Hay, carrots, bread and roses
Subsistence and surplus wages in Sraffa’s papers
Introduction
This paper focuses on the concepts of subsistence and surplus wages as proposed by Sraffa. Wages have a double component: they are both a cost of production and a share of net income — an ambivalence that reflects deep tensions inherent in the nature of wage labour as a commodity that needs to be reproduced, but also can negotiate its own price. This ambivalence conceals a double perspective: for capitalists, wages are a cost, whereas, for workers, they are the entitlement to the means of subsistence. This means that within the definition of wages is hidden a conflict of interests: understood as a cost, workers’ consumption is merely functional to the production and realization of profit, but, from the workers’ point of view, wages are means of their well-being (Sen, 1985).
The ambivalence inherent in the double component of wages makes it possible to see the nature of wage negotiation, the terms of the ground of confrontation, the social subjects involved and the radical nature of the conflict. The double component is in reality the tip of an iceberg, which the study of Sraffa’s papers enables us to discern. To grasp the depth of this conflict, we must be clear about the questions inherent in subsistence and the forces acting within it. It is in fact in the historical changes in normal living conditions for the whole labouring population (waged workers and those who depend on their wages) that we can trace the dynamics of the capitalist system. In order to fully grasp the ambivalence of wages, it is important to work both on Sraffa’s book and on his archive papers.
In Production of commodities by means of commodities: prelude to a critique of economic theory,1 Sraffa excludes subsistence commodities from the category of basic commodities that, by definition, enter the production of all other commodities (Sraffa, 1960, pp. 7–10). By so doing, he partly modifies the wage theory adopted by the surplus-approach theorists, from Petty to Marx, though he relies on them for the constructive part of his theory.2 The result of this decision is the removal of two fundamental processes from the direct focus of the analysis of surplus and relative prices: (a) the production of wage commodities, traditionally symbolized by corn in the works of classical economists; and (b) the subsistence process, i.e. the social reproduction of the labouring population, which classical political economists defined as a bundle of commodities conventionally necessary to reproduce the ‘race’ that supplies the commodity labour.
In Production of commodities, the idea of wages as an inventory of conventional necessaries is replaced by that of a variable proportion of the net product. Actually, when surplus is introduced, wages are, at first, given the twofold character of productive consumption and net income, but later, only the character of net income. Sraffa asserts, however, that the system of price equations, formulated in terms of net wages, can be extended to the ‘more appropriate, if unconventional, interpretation’ in which subsistence wages are also included (ibid., p. 10).3 The exclusion of wage goods from the basic commodities also makes it possible for wages to be paid ‘post factum’, at the end of the process of production (ibid., p. 13). In fact, this is not bread eaten before going to work, including childhood, but an additional loaf of bread, or a rose, to be enjoyed as surplus.
In this context, Sraffa’s archive papers provide material that can help us to capture the implications inherent in subsistence wages, seen by classical economists as the indicator of a sustainable and necessary state of the process of social reproduction of the labouring population, and we speculate on the historical and analytical relationship between surplus and subsistence wages.4 The archive papers shed light on the classical concept of wages as a real physical cost necessary to form and maintain workers’ productive capabilities, to reproduce the labouring population in a social context, and to meet the reproductive conditions of the economic system as a whole. From this follows the analytical method of ‘physical costs’, which Sraffa contrasts with the method used by the marginalist economists who define the real cost of labour as disutility and sacrifice.
From his first critical comments on Cannan’s Theories of production and distribution, dated February 1923 (D1/67.1.3), to the notes in a file marked by Garegnani as ‘Gathered by Sraffa in preparation for a work subsequent to Production of commodities’, dated 1945–67 (D3/12/42), Sraffa looks into the question of subsistence and its analytical implications for the concept of the cost of labour. The key words for understanding the nature of wages in the physicalcosts method are: ‘enable’ and ‘necessary’. As an illustration of their meaning, Sraffa in a ‘Notebook’ dated December 1927–March 1928, entitled ‘Looms’ and marked ‘IMPORTANT’ writes:
Example, carrots are necessary if we want a donkey to work. But there are two sorts of carrots: those which we must have given to it before in order to enable it to work (otherwise it would be dead) and those you must show to it and promise to it in order to induce it to work.
There is a great difference between the two: the first is a definite number or weight of real carrots, determined by physiological condition, and since they have been actually consumed, it has been possible to weigh them and to know exactly to the ounce their quantity: no tricks can be played about them.
The others are different, they needn’t even be real carrots — It may be a mashed paper [sic] carrot, rubbed against real carrots to take up the smell, which we simply show to the donkey, or it may be a stick […]
Now economics deals with mashed paper carrots and whips, P.E. dealt with real costs.
(D3/12/10.61.1–3).
In our view, the notion of subsistence wages is central, both for the full understanding of the classical surplus approach as a method of physical costs, and for the critique of neoclassical utility theory, considered by Sraffa the ‘black night’ of political economy (D3/12/7, 160). Sraffa is well aware that the role of subsistence is not limited to backward living conditions and economies or to a list of basic necessaries. In a letter to Garegnani of 13 March 1962 (D3/12/III.147), Sraffa denies that he is thinking of a clear sequence from subsistence to surplus consumption, and also indicates some readiness to change the relation between rate of profit and rate of surplus wages:
Here we get into a difficult matter, and while I would be glad to talk with you about the question of determination of wages, it is too complex for a letter. First of all, ‘subsistence’ has never meant pure ‘physiological necessity’ (whatever that means), but always includes social and historical or habitual necessity. This is clear, if only because when wages are reduced people will often give up something physiologically necessary before giving up a ‘superfluous’ thing like alcohol, smoking etc. And then when a ‘standard’ level has prevailed for a certain time it becomes necessary — if you want the result.
[…] On the other hand, I am convinced that maintenance of the interest rate by the bank or stock exchange has had its part in determining the distribution of income between social classes: because it necessarily affects anyone who loans or borrows.
[…]I have no intention of proposing another mechanical theory which, in one form or another, would support the idea that distribution is determined by natural, technical, or even accidental circumstances that would render futile any action on either side aimed at modifying it. In conclusion I would say that in the review it will be best not to insist too much on the obiter dictum of the monetary rate of interest.
(D3/12/111.149–150, translated from Italian)
The present study consists of four sections. In the first, Sraffa’s papers are analysed with regard to the question of the real cost of labour: (a) in the neoclassical and (b) in the surplus theories. In the second, we discuss the process by which Sraffa formed the decision to exclude subsistence from the basic commodities. In the third, we focus on the ambivalence of wages as costs and net income. In the fourth, we draw some conclusions about the relation between subsistence and surplus wages and its modern relevance.
The real cost of labour in the archive papers
At the end of the 1920s, Sraffa began working on Production of commodities and, at that stage, probably chose his point of attack to produce the greatest critical effect on marginal utility theory, using the toolbox of a classical surplus approach.5 This point is the question of the determination of surplus in the productive process, its distribution at institutional level, and its possible objective measurement in a system of relative prices of heterogeneous commodities. This line of attack enables him, on the one hand, to take up the argument of the measure of surplus where Ricardo and Marx had left it, and, on the other hand, to show the logical inconsistency of the measure of the quantity of capital, at the basis of neoclassical economics. In doing so, Sraffa omits the component of subsistence wages from the determination of surplus and relative prices, but clearly brings out the institutional distribution of the surplus and the variability of wages.
A key element in understanding the notion of subsistence and surplus wages is Sraffa’s work on the history of economic thought, which he considered indispensable for comprehension of any analytical question. According to Sraffa, conflicting interests inevitably affect the concepts, the tools and the relative importance of the problems. Thus, there is a risk of distortion with the passage of time, as theories are used without their original meaning and context being remembered. Nevertheless, theories deposit concepts and analytical tools as a patrimony to use and criticize, regardless of the motivations from which they originated (D2/4.3.2). Moreover, in determining the order and weight of their arguments, theorists interact with the common sense and prejudices of the public about practical problems and their possible solutions. Especially in the case of the theory of value and distribution, the rhetorical aspects of the arguments are essential for understanding and modifying current conceptions. In this regard, Sraffa’s words offer an insightful warning:
Thus every economist tends to frame his theories in such a way that certain elements acquire in them importance which is entirely out of proportion to the part they play in real life, but reflects the necessity ofin which the economist has been of opposing oppositeobsolete theories or popular prejudices. And when the theory has crystallized and we have forgotten the way in which it has grown, we are often inclined to over-estimate the importance of certain elements simply because for long-forgotten historical reasons they play a very large part in accepted economic theory.
A further disturbing element is that in the background of every theory of value there is a theory of distribution. The real problem to be solved by a theory of value, that is: ‘Why is a commodity exchanged with another in a given ratio?’ is constantly transformed into the entirely different one: ‘How is the price received for the product distributed between the factors of production?’ […] There is a continuous attempt at visualizing in the microcosm of any one particular commodity a process which takes place only in all commodities as a whole, considered simultaneously, that is in society as a whole.
And often theories of distribution in their turn are meant not so much as a means of analyzing the actual process through which the product is distributed between different classes, as for showing either that the present system is wrong and should be changed, or that it is right and it should be preserved. Thus it becomes an analysis of what is itthe theory becomes a form of propaganda for what ought to be.
(D2/4.3.3–4)
To understand the theories of wages and distribution, one must begin with the fact that, in their analyses, all economists, classical and neoclassical, are talking about the functioning of a capitalist system in which workers are recognized as means of production. In this sense, the costs of social reproduction of labour ought to be analysed like the costs of reproduction of slaves, horses and machines. In a document called ‘Real physical costs’, Sraffa remarks:
Capitalism is very ill adapted for the analysis of those parts of econ[omic] theory which are common to any economic system. It is intermediate between slavery and socialism. The best suitable is a community where all labour is performed by slaves. Such a comm[unity] exibits clearly the distinction between real cost of labour and wages. Horse & slave, equal: they receive food necessary for efficiency, and all profits go to the employer.
(D3/12/42, 40)
As a matter of fact, all the difficulties in the theory of wages are due to the ambivalence arising from the fact that wage labour is both free and ‘commanded’. This ambivalence is reflected in all the questions linked with wages in both the theoretical traditions, classical and neoclassical. The element of freedom in waged labour takes above all the form of self-management of reproduction; it is this self-management that makes waged workers different from slaves.6 In the case of slaves, the owner is assumed to control their reproduction directly, as with horses, controlling diet, shelter, cohabitation and number of children, whereas the capitalist does not administer the reproduction of workers and does not directly dictate the number of children, cohabitation, separations, gender relationships, diet, housing, etc.7
In general, we can say that the concepts of subsistence and surplus wages are present in both theories, but they are explained and positioned differently. On the one hand, the neoclassical economists see subsistence simply as a downward rigidity of wages, analytically marginal and introduced for reasons of social equity in the presence of poverty, or as an inducement to efficiency that does not challenge the general theory of labour supply-and-demand utility functions. On the other hand, the classical economists see it as the normal price, necessary for enabling labourers to work and reproduce their race. Conversely, surplus wages are seen as the general case by the neoclassicals and a particular case by the classicals.8
Though we cannot here go into Sraffa’s detailed survey of neoclassical and classical literature on wages, it is worth noting a few analytical points, indicated in his papers, in which the question of subsistence plays an important role in neoclassical economics, and surplus theories.
Neoclassical economics
The cynicism of treating workers like horses is the point on which Marshall clearly demurs. In the first chapter of the sixth book of the Principles of economics, where he deals with the distribution of income, he explicitly states:
The keynote of this Book is in the fact that free human beings are not brought up to their work on the same principles as a machine, a horse, or a slave.
(Marshall, 1920, p. 504)9
Apart from humanitarian goodwill, however, the problem is to see how the question of social reproduction of the labouring population fits into the neoclassical theoretical scheme, especially with regard to Marshall’s attempt to link the theory of demand, based on marginal utility, with the classical theory of costs of production. Marshall’s synthesis becomes impossible precisely because of the profound difference in the meaning assigned to the real cost of labour in the two theoretical contexts (D3/12/7.105–6).
In the case of the labour supply, the concept of marginal utility is merely ‘thin air’. In the neoclassical approach in general, according to Sraffa, measurements are statistically impossible, and systematic functional relationships between prices and quantities, based on marginal utility, are illusions that only apparently find a real sense through money. Bentham, says Sraffa, had already indicated this solution to get out of certain difficulties linked with the measurability of pleasure, and the same is done by the modern economists who use money as an indicator of utility (D1/11/1.7).10 According to Sraffa, the marginalist formulation of the utility functions forces us to assume a choice between commodities, one with solely positive utility, the other with only negative utility. In Marshall’s view of the labour supply, for example, work is only disutility and consumption only utility. ‘This is too simple’, says Sraffa in a document, dated in the catalogue as pre-1928, and entitled ‘Scissors in ultimate conditions’ as: ‘In every act of consumption there is a hidden element of disutility as in every toil there is an element of utility […] there are no unmixed pleasures and there are no unmixed evils’ (D1/13.5.1). A document dated Paris, February 1923, before his Cambridge period, contains reading notes on Cannan’s Theories of production and distribution, in which Sraffa uses references to real daily life as a basis for calling for a return to common sense in the notion of labour supply. In particular, ‘it is absurd to reduce amounts of food to infinitesimal quantities, as below a certain limit of size commodities are useless’ (D1/67.1.3). Hence, there exist critical quantities that depend on the function of the commodity and the rhythm of life that regulates the timing of consumption; besides, the utility of successive units might even increase, and it can happen that ‘l’appetito vien mangiando’ (ibid.).11 The utility of income also depends on context and, hence, above all, on living conditions of those who receive the income.12
In reality, the hypothesis of a continuous choice between income and work, on the basis of utilitarian axioms, is contradicted if one explicitly takes account of the physical process of subsistence and conventional standards of living (D1/13.1–6). For example, with an increase in price, the supply of labour may diminish because other family members leave the labour market owing to social conventions regulating women’s work (D1/11.6). Or else, with a fall in wages, the labour supply may increase because the wage can no longer guarantee the habitual standard of living; or, on the contrary, with a wage increase, the supply may fall because the habitual standard has been reached (D1/11.90.2).
The question of subsistence as a necessary cost of production re-emerges as an unresolved muddle when it comes to defining net income; on this question, there is no agreement among neoclassical economists. In fact, as a cost it should by definition not be included in the net income. Thus, in the report of the British Association on the Common Measure of Value in Direct Taxation, compiled in 1878 by a commission including Jevons, it is asserted that: ‘As the horse has to be clothed and stabled, so the productive labourer has to be clothed and housed’. This assertion is cited and criticized by Edgeworth in the entry on ‘Income’ that he wrote for The Palgrave dictionary (1906, p. 374), as reported by Sraffa (D3/12/42.36).
The costs of reproduction of labour emerge as an analytical problem also in the debate on the value of emigrants in the Giornale degli economisti between 1904 and 1905 (D1/60.1). Pareto entered the debate with an article entitled ‘Il costo di produzione dell’uomo e il valore economico degli emigranti’,13 in which he advises to include the costs of raising labourers that migrate in the national expenditure (Pareto, 1905, p. 326). In a previous article in the same journal, Coletti had mentioned also the economic value of caring for children:
An important class of elements which the statistics fail to show is the care devoted by the mother and other family members to the upbringing of children […]
The great influence of this element of cost, that is to say the different degrees of care and affection used in bringing up children, is indicated […] by the obviously greater mortality of illegitimate, neglected children compared with legitimate children.
(Coletti, 1905, p. 264, in Italian)
Pigou and Marshall also take up these questions, although when defining national income Marshall says it is impossible to make unpaid domestic work visible (Marshall, 1920, p. 524), and Pigou excludes from national income ‘food and clothes essential for the maintenance intact of the labour force’ (Pigou, 1946, p. 4).14
The surplus theories
In the classical surplus approach, the reproduction of labour is seen as a physical and social process whereby flesh and blood men and women, embedded in social communities, at particular times and places, are made able to work (Picchio, 1992, pp. 8–29).
The clearest formulation of subsistence in a surplus approach is given by the Physiocrats. This is referred to by later surplus theorists, who use various circular flows to represent the reproduction of the system. For the Physiocrats, the correct definition of costs is central for reaching a correct notion of the surplus. First of all, it is clear that rent is not one of the costs, as it is mostly based on an institutional system of property of a scarce non-reproducible factor and does not reflect a contribution to the process of production. In their view, moreover, profit is not yet fully distinguished from earnings from work, and thus is explained as the conventional level of subsistence for farmers. Smith introduces the distinction between wages and profit, but his view of profit is ambiguous: it appears both as surplus in the real system and as a necessary addition within his price system. For Ricardo, profit is consistently a residuum and, hence, by definition not a cost of production. His general trade-off between profit and wages is based on this clarity. The surplus is visible as long as the product and the costs are measured in physical terms (corn). The confusion starts when labour begins to be measured in time or abstract energy. In fact, the measure of value that Sraffa seems to like the most is that of Petty, who says: ‘the day’s food of an adult Man, at a Medium, and not the day’s labour, is the common measure of value’ (Petty, 1972 [1691], p. 65). Sraffa asserts, in a document written in the summer of 1929, that the erroneous result of measuring value in terms of abstract labour arises through a series of ‘small errors’:
The fatal error of Smith, Ricardo, Marx has been to regard ‘labour’ as quantity, to be measured in hours or in kilowatts of human energy, and thus commensurated to value […]
All trouble seems to have been caused by small initial errors, which have cumulated in deductions (e.g. food for worker = quantity of labour, is nearly true).
(D3/12/11.36)
In the document entitled ‘Degeneration of cost and value’, he again takes up this question, viewing it as the origin of a much more radical shift, brought about first by the wage-fund theorists and then by the marginalists. It is useful to follow Sraffa in the path leading to the modern definition of the cost of labour:
Smith & Ricardo & Marx indeed began to corrupt the whole idea of cost — from food to labour. But their notion was still near enough to be in many cases equivalent.
The decomposition went on at a terrific speed from 1820 to 1870: Senior’s abstinence and Mill’s mess of the whole thing. Cairnes brought it to the final stage ‘sacrifice’ […]
Simultaneously a much bigger step was taken in the process of shifting the basis of value from physical to psychical processes: Jevons, Menger, Walras.
This was an enormous breach with the tradition of P.E.; in fact this has meant the destruction of classical P.E. and the substitution for it, under the old name, of the calculus of Pleasure & Pain (Hedonistic).
(D3/12/4.2.1)
The central role given by classical economists to human subsistence in the analysis of relative prices and distribution has methodological implications that Sraffa shares even when he does not give direct attention to it. In the papers it emerges that Sraffa is fully aware that the question of value is not reducible to a problem of measure: it is also a problem of method and of vision. As regards to method, it is worth noting the comment written, in February 1923, to Cannan’s introduction:
From his first words C. sets himself the task of searching for causes: but why not for the relationships of the nature and the structure of institutions and phenomena? Can he possibly manage to explain everything as a chain of cause and effect?
(D1/67.1.1; translated from Italian)
In the preparatory notes for his lectures of 1928–31 (D3/12/4), Sraffa takes up these methodological aspects again, focusing on the concept of cause, in the metaphysical sense or as determinant in the cognitive and mathematical sense (D3/12/4.6), and assessing the relative importance of these viewpoints.15 The analytical foundation is treated as an intellectual gymnastic ‘which may give us some pleasure because it suits our habits, it clears up some relations, but tells us nothing about the nature of things’ (D3/12/4.14). The really important aspect of the analysis is its historical and social significance, which is:
the truly important, that which gives us a real insight into the mystery of the human mind and understanding of the deep unknown relations of individuals between themselves and between the individual and society (the social, or rather the class mind).
It is terrific to contemplate the abysmal gulf of incomprehension that has opened between us and the classical economists.
(D3/12/4.14)
The definition of metaphysics given by Sraffa in the same document is also significant:
by metaphysics here I mean, I suppose, the emotions that are associated with our terminology and frames (schemi mentali) — that is, what is absolutely necessary to make the theory living (lebendig) capable of assimilation and at all intelligible.
(D3/12/4.15.1)
Thus the metaphysical aspects are not dismissed, but rather, only at the level of measure, the question of value has to be freed from political and ethical judgements. Values belong most of all to the philosophical vision and anthropology of the social system (D3/12/7.161), nevertheless, the measure of exchange values can be dealt with a system of relative prices that express a self-sustaining exchange system which allows for a sustainable process of production of commodities by means of commodities.16
A few notes on the route leading to wages in Production of commodities
As we mentioned in the first section, Sraffa sacrifices the given subsistence wage as an inventory of commodities, and instead expresses wages as a variable proportion of the net product. In the archive, this change in the notion of wages appears explicitly in a document dated 1–1–43, entitled ‘Transition from 2nd and 3rd equations, i.e. Replacement of wages as constant inventory as wages as variable w’ (D3/12/33.90). This is worth quoting at length, because it is crucial for the question of physical costs, and important for the distinction between subsistence as a list of commodities and net wages as a proportion of net product:
We have represented wages as a list of commodities, each in a specified quantity in units of weight, length, etc. Now we propose to regard wages as variable. We could do this by regarding the quantities of the commodities entering into wages as variable. Then wages could vary in a large number of ways: the quantity of one commodity could increase in a given proportion, that of another in a different proportion while that of a third decreased, and a fourth commodity hitherto not consumed by the working class might be added to the list in a given quantity. We could give to these variables the value which they had in each situation, & an incidental result of the solution would be the total price of the commodities entering into the wages. In the present inquiry however we are only interested in this last quantity, the price of wages, & not in the different wages in which they may be spent.17
For each set of values of these numerous variables there would be only one total price of wages: but many possible sets could correspond to the same total price.
(D3/12/33.90.1a)
This passage may provide an indication for a possible ‘easy’ adaptation (according to Sraffa) of the system of relative prices, to the already mentioned ‘more appropriate, if unconventional, interpretation’ of wages as subsistence (Sraffa, 1960, p. 10). It is marked in the margin with a wavy line, which might indicate a lack of total conviction. Whereas there are no signs of uncertainly in the second part of the passage, which says:
Now we propose to regard wages as variable. But instead of doing this by regarding the quantity of each commodity entering into wages as an independent variable (which would lead us into an enquiry of the ways in which wages are spent) we shall consider a single variable, the price, in terms of the commodities chosen as standard of the collection of the commodities which enter into the workers’ consumption. The list of commodities can vary in many ways and to each list there corresponds one value of w: but to any one value of w there corresponds a large number of lists of commodities (always within the limits of those produced as we do not now consider changes in production).
(D3/12/33.90–.a-b)
The aggregate W (wL) also conceals possible variations in quality of the commodities included in the lists, as it deals only with the proportion of surplus distributed to wages. This question is dismissed as of ‘no interest except to the shopkeepers’ (D3/12/33.91.) and, we would add, to housewives. In a document dated 17–1–43, Sraffa acknowledges the necessity of changing the first equations if wages are considered a dependent variable and writes:
In the first equations, wages are lots of commodities. If now we want to regard them as variable, it is clear that there are sorts sources of variation: the quantity of labour may change + the wage per man may change, in our present notation these two are not distinguished. So we must explicitly make appear the two quantities, as w and L.
(ibid.)
In this regard we could note that the quality of forces affecting respectively w and L are quite different, although partly related; w, as standard of living, belongs to the sphere of social reproduction that includes habits, tastes, social conventions and institutions such as the family and the state, while L (employment) depends on production, technologies and the rate of accumulation.18
The preceding passages show Sraffa’s awareness of some of the difficulties involved in returning to physical subsistence wages, necessarily specified in lists of commodities. These analytical difficulties might have been the reason why he excluded the physical component of conventional necessaries from basic commodities. This exclusion, moreover, makes it more plausible to use the proportion of surplus wages as a dependent variable, as Sraffa explicitly recognizes in an important document entitled ‘Scaffolding’, dated September 1956 (D3/12/68). In this case too, it is worth recalling the textual formulation that gives an account of the decision:
On the other hand, with the wage measured in abstract St[andard] c[ommodity], it becomes awkward and unrealistic to continue to regard it as an ind[ependent] variable a practice which originates from a wage which consists of the nec[essaries] of sub[sistence]; it will therefore be convenient to replace in that position with the rate of profits.
(D3/12/68.1)
However, the choice to define the surplus wage as a dependent variable could also be explained by a certain reluctance to specify the actual conditions of material and social reproduction of the labouring population, and the preference for monetary theory, a field in which modern economists, including Sraffa, felt more at ease. On monetary theory, Sraffa had stated, in a document going back to the end of the 1920s:
Among the different parts of economics, the monetary part is one of the least inexact: the theory of money is the one which perhaps neglects the fewest essential facts. This is partly due to the relative simplicity of the material, partly to the fact that because of its practical importance it has been longer studied and more extensively elaborated. Consequently monetary theory (together with that of finance) is perhaps the only economic theory that practitioners must take account of.
(D2/1.2, in Italian)
Lastly, Sraffa is aware of the heroic simplifications required in order to reach clear definitions, and of the reductionism necessarily involved in a process of abstraction. But he thinks the classical concepts are sufficiently clear and rooted in experience to constitute a usable and fruitful analytical tool for studying certain basic aspects of the capitalist economic system. In a document, already mentioned, that includes some notes collected by Garegnani under the rubric of ‘material prepared for a possible future work’, Sraffa writes:
The difficulty of distinguishing in the total money cost of a thing what is real cost and what is surplus may be very great in practice, but it is not greater to conceive than other similar distinctions that are accepted in every economic theory in respect to rent, interest, etc.
The difficulty is, what is ‘necessary’ food, shelter etc., to be given to the worker in order just to enable (not to induce him, as this involves the possibility of alternative employment, in the widest sense, including leisure, on his part) to produce a thing is certainly not greater than that of distinguishing, in the total payments made to landlords, what is real rent and what is interest and depreciation of capital sunk into the land, or in the total sum paid for the hire of a horse what is food, shelter & depreciation of the horse and what is interest on capital.
(D3/12/42.35)
The ambivalence of wages
Many of the problems inherent in the analysis of wages derive from the fact that two meanings converge in its definition: cost and income. The two meanings can be reconciled, in a capitalist system, only if income is considered gross income or if subsistence and net wages are seen as different components. The ambivalence of wages, nevertheless, is rooted in the historical nature of wage labour. According to the classical economists, the analysis of the conditions of self-replacement of the system must perforce start out from the production of the subsistence goods required to enable workers to work and reproduce as a race. However, in a capitalist system, the workers are means, and the commodities conventionally needed for their subsistence are to be considered as capital. But horses and other commodities do not receive income. If they did, they would belong to the nation understood as a community of citizens enjoying rights, as suggested by Cannan in a note to Smith’s Wealth (Smith, 1976b, p. 57), taken up by Sraffa (D3/11/37.6). Wages as income refers, then, to a constitutional plane of citizenship that places wage labour in the sphere of commodities and labourers in that of rights.
In the papers, Sraffa more than once calls attention to the ambivalence of wages as income and cost (D1/5.83), returning it to the historical context in which are embedded the distribution of income among classes and the very notion of value. The conceptualization of wages depends on the nature of capitalist wage labour and the inherent profound ambivalence between individual freedom and structural constraint to work (D3/12/42.40).
The difficulties of distinguishing between a process of reproduction of human productive units and a process of income distribution are aggravated by the fact that subsistence necessaries are defined by social conventions regarding bodies, consumption, gender relations, generations, life cycles, cultures, places etc. Moreover, the conceptualization of what is conventionally necessary depends on the perspective of the subjects observing the economic system. For instance, in a capitalist system, there is also a ‘necessity’ for profit, conventionally acknowledged as recognition that its rule rules. In this connection Sraffa notes:
the distinction between wages & interest is wrong therefore: they are both necessary.
How then did we reach the previous conclusion? Is it wrong?
Our point of view was wrong. We were looking at it from the point of view of ‘what can be changed’. Wages we found are dependent upon physiological laws that cannot be changed, and habits that can hardly be deliberately changed, if at all. Interest depends upon civil law which can be changed.
But by whom and how?
(D3/12/7.42)
As a matter of fact, the two components of wages, subsistence and net income, express two different economic problems. On the one hand, the analysis of production demands that the conditions of reproducibility of the system be made explicit; on the other hand, the distribution of income reflects the historical fact that the wage workers are not horses or slaves. In a document called ‘Definition of net wages’, dated 1928–31 in the archive, Sraffa writes:
The real problems of distribution arise in the second part that constitutes a ‘surplus’. This part is distributed among the factors in proportions that are largely determined by causes other than the mode of production.
We have thus, implicitly, determined the two conditions essential for constructing a theory of wages. A) the property of the capital must be invested in persons other than those of the workers: otherwise, each worker—capitalist—landowner would be master of his own product: and thus distribution would once again be predetermined by production, there would be problems of exchange, not of genuine distribution. B) the worker must be free; i.e. must himself be the owner of his working strength. Slaves receive from the national product only the portion sufficient to keep them working, at most reproducing: they receive nothing from the surplus.
Moreover, what they receive is not part of the net national income, and the same is true of the fuel employed to drive the machines.
(D1/60.11)
The two components of wages thus reflect a conflictual class relation — as Sraffa recognized in a document of 1928 entitled ‘Notes, essential on industries using hypothetical examples, with a note on language’. The inherent conflict is made plain, indeed, in the definition of what is necessary, definable only by indicating the class perspective. Precisely, in a document entitled ‘Surplus product’ Sraffa says:
The study of the ‘surplus product’ is the true object of economics: the great difficulty of the matter is that this object either vanishes or remains unexplained. It is a typical problem to be handled dialectically.
This notion is connected with that of ‘necessity’, & ‘necessity’ has only a definite meaning from a given point of view, which must be explicitly stated, & then adhered to consistently.
[…] Therefore, according to what an economist selects as a ‘subject’ of his economy (usually identifies himself with it) the surplus will be different.
The standpoint of capitalist society itself is that of the ruling class & therefore the surplus is composed of rent, interest and profit.
(D3/12/7.161.1)
In a dialectical perspective, it must be remarked that both subsistence and surplus wages belong to the anthropology and politics of the wellbeing of the labouring population. The aggregate of the ‘wages fund’ (W) is determined by the normal living conditions of the labouring population (w), multiplied by the quantity of labour employed (L). These two dimensions — normal living conditions of the labouring population reflected in wages and the actual use of the employed labour force — reflect different social subjects that negotiate in different social spheres. The politics of living conditions is in fact kept historically separate from the politics of working conditions, although differences and separations, highly gendered, usually pass unnoticed, and standards of living are seen as adaptive to markets and technology. If the economic system is extended to include the process of social reproduction, the relationship appears to be much more complex, but wages reflect this very complexity.
Conclusions: the relationship between subsistence and surplus wages
In the papers, Sraffa offers us a guide for recovering the concept of subsistence wages, and in his book provides a system of relative prices that brings out both the technical physical relationships between inputs and outputs and the institutional distribution of net income. He separates distribution from the false technical objectivity and mystical evanescence of the marginal utility of ‘factors of production’, and restores it to the ‘push and pull’ of institutional power relations between antagonistic classes. Class opposition, however, if we keep living conditions under direct focus, can be seen as based on the whole process of social reproduction in its many dimensions. Waged men and women, and those who depend on their wages, are social and political subjects who can struggle to gain a higher real wage (more and better commodities) but also for justice, rights and greater control over their lives. To improve their individual and social well-being, they have to negotiate over the price of labour, and this fact makes the labour market a very different institution from other markets. Moreover, ‘conventional necessaries’ that define the efficiency component of living standards are not separable from the surplus (luxury) component. They are not separable in different lists. For instance, music and games have always been part of subsistence.
Sraffa shows that, in the relationship between subsistence and surplus wages, we can find: an ambivalence, a double and conflicting perspective, a different metric (list or proportion), a social conflict over distribution, a tension between means and ends, a possible contradiction. The complexity of human passions at play in the process of making a life does not exclude the relevance of their material dimensions. At the end of the day, the labouring population needs food, clothes, houses, hospital beds, books, movies, records, trains, computers, etc. Effective consumption can be divided for analytical purposes into conventionally necessary consumption (social capital) and surplus (luxury). How to distinguish between the two must be left to a historical, political and anthropological enquiry. Thus, historically, we can find it realistic that profit and interest, i.e. profit-holders and bankers, have the power of imposing, within certain efficiency limits, their perspective on what is socially necessary and on distribution of surplus, but it is also possible, as Sraffa recognizes, that the lead in the sharing rule could move into the hands of the labourers, with major dynamic effects.
If the subsistence component is kept visible in the analytical framework, wages are defined, in our opinion, by four fundamental processes: (1) production, as labour is a necessary input; (2) exchange, as they are a price established in the labour market; (3) distribution, as they are a share of national income; and, last but by no means least, (4) social reproduction of the labouring population, which is the process that enables people to enter the labour market. To state that wages cannot be lower than the level of conventionally necessary consumption means anchoring them to a complex historical and anthropological process that marks the classical theory of the value of labour.19 By focusing on subsistence wages, within the Ricardian and Marxian core of values and prices of production, the whole system is given space for a multidimensional methodology, based on the nature of the individual and of his/her relationship with society, that gives a central place to bodies, minds and passions. This multidimensional approach is rooted in the foundations of classical economic thought, where the subsistence of the labouring population is brought together with the conditions of production and exchange of other commodities, and where living conditions of the labouring population are dealt with at the level of the structural self-replacing state of the economy.
The notion of subsistence wages also enables us to identify a deep fracture running through the whole capitalist system, caused by the tension between wages as normal costs of social reproduction and profit. The roots of this tension lie precisely in the difficulty of containing the labouring population’s material and cultural conditions of life and gender relationships within the limits of the reproduction of ‘human capital’ — i.e. the necessary costs of maintaining efficiency. The component of capital as conventional necessaries, clearly indicated by Ricardo in terms of ‘food and clothing’ (Ricardo, 1951, p. 95), has been largely ignored by economists, who have concentrated mostly on fixed capital and its technological modifications. Smith and Ricardo were still in a ‘moral economy’ context and considered the labouring population as a social subject, without political representation but with a considerable power of rebellion focused directly on living conditions — as for example in the case of the food riots.20 In this connection, it is worth remembering that, in the classical scheme, profit is the residuum of product left over after subtracting all that goes to the labouring population, under various headings, including the Poor Law (Ricardo, 1951, p. 108).21 In this wider context, the definition of what is conventionally necessary and what is surplus social wage obviously becomes much more complicated, but maintains all its importance for the definition of capital, in this case social capital.
Sraffa’s numerous and fruitful notes on wages, partly contained in the fragments cited in this paper, enable us to rescue an abandoned line of enquiry, that of wages as normal costs of social reproduction of the labouring population. This line of enquiry discloses deep tensions and indicates dynamic forces at play, both at the level of the institutional distribution of surplus and at the level of epochal changes in gender relationships, habits, tastes and culture of the modes of subsistence. Many problems are still open, as the classical assumption that subsistence is a necessary input brings us back to some questions that need to be conceptualized and better specified. For example: how to define the standard of living? What has to be considered capital? Whose subsistence should be considered capital? What process of social reproduction of the population has to be taken as normal? What is the relation between the individual (male and female) and society? What is the role of the family and state? The theory of subsistence wages, then, is only a point of departure that poses fundamental questions for the understanding of the dynamic structure of the economic system and leads us to focus on the whole process of ‘enabling’ individuals (women and men) to enter the waged labour market and its inherent tensions with the effort to compose their lives.
Notes
* I am grateful to Alfons Barcelò, Heinz Kurz, Annalisa Rosselli and Fernando Vianello for their precious comments, and Pierangelo Garegnani for allowing permission to publish the Sraffa documents cited in this paper. I also thank Joan Hall for translation and the staff of the Wren Library in Cambridge for their patience and kindness.
** The words and phrases in this paper that have been struck through and re-written are reproductions of the original text in Sraffa’s notes.
1 Henceforth Production of commodities.
2 On the connection between classical surplus theorists and Production of commodities, see Sraffa (1960, pp. 93–5).
3 The relationship between net and subsistence wages in Production of commodities is examined in Roncaglia (1974) and Pivetti (2000). In our opinion, they overlook the analytical complexity of the concept of subsistence, its importance in the foundations of the surplus theories, its dynamics and its policy implications.
4 With regard to the chronology of the papers, we date the papers according to the catalogue and we do not trace a development of Sraffa’s thought on the notion of subsistence itself. What changes is the position of subsistence, as coventional necessaries, in relation to surplus wages. In respect of the definition of the two components of wages, of particular importance are the documents D1/58 and D1/60 (1928–1931). For a guide to the papers and the catalogue, see Garegnani (1998), Kurz (1998), De Vivo (2000), Smith (2000).
5 In our view, this sense of strategic choice could be well expressed by a sentence from Whitehead’s An introduction to mathematics, marked by Sraffa in the margin of the copy he possessed (PS, 3500):
Operations of thought are like cavalry charges in a battle — they are strictly limited in number, they require fresh horses, and must only be made at decisive moments.
(Whitehead, 1911, p. 61)
Reading the papers, we see that the preparation of the battle was long and went through different phases: (1) the choice of the point of attack and of the tool box, (2) the long and troublesome adaptation of the classical analytical tools to correct former errors in the measure of value, and (3) the publication of Production of commodities, a decisive moment for the reappraisal of Marxian studies. The political stand was chosen by Sraffa as a young man in Turin (Naldi, 2000).
6 In the separation of production and social reproduction, Smith identifies the source of a significant reduction in the costs of reproduction of wage labour relative to slavery (Smith, 1976a, pp. 183–4).
On the capitalist relationship between production and social reproduction and its importance for the analysis of the labour market, see Picchio (1992).
7 This aspect of direct control over the reproduction of slaves and its implications for the relation between men and women is perceived by Smith in the Lectures on jurisprudence of 1762–3 (Smith, 1978, pp. 175–82). To this regard, the following text is revealing of the material and social basis of sexual relationships:
When a man takes a wife she comes to be alltogether under his protection; she owes her safety and maintenance (especially in the lower ranks) entirely to her husband, and from this dependance it is that she is thought to be bound to be faithfull and constant to him. But a female slave who cohabits with a male has no such obbligation; she is not maintained by his labour, nor defended by him nor anyway supported […] Many other things render their cohabitation precarious: The duration of it does not depend on themselves but on their master […] if he thinks he has not the profit he might have from the slaves the female would bear him, he may take her from her present mate and give her to another. The slave in this manner is deprived of all the comforts and can have but very little of the parental affections of a parent […] The slaves were […] altogether dependent on others for their lives and property and deprived of their liberty and cut out of the consolations of marriage, for we may justly say they had no wives […] Slaves were of all others the most dependent and uncertain of their subsistence.
(Smith, 1978, pp. 178–9)
In the passage from slavery to paid labour in the US, freedom was mostly intended as control over reproduction and the right to have a family (Stanley, 1999).
8 On the presence of surplus wages in classical economists, see Pivetti (2000, pp. 300–1).
9 As Sraffa notes, Marshall, partly and in an indirect way, retrieves the role of living standard in the relation between wages and efficiency, but reduces it to an incentive to produce that does not challenge the supply-and-demand wages theory (D3/12/42.44–5).
10 We might say that utilities are phantoms that can be grasped conceptually only by transposing them onto the level of money income and, hence, of prices — which in fact are precisely what the theory is supposed to explain. Thus, as in a fairy tale, in the neoclassical theory, relative prices act like a sheet thrown over the phantoms to make them visible.
11 ‘Appetite comes with eating’.
12 This relationship between context and the utility of income was noted by Bernoulli, whom Sraffa reads in Latin (D/11.1)
13 ‘The cost of production of man and the economic value of emigrants’.
14 In this connection, it may be noted that the problem of the contribution of the unpaid labour of social reproduction once more rears its head, owing to the fact that in many industrialized countries the national bureaux of statistics measure this contribution and show that its total is generally and persistently slightly greater than the total of paid work, of men and women. On this see Picchio (2003).
15 In another document Sraffa specifies the meanings of different concepts:
Cause in metaphysical sense
Measure in sense of meter of value [value later underlined with a wavy line]
Determinant in sense principium conoscendi
Determinant in mathematical sense, given a we know b in an equation a + b = 0
Indicative of value [added later in pencil]
(D1/67.1.1)
16 On certain methodological aspects, such as the breakdown of the analysis into subsystems, Sraffa seems to us to recall Whitehead. For intance, in the division of the system as a whole into analytical subsystems defined by different levels of complexity:
a fundamental concept which is essential to scientific theory; I mean the concept of an ideally isolated system. This conception embodies a fundamental character of things, without which science, or indeed any knowledge on the part of finite intellects, would be impossible. The ‘isolated system’ is not a solipsist system, apart from which there would be nonentiy. It is isolated as within the universe. This means that there are truths respecting this system which require reference only to the remainder of things by way of a uniform systematic scheme of relationships. Thus the conception of an isolated system is not the conception of substantial independence from the remainder of things, but of freedom from casual contingent dependence upon detailed items within the rest of the universe. Further, this freedom from casual dependence is required only in respect to certain abstract characteristics which attach to the isolated system, and not in respect to the system in its full concreteness.
(Whitehead, 1927, pp. 58–9)
On these methodological aspects, see Ginzburg (2000, p. 115).
Regarding Sraffa in relation to Whitehead, we have found only a reference, to Science in the modern world (D3/12/10.25). Sraffa’s library contains two books by Whitehead, with frequent marginal annotations: Introductory mathematics (1911) and Science and the modern world (1927).
17 It is interesting to note the unusual term ‘price of wages’ employed by Sraffa. The term is sometimes used by Ricardo to indicate the difference between wages as workers’ subsistence, given by a list of goods and by an absolute level, and wages as real cost of production for the capitalist, given by a proportion between product and costs. For a detailed analysis of the Ricardian notion of the ‘price of wages’, see Gerke (2003).
Actually, in both cases, wages can be expressed as a proportion, with the difference that subsistence wages are a proportion of gross income, and surplus wages a proportion of net income. They can also both be expressed as a list of commodities, the difference is that subsistence commodities are to be understood as necessary consumption used in production, and consumption out of surplus-wages as final consumption. In fact, wages are always perceived by labourers as an inventory of commodities and by capitalists as a proportion.
18 Sraffa notes also that the meaning of the two rates of surplus is different, in particular: ‘Rate of wages is “commodity per head per unit of time”. Rate of interest is “money per cent per unit of money”’ (D3/12/7.141).
91 In this regard, it is interesting to note that Sraffa was well aware of the anthropological aspects inherent in the question of the nature of value. In this connection, his references to Malinowsky (1929) are very interesting (D3/12/7.12).
20 On the ‘moral economy’ of those struggles, see Thompson (1971).
21 It is important to note that, in modern welfare states, which partly descend from the Poor Laws (Picchio, 1986, 1992, pp. 57–73), public services for health and education, given their universalist content, have been the most effective means of redistribution between and within classes (Piketty, 2002, pp. 144–5).
References
Coletti, F., 1905, ‘Il costo di produzione dell’uomo e il valore economico degli emigranti’, in Il giornale degli economisti, vol. XXX, febbraio.
De Vivo, G., 2000, ‘Produzione di merci a mezzo di merci: note sul percorso intellettuale di Sraffa’, in M. Pivetti, a cura di, Piero Sraffa: contributi per una biografia intellettuale, Roma, Carocci.
Edgeworth, F.Y., 1906, ‘Income’, in I. Palgrave, ed., Dictionary of Political Economy, London, Macmillan.
Garagnani, P., 1998, ‘Sui manoscritti di Pierro Sraffa e il lavoro per la loro pubblicazione’, Rivista Italiana degli Economisti, vol. 3, 1, April, pp. 151–6.
Gehrke, C., 2003, ‘“Prices of wages”: a curious phrase’, paper presented at the ESHET Conference, Paris, February, 2003.
Ginzburg, A., 2000, ‘Sraffa e l’analisi sociale: alcune note metodologiche’, in M. Pivetti, a cura di, Piero Sraffa: contributi per una biografia intellettuale, Roma, Carocci.
Kurz, H., 1998, ‘Against the current: Sraffa’s unpublished manuscripts and the history of economic thought’, in The European Journal of the History of Economic Thought, vol. 5, 3, pp. 437–51.
Malinowsky, B., 1929, The sexual life of savages, London, Routledge.
Marshall, A., 1920, Principles of economics, London, Macmillan.
Naldi, N., 2000, ‘The friendship between Piero Sraffa and Antonio Gramsci in the years 1919-1927’, in European Journal of the History of Economic Thought, vol. 7, 1.
Pareto, V., 1905, ‘Il costo di produzione dell’uomo e il valore economico degli emigranti’, in Il giornale degli economisti, vol. XXX, aprile.
Petty, W., 1691 [1672], Political anatomy of Ireland, London.
Picchio, A., 1986, ‘Poor laws’, in J. Eatwell, M. Milgate and P. Newman, eds, The new Palgrave dictionary of political economy, London, Macmillan.
Picchio, A., 1992, Social reproduction, the political economy of the labour market, Cambridge, Cambridge University Press.
Picchio, A., ed., 2003, Unpaid work and the economy, London, Routledge.
Pigou, A.C., 1946, Income: an introduction to economics, London, Macmillan.
Piketty, T., 2002, L’Economie des inégalités, Paris, Édition la Découverte.
Pivetti, M., 2000, ‘Il concetto di salario come ‘costo e sovrappiù’, in M. Pivetti, a cura di, Piero Sraffa: contributi per una biografia intellettuale, Roma, Carocci.
Ricardo, D., 1951, On the principles of political economy and taxation, P. Sraffa, ed., with the collaboration of M. Dobb, The works and correspondence of David Ricardo, vol. I, Cambridge, Cambridge University Press.
Roncaglia, 1974, ‘Labour-power, subsistence wage and the rate of wages’, Australian Economic Papers, vol. 13.
Sen, A., 1985, Commodities and capabilities, Amsterdam, North Holland.
Smith, A., 1976a, An enquiry into the nature and causes of the Wealth of Nations, R.H. Campbell, A.S. Skinner and W.B. Todd, eds, Oxford, Oxford University Press.
Smith, A., 1976b, An Enquiry into the Nature and Causes of the Wealth of Nations, E. Cannan, ed., with a new Preface by George J. Stigler, Chicago, University of Chicago Press.
Smith, A., 1978, Lectures on jurisprudence, R.L. Meek, D.D. Raphael and P.G. Stein, eds, Oxford, Oxford University Press.
Smith, J., 2000, ‘Le carte di Sraffa presso la biblioteca del Trinity’, in M. Pivetti, a cura di, Piero Sraffa: contributi per una biografia intellettuale, Roma, Carocci.
Sraffa, P., 1960, Production of commodities by means of commodities: a prelude to a critique of economic theory, Cambridge, Cambridge University Press.
Stanley, A.D., 1999, ‘“We did not separate man and wife, but all had to work”: freedom and dependence in the aftermath of slave emancipation’, in S.L. Engerman, ed., Stanford, Stanford University Press.
Thompson, E.P., 1971, ‘The moral economy of the English crowd’, in Past and present, vol. 50.
Whitehead, A.N., 1911, An introduction to mathematics, Home University Library, London, Williams and Norgate.
Whitehead, A.N., 1927, Science and the modern world, Cambridge, Cambridge University Press.