84

).* = L(Pj)-I(Pj)(I-A)-I(Pj)-1
).* {L(I-A)-I}(P)-I
4
).* = ).(Pj)-I
Thus, the j-th element At = ))Pj. That is, each element of the row vector). * is Marx, Sraffa and the N eo-Classicals
in fact the ratio of total labour requirements per unit output. Clearly, this
ratio is independent of any choice of the unit of output (lbs., tons, etc.) in Context
4. The preceding results point to a simple way of deriving the data
necessary for our calculations. Beginning with the empirical input-output Hector Guillen Romero
matrix A* and the corresponding vector L* direct labour requirements per
dollar of output, we can immediately calculate ).*, total labour requirements
per dollar of each sector's output. These correspond to the data we used from I. Introduction
Leontief. The elements of).* are ))Pj. Hence if we know the gross sales p/Xj for
each sector, we can immediately derive the total labour requirements )'h
which correspond to these sales (even though we do not at any time actually Although the terms 'neo-Ricardian school' or 'Cambridge school' are
define any units of output x). sometimes applied to the contribution on growth theory of Joan
Robinson, Nicholas Kaldor, Luigi Pasinetti and others, in this work I
AjXj -- p; PjXj -
)'j -
Aj*(PjXj) deal almost exclusively with the theory of prices elaborated by Piero
SratTa. 1 The school's methodological foundations can be found in
The last operation gives us total labour requirements Ah in worker-years and the works of Dmitriev and von Bortkiewicz, who wrote at the turn of
total prices (gross sales) pjX j in dollars. the century.2 These authors' importance has grown recently following
the publication of SratTa's work.
5. Two data sets were used, in which Aj = AjXj and Pj = pjXj are derived in The neo-Ricardian school can be assessed either in relation to
manner indicated in 4 above. Defining the average value of the dollar as a == Marxist value theory or in relation to neo-classical theory, whether in
([ A)I(I P), we can then use this to define total direct prices p jO = (l/a)A j. its vulgar form (Jevons, Menger, Marshall) or its general equilibrium
Finally, both Pf and P j are expressed as prices relative to their respective form (Arrow, Hahn, Malinvaud, Walras).
average prices pO = (I P?/N) and P = (I PiN). Note that by construction,
pO=P.
The first data set is based on Leontiers 1947 data, from W. Leontief, J. N eo-classicals and N eo- Ricardians
Input-Output Economics, Oxford, New York, 1966, appendix III, pp.
129-133. Total Sales P j were taken from US 1947 input-output table, 192- The neo-Ricardian school rejects subjective individualism and the
order. role of supply and demand in the determination of income distri-
The second set was provided by Edward Wolff of New York University. In bution. A recognition of the class division of society is central to its
this data, the direct labour requirements vector was computed in two ways: analysis. 3 It makes an internal critique of vulgar economics, showing
first, in worker-years of undifferentiated labour requirements; and second, in that many of its propositions are inconsistent with its own assump-
a skill-weighted index of worker-years where relative wages were used as tions. In particular SratTa shows that neo-classical capital theory is
weights (for lack of better indexes). The latter data are the ones actually incoherent and indeterminate in its vulgar version. This critical wing,
shown, but the regression results are substantially the same with either set. fully developed since SratTa's book appeared, seemed to have
Lastly, both the graphs and regressions leave out the real estate and rental
sector, since on theoretical grounds within both the Ricardian and Marxist culminated in the works of the Cambridge school on macro economic
theories of rent, though the magnitude of rent can be derived from value production functions and problems related to the choice of
relations it is not related to any labour-time expended in the collection of techniques. 4
rent (in the real estate and rental sector). Once again, however, this makes While these attacks are fundamentally internal, the neo-Ricardians
little difference to the log-regression results. make a basically external critique of general equilibrium theory.

85
86 Guillen 87

Indeed, as A. Medio recognizes,s general equilibrium theory is accept, with Dobb, that Marxist value theory is essentially a theory
unaffected by the attacks levelled at the vulgar [Link] o~ the about 'conditions of production' in which relations of exchange have
logical plane. His argument is that concepts such as capI~al, profit, or a subsidiary or almost superfluous function, we can easily derive the
interest play no essential role in this theory, which IS unable to similarity between Marx and Sraffa.
confront socially important problems because it lacks the necessary A very similar line has been followed by other Italian authors. For
conceptual tools. For Medio, its weakest [Link] points a~e Alessandro Roncaglia,9 Marxist theory offers an underlying inter-
the individualism used in the study of the behavIOur of economic pretation of capitalist society through which Sraffian concepts such
agents, the technologism used in studying the production process, as commodity, price, wage and profit can be understood. The
and the concept of an economic system as an exchange economy. The problem of relative prices, which Sraffa confronts, is for this reason
concepts of class and social production relations .naturall~ d? not described as 'internal' to the study of the capitalist system, pre-
enter equilibrium theory. It is, furthermore, essentially static, III ~he supposing the institutional framework studied by Marx.1 0 Thus for
sense that it has no means of studying processes~the laws of motIOn Roncaglia, many of Sraffa's concepts only find their fullest ex-
which lead the system from one state to another. planation in Marx's more general analysis. Sraffa, he says, is writing
after Marx, and can thus presuppose the Marxist analysis of
2. Marxists and neo-Ricardians capitalist society without having to restate it. Roncaglia even claims
that thanks to Sraffa Marxism now has a more scientific foundation.
Many economists have dealt with neo- Ricardian [Link] as if it were in
continuity with the Marxist tradition, a view decreaslllgly acceptable
3. The Importance of Studying Neo-Ricardianism
to the Sraffians themselves and which this paper sets out to refute.
Ronald Meek and Maurice Dobb have hailed Sraffa for Bourgeois economic thought is divided into two camps, neo-classical
'rehabilitating' Marx. Meek, in his introduction to the second edition and neo-Ricardian. The neo-Ricardians have rejected marginalist
of his Studies in the Labour Theory of Value 6 presents Sraffa's system positions en bloc after an internal critique, and have returned to the
so as to show how some of its basic elements can be adapted and used earliest formulations of bourgeois economics, particularly to those of
by modern Marxists. Meek presents a sequence of five ~od~ls .of the the physiocrats, of Smith and of Ricardo. We are dealing here with
Sraffa type linked by a kind of 'logical-historical' analYSIS slm~lar to what Sergio Latouche calls the reswitching of dominant ideologies. I I
Marx's. From the outset he claims to show that a modern Marxist can The neo-Ricardian position is not yet hegemonic within dominant
reformulate and develop Marx's original theory with commodities economic thinking, but is making progress-although in recent years
themselves, rather than their values, taken as 'concrete prior the neo-classical school has been in the ascendant under the pressure
magnitudes'. For Meek the transformation problem can be prope.r~y of conservative policies and monetarism. The critical presentation of
resolved only by postulating, in one form or another, speCifiC neo-Ricardian theory is very important, not just because this theory
interrelations between inputs and production. Finally, he concludes plays a role in dominant economic thinking but because it has begun
that Sraffa's procedure reflects Marx's basic idea, th~t prices. and to win some influence within the workers' movement. Sraffa's
incomes are in the last instance determined by productIOn relatIOns, emulators, for instance, are advisors to the FLM, the most important
more clearly and effectively than Marx's.7 industrial trade-union federation in Italy. Moreover the 'Cambridge
Maurice Dobb holds a similar view. 8 He affirms that what is School' has its defenders in the heart of the Italian Communist Party
particularly striking (some would say revolutionary) about the and among many on the far left. 12
Sraffian system is its rehabilitation of the Ricardo-Marx approach to Just as alarming were views expressed by participants in a value
problems of distribution from the production side, so that relative theory seminar at Modena University in 1978. Most accepted that
prices are independent of the pattern of consumption and demand. Marxist value theory was invalid, and the entire discussion turned on
From Sraffa onward there have been many mathematical models whether this jeopardized the whole of Marx's theoretical edifice or
in which relative prices are derived directly from the conditions of only part of it. The commonest positions were those of Garegnani and
production without being affected by the pattern of demand. If we Coletti. For Garegnani, essentially, the labour theory of value does
88 Guillen 89
not apply. The concept of surplus can supplant that of v~lue and can
reduces to that of finding the relative prices of the various com-
be used, as by Sraffa, to determine prices and the profIt rate. 13 In
modities. These prices should be such that, respecting the law of
Garegnani's view this does not prevent us referring ~ack to Capi~al f~)f
equality between the 'values' of production (prices of phys~cal
an explanation of phenomena such as accumulatIOn and capItalIst
quantities of products) and the 'values' of costs (prices of physIcal
crisis. Coletti agrees that Marxist value theory is invalid but argues.
quantities of means of production), the initial position of the system
more coherently than Garegnani, that this calls into question the
can be restored. The system of prices of production can be formalized
greater part of Marx's analysis, since one can neither define the
as follows:
concept of commodity nor that of competition without the concept of
Let us denote by A, B, ... , K the quantities of commodities
value in Marx's system. 14 Another position worth considering for its
a, b, ... , k produced annually. Let us denote by Aa, B b, ... , Ka the
implications is Napoleoni's. He holds that the autonomy of th~ value
quantities of these commodities needed for the production, of A, and
category in Marx is such that, even if it cannot be used to derIve the
similarly for B, ... , K. The amounts produced and the technological
rate of profit or prices, it can and must be discussed at a philosophical
requirements of production are given as data. The unknowns to be
level. 15
determined are the prices of the k commodities Pa' Pb' ... , P k • As we
Although neo- Ricardian theory can be assessed either in relation to
can see these prices result from the technology and they play no role in
Marxist value theory or to neo-classical theory, in this essay I
the assignment of resources. The only thing they do is signal those
concentrate almost exclusively on its relation to Marxist theory. The
'values' of units of the commodities, a, b, ... , k which, if adopted, will
essay is divided into two parts. In the first part I give a summary of
allow the initial position of the system to be re-established. This
Sraffa's scheme, and in the second I assess its relation to Marxist subsistence system can be written as follows:
analysis.
AaPa + BaP b+ ... + KaP k = AP a
II. Piero Sraffa's System of Prices of Production AbP a +BbP b+ ... +KbP k = BP b

Piero Sraffa's work is essentially a study of prices of production and

the influence on them of distributional variables such as the wage rate
and the profit rate. He does not take into account problems Since by the definition of subsistence economy A = Aa + Ab +
concerning production and employment levels, income distribution, ... + Ako and the same for B· .. K, anyone of the equations can be
growth and such like. 16 This is because he wants to concentrate on an deduced from the sum of the others. Therefore this system contains
economic system whose properties do not depend on variations in the k -- 1 independent equations which determine k - 1 relative prices.
scale of production or in the proportion of its 'factors'. This precise The prices are expressed in terms of a commodity chosen as unit of
limitation on the object of study renders it susceptible to an 'exact' measurement, whose price is set equal to 1.
treatment, in the same sense as in the mathematical sciences. One of the effects of the appearance of a surplus is the introduction
Sraffa's analysis is developed in four stages. He first presents us of the notion of luxury, or non-basic goods. Luxury products are
with a perfectly closed productive process, that is to say a process in those used neither as means of production nor as means of su bsistence
which the same commodities appear both as means of production and in the production of others. Obviously, this category of product could
as products. Furthermore, the quantities produced of each good not exist in a subsistence model, since the surplus being by definition
(which are taken as given) are exactly equal to the quantities used as zero, every commodity played the role of product and instrument of
means of production. In this 'subsistence' economy, there is no production at one and the same time.
surplus, but simply the minimum necessary for the economy to Luxury products do not figure in the determination of the system.
reproduce itself in the same' "form and with the same dimensions. Their role is purely passive. If, for example. says Sraffa, we were to
Labour (and its wage) are represented as any other commodity, that is eliminate one equation corresponding to the production of a luxury
to say they form part of the means of production. 1 7 The problem thus good, we would eliminate an unknown (the price of this good) which
90 Guillen 91

would appear only in this equation and the remaining equations those goods needed for the workers' subsistence as means of
would continue to form a determinate system which would admit of production, like fuel, and deal with the variable element as part of the
the same solutions as the previous one. This would not happen if we surplus of the system. We then have a net or surplus product which
were to eliminate a non-luxury good, which Sraffa designates a basic divides up into wages and profits. The wage rate and the profit rate
product (a good which enters into the production of every good, are not simultaneously determined by this equation system, so that
either directly or indirectly). one of the magnitudes has to be externally determined and the other
In the second stage Sraffa introduces a surplus. He maintains the will be determined as a function of it. In this case, as Sraffa indicates
hypothesis that inputs and outputs are made up of the same 'the system can move with one degree of freedom; and if one of the
commodities. The difference is that the technological structure is such variables is fixed the others will be fixed toO.'18
that the quantity of each good produced can be equal to or greater If we call the given quantities of labour of uniform quality La'
than that used as means of production. The 'value' of production Lb, ... ,Lk (not labour power as Altvater, Hoffman and Semmler
(price of the physical quantity produced) will be greater than that of erroneously do !)19 and if we denote by w the wage per unit of labour,
the costs (price of the physical quantity consumed) because there is a the system of production prices appears in the following form:
surplus. The equation system which expresses this schema simul-
taneously determines the set of relative prices and the general rate of (AaPa +BaP b+' .. + K aP k)(1 +r)+Law AP a
profit. This general rate of profit manifests the fact that the surplus (or (AbP a +BbP b+' .. + K bP k)(1 +r)+Lbw BP b
profit) is distributed to each productive activity in proportion to the
'value' of means of production used up. As in the preceding stage, (2)
labour is not present in its direct form, but only through the
commodities consumed by workers. The workers, and these com-
modities, appear as part of the means of production. If we set the price of the net total product equal to 1, we will have the
In this case we denote the rate of profit by r, and the system of prices following additional equation:
of production can be written as follows:
[A-(Aa + Ab + ... + Ak)]P a +[B -(Ba +Bb + ... +Bk)]P b+ ...
(AaPa+BaP b +" '+KaP k)(1+r) = AP a
+[K-(Ka+Kb+" . + Kk)]P k = 1 (3)
(Ab~a+Bb~b+" '+Kb~k)(1:+r) = B~b
(1) The system will now consist of k + 1 equations and k + 2 variables
(k prices, the wage wand the rate of profit r). The system has one
degree offreedom: if we now fix the level of one of the distributional
Since A ~ Aa + Ab + ... + Ak; B ~ Ba + Bb + ... + Bko ••• , with at least variables (wage or profit rate) as an independent variable, we can
one strict inequality, we have k independent equations which simul- determine prices and the other distributional variables.
taneously determine the k - 1 relative prices and the rate of profit.
In the third stage, Straffa changes the assumptions concerning The Search for the Standard Commodity
wages which he made in the two previous stages. Previously it was
assumed that wages were represented by the means of subsistence One of the important aspects of Sraffa's work is that it introduces a
needed by the workers, so that they appeared in the system on the very special unit of measurement of value. In order to understand this
same level as vehicles or fuel consumption. But in reality wages can measure we must refer to Ricardo's aims in explaining the general rate
contain not just the ever-present element of subsistence (which is of profit beginning from the rate of profit in agriculture.
constant) but also a share of the surplus (which is variable). If we assume as Ricardo did in his famous essay of 1815 20 that in
In this situation, the most correct procedure would be to divide the agriculture only one good is produced---corn-and that the sub-
wage into its two component parts, that is to say, to continue to treat sistence of agricultural workers consists only of com, then the rate of
92 Guillen 93

profit in agriculture, that is to say the relation between the surplus capitalist economy. In fact for him, if the natural price of a commodity
agricultural product and what was set aside for the agricultural were equal to the amount of wages paid in order to obtain it,
labourers' means of subsistence can be calculated directly as such, everything would be very simple. An object for which twice as much
without recourse to the price of goods, provided that the terms of the had to be paid would necessarily be an object into which twice as
relation are physically homogeneous. Thus, in the agricultural sector, much labour had gone. But- the price includes capital's profit.
the rate of profit does not vary as a result of changes in real wages. But Can we say that the profit on capital is the remuneration of a kind
since the rate of profit has to be the same in all productive activities, of work, that ofthefirm's management? No, because for Smith profits
the price system must be such as to allow equality of rates of profit in and wages are determined by completely different principles. Profits
other sectors with that which holds in agriculture. depend solely on the amount of capital employed and are greater or
Malthus exposed an important defect in this reasoning. There is no small depending on the size of this capital. Smith did not think it
economic sector, not even agriculture, where both capital advanced possible to maintain that the exchange value of products would be
and all the results of production consists of a single product. Wages determined by labour costs in the economy he contemplated. But
are not made up of com alone. Workers consume manufactured intending to maintain a relation between exchange value and labour,
goods. This means that the calculation of the rate of profit involves a he finally declared that the normal price of each object corresponded
comparison of aggregate of heterogenous goods, with the product, to the amount oflabour which could be 'commanded', that is, bought
wages and total investment. But in order to compare these with the object-a solution which failed to overcome the problem of
heterogenous goods they must be converted into common units. the dependence of prices and profit rates.
To overcome this objection, Ricardo sought a unit of value capable To avoid this complication of the interdependence of relative prices
of measuring, as if dealing with a single quantity, the heterogeneous and the rate of profit, Ricardo tried to find an 'invariable measure of
mass of produced goods. His theory required a measure of value which values'. He scouted a number of possibilities, one of which was the
would make it possible to convert into a homogeneous unit of elaboration of a theory of labour-values, but he realized that labour-
measurement, analogous to the unit of com in his simple model, the values did not precisely reflect relative prices. He also attempted to
heterogeneous groups of goods divided up in the form of rent, profit take an 'average' good as a standard, but realized that only limited
and wages. Now, the heterogeneous goods could be converted into a progress could be made in this direction. In fact, as he recorded in his
homogeneous measurement as a function of their relations of Principles, 'when commodities varied in relative value it would be
exchange on the market, that is to say, as a function of their relative desirable to have the means of ascertaining which of them fell and
prices. But from this a complication arises, since these prices depend which rose in real value, and this can be effected only by comparing
on the rate of profit. them one after another with some invariable standard of measure of
In confronting this same problem, Adam Smith had previously value, which should itself be subject to none of the fluctuations to
tried to formulate a theory of labour value. 21 In order to explain which other commodities are exposed'. Unfortunately, he adds 'Of
exchange-value, he begins from a hypothetical society in which such a measure it is impossible to be possessed, because there is no
everyone works and exchanges the products of their labour. In such a commodity which is not itself exposed to the same variations as the
society, according to Smith, products would be exchanged in things the value of which is to be ascertained; that is, there is none
proportion to the quantity of labour needed for their production. If which is not subject to require more or less labour for its pro-
this did not happen, some members of society would lose out and the duction.'22 This idea was maintained in his final written work, shortly
exchange system could not function. In constructing this model, before his death, in which he confessed that 'there is no perfect
Smith tried to refer to a primitive type of society. But contemporary measure of value in nature'.23
sociologists have shown that primitive exchange differs greatly from Ricardo's idea of using an 'average good' as a standard of value has
Smith's conception. In reality, we have had to wait until capitalist resurfaced with Sraffa. Sraffa shows how such a good can be
economies came into being to find a way of life in which labour conceived as a composite good and used in the analysis of income
expended in production determined exchange relations. Smith, on the distribution over a given period in an economy which produces
other hand, thought that his explanation of value was not valid for a reproducible goods. 24 He analyses the effects of a variation in wages
 Guillen 95
94

on prices and on the rate of profit, taking into account the fact that the (Aaqa + Abqb + ... + Akqk)(l + R) Aqa
rate of profit is the same in all branches-the hypothesis of the
(Baqa + Bbqb + ... + Bkqk)(l + R) Bqb
equalization of profit rates.
He supposes furthermore that production methods do not change (4)
and that quantities produced remain given. Under these conditions
Sraffa seeks a commodity which, although it will be no less
susceptible than any other to rises or falls in price with respect to
We can define the units in which the multipliers must be expressed
individual commodities,25 as a result of movements in the wage rate,
thanks to an additional equation which incorporates the condition
will be such that we would know with certainty that 'any such
that the quantity of labour embodied in the standard system is the
fluctuations would originate exclusively in the peculiarities of pro-
same as in the actual system being studied:
duction of the commodity which was being compared with it, and not
in its own.'26 (5)
'It is not likely that an individual commodity could be found which
possessed even approximately the necessary requisites', he notes. 27 Here we have a system of k + 1 equations which determine the k
But a 'composite commodity' could be produced, that is, an aggregate multipliers and R.
of commodities such that the commodities which compose it also Solving this equation system, we obtain a set of numbers for the
figure, in the same proportions, in the means of production of the mulipliers (q~, q~, ... , q~) which Sraffa applies to the equations of the
aggregate. Sraffa calls this aggregate the standard commodity and production system,29 converting them into a standard system in the
uses the term standard system to refer to the set of industries following manner:
concerned, taken in the proportions needed to produce the standard
commodity. q~[(AaPa + BaDb + ... + KaPk){l + r) + Law] = q~APa

q~[(AbPa+BbPb+" '+KbPk){l+r)+Lbw] = q~BPb

(6)

The Standard System From here Sraffa derives the standard national income, which he
adopts as a unit of measure of wages and prices for the original
The formal method of construction of the standard commodity is production system.
equivalent to taking a set of K appropriate multipliers which can be The equation which tells us that the price of the net product is equal
designated qa' qb' ... , qk and applying them respectively to the to 1 is replaced by the following, in which the q' represent known
equations of production of the commodities A, B, ... , K. 'The numbers, whilst the p are variables:
multipliers', says Sraffa, 'must be such that the resulting quantities of
the various commodities will bear the same proportions to one [q~A - (q~Aa + q~Ab + ... + q~Ak)]P a +
another on the right-hand sides of the equations (as products) as they
[q~B-(q~Ba+q~Bb+ ... +q~Bk)]Pb+ ... +
do on the aggregate of the left hand sides (as means of production).'2s
This implies that the percentage by which the production of a [q~K-(q~Ka+q~Kb+ "'q~Kk)]Pk = 1 (7)
commodity exceeds the quantity entering as means of production will
be equal for all commodities. This percentage Sraffa calls the !his 'composite' commodity is the standard of wages and prices. The
standard relation, and denotes it by R. Introduction of the normalization conditions into the standard
Under these conditions the system q can be written down as system means that there are now k + 1 equations (K price equations
follows: and 1 normalization equation) and k + 2 unknowns (k prices and two
96 Guillen 97

distributional variables}. The system still has one degree of freedom. system of equations which, with given wages, determines the price of
In the standard system, the relation between net product and all products, the rate of profit, and the rent on land of different quality.
means of production can be calculated in physical terms, given that
we are dealing with two aggregates in which the commodities making
them up are the same. In this framework, the standard commodity
III. Problems of Interpretation
plays the role of Ricardo's 'com'. With the help of the standard
commodity, Sraffa only resolves part of the problem which Ricardo
1. SrajJa's Theory of Prices of Production in Relation to
fails to solve in passing from com to embodied labour. In effect,
Marginalist Theory
Ricardo was looking for a standard which would be invariable, both
in relation to changes in the conditions of production of commodities, Attempts have been made to characterize Sraffa's theory using
and in relation to changes in the distribution of incomes under fixed marginalist categories. In doing this some authors have committed
production conditions. Sraffa abandons the search for an invariable many errors. For example. H. G. Johnson has indicated that Sraffa's
commodity with respect to variations in the conditions of production, theory represents a system of 'incomplete' general equilibrium, in the
and his analysis leads him to the search only for a standard of prices sense that he down plays consumption (demand) by speaking only of
which would be 'invariable' when income distribution changes, production (supply).30 The same idea is expressed by R. F. Harrod
taking production conditions as given. when he asserts that 'the most notable features of Sraffa's book is that
Thus, as the rate of profit obtaining in com cultivation was, for he makes no reference to the scale or elasticity of demand for final
Ricardo, the rate of profit which would impose itself as a general rate products, when one of the central themes that he takes up is the
of profit, Sraffa shows in the same manner that if the wage is expressed determination of prices ... It is surprising', he adds 'that a price
in terms of the standard commodity then the rate of profit which, in system can be determined without reference to final demand.'31 Joan
the standard system, is obtained as a ratio between quantities of Robinson reflects along the same lines, when she asserts that 'he has
commodities, will give rise in the actual system to a relation between only given us half an equilibrium system.'32
aggregate values. More specifically, if we recall that R is the relation However, the reference to general economic equilibrium is
established in the standard system between the net product and the incorrect. Sraffa looks at prices of production, determined on the
means of production, and that it is therefore the maximum rate of basis of a hypothesis that profit rates equalize. He is interested in a
profit for the real system, and that w, the wage, is expressed in terms of different problem from the marginalist problem of 'equilibrium
the standard commodity (remember that, as with Sraffa, the total prices' which would guarantee the equality of supply and demand. 33
quantity of labour is set equal to one, wage and wage rate coinciding), Just as it is wrong to speak of an 'equilibrium system' in the
the prevailing rate of profit in the real system is given by r = R(l- w). traditional, that is 'marginalist' sense of the term, it is also incorrect to
After showing that the standard system is unique, Sraffa makes speak of a 'general system'. Sraffa takes into consideration in his
abundant use of the relation between wage rate and profit rate to deal analysis only those factors necessary for the resolution of his problem.
with many theoretical problems. Thus, for example, he analyses the For this reason he separates out all elements which by definition
case in which commodities are produced with means of production exercise no influence upon prices of production, or whose influence is
which have been produced at different points in the past (and thus in exercised via distribution, technology and the scale of production,
succession) so as to show that the profit element in the prices of these elements which are given in Sraffa's system. It is thus not possible
means of production is different, and he asks how relative prices of to speak of a partial analysis in the neo-classical sense, since Sraffa
commodities will change with changes in the rate of profit. does not concentrate on one part of an economic system with the aim
In the second part of his book, Sraffa studies new problems which of producing an approximate solution to a problem whose true
arise out of looking at the existence of branches of industry with solution can only be ascertained within the framework of a more
multiple outputs Uoint production) and fixed capital. Sraffa further general analysis. Rather, he considers all those elements necessary for
introduces land into his analysis and constructs a more complex the solution of the problem as it is posed.
98 Guillen 99

What we have just said is valid for the distinction which neo- quantities produced and used in production, so that prices do not
classical economists frequently make between static analysis and depend on the forces of supply and demand. On the other hand. these
dynamic analysis. According to the neo-classicals. d~namic ana~ysis is forces play a central role in the neo-classical system, since they
characterized. essentially. by the inclusion of vanables relatmg to simultaneously determine quantities and prices, in order to finalize the
different times. More specifically. Harrod asserts that dynamic equilibrium process. This calls for a hypothesis about returns, and
analysis includes 'propositions in which the rate of growth appears as about the form of the demand functions, before prices are known. For
an unknown quantity.'34 This idea is reiterated by Hicks, who asserts the neo-Ricardians, the hypothesis about returns is only important
that in order to make a dynamic analysis every magnitude must be for the theory of accumulation, but is logically independent of price
dated. 35 From this point of view. we can define a static analysis as an theory. Therefore, the neo-classicals insist on consumers' preferences,
economic analysis which is not dynamic. By this definition Sraffa's on utility and more generally on the problems of individual choice.
analysis of prices of production would be static. However, if we These quantities are of little interest to the neo-Ricardians who
attempt a positive definition, looking at the 'static theories' of the construct an analysis in terms of social classes (even though their
neo-classicals. we can see that what typifies them is their atemporal concept of social class is far removed from that of the Marxists), and
context. They try to interpret the values of the variables under are not interested in the logic of individual behaviour.
consideration as equilibrium solutions to the economic system under Everything that we have just said makes Harrod's aims in respect of
consideration. Sraffa illusory, in the sense that they attempt to establish inter-
From this point of view it is more correct to say that Sraffa does connections between the Sraffian system and the neo-classical
not make a static analysis but 'photographs' a moment of growth, system instead of dealing with the Sraffian system as a 'prelude to the
which is very different. He makes no abstraction from ti~e, since th.e critique of economic theory'.n From our point of view, the two
moment under consideration is determined by past hIstory and IS systems can only coexist with the greatest of difficulty.
limited to generating the following moment in time .. Thu.s [Link]
errors of interpretation of Sraffa's work have theIr ongI~ m .an
insufficient comprehension of the difference between neo-RIcardian 2. Sraffa's Theory of Prices of Production in Relation to Marxist
economics and neo-classical economics. Analysis
Neo-Ricardian economics is that which, on the basis of the
existence ofa physical surplus, studies its distribution thanks to a price A. The Standard Commodity and the Status of Labour in the N eo-
system with the restriction that the [Link] [Link] must Ricardian Economy
reproduce itself. In all this, the hypotheSIS of equahzatton of profit
rates plays a decisive role. Neo-classical economic~, us!ng the c~ncept The standard commodity is the central element in the analysis of
of factor of production, leads towards the determmatton of pnces of prices of production. We know that the standard system is built from
goods and services of the factors of production, and therefore the ~ate the initial production system by taking as data the quantities
of profit, since capital is conceived as a factor o~ productIOn, produced and the conditions of production. To each production
corresponding to an equilibrium amongst all economIC agents. system there corresponds a unique standard system. It is necessary to
The neo-Ricardian school rests on the notions of surplus and ascertain, Benetti asserts, 38 whether Sraffa's standard commodity
reproduction, whilst the neo-classical school rests on the notions of lets us study the process through which prices reach the levels defined
factor of production and equilibrium. The logical structures of both by the system of prices of production. This calls for a price standard
schools, like the categories they use, are very different. Thus, for defined when the profit rate is not equalized, since the system is in
example, the concept of profit, for the neo-Ricardians, is not the motion towards a final state not yet attained. In spite of what has been
reward due to a factor of production but a part of the surplus. achieved in this respect, such a construction is far from having been
Equally. the notions of capital and wage do not have the same demonstrated.
meaning for the neo-classicals and the neo- Ricardians. 36 Even if it were possible, we would have to compare the system at
The existence of a physical surplus implies prior knowledge of the two distinct points in time to make sense of the process of price
 Guillen 101
100
formation (or profit equalization). This is necessary in order to unit of measure of the wage is consumed or not consumed by the
determine the tendencies of prices and profit rates in different workers, changes in the wage will necessarily be translated into
branches of production to rise orfall. But at least one of the data of the changes in the quantities of commodities consumed by the workers
system----quantities produced-varies during this process. Th~s, we and, therefore, into changes in the production system. The problem
obtain different standards corresponding to different moments m the that we indicated previously simply reappears. 43
process. 'The comparison of prices determined at different points. in An intermediate possibility is to resort to the artifice suggested by
time is therefore impossible. That is, Sraffa's standard commodIty Pasinetti, although previously suggested by Sraffa of considering the
does not let us take into account a fundamental aspect of capitalist wage rate as divided into two parts: one portion necessary for
. ..
practIce----{;ompetItIOn.
,39
. . subsistence which is comparable with those commodities con-
Further, even accepting the uniformity of profit rates, It IS necessary stituting means of production (their composition being rigidly
to examine the suppositions on which the construction of the determined by biological necessities), and the other part forming part
standard commodity is based. The function of this commodity is to of the surplus. But, if this distinction is accepted, the subsistence
compare prices corresponding to a single production [Link] in portion of the wage rate will acquire the same status as a technical
different distributional states. It is built under the hypothesIs that datum and will be included in the matrix of technical coefficients. This
labour does not form part of advanced capital. For some authors this leads to a reinterpretation of Sraffa's system, making w refer only to
hypothesis is of no great importance. Thus Maurice ~obb assert.s that that part of the wage which forms part of the surplus. In this case when
'this is done merely for convenience in order to defme the maXImum w varies, no technical datum will change and prices corresponding to
profit of a standard commodity and demonstrate the effect of changes different wage levels can be compared. 44 Sraffa's standard com-
in the wage-profit couple on relative prices.' 'In Principle', adds modity will exist and the movement of relative prices will be
Dobb, 'nothing is implied by this change.'4o intelligible. However. it should be noted that the construction of
F. van de Velde argues that labour is omitted from advanced Sraffa's standard commodity suggests that at least part of the wage
capital only because 'the relation between profit rate and wage rate forms part of the net product.
appears simpler and clearer.'41 Henri Denis indicates [Link] 'for reasons In order to avoid ambiguities it should be made clear that the
that are not explicit (which are, I think, of a mathematIcal character) problem is not that the wage is paid postfactum in Sraffa, as several of
Sraffa eliminates variable capital, saying: we shall suppose that his critics have suggested,45 but the fact that the wage does not form
wages are paid after the produce has been sold----{;onsequently out of part of advanced capital. These are two different things and therefore
the income from the period under consideration, for example out of should not be confused. 46 For Marx wages are paid post factum and
the income for the year.'42 can still form part of advanced capital. 47 Thus, in the majority of
We shall study the consequences of suppressing this hypothesis, by branches of economic activity the wage is paid post factum but time
treating the wage as part of advanced capital. There are two must still pass until the capitalist can realize his merchandise, that is
possibilities: to treat it in the Ricardian manner as a bundle o~ goods to say, recuperate advanced capital (including variable capital) and
or a good, or consider it only as a price with ~ [Link]~ dIffe~ent obtain his surplus value.
status from that of a commodity price. In the fIrst case, It IS ObVIOUS The idea that at least part of the wage must be a fraction of the net
that any change in the wage will involve a change in the technical product is an indispensible condition for the existence of a standard
coefficients representing the quantities of wage-good. Thus, a dat~m commodity in Sraffa's sense. That is to say, Sraffa's system is only
in the system must be changed. Since there is a one-to-one relatIOn acceptable if at least part of the wage is solely considered as a
between standard systems and production systems there cannot be a distributional category. But, this is impossible, since the wage is only a
single standard system. It is impossible to compare prices corre- distributional category because it forms part of advanced capital, and
sponding to different wage levels. In this case, Sraffa's wage-standard is hence a production category.
does not exist and the movement of relative price is unintelligible. When the essential link between production and distribution is
We get into the same difficulties in the second case, [Link]~ the broken, as is the case with Sraffa, the 'wage' variable can no longer
wage in the form of a price. Whether the commodity whose prIce IS the designate the wage in its proper specificity. It could be interpreted
102
Guillen 103
as some kind of deduction from the net product: for example a tax on
have comprised, essentially, an explanatory theory of profit rates.
the net product of each branch, fixed at a un~form rate (the wa~e rate)
There are in existence several attempts to explain wage rate, but all
on the basis of a different amount (the quantIty oflabour used) m each
they do is make a vague allusion to the class struggle, avoiding the
branch. .
problem under study. Moreover they encounter the awkward
Furthermore, profit no longer appears as the means to an ulten~r
problem that wage bargaining by trade unions can have no meaning
accumulation, but is reduced to simple purchasing power. ~or ~h~s before the price system is known.
reason, it appears as identical in nature to the w~ge: fro?I which It IS
Attempts to fix the profit rate have followed four principal courses.
distinguished only by its spe~ific mo?e o~ ?Istnbut~on. between
The first possibility is to close the model following Sraffa's
branches. Profit is no longer defmed by ItS ongm (explOitatIOn of.t~e
suggestions in his work. He points out that the 'rate of profit, as a
labour force) nor by its destiny (accumulation): it is present, bu~ It IS
ratio, has a significance which is independent of any prices and can
not known from where it comes or where it goes. Wage and profit are
well be "given" before the prices are fixed.' To that extent, he adds, it is
neither distinguished by their origin nor by their destiny, but appear
'susceptible of being determined from outside of the production
as a pure purchasing power, as two masses of the same formless
system, in particular by the level of the money rates of interest.'5o
substance, distinguished from each other only because ~hey are
However this solution does not stand up to a careful study of the facts.
distributed between branches in two different ways. From this we can
Explaining the profit rate by means of the rate of interest only
deduce that Sraffa's system is unable to reproduce capitalism's
displaces the problem. What determines the rate of interest?
essential characteristic-the wage relation. The relation between
Moreover this solution assumes that the profit rate is regulated by
production and distribution is broken, and Marx's analysis of this
the monetary rate of interest. This last hypothesis can clearly be
relation bypassed. .
defended. Competition between capitalists guarantees that the profit
For Marx, 'The structure of distribution is completely determmed
rate is uniform and cannot permanently exceed the rate of interest.
by the structure of production. Distribution is itself a product of
But the correspondence between interest rates and profit rates is far
production, not only in its object, ~n. that on!y the results.?f
from close since the interest rate depends on many other factors. The
production can be distributed, but also m It~ form, m tha~ ~he specIfic
creation of a causal chain between interest rate and profit rate
kind of participation in production [Link] speclf~c f?rm~ of
presupposes that the interest rate can be fixed independently of the
distribution.'48 But production is also determmed by dlstnbutlOn.
profit rate, for example by the policy of a central bank, and that then
For example, 'A conquering people divides the land among the
the central bank takes such control over the firms that their rate of
conquerors, thus imposes a certain distribution and f~rms of property
profit is tightly linked to the rate of interest. As we shall see, the causal
in land and thus determines production. Or It enslaves the
conquer~d, and so makes slave-labour the foundation of production. chain is too subject to over-restrictive conditions to be really
acceptable.
Or a people rises in revolution and smas~es the gn~at !an~ed est~tes
A second solution, more commonly accepted, is to close the system
into small parcels, and hence, by this new dlstnbutlOn, gives
· a new ch arac t er. '49 with the Cambridge relation r = g/Sc, in which r is the rate of profit, g
pro d uctlOn
the rate of growth and Sc the capitalist propensity to save. 51 This
solution leads to many difficulties. In the first place we have to accept
B. The Closure of SrafJa's System a movement of causality from the rate of growth to the rate of profit,
The system established by Piero Sraffa is formally [Link], ~iven which is not at all evident and indeed seems without foundation,
production levels, producti?n methods a~d one of the dlstr~butlOnal involving as it does a simple dynamic equilibrium equation. Second
variables, prices of productIOn are determmed. However, this closure this solution rests on various assumptions: the rate of growth .is
is obtained by fixing one of the distributional variables (the rate of independent of the real wage, investment is financed by a fixed part of
profit or the wage rate) at an arbitr~ry level. AJogical closure of the profits and returns to scale are constant. Finally, it leaves unsolved
system demands either a the~ry which dete:mmes th.e w~ge .rate,. or the problem of what determines the rate of growth.
one which determines the profit rate. The mam efforts m this dlfectlOn A third solution assumes that the entrepreneurs make use of what is
known as a 'normal' rate of profit. This is the rate used by the
104
Guillen 105

entrepreneurs in their provisional economic calculations, which fixes This initial scheme in value terms is expressed in hours of labour and
the level of the profit rate. This idea is also not free from serious not in monetary terms. 54
difficulties. Nothing justifies the a priori assertion that the entre- Calculating the general rate of profit r as the relation of total
preneurs' understanding is so advanced that they all have the same surplus value S to the total capital advanced (C + V) we can specify the
idea of what constitutes a normal profit. This profoundly subjective system of prices of production:
element is being introduced into a price theory which is supposed to
be 'objective'. We must not make the determination of prices depend I: C 1 + Vi +r(C 1 + Vi) = G 1
on such a 'volatile' element as the entrepreneurs' 'animal spirits'. II: Cz+Vz+r(Cz+V z ) Gz
= (9)
Afourth solution makes the profit rate depend on the relation of
forces between social classes, that is to say, on the class struggle. With III: C 3 + V3 +r(C 3 + V 3) = G 3
this, according to some authors,5z we can reintroduce political
considerations into economics. Furthermore, it is thought that the C + V +r(C + V ) = G
authors who are turning to this solution are revitalizing the Marxist
approach to distribution. 53 This solution has the defect that it cannot We can verify that r(C + V) = S, that is that the total profits equal total
explain exactly what role is played by the class struggle in the surplus value, and furthermore the W = r, that is to say, total value is
determination of the rate of profit, nor can it be precise as to what equal to total price of production.
level, as a function of this struggle, the rate of profit must be According to his 'correctors' Marx's equations for prices of
established. production are logically incorrect since what enters into the price of
After reading the neo-Ricardians analyses one gets the feeling that production of a commodity, its cost of production, must also be
once the means of production have been deployed, the produce ofthe calculated in price of production terms. Inputs should not be
economy can be distributed in any way between capitalists and measured in terms of values but prices of production. In Marx's
workers without affecting the mode of production as such. In proposed solution, the same commodity is evaluated in two different
summary, the attempt to return to the class struggle is more of an alibi ways: as an input, that is to say as an element of the prices of
than the outline of a real solution. production, it is evaluated in value terms; as a product, that is to say,
as a result of the production process, it is evaluated in terms of prices
C. Sraffa and the Transformation Problem of production.
Sraffa's system of price of production is the logical result of a certain The solution proposed by many of the 'correctors' such as von
understanding of the problem of transforming values into prices of Bortkiewicz, Sweezy and Sraffa, though not Steedman, begins from
production. More specifically, it constitutes the logical and only a value system in which the conditions of simple reproduction hold.
result of the 'corrections' brought by Claudio Napoleoni into Marx's
scheme of prices of production, based on the work of von Bortkiewicz. I: C 1 +V 1 +S 1 C 1 +C z +C 3
The Marxist procedure for transformation is presented by Marx's
'correctors' in the following manner. Designating constant capital by II: Cz+Vz+S z V 1 +V Z+V 3 (10)
C and variable capital by V, surplus value by S and value by Wand III: C 3 + V3 +S3 Sl+ S Z+ S3
dividing the economy into three sectors, sector I producing means of
production, sector II producing wage goods and sector III producing In this system, as we have already said, according to Marx's
luxury goods, we get the following system in value terms: 'correctors', values are measured in quantities of labour. 55 In effect,
I:C 1 +V 1 +S 1 =W 1 the value substance, abstract human labour, is replaced by its
magnitude, units of labour time, and capital is simply reduced to
II: C Z+V Z+S 2 = W z (8)
inputs of labour time.
III: C 3 +V 3 +S 3 = W3 Let us suppose that the price of production of a unit of constant
C +V +S = W capital is x times its value, the price of production of a unit of variable
106 Guillen 107

capital is y times its value, and the price of production of a unit of (LaaP a + LahP b+ ... + Lak P k )(1 + 3)
luxury articles is z times its value. Furthermore, if we represent the
general rate of profit as r which is not defined from the value system (LbaP a + LbhP b + ... + L hk P k)(1 + r) (12)
but from the price of production system, we can write down the
following system for prices of production:

The system consists of k equations and k + 1 unknowns to be

determined (k coefficients of transformation and the rate of profit).
But since we have to make one coefficient of transformation equal to 1
the system can be determined without difficulty.
Since the L jh and the L j are the only data of the problem, one might
think that a knowledge of values is the logically prior condition to a
knowledge of prices and the rate of profit. However, if we look at
things more closely~say Marx's 'correctors' ~we can see that values
We have three equations and four unknowns to determine (the three can be replaced by physical quantities of commodities without the
coefficients of transformation and the rate of profit, that is x, y, z, r). logic of the system being altered.
Setting z = 1, the system can easily be solved. In effect, the matrix of values which can be written as:
But Marx's 'correctors' correct not only Marx but von Bortkiewicz.
Suppose constant capital comprises two commodities, a tractor and a LaaLab ... Lak]
thresher, whose values are C 1 and C 2 , where C 1 and C 2 add up to C.
The ratio of prices of production, resulting from von Bortkiewicz's [ LbaLbb ... Lbk
.
...
. ..
schema is C 1 XjC 2 X, equal to the relation between the values Ct/C 2 . In . .
the words of one of von Bortkiewicz's 'correctors': 'I assume that the LkaLkh ... Lkk
commodities which make up this constant capital exchange, amongst
themselves, according to values and not according to prices, because I can be replaced by the matrix of physical requirements of com-
apply a single coefficient of transformation of values into prices to the modities, that is to say by:
whole aggregate: which means that, in the interior of this aggregate, I
assume that the relations of exchange between commodities are those AaBa' .. Ka
which correspond to relations between values.'56 To obviate the
difficulty it is enough to rewrite the system 'but in such a way that the AbBb' .. Kb
equations refer always not to aggregates of commodities, but only to
individualcommodities.'57 To produce this effect we shall denote by
L jj the value of commodity j which is used as input in the production
of commodity i. Put another way, let us use L jj to denote the quantity
of labour contained in commodity j which is needed as input to For its part the vector [Ll' L2, ••• , Lk ] can be substituted for by the
produce that quantity of commodity i which incorporates an amount vector [A, B, ... , K]. Further still, if we define the units of physical
L j of labour. On the other side let us denote by Pa, Pb' ... , P k the quantites of commodities properly, that is, if we define a physical unit
coefficients of transformation of values into prices. These coefficients as the quantity of commodity which incorporates a unit of value, we
of transformation can be interpreted as prices of a unit of value. obtain the same numerical values for the two matrices and the two
With these specifications we can write down the following trans- vectors.
formation schema: This brings us, without difficulty to the system of prices of
 Guillen 109
J08
production presented by Sraffa at the beginning of the second relation between the two spheres, so essential for the explanation of
profit, is broken. Sraffa's posture, precisely by suppressing a specific
chapter of his book:
and essential term of Marx's analysis-value-the only term which
(AaPa + BaP b + ... + KaPk)(l + r) = APa captures and unifies the social complexity of capitalist economic
reality, is orthogonal to Marx's and can in no way be considered as
(AbPa+BbP b + ... + KbPk)(l +r)=BPb complementary to it. 59 Because of everything that has been said,
those who think that the history of the transformation problem has
ended with Sraffa should be considered as being more in continuity
with the Ricardian problematic than the Marxist.
Napoleoni, like all the neo-Ricardians, has not captured the nature
With this, the neo-Ricardians think that Sraffa has resolved the of abstract labour as social labour and the magnitude of value as
debate over transformation. Thus Napoleoni, who now proclaims the socially necessary labour time. He does not understand that money is
Marxist theory of labour to have a purely philosophical validity, the materialization of abstract universal labour time and that
says 'I have mentioned Sraffa's results as a confirmation of the capitalist society necessarily creates its own measure of value. As
possibility of determining prices and the share of profits in this way, Altvater, Hoffman and Semmler correctly argue: 'Sraffa abandons
independently of value. Furthermore, what Sraffa says may justifiably the analysis ofform which constitutes Marx's special contribution in
be taken as a full stop in the history of the transformation problem: 5 8 relation to Ricardo' ... From the moment value ceases to be directly
That is to say, for the neo-Ricardians Marx's transformation is social, the value of commodities can no longer be reflected in a
superfluous because prices of production and profits can be obtained directly comprehensible manner, since the value of each merchandise
without reference to value or surplus value. For the neo-Ricardians it must be expressed in terms of use values of other commodities. The
is an irrelevant detour to begin with values and transform them into universal equivalent is converted, in this exchange, into commodity-
prices of production. However, with Sraffa we situate ourselves money, so that "labour value" disappears as a measurable expression
within the system of prices of production and get a satisfactory theory, of human labour time ... :60
on the logical plane, of relations of exchange of commodities on the In commodity-producing societies it is a thing, for example gold as
basis of the uniformity of the rate of profit, but we lose any a money commodity, which takes on the task of representing value, so
comprehension of the nature of commodity, the origin of profit or the that Marx considered an understanding of the category of money a
social relations of production. Accepting Sraffa's theory as a solution prerequisite to understanding the essence of value. Those who have
to the transformation problem is to fail to understand Marx's not understood this and continue believing in the old utopia of
particular problem and, instead to pick up directly where Ricardo left 'labour-money' will benefit greatly from reapplying themselves to the
Grundrisse where Marx criticized such proposals forthrightly.
off.
The positions defended by participants at the Modena 1978 For the neo-Ricardians the transformation of values into prices is
University seminar, far from solving the problem, represent an an irrelevant detour, because prices of production can be calculated
evasion of it because they suppress one of its terms. The correct directly without reference to value. For the neo- Ricardian cu rrent, the
determination, on the logical plane, of prices of production by Sraffa important object is a theory of prices and since they see their concept
is carried out without any reference to Marx's theory of labour value. of value as unnecessary for the calculation of prices, they conclude by
The initial data are physical quantities of reproducible commodities rejecting the relevance of Marxist method.
which figure in inputs and products, and a law of distribution (we give The neo-Ricardians claim to situate themselves immediately at a
ourselves a variable which states distribution between wages and shallow level of abstraction by dealing with prices of production, that
profits, and a norm for distributing the global profit between various is to say, making competition between capitals intervene
branches of production). The simple definition of prices of production immediately. They forget that Marx began from a sufficiently deep
is enough to determine them: the arrangement of prices is therefore, in level of abstraction (studying capital and its forms of existence in
fact, totally independent of the world of Marx's values, and the general) in order to approximate progressively to the concrete reality
J10 Guillen 111

which for the vulgar conception constitutes its point of departure. 61 3. Some Final Considerations in Relation to Piero SrajJa's Theory
If we jettison the concept of value we also have to abandon the
concept of surplus value, and therefore, the concept of rate of Sraffa defines production in isolation, in terms of technical relations
exploitation, since it is a relation expressed in value terms. However, but he makes no reference to social relations in the process of
Garegnani thinks, as did Bernstein long ago,62 that dropping the production. Apart from pointing out how commodities are actually
concept of value does not mean dropping the notion of exploitation. used to produce commodities in a capitalist society, Sraffa has
For Garegnani, 'the proposition which refers to the existence of the constructed an imaginery world in which things (use values) produce
exploitation oflabour in a capitalist society does not at all depend on things (use values).
the validity of the labour theory of value, but on the validity of the One of the most important differences between Marx and the neo-
whole theoretical proposition founded on the notion of surplus.'63 Ricardian is that the neo-Ricardians use the term surplus in place of
Thus for him a serf is exploited by the feudal lord only because he the category of surplus value used by Marxists. This is more than a
cannot appropriate the whole of what he produces, and this is semantic difference, since their practice of referring to the surplus is a
independent of any concept of value. reflection of the fundamental difference between their approach and
But what Garegnani does not understand is that Marx was fully Marx's. This conception of surplus is clearly presented in Sraffa's
aware that surplus labour is as old as the history of human work. In the first phrase of the second chapter of his work he asserts
civilization,64 even though the product of this labour assumes the that 'the economy produces more than the minimum necessary for
form of surplus value only when the owner of the means of production replacement and there is a surplus to be distributed.'6 7 This comes as
encounters a free labourer as the object of exploitation and exploits a surprise, since the book's first chapter deals with 'an extremely
her or him with the object of producing commodities, that is to say, simple society which produces just enough to maintain itself68 and
when the means of production take on the specific form of capital. nowhere does Sraffa tell us how the surplus repeatedly emerges.
Because of this it is clear that the particular, capitalist form of Since Sraffa does not see social relations in the production process,
exploitation can only be understood through the Marxist categories there is nothing in his discussion of the surplus comparable to the
of value and surplus value. Garegnani limits himself to the general Marxist concept of capital as a coercive relation, thanks to which the
and, therefore, totally diffuse idea of explotation as such, without working class is obliged to work more than is prescribed by the
dealing with the analysis of the specifically capitalist mode of narrow limits of its vital needs.
exploitation. He forgets the specific economic form in which surplus When Sraffa elaborates his understanding of the surplus, the
labour is extracted from the direct producer, and he forgets that under difference between his approach and Marx's becomes clearer.
capitalism social relations between persons appear as detached from Consider, for example, the definition of surplus which Sraffa offers us
social relations between things, between the products of labour. ~sing national income terminology: 'The national income of a system
To the extent that I base myself on orthodox Marxism in Lukacs's m a self-reproducing state consists of the set of commodities which are
sense. I have the 'scientific conviction that in dialectical Marxism the left over when from the national product we have removed item by
correct method of investigation has been discovered, that this method item the articles which go to replace the means of production used up
can only be extended, amplified or deepened in the spirit of its in all the industries.'69
founders'.65 For this reason I think that all the attempts made by the According to Frank Roosevelt, three senses in which the Sraffian
neo-Ricardians to transcend or correct Marxism have only led to concept of surplus differs from the Marxist concept of surplus value
superficial deformations, to triviality and to coarse eclecticism in the ~an be discerned in this definition. 70 In the first place, Sraffa's surplus
style of Garegnani, when he proclaims that the explanation of profit IS a physical, rather than a value, phenomenon. It is the set of
and prices can be obtained from Sraffa but that in order to explain 'commodities' (read: use values) that remain after subtracting from
accumulation or crisis one must return to Marx's Capital. 66 From my the total produce of the economy those 'articles' which are necessary
point of view the Marxist and neo-Ricardian frameworks can only be in order to replace those which have been used in production. In the
reconciled with the greatest difficulty. second place, both its existence and its precise magnitude are
 Guillen J13
J12
technologically determined. In Sraffa's syste~, the r~placem~nt Since the neo- Ricardians consider the surplus as a relation between
requirements of an economy are fixed by technical. relatl?ns w~Ich things, they are incapable of understanding that its existence reflects a
exist in each branch. These tell us how much of each mput IS reqUIred real struggle between social classes at the level of production. They
to produce given amounts of each product. Once we know the refer to the class struggle only in relation to the distribution of the
technological characteristics of a society we can say whether a surplus surplus once it has been produced. The neo-Ricardian and Marxist
exists or not and how big it is. schools have a very different understanding of the nature of
Third, Sraffa's surplus, in contrast with Marx's concept of [Link]~s exploitation. For Marx, exploitation is the extraction of surplus
value, includes that part of the product of the economy WhICh IS labour in the production process. For the neo-Ricardians, it only has
consumed by workers. As we saw in the definition given above, only to do with the mode in which the social product is distributed. The
those products needed to replace the means of production are tendency of the neo-Ricardians to focus solely on the distribution of
subtracted from the total product. The remaining products of the the product can be seen as another manifestation of commodity
economy are included in the surplus and in workers' c?[Link], fetishism. Instead of concerning themselves with the elimination of
just as the part of the total product collected by the capItahsts forms waged labour, they confine their attention to things like increasing the
part of the surplus. 71 bargaining power of the workers. This leads to an emphasis on
From the Marxist point of view, Sraffa's treatment of the surplus changing the distribution of income in favour of the workers more
mystifies the actual relations of capitalist production. In effect, its than changing the mode of production as such.
presentation of the surplus as a physical phenomen~n ~bscures the As Marx pointed out: 'Trade Unions work well as centres of
significance of the fact that all the products of a capItahs.t econo~y resistance against the encroachments of capital. They fail partially
appear as values. After reading Sraffa, one can have the ImpressIOn from an injudicious use of their power. They fail generally from
that there is really no difference between the surplus produced ?y a limiting themselves to a guerrilla war against the effects of the existing
capitalist society and that produced by any other t~pe of socIet~. system, instead of simultaneously trying to change it, instead of using
Furthermore, one of the most serious defects of Sraffa s treatment IS their organized forces as a lever for the final emancipation of the
that by including the workers' consumption in with the surplus, he working class, that is to say, the ultimate abolition of the wages
obscures the Marxist distinction between necessary and suplus system'. 72
labour. Marx did not include the workers' consumption in with Thanks to neo-Ricardianism, therefore neo-classical economics
surplus value because he wanted to bring out ~n the one hand the has been su bjected to a series of withering critiques, while traditional
relation between surplus value and the value receIved by workers, and reformism of the Fabian variety has acquired a more 'scientific'
on the other, the two parts of the labour time of the workers. Marx foundation. But in resuming the Ricardian tradition as if it stood in
treats the value received by the workers as the product of necessary diametrical opposition to Marxism, rather than being the highest
labour and connects surplus value to surplus labour. stage of classical economics antecedent to Marx's 'critique of political
Sraffa never distinguishes between necessary labour and surp!us economy', latter-day neo-Ricardians have suppressed the actual,
labour. For him, there is no difference between the labour WhICh historical dialectic of classical economics' transcendence. In effect,
produces a surplus and that which only replaces the used up means of they have turned back the theoretical and political clock. Imagining
production. His failure to distinguish between surplus labour ~nd themselves to be the pioneers of a 'post-Marxist' era in political
necessary labour and his treatment of the surplus as .a physIcal economy they have only succeeded in returning to a pre-Marxist past.
phenomenon, leads him to say that the produced surplus IS a surpl~s
of things more than of labour. Put another way, the surplus. m
Sraffa's system is not a relation between people but a relatIOn
between two sets of products, one comprising the total product of the
economy and the other comprising used up means of production. As
Frank Roosevelt points out; the Sraffian concept of the surplus can
be considered as an example of commodity fetishism.