Debunking Steve Keen on Marx. Part I
by Jonathon Collerson
* Steve Keen teaches economics at the University of Western Sydney. He makes a lot of noise about having done some damage to Karl Marx’s concept of value. Both Guy Rundle and Eric Aarons have cited Keen as an authority on Marx, and particularly on the redundancy of Marx’s concept of value. So, showing that his account of Marx is simply wrong has its place. Part I goes over Keen’s reproduction of Ian Steedman’s account of the supposed transformation problem — which is supposed to demonstrate a logical flaw in Marx’s concept of value — in his Debunking Economics (2001). Part II will deal with his claim that Marx’s concept of value implies that means of production create new value; or that human activity is not a necessary condition of value’s existence.
1.
Steve Keen’s intervention in economics involves an attempt to refute Marx’s concept of value. He produced a 100-page Masters Thesis on Marx’s concept, in the early 1990s, that he later condensed into a series of academic journal articles and a chapter of his well-received book, Debunking Economics: The Naked Emporer of the Social Sciences (Keen 2001). Keen reproduces Ian Steedman’s allegation of internal inconsistency in Marx’s transformation of values into prices of production in Chapter 9 of Capital III (Steedman 1977: 29-49; Keen 2001: 281-86). Steedman’s argument was widely taken to have put an end to Marx’s concept of value in the 1970s.
Steedman directly follows Ladislaus von Bortkeiwicz’s argument that Chapter Nine of Capital III is internally inconsistent because the prices of commodities leaving production are greater than the values of commodities entering production. He argued that, if inputs and outputs did not share the same prices, sales would not equal purchases and reproduction would breakdown. Steedman begins his representation of this argument by emphasising a “physical quantities approach” to his analysis. He provides the following representation of a three-department economy producing iron, gold and corn; iron and corn are measured in tons, gold in kilograms and labour in hours. For simplicity, $1 the equivalent of one hour of labour.
Table 1: Physical quantities
See: Steedman 1977:38.
Steedman derives the ratios of labour to iron, corn and gold and calls these ratios ‘labour values’; denoted li,lc and lg. For instance, in the first row 56 hours of labour use 28 tons of iron to produce 56 tons of iron. Thus the ‘labour value’ of iron is
28li + 56 = 56li, or li = 2.
The ‘labour values’ of gold and corn are 1 and 4. The ‘labour-value’ of labour-power is derived by multiplying the means of subsistence, assumed to be 5 tons of corn, by its ‘labour value’:
v = 5lc = 5(4) = 20.
The total living labour is 80 hours, so surplus labour over the value of means of subsistence is:
80 – 20 = 60 hours.
The rate of surplus-value can then be derived:
s / v = 60 / 20 = 3, or 300%.
Steedman transforms Table 1 into ‘labour values’ by multiplying these values through the physical quantities columns. For instance, variable capital in each department is derived from the real wage. If the total real wage is 5 tons of corn, and this produces 80 hours of living labour, each hour is the equivalent of
5/80 = 0.0625 tons.
56, 16 and 8 hours are worked in the three departments, so real wages for each department are:
56(0.0625) = 3.5 tons
16(0.0625) = 1 ton
8(0.0625) = 0.5 tons.
‘Labour values’ are then derived by multiplying these real wages by the ‘labour value’ of corn:
3.5(4) = 14
1(4) = 4
0.5(4) = 2.
The total variable capital, or the value of labour-power, is 20.
2.
Table 2 shows Steedman’s physical quantities transformed into ‘labour values’: c is constant capital, v is variable capital, s is surplus-value = v(s/v) and w is the total value output = c + v + s. Taking these as data, we can derive further information: π is the average profit on investment = s([c + v] / [total c + total v]), p is the output’s price of production = c + v + π, s/(c + v) is the value rate of profit and π/(c + v) is the average rate of profit.
Table 2: ‘Labour values’.
Derived from: Steedman 1977:42 and Kliman 2007:150.
Table 2 is supposed to contain the internal inconsistency produced by Marx’s concept of value. The clue for Steedman is that, “both iron and corn appear to have different exchange values when sold as output from when they are purchased as inputs”. He continues: “This is nonsensical since sale and purchase are two aspects of the same transaction” (Steedman 1977:43-4). He concludes that inputs must also be given at prices of production so that both “aspects of the same transaction” are equal.
Keen elaborates this point by showing that when the iron industry buys means of production and labour-power to reproduce itself its cost-price will be
28(101.82/56) + 3.5(37.82/8) = $67.46
in the next period. This leaves a revenue of $34.36, which is $2.54 greater than its profit of $31.82 in this first period. “Clearly there is an inconsistency,” Keen concludes from this difference. “What is supposed to be an equilibrium (and therefore stationary) turns out not to be stationary at all” (Keen 2001:285). To amend Marx’s inconsistency, ‘labour values’ are discarded and Keen reproduces Steedman’s equilibrium prices for iron, gold and corn, derived as ratios of physical quantities, which allow one set of prices to regulate the inputs and outputs of each department (Steedman 1977:45-7; Keen 2001:285-6).
3.
Table 3 follows Andrew Kliman’s refutation of Bortkiewicz, by introducing non-reversible, chronological time (Kliman 2007:149-52):
Table 3. Values.
Derived from: Steedman 1977:42 and Kliman 2007:150.
Inputs to Period 2 are the production prices determined in Period 1. The physical quantities of Table 1 haven’t changed. The same quanties are consumed and produced, and the same hours are worked. But the value of iron has decreased to $101.82 and corn has increased to $23.64 (five tons of corn are required to reproduce labour-power and the remaining three tons are consumed by capitalists). Revenues, r, totalling $66.54 remain after both constant and variable capital is advanced.
We saw, above, that Keen collapsed (or confused) revenue and profit and saw this as disrupting reproduction. But revenues are not profits; revenue is an entirely distinct concept. Revenues fund capitalists’ consumption of total gold output and three tons of corn: they do not enter reproduction (Marx 1990:738). Profits are determined on the basis of the new value composition of capital and not the organic composition of the previous period, as Keen suggests. The increased value of corn necessarily changes the ratio of surplus-value to variable capital.
Keeping in mind that we’re calling $1 the equivalent of one hour of labour, surplus labour in Period 2 is
80 – 23.64 = 56.36 hours.
The rate of surplus-value for Period 2 can then be derived:
s / v = 56.36 / 23.64 = 238.41%.
Both surplus labour and the rate of surplus-value have declined in Period 2 and less surplus-value is generated as a result. The average profit for the iron industry is derived by dividing its composition (c + v) by the total social capital (total c + total v) and expressing this ratio as a part of the total surplus value (total s):
56.36([50.91 + 16.55] / [101.82 + 23.64]) = $30.31.
The difference between profit in Period 2 and Period 1 is
30.31 – 31.82 = –$1.51.
This doesn’t indicate any inconsistency, but is consistent with the increased value of means of subsistence and, thus, the value of labour-power. The difference of $1.51 simply expresses a decreased rate of surplus-value and, against Keen, has nothing to do with revenue.
Once capital has been advanced and revenue spent, the total social product has been reproduced. Total purchases in Period 2 are the total social capital (total c + total v) plus total revenue (r):
125.46 + 66.54 = $192.
This equals the total production price for Period 1: the total social capital invested plus the total profit (π):
132 + 60 = $192.
Finally, in Period 2 all sales equal their purchases:
Iron: (corn c + gold c) – (iron v + iron r) = (21.82 + 29.09) – (16.55 + 34.36) = $0
Gold: (iron r + corn r) – (gold c + gold v + 3 tons of corn) = (34.36 + 13.64) – (29.09 + 4.73 + 14.19) = $0
Corn: (iron v + gold v + 3 tons of corn) – (corn c + corn r) = (16.55 + 4.73 + 14.19) – (21.82 + 13.64) = $0
In other words, demand meets supply exactly. There is no excess value and, thus, reproduction cannot expand. The total social product has been bought and sold at production prices that diverge from input values. The model is stationary. There is no demonstrable inconsistency.
4.
The transformation of values into prices is only a problem on the assumption that the prices of commodities entering production and prices of commodities leaving production, “are two aspects of the same transaction”. But input and output prices inform two separate transactions made at separate points in time. The point of the production process is the augmentation of value. Marx notes that it would be nonsensical to invest capital if that capital value did not expand. Thus, the value that issues from the circuit of a capital is not equal to the capital value invested (c + v), but to the total value output (w); viz. the value that constant capital represents plus a new value that both replaces the variable capital and produces a surplus-value beyond this (c + v + s). Any individual capital is always under the condition of the total social capital. Competition among capitals distributes the total surplus value across individual capitals to give the average profit on investment (π). For instance, the iron industry produced a surplus-value of $39.45 in Period 2, but only realised a profit of $31.82. Competition with other capitals forced it to forfeit a value of $7.63. The addition of the average profit to the capital invested finally gives the prices of production paid for inputs in the next period (c + v + π).
Simultaneity suggests that inputs are bought at prices that their consumption and circulation have yet to generate. Prices are simultaneously generated for inputs and outputs. But this means that inputs are constrained to prices of production that logically cannot yet exist, if we allow non-reversible time. This renders a key aspect Keen’s intervention into economics inconsistent. He both rejects simultaneity and follows Steedman in using this method in his attempt to refute Marx. “If economics is to have any relevance to the real world—if economics is even to be internally consistent—then it must be formulated in ways that do not assume equilibrium,” Keen says. “Time, and dynamic analysis, must finally make an appearance in economic analysis” (Keen 2001:175). The exclusion of time and disequilibrium is a question of scientific method for Keen: “A theory may well draw power from ‘unrealistic’ assumptions if those assumptions assert, rightly, that some factors are unimportant in determining the phenomena under investigation,” he says. “But it will be hobbled if those assumptions specify a domain of the theory and real world phenomena are outside that domain” (Keen 2001:153). Time is a “real world phenomena” that is outside the theory-domain given in the transformation problem. In its place, constant physical quantities and equilibrium prices are assumed.
Keen goes further. He scolds those interpreters of Marx that insist that time “make an appearance” and equilibrium be discarded. “The latest attempts [to refute the charge of internal inconsistency] argue that, since Marx’s theory was actually dynamic rather than static, the transformation problem should be solvable in a dynamic model,” Keen says, then immediately adds, “Nice try guys, but you really shouldn’t bother” (Keen 2001:288). Is this not a classic example of moving outside an argument, if not into ad-hominem, when your opponent nulifies your claims? And yet Keen argues that temporal interpretations of Marx must have greater “relevance to the real world”, given that they allows time to exist, than his own static interpretation. In other words, Keen must reject the first part of his own refutation of Marx.
Bibliography
Keen S (2001) Debunking Economics: The Naked Emperor of the Social Sciences, Zed Books, London and New York.
Kliman A (2007) Reclaiming Marx’s Capital: A Refutation of the Myth of Inconsistency, Lexington Books, Lanham, Maryland.
Marx, Karl (1990) Capital I, Penguin, London.
Steedman I (1977) Marx after Sraffa, New Left Books, London.
[...] Here is a working-up of Marx’s ideas from Chapter 9 of Capital III. It is the first half of a response to Steve Keen’s absurd criticisms of Marx. It might be helpful in working through Chapter 9 & 10. LikeBe the first to like this post. [...]
Kliman’s so-called temporal system relies on an algebraic trick. If prices at input are not the same as prices at output then effectively, in an n-good economy/static equilibrium price matrix (the system Steedman uses and Kliman bastardises), you have 2n unknowns and thus the entire system is indeterminate. Kliman’s system can be used to demonstrate that Marx’s fundamental equalities hold, but it could also be used to demonstrate that the price of the moon is equal to the price of a kilo of green cheese.
Kliman tries to get around this by arguing that the price of initial inputs in period 1 is “a given” but this is just arguing in a circle, assuming what he is trying to demonstrate. His idea that “temporal” means single-step year long production cycles is, frankly, weird.
Ironically Kliman could escape from this particular algebraic mess by embracing a strong version of the labour theory of price which makes all prices directly equal to labour content; he’d then have the same number of equations as unknowns; but for reasons unknown he’s dedicated to attacking Shaik, Cottrell, etc and others who advance this strong equivalence.
You also seem to assume that Keen is in agreement with Steedman about the form that economics ought to take. In fact he is very critical of Steedman’s static equilibrium models as well as Klimans. See e.g. Keen “Answers (and questions) for Sraffians (and Kaleckians)”, Review of Political Economy, Vol 10, No.1, 1998, which argues that Steedman/Sraffa long-run prices will never be arrived at in a dynamic economic system. I think Keen reads Sraffa in accordance with PoC’s subtitle “prelude to a critique of economic theory”; Sraffa’s work is first and foremost a hatchet job, not an alternative school per se.
Rather Steve has constructed a dynamical system based on Godwin’s version of Marx’ debt crisis theory, which is (I think more than Steve is willing to admit) quite similar to the forms of dynamical Marxism put forward in Farjoun & Machover’s “Laws of Chaos” and Wright, Cockshott, Cottrell, et al’s “Classical Econophysics”. Keen’s paper “A Marx for post-keynesians” is a pretty clear olive branch which mediates between these two views.
Thanks for your comments, James.
I’m not aware of all the literature you are referring to, so I can only make a small comment back.
I don’t think there is anything wrong with Kliman’s procedure here. Alan Freeman makes a similar case; I used Kliman simply because he was more current in my reading, I could have modeled it after Freeman with the same conclusion. The point is that numbers aren’t what is particularly compelling for me, though it is nice that the proof works. What is compelling for me is that they they are consistent with the sort of materialist philosophy that Marx was working on. I think the numbers are a subaltern proof of the materialism rather than the starting point for it. In this way I think the comment Kliman makes about data being given is significant, and ought not be dismissed lightly. To put it simply, the total value in the system is unknowable, because it must contend with the aleatory conditions of its own circulation. The actual values that capitals take are totally uncertain and can’t be guessed at (the GFC was an example of the result of guessing at values). Thus you have to take what is given as data rather than assume future values as data.
On another point, I am aware that Keen is actually opposed to Steedman’s sort of economics. I mention this in the paper as suggesting something internally inconsistent in his use of Steedman and his opposition to reading Marx temporally.
That is as far as I can go, without overstepping my discipline. It is
ironic that this part of my argument against Keen, is really the second
bit – I’ve developed the philosophical part but haven’t felt compelled
to publish it. Keen doesn’t understand Marx concepts at a very basic level; for example the identity of constant capital and means of production or the exclusion of a distinct concept of value in favour of his ‘dialectic’ of exchange-value and use-value, to name two errors.
Finally, I’m unsure about about the price of the moon and don’t eat dairy.
Thanks for this article, i´m curious on how Steve Keen (or anybody) can affirm such things as the ones aimed at the second part of this response. I suppose it´s based on misunderstandings, judging by your previous comment. Anyway, i´m looking forward to the second part.
Do you still plan to post Part 2? I would be very interested in reading it.
Also, is this numerical example supposed to represent an equilibrium system? Because the labor value of iron along with all c, s, w, pi, & p values are monotonically decreasing and, if you carry Table 3 forward far enough, you end up in negative territory.
Thanks for your comment. I produced a draft of part two, but I really wanted to avoid a ‘what Marx really said’ response, as I don’t think Keen really cares what Marx really said. E.g. he doesn’t at all acknowledge that means of production and constant capital are different concepts; the first refers to material objects (computers, pricing guns, tractors, &c.) the second to the value that those material objects bear at a particular instant. This leads Keen to think that when Marx talks of an increase in the power of a set of means of production, i.e. that they are materially more productive, that he is talking about means of production producing value. By definition this is ruled out in Marx, value signifies human activity, and yet Keen bizarrely reads this into Marx’s mysterious dialectics (which is another bizarre avenue). On the second point, the model isn’t at all an equilibrium; but more importantly it is a representation, rather than something that could have causal power in the reproduction of the capitals it signifies. What those capitals do determines the model not the other way around. That said, I’m not an economist. All that the tables need to do here is show one way in which there is no transformation problem.
If this matter is still of interest to you: this right here is probably the most comprehensive critique of Keen’s treatment of Marx I’ve encountered.
Thanks Helund. I think have seen that somewhere. But, yes, admittedly, I don’t really think Steve Keen’s stuff is worth the effort. If you every meet him, he is quite provocative, arrogant, you have this urge to put him in his place, and his stuff on Marx such a joke that is asking for a take down, but really unless one’s field is economics, and particularly heterodox economics, I don’t think it quite matters what Keen is saying. This was really just a way for me to read those chapters in Capital III, with a problem to work through. I really benefited from the exercise, with the aid of Kliman’s book.
As indicated in the comments beneath that article, I actually have met him. I agree he does come off as quite haughty while soapboxing, though on a personal level he seemed nice enough. I mean, he didn’t seem like an unkind person or any such thing. And he has assembled a pretty thoroughly damning critique of marginalist orthodoxy in his book. I’ve seen a number of neoclassical economists who disagree with him who nevertheless concede his book as a solid dismantling of neoclassical pedagogy, at the least, particularly at early and intermediate levels.
But yeah, his treatment of Marxian econ is surprisingly slipshod in light of both his thesis and his published work on the subject, and his “renegade” conceit is pretty off-putting.
[...] model is simply a misinterpretation of Marx, and it is not even necessary to go into the maths. There is, of course, a possibility that this is an overly superficial interpretation and I am [...]
[...] have more sympathy with the LTV (mostly because its proponents seem to have coherent responses to every criticism thrown at it), but I remain unconvinced. The defences of the labour theory of [...]