Brief notes on the “Marxian economists” vs. Marx

A guest post (from the end of last year) by Guglielmo Carchedi on Michael Roberts’s blog^[CR24] makes an attempt to refute the framework put forth in Emmanuel Farjoun & Moshé Machover’s Laws of Chaos (1983) (henceforth LoC).^[FM83] Paul Cockshott responds on his own blog^[Coc24] in defence of LoC’s approach, and Carchedi fires back with an addendum.

All involved are avowed “Marxists”, and the result is generally illustrative of the level of (mis)understanding of Marx’s own work that’s typical of so‐called “Marxists”. By investigating said exchange, this post begins a brief series, which will only touch upon LoC in the second instalment; that is, the present post neither attacks nor defends LoC.

After an introduction & a very brief outline of part ii of Capital, Vol. 3^[1] (concerning the transformation of profit into average profit), Carchedi argues against what he sees as LoC’s value theory (emphases in the original):

A commodity acquires its individual value in the process of production. This is transformed into its social value in a different process, distribution. These are two different temporal processes taking place at different times, a hardly challengeable fact. Yet, the critique rests on an implicit assumption: that production and distribution take place contemporaneously, i.e. that time does not exist.

Already, something here is deeply confused. Carchedi’s basic distinction is binary: “individual” vs. “social” value. Given his description of the former, it must be equated with the specific, concrete labour‐time spent in production. But what of the latter? On the face of it, “social value” is pure tautology: value in the relevant sense is inherently social. To think otherwise is commodity fetishism.^[2] Thus, so far, Carchedi’s “individual value” isn’t value at all — it’s concrete labour‐time.

Because (as we’re about to see) Carchedi has no theory of value, he needs to rely on a temporal gap between production & “distribution” to be able to account for even the most everyday value & price determinations. For now, we merely note that Marx has a very different idea of how “challengeable” this temporal gap is when it comes to the transport industry.^[3]

Carchedi continues:

The value of a commodity can disappear, but only if the commodity is destroyed (by humans or by natural forces). As long as it is not destroyed and it is sold, it must realize its value. If it realizes more or less than its value, we must account for the difference.

He believes that the thing literally has value, as if it were inhabited by some vapourous value‐substance; impregnated with value in such a way that it can realise this already‐existing(!) value in exchange. Thus we, apparently, need only account for how the thing is sold at a price that doesn’t correspond exactly to its value:

Suppose that the labour used for $O$ at $t_1$ is 10⁢ hours and that the labour used for $R$ at $t_2$ is 8⁢ hours. The replacement thesis would argue that $O$ sells at 8 [hours…?]. But what happens to a value of 2? Has it evaporated?

No; there’s nothing to evaporate. Value is constituted by a social web of exchange‐relations; a thing is in no way destined to “have” value just because a certain quantity of (concrete!) labour went into its production. This vantagepoint wrongly assumes that…:

• …the conditions under which a thing is produced are directly observable to potential buyers.

• …use‐values are directly physical properties of things. As one counterexample, consider a thing that has been obsoleted by a technological advancement — its use‐value is voided, in spite of it suffering no physical alterations.

  To say that a commodity has use‐value is simply to assert that it satisfies some kind of social need. [I]f the commodity in question is produced on a scale that exceeds the social need at the time, a part of the society’s labour‐time is wasted, and the mass of commodities in question then represents on the market a much smaller quantity of social labour than it actually contains.

• …worst of all, any product of labour is always‐already a commodity.^[2] We thus lose the basic insight of Marx’s work: that the commodity‐form, & especially the absolute generalisation of commodity production — that is, capitalism — are historically contingent.

Marx is clear on the matter that Carchedi is here stumbling over; not only in Vol. 1, but also Vol. 3 (emphasis in the original):

The value of any commodity — and thus also of the commodities which capital consists of — is determined not by the necessary labour‐time that it itself contains, but by the socially necessary labour‐time required for its reproduction. This reproduction may differ from the conditions of its original production by taking place under easier or more difficult circumstances. If the changed circumstances mean that twice as much time, or alternatively only half as much, is required for the same physical capital to be reproduced, then given an unchanged value of money, this capital, if it was previously worth £100, would now be worth £200, or alternatively £50.

The abstract labour‐time that’s socially necessary to produce a given commodity‐type is determined through collective social experience, mediated by exchange. A shirt does not “remember” that it was produced during a year of bad cotton harvests; when a better harvest comes around, it will still just be a shirt.

But let’s allow Carchedi to solve his own little riddle:

What happens in reality is that at $t_2$ both $O$ and $R$ compete on the market. Due to price competition, they sell more or less at the same price to other sectors and realize the same value, e.g. 9. If $O$ has cost 10⁢ h but realizes 9⁢ h, it loses 1⁢ h. If $O$ has cost 8⁢ h and realizes 9⁢ h, it gains 1. What $O$ loses is realized by $R$, and vice versa. But notice that they do not lose to and gain from each other directly because they do not buy from and sell to each other. Rather, they exchange with other commodities whose sectors act as a clearing house.

We can now pinpoint the exact location of Carchedi’s confusion. When he says “social value”, he means market value.^[4] We’re left with a theory having the following structure:

Use‐value

The physical thing itself.

Individual value

The concrete labour‐time embodied in the commodity.

Market value

The price (Carchedi specifically uses “price” in the above quotation, and insists earlier that value is always realised unless the use‐value is destroyed or unsold) at which a commodity actually sells, equal to the average of the individual values of the commodities of that type currently on the market.

It’s hopefully clear that this bears precious little relation to Marx’s own theory. We’ll put aside that Carchedi’s insistence on value being realised, & his conflation of value with price, are inconsistent with Marx’s analysis of supply & demand^[5] (in addition to being conceptually confused). Instead, we’ll focus on the links — absent from Carchedi’s misunderstanding — that connect the core of Marx’s value theory to market value.^[4]

It’s true that Marx occasionally uses the term “individual value” — but exclusively when competition between capitals is the focus of analysis. This logical step is only valid when capitalist production (and, a fortiori, commodity production) is highly developed, such that both of the following obtain:

1. Intercapital competition forces producers — represented by capitalists — to thoroughly economise their production‐processes, so that e.g. a capitalist who “has a foible for using golden spindles instead of steel ones”^[6] isn’t economically viable.
2. The means of production are viewed not primarily as use‐values for concrete labour purposes, but as autonomous (i.e. disconnected from the exchange moments that actually constitute value) exchange‐values for valorisation purposes.

Marx already anticipates this latter point in ch. 1 of Vol. 1,^[7] but enunciates it most clearly in the Resultate (emphases in the original):

Apart from the service they perform as use‐values in the process of production, all the means of labour that have been produced now also serve as ingredients in the valorization process. Where they are not changed into actual money, they are converted into accounting money; in short, they are used as exchange‐values and the element of value that they add to the product in one way or another is precisely calculated. […] In other words, then, inasmuch as the commodity is treated as an autonomous exchange‐value, it acts as money. Thus since wheat, hay, cattle, seed of all kinds, etc. are sold as commodities — and since without the sale they cannot be regarded as products — it follows that they enter production as commodities, i.e. as money. […] They have long since become commodities in [the capitalist’s] eyes, since they are articles, means of labour, that are at the same time values forming part of his capital.

By gutting the essence of Marx’s value theory, Carchedi is left with merely a theory of prices — and a pitiful one, at that. Consequently, he feels the need to defend himself against a so‐called “infinite regression critique”.

The defence proceeds with series of eight formulae that, in a parade of tautologies, simply assert that value = money = price. We’re then treated to a lovely chart ostensibly providing empiric evidence of this crude equation, with no methodology other than “Source: author’s computation”. Because said computation presumably uses the aforementioned formulae, it directly begs the question by effectively defining the correlation to be 100% at the beginning of each time‐step (with the data spanning 68 such beginnings).

It behoves us at this point to leave poor Carchedi alone; the rest of his paper largely consists of fumbling around with buzzwords like “chaos” & “dialectics”, with some of the misunderstandings touched upon above sprinkled in for good measure.

Given the structure of Cockshott’s response,^[Coc24] the centrepiece is evidently his 2013 paper in World Review of Political Economy.^[Coc13] We’d therefore be remiss not to touch upon this paper.

The following critique of [Coc13] isn’t meant to imply any defense of Marx’s law of the tendential fall of the profit‐rate (LTFPR).

Because one of the most common objections to LTFPR from the neoricardian standpoint is Okishio’s theorem, Cockshott is quick to bust out the input–output tables.

Table 2 is set out so that certain constraints required for reproduction are met. These were analyzed by Marx when he looked at simple reproduction in Chapter 20 of Capital Volume 2. In particular we need to ensure that the total surplus value equals the total output of luxuries, that total wages equal the total output of the subsistence goods, […]

If we actually consult the above‐cited chapter of Vol. 2,^[8] however, we predictably find that Cockshott’s fantasy capitalists are starving! It’s simply not possible to survive on luxury goods alone, as much as I’m sure the capitalists would if they could.

In any case, the conclusion of the paper is the following formula:

Where:

$r_{\mathrm{e}}$

Profit‐rate (for the total social capital) at equilibrium.

$t$

Rate of cheapening of the components of constant capital (“technical progress”).

$g$

Growth‐rate of the employed workforce.

$\mathrm{\alpha }$

Accumulation‐rate.

The first thing we notice is that the use of $g$ has some strange implications. Not only is it assumed by definition that the requisite capital (constant & variable) exists to employ the workforce that $g$ represents the quantitative growth of, but it’s also conversely assumed (but no longer by definition) that there are enough workers to valorise the existing — including the freshly‐accumulated — capital.

Moreover, this same array of dissociated definitions says nothing in particular about the composition of capital, neither technical nor organic. This is rather surprising given that the paper’s explicit starting‐point is Capital, where composition of capital is clearly a major part of Marx’s arguments.

The mathematics of this formula are also unclear. As near as I can tell, the earlier version of the formula, $r_{\mathrm{e}}=g∕\mathrm{\alpha }$, is mathematically valid; but I can’t say the same for the final version (which adds $t$). To see why, consider an organic composition of ${50}_c+{50}_v$ with a rate of surplus‐value of 200%, yielding ${100}_s$. It follows that the initial profit‐rate $r_0={100}_s∕\left({50}_c+{50}_v\right)={100}_s∕{100}_C=\mathrm{100\%}$. For this to be the equilibrium profit‐rate (i.e. $r_0=r_{\mathrm{e}}$), we might consider, for example, $t≔\frac{1}{5}$, $g≔0$, and $\mathrm{\alpha }≔\frac{1}{5}$. This causes the ${50}_c$ to be devalued to ${40}_c$, and causes the total capital to increase in value terms to ${40}_c+{20}_{\mathrm{\alpha }}+{50}_v={110}_C$. Because $g=0$, nothing else changes; in particular, the variable capital is unchanged, and ${100}_s$ is still extracted. The new profit‐rate, then, is $r_1={100}_s∕{110}_C\approx \mathrm{90.9\%}\ne \mathrm{100\%}$. It seems we weren’t at equilibrium, after all.

However, the real justification here is the charts. These make the predictive power of the above formula appear quite impressive indeed. But what’s actually going on in these charts?

• Where’s even the most cursory of statistical analysis?
• Where are these numbers coming from? We’re told that the raw data are to be found in the Penn World Table, but that’s about it. To give but one example: how exactly are values of $t$ divined from these data?
• What’s the significance of the chosen countries & time‐periods?

Taking a step back, there are some even bigger problems with the overall approach here. The ultimate structure of Cockshott’s argument is that we can disembody particular statistical measures ($t$, $g$, $\mathrm{\alpha }$) from their social places & significances and, without reïnterconnecting them, plug them into a formula that will make ostensibly decent predictions on a two‐to‐three‐year timescale.

If the inputs to this formula take on their values apparently by magic, and have no particular relationship to one another, then in what way is this an improvement on Marx’s treatment? For instance, defining $g$ essentially in terms of population growth prima facie ahistoricises this predictor of the global profit‐rate.

Allow me to make an analogy to the medical world. We hope that our treatments are safe & effective, and to that end, we want good empiric evidence for that safety & efficacy. Hence evidence‐based medicine. But without an understanding of the human body that can plausibly be called scientific (whence science‐based medicine), evidence for the efficacy/safety of a treatment is evidence only for that treatment, as it was used in the studies in question. The relationship of the treatment to the inner mechanisms of the human body is thus mystified.^[9]

The paper in question is concerned with the predictor $r_{\mathrm{e}}$, insofar as it predicts the movement of a surface‐level observable. Any theoretic plausibility of said predictor is rubbed off from Cockshott’s cherrypicked version of Marx’s arguments with which the paper begins. This is a serious deficiency, considering that the paper has been proffered in response to Carchedi’s assertion of LTFPR as a crucially important element in the theorisation of crisis (& of various other unsavoury phenomena). How is this supposed to improve upon Marx and thus answer Carchedi? By making trivial econometric predictions? Evidence‐based economics, indeed.

And, speaking of crisis, how is this model supposed to be affected by, for example, overproduction (itself partly a result of $\mathrm{\alpha }$)? As much as Cockshott is critical of Okishio here, he nonetheless commits a similar error. Where Okishio assumes a physically‐fixed basket of “wage goods”, this is in ignorance of the use‐values of means of consumption;^[10] the “basket”, if there is such a thing, must change with social changes in individual needs due to new technology, cheapening of means of consumption, shifting cultural norms, etc., etc.. For both Cockshott & Okishio, the fact that commodities produced must generally be consumed in order for reproduction to proceed normally isn’t relevant, except insofar as the requisite exchange‐value relations must be satisfied (as in Vol. 2, part iii).

We might think of this model as a work of “dynamics” inasmuch as it seems to be dealing with timeseries & rates over time: $t$, $g$, and $\mathrm{\alpha }$. However, the underlying logic is, as made explicit in the paper itself, essentially simultaneous linear equations (basically, linear algebra). The implication is that harmonious equilibrium is assumed, “chaos” is neither present nor necessary (this being important mostly because Carchedi’s most fundamental objection is to exactly this “chaos”/“probability” aspect), and $r_{\mathrm{e}}$ is just a single number to which the actual total capital must converge… somehow.

Ultimately, this yields a viewpoint from which any crisis must be little more than a reaction to (unspecified or underspecified) variables outwith the purview of the model.

It would seem that [Coc13] has some sociopolitical implications. But we’re largely left to wonder what those are — with the exception of “it shows the antagonistic and in the long term reactionary relation that the capitalist class has with the development of the productive forces”, which is already in Marx anyway — as a result of the partial abandonment of socioëconomic categories & links, in favour of trivial econometric predictions.

Continuing on with Cockshott’s reply to Carchedi, we arrive at the question of the importance of production‐prices (& the general profit‐rate with which they’re intimately tied up) to Marx’s theories. We’re thus treated to this little gem:

If it was so central, it would have come on early. Instead, it’s relegated to something that he didn’t bother actually publishing himself.

First of all, Marx died. That’s why he “didn’t bother actually publishing” Vols. 2, 3, & 4^[11] himself.

More to the point, Capital is logically structured,^[12] proceeding from the most one‐sided & abstract germ of the theory to progressively more sophisticated versions, peeling back layers of abstraction/assumption (perhaps Cockshott will understand better if we use the term “successive approximation”) as it goes on.

There’s a method to the madness, and that method is to start from a singular capital (Vol. 1), then introduce the exchange between capitals (Vol. 2), then introduce competition between many independent capitals (Vol. 3), and finally conclude with a history of approaches to the problems now formulated & treated (Vol. 4^[11]).

Marx already has this structure in mind throughout Vol. 1 (according to Cockshott, the only volume), where he, for example, warns the reader many times that the assumption of commodities being sold at their values is being made. This assumption is made explicitly in Vol. 2 as well, and only Vol. 3 finally discharges it by introducing production‐prices.

Moreover, Cockshott’s view of Capital implies that the discussions of commercial capital, money‐dealing capital, interest, rent, etc. — all of which are treated only after the exposition of production‐prices (and of LTFPR, for that matter) — are unimportant as well. To be sure, it’s possible to adjust Marx’s treatment of these subjects to an alternative understanding — or even rejection — of production‐prices; but Cockshott seems intent on throwing the baby out with the bathwater.

In any event, Cockshott concludes the reply by emphasising that, in reality, profit‐rates are far from being the same for each capital. But this anticipates the next instalment of this series, which will review LoC — so this is a good place to stop for now.