx never mentioned one of the factors that affect TSSI prices of production when discussing why prices of production change.
It is true that Moseley manages to type out an additional reason to deny the validity of the temporal single-system interpretation of prices of production: “Marx’s prices of production are long-run equilibrium prices which implies that input prices = output prices.” But he’s simply begging the question, assuming what he needs to prove. If the TSSI is valid, then Marx’s prices of production exist even when input prices don’t equal output prices. So, in order to arrive at his “conclusion” that the TSSI is invalid because its prices of production exist even when input prices don’t equal output prices, Moseley has to first ASSUME that this makes it invalid. It’s a straightforward petition principii.
In any case, Moseley is wrong to allege that TSSI prices of production change “even if all the conditions of the production of value and surplus-value remain the same.” This claim is based on his strange interpretation of examples in which inputs initially exchange at their values while subsequent input prices and output prices are prices of production. He thinks that this implies that “prices of production change every period, not because of changes in the conditions of production of value and surplus-value, but because input prices ≠ output prices.”
It does not. The cause of Moseley’s error here is his lack of curiosity about why the inputs initially exchange at their values.
Consider what would happen if the two sectors in the example had identical compositions of capital, etc., in the period immediately preceding the example, but then sector 1 adopts labor-saving technical change in the first period of the example. Then, if profit rates are equalized before and after, the commodities will exchange at their values at the start of the example but at prices of production that differ from values at the end of the first period of the example (and thereafter).
Moseley’s claim that TSSI prices of production change “even if all the conditions of the production of value and surplus-value remain the same” may also be based on the belief that causes cannot lead to persistent effects. That belief is false, as I discussed in a previous comment. Even if a roof collapses well after the rain has stopped, the rainstorm may indeed be the cause of the collapse. And even if prices continue to change well after a change in the conditions of the production of value and surplus-value, the changed conditions of production may indeed be the cause of price changes.
Reply to Kliman May 4
Convergence of the iterative method
The key point in my argument about the convergence property of Kliman’s iterative method (that “temporally determined” prices of production converge to static equilibrium prices of production when the productivity increases in individual industries stop) (April 3, April 14, April 16 11:25am, May 1) is to distinguish between the price of good A as input [P(A)] and the price of other goods that are produced with Good A is input [P(X)] (see below for a repeat of my argument).
Kliman’s first comment on my argument (April 10) was in terms of a one-commodity model, which is not an appropriate response to my argument because one cannot distinguish between good A as input and goods X as outputs.
Kliman’s second comment on my argument (April 14 5:07 pm)) was in terms of a two-interlocked-commodities model (both goods are inputs to itself and to the other good), which is also not an appropriate response to my argument because again one cannot distinguish between good A as input and goods X as outputs.
Kliman’s third comment on my argument (April 15 6:35 pm) was in terms of a one-SET-of-commodities model, which again is not an appropriate response to my argument because one cannot distinguish between good A as input and goods X as outputs. A one-SET-of-commodities model in which the prices of individual goods are not determined individually and in which the set of commodities is an input (and the only input) to the same set of commodities is essentially the same as a one-commodity model.
Kliman’s fourth comment (May 4) argued that I misinterpreted the subscripts in his April 15 comment as two goods, but they refer to two time periods. I did not misinterpret his subscripts as time periods, but I did misinterpret his one-set-of-commodities model as a one-commodity model. But as noted in the previous paragraph, his one-set-of-commodities model is essentially the same as a one-commodity model.
So Kliman still has not engaged my argument (about the convergence property of his iterative interpretation) in the terms of my argument. Please do! Please respond to the latest version on May 1, which I repeat below:
REPEAT FROM MAY 1
In the following, I will use the following abbreviations:
HCPP for prices of production based on historical cost
CCPP for prices of production based on current cost
(parenthetical paragraph deleted)
1. A productivity increase in an industry will lower the price of production of that good (call it Good A and its price p(A)) IN THAT PERIOD. If good A is a final good (i.e. not an input to other goods), then the reduction of p(A) will have no effect on the price of all other goods. And that is the end of the story. Since there was no productivity increase in the production of the inputs of Good A, the HC of its inputs = their CC and thus HCPP(A) = CCPP(A). Market prices will fluctuate around the new PP according to the accidental causes of S and D.
2. If Good A is an input to the production of other goods (call them collectively Good X), then (according to the TSSI) in the first period the HC of the inputs to Good X will be > the CC of the inputs to Good X and HCPP (X) will be > CCPP (X).
3. However, *if there is no further increase of productivity* in the production of Good A in the next period, then (according to the TSSI) the HC of the inputs to Good X will be = to the CC of the inputs to Good X and HCPP will be = to CCPP. Thus the HCPP of Good X converges to its CCPP after one period.
4. If productivity increases again in the production of Good A in the second period, then the second period would be a repetition of the first period, with both HCPP(X) and CCPP(X) decreasing further and continuing to be ≠.
However, as soon as the increase of productivity in the production of Good A stops for a period, then (according to the TSSI) the HC of the inputs to Good X will be = to the CC of the inputs to Good X and HCPP(X) will be = to CCPP(X). The convergence is within one period.
5. While the productivity increases are happening, market prices MIGHT fluctuate around the HCPP if the perception of capitalists is slow to catch up to the new reality, but the HCPP itself converges to the CCHP as soon as there is no productivity increase, so that at that point market prices are fluctuating around CCPP, and the CCPP is the true long-run center of gravity prices, which will remain in effect until there is a new productivity increase.
Although I haven’t engaged your argument “in the terms of [your] argument,” I’ve proved that it is false nonetheless.
My proofs pertain to any set of prices and thus to any number of goods (1, 2, … 14,389, …). And the proofs hold true for any input-output relations or lack of input-output relations among the goods.
So whatever “model” you dream up, the proofs apply inter alia to it.
Your step 1 is obviously false. Assume that A1 = A2 = L1 = L2 = 2, that X1 = X2 = 4, and that the money wage per unit of L = 0.5. (A1 and A2 are quantities of Good 1 used to produce Goods 1 and 2, respectively.) Note that Good 2 isn’t an input into anything. Also assume that prices of production are determined in accordance with Marx’s value theory as understood by the TSSI.
Now if the initial input prices are p1 = p2 =1, and rates of profit are equalized, then the temporally determined output prices are also p1 = p2 =1.
Imagine that L2 falls from 2 to 0, and all else remains equal. In this next period, the input prices are of course p1 = p2 =1. If rates of profit are equalized once again, then the temporally determined output prices are p1 = 0.9, p2 = 0.6.
So it is false to claim that “If good A is a final good (i.e. not an input to other goods), then the reduction of p(A) will have no effect on the price of all other goods. And that is the end of the story.” In this example, Good 2 was not an input into other goods, but the reduction of p2 DID have an effect on the price of Good 1 (because it reduced the general rate of profit that in turn caused p1 to fall). And that is the end of the story.
I have a VERY hard time understanding the reasoning that has led you to the other false conclusions, but I suspect that the error you’ve made here may be contributing to the other errors.
Reply to Kliman May 7
convergence property of Kliman’s iterative method
But again Kliman’s example in this comment of an increase of productivity in a final good sector is a two-commodity example in which the increase of productivity in one good has a very significant effect on the general rate of profit (declining from .33 to . 2). In reality, this is not the case; the effect of an increase of productivity in one industry on the general rate of profit is usually negligible. The way an increase of productivity in one industry has a significant effect on the prices of other goods is as an input to the other goods, and that was the main point of my argument.
The main point of my argument is this: According to the TSSI, if there is no further increase of productivity in the second period in the industry that produces the input (and everything else remains constant), then the price of the input at the end of period 2 will be equal to the price of that input at the beginning of period 2. [Right?] From this “catching up” of the historical cost of the input to its current cost in the second period, it follows that the historical cost prices of production of the goods produced with this input will be equal to their current cost prices of production. That is, the historical cost prices of production of the goods produced with this input CONVERGE to their current cost price of production in the second period or in any period in which there is no productivity increase.
True or false?
Reply to Kliman May 5, 3:39 pm.
Three points:
1. TSSI prices of production must change in order to equalize the rate of profit
In my comment of April 16, I agreed that a change in the relative proportions of capitals with unequal compositions of capital can change prices of production (as industry total) and also can change unit prices of production.
But I argued in that comment that this additional factor does not support the TSSI interpretation of prices of production that change every period even if productivity and wages – and the scale of production – and the distribution of capital across industries – all remain constant. All these causes of changes in prices of production are *conditions of production*, and affect the *production of value and surplus-value*. Kliman’s prices of production change every period, not because of changes in the conditions of production of value and surplus-value, but because input prices ≠ output prices, and thus prices of production must change every period in order to equalize the rate of profit.
In Kliman’s latest comment, he quoted this last sentence, but he left off that last part of the sentence which explains why his prices of production MUST CHANGE when input prices ≠ output prices – in order to equalize the rate of profit.
That is, even if all the conditions of production remain invariant, if input prices ≠ output prices, the TSSI prices of production will change – and not only will change, but MUST CHANGE – in order to equalize the rate of profit. If the TSSI prices of production don’t change, then the rates of profit in the next period would not be equal. I explained this necessity of the TSSI prices of production to change, even if all the conditions of production remain the same in Chapter 9 of my book, with reference to Kliman-McGlone 1988, as follows:
“The reason why the TSSI prices of production continue to change from period to period is because input prices ≠ output prices and the on-going equalization of the rate of profit and the transformation process itself. The prices of production of the output at the end of the first period become the prices of the inputs at the beginning of the second period. If prices of production of the outputs were to remain constant, i.e. if the prices of production of the output of the second period were equal to the prices of production of the output of the first period, then rates of profit in the second period would not be equal across industries. Therefore, in order to equalise the rate of profit in the second period, the prices of production of the output of the second period MUST CHANGE and must be different from the prices of production of the output of the first period.
“Similar logic applies to future periods, until the prices of production of the output eventually converge to long-run equilibrium prices and input prices = output prices. In each period, the prices of the inputs are not equal to prices of the output, which implies that the inputs prices in the next period will be different from input prices of the current period. If the output prices in the next period were to remain the same as in the current period, while the input prices changed, then the rates of profit across industries would not be equal. In order to equalise the profit rate, the ‘prices of production’ of the outputs of the next period MUST CONTINUE TO CHANGE. Therefore, the TSSI prices of production continue to change from period to period as a result of input prices ≠ output prices and the ongoing equalization of the rate of profit and the transformation of output prices into prices of prices of production, even though it is assumed that the productivity of labour and the real wage remain constant.” (p. 298; emphasis added)
2. Why start with values?
Then Kliman criticized me for “lack of curiosity” about why inputs initially exchange at values. To explain this, he assumes another two-sector example (again!) and the two sectors are assumed to have *EQUAL compositions of capital* in period 1, and that is the reason why the inputs initially exchange at values.
Kliman also assumes *labor-saving technological change* in Sector 1, so that, if rates of profit are to be equalized, outputs must exchange at their prices of production.
But this is not the way that Marx theorized prices of production; Marx assumed *UNEQUAL compositions of capital and *no technological change*. And these have been the assumptions of the entire century-long debate over the transformation, including in the Kliman-McGlone articles. In their 1988 article, they wrote:
“(Solely in order to facilitate comparison with ‘transformation problems’ ‘solutions’, we begin without any ‘errors in the past’; i.e. initial values are equal to the values of means of productin and labour-power.)” (p. 72)
This procedure is of course acceptable, but it has nothing to do with EQUAL compositions of capital and in fact UNEQUAL compositions of capital are assumed their first period.
3. “persistent effects”
Finally, Kliman argues that I fail to take into account “persistent effects” – which presumably means effects that last for more than one period (e.g. the roof and the rainstorm!). Kliman argued:
“Even if prices continue to change well after a change in the conditions of production of value and surplus-value, the changed conditions of production may indeed be the cause of price changes.”
But this is not the case in the Kliman-McGlone papers. As already mentioned, technology is assumed constant throughout all periods in these papers. There can be no “persistent effects” of technological change if there is no technological change. And yet the TSSI prices of production change every period until they converge to the static equilibrium prices of production.
P.S. I am glad that Kliman did not try to defend his earlier ludicrous interpretation of “productivity” – which is a purely physical use-value concept in Marx’s theory – but is defined by Kliman to include exchange-values and prices. And thus a “change of productivity” could also mean a change in the prices of the inputs, even though physical productivity remains the same!
Reply to Fred Moseley’s comment of Mon, 8th May, 2017, 8:21 am.
He wrote, “But again Kliman’s example in this comment of an increase of productivity in a final good sector is a two-commodity example in which the increase of productivity in one good has a very significant effect on the general rate of profit (declining from .33 to . 2). In reality, this is not the case ….”
That doesn’t matter. It’s a completely irrelevant and bogus point, introduced simply in order to try to prove to me once again that Marxian economics means never having to say you’re sorry.
It’s completely irrelevant because even a single counter-example is sufficient to disprove, once at for all–no ifs, ands, or buts–a proposition that allegedly holds true universally, like Moseley’s proposition that “If good A is a final good (i.e. not an input to other goods), then the reduction of p(A) will have no effect on the price of all other goods. And that is the end of the story.”
Moseley knows very well that a single counter-example is sufficient, because Alan Freeman and I pointed it out to him at length in our 2009 paper, “Moseley and Rieu’s ‘Useless and Erroneous Activity,” http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.605.7033&rep=rep1&type=pdf
The paper’s title comes from the following statement by the late Finnish mathematician Perrti Lounesto:
So, since Moseley knows very well that a single counter-example is sufficient to disprove his proposition, and he knows very well that I have provided such a counter-example, why doesn’t he just admit that his proposition is false? The answer, of course, is that he is dead set on drumming into me the lesson that Marxian economics means never having to say your sorry.
The thing is, I’m an opponent of the “Marxian economists” and the rest of the post-truth, “alternative facts” world that has been brought to political power with the election of Trump. So I prefer to die rather than learn the lesson Moseley wants to teach me. Until he admits that my counter-example has disproven his proposition–once at for all, no ifs, ands, or buts–my responses to him will consist entirely of requests that he admit this. The future of truth is at stake.
I join in Kliman’s request that Moseley admit he is wrong, simply wrong. You don’t have to be a sophisticated economist or logician to know that one counter-example disproves a proposition.
This “debate” (in a genuine debate, Moseley would engage the issues honestly) has been going on far too long–because Moseley keeps throwing up red herrings, changing the subject, misrepresenting what has been conceded by his own prior statements, etc. To what end is he conducting this endurance contest? Apparently, just so he can keep saying that the TSSI has not been proved. Why doesn’t he insist that global warming has not been proved while he’s at it?
At a time when ALL truth is in jeopardy, I call for intellectual honesty by leftist scholars.
Reply to Kliman (May 8)
I didn’t receive an email notice about Kliman’s comment and I just discovered it this morning.
My argument (May 1 and 8) was not a mathematical theorem. My argument was a statement about reality – that the effect of an increase of productivity in one industry on the general rate of profit is “usually negligible”. One counter-example does not disprove “usually”. Kliman quoted part of a sentence of mine: “In reality, this is not the case …”. But he left off the rest of the sentence! “… the effect of an increase of productivity in one industry on the general rate of profit is USUALLY negligible”. (who’s being dishonest?)
Besides, Kliman’s counter-example is completely unrealistic and thus is not even evidence against my statement about reality, to say nothing about “disprove”.
But, as I said before, this sentence was in a paragraph about final goods, which was not the main point of my argument. The main point of my argument had to do an increase of productivity in an industry that produces INPUTS used in the production of other goods. I repeat that argument here and ask again for a response.
REPEAT OF ARGUMENT
In the following, I will use the following abbreviations:
HCPP for prices of production based on historical cost
CCPP for prices of production based on current cost
1. A productivity increase in an industry will lower the price of production of that good (call it Good A and its price p(A)) IN THAT PERIOD.
2. If Good A is an input to the production of other goods (call them collectively Good X), then (according to the TSSI) in the first period the HC of the inputs to Good X will be > the CC of the inputs to Good X and HCPP (X) will be > CCPP (X).
3. However, *if there is no further increase of productivity* in the next period in the production of Good A, then (according to the TSSI) the price of production of Good A will remain the same, and the HC of the inputs to Good X will be = to the CC of the inputs to Good X, and HCPP will be = to CCPP. Thus the HCPP of Good X CONVERGES to its CCPP after one period.
4. If productivity increases again in the production of Good A in the second period, then the second period would be a repetition of the first period, with both HCPP(X) and CCPP(X) decreasing further and continuing to be ≠.
However, as soon as the increase of productivity in the production of Good A stops for a period, then (according to the TSSI) the price of production of Good A will remain the same, and the HC of the inputs to Good X will be = to the CC of the inputs to Good X, and HCPP(X) will be = to CCPP(X). The convergence is within one period.
5. While the productivity increases are happening, market prices MIGHT fluctuate around the HCPP if the perception of capitalists is slow to catch up to the new reality, but the HCPP itself converges to the CCHP as soon as there is no productivity increase, so that from that point forward market prices are fluctuating around CCPP, and the CCPP is the true long-run center of gravity prices, which will remain in effect until there is a new productivity increase.
Andrew, true or false?
In his latest act of dishonesty, Moseley (Sat, 13th May 2017 10:38 am) provides us with what he falsely claims is a repeat of his argument (“REPEAT OF ARGUMENT”). It is not.
Moseley has doctored his argument so as to eliminate the proposition that my counter-example disproved: “If good A is a final good (i.e. not an input to other goods), then the reduction of p(A) will have no effect on the price of all other goods. And that is the end of the story.”
The following PDF, which puts the May 7 and May 13 versions of his argument side by side, is proof positive that Moseley has doctored his argument:
http://marxisthumanistinitiative.org/wp-content/uploads/2017/05/Moseley-Rewrites-History.pdf.
The proposition in the May 7 version that I disproved by means of a counterexample is GONE!
The purpose of this deplorable rewriting of history is to try to drum into me the lesson that Marxian economics means never having to say you’re sorry.
The thing is, I’m an opponent of the “Marxian economists” and the rest of the post-truth, “alternative facts” world that has been brought to political power with the election of Trump. So I prefer to die rather than learn the lesson Moseley wants to teach me. Until he admits that my counter-example has disproven his proposition–once at for all, no ifs, ands, or buts–my responses to him will consist entirely of requests that he admit this. The future of truth is at stake.
Yes, I over-stated my argument about final goods in my May 1 comment. And I revised it in my May 8 comment (“usually negligible”), which is the one Kliman quoted in his May 8 comment, which left off the last half of the key sentence and to which I was responding in my May 14 comment. In the paragraph before repeating my argument about inputs in my May 14 comment, I stated:
“But, as I said before, this sentence was in a paragraph about final goods, which was not the main point of my argument. The main point of my argument had to do an increase of productivity in an industry that produces INPUTS used in the production of other goods. I repeat that argument here and ask again for a response.”
The phrase that argument” in the last sentence refers to the argument about inputs, not final goods.
And it is still true that Kliman’s two-commodity example about final goods was not even a counter-example to my original argument which was about reality because his example was so unrealistic. But for the sake of discussion, I retract the original argument about final goods (which I have already revised).
So can we discuss the case of an increase of productivity in the production of INPUTS to other goods, which was the main point of my recent comments?
I appreciate that Fred Moseley (Mon, 15th May 2017 8:06 am) has finally conceded that my counterexample (of Sun, 7th May 2017 10:27 am) disproved his claim that “If good A is a final good (i.e. not an input to other goods), then the reduction of p(A) will have no effect on the price of all other goods. And that is the end of the story.”
But his “interpretation” of the following comment of his is simply false:
Moseley “interprets” that comment of his as follows:
That is just plain false. The original comment clearly distinguished between the “main point of my argument” and “that argument” as such, and Moseley claimed that what he was “repeat[ing]” was “that argument,” not the “main point” of the argument.
Now we can move on.
This is Moseley’s revised argument:
The following example disproves point 3, the 2d paragraph of point 4, and point 5: http://marxisthumanistinitiative.org/wp-content/uploads/2017/05/disproof-of-additional-claims.xlsx
I know what I meant by “that argument” and I think it is clear from the paragraph that Kliman quoted. That paragraph was clearly about moving on from the final good case in the previous paragraph to the input case. The argument that I repeatedly asked for a response to, including in this paragraph, was the input case.
I hope that “moving on” includes responding to my argument about the convergence property of the iterative method.
5/12 further reply to Kliman May 1 and other recent comments
Kliman presents the disagreement between us in our recent discussion as the following:
(1) he analyzes the actual capitalist economy with technological change
(2) I analyze the capitalist economy without technological change
(1) is not true. Yes, he assumes technological change in his model (to an exaggerated degree), but he also assumes *only two interlocked commodities* in this “actual capitalist economy”, so it is far from “actual”.
(2) is both true and not true. It is true in the sense that I argue that Marx’s theory of prices of production, as presented in Part 2 of Volume 3 (which is what my book is about), *assumes a given technology*. I don’t think there can be any doubt about that. No technological change takes place in his period of analysis. Marx added that the main cause of changes in prices of production is technological change, but the basic theory of prices of production assumes a given technology.
And century-long debate over Marx’s theory of prices of production, from Bortkiewicz on, has assumed given and constant technology, including the Kliman-McGlone papers (1988 and 1996). My book is a response to the Bortkiewicz critique. I argue that, with productivity assumed to be constant, Marx did not “fail to transform the inputs”, because the inputs are not supposed to be transformed! The inputs – constant capital and variable capital – are the SAME – and are supposed to be the same – in both the macro theory of the total surplus-value in Volume 1 and in the micro theory of prices of production in Volume 3 – the quantities of money capital advanced to purchase means of production and labor-power at the beginning of the circuit of money capital (M – C … P …), which are taken as given at both levels of abstraction.
However, (2) is not true in our recent discussion of Kliman’s two-commodity Excel model. For the sake of this discussion, I consider Kliman’s temporally determined prices of production (TD PP), including with technological change (increasing productivity), but I argue that it is not realistic to assume technological change occurs in every industry in every period. I am not assuming “no technological change”, but rather that technological change takes place sporadically in individual industries and there are many periods in which there is no technological change in an industry and in industries that produce its inputs.
And I argue that in these periods of constant technology, then the TD PP will gravitate in his model toward the static equilibrium price of production (SE PP), and thus the SE PP is the true long-run center of gravity price in his model, around which both market prices and the TD PP fluctuate. This convergence is a property – and a well-known property – of the iterative method of solving a system of equations. This convergence may be interrupted by more periods of technological change, but the convergence will resume as soon as technology settles down.
Andrew (I ask again): do you agree or disagree with this gravitation of TD PP to SE PP in periods of no technological change?
Unrealistic assumptions
I have also criticized Kliman’s two-commodity model for its extremely unrealistic assumptions: only two goods, and thus productivity changes in one sector has a significant effect on the general rate of profit, each good an input to itself and to the other good, the quantities of the two goods are locked into a reciprocal relationship with each other, and the market prices of these two goods are determined in an ad hoc and inter-related way.
Kliman asserts that these assumptions are “for convenience only” and do not affect the essential points.
But this is not true. These unrealistic assumptions have the effect that of magnifying and distorting the effects of technological change in one industry on the general rate of profit and the PP of the other commodity and prolonging the time required for the TD PP to converge to SE PP; for example in my revision to Kliman’s model to hold productivity constant in both sectors for 8 periods, it took 4 periods for the TD PP to converge to the SE PP because of these unrealistic assumptions.
Kliman has complained that I have the “persistent and annoying habit of questioning [his] assumptions”. I would say that Kliman has the annoying habit of presenting his arguments in terms of extremely simplified one or two commodity models that have little or no relation to reality, so I criticize the very unrealistic assumptions on which his simple models are based and which make the conclusions of these simple models inapplicable to the real capitalist economy.
Fixed capital
Another unrealistic assumption of Kliman’s original two commodity model is that there is *no fixed capital* (i.e. no machines), even though productivity is assumed to increase 4% every year. Kliman responded to this criticism with another two sector model in which one good is fixed constant capital (e.g. a machine) and the other good is circulating constant capital.
However, this model has the same defects of very unrealistic assumptions as the circulating capital only model: only two goods, and thus productivity changes in each sector has a significant effect on the rate of profit, each good an input to itself and to the other good, the quantities of the two goods are locked into a reciprocal relationship with each other, and the market prices of these two goods are determined in an ad hoc and inter-related way.
Increases in productivity (↓L) are assumed in this model to occur every fourth period in each sector, staggered so that there is an increase of productivity occurs for both goods in every other period (because each good is an input to the other good) and no increase of productivity occurs in the alternating periods. This frequency and alternating pattern of productivity increases, along with the unrealistic assumptions listed above, delays and distorts the underlying tendency of TD PP to converge to SE PP.
In addition, the fixed capital model assumes that technological change occurs with the *same type of fixed capital good* (i.e. the same type of machine). But, as I have argued before, an increase of productivity usually occurs with a *change to a new type of machine*, which is not included in Kliman’s model and I don’t see how it could be included because the quantities of different types of machines cannot be compared.
In addition, productivity increases occur with an increase in the ratio F/L (fixed capital good to labor). But with many types of machines, such a continual increase in the ratio F/L is not possible because there is a relatively fixed relation between the machine and the quantity of labor required to operate it (i.e. “fixed coefficients of production”).
In addition, the model assumes that the fixed capital good lasts forever, so there is no depreciation component of the prices of the commodities. And more importantly the original historical cost of the machine purchased in previous periods stays in the denominator of the rate of profit forever, even though the machine gets cheaper and cheaper as time goes on. This causes the rate of profit to continue to decline indefinitely, in contrast to the circulating capital model in which the rate of profit levels off after a few periods.
Therefore, Kliman’s fixed capital model is certainly not a realistic way to analyze the effect of technological change (or anything else).
A very different model
Instead of responding to these criticisms of his two commodity models (“never having to say you’re sorry”?), Kliman presented a very different model, with the following features:
1. instead of 2 sectors, there are 10,000 industries!
2. instead of 50 periods, there is only one period.
3. instead of different prices of INPUTS (historical cost and current cost), the prices of the inputs are assumed to be the SAME in both TD PP and SE PP, and the output prices are assumed to be DIFFERENT because of an increase of productivity during the period. The TD PP (output) is assumed to be equal to the CURRENT (i.e. lower) price of the output and the SE PP (output) is assumed to be equal to the HISTORICAL (i.e. unchanged) price of the output. This is the opposite of the usual TSSI models in which the TSSI is associated with the historical cost of inputs and the SE interpretation is associated with the current cost of the inputs. As a result, the TD PP is in the two commodity models.
I don’t understand the rationale of why the SE PP of the output is assumed to = historical costs and not fall to the current price, since the SE interpretation assumes that the input prices = current prices. It doesn’t make sense. Please explain.
4. since there is only one period, prices of production are not determined by the iterative method, contrary to all of Kliman’s previous work.
5. since there is only one period, the model does not include fluctuations of market prices from one period to the next.
Kliman argues that “one can easily generate additional periods by pressing ‘F9’ and thereby generating a whole new set of capital advances.” But in this case, the second period would have NO CONNECTION whatsoever with the first period. The second period would not start with the output prices of the first period, but would instead start with another randomly distributed set of capital advances. And the quantities of capital advanced and output produced in the second period would not adjust to differences in the individual industries rates of profit in the first period.
I have not yet had enough time to fully digest this completely different and idiosyncratic model, so I’ll stop there for now, except for two questions of clarification for Andrew:
1. You say that the temporally determined actual rates of profit (rj)are normally distributed. Around a mean of what? The temporally determined average rate of profit?
2. What is the meaning of njCj in your equation for Pj?
nj is a normally distributed random number around a mean of 1 and Cj is the capital advanced which is randomly generated with a uniform distribution. So what is njCj?
Thanks.
I should also mention that soon I am going to have to take a break from this discussion for the summer (probably time for a break anyway). I will be in China in June and July and will teach a course at Renmin U. in Beijing, and I have a busy schedule in August (Rio) and September (London and Berlin).
A reply to Fred Moseley’s comment of Tue, 16th May 2017 8:06 am.
He writes, “The argument that I repeatedly asked for a response to, including in this paragraph, was the input case.”
Not so. What he wrote earlier was this: “The main point of my argument had to do an increase of productivity in an industry that produces INPUTS used in the production of other goods.”
Thus, the “input case” was not the argument that he “repeatedly asked for a response to.” It was the alleged “main point” of that argument. But he falsely claimed to “REPEAT” the argument, when in fact he doctored the argument in order to remove the “non-main” point that my counter-example disproved.
He also writes, “I hope that ‘moving on’ includes responding to my argument about the convergence property of the iterative method.” I’m not sure what part of my comment
“The following example disproves point 3, the 2d paragraph of point 4, and point 5: http://marxisthumanistinitiative.org/wp-content/uploads/2017/05/disproof-of-additional-claims.xlsx”
he doesn’t understand.
The points disproven in that example (posted here on Mon, 15th May 2017 9:58 am) are precisely points made by Moseley “about the convergence property of the iterative method”:
“the HCPP of Good X CONVERGES to its CCPP after one period” (point 3);
“as soon as the increase of productivity in the production of Good A stops for a period, then (according to the TSSI) the price of production of Good A will remain the same, and the HC of the inputs to Good X will be = to the CC of the inputs to Good X, and HCPP(X) will be = to CCPP(X). The convergence is within one period” (2d paragraph of point 4);
“the HCPP itself converges to the CCHP as soon as there is no productivity increase, so that from that point forward market prices are fluctuating around CCPP” (point 5).
A reply to part of Fred Moseley’s comment of Tue, 16th May 2017 3:08 pm.
He writes, “Instead of responding to these criticisms of his two commodity models (“never having to say you’re sorry”?), Kliman presented a very different model ….”
First, as I have repeatedly emphasized, these are not models, but examples:
Second, contrary to Moseley’s claim that I did not “respond [ ] to these criticisms of his two commodity models [sic] (‘never having to say you’re sorry’?), I did indeed respond, quite directly:
A further response to Fred Moseley’s comment of Tue, 16th May 2017 3:08 pm.
He complains that “Kliman has the annoying habit of presenting his arguments in terms of extremely simplified one or two commodity models [sic] that have little or no relation to reality, so I criticize the very unrealistic assumptions on which his simple models [sic] are based and which make the conclusions of these simple models [sic] inapplicable to the real capitalist economy.”
As I have repeatedly emphasized here, there are no models here, just arguments that do not depend in any way on any particular modeling assumptions, and examples that illustrate the arguments.
Moseley picks at incidental features of the examples in order to avoid dealing with the logic of the arguments. Note that his comment goes into great detail about my May 3 spreadsheet example, which he mischaracterizes as a model, but totally avoids the argument that preceded it, and which is completely independent of any modeling assumptions.
Here is the argument once again. If Moseley objects to any feature of the example–and even if he doesn’t–he should engage with the actual argument and its logic:
AFTER putting this argument forward, I then made clear that the spreadsheet example is only an illustration of the argument, not the argument itself:
Reply to Kliman (May 16, 6:20 pm)
Kliman’s two commodity example does not disprove my argument about the convergence tendency of the iterative method. My argument had to do with a commodity A which is an input to the production of other commodities X. A is not an input to itself and the X goods are not inputs to A nor to themselves. In this case, the TD PPs of X converge to the SE PPs in one period in which there is no technological change. Kliman has not shown a flaw in this argument.
In Kliman’s two commodity example, on the other hand, each good is an input to itself and to the other good and each good has a significant effect on the general rate of profit. As I have explained in previous comments, these unrealistic interconnections slow down the convergence of TD PP to SE PP, as price changes reverberate back and forth between the two interlocked goods. But these interconnections do not eliminate the convergence. In Kliman’s example, the convergence starts immediately in period 2 and the TD PPs are within 1% of the SE PPs by period 7.
But my example is much more realistic and shows that convergence generally happens faster than that.
In his comment of Fri, 19th May 2017 8:58 am, Fred Moseley wrote,
Moseley is either lying or delusional (or both).
He put forward his “argument about the convergence tendency of the iterative method” in a comment of Mon, 1st May 2017 8:19 am. He repeated it in a comment of Sun, 7th May 2017 9:00 am. He then produced a revised version of the argument in a comment of Sat, 13th May 2017 10:38 am. None of these versions of his argument impose the ridiculous restrictions that “commodity A … is not an input to itself and the X goods are not inputs to A nor to themselves.”
My spreadsheet example (http://marxisthumanistinitiative.org/wp-content/uploads/2017/05/disproof-of-additional-claims.xlsx) does indeed disprove the “argument about the convergence tendency of the iterative method” contained in Moseley’s comments (of Mon, 1st May 2017 8:19 am; Sun, 7th May 2017 9:00 am; and Sat, 13th May 2017 10:38 am), specifically point 3, the 2d paragraph of point 4, and point 5 of that argument.
I thus call on Moseley to retract his false claim that his “argument had to do with a commodity A which … is not an input to itself and the X goods are not inputs to A nor to themselves.” And I call on him to retract his false claim that my spreadsheet example “does not disprove” his “argument about the convergence tendency of the iterative method” contained in the above-mentioned comments.
For what it’s worth, the following spreadsheet example shows that the claims contained in point 3, the 2d paragraph of point 4, and point 5 of Moseley’s argument would still have been false if he had imposed the ridiculous restrictions that “commodity A … is not an input to itself and the X goods are not inputs to A nor to themselves”:
http://marxisthumanistinitiative.org/wp-content/uploads/2017/05/Nope-youre-still-wrong.-Time-to-give-up..xlsx
Once again, the claims contained in point 3, the 2d paragraph of point 4, and point 5 do not hold true: convergence of temporally-determined prices to static-equilibrium prices is not complete after one period.
Note that the example has to assume that Department I’s product (“commodity A”) is produced without any means of production (i.e., by living labor alone), since that is what Moseley’s allegedly “realistic” restrictions (“commodity A … is not an input to itself and the X goods are not inputs to A”) stipulate. Also note that, in order to prevent Moseley from dodging the fact that the counterexample disproves his claims, by complaining that purchases of Department II’s product (the “X goods”) by workers in Department I turn the X goods into inputs to A, the example does not assume that workers purchase Department II’s product.
Reply to Kliman May 20
Kliman started his comment with the usual insult, which I don’t appreciate:
“Moseley is either lying or delusional (or both).”
Andrew, could we please just stick to the issues without the ad hominem?
My assumptions (A is not an input to itself and the goods X that A is an input for are not themselves inputs to A) are the most common cases (not “ridiculous restrictions”) and I didn’t think to state them explicitly. But that is what I had in mind and this argument is not disproven by your first two-commodity example, in which both goods are inputs to themselves and to the other good.
Your second example satisfies those conditions, and it takes 3 periods for the TD PP to converge within 1% of the SE PP. The reason it takes a little longer to converge is that (because there are only two goods) the cheaper constant capital for Good 2 increases the rate of profit for Sector 2 and thus increases the general rate of profit 15% (from 90% to 114%). In terms of my argument, Good 2 is actually a group of goods that Good 1 is an input for, and if this group is large enough, then an increase in its rate of profit could have a significant effect on the general rate of profit and thus could delay the convergence somewhat. But this would be an exception. In most cases, the group of goods X is small and its higher rate of profit would not have a significant effect on the general rate of profit and convergence would happen quicker. But the main point of my argument is not the one period, but the convergence property of the iterative method, which is illustrated by all of Kliman’s examples.
REPLY TO KLIMAN’S 10,000 INDUSTRIES EXAMPLE
Kliman is trying to prove that temporally determined actual prices (TDPi) fluctuate around the temporally determined price of production (TDPPi), and not around static equilibrium prices of production (SEPPi). However, the argument and this example are in term of ONE PERIOD ONLY, so they don’t prove anything directly about fluctuations of actual prices across multiple periods.
Kliman argues that “one can easily generate additional periods by pressing ‘F9’ and thereby generating a whole new set of capital advances.”
But in this case, the second period would have NO CONNECTION whatsoever with the first period. The second period would not start with the output prices of the first period, but would instead start with another randomly generated set of capital advances. And the quantities of capital advanced and output produced in the second period would not adjust to differences in the individual industries rates of profit in the first period. Thus “pressing F9” does not provide the basis for an explanation of fluctuations of actual prices around long run center of gravity prices in future periods.
Kliman also argues that one can infer that the TDPi’s fluctuate around the TDPPi’s, and not around SEPPi’s, from the fact that in this single period 80% of temporarily determined industry rates of profit (TDRi’s) are below the corresponding static equilibrium rate of profit (SER*) and the TDRi’s are evenly divided between half above the corresponding temporarily determined rate of profit (TDR*) and half below, which implies that 80% of the TDPi’s are below the corresponding SEPPi’s and are evenly divided above and below the corresponding TDPPi’s (because the capital advanced is assumed to be the same for all three sets of prices).
Kliman argues: “None of these results are compatible with the idea that static-equilibrium prices and the static-equilibrium rate of profit are centers of gravity. All of them are compatible with the idea that the actual, temporally-determined prices of production and the general rate of profit are centers of gravity.”
(It should be noted that this relation between the TDPPi’s and the SEPPi’s in this example is opposite to all previous TSSI writings in which the TDPPi’s are > SEPPi’s. The reason for this opposite relation is different assumptions about the prices of inputs and outputs. Instead of different prices of INPUTS (historical cost and current cost), the prices of the inputs are assumed to be the SAME in both TDPP and SEPP, and the output prices are assumed to be DIFFERENT because of an increase of productivity during the period. The TDPP (output) is assumed to be equal to the CURRENT (i.e. lower) price of the output and the SEPP (output) is assumed to be equal to the HISTORICAL (i.e. unchanged) price of the output. This is the opposite of the usual TSSI works in which the TSSI is associated with the historical cost of inputs and the SE interpretation is associated with the current cost of the inputs. These different assumptions have important implications, as we will see below.)
However, Kliman’s results in this example are due to the fact that the TDRi’s are ASSUMED to be normally distributed around TDR*. If the TDRi’s are assumed to be normally distributed around TDR*, then yes indeed half the TDRi’s will be above the TDR* and have will be below, and similarly for TDPi’s and TDPPi’s, by the definition of a normal distribution.
Furthermore, this ad hoc method of determining the TDRi’s and TDR* is entirely different from the method of determining these variables in Kliman’s recent two-commodity examples that we have discussed. In the two-commodity examples, the TDRi’s are DETERMINED FIRST, prior to and independently of TDR* (and the TDRi’s depend primarily on the relation between input prices and output prices in each particular industry), and then the TDR* is derived as a kind of weighted average of the pre-determined TDRi’s (by the relation between input prices and output prices in all industries together). However, in the recent example with 10,000 industries, the logic of determination is essentially the opposite: the TDR* is determined first (by aggregate quantities) and then the TDRi’s are assumed to be random numbers that are normally distributed around the mean of the pre-determined TDR*.
Kliman emphasizes that the weighted average of the TDRi’s is the TDR*:
“The weighted average of the industry-level rates of profit TURNS OUT TO BE EXACTLY EQUAL to the temporally-determined general rate of profit …” (emphasis added)
But this is not true. A weighted average implies that the TDRi’s are determined independently of TDR* and then the TDR* is derived from the pre-determined TDRi’s. But this is not true in Kliman’s 10K industries example.
The phrase “turns out to be exactly equal …” also suggests that the TDRi’s are determined independently of TDR* and their weighted average “turns out to be exactly equal …”. This result is not surprising, since that is what is assumed – the weighted average of the individual values of a normal distribution (with the probabilities of occurrence as the weights) is the mean of a normal distribution.
There is a hint of this different logic of determination in a change of Kliman’s terms for the rate of profit – from “average” rate of profit in the two-commodity examples to the “general” rate of profit in the 10K industries example.
And similarly with prices. Kliman’s result that 80% of the TDPi’s are below the corresponding SEPPi’s and are evenly divided above and below the corresponding TDPPi’s is due to the fact that TDRi’s are assumed to be normally distributed around the TDR* and the capital advanced is assumed to be the same for all the prices.
And again the TDPi’s and TDPPi’s are determined in this example in an entirely different way from the method of determining TDRi’s and TDR* in the two-commodity examples. In the two-commodity examples, the TDPi’s are determined independently of the TDPPi’s (mainly by the labor-times in each particular industry), and then the TDPPi’s are derived by adding the average industry profit to the prices of the inputs. However, in the 10K industries example, the TDPPi’s are determined first and then the TDPi’s are assumed to be evenly distributed around the pre-determined TDPPi’s.
It should also be noted that the prices in this example are NOT UNIT PRICES, but are instead TOTAL PRICES in each industry (i.e. unit price x quantity of output). Therefore, this example does not explain fluctuations of unit prices around long run center of gravity unit prices, which is what I thought the debate was about (this is what the previous two-commodity examples were about).
We can also see that the “actual” industry prices in this example (the TDPi’s) are NOT REALLY ACTUAL PRICES, based on the actual labor-times and conditions of production in each industry, but are instead just random numbers. So this example does not explain how actual industry prices are determined and how actual industry prices fluctuate around long run center of gravity prices, because the TDPi’s are random numbers, with no connection with the random numbers of the next period or the period after that, etc.
Kliman also emphasizes the sum of the TDPi’s “turn out to be exactly equal” to the sum of the TDPPi’s”:
“The results are quite clear … The total economy-wide price TURNS OUT TO BE EXACTLY EQUAL to the total temporally-determined economy-wide price of production …” (emphasis added)
But this aggregate equality is ASSUMED! Kliman states this assumption twice – both before and after the above statement that this aggregate equality is a result.
Kliman first stated:
“The actual industry-level output prices satisfy the condition that (a) total economy wide-ride price equals the total temporally-determined economy-wide price of production …”
This assumption is almost word for word the same as the above “result”, except that “equals” in the assumption becomes “turns out to be exactly equal” in the “result”.
And then 10 lines after stating that this aggregate equality is a result, he states the assumption again:
“Note that these results depend on no assumptions apart from: …
2. The temporally-determined economy-wide total price equals the total temporally-determined, economy-wide total price of production.”
Again the language is almost “exactly equal”.
So these two sums are assumed to be equal and then “it turns out” that these two sums are exactly equal! Kliman emphasizes that this aggregate invariance is what makes his example “work”. But he does not mention that this aggregate invariance is working overtime – both as assumption and as conclusion.
Furthermore, there is another important issue here: Kliman has agreed in the discussion of the two-commodity examples that there are many industries in which there is NO DECLINE OF INPUT PRICES in a given period, and in these industries there is NO DECLINE OF THE RATE OF PROFIT due to this cause. Therefore, in order to analyze the effects of a decrease in the prices of inputs Kliman should distinguish between two types of industries:
(1) industries which use inputs that become cheaper during the period of analysis
(2) industries which use inputs that do not become cheaper during the period of analysis.
In type (2) industries, there would be no difference between the historical cost of inputs and the current cost of inputs, and thus (according to the TSSI) there would be no decline in the TD rate of profit (TDRi) in these industries. However, there would be a decline in the TD general rate of profit (TDR*) because the rate of profit has fallen in type (1) industries. On the other hand, the static equilibrium rate of profit (SER*) would not fall because the inputs in type (1) industries are valued at current cost. Thus, in type (2) industries, the TDRi and the SER* are similar in that neither of these rates of profit falls, but the TDR* does fall and this is different from the other two rates of profit.
These differences are repeated for the prices in type (2) industries. The TDPi and the SEPP don’t change, but the TDPP falls (because the TDR* falls).
These conclusions are very different from Kliman’s 80% results.
In type (1) industries, the historical cost of inputs would be > the current costs of inputs, so (according to the TSSI) the TDRi falls and the TDR* also falls, but the SER* does not fall (because inputs are valued at current costs). On the other hand, with respect to prices, the TDPi and the SEPPi remain the same, but the TDPPi falls (due to the fall in the TDR*).
Furthermore, in the next period, if there is no further decline in the price of inputs in type (1) industries, then (according to the TSSI) the historical cost of these inputs would adjust to their current costs and thus the TDPPi would CONVERGE to the SEPPi, as we have seen in the two-commodity examples. The difference between the historical cost and the current cost of inputs is temporary, and is eliminated as soon as the prices of the inputs in an industry do not change in a period.
Therefore, even though the TDPi’s may have been closer to the TDPPi’s in the first period, that does not imply that the TDPi’s will fluctuate around the TDPPi’s, because the TDPPi’s themselves converge to the corresponding the SEPPi’s in the next period (or as soon as the prices of inputs stay the same).
Also in the next period, if inputs become cheaper in another set of industries, then these same effects would be repeated in the next period, and the period after that, etc.
This CONVERGENCE of the TDPPi’s to the SEPPi’s is obscured in Kliman’s example because, as noted above, he assumes in this model that the reason inputs prices are > output prices is not because input prices are valued in historical costs (an in all previous TSSI work) but instead because TDPPi’s are valued at current prices and SEPPi’s are valued in historical prices! As a result, the adjustment of the TD historical costs of inputs to their current costs disappears and so does the resulting convergence of the TDPPi’s to the SEPPi’s.
Therefore, in this 10K example, it is still true that changes in input prices in individual industries are sporadic and not continuous, and when the input prices in an industry are stable, then the TDPPi’s converge to the corresponding SEPPi’s and the latter are the true long-run center of gravity prices in all industries.
Therefore, I conclude that Kliman has not proved that temporally determined actual prices (TDPi) fluctuate around temporally determined prices of production (TDPPi) and not around static equilibrium prices of production (SEPPi), neither in his two-commodity examples, nor in his 10,000 industries example.
Just like he earlier also failed to prove that my interpretation of Marx’s rate of profit is the same as Sraffa’s rate of profit, and like he failed to prove that my interpretation of Marx’s rate of profit is determined by his “physicalist” quantities, since his “physicalist” quantities are themselves derived from my monetary rate of profit (i.e. his argument is based on circular reasoning).
On the other hand, I have shown that Kliman’s temporally determined prices of production (TDPPi) CONVERGE to static equilibrium prices of production (SEPPi) in any period in which the prices of inputs in a given industry remain constant.
And I have also shown that Kliman’s interpretation of temporally determined prices of production based on the iterative method – which change every period, even if the conditions of production remain the same (until they converge to the static equilibrium prices of production) – contradicts Marx’s prices of production which change only if the conditions of production change.
As I mentioned before, I am leaving for China next week and I will take a break from this discussion until the Fall. In the Fall, I plan to write a summary of our discussion thus far and would be happy to continue the discussion.
My book Money and Totality is out in paperback from Haymarket Press (for $20):
https://www.haymarketbooks.org/books/1025-money-and-totality
I hope that readers of this discussion will be interested enough to read the book and let me know what you think (fmoseley@mtholyoke.edu).
“Money and Totality is probably the best book on Marxist economic theory this year and for this century so far.”
Michael Roberts, author of The Long Depression.
“This book … is surely one of the most important contributions to Marxian scholarship published in our time.”
Tony Smith, Iowa State University
One more comment and a few questions for clarification before I go:
1. In the equation sheet of his spreadsheet, Kliman gives the following equation for individual industry prices:
Pj = (njCj)/∑njCj) (∑Pj)
This equation appears to be circular reasoning, since Pj is on both the LHS and the RHS of the equation. Kliman does not say so explicitly, but we can tell from the Excel cells that circular reasoning is avoided by substituting the total temporally determined industry PRICE OF PRODUCTION (∑PPj) for the total temporally determined industry PRICE (∑Pj) on the RHS of the equation; i.e. by ASSUMING: ∑Pj = ∑PPj. (right?) This is the “aggregate invariance” that Kliman emphasizes:
“The total economy-wide price [∑Pj] TURNS OUT TO BE EXACTLY EQUAL to the total temporally-determined economy-wide price of production [∑PPj] …” (emphasis and brackets added)
This is precisely where the aggregate invariance is ASSUMED – in the determination of each and every Pj. That is why it is illogical to argue that (∑Pj) “turns out to be exactly equal” to (∑PPj), as if that is an independently derived result, because the two are assumed to be equal to begin with.
2. Kliman states below this equation:
“I assume that, due to labor-saving technological change, that the (weighted) average of output prices is 4% less that the (weighted) average of input prices.”
But what exactly is the “weighted average” of output prices (or input prices)? As we have seen, prices in this example are not unit prices, but are instead the total price in each industry (i.e. unit price x quantity of output). So “output prices” must mean total industry prices. But what is the economic significance of the weighted average of the total prices in each industry? And what does this have to do with Marx’s theory?
And what are the weights that are assigned to the total price in each industry? The weights cannot be quantities because they are already taken into account in the determination of the total industry price. The weights appear to be the probabilities in a normal distribution of random numbers. (right?) This is a reminder that the Pj’s in Kliman’s example are NOT ACTUAL total industry prices, but are instead only random numbers.
3. Andrew, please express algebraically the “weighted average of output prices”. Thanks.
I’ve just received a question (via e-mail) about whether I am “planning to continue the discussion with Moseley, as he’s had the last word since June?”
But has he actually “had the last word since June”?
In one sense, yes, since the latest comments here were Moseley’s comments from early June.
But in another sense, Moseley has never had the last word.
The last word on the topic continues to be my completely general argument that static-equilibrium prices and the static-equilibrium rate of profit simply cannot be centers of gravitation of actual market prices and rates of profit when prices tend to trend downward because of technological change.
It’s the last word because it is completely general–quibbles about the “assumptions” of my “model” cannot challenge it in any way, since it is free from restrictive assumptions.
Moseley has still not responded to this completely general argument, which I first put forward (above) on May 3 and have repeated since then.
It goes as follows:
Let me try something simple.
Andrew’s model is a model of ongoing period-per-period change. Such a model does never reach a static equilibrium state. Here, categories like static equilibrium PP do not make sense
and comparing this non-existing SEPP with something else like TDPPi is meaningless.
Fred probably has a model in mind where change only occurs in beginning periods and the following periods have to cope with the input price effect of it. Therefore for these periods its
TDPPi differ from future SEPPi that only exist when static equilibrium state is reached; however the TDPPi converge towards them. In equilibrium state TDPPi equal SEPPi and market prices fluctuate around SEPPi.
Let me try also something simple:
Static equilibrium is a delusion.
“There is a continual movement of growth in productive forces, of
destruction in social relations, of formation in ideas; the only
immutable thing is the abstraction of movement — mors immortalis.” (Marx, The Poverty of Philosophy)
I may have solved one the controversies here; if we go back to Part 2: Kliman’s contested demonstration that Moseley’s interpretation of Marx also has changes in the prices of production (assuming input prices = output prices) despite the level of productivity and wages remaining constant.
Moseley alleges that the change is not relevant to his interpretation because his claim was about the fixedness of the prices of production was only in a “restricted [sic] sense”, not in a “general [sic] sense”. His claim is that for a given amount of capital (total; fixed + variable) and a given level of production, the prices of production on his interpretation do not change.
Let’s Kliman’s demonstration in Part 2, here http://marxisthumanistinitiative.org/wp-content/uploads/2016/05/All-Value-Form-Part-2.pdf
Before the change in size, the prices per unit were 2 for sector 1, and 0.8 for sector 2.
And after, the prices per unit were 2.5 for sector 1, and 1 for sector 2.
If the “scale of production” could explain the difference in the prices of production, then the prices *per unit* (that is, controlling for the scale of production) would remain the same: but they don’t.
Perhaps Moseley’s other stipulation-the amount of (total) capital per 100-can explain the change.
If we note that the total capital in sector 1 before the change is 96+4=100, the total capital in sector 2 before the change is 12+8=20, the total capital in sector 1 after the change is 120+5=125, and the total capital in sector 2 after the change is 30+20=50, then we can control for the amount of capital.
Dividing we get 2/100=0.02 for sector 1 before the change, 0.8/20=0.04 for sector 2 before the change.
And 2.5/125=0.02 for sector 1 before the change and 1/50=0.02 for sector 2 after the change.
Interestingly enough, the prices of production for *sector 1* remain the same after controlling for the amount of fixed capital, but the prices of production for *sector 2* do not.
If Moseley’s interpretation were to be internally consistent, then controlling for the scale of production (per-commodity price) and fixed capital (the division above) would have equalized the prices of production in ***all*** sectors, not just one.
I think this demonstrates that Professor Moseley’s interpretation is untenable, as his own prices of production change despite controls for his proposed explanations.
This is a very interesting comment, Musicotic. And I thank you for raising an issue that I hadn’t thought about.
I don’t understand some of your computations. I think 2/100=0.02, 0.8/20=0.04, etc. are dividing per-unit prices by the total sectoral capital advances. I don’t understand what the results of the division are, nor what they imply.
BUT YOU ARE ABSOLUTELY RIGHT IN ANY CASE: Moseley’s (per-unit) prices of production do change, despite no change in “productivity” or real wages, EVEN FOR A GIVEN AMOUNT OF ADVANCED CAPITAL.
A moment’s thought makes this obvious. When we scale the total sectoral price of production up or down, to put it on the basis of 100 units of capital (or whatever number), we also have to scale the total sectoral physical output up or down to the same degree, to put it, too, on the basis of 100 units of capital (or whatever number). Thus both the numerator (total sectoral price) AND the denominator (total sectoral physical output) of the per-unit price increase or decrease by the same percentage, leaving the per-unit price unchanged. In other words, putting everything on the basis of a capital advance of 100 (or whatever number) makes no difference whatsoever.
That’s the general proof. For the example in question, here are the numbers:
.
RAW DATA
original capital, sector 1: $100
original total price of output, sector 1: $120
original total physical output, sector 1: 60
original capital, sector 2: $20
original total price of output, sector 2: $24
original total physical output, sector 2: 30
new capital, sector 1: $125
new total price of output, sector 1: $150
new total physical output, sector 1: 60
new capital, sector 2: $50
new total price of output, sector 2: $60
new total physical output, sector 2: 60
.
PER $100 OF ADVANCED CAPITAL
original capital, sector 1: $100
original total price of output, sector 1: $120
original total physical output, sector 1: 60
original capital, sector 2: $100
original total price of output, sector 2: $120
original total physical output, sector 2: 150
new capital, sector 1: $100
new total price of output, sector 1: $120
new total physical output, sector 1: 48
new capital, sector 2: $100
new total price of output, sector 2: $120
new total physical output, sector 2: 120
.
Using the latter set of figures, when everything is on the basis of capital advances of 100, and dividing the total prices of outputs by the associated total physical outputs, we get:
.
MOSELEY’S “RESTRICTED-SENSE” PER-UNIT PRICES OF PRODUCTION
original, sector 1: 120/60 = 2
original, sector 2: 120/150 = 0.8
new, sector 1: 120/48 = 2.5
new, sector 2: 120/120 = 1
.
Hence, Moseley’s (per-unit) prices of production do change, despite no change in “productivity” or real wages, EVEN FOR A GIVEN AMOUNT OF ADVANCED CAPITAL.
Distribution of capital across industries with unequal compositions of capital
In Kliman’s two-sector example, the rate of profit and industry prices of production (“total price of output”) and unit prices of production are determined by the Sraffian method of simultaneous equations, not by my interpretation of Marx’s theory, and his results show that both the industry price of production and the unit price of production of both commodities change even though the productivity of labor and the real wage remain the same. (The rate of profit remains the same; more on this important point below; relative unit prices also remain the same.) And Kliman argued that this additional cause of changes in prices of production that was not mentioned by Marx implies there could be other possible causes that were not mentioned by Marx, including the ongoing transformation of values into prices of production in the TSSI interpretation.
According to Marx’s theory, the distribution of capital across industries with unequal compositions of capital affects the rate of profit because it affects the quantity of labor and surplus-labor in relation to the total capital in the economy as a whole and thus indirectly affects prices of production. Marx discussed the effect of the distribution of capital on the rate of profit in Chapter 9, pp. 261-63. But a few pages later, when he discussed changes in the rate of profit that cause changes in prices of production, he did not mention the distribution of capital as a cause of a change in the rate of profit. A change in the distribution of capital is a minor cause of a change in the rate of profit, and a change in the rate of profit affects prices of production slowly and over long periods of time. Therefore, a change in the distribution of capital is a minor indirect cause of changes in prices of production, and one can understand why Marx overlooked this minor cause in his discussion of causes of changes in prices of production. However, one can easily add this very minor cause based on Marx’s discussion a few pages earlier.
On the other hand, according to the TSSI, the ongoing transformation of values into prices of production is not a minor indirect cause of changes in prices of production over long periods of time, but is instead supposed to happen *in every period and in all industries* (until convergence to long-run equilibrium prices). So it is not reasonable that Marx would have overlooked such a universal and continuous cause of changes in prices of production in his repeated discussions of causes of changes in prices of production.
As mentioned, Kliman’s example is based on Sraffian theory, not Marx’ theory, and therefore Kliman’s conclusion about the effect of a change in the distribution of capital on the rate of profit is not the same conclusion as in Marx’s theory. As just discussed, according to Marx’s theory, a change in the distribution of capital across industries with differing compositions of capital *affects the rate of profit* because it affects the quantity of labor and surplus-labor in relation to the total capital in the economy as a whole. In Kliman’s example, Branch 2 is more labor-intensive (i.e. a lower composition of capital). According to Marx’s theory, a higher proportion of the more labor-intensive Branch 2 would *increase the rate of profit* because there is relatively more labor and more surplus labor in the economy as a whole.
According to Sraffian theory, on the other hand, this shift in the proportions of capital and output has *no effect on the rate if profit* because in Sraffian theory the rate of profit is not determined by surplus labor, but is instead determined by the physical input-output coefficients, and those physical coefficients do not change as a result of the changing proportions between industries. (Sraffa used this property of an input-output system of equations to construct his standard system. The standard system is constructed from the actual system by changing the proportions of the actual system to be the same as the standard commodity and the standard system has the same rate of profit as the actual system.) In Kliman’s example, the rate of profit remains 20% after Branch 2 doubles in size.
This is another difference between (my interpretation of) Marx’s theory and Sraffian theory of the rate of profit (in addition to the effect on the rate of profit of technological change, of changes in the turnover time of capital, of luxury goods, of the devaluation of capital). The difference between the two theories in the case of the effect of a change in the distribution of capital on the rate of profit is clear and obvious:
Marx: change in the rate of profit vs. Sraffa: no change in the rate of profit.
I understand this difference more clearly from Kliman’s example.
Stripped to its essentials, Fred Moseley’s latest comment is a crucially important admission that one of his most important claims is just plain false.
He has frequently alleged, over the course of many, many years, that there are TWO and ONLY TWO causes of changes in (per-unit) prices of production, according to Marx’s theory: changes in technology and changes in the real wage rate.
Moseley’s latest comment concedes that this longstanding pillar of his interpretation is false because, as he has belatedly discovered, “the distribution of capital across industries with unequal compositions of capital … affects prices of production. … a change in the distribution of capital is a … cause of changes in prices of production”!
Furthermore, Moseley’s false claim that changes in technology and real wages are the sole causes of changes in Marx’s prices of production has been an exceptionally important basis for his allegation that the TSSI is an incorrect interpretation of Marx’s value theory. He has frequently argued, over the course of many, many years, that because the TSSI implies that there are additional causes of changes in prices of production, it must be considered an incorrect interpretation. But Moseley’s latest comment is tantamount to an admission that this longstanding argument of his against the TSSI has been incorrect—-all along.
Clinging to a weak argument
Kliman is clinging to a very weak argument to support his multi-period interpretation of the transformation problem. He argues that the fact that the distribution of capital has an effect on prices of production that was not mentioned by Marx in his discussions of causes of changes in prices of production opens the possibility that there are other possible causes of changes in prices of production that also were not mentioned by Marx, including the TSSI interpretation that involves changes in prices of production in every industry and in every period.
This logical leap is far-fetched. The distribution of capital is a very minor cause of changes in prices of production. The distribution of capital affects prices of production only indirectly through its effect on the rate of profit, and changes in the distribution of capital is a minor infrequent cause of changes in the rate of profit; and changes in the rate of profit affects prices of production only slowly and over long periods of time. So it is not surprising that Marx overlooked the distribution of capital in his discussions of causes of changes in prices of production.
According to the TSSI, on the other hand, prices of production are supposed to change in every period and in all industries (until convergence to long-run equilibrium prices). The TSSI multi-period transformation of values into prices of production is not a minor indirect infrequent cause of changes in prices of production, but is instead the whole process of the transformation itself, involving changes in every industry in every period (actual historical periods). If Marx intended the transformation process to be a multi-period process, surely he would have said something, and probably a lot, about such a multi-period process. But Marx never said anything about the transformation being a multi-period process.
Why do TSSI prices of production change?
In addition, there is a related serious problem with the TSSI interpretation: *why do TSSI prices of production* change from the beginning of a period to the end of the same period, even though the productivity of labor (and everything else) do not change? What are the causes of the change of TSSI prices of production between the beginning and the end of the same period?
The TSSI assumes that inputs are purchased at their last period prices of production (which are not long-run equilibrium prices). These same input-goods are produced as outputs in the current period. If these newly produced input-goods were sold at the same price at which they were purchased at the beginning of the period, then rates of profit would be unequal. Therefore, *something must happen during the period to cause prices to change* in such a way to equalize the rate of profit. What happens?
The usual assumption about the equalization of the profit rate (from Adam Smith on) is that rates of profit are equalized through a process in which capitals are transferred from industries with lower than average rate of profit to industries of higher than average rate of profit, and these transfers of capital would cause the quantities of output produced in these industries to change, which in term would cause the prices of these goods to change and to gravitate toward long-run equilibrium prices. In summary: transfers of capital → changes of output → changes of prices.
But *none of this happens in the TSSI interpretation of the transformation*, except the change of prices. There are no transfers of capital, and no changes of quantities produced. The quantities of inputs and outputs remain the same in all periods (simple reproduction is assumed). Nothing changes in the TSSI transformation except prices. But then the TSSI provides *no explanation of the causes of the changes in the prices of input-goods* from the beginning of a period to the end of the same period, even though nothing else changes.
Another difference between my interpretation of Marx’s theory and Sraffa’s theory
Finally, Kliman did not respond in his comment to the other main point of my previous comment – that the different effects of a change in the distribution of capital on the rate of profit is another way in which my interpretation of Marx’s theory of the rate of profit is different from the Sraffian theory of the rate of profit, contrary to Kliman’s long-standing argument that my interpretation is the same as Sraffian theory:
Marx: change in the rate of profit
vs. Sraffa: no change in the rate of profit.
Stop making crap up, Fred.
It is simply untrue that “Kliman is clinging to a very weak argument … that the fact that the distribution of capital has an effect on prices of production that was not mentioned by Marx in his discussions of causes of changes in prices of production opens the possibility that there are other possible causes of changes in prices of production that also were not mentioned by Marx, including the TSSI interpretation that involves changes in prices of production in every industry and in every period.”
That’s not my argument. You know well what my actual argument is, and it has NOTHING to do with Marx supposedly forgetting to mention a cause of changes in prices of production! Contrary to you, who charge him with forgetfulness (which is actually just a polite way of charging him with internal inconsistency, as in the “Marx forgot to transform the inputs” garbage), I see NO evidence that he forgot to mention a cause of changes in prices of production. None whatsoever. I think he said what he meant, not forgetting anything (and certainly not something he had written only a few pages earlier!): a commodity’s price of production can change for ONLY TWO reasons; either the general rate of profit changes or the productivity of labor changes.
My actual argument, as you well know–because you quote it in your book!–is this:
Please withdraw the nonsense about the very weak argument that I allegedly cling to, but in fact do not.
Fred, once we set aside your unsubstantiated and baseless dodge about Marx’s supposed “forgetfulness” about what causes changes in prices of production, the actual issue here is what Marx meant by “productivity” in the passage in question. If he meant “technology”––as you have alleged, despite the clear evidence that Marx used ‘change . . . in productivity’ and ‘change in value’ synonymously in this passage!––then you must repudiate your interpretation of Marx’s value theory, because your interpretation (like the TSSI) implies that prices of production can indeed change in the absence of changes in technology (or real wages).
That’s what I’ve shown. And you have conceded that I’m right. And nothing in your response dares to challenges my comment, above, which pointed out that you have indeed conceded this.
Your other option is to admit that “productivity” in this passage isn’t synonymous with “technology,” in which case you must repudiate your charge that the TSSI is an incorrect interpretation because it (like your own interpretation!) implies that prices of production can change in the absence of changes in technology (or real wages).
It’s your choice.
Fred, I have to say that your behavior is a perfect example of why you “Marxist economists” literally make me sick. Everything is always Marx’s fault, never your own fault.
Most of them say that “Marx forgot to transform the inputs.” Now you pile on and say that, when he stated categorically that there are two and only two causes of changes in prices of production, Marx “forgot” that there are other causes! Quite a forgetful guy. “They have nothing to lose but their … um … um … keys? … um, minds? … train of thought?”
You people never allow yourselves–or the rest of us!–to consider even the possibility that the difficulties you face when trying to interpret Marx are weaknesses in your interpretive frameworks rather than defects in his arguments, and that you need to abandon these interpretive frameworks and start over from scratch.
And in the present case, I should point out, your “Marx forgot” gambit is just all too convenient, self-serving, ad hoc, and redolent of “our-line’s-been-changed-again”-ism. When you thought that prices of production in your interpretation do not change except when technology or real wages change, you forcefully asserted that Marx’s statement about two and only two causes of changes in prices of production was definitive, categorical, the last word. And you tried to use this last word as a bludgeon to beat back the TSSI. That went on for decades. No hint from you throughout that time that the statement was perhaps not quite the last word.
But then I demonstrated that prices of production in your interpretation DO change even when technology or real wages DON’T change. So, by no accident whatever, you quickly concocted a new argument about Marx’s statement, diametrically opposed to your original one in a crucial respect: no, that statement wasn’t the last word; Marx temporarily “forgot” about one additional cause of changes in prices of production. And by no accident whatever, the one cause that he “forgot” about just happens to be the thing that causes changes in your prices of production when technology and real wages don’t change. HOW CONVENIENT!
What we need are good-faith efforts to understand Marx’s reasoning in and for itself, not bad-faith efforts to use Marx as rhetorical support for what we ourselves want to say. Unfortunately, the latter is the near-universal practice of the “Marxist economists” and indeed most academic Marxism throughout the social sciences and humanities.
As I said, this literally makes me sick. I can see why some people care more about what they themselves want to say than about getting right what Marx said. What makes me sick is the refusal, despite that aim, to leave Marx alone, to let him speak for himself and have his theory, which differs from one’s own theory … the arrogant compulsion to use Marx as a tool in the service of one’s own theory or approach and a tool one wields to advance one’s own aims.
No responses
I first point out that in Kliman’s latest comment he did not respond to two of the main points of my previous comment:
1. The TSSI interpretation of the transformation problem provides no explanation of what causes prices of production to change from the beginning of a period to the end of the same period, even though nothing else in the economy changes, including productivity and quantities of outputs.
2. The effect of a change in the distribution of capital on the rate of profit is another difference between my interpretation of Marx’s theory and Sraffa’s theory (Marx: no change in the rate of profit vs. Sraffa: change in the rate of profit)
Not clear what Kliman is saying about the effect of the distribution of capital on prices of production
In Kliman’s first comment on Tuesday, he quoted me summarizing his argument:
“Kliman is clinging to a very weak argument … that the fact that the distribution of capital has an effect on prices of production that was not mentioned by Marx in his discussions of causes of changes in prices of production opens the possibility that there are other possible causes of changes in prices of production that also were not mentioned by Marx, including the TSSI interpretation that involves changes in prices of production in every industry and in every period.”
And then he said: “That’s not my argument.”
Andrew, if that is not your argument about the effect of the distribution of capital on prices of production and how that relates to the TSSI interpretation of the transformation problem, then what is your argument about the distribution of capital?
Then Kliman said: “I see NO evidence that he forgot to mention a cause of changes in prices of production. None whatsoever.”
But Marx discussed the effect of the distribution of the distribution of capital on the rate of profit in Chapter 9 of Volume 3, pp. 261-63, and changes in the rate of profit causes changes in prices of production. But a few pages later he did not mention the distribution of capital as a cause of a change in the rate of profit which in turn causes a change in prices of production. Isn’t that evidence that he overlooked this minor indirect cause of changes in prices of production?
Kliman continued:
“I think he [Marx] said what he meant, not forgetting anything (and certainly not something he had written only a few pages earlier!): commodity’s price of production can change for ONLY TWO reasons; either the general rate of profit changes or the productivity of labor changes.”
But Marx also discussed a number of causes of changes in the general rate of profit (e.g. change in the composition of capital or the rate of surplus-value), but he did not mention a change in the distribution of capital.
Kliman’s erroneous definition of the “productivity of labor”
Throughout the three volumes of Capital, Marx consistently defined the “productivity of labor” in purely physical terms – as the ratio of the quantity of output produced for a given amount of labor (e.g. per labor hour), independent of the prices of inputs. For example, Marx first introduced his concept of “productivity” in Chapter 1 of Volume 1 in purely physical terms:
By “productivity” of course, we always mean the productivity of concrete useful labor …
Useful labor becomes, therefore, a more or less abundant source of products in direct
proportion as productivity rises or falls. (p. 137)
And in Chapter 12, on relative surplus-value and capitalism’s inherent tendency to technological change, an increase in the productivity of labor is defined as a reduction in the labor-time necessary to produce a commodity and is explained as the result of “technological revolutions” and “alterations of the labor process”:
By an *increase in the productivity of labor*, we mean an *alteration in the labor
process* of such a kind as to *shorten the labor-time socially necessary for the
production* of a commodity, and to endow a given quantity of labor with the power of
producing a *greater quantity of use-value…*
The technical and social conditions of the process [of production] and consequently the
*mode of production itself must be revolutionized* before the productivity of labor can
be increased. (p. 431)
And in all the passages that I have quoted concerning the causes of changes in prices of production, a change in the productivity of labor is always described as a change in the labor-time required to produce commodities. Two examples:
Once the cost-prices of commodities in the various branches of production are
established, they rise or fall relatively to each other with any change in the values of
the commodities. *If the productivity of labor rises*, the *labor-time required for the
production* of a particular commodity decreases and therefore its value falls; whether
this change occurs in the labor used in the final process or in the constant capital,
the cost-price of this commodity must also fall correspondingly. (TSV.II. p. 215;
(2) The general rate of profit remains unaltered. In this case the production price of
a commodity can change only because its *value has altered*; because *more or less labor
is required for its actual reproduction*, whether because of a *change in the
productivity of labor* that produces the commodity in its final form, or in that of the
labor producing those commodities that go towards producing it. The price of production
of cotton yarn may fall either because raw cotton is produced more cheaply, or because
the work of spinning has become more productive as a result of better machinery.
(C.III. p. 308)
I hope that these key passages are sufficient to show that Marx’s concept of productivity is defined in purely physical terms, independent of input prices.
Kliman, on the other hand, argues that a “change in productivity” could also mean a change in the price of inputs *without a change in physical productivity* because “change in productivity” is synonymous with “change in value” and “change in value” could be due to a change in the price of inputs. And the change in the price of inputs without a change of physical productivity could be due to the ongoing transformation of values into prices of production as in the TSSI interpretation.
A logical diagram of Kliman’s argument is:
TSSI transformation → Δ P(inputs) → Δ value → Δ “productivity” → Δ price of pd.
*all this without a change of physical productivity*!
But Marx never said anything about a change in productivity resulting from a change in the price of inputs without a change in physical productivity. Instead, the direction of causation is always the other way around: changes of the prices of the means of production are always caused by changes in the physical productivity of labor (including in the above two passages). And changes in the prices of the means of production in turn cause changes in both the values and the prices of production of commodities.
A diagram of Marx’s logic is:
Δ productivity → Δ P(inputs) → Δ value and Δ price of pd.
where *productivity means physical productivity*
In Kliman’s first discussion of the effect of the distribution of capital on the rate of profit, in Part 2 of his comments on my book, he stated that:
“… I first need to point out that my demonstration does not resort to any *terminological subterfuges*. In other words, the demonstration defines productivity … exactly as Moseley defines them.” [i.e. in terms of physical quantities]
“Terminological subterfuge” is a good description of Kliman’s attempt to define “changes in productivity” to include changes in the prices of inputs without a change of physical productivity. Kliman’s statement seems to be a tongue-in-cheek reference to his prior “productivity” argument in his 2007 book; but it is not a joke. Kliman’s “productivity” argument is a subterfuge.
Finally, as discussed in my previous comment, the TSSI does not actually explain why prices change from period to period even though physical productivity remains the same; so Kliman’s “productivity” argument is doubly invalid.
Fred Moseley has now alleged three times (Aug. 3, Aug. 6, and Aug. 10) that “[t]he effect of a change in the distribution of capital on the rate of profit is another difference between my interpretation of Marx’s theory and Sraffa’s theory (Marx: no change in the rate of profit vs. Sraffa: change in the rate of profit).”
I note that this claim keeps being made without EVER being demonstrated.
I’ll believe it when I see it, i.e., when it’s demonstrated.
I’m confident that I’ll never see such a demonstration.
To demonstrate it, Moseley “merely” needs to produce a numerical example in which:
(1) the distribution of advanced capital varies across sectors;
(2) the compositions of capital differ across sectors (so that we can distinguish between what happens to prices of production and what happens to values);
(3) the physical techniques of production utilized in the different sectors, and the real wage per unit of living labor, are shown to have remained constant (so that we can assess whether Moseley’s rate of profit changes when the distribution of capital–and it alone–changes);
and, MOST IMPORTANTLY,
(4) THE PER-UNIT PRICES OF OUTPUTS EQUAL THE PER-UNIT PRICES OF INPUTS.
I am confident that he cannot produce such an example. But let’s see. Hic Rhodus, Hic Salta, Fred.
“Once” he produces such an example, I’ll be happy to return to other issues raised in this discussion. 😉
“Let Marx speak for himself”
Marx’s own example of the effect of the distribution of capital on the general rate of profit is given on pp. 261-63 of Volume 3, which I referred to in a previous comment. His example has four capitals (A, B, C, D) with different compositions of capital, and with the initial distribution of capital, the rate of profit is 22.5%.
C V C+V S
A 75 25 100 25
B 60 40 100 40
C 85 15 100 15
D 90 10 100 10
Total 310 90 400 90
Rate of profit: 90 / 400 = 22.5%
Then with a different proportional distribution of capital, the rate of profit falls to 13.1%.
C V C+V S
A 150 50 200 50
B 180 120 300 120
C 850 150 1000 150
D 3600 400 4000 400
Total 4780 720 5500 720
Rate of profit: 720 / 5500 = 13.1%
The change in the distribution of capital results in a higher proportion of the total capital having a higher composition of capital (e.g. C and D) and thus the total capital has a lower rate of profit.
Marx summarized his discussion as follows:
“The general rate of profit is determined therefore by two factors:
(1) the *organic composition of the capitals* in the various spheres of production, i.e. the different rates of profit in the particular spheres;
(2) the *distribution of the total social capital between these different spheres*, i.e. the relative magnitudes of the capitals invested in each particular sphere, and hence at a particular rate of profit; i.e. the relative share of the total social capital swallowed up by each particular sphere of production.”
Physical quantities and unit prices have nothing to do with the determination of the rate of profit in Marx’s theory. The rate of profit is determined at the macro level of abstraction by the rate of surplus-value, the composition of capital in each industry, and the distribution of capital across industries with different compositions of capital. The rate of profit is determined prior to and independent of unit prices.
In Sraffian theory, on the other hand, *the rate of profit does not depend on the distribution of capital*, as Kliman’s example in Part 2 illustrates (the rate of profit remains at 20%). The rate of profit in Sraffian theory does not depend on the rate of surplus-value and the composition of capital but instead depends on the physical input-output coefficients, which remain the same when the distribution of capital changes. The rate of profit in Sraffian theory is determined simultaneously with unit prices.
Thus there is a clear difference between my interpretation of Marx’s theory and Sraffian theory in the case the distribution of capital and the effect of a change in the distribution of capital on the rate of profit. And also a clear difference on the relation between the rate of profit and unit prices. The basic reason for these differences is that Marx’s theory is based on the labor theory of value and surplus-value and Sraffian theory is not.
My reply is here: https://marxisthumanistinitiative.org/miscellaneous/moseley-fails-to-demonstrate-false-claim.html
MY INTERPRETATION OF MARX’S THEORY OF THE RATE OF PROFIT
Kliman accused me of “equating” myself and Marx and “bait and switch”. So I want to be clearer this time about what is my interpretation of Marx’s theory on the issue of the effect of the distribution of capital on the rate of profit.
I start with Marx’s example on pp. 293-95 of Vol. 3, discussed in my last comment.
In this example, the rate of profit is determined as an aggregate ratio at the macro level of abstraction. The quantities of C and V in each industry are taken as given and the quantity of S in each industry is assumed to be = V (i.e. the rate of surplus-value = 1.0) Then these industry quantities of C, V, and S are added up to obtain the totals of C, V, and S for the economy as a whole. Then the rate of profit is determined by dividing the total S by the total (C+V).
It follows from this macro labor theory of value and surplus-value that the rate of profit depends in part on the distribution of capital across industries with unequal compositions of capital (different ratios of C to V) because the S in each industry is determined by the V in each industry, not by the total (C+V). Marx’s example on these pages, which I reproduced in detail in my last comment, illustrates this effect. Marx compared two cases with four capitals and these capitals have different compositions of capital. In the second case, capitals with a higher composition of capital (i.e. relatively less variable capital and therefore relatively less surplus-value) comprise a higher percentage of the total capital than the first case and therefore the second case has a lower rate of profit.
*That is my interpretation of Marx’s theory the determination of the rate of profit* –
at the macro level of abstraction, prior to the determination of prices of production.
The rate of profit in my interpretation of Marx’s theory is determined by the aggregate ratios of C/V and S/V and the distribution of capital across industries, as summarized above (and also on the turnover time of capital, which Marx abstracts from here). According to my interpretation of Marx’s theory, the rate of profit is determined independently of physical quantities and unit prices.
My interpretation of Marx’s theory of prices of production (derived at a later micro level of abstraction), does assume that the prices of production of the means of production as inputs are equal to the prices of production of the means of production as outputs. But that equality has nothing to do with the determination of the rate of profit in my interpretation of Marx’s theory, which is determined at the macro level of abstraction prior to prices of production. That equality is not a “condition” in my interpretation of Marx’s theory of the rate of profit, contrary to Kliman’s claim. The determination of the rate of profit in my interpretation of Marx’s theory is summarized above and unit prices are not even mentioned, let alone this equality as a “condition” of determination of the rate of profit.
Rather the equality between the prices of production of the means of production as inputs and the prices of production of the means of production as outputs in my interpretation of Marx’s prices of production is because, according to my interpretation, Marx’s prices of production are *long-run equilibrium prices* and long-run equilibrium prices have this property (i.e. prices of production of the means of production are the same as both inputs and outputs). But that equality has nothing to do with my interpretation of Marx’s theory of the determination of the rate of profit which is determined prior to and independent of the determination of prices of production. (I have presented substantial textual evidence on numerous occasions to support this interpretation of Marx’s prices of production as long-run equilibrium prices; see for example: https://www.academia.edu/27678884/Marxs_Concept_of_Prices_of_Production_Long-Run_Center_of_Gravity_Prices)
It is of course possible that my interpretation of prices of production as long-run equilibrium is mistaken, but that is my (well-supported) interpretation and that is the reason I specify that input prices = output prices in my interpretation, not because prices are determined simultaneously with the rate of profit nor because this equality is a condition for the determination of the rate of profit.
It should also be pointed out that this equality between input prices and output prices does not apply to a given real wage because there is no such thing in reality of a given real wage. Workers are not paid a given bundle of wage goods (and the same bundle for all workers); instead workers are paid a money wage and each worker decides individually which goods to purchase and some workers decide to save a portion of their wage.
Kliman argued that in the example I presented from Chapter 9, the change in the rate of profit might be due to technological change. But Marx’s example clearly assumes constant technology, as is indicated by the fact that the organic composition of capital (C/V) is the same for all capitals in both cases. And Marx said nothing about a change of technology. The only thing is different between the two cases in Marx’s example is the distribution of capital.
And that is my interpretation of Marx’s theory of effect of the distribution of capital on the rate of profit.
Andrew, what is your interpretation of Marx’s theory of the effect of the distribution of capital on the rate of profit?
This just-published article is my response to Fred Moseley’s comment of August 27:
https://marxisthumanistinitiative.org/miscellaneous/moseleys-tangled-web-using-a-second-false-and-unsubstantiated-claim-to-demonstrate-the-first-one.html
MY DEMONSTRATION
In his latest comment, Kliman argues that my example (taken from Marx) is too general and incomplete and the change in the rate of profit could be due to technological change and/or a change in the real wage.
But Kliman does not mention two crucial details in Marx’s example that suggest that technology and the real wage have not changed – *both the composition of capital and the rate of surplus-value remain the same in all industries*. A change of technology would generally change both the composition of capital and the rate of surplus-value; so the fact that neither of these determinants of the rate of profit changes suggests that technology has not changed. Similarly, a change in the real wage would generally change the rate of surplus-value; so the fact that the rate of surplus-value has not changed suggests that the real wage has not changed.
Furthermore, even if technology or the real wage did change, but had no effect on the composition of capital or the rate of surplus-value (as in Marx’s example), then these changes would not cause the decline in the rate of profit in this example. Technology and the real wage affect the rate of profit only through their effects on the composition of capital or the rate of surplus-value. The only determinant of the rate of profit that is changing in this example is the distribution of capital which must be the cause of the decline in the rate of profit.
The rate of profit function in Marx’s macro theory of the rate of profit could be written in general terms as:
RP = f (CC, RS, DC) where DC in the distribution of capital
In Marx’s example, CC and RS remain the same, but DC and RP change. Therefore, the cause of the change in RP must be a change in DC. That is my “demonstration”.
So this is another difference between my interpretation of Marx’s theory and Sraffian theory, in which the distribution of capital has no effect on the rate of profit.
A Sept. 29 article of mine, elsewhere on this website, responds to Fred Moseley’s comment of Sept. 24.
Please see my reply to Kliman’s Sept. 29 article elsewhere on this website.