Saving labour or capital?

Does the rate of profit tend to fall? The traditional Marxist argument upon which this theoretical proposition is based is badly flawed, argues Moshe Machover

Comrade Hillel Ticktin has argued that the present global crisis in not due to a fall in the average rate of profit.[1] Moreover, he is also quite sceptical (to say the least) of the claim, made by Andrew Kliman and others, that the average rate of profit has actually displayed in recent times a tendency to fall. I think he is right on both counts. However, I would go much further: I wish to refute the traditional Marxist theoretical argument claiming to prove that the average rate of profit has a long-term tendency to fall.

A faulty chain

That traditional argument consists of the following chain of propositions:

A. Under capitalism, the productivity of labour tends to rise due to technological innovation, motivated by competition.

B. Rising labour productivity due to technological change tends to increase the organic composition of capital.

C. The average rate of profit has an inverse tendency to that of overall (economy-wide) organic composition; other things being equal, a rise in the overall organic composition of capital tends to lower the average rate of profit.

(Actually, the claim that the average rate of profit has an inherent tendency to decline was made before Marx, by David Ricardo. Marx accepted Ricardo’s claim, but thought he could provide a theoretical proof of it using the chain of reasoning just outlined in simplified form. Marx’s original presentation is in chapters 13-15 of volume 3 of Capital.)

I assert that, while propositions A and C are true, proposition B is false; and therefore the entire argument breaks down.

Before I proceed to disprove proposition B, I must explain the terms I will use. First, organic composition. There is no generally agreed definition of this term. Marx himself uses it in several, somewhat different senses, not exactly equivalent to one another. But for the purpose of the present discussion it does not really matter which of these senses is used: proposition B does not hold under any of them. However, to fix ideas, I mean by organic composition (on a given day) the value of invested capital used in or for production during that day (including means of production, buildings, stocks of raw material waiting to be used, half-finished products, and finished products waiting to be sold), divided by the amount of labour (measured, say, in worker-hours) employed during that day in putting that capital in motion.[2] Organic composition can refer to a given firm, or to a given sector, or even to a whole economy. In any case, it is time-dependent: it changes over time.

Second, I just referred to the value of capital goods. By the value of a commodity I mean its exchange value in Marx’s sense of the term: the amount of labour necessary to reproduce it. However, for the purpose of the present discussion it would make little difference if we understand value as price, measured, say, in pounds or dollars (adjusted for inflation - ie, so-called constant price). The truth or otherwise of propositions B and C is not affected. The reason for this is that, while there is no determinate relationship between the price of a single commodity and its value, there is a statistical relationship: if we take large, diverse samples (‘baskets’) of commodities, the ratio between their respective total prices and total values is very nearly constant.[3]

Third, labour productivity is hard to define (and is possibly meaningless) as a global quantity. But it has a clear meaning in relation to each specific type of product: it is the inverse of the value of a unit of this type of product. For example, if the total (direct and indirect) labour time needed to produce a car of a given type is halved, this means that the productivity of labour producing this type of car is doubled.

Productivity and organic composition

Analytically speaking, technological advance, leading to increased productivity of labour, can take place in two ways: saving labour and saving capital.

A labour-saving technological change increases productivity (other things being equal) by reducing the amount of labour used directly in producing each unit of output.

A capital-saving technological change replaces means of production, or raw materials, or other non-labour inputs by less valuable ones. Other things being equal, this increases the productivity of labour because the value of each unit of output consists, in addition to the amount of ‘living’ labour employed directly in producing it, also of the amount of ‘dead’ labour, the value of the material inputs used up (ie, consumed) in producing it.

In reality, these two kinds of technological changes often occur together. It is even possible for the productivity of labour to increase by a technological change that is labour-saving and involves negative capital saving, or is labour-wasting and capital-saving. For example, an increase in the value of the raw material used for a given product can be more than offset by reducing the amount of labour per unit of output. Or, conversely, an increase in the amount of labour per unit can be more than offset by using much cheaper raw materials. Labour productivity will increase if the net result is a reduction of the total amount of living and dead labour embedded in each unit of output.

But for the purpose of analysis it is convenient to separate the two kinds of technological change. This is why I keep saying “other things being equal”.

Now observe that, other things being equal, a labour-saving technological change tends to increase the organic composition of the capital of the firm in which the change takes place. This is because such a change increases the ratio between the value of invested capital and the amount of employed labour. The opposite holds for capital-saving technological change: it tends to reduce the above-mentioned ratio, and hence the organic composition of the capital of the given firm.

In the early stages of the agricultural and industrial revolutions, technological changes were predominantly labour-saving. Think, for example, of the introduction of agricultural machinery, or of the replacement of the hand loom by the power loom. These obviously resulted in an increase of organic composition of capital in these branches of production, and hence also contributed to increasing the overall organic composition in the entire economy. It was quite natural to assume in the 19th century that technological change would always be of this type. Marx took it for granted, and hence believed in the truth of proposition B.

But in fact not all technological change is of this kind. As an illustration, let us take one of the oldest branches of capitalist production, where mass production was pioneered: printing. This industry has undergone at least two major waves of technological innovation. For several centuries following the invention of printing with movable types, printing was done sheet by sheet using a hand press: the press had to be tightened and loosened for each sheet of paper. Typesetting was done manually, each individual letter-type placed by hand in a printer’s form. Then came a series of changes: the hand press was replaced by a fast, mechanical press into which the paper was fed automatically; and most typesetting was now done by linotype (and later monotype) machines, using hot lead. These machines (as well as the lead needed for machine typesetting) were very expensive, but they made possible a huge saving in labour. Undoubtedly, these technological changes resulted in a considerable increase in the organic composition of capital in the printing industry, and hence contributed to its increase in the entire economy.

However, quite recently printing has been revolutionised by capital-saving changes. The Weekly Worker is typeset on a desk computer - very much cheaper than a hot-metal typesetting machine (plus the lead needed for it). Printing is done on electronic photo-printers, which are considerably cheaper than the old printing machines, especially where not very large print runs are concerned. For a small number of copies, most of us use domestic desk printers, which are very much cheaper than the machines needed to produce a similar quantity and quality of output using the technology of yesterday. These recent changes have involved little if any saving in direct labour per unit of output; the saving is predominantly in capital (indirect labour).

Thus the effect of increased productivity of labour in a given branch of production on the organic composition in that branch is in principle indeterminate: the latter can go up or down, or be unchanged - depending on whether the new technology is labour-saving or capital-saving. And there is no law that says that technological change must always, or for the most part, be labour-saving.

Indirect effects

In addition, there are technological changes that tend to reduce the organic composition - not necessarily in the industries in which these changes take place, but in other branches of production, and hence in the economy as a whole. This kind of change was in fact mentioned in passing by Hillel Ticktin in his article.

I am referring, of course, to technological changes that increase labour productivity in what Marx called ‘department one’, which produces means of production (including raw materials). Suppose that labour productivity increases (as it always tends to do) in the industry that produces printing machines. Whether this is achieved by labour-saving or capital-saving in this industry, it will in any case tend to reduce the organic composition of capital in the printing industry. Similarly, an increase in the productivity of labour in steel production, no matter how it is achieved, will reduce the organic composition in industries that use steel as input; for example, the car manufacturing industry. Consequently, sufficiently large increases in labour productivity in department one, no matter how they are achieved, will tend to reduce the overall organic composition in the entire economy. Therefore, even if all technological change were labour-saving, overall organic composition can still go down, provided the productivity of labour in department one increases sufficiently fast.

Since it is impossible to predict the course of future technological change - whether it will be mostly labour-saving or capital-saving, and how much of it will take place in department one, proposition B, and hence the traditional Marxist argument for the historical tendency of the average rate of profit to decline, is untenable.

Of course, this does not necessarily mean that the average rate of profit does not in fact tend to decline in the long run. This is an empirical question. My feeling is that, based on existing data, the evidence for it is quite weak. But it is quite possible that such a tendency will emerge in future. It is even conceivable that there is some theoretical argument pointing at such a future tendency; but, if so, it is not the traditional argument, which is fallacious, as I have shown.


1. ‘The theory of capitalist disintegration’ Weekly Worker September 8.

2. Another definition of this term takes, instead of the amount of labour, the wages paid for that labour. But this will not make any difference to the present discussion, provided we assume that the wage rate does not change very much in value terms.

3. For details, see E Farjoun, M Machover Laws of chaos London 1983. This book also contains a detailed technical analysis of the behaviour of the rate of profit.