Letter to Frank S. about the labor theory of value

(Focusing on the ability of machines to contribute
to surplus value, and Say's "Law")

Dear Frank,

Thanks very much for commenting on my "review" of Keen’s Debunking Economics. I’m also glad that you’re working on some more thorough comments, because I think it will probably take several in-depth back-and-forths between us before we can get to the bottom of things.


1.   Not a proper "review".

I agree that my "review" is not really a review, properly speaking. It is only a partial review of just one of Keen’s chapters (his critique of "Marxian economics"), and even then focuses mostly on the labor theory of value (LTV). Moreover, it is as much an essay occasioned by my reading of Keen’s book as it is an actual review, even of that one chapter. That’s why I entitled it "Steve Keen on Marxist Economics, Together with a Mini Essay on the Labor Theory of Value".

The vast bulk of Keen’s book, by the way, is not relevant to Marxist political economy at all. It is instead a thorough trashing of neoclassical bourgeois economic theory from within the general bourgeois perspective (albeit the liberal Keynesian wing). It brings out many of the inconsistencies, bad logic, false assumptions, and so forth, in neoclassical economics. I have no particular interest in reviewing all that because—for one thing—I don’t take neoclassical economics that seriously to begin with. And for another thing, I agree with most of Keen’s comments against neoclassical economics. The basic thing I have to say about Keen’s assault against that dominant bourgeois paradigm is just to recommend the book to anyone who does take any of that stuff seriously.

The main reason I read books—any book—is to provoke my own thinking. And I really don’t see anything wrong with that. Why must we always put on blinders and focus in the direction the author wants to go? Why not let our thoughts go where they may? I’ve read some books that were totally worthless to me except for one or two tangential points that really set me thinking deeply about something. And that made those books valuable to me, no matter how dubious their value is in general. So I guess I reject your criticism that my "review" wasn’t really a proper review. I wasn’t trying to sell it as one.

It wasn’t Keen’s book, by the way, which convinced me of the argument about the LTV that I put forth in my essay. For years—ever since I first read Marx’s "Wages, Price and Profit" and started learning about the LTV—I have been uneasy about the implications of the theoretical continuum from simple machines through complex computer-controlled machines to human beings ourselves. All along I’ve refused to admit that an "android" (if one existed) couldn’t create surplus value. But I never sat down before and tried to systematically work out what this might mean for the LTV. Keen just provoked me to bite the bullet, and begin to try to consistently work out my previously undeveloped notions in this area.

Yes, it is dangerous to do something like this. You can end up some place where you didn’t intend to go. "Let us admit at once the case of the conservative; once you start to think there is no telling where you will come out." [John Dewey, quoted from memory. I suppose I have to apologize for quoting somebody like that. But I’ll quote anybody, even the devil, when he is right.] But, actually, I don’t think my thoughts have ended up opposing Marx so much, as improving on Marx’s somewhat faulty version of the LTV.


2.   The "android argument".

I think you misunderstand what I was trying to do with what I called the "android argument". My point was not to "suppose" that androids will be built (though I do believe that will probably happen), and then to argue that "therefore" we have to modify Marx’s version of the LTV. Instead I was using the android thought experiment to try to soften resistance to some principles or propositions which I hold to be true whether androids are ever constructed or not. Some of these propositions (which collectively form my overall argument) are:


1)   There is nothing mysterious nor magical about human labor that can make it alone capable of generating surplus value in a system of capitalist production. (Marx never stated just why only human beings supposedly have this capability, and thus left it all very mysterious.) There must be a good answer to the question, "Why is human labor able to produce surplus value?" It is incumbent on Marxist political economy to state precisely what the answer to this question is.

2)   Furthermore, it is not human creativity or intelligence that explains why human labor can generate surplus value. (Some labor that generates surplus value requires very little intelligence and shows virtually no creativity, such as certain types of extremely repetitive assembly line work. And, on the other hand, some machinery already demonstrates considerable intelligence and sometimes even some creativity—though artificial creativity is much rarer so far, and is mostly confined to software, such as certain expert systems.)

3)   Moreover, there is no single specific human capability which, when exercised, alone is able to generate surplus value. (On the contrary, there is an indefinitely large number of actual things human workers can do which can generate surplus value, and new ones are constantly being developed as production processes change.)

4)   Moreover, even now a great many of these specific things which humans can do, and which generate surplus value, can also be done by machines. (And in historical terms, machines are being rapidly improved in their abilities to replace human beings in more and more types of work.)

5)   All the sorts of things that human beings can do which create surplus value come in degrees, from very crude and primitive efforts to the highly skilled and sophisticated. This is why we first see machines replacing the cruder forms of human labor, and then—as the machines are improved—ever more sophisticated forms of human labor of the same type.

6)   The thing that really explains why human beings are able to generate surplus value is simply that human beings can be "used over and over" in the capitalist production process. (I.e., they are not used up in the production of any commodity, as raw materials are.)

7)   But if the last principle is true, then anything else that is used in the production process, without being used up in each output commodity (i.e., tools and machinery), should also contribute to surplus value.

8)   However, machines and tools themselves were created by past human labor, and in fact all commodities that presently exist are in the final analysis the result of the application of direct or indirect human labor in various forms to the natural products of the world around us.

9)   It is in general impossible to quantitatively compare goods (or commodities) by comparing their use values, since use values are so diverse, tend to be unrelated each other, and depend on the diverse and ever changing needs and subjective desires of different individuals. (Whether a loaf of bread or a warm coat is of greater value to a person depends on the situation.)

10)   But there is an underlying basis, and only one such feasible basis, for quantifying the difference in exchange values for different commodities: the socially necessary labor times incorporated into them. (So far this means just human labor, though if androids are ever constructed, or if sentient aliens were to show up on earth and also become workers, it would then include their labor.) Moreover, the labor considered here must include not only the direct labor, but also the properly apportioned indirect, or past labor contributed both in the form of raw materials and tools & machinery. (The actual exchange values we see, however, may be adjusted from this total socially necessary labor time for various reasons, but the underlying labor time still forms the primary basis for these exchange values.)

11)   Thus the labor theory of value (LTV) remains essential in political economy. (Systems, like Sraffa’s, that derive output prices directly from input quantities and prices (including labor-power prices) may be technically possible, but are irrelevant here. Political economy needs to explain a lot more than one price in terms of others. For example it needs to explain how exploitation occurs and the deep reasons for capitalist economic crises. (Moreover, all prices themselves implicitly relate to socially necessary labor times; the fact that a car costs $20,000 is meaningless except within the framework where workers are paid definite amounts for an hour of their labor time.) Sraffaian economics obscures such things. We need to stay true to a profoundly political economics, a political economy that clearly exposes the genuine human relationships at the bottom of things.)

12)   But Marx’s specific version of the LTV needs to be modified in one major way: It is not only current labor which produces surplus value in a system of capitalist production, but also the continuing use of past labor (embodied in tools and machines). Modifying the LTV in this way allows us to resolve otherwise insurmountable logical and conceptual problems in traditional Marxist political economy. This revised LTV is far more coherent and defensible, but retains the essential aspects of the traditional Marxist explanations for exploitation, the source of capitalist profits, the underlying source of capitalist overproduction crises as arising from the exploitation of labor in the capitalist production process itself, and so forth.


I could add to this list of principles/propositions/arguments (and refine or expand on them), but this is good enough for now. These propositions collectively are my real argument. (In my first review/essay their purpose may have seemed simply to provide support for the "android argument", but actually the android argument was supposed to provide support for them.)

Thus contrary to what you took me to be saying, my argument is not based on speculations about androids, even if that may be what initially led me to it. If you want to shoot down my overall argument and position, you need to focus not on androids, or anybody’s suppositions about them, but on the specific propositions I listed above.

For example, if you think that there is something very special about human labor, and not only human labor, but specifically direct (or current) human labor, that makes it alone capable of generating surplus value, you need to spell out what that is. Or else try to say why it is not necessary to give an answer to this question.

Personally, I don’t know of any good arguments against any of the principles or propositions above. I don’t say there definitely aren’t any! Only that I don’t know of any, and haven’t been able to think of any.

I feel very confident about most of these propositions, and it would take a lot to convince me they are fundamentally wrong. Proposition 6), however, is a departure and more of a hypothesis. I am fairly confident that it is true, but I am willing to entertain alternative hypotheses, if others can come up with them. Moreover, the best case against any hypothesis is usually made by presenting a superior hypothesis. If you think my propositions 6) and 7) are wrong, what would you replace them with?


3.   Socially necessary labor times & competition.

In your letter you state that


Socially necessary labor times is [sic] constantly changing perforce of the competition of capitals inherent in capitalism. If a particular enterprise comes up with a more efficient way of making a better mouse trap by use of more advanced machinery or productive technique it will make extra profit until the other capitals catch-up. But that short term extra profitability comes, for Marx, out of the total surplus value aggregated by the whole of the bourgeoisie. What comes of this temporary advantage by one particular capital is that other capitals adopt this machinery/productive technique (the hammer vs. the rock) in order to avoid the pain of extinction in competition. The amount of socially necessary labor time to produce that use value is cheapened in the process as the new standard becomes adopted industry wide. Your exposition confuses this temporary super-profitability of particular capitals with surplus value. You attribute such surplus profit to the machine rather than to the running contradiction between older and newer standards of necessary labor time. You simultaneously confuse profit with value (and surplus value with super profit).


There is some meat to chew on here.

First of all, I agree that the introduction of machinery, or its improvement, or the introduction of better production techniques, create new standards of socially necessary labor time for the production of commodities that can be made via the use of that new machinery or improved technique. New labor becomes more productive in this way.

But I maintain that one of the principal means by which this happens is that new labor is able to more effectively make use of past labor. Thus if a new type of machine is built, the labor that went into making that machine may henceforth be used multiple times, over and over. If a machine is improved, then the labor that went to make the improved machine is more productive when that machine is used later on. Or if the machine is improved by making it last longer, then the labor that went into making the machine can be reused more times until the machine wears out and needs to be replaced.

So we already see that far from being an objection to my theory, the fact that the socially necessary labor time necessary to produce commodities falls (with improved machinery, etc.) is quite compatible with my theory, a necessary consequence of it, and is even assumed by it. Anything that allows the reuse of past labor more times than before, or allows the more effective use of past labor, will of course result in the reduction of the socially necessary labor time required to produce the end commodities in question.

Creating a machine in the first place means allowing past labor to be used over and over again in the present and the future. This is why machines must be said to contribute surplus value to the final output commodity. Machines allow us to use past labor in the current production process. But if we are now using past labor alongside current labor, then we must add both together to determine the total socially necessary labor time that is being added to the current output commodity. The worker who uses the machine contributes through his/her labor some of the surplus value which gets incorporated into the output commodity, but the rest comes (indirectly) from the labor of the workers who built the machine in the first place.

When a machine is improved, then a greater proportion of the surplus value incorporated in the final output commodity comes from the machine rather than from the worker who uses the machine. If an "absolutely automatic" machine were ever built (an android or otherwise), one that never again required any human tending or control, then it would be contributing 100% of the surplus value to the output commodities it produced. (However, if an android used a milling machine, then part of the surplus value in the output commodities would come from the milling machine, and part from the android. In other words, in actuality most commodities are produced by multiple machines, and—at present—multiple human workers.)

But while the proportion of surplus value contributed by an improved machine to the final output commodity will increase (relative to that from the worker using the machine), it is quite possible—and even typical—that the actual amount of surplus value the machine contributes per output unit might fall—at least once competition has its effect. (Of course this means the amount of surplus value contributed per output unit by the worker using the machine must fall too.) Thus suppose that individual new widget machines across the entire widget industry each produce widgets with much less human labor required in tending them, but that competition drives down the selling price so that there is only a fraction of the profit in each widget as before. This lower profit per widget implies less total surplus value per widget. If the widget market remains constant (in terms of units sold per annum), this also implies less total surplus value for all the widgets produced after the universal adoption of the widget machines.

But we see here that while improved machines and improved technique, together with the competition of many capitals, can in fact lower the amount of surplus value in the output commodities, this in no way affects the basic question at issue—whether machines also contribute to what surplus value there still is in each output commodity.

My argument in no way relies on the temporary extra profit that one company may derive by being the first to introduce some new machinery. Indeed, my argument is just the same whether we are talking about a competitive industry or not. If there is competition then the amount of surplus value per commodity will likely fall for the commodities produced with improved machinery or technique, but the fact that the machinery contributes to the surplus value incorporated in the output commodity is not affected by this. (Indeed, as I mentioned above, the proportion of the surplus value contributed by the improved machinery to the output product actually increases in any event!)

So, Frank, if you think you are on to something here, you need to spell it all out in much greater detail before I will agree that it damages my position in any way.


4.   Ultra-imperialism, nonproductive labor, etc.

You continue:


Note here that failure to take into account the centrality of competition of capitals also plays a role in your subsequent errors (in my opinion) regarding crisis theory. This "oversight" is most fundamentally what is erroneous in Kautsky's "economic" argument that capitalism can and has gone on to an "ultra" stage.

Once one gets beyond baby-talk Marxist pedagogy one begins to realize that surplus value created for the whole social formation by productive labor is distributed all over the place. For example, we know that there are many "costs of circulation" in the system. These costs ultimately are paid for out of surplus value. To make this point all the more stark consider the situation of workers who don't create commodities but who are nonetheless vital to their circulation [and the concomitant realization of profits by their employers and to the completion of the circuits that are necessary to capitalist economy].

Proletarians employed in such circulation activities are paid out of the surplus value created by "productive" labor elsewhere in the economy. Even though these workers are paid a value only equivalent to the cost of their reproduction as workers, that cost is itself paid out of the surplus value created by productive laborers. This is Marx's view. Also paid out of surplus value created elsewhere is the remuneration (viz.—profits) of capitalists who employ these "nonproductive" workers. Please think about this in your argument.

This broad distribution of surplus value does not go one-on-one with profitability but the gravity of particular profitabilities does have an effect on the distribution of surplus value within an economy. Marx is quite right in making central how this works as a whole rather than making it a question of how a particular worker or group of workers gets "screwed out of" the value they create. Your argument would be better informed if it took this viewpoint into consideration.


All of the above mystifies me; not because I disagree with most of these points, but because you seem to think I do. I just don’t see how any of this bears on my argument about machines being able to contribute to surplus value.

My version of crisis theory has nothing to do with Kautsky’s theory of ultra-imperialism, which I have never supported in any way. (Indeed, I’ve criticized those like Alejandro Montaño, who argued that "California can’t nuke New York"—thus implying that on an international level too there could not possibly be a nuclear war because of the supposed super-tight integration of the global economy.)

However, I have argued, and still think, that even state capitalism of the Soviet revisionist type ("one big corporation"), is subject to economic crises. And I did say that even if (contrary to what is actually possible) there was only one giant capitalist corporation that ran the whole world, there would still be economic crises (as long as there was still a capitalist labor market—hired workers). In short I do claim that the basic reason for crises has little to do with there being "many capitals", but rather something that inevitably develops out of the exploitation of labor under capitalist production (the extraction of surplus value, or in other words the contradiction between the social nature of production and private expropriation). And this remains just as true whether or not past labor (in the form of machines) also contributes to surplus value. But I think I should avoid getting any further into this huge and rather tangential subject of crisis theory at the moment.

I don’t disagree with what you said about productive vs. unproductive labor, or about any other way that surplus value gets redistributed in ways that are not initially obvious. I have no problem with any of that. But I don’t see its relevance to the present discussion. What we are talking about is not how surplus value gets redistributed, but how it gets created in the capitalist production process (and specifically the role of machines in this).

Again, if you think your argument here is damaging to my position, you need to bring out much more clearly how that is so.


5.   Confusion of use value and exchange value.

You continue:


Your thesis makes the temporary advantage of the innovative entrepreneur into an element of the system creating surplus value. It is not. The super-profit of that capitalist comes out of the totality of surplus value created by this dynamic system of capitalism, not out of "surplus value" created by the new machines. Since such innovation is inherent in capitalist competition, it is part of the process, not an exception wherein machinery creates surplus value.

You ask to be told of what one might call embarrassment factors. What I find most pertinent and egregious in this regard is the [that?] you confuse (1) profits with surplus value; and (2) use value with exchange value. Your version of Marxist economics is frightfully wrong and wrong headed in both these matters. For example, please look over how your essay says "use value" when it is actually attempting to talk about exchange value.


The first paragraph here [i.e., your point (1) in the second paragraph] seems to be a repetition of the first argument which I already dealt with earlier. I don’t see that I have done what you say I have—in any way! If you still think my argument amounts to this (despite the way it seems to me), you need to demonstrate that.

With regard to your point (2), about the confusion of use value and exchange value, I have to admit there are many cases of this in my first essay. In looking back over it I discovered for example the following sentence in section 4: "Marx’s version of the LTV claims that only human labor can create surplus value (i.e., use value beyond that which is used up in the production process), and specifically that machinery, no matter how complicated and sophisticated it is, cannot do this." That is indeed a wrong definition of surplus value. Surplus value is not "use value" beyond that which is used up, but rather what Marx usually just called "value" beyond that which is used up; that is, labor value which contributes to the final exchange value. (It is not the entire exchange value of the commodity, of course, since some of that exchange value is also transferred from the raw materials and tools and machinery used up—both Marx and I agree on that much.) Of course this new surplus value, this new net value (as part of the total exchange value of the commodity), is only possible because the production process created a new use value (i.e., a new commodity for sale which has an actual use—at least to somebody). Thus the labor process does create new use values, and one can even say that the creation of surplus value is only possible because of the simultaneous creation of new use value. But nevertheless I was being very sloppy here in saying that surplus value itself is "use value" or a part of it. For the capitalist, the entire worth of the item produced lies in its exchange value, not its use value.

Notice, however, that if I correct my erroneous statement by removing the word ‘use’ from the brief definition of surplus value, it in no way affects the validity of my basic argument about machines being able to contribute to surplus value (via the reuse of past labor). This correction merely makes what I was trying to say more coherent, and thus more plausible, as far as I can see.

I found many more examples where I referred to "use value" when I should have said either just "value" or "exchange value". In fact, in most of the places where I referred to "use value" the word ‘use’ should simply be crossed out. One of the worst examples of this use value/exchange value confusion is the following sentence from section 8 of my essay: "Consequently there needed to be a means of systematically deriving actual prices (or ‘exchange values’) from use values, or put the other way around, a method of ‘transforming’ use values into prices." That sentence should be completely rewritten as: "Consequently there needed to be a means of systematically deriving actual prices from labor-equivalent values, or put the other way around, a method of ‘transforming’ values into prices." Notice, however, that once again correcting this terrible sloppiness in no way undercuts my argument about machines contributing to surplus value.

In addition to finding 9 or 10 cases of this use/exchange confusion, in rereading my first essay I also found several places where I incorrectly referred to labor power when I should have said just labor, and a variety of other slips and embarrassments. By way of partial explanation, I was busy busting my brain, trying to think out some other issues in a way I never had before, and that essay was pretty much a record of my thought process—with all the sloppiness one expects to find in actual human thought processes before they later get tidied up. Of course I can’t justify posting that essay or sending it to you without looking it over more carefully. My deep apologies!

I can readily see how such great sloppiness in my original review/essay might have led you to misunderstand even what the exact theory that I was defending was, and to suggest that my theory was a result of confusion over use value/exchange value, and the like. But I have now corrected all these confusions (or at least the ones I’ve spotted so far), and it seems to me that this has only clarified my meaning and strengthening my basic argument.

Please take a look at my revised version and let me know if you think there are places where I have still confused use value and exchange value, or made any other gross mistakes of that sort. And especially point out any cases where you think my version of the LTV depends in any essential way on such confusions. Because I just don’t see that it does.


6.   Say’s "Law".

You state that:


I think you have here again confused two points. Say is wrong because he thinks that capitalism is perpetually regenerating and ever more enhancing of the public interest aside from exogenous factors. You take this to vitiate the point that both Marx and Lenin hold: that generally the productive circuit of capital is principal over the other circuits of capital.


No I don’t. I assume you are talking about the M-C-M' circuit (or M-C-C'-M', as many people prefer) being principal over the C-M-C' circuit, as far as understanding the essence of capitalist production, the source of surplus value and profit, the fundamental reason for crises, and the like. I agree with Marx completely about all that. (In fact this was exactly Marx’s central argument against Say’s Law.) Or do you mean something else here?

I am curious as to what you think Say’s "Law" or "Principle" is, exactly? Did you mean your phrase "capitalism is perpetually regenerating and ever more enhancing of the public interest aside from exogenous factors" to be a statement or definition of Say’s "Law"?

The usual and most concise statement of Say’s "Law" is that "supply creates its own demand". [See the Routledge Dictionary of Economics (1995), or just about any other contemporary reference.] This is much more specific than simply saying "capitalism is perpetually regenerating"—which can be taken to mean a variety of things (a few of which are even true).

The basic issue that Say’s "Law" is addressing is whether or not overproduction is possible, or—as it used to be expressed back then—whether or not "gluts" are possible. Say unequivocally said no, gluts are not possible (except perhaps for exogenous factors—and Say was loath to admit even that possibility) because the very process of production creates not only the commodities to be sold, but equivalent wealth in other hands which can and will be exchanged for it. Or, more concisely once again, that supply creates its own demand.

Marx ridiculed Say, Ricardo, James Mill, John Stuart Mill, and all the others who accepted the validity of this principle or variations of it. He said, for example,


The conception (which really belongs to [James] Mill), adopted by Ricardo from the tedious Say (and to which we shall return when we discuss that miserable individual), that overproduction is not possible, is based on the proposition that products are exchanged against products, or as Mill put it, on the "metaphysical equilibrium of sellers and buyers", and this led to [the conclusion] that demand is determined only by production, or also that demand and supply are identical. The same proposition exists also in the form, which Ricardo liked particularly, that any amount of capital can be employed productively in any country. [Theories of Surplus-Value, Part II, (Moscow: 1968), p. 493. Emphasis in original.]


Thus the erroneous conception of Say, Ricardo, and the rest, is in fact based on the false assumption that the C-M-C' circuit is the principal circuit of capital. And this is indeed one of the reasons why Marx emphasized the importance of focusing on the M-C-M' circuit instead. As I said, I wholeheartedly agree with Marx on all of this.

But most of Marx’s criticism of Say’s Principle (as Marx tends to refer to it) occurs in Theories of Surplus Value, especially volume 2. (This is part of what Marx intended to be the 4th volume of Capital.) Unfortunately, TSV, vol. II was not published—even in German—until 1905. (Vol. I was published in 1904, and vol. III was not published until 1910.) Prior to that time most other Marxist writers on political economy—including Lenin—did not have access to Marx’s criticism of Say and his principle and of those like Ricardo who borrowed it. This was a serious problem since I really believe that people who don’t carefully read Theories of Surplus Value generally tend to misunderstand Marx’s view on many very basic points.

And specifically I think Lenin got a very skewed understanding of Marx’s political economy because of this, which is reflected in his early writings on political economy. (See below.) Worse yet, the Communist movement since Lenin’s day has often followed Lenin in his early misunderstandings. In particular, the RCP has done so, and continues to do so to this day. One illustration of this is the statement in AID that "The expansion of capital requires the continual perfecting of the division of labor and generates its own demand and markets." [P. 260; my emphasis.] The last six words here are nothing else than an affirmation of Say’s erroneous "Law". This is not a one-sentence aberration. This principle, Say’s "Law", is the initial assumption that leads the RCP into insisting that capitalist crises can only arise from the anarchy of many capitals. Thus the crisis theory in AID is fundamentally wrong, even though it is true that the various forms of the anarchy of production under capitalism can in fact aggravate the much more basic problem of overproduction of capital. Since the main theme of AID is its crisis theory, and the application of that theory to the U.S. of the 1980s, the book as a whole is extremely damaged by this initial assumption of Say’s "Law". Moreover, this assumption has led to many other errors, such as a complete failure to understand the political economy of the Soviet Union in the revisionist period. Lotta has insisted, for example, that Soviet economic problems were also fundamentally the result of the "hidden" anarchy of many capitals there.

Your letter continues:


I think it is time for you to fess up and cite which "early works" of Lenin are you talking about and what particular quotes you find that support your view that he clung to Say's Law up to some point (again, please specify where, when and how).


Well, I did mention the most important place where Lenin supported Say’s "Law" in a letter to you on 2/25/01, namely Lenin’s 1897 pamphlet A Characterization of Economic Romanticism. Sometime after that I wrote up the specific citations in either a letter to somebody or else an essay, but I haven’t been able to locate it again. (I might not have sent you a copy at the time because of your lack of an Internet connection.)

Nowhere in his writings, I believe, does Lenin mention "Say’s Law" by name. (At least there is no entry for this in the Subject Index volume to Lenin’s Collected Works, and I don’t recall seeing any such passage myself while reading Lenin.) But there are references to Say himself in volumes 2, 3 and 4. Of course it is possible to support Say’s "Law" without naming Say, as the RCP did in AID. Anyway, here are some of the passages I managed to find this time around:

In A Characterization of Economic Romanticism (1897), in LCW vol. 2:


1)   "The more rapid the process of accumulation, i.e., the excess of production over consumption, the better, taught the classical economists, who, though they were not clear about the process of the social production of capital, and though they were unable to free themselves from Adam Smith's mistaken view that the social product consists of two parts, nevertheless advanced the perfectly correct idea that production creates a market for itself and itself determines consumption." [LCW 2:148] The last part of this, which I have put into bold type, is a clear affirmation of Say's "Law", and note that Lenin calls it "perfectly correct" when it fact it is perfectly incorrect according to Marx.

2)   In the next paragraph, Lenin repeats: "Further, the failure to understand that production creates a market for itself leads to the doctrine that surplus-value cannot be realized." [LCW 2:148; bold added.] Again, this is a rather clear affirmation of what we now call "Say's Law".

3)   Several pages later, Lenin says: "His [Sismondi's] assertion that rapid accumulation leads to disaster is absolutely wrong and is solely the result of his failure to understand accumulation, as are his repeated statements and demands that production must not outstrip consumption, because consumption determines production. Actually, the very opposite is the case, and Sismondi simply turns his back on reality in its specific, historically determined form and substitutes petty-bourgeois moralizing for an analysis." [LCW 2:156, bold added.] This example may not be quite as clear as the two above, but to my mind it is another at least implicit affirmation of Say's "Law". I.e., he denies that "consumption determines production" (in which he is correct), but seems to argue that the opposite is the case, that "production determines consumption", which presumably means that everything which gets produced must be consumed. This just ain't so, no matter who says it.


Lenin continued on this general theme in other essays but in a much more implicit fashion. (See for example "A Note on the Question of the Market Theory" in vol. 4, esp. pp. 58-60.) Because of his remarks of the sort quoted above from A Characterization of Economic Romanticism, the Narodniks he had been attacking found an opening for a counter-attack. They accused him of adopting the views of Say and Ricardo against Marx. In an important essay on this whole general topic, "Once More on the Theory of Realization" (in LCW vol. 4), Lenin responds to this as follows:


8. Struve says that I ascribed to Marx the bourgeois-apologetic theory of Say-Ricardo, the theory of harmony between production and consumption, a theory that is in howling contradiction to Marx's theory of the evolution and eventual disappearance of capitalism; that therefore, my "perfectly correct argument" that Marx, in both the second and third volumes, stressed the contradiction, inherent in capitalism, between the unlimited expansion of production and the limited consumption on the part of the masses of the people, "jettisons that theory of realization … whose defender" I am "in other cases."

This statement of Struve is likewise untrue and derives likewise from the above-mentioned misunderstanding to which he has become subject.

Whence comes Struve's assumption that I do not understand the theory of realization as an analysis of the process of reproduction and circulation of the aggregate social capital, but as the theory which says only that products are exchanged for products, a theory which preaches the harmony of production and consumption? … [LCW 4:83]


So we see here that the Narodniks were using the exact same charge against Lenin in this case as you were using against me, Frank, and that Marx used against Ricardo and Say—namely that Lenin was thinking in terms of C-M-C' rather than M-C-M'. The charge against me was invalid; it is unclear to me if this specific Narodnik charge against Lenin was valid or not. (It was certainly true of Say and Ricardo themselves.) At any rate Lenin denied it. But though he thinks the Narodniks have misconstrued his (and Marx’s) viewpoint, he still does not change what he said. While trying to disassociate himself from Say and Ricardo, he still does not explicitly disassociate himself from what we now call Say’s "Law".

I need to briefly mention how Lenin got himself into this corner in the first place. Lenin and other Marxists were in a big theoretical struggle with the Narodniks. The ironic situation here was that Lenin was defending capitalism against the (mostly) ignorant attacks of the populists against it. They denied that capitalism was even possible in Russia. They claimed that capitalism was an "artificial" creation, that it required foreign markets (which Russia didn’t have), and—more to the point—that capitalism could thus not possibly develop in Russia because it would find itself in an intractable overproduction crisis right from the very outset. On this last point they were claiming Marx’s support, and this is what stirred Lenin’s greatest ire.

Lenin wrote a whole book, and numerous supplementary articles to show just how wrong the Narodniks were. He pointed out that, contrary to what the Narodniks thought, capitalism had already developed to a considerable degree in Russia, and that it was fast developing further. He argued that this was a progressive development; that it was leading to the development of a revolutionary proletariat; and that the Narodnik conception of going directly to socialism via the declining peasant "communes" was a hopelessly utopian idea.

Lenin also correctly argued that, while Russian capitalism would of course suffer periodic crises, it could still continue to develop overall. He argued that while it was correct that the development of production under capitalism goes on at a faster rate than the development of consumption by the masses, there is no necessary problem here because the capitalists use the remaining surplus value to expand the productive forces. He admitted that there is an absurdity here, a real contradiction. In his book The Development of Capitalism in Russia (1899) he stated:


The development of production (and, consequently, of the home market) chiefly on account of means of production seems paradoxical and undoubtedly constitutes a contradiction. It is real "production as an end in itself"—the expansion of production without a corresponding expansion of consumption. But it is a contradiction not of doctrine, but of actual life; it is the sort of contradiction that corresponds to the very nature of capitalism and to the other contradictions of this system of social economy. [LCW 3:56]


In all this, Lenin was correct. But in his haste to shoot down the Narodniks, and because his own knowledge of Marx was incomplete (though through no fault of his own), he went beyond this, and in doing so went a bit too far. Lenin, at least in 1897, and like the RCP today, thought that what we now call Say’s "Law" was valid, and he used that invalid argument as a supplement to his many valid arguments against the Narodniks.

But as with the RCP today, this agreement with Say’s "Law" had a lot of additional negative ramifications. It meant that Lenin focused way too much on the anarchy of production when explaining economic crises, for example, and did not truly understand that the real contradiction that he later admitted existed in the capitalist’s "production as an end in itself" was itself the most fundamental cause of crises. In short, "production as an end in itself" only works for a while. Then, when it becomes obvious even to the capitalists that they have built up the productive forces way beyond what can really be justified, there is a crisis, and the necessity to write off (or otherwise destroy) a large part of that excess capital. This clears the ground for a new cycle of accumulation.

If Say’s "Law" were just some abstract error, with no ramifications elsewhere, then it wouldn’t be worth bothering about. But in fact a person’s attitude towards Say’s "Law" is one of the key theoretical starting points in political economy. If somebody gets that wrong, they inevitably get a whole lot more wrong.


7.   Wrap-up.

Well, it’s well past time to wrap up this over-long letter. You state that "All the subsequent economic history of the 20th century bears him [Lenin] out—crisis has principally been driven by the overproduction of capital." With that conclusion I have not the slightest disagreement. Do you think I somehow do?

I should note that Lenin’s "Imperialism", which strongly puts forward this view, comes from Lenin’s later position on political economy, which de-emphasized "anarchy" and more strongly emphasized overproduction, and the tendency toward the intractable nature of overproduction in the imperialist era.

Your final observation is that I should rethink my claim that the restructuring of capital in the imperialist era is principally caused by the physical destruction of capital (you say "commodities", rather than "capital" here; I wouldn’t put it that way). Instead the principal means is still through the forcible readjustment of value relations, you say.

I need to think about this some more. It is clear that a massive amount of such readjustment of value relations of capital happens all the time, and especially during crises. But you yourself hold to the RCP view that the biggest of these crises and (temporary) resolutions occur at nodal points like world wars, and of course at least in World War II there was a colossal physical destruction of capital.

Moreover, the most recent "nodal point", according to the RCP, the fall of the revisionist Soviet Union and its empire, did not see a simultaneous resolution of the overproduction crisis that began in the early 1970s and which not only continues today, but is growing worse. (I disagree with the RCP claim that the existing economic contradictions were mostly resolved circa 1990, and that an essentially new economic spiral began along with a new political spiral. They were misled by the brief "New Economy" boom in the U.S., as well as by their erroneous spiral/conjunction theory itself. See my "Notes on Notes on Political Economy" for more on this.) So in my view there have been three decades of restructuring mostly by the forcible readjustment of value relations, and this has not resolved the long overproduction crisis. Instead, at the present time it shows signs of becoming qualitatively worse. This reinforces my theory that the mere readjustment of value relations is no longer sufficient in the monopoly/imperialist era to fully resolve crises.

One last thing for me. I am anxious that our disagreements over whether machines can contribute to surplus value and other points of theoretical political economy—which are very likely to continue for some time—not drive any wedge between our friendship, or prevent us from collectively working on other more immediately important political tasks. I think you agree that we should be able to debate abstract theoretical questions without any personal animosity developing.

Best wishes,


—Scott (12/8/03)
    (With a bit more editing on 12/9/03.)



Marxist Political Economy Home Page
Scott H.'s Home Page on MASSLINE.ORG
MASSLINE.ORG Home Page