MARX BOOK DAY FOUR

 Modern analysis has made it possible to determine Ricardo’s theory of natural price works in System C. Does it work in every economy engaged in reproducing itself cyclically in the manner of System C? As a test of the theory, let us suppose that a more efficient technique has been discovered by the farmers in System C for producing corn. Previously, 100 units of labour were required to be combined with 2 units of corn and 16 units of iron to produce 300 units of corn. Henceforward, the same capital stock can produce the same output when combined with only 50 units of labour. (Once again, I stress that technological realism plays no role in this discussion.) The result, which we shall label System D, is as given in Table 4, assuming that subsistence remains unchanged.

 

System D

                                    Labour Input   Corn Input  Iron Input  Books Input      Output

Labour Sector                                          32               16               0                   160

Corn Sector                      50                     2                16               0                   300

Iron Sector                       90                      9               12               0                     90

Books Sector                   20                      1                 2                2                    40

Total Input                    160                    44               46               2

 

In order to calculate the labour values of corn, iron, and books in System D, we must form new labour value equations, for the conditions of production in the corn sector have now changed. 

 

Following the procedure by which we formulated equations (2) through (4), we obtain:

 

50 + 2λc + 16λi + 0λb  = 300λc, (2’) 

90 + 9λc + 12λi + 0λb  = 90λi (3’) 

 

and 

 

20 + λc + 2λi + 2λb  = 40λb. (4’) 

 

When we solve equations (2’) through (4’) in order to obtain the labour values of corn, iron, and theology books in System D, using the same algebraic manipulations that gave us the labour values in System C, we find that: 

 

λc ≅ 2312 

 

λi ≅ 1.1811 

 

λb ≅ 5946 

 

The change in the technology of corn production has altered not only the labour value of corn, but the labour value as well of iron and books. In each case, the new value is lower than the old one. A moment’s reflection will show why this is so. The “labour value” of corn is simply the quantity of labour directly or indirectly required to produce a unit of corn. If a new technique of corn production is introduced which lowers the quantity of some commodity required as input, then clearly less labour will be required for the production of the corn output, because part of the labour required is precisely the labour embodied in the inputs. But with less labour required for the production of the same output of corn, the labour embodied in each unit of corn output will be smaller. Now, corn is required as an input into the iron and books industries. Hence, in those industries as well, less labour will be required indirectly for the same output, and so the labour value of iron and books will fall as well. In this way, a change in the conditions (or “facility”) of production in one industry radiates throughout the entire economy. 

 

To calculate the relative prices in System D, we must form new price equations, reflecting the altered conditions of production in the corn industry. Following the same procedure as before, with corn chosen as “money” or numeraire, we have: 

 

(50w + 2 + 16pi)(l + π) = 300, (5*) 

 

(90w + 9 + 12pi)( l + π) = 90pi, (6*) 

 

and 

 

(20w + 1 + 2pi  + 2pb)(l + π) = 40pb. (7*) 

 

When we attempt to calculate the relative prices and the profit rate in System D, we encounter a disturbing problem. In manipulating equations (5*) through (7*), we find that the wage, w, does not drop out of the calculation. In fact, the prices and the profit rate do not become determinate until we specify the wage. Following the same procedure used above, we will fix the real wage at .2 units of corn and .1 units of iron. If we substitute into our equations the expression w = .2pc + .1 pi and solve for the values of the variables, we obtain: 

 

pc = 1 

 

pi  ≅ 3.869 

 

pb  ≅ 1.962 

 

w ≅  .5869 

 

π ≅ 2.217 

 

We can check to, see whether Ricardo’s theory of natural price holds for System D; as it did for System C, by substituting the labour values and prices just obtained into equation (1). When we do, we find that: 

 

(pc/λc) ≅ 4.325 

 

(Pi/λi) ≅ 3.276 

 

where has Ricardo’s theory gone wrong?  Let us follow Ricardo’s own thinking in this matter, for he knew full well that prices are not always proportional to labour values, and he even saw essentially why they are not. 

 

As we have seen, in the heading of section three of the chapter on value in Principles, Ricardo announces the theory that the natural prices of commodities are determined by the labour indirectly as well as directly bestowed on them in the process of their production. Immediately following section three, Ricardo acknowledges the existence of other factors that affect natural price. The headings of sections four and five tell the story: 

 

Section IV 

 

The principle that the quantity of labour bestowed on the production of commodities regulates their relative value, considerably modified by the employment of machinery and other fixed and durable capital. 

 

Section V 

 

The principle that value does not vary with the rise or fall of wages, modified also by the unequal durability of capital, and by the unequal rapidity with which it is returned to its employer. 

 

Ricardo states the point exactly correctly when he says: “If men employed no machinery in production but labour only, and were all the same length of time before they brought their commodities to market, the exchangeable value of their goods would be precisely in proportion to the quantity of labour employed.”

 

 The problem, to put the point abstractly, is that the production of commodities requires time in addition to nature and human labour. In the calculation of the labour values – the quantities of labour directly and indirectly required for production – no account is taken of the time that must elapse between the moment when the bestowing of labour commences and the moment when the commodity appears on the market, available for sale. But in the rational calculations of prudent entrepreneurs, time plays a central role. A rate of return on capital invested is a percentage increment per period of time elapsed. If the ruling profit rate is 10 percent, the prudent capitalist will seek to earn 10 percent per annum on the value of capital invested, which is 21 percent in two years, 33.1 percent in three years, and so forth. Those commodities produced in a more “roundabout” manner, to use Böhm-Bawerk’s felicitous phrase, must sell at prices that deviate more significantly from their labour values in order for their producers to earn the society-wide rate of profit. 

 

Ricardo was painfully aware of the fact that the failure of prices to be proportional to labour values undermined his entire economic theory. So long as prices are proportional to labour values, and so long therefore as the distributional variables, w and π, are independent of the prices of commodities, it is possible unambiguously to analyse the distribution of the social surplus. We can separate out the effects on the economy of a change in technology, which will show up as a change in relative prices, and a change in the distribution of the surplus, which will show up as a rise of the wage and a fall of the profit rate, or vice versa. Since labour values are derived directly from the technical conditions of production, they are unaffected by the way in which the surplus is divided between workers and capitalists – which is to say, they are independent of the values taken on by w and π. If prices are proportional to these labour values, then obviously prices as well will be independent of the wage and profit rate. 

 

In a system of this sort, a rise in the wage is unambiguously an improvement in the real (i.e., physical) income of the workers. Changes in relative prices will be a consequence solely of changes in the facility of production of commodities. Thus, it will be possible always to distinguish between objective or technical changes on the one hand and social or distributional changes on the other. To put the same point differently, it will always be possible to distinguish between a redistribution of the existing physical surplus and a technological change that alters the size and composition of the surplus. 

 

But as our analysis of System D reveals, this lovely idea of Ricardo’s is doomed to failure, as he himself knew. Although there is a good deal more of importance in Ricardo’s political economy, including his theory of rent and his analysis of the determination of the real wage, from a strictly theoretical standpoint, Ricardo’s labour theory of natural price comes to a dead halt right here. Just as Adam Smith was unable theoretically to extend his correct analysis of the early and rude state to the case in which the accumulation of stock and the appropriation of land has taken place, so Ricardo is unable to provide an adequate analysis of the deviation of natural prices from labour values as a consequence of unequal times that elapse between the bestowal of labour on the production of commodities and their realisation, or sale, in the market. 

 

What is it about some systems that makes their natural prices independent of the distributional variables and proportional to the labour values? Why is it that prices are proportional to labour values in System C, but are not in System D and so many other systems besides? Is there something peculiar about the structure of System C that yields this proportionality? Because we are dealing at so high a level of abstraction, in which our only data are the quantities of labour and commodities required as inputs in each line of production, the answer can only lie in some quantitative feature of the relative proportions in which the different inputs are combined. 

 

In the corn sector of System C, 100 units of labour are bestowed directly on the production of the corn output. The non-labour inputs consist of 2 units of corn and 16 units of iron. These embody 2(.4) + 16(1.2) = 20 units of labour, which is thus indirectly required for production. The ratio of labour directly required to labour indirectly required is 5:1. Analogous calculations reveal that the same ratio, 5:1, obtains in the iron and theology books sectors as well. Intuitively, we might guess that when the ratio of labour directly required to labour indirectly required is the same in all lines of production, then variations in the wage rate, and consequently in the profit rate, will affect all commodities proportionally, and hence will not alter the price of one commodity relative to another. An increase in the profit rate will, of course, work itself out more heavily the farther back in the past we carry our calculations, but since the proportion of labour directly to labour indirectly bestowed is the same for all commodities, no differential advantage will accrue to one commodity relative to another.

 

In System D,  we find  that the ratios of labour directly required to labour indirectly required in the several sectors are widely divergent. In System D, the ratio of direct labour inputs to embodied labour in the corn sector is roughly 2.58 to 1, whereas the ratio in the iron sector is 5.537 to 1.  The fact is that our general theoretical intuition is correct. In linear models like those we have been examining, so long as there is a positive rate of profit, natural prices are proportional to labour values if and only if the proportion of labour directly required to labour indirectly required is the same in all lines of production.  Ricardo’s labour theory of natural price holds true for all and only those surplus-producing economies with positive rates of profit in which the ratio of labour directly bestowed on commodities in production to labour embodied in non- labour inputs and thereby indirectly bestowed on, or transmitted to, commodities in production is the same in all sectors. 

 

Thus far, our result may appear to be little more than a theoretical curiosity, for the condition that direct to indirect labour ratios be the same in all industries is thoroughly contrary to economic experience, and without apparent significance. Since it is generally the case that agriculture is relatively labour intensive and industry is relatively capital intensive, we would never expect to find an actual economy that even approximated equal ratios of direct and indirect labour across the board. 

 

The result takes on a much deeper historical and theoretical significance however, when we discover, as we shall presently, that it serves as the starting point for Marx’s theoretical investigations. The condition of equality of proportion of direct to indirect labour is arithmetically equivalent to what, in the terminology of Karl Marx, is called “equal organic composition of capital.” Beginning precisely at the point where Ricardo’s theory fails, Marx writes all of volume one of Capital from the point of view of, or on the assumption of, the proportionality of prices to labour values. Only after he has thoroughly explored this theoretical terrain does he go on, in volume three, to examine the general case in which prices deviate from labour values. Why he should choose to adopt this course will turn out to be a key to understanding his theory of capitalist exploitation. 

 

With the formulation of the conditions under which Ricardo’s theory of natural price fails, we have come to the theoretical dead end of his system. Ricardo himself devoted his last several years to unsuccessful attempts to analyse the problem, and in an unfinished essay written during the final weeks of his life, we can see him still turning the puzzle this way and that.  Before turning to Marx, however, we must devote a few pages to the two portions of Ricardo’s system that we have thus far passed over in silence, namely his theory of rent and his theory of the determination of the wage. 

 

                                                Ricardo’s Theory of Rent 

 

Our analytical story began with Adam Smith’s observation that commodities fail to exchange in proportion to the labour directly required for their production once we acknowledge accumulations of capital stock and private appropriation of land. We have now completed our examination of Ricardo’s theoretical attempt to take account of the accumulation of capital, but we have not yet considered how he proposed to handle the fact of privately appropriated land. Strictly speaking, the Ricardian theory of rent is not directly relevant to our story, for the problem of rent is a side issue for Marx. Nevertheless, Ricardo’s solution to the puzzle of rent is one of the analytical high points of the classical period, and is well worth a slight detour. 

 

All commodities are produced, directly or indirectly, by the mixing of human labour with nature. Nature is, as Locke put it, a free gift from God, and so long as human need – or, more accurately, effective market demand – falls short of the abundant fecundity of God’s gift, the part played by nature in production has no economic significance. When fertile land is plentiful, it will make no difference should one man lay claim to a portion of it and seek, by force of arms or writ of law, to exclude others from it. The rest will simply move on to new land, equally fertile, and mix their labour with it. Nor will anyone be willing to pay so much as a farthing for the sheer privilege of working a piece of land, so long as equally good land remains unclaimed. 

 

Indeed, even if all the fertile land has been appropriated, Ricardo supposes, virtually no rent will be paid so long as three conditions are fulfilled: first, no single landholding must comprise so large a share of the whole that some cultivation of it will be required to satisfy existing effective demand (this guarantees that no single landowner has a stranglehold on the market for corn). Second, landlords and entrepreneurs must be perfectly self-interestedly rational (this guarantees that we can make a priori predictions of their behaviour purely on the basis of a calculation of their economic interest). And third, no collusion must take place either among landowners or among entrepreneurs (this eliminates the possibility of monopolies or monopsonies that destroy the effects of market competition). If a landowner were to attempt to charge a rent for the use of his land, competition from other landowners, whose holdings were earning no rent and had no alternative uses, would drive the prevailing rent virtually to zero. 

 

Let us now suppose that through population growth, the demand for food rises until all the available land of the best quality is under cultivation. As demand continues to grow, the market price of corn will rise above its natural price, and the entrepreneurs who have invested in corn production will earn a super-profit. At this point, two things will happen. First, it will become profitable, for the first time, to bring into cultivation less fertile lands, lands which require a greater application of capital (more labour, more fertiliser, more tools) per unit of output. It will be profitable because even though a bushel of corn grown and harvested on this land costs more to produce, still the unnaturally elevated price of corn permits entrepreneurs consigned to the less profitable land to earn at least the going rate of profit, and possibly more. But second, these newly arrived entrepreneurs, seeing the much greater rate of return earned by those farming the original, more fertile land, will offer to pay a rent to the owners of this land for its use, because they can afford to pay a rent and still do better than on the less fertile land, for the use of which they pay nothing. 

 

Competition over time equilibrates the system so that the rent paid on the less fertile land just absorbs the extra return that would otherwise accrue to the entrepreneurs who grow their corn on it. In this way, a single system-wide profit rate is re-established. Four things have changed: more corn is being grown; some of the corn is being grown with a new, less efficient technique; a rent is being paid on the most fertile land; and the price of corn has risen. 

 

In fact – and this is for Ricardo the theoretical point of the entire exercise – the price of corn is determined by the technical conditions (or “facility of production”) on the least fertile land. It is those conditions that determine how much an entrepreneur will be willing to pay in order to shift his production to the more fertile land, and it is thus those conditions that determine how large a rent will be paid for the more fertile land. 

 

Now, in a competitive economy, there can be only one natural price for each commodity, and only one profit rate. By hypothesis, no rent is paid on the less fertile land. Hence, rent plays no role in the price of the corn grown on that land. But since the price charged for that corn must be the single price at which all corn sells, it follows that rent plays no role at all in the determination of the price of corn. And that being so, rent plays no role at all in the determination of any price in the economy! 

 

This is a remarkable and thoroughly counterintuitive conclusion, which merits a bit of reflection. To the entrepreneur who must pay for his seed, his tools, his machines, his labour, and for the use of the land on which he raises his crop, it certainly appears that rent plays a role in the determination of price. The entrepreneur who prices his output without allowing for rental charges will quickly go broke, no matter what David Ricardo says. If rent is not a cost of production, what then is it? 

 

Ricardo’s answer, quite simply, is that rent is a deduction from profits. It is a diversion to the landlords of a portion of the profits earned by the entrepreneur. Adam Smith and the others have the matter exactly backwards. The price of corn is not driven up by the rentals paid to the landlords. No doubt the landlords will extract whatever rent they can, but in a competitive market entrepreneurs will consent to the payment of a rent only if they can thereby gain access to fertile land for the growing of corn which can be sold dear on the market. As Ricardo says in one of the best-known tags of classical political economy, “Corn is not high because a rent is paid, but a rent is paid because corn is high.” 

 

Ricardo’s theory of rent is thus, among other things, an addendum to his answer to the question: Who gets the surplus? In an economy in which arable land is scarce, a portion of the surplus is appropriated by the landlords in the form of the rentals they charge for the use of their land. The market is the mechanism by which they get their rentals, and the size of their share of the surplus is determined by the relationship between the conditions of production on the most fertile land and on land of lesser fertility. 

 

In real terms, what has taken place is a transfer of a portion of the entrepreneurial profit into the hands of the landlords, who collectively hold a monopoly of a scarce resource – fertile land. The money wage has changed, but by an amount just sufficient to permit workers to purchase the same market basket of goods at the new prices. It follows that there is no real opposition of interest between the landed class and the working class. The real conflict is between the landed gentry and the capitalist class. 

 

As the demand for corn grows, it is quite possible that all of the second-quality land will be brought under cultivation without satisfying demand. Even less fertile land will then be cultivated. The even less facile mode of production employed on the least fertile land will result in a still higher price for corn, and this in turn will increase the rent on the most fertile land and introduce for the first time a rent on the land of second quality. The rentals will differ by just enough to maintain a single economy-wide price for corn. The profit rate will of course fall. In this way, a schedule of graduated rents can come into existence on lands of decreasing fertility. Always, there will be no rent paid on the least fertile land, and it will be the conditions of production on that land that will determine the price of corn. 

 

The reason for Ricardo’s fear of a “stationary state” emerges clearly from this analysis. Ricardo, like Smith, was persuaded that economic growth would come only from the productive investment of the revenues of the capitalist class, not from the expenditures of rental income, which he saw as being unproductively consumed in the maintenance of a grand style of manorial living. As population growth increased the demand for food, ever less fertile land would be called into cultivation. Rentals would rise and the profit rate would fall, and a larger and larger share of the social surplus would be transferred to the unproductive consumption of the landlords. The portion of the surplus devoted to new investment would shrink until, at the horrible limit, the entire annual surplus would go for rent, and scarcely enough would remain in profits to encourage entrepreneurs even to undertake simple reproduction.

 

Ricardo’s Theory of Wage Labour

  

Labour, like all other things which are purchased or sold, and which may be increased or diminished in quantity, has its natural and its market price. The natural price of labour is that price which is necessary to enable the labourers, one with another, to subsist and to perpetuate their race, without either increase or diminution.

 

The central fact of capitalism is the historic separation of the working class from the means of production, and the consequent emergence of wage labour. Dobb has remarked that Ricardo’s argument against Adam Smith’s “adding-up” theory of price “turned on his bringing money itself within the circle of commodities, and in doing so postulating that the price of any commodity or group of commodities can only rise if more labour is required to produce it relatively to the amount of labour required to produce an ounce of gold.” It could even more pointedly be argued that the heart of Ricardo’s theoretical advance lay in his bringing labour itself within the circle of commodities. 

 

Everything in Ricardo’s discussion in the chapter “On Wages” makes it clear that labour is to be analysed as a produced commodity. The quantity of labour available on the market adjusts itself to long-run fluctuations in effective demand, just as does the quantity of corn or iron. And the natural price of labour is determined by the quantity of labour directly and indirectly bestowed upon its production. Clearly, the treatment of labour itself as a produced commodity, both by society and by the theorists of political economy, constitutes a development of enormous historical and theoretical importance. Since this theoretical and historical development lies at the heart of Capital, we shall postpone an extended discussion of it until we come to Marx’s political economy. At this point in our story, let us confine ourselves to those points that are essential to an understanding of Ricardo’s theory. 

 

Ricardo was able to incorporate wage labour into an analysis of the value-determination of produced commodities by adopting Thomas Malthus’ views concerning the growth and fluctuation of population in response to variations in the real wage. In the opening paragraphs of the Principles, Ricardo observes that there are some commodities “the value of which is determined by their scarcity alone. No labour,” he says, “can increase the quantity of such goods, and therefore their value cannot be lowered by an increased supply.” Ricardo cites the now classic examples of “some rare statues and pictures, scarce books and coins, [and] wines of a peculiar quality, which can be made only from grapes grown on a particular soil.” But he brushes these sorts of commodities aside impatiently, noting that they “form a very small part of the mass of commodities daily exchanged in the market.” Like his fellow analysts of the burgeoning capitalism whose revolutionary productivity was even then pouring forth such heaps of commodities as had never been seen before, Ricardo focuses his theoretical attention on common consumer and capital goods – corn, iron, linen, woolens – not on the luxury goods that had formed the predominant part of the trade of an earlier age. Rather than argue the matter at length, Ricardo simply lays it down as a stipulated constraint on his system that “in speaking then of commodities, of their exchangeable value, and of the laws which regulate their relative prices, we mean always such commodities only as can be increased in quantity by the exertion of human industry, and on the production of which competition operates without restraint.”

 

There are, however, two items on the market whose quantity seems, at first sight, not to be capable of increase at will, but whose importance to the economy does not permit us to consign them to the residual category of old pictures and fine wines, namely land and labour. The available acreage of arable land and the labour force seem to be fixed quantities that must be treated as parameters of any economic model, not as variables. Ricardo’s theory of value, or indeed any theory of value that explains price as determined by cost of production, depends upon a successful treatment of the prices of land and labour – rent and the wage. 

 

Ricardo has handled rent by adopting the West/Torrens/Malthus analysis, according to which the rental charged by landlords plays no role in the determination of price. To deal with labour, Ricardo takes up Malthus’ theory that population adjusts itself to food supply. The wage is construed as the price of labour, and a cost-of-production account is given of the determination of that price. When excess demand for labour (during a period of rapid economic expansion, for example) drives the market price for labour above its natural price for any significant period of time, the working class responds by bearing and rearing more children. After a period of disequilibrium, the augmented supply of labour drives the market price down to its natural price. 

 

In weighing the plausibility of the Ricardo/Malthus theory of the wage, particularly in comparison with Marx’s alternative account of the “reserve army of the unemployed,” we ought to keep in mind that Ricardo was writing at the beginning of the nineteenth century, at a time when the secular growth of the labour force was more significant than short-term fluctuations attendant upon business booms and busts. 

 

Prudent capitalists choose the most efficient techniques available for the production of their commodities, combining just as little corn, iron, linen, and labour as is absolutely required by the technology of the period. Competition ensures that there will be no waste, for the entrepreneur who allows his tools and raw materials to be used with less than maximal efficiency will be driven to the wall by competition in the market. 

 

So too, we are to suppose, workers adopt a technique for the reproduction of their labour – a standard of living, we call it – that is maximally efficient. The worker who insists on eating steak rather than potatoes, and who prices her product – her labour – to suit will find herself unable to sell her product on the market. She will be out of a job, in short. Thus, the natural or equilibrium price of labour will be a wage just sufficient to support a subsistence standard of living. 

 

But to this Swiftian tale, Ricardo adds the crucially important qualification that habit and custom, history and culture, in part determine what will count, at any moment, as subsistence. One passage will suffice to represent Ricardo’s views on this subject, which are not very systematically developed: 

 

It is not to be understood that the natural price of labour, estimated even in food and necessaries, is absolutely fixed and constant. It varies at different times in the same country, and very materially differs in different countries. It essentially depends on the habits and customs of the people. An English labourer would consider his wages under their natural rate, and too scanty to support a family, if they enabled him to purchase no other food than potatoes, and to live in no better habitation than a mud cabin; yet these moderate demands of nature are often deemed sufficient in countries where “man’s life is cheap,” and his wants easily satisfied.. Many of the conveniences now enjoyed in an English cottage, would have been thought luxuries at an earlier period of our history.

 

Ricardo shows no awareness that the collective definition of subsistence might be a matter over which classes could struggle, nor, needless to say, does he evince any recognition of the deep epistemological problems posed by the possibility that the cognition of certain elements of social reality is an object of class conflict. 

 

One of the peculiar consequences of Ricardo’s assimilation of the reproduction of the working class to the entrepreneurial production of commodities is that in his theoretical model, workers are the only producers who have a significant positive interest in employing sub-optimal methods of production insofar as they are able! From an economic point of view, an improvement in the standard of living of the working class is a step backward to a less efficient method of production. A higher standard of living means less capital available for investment and growth. 

 

Piero Sraffa, in his modern reconstruction of the Ricardian perspective, adopts an alternative way of analysing the price of labour. He treats all wage payments as a distribution of a portion of the surplus, in the manner of neoclassical theory but not of classical or Marxian theory. He then studies the relationship between the wage and the profit rate, allowing each to vary from zero to its maximum magnitude. If we look back at System C, for example, we can calculate that π reaches its maximum at π = 5, when w = 0. When π = 0, w = 2.5 units of corn. This mode of analysis has a number of implications, all of which seem to me to be unfortunate for a fruitful exploration of capitalist economies. 

 

The first is that as far as the formal structure of the model is concerned, the wage rate can in principle sink to zero, leaving the workers to live on air. A rise in the wage above zero is construed as a reflection of some positive measure of political or bargaining power of the working class. But this is contrary to capitalist reality. As Ricardo and Marx quite correctly observe, at any specific moment in history there is a physically and socially defined conception of subsistence that determines a floor below which the real wage may not for long be allowed to fall. 

 

Viewed from the perspective of Ricardo and Marx, labour/management bargaining takes two quite distinct forms. The first is a struggle over the distribution of what is acknowledged to be surplus, with capital seeking to drive labour’s wage down to subsistence, and labour seeking to raise the wage above subsistence. The second is a straggle over the social definition of reality itself – a struggle over what shall count as part of subsistence. Are medical services necessary, or are they a luxury? Are pensions necessary, or are they too luxuries? Is meat a necessary part of a working-class diet? And so on. The periodic redefinition by the U.S. Bureau of Labour Statistics of a poverty-level standard of living for a family of four is merely the latest and most sophisticated version of this old struggle. 

 

When scarcity of labour drives the market wage above the natural wage (as defined by the conditions of subsistence), workers strive to preserve their gain by redefining “subsistence.” Capitalists meanwhile seek to drive down the natural wage by branding costly elements of the working-class real wage bundle as luxuries that are unnecessary and inimical to a satisfactory rate of capital accumulation. The importance of this direction of analysis lies precisely in its identification of the ways in which the collective social definition of reality supersedes merely technical production relations. This in turn introduces an historical dimension into what is otherwise a timeless analysis of objective input/output proportions. Sraffa’s analysis confuses the two quite different forms of labour/capital struggle. 

 

The second consequence of treating wages as a distribution from surplus is that wage goods (food, clothing, and shelter) must be construed formally as luxuries rather than as inputs into production. Food, on this construal, is not directly or indirectly required for the production of all commodities, even though labour is directly required in all industries, because it is in theory possible for workers to consume nothing at all. It follows that a change in the facility of production of corn will have no effect on the profit rate. Aside from the fact that this consequence is utterly false to Ricardo’s intentions and underlying ideas, it is, I suggest, a quite unfortunate analytical implication to build into one’s basic model. 

 

Third, the Sraffa line of analysis construes labour as a scarce good whose supply is determined outside the economy, not as a produced commodity whose natural price is governed by the conditions of its reproduction and whose available quantity is regulated by market forces. In short, labour is, in Sraffa’s model, exactly like land. Now, the Sraffa/Ricardo analysis of rent portrays it as a distribution of a portion of the surplus to landowners as a consequence of the scarcity of land of the best quality. When land of the best quality is not in short supply, competition together with the rational self-interest of the landowners drives rental charges to zero. But as far as any analytical feature of Sraffa’s formal model is concerned, labour ought therefore to earn no wage at all. The workers, like the landowners, are the possessors of a factor of production that is in excess supply and has no alternative uses. In therefore ought to bear a zero price. 

 

More generally, insofar as we distinguish skill levels, and admit alternate techniques of production using labour of differing skill levels, we can suppose that when all the most highly skilled workers have been employed at vanishingly low wages, a rent will arise on their scarce skills, and they will begin to receive this wage while their less-skilled brethren receive no wage at all for unskilled labour. The wage, like ground rent, will thus be seen as a distribution of the surplus, having no effect on prices. On this analysis, which is indeed implicit in the Sraffa approach, the only difference between landlords and labourers is the extra-economic fact that workers organise to drive the wage above its natural economic level, for example by unionisation or other techniques for creating artificial scarcity, whereas landlords do not. 

 

This is certainly a consistent way of handling labour and the wage, but it is, I suggest, quite unsatisfactory. First of all, it entirely ignores the purely physical or biological preconditions for the survival of the workers. Second, it belies the central fact that what is commonly accepted as a subsistence level of existence is socially determined. Third, it loses entirely the analytically powerful idea that the wage is related to the cost of reproduction of labour, an idea which is central to the entire classical tradition. For these reasons, I believe Sraffa is wrong to adopt the analytical line he does, and in the later portions of this book, I shall follow the Ricardo/Marx proposal to treat the wage as determined by the cost of reproduction of labour