Roncaglia:

THE REDUCTION OF COMPLEX LABOUR TO SIMPLE LABOUR*

Alessandro Roncaglia

One of the most common simplifying assumptions in economic theory is the exist-
ence of a homogeneous, qualitatively undifferentiated, labour force. In reality,
however, diverse types of qualitatively different labour exist. In order to
show different theories based on the assumption of a homogeneous labour force as
simplified representations of reality, and not as imaginary constructions, it is
necessary to show that there exists a method of aggregation of diverse types of
labour that will not influence the results of the analyses. For different types
of analysis there will obviously be different methods of aggregation; the validity
of any method of aggregation should be judged exclusively with reference to the
objectives of the analysis in question. This fact is, however, too often for-
gotten and the methods of aggregation adopted by the classical economists and
Marx have been criticised because they are not generally valid for the resolution
of all the problems presented by the existence of heterogeneous labour. In
particular the problem of aggregation (or in the terms of the Classics, the re-
duction of complex to simple labour) has been confused with the analysis of the
structure of wage differentials.
The classical economists maintained, in general, that the qualitative differences
in the labour force could be reduced to quantitive differences. They thus looked
for a quantitative element that would allow the determination of coefficients for
the reduction of complex to simple labour. There were two methods commonly adop-
ted by the classical economists to make the diverse quality of labour homogeneous
and thus additive: (a) wage differentials, taken as given, and (b) the reduction
of qualitative differences to differences in the costs of training.
The two procedures have diverse significance and the classical economists used
them alternatively to treat diverse problems. The first method can be used when '
the structure of wages is determined independently of the variables under con-
sideration for analysis and reasonably stable for the period under consideration.
This, however, is not always the case. For example, as we shall see more clearly
below, in the context of a theory of value such as the Marxian the assumption
given of wage differentials would involve circular reasoning. In addition, the
empirical •evidence only partially supports the assumption that the structure of
wages is unchanged over time. The second method, based on the cost of training,
is, on the contrary, more appropriate for Marxian analysis, but would require man-
ifestly unrealistic assumptions to be used in the analysis of actual market
phenomena.
The two methods cited, however, do not provide an explanation of such wage
differentials as one finds in reality; the first takes as given what should be
the object of investigation; the second gives an explanation that we shall see to
be at least partial and subject to highly restrictive assumptions. The problem
of wage differentials, however, is going to be held to one side in line with the
principle that what concerns us is only the validity of the methods adopted for
the aggregation of heterogeneous labour by the Classical economists and Marx.

*I would like to thank M. Ridolfi, R. Rowthorn, L. Spaventa, P. Sylos Labini and

M. Tonveronachi for their useful comments.
This paper originally appeared in Italian in Note Economiche, 1973, no.3 under
the title 'La Riduzione di Lavoro Complesso a Lavoro Semplice'. Translation by
J.A. Kregel.
Roncaglia 2
II
Adam Smith is frequently cited as one of the first of the proponents of the
concept of human capital. In his examination of the circumstances which influence
the difference in wages between different occupations he stresses the degree of
agreeableness, peril, responsibility and prospect of success associated with
the work, as well as the cost necessary to obtain the necessary qualification:
A man educated at the expense of much labour and time to any of those
employments which require extraordinary dexterity and skill, may be
compared to one of those expensive machines. The work.which he learns
to perform, it must be expected, over and above the usual wages of
common labour, will replace to him the whole expense of his education,
with at least the ordinary profits of an equally valuable capital.
(Smith, Wealth of Nations, McCulloch Edition, p,46)
Smith, however, is very cautious in relating the differences in wages to the dif-
ferences in degree of qualification. In addition to the fact that this is only
one of a number of factors associated with wage differentials such an association
would require the operation of free competition, that the market should be in
equilibrium and that the division of labour should have achieved an advanced
stage:
In order, however, that this equality may take place in the whole of
their advantages or disadvantages, three things are requisite, even where
there is the most perfect freedom. First, the employment must be well
known and long established in the neighborhood; secondly, they must be in
their ordinary or what may be called their natural state; and, thirdly,
they must be the sole or principal employments of those who occupy them.
(ibid., p52)
Smith remarks that restrictions to competition imposed by law as well as restric-
tions due to tradition are particularly numerous in the labour market.
In order to make labour homogenous Smith appears to consider the reference to
wage differences as given by the market as most appropriate to his purposes:
But it is not easy to find any aggregate measure either of hardship or
ingenuity. In exchanging, indeed, the different productions of different
sorts of labour for one another, some allowance is commonly made for both.
It is adjusted, however, not by any accurate measure, but by the higgling
and bargaining of the market, according to that sort of rough equality
which, though not exact, is sufficient for carrying on the business of
common life.
(ibid., p52)
To the extent that wage differentials are Influenced by law and tradition they
will be relatively stable and will be subject to slow modification in the long
run. Thus while changes in the general economic situation of a country may ,
influence the distributive variables (wage and rate of profit) it will influence
them in a uniform manner in the various sectors:
Such revolutions in the public welfare, though they affect the gene/al
rates both of wages and profit, must in the end affect them equally in
all different employments. The proportion between them, therefore, must
remain the same, and cannot well be altered, at least for any consider-
able time, by any such revolutions.
(ibid., p66)
The principal objective of Smith's analysis was the description of the process
of accumulation and development. To the extent that the structure of wages does
not vary in a given historical period one could use it as a base for the quantif-
ication
 Roncaglia 3

of the different types of labour and one could express this heterogeneity in a
unique variable. The stability of the structure of wages presumed in Smith's
analysis thus requires analysis of the long period.
The case of Ricardo is different. By appealing to the authority of Smith he
tried to exclude the heterogeneity of labour in the study of the relation of ex-
change of goods:
...whatever inequality there might originally have been in them,
whatever the ingenuity, skill or time necessary for the acquirement of
one species of manual dexterity more than another, it continues nearly
the same from one generation to another; or at least that the variation
is very inconsiderable from year to year, and therefore, can have little
effect, for short periods, on the relative value of commodities. (1)
(Ricardo, Principles, Sraffa ed. p22 )
Now, if wage differentials are determined to any significant degree by custom
and tradition it is plausible that they be taken as reasonably stable in a long
period situation.( 2 ) But this stability is not guaranteed absolutely when a
shorter time horizon is considered and forces unique to any single sector could
come into play. It is in this sense that Mill pointed out that different
types of labour could be considered as distinct one from the other; the ratios
of exchange of different goods could thus vary, even with an unchanged technol-
ogy, as a consequence of variations in wage differentials.
For completeness we should also point out that the method adopted by Smith and
Ricardo was later taken UP by Keynes, for whom the assumption of fixed wage
differentials was nothing more than a logical extension of the assumption of
fixed relative prices upon which the Keynesian macroeconomic model is based:
For, in so far as different grades and kinds of labour and salaried
assistance enjoy a more or less fixed relative remuneration, the
quantity of employment can be sufficiently defined for our purpose
by taking an hour's employment of special labour in proportion to its
remuneration.
(General Theory, p.41)
Keynes' analysis, of course, concerns the short period, but for Keynes the
stability of wage differentials is, as he makes clear, a simplifying assump-
tion that could be abandoned without difficulty.

III
For Marx the problem is taken differently. With his theory of value he proposed
an explanation of the existence and the nature of the relevant economic magni-
tudes (profit, wage, price, etc.) that characterise capitalistic society. From
this point of view it would have been contradictory to make reference to market
phenomena in order to homogenise heterogenous labour; the phenomena observed in
the market could not be assumed as given, and thus presupposed, in an analysis
that intended to explain their existence.
On the other hand the cost of training presented itself as a natural development
of the concept of labour-power that was introduced into economic analysis by
Marx. Once labour-power is shown to be a commodity whose value depends on its
'cost of production', there is no difficulty in extending this line of thought
to link value and 'cost of production' of different qualities of labour-power.
This is just the relation that Marx suggests in the following passage:
All labour of a higher or more complicated character than average
labour is expenditure of labour-power of a more costly kind, labour-
power whose production has cost more time' and labour, and which
Roncaglia 4

therefore has a higher value, than unskilled or simple labour power.

This power being of higher value, its consumption is labour of a
higher class, labour that creates in equal times proportionally higher
value than unskilled labour does.
(Capital, I, p.197, Lawrence and Wishart edition)
This view is perfectly rigorous and satisfactory; as we shall see below, such
a method allows one to obtain a significant, uniquely determined solution for
the coefficients of reduction for different types of complex labour to simple
labour. Actually, the transformation from these coefficients to the corres-
ponding wage differentials is a problem that is similar to and no more difficult
than the transformation from value to price that arises with other commodities,
and are the logical corollaries to two different levels of analysis (see VI).
Marx did not worry too much about the problem, and limited himself to a foot-
note indicating his lack of interest in it (to this effect Marx cites Mill:
'The great class who have nothing to give for food but ordinary labour are the
great bulk of the people.') besides, at least in part, the problem is a false
one:
The distinction between skilled and unskilled labour rests in part
on pure illusion, or, to say the least, on distinctions that have
long ceased to be real, and that survive only by virtue of a tradit-
ional convention; in part on the helpless conditionof some groups
of the working class, a condition that prevents them from exacting
equally with the rest the value of their labour-power.
(ibid., p.197, n.1)
The problem of the reduction of complex to simple labour had, however, already
been brought up by Marx at the beginning of Capital in a passage that is highly
ambiguous, and upon which his critics, especially BOhm-Bawerk, have based
their charge of circular reasoning.
Skilled labour counts only as simple labour intensified, or rather,
as multiplied simple labour, ,i-given quantity of skilled being con-
sidered equal to a greater quantity of simple labour. Experience
shows that this reduction is constantly being made. A commodity may
be a product of the most skilled labour, but its value, by equating
it to the product of simple unskilled labour, represents a definite
quantity of the latter labour alone. The different proportions in
which different sorts of labour are reduced to unskilled labour as
their standard, are established by a social process that goes on
behind the backs of the producers, and, consequently, appear to be
fixed by custom.
(ibid., p•44 (3) )
In a footnote to this passage Marx points out explicitly that he is not talking
about the determination of the wage or wage differentials.( 4 ) The apparent
error in the passage would appear to have grave consequences: the value of com-
modities is determined by the quantity of simple labour contained in them, and
this quantity would be deduced at the same time from the value of commodities
-- a glaring example of circular reasoning! If one is willing to accept the
interpretation of Biihm-Bawerk there is nothing left except to believe that Marx
is guilty of a mental slip that it would be possible to correct on the basis of
a suggestion that Marx himself makes in relation to the cost of training. How-
ever, it seems plausible to reject the interpretation of BOhm-Bawerk by demon-
strating that in the passage cited Marx indeed was saying something quite differ-
ent from what has been attributed to him.
This problem of interpretation, it should be observed, in fact relates only to
the history of economic thought. Once it has been established that the use of
 Roncaglia 5

costs of training provides an adequate method to make the reduction from complex
to simple labour within Marx's theory it makes little analytical difference whether
or not this method coexists in Marx's work along with an erroneous method that is
vitiated by circular reasoning.. As with all problems of interpretation it is
difficult to give a definitive answer. We will content ourselves here simply with
the fact that there is at least one element that would allow us to interpret the
previous passage in a manner that is consistent with the rest of Marx's analysis.
This element is to be found in a passage of the Critique of Political Economy:
This reduction (from complex to simple labour) appears to be an
abstraction, but it is an abstraction which is made everyday in the
social process of production. ...Labour, thus measured by time, does
not seem, indeed to be the labour of different persons, but on the
contrary the different working individuals seem to be mere organs of
this labour. ...The Laws governing this reduction do not concern us
here. It is, however, clear that the reduction is made, for, as ex-
change-value, the product of highly skilled labour is equivalent, in
definite proportions to the product of simple average labour; thus being
equated to a certain amount of simple labour.
(Critique, Dobb edition, pp.30-1)
Marx is thus saying: we are not speaking of a method of reduction here, but the
problem will be treated later. But in the Critique he never returns to the problem.
Now, as is well known, this work corresponds, in outline and in the categories
discussed, to the first three chapters of Capital, Volume I. One might be thus
led to suppose that the 'laws that regulate this reduction' would be those indicat-
ed at the end of chapter five of Capital, that is, in the passage cited at the
beginning of this section in which Marx refers explicitly to the costs of train-
ing. The passage of Capital discussed by B8hm-Bawerk corresponds, according to
this line of argument, to the passage just cited from the Critique. In this
passage as in the first Marx denies that he is talking about the method of reduc-
tion'. Marx is saying, above all, that the abstraction of the qualitative differ-
ence between the different types of labour is something more than just a mental
exercise( 5 ) and is in fact a process that in particular historical periods is
verified by reality. In other words, in a capitalist system the existence of
the market, implying the existence of a quantitative relation between commodities,
implies as well a quantitative relation (according to a law not yet specified)
between the different types of labour. Thus, in a system dedicated to the prod-
uction of commodities for the market, labour is going to be considered as social
labour, which should supercede the various characteristics of any single worker.
The reasoning is not circular for the demonstration of the existence of a
quantitative relation is something different from the specification of its char-
acteristics: the 'law of reduction', in our explanation, is determined without
any circular reasoning through the concept of the costs of training. As has been
noted, this is only one of the possible interpretations of Marx's controversial
passage (and to my thinking the most reasonable), (6 ) but whether this or B8hm-
Bawerk's interpretation is correct, it still remains true that the method of train-
ing costs provides an appropriate solution to the problem of the aggregation of
different types of labour-power within Marx's theoretical scheme.

IV
In order to calculate the reduction coefficients for the transformation of com-
plex to simple labour it is only necessary to refer to the direct and indirect
labour employed in the process of training (including the labour to be trained,
but excluding that contained in its subsistence).
Using the notation used in Sraffa's Production of Commodities by Means of Com-
modities( 7 ) we can represent the complex of processes of production employed in an
 •▪

Roncaglia 6

economic system at a given moment of time, considering at the same time the
training processes of the labour force (let us suppose for simplicity the
absence of joint processes, and thus the absence of fixed capital, and labour-
power used for more than one period).

A + + K + L + + L + -4- A
a a a a Ka aa
O 00 0000V 00•0000000 (1)

Ak + .. + Kk + aLk + + KLk + aLk K

A + + K + L + + L + L L
a a ac Ka a a a
O 000 01000 00•00000000 (2)
A + +K+L+ +L+L L
K a K KK OK

The expressions grouped under (1) represent the processes of production for
physical goods; Ij indicates the quantitY of good i necessary for the production
of quantity J of good j; 0,j indicates the quantity of complex labour of type
necessary to produce J of good j. The expressions grouped under (2) represent
the training processes for the labour force: 1E is the quantity of good i necessary
to obtain quantity 0, of complex labour of type 6; yLE is the quantity of com-
plex labour of type y necessary to obtain quantity EL of complex labour of type c.
aLi and GLE are quantities of simple labour necessary to produce quantity J of
good j and quantity e l, of labour type c, respectively (i,j=a,...,k;
Let us suppose that the economic system under consideration be in a self-
replacing state such that for all goods and for all types of labour the quantity
used up in production is less than or equal to the quantity produced. The aim
of our analysis, namely the determination of the labour-values at a given point in
time, does not require any assumption concerning the possible changes in the magni-
tudes of the variables considered over time. It is thus not necessary to specify
that the system is in a state of either simple or enlarged reproduction.
Let us indicate the quantity of simple labour directly and indirectly embodied
in a unit of commodity i (i.e. its value) by Xi; and the coefficient of reduction
of complex labour of type E to simple labour by p E . The value of all commodities
is equal to the sum of the simple labour quantities embodied in their means of
production, that is to the sum of their values plus 'the simple labour directly
employed in their production which is obtained by reducing the various types of
complex labour to simple labour by means of the reduction coefficients. We can
thus transform the system (1)-(2) into the following equation system:

X A + O 0 K + 1 L + + p L + = X A
a a k a a a a KK a ca a
00 0000 00000 0 O• • 00 00 0 00 00 0000 0 (3)
Aa A. 000 Ak Kk P a aLk °°" aLk = Ak K

AA+ + X K +pL+ +pL+L= 1

a a k a a CC a K K el U pa 6
O 000 0•00000 000.000 000 0 (4)
X A + +X K+p L+ +pL+L =pL
a K k K a K KKK OK K K
 Roncagli

The system composed of groups of equations (3)-(4), that is k + K equations,

will be sufficient to determine the k + K unknowns A
It is necessary to make some observations on this solution. First, the two
groups of equations (3)-(4) are simultaneously determined; in other words the
reduction of complex to simttle labour is carried out at the same time as the de-
termination of the value of the different commodities. Only if l c = 0 for i = a,
...,k and e = a,...,K; that is if the training process requires only labour
time (simple and complex), system (4) could be solved independently of system (3)
and the reduction be made before the determination of values.
Second, in the preceding it is implicitly assumed that the length of the work-
ing day is given, at least for skilled labour. In fact the labour inputs in a
given process of production are necessarily tied to the specified working time
(e.g. number of hours) necessary for each unit produced, while the process of
training concerns not hours of work but labour-power (i.e. a man-year of given
skill, if a year is the duration of labour power); to pass from a quantity
expressed in labour-time (e.g. hours) to a quantity expressed in terms of labour-
power (e.g. a skilled labour-year) it is necessary to specify the time in hours
of the working year, or, in more usual terms, the length of the working day.
One can thus choose to express the product of the training process either in terms
of labour-time, as we have done above, or alternatively, one can express all the
various inputs of labour in terms of labour-power, that is, in man-years.
Third, one could easily relax the assumption concerning joint production which
requires, among other things, that labour power can be used for only one year.
In fact we could choose either to make a simple extension of the equation system
(similar to that made by Sraffa in his treatment of joint production) or to use
Sraffa's concept of sub-systems, or, given that the time element has no meaning
in the absence of profit rates, to create a period of production equal to the
lowest common denominator of the various fixed capital goods and the various
types of labour-power lasting for more than one period, which permits the rein-
troduction of simple production when there is no joint production in the technical
sense of the term. For the fixed capital goods already in use for one or more
periods, and for technically joint products, it is possible to come up against the
problem of values (quantities of embodied labour) which are negative; but the
third of the possibilities given above allows one to confirm that the problem
could not arise, in a self-replacing system, in the case of production processes
which produce a single product, either a fixed capital good (or complex labour
power that is used for more than one period) or if the production process requires
the use of fixed capital and/or complex labour power for more than one period.
We should recall here that if the process of training lasts more than one period,
but the skilled labour-power is then used for only one period it is not necessary
to consider joint production; all processes of training can be sub-divided into
a number of parallel processes, each one of them a stage in the process of
training. The labour power that emerges semi-trained from the first stage will
then appear as means of production for the second stage and so forth until the
completion of all stages of training. The case of labour-power that requires more
than one year for its training, and which is then used for more than one period
is just a combination of the two cases already considered. These cases, as we
have seen, do not raise substantial complications in relation to the assumption
that the labour-power is trained in one year and then expended for only one year,
which is the unrealistic assumption adopted initially. Thus we can accept the
results of analysis based on that assumption.
We remark finally that the procedure of reduction outlined above depends only on
the knowledge of the technical coefficients; price, rates of wages and profits
are not needed in any way. The method suggested by Marx does not then imply any
form of circularity.
Roncaglia 8

Having examined the method of reduction of complex to simple labour proposed by

Marx, and having found it to be consistent with the intention of Marxian theory
it will now be easier to separate out the weak points of the critique mounted
against Marx in relation to the problem.
The best known is that of Bbhm-Bawerk which has been uncritically repeated by
numerous authors.
In the first place, Balm attacks the reduction of qualitative differences in
the different types of labour to quantitative differences as 'artificial'. In
the second place, basing himself on the ambiguous passage of Capital (discussed
in III above) he accuses Marx of circular reasoning because Marx appears to
deduce the coefficients of reduction of complex to simple labour from the
exchange values of commodities. Finally, Bohm-Bawerk attributes to Grabski, (9 )
an obscure follower of Marx's, the Marxian method of reduction described in
III and developed in IV, and on account of this obscurity the illustrious Austrian
economist felt himself justified in stopping short of considering the method,
simply remarking that wage differentials that occurred in reality did not corres-
pond to observed differences in costs of training:
I do not think it will need many words to show clearly the complete
inadequacy also of this explanation. I have nothing to say against the
view that to labour in actual operation should be added the quota due to
the acquirement of labour power. But it is clear that the difference in
value of skilled labour as opposed to unskilled labour could only then be
explained by reference to this additional quota if the amount of the latter
correspond to the amount of that difference.
(Karl Marx and the Close of His System, Sweezy
ed. p84)
and that, maintains Bbhm-Bawerk, is not true (ibid. p85).
The reply to the first criticism is to be found in the same passage of Marx
that the Austrian economist used as a basis for his charge of circularity if
the interpretation that we have given to this passage in IV above is correct:
the reduction from qualitative differences to quantitative differences is
not an arbitrary artifice, but corresponds to the real fact that in a capital-
istic system the market imposes the existence of quantitative relations among
the different commodities as well as among the different types of labour that
produce them. The second criticism (the accusation of circularity) falls,
as we have seen, when one uses the costs of training of the various types of
labour-power as the basis of the reduction coefficients; ( 10) whether or not
the accusation of circularity is valid in relation to a particular passage of
Capital thus has no great importance to the problem at issue. As for the third
criticism, the non-correspondence of wage-differentials and reduction coefficients
determined on the basis of differences in costs of training, it should be
sufficient to recall that values and coefficients of reduction on the one hand,
and prices and wage differentials on the other, belong respectively to two
different levels of analysis. The non-correspondence of differentials and coef-
ficients thus poses a problem that is no different from the non-correspondence
of prices of production and labour values (this criticism has recently reappear-
ed and will be examined in more detail in VI below).
(11)
Later criticism has been simple repetition of Bbhm-Bawerk's position.
Thus Bortkiewicz held that 'Marx had treated this problem (of reduction) in an
unsatisfactory manner' as well as viewing the problem as of little relevance.
He seemed to find reference to the cost of training acceptable, but was reserved
 Roncaglio 9

in his' approval of this 'attempt' which he attributed to Hilferding, Grabski, and

Dietzel, because of his desire to be 'faithful' to Marx. In fact, several Marx-
ists of the period seem to have forgotten Marx's own suggestion and thus felt
obliged to propose alternative methods of reduction. Liebknecht, for example pro-
poses if with great reservation - reference to the maintenance costs of differ-
ent groups of workers, determined physiologically. (12) A few years later Boudin
proposes the differences in productivity between complex and simple labour, which
in practice, is the same as reducing all types of labour to 'efficiency units':
"A skilled worker produces in a given period of time more than an unskilled
worker...The value of the commodity...will be the amount of average normal labour
necessary for its production." (The Theoretical System of Karl Marx, Chicago, 1907,
New York, 1967, p116) and the greater value of skilled labour-power will be seen
to be deduced from its greater productivity. But because that higher productivity
will be measured in physical terms, as is necessary if one is to avoid circular
reasoning, it will be necessary that simple and complex labour are perfectly sub-
stitutable, an hypothesis wildly unrealistic that Marx would never have admitted.
Among the more recent of the followers of Bbhm-Bawerkian criticism is Samuelson,
who in a recent article attributes to Marx (and to Ricardo) the method of efficiency
units 'put forward by Boudin and who reaffirms the impossibility of comparing differ-
ent types of qualitatively different labour in an argument with a 'fictitious'
Marx:
However, the efficiency-unit device will work empirically at best only as
an approximation. Natural differences show a Gaussian-like spread. 'A man's
a man for all that' is a proper legal dictum. But a woman is not a man, and
men are not at any age homozygous twins. Thus, let women be three times as
efficient in beaver production and two times as efficient in deer production.
How do we get our new quantum of 'socially necessary labour' ? The answers
are, on reflection, clear. Without the Walrasian conditions of equilibrium
...little progress is possible. (Journal of Economic Literature, 1971 pp.404-5)

The reference to Walras indicates that for Samuelson the differences between dif-
ferent types of labour (men and women) are a fact of nature. We are thus told
to abandon the point of view of the Classics, that is the consideration of commod-
ities as exclusively produced and reproducible goods and return instead to the
Walrasian methodology which is most appropriate to the problem of the optimal
allocation of given scarce resources. But even if natural talents are randomly
distributed amongst the population this does not constitute an objection to the
Marxian theory of value: it is sufficient to refer to the mean as Marx himself
proposed. Difficulty with this approach would only arise if people with partic-
ular abilities were concentrated in particular occupations. But, as Rowthorn
has observed:
The problem should not, however, be exaggerated. Mechanisation, automation
and other changes in methods of production have already )reduced dramatic-
ally the importance of such special capabilities as great physical strength
or manual dexterity, and further changes in this direction will continue to
occur in the future. Specific intellectual and artistic natural abilities
will doubtless remain important in certain restricted areas of economic
activity. But their overall significance is not and probably never was very
great. (The Reduction of Skilled to Unskilled Labour' CSEB 8)
If the problem in practice reduces to artists we can turn in substance to those
categories of goods (paintings, statues, etc.) 'whose value is determined ex-
clusively by their scarcity'. But, as Ricardo observed, 'such commodities com-
prise a very small part of the mass of commodities exchanged daily on the market',
and could easily be left out of an analysis of the fundamental mechanisms of
the functioning of the social system. We should, however, appreciate Samuelson's
Roncaglia 10

ability to pick his examples: the sculptor used as an example by,Bdhm-bawerk and
which can be considered a marginal case, is replaced by a case in which natural
ability would have a considerable practical relevance.( 13 ) In reality, for the
greater number of occupations, if not for all, it makes no difference if the
labour is male or female; particular pieces of legislation which may favour female
labour (time off for child birth, early retirement, etc.) create a uniform diff-
erence between male and female labour in all sectors of the economy and which can
be easily taken into account with the proposed method on the basis of demographic
statistics (while the use of Walras' method would require differential advantages
for the various employments). As for the example of Samuelson as well as that
of Bbhm-Bawerk, the reality of mass production undoubtedly confirms Marx's
approach when we remember that all wage labourers are in substantially the same
position as they face capital.

VI

The method of reducing complex to simple labour proposed by Marx is thus perfect-
ly adequate, emerging unscathed from the various criticisms that have been levelled
against it. Once this method is accepted, however, another problem immediately
emerges: the specification of the relation that exists between the coefficients of
reduction and wage differentials.
This question, already raised by Bdhm-Bawerk and Joseph, has been recently put
forward by Marishima. After having reduced complex to simple labour with a method
analogous to that prposed here (see IV) Morishima observes that there is no reas-
on why wage differentials should be proportional to the coefficients of reduction:
'Then we may have several groups of workers exploited at different rates, in con-
tradiction to Marx's two-class view of the capitalist economy.' Morishima con-
siders this an insurmountable problem for Marx's theory of value.
Let us look at the problem in more detail. If we leave aside the differences
between different types of labour-power which do not result from differences in
costs of training, and if we suppose that the type of training has been carried
out by the labourer himself, and that perfect competition rules on the market,
it would be possible to determine the wage differentials by appropriately enlar-
ging the system, of equations used by Sraffa to determine relative prices:( 14 )

(A p +...+ Kp)(1+r) (71 w +...+K La i(w + aa

L )w = Ap a
a a ak a ac (5)
••• ••• ••• ••• •••

(Akp a +...+ Kkpk )(1+r) + (c .L k +...+ K Lkw ic + a Lk )w = Kpk

(A p. +...+ K p k )(1+r) + a( aLa w +...+KaK

L w + L')w = aLv w a -1) ( (6)
a a a
••• ••• ••• ••

(A, p +...+ K K p )(1+r) + ( KL aw +..;+KLK K

w + )w = Lw(w -1)
Ka - 0K

p ,...,p k are the prices; w ,...,w K the wage differentials (the coefficients by
which it is necessary to multiply the wage associated with simple labour in order
to find the wage appropriate to complex labour); r, the rate of profit; w, the
wage rate for simple labour. The equations (5) show the fact that the product
prices are equal to the prices of the means of production plus profit, plus wages
(paid post factum) for the various types of labour employed. The equations (6) show
in the same manner the costs of qualification: the wage for complex labour (=Wwe)
 Roncaglia 11

should be such as to take into account the costs of training (=w(w e -1)) as well as
the wage normally associated with simple labour (w).( 15 ) We then have k+K inde-
pendent equations; and given one of the distributive variables, w or r we can
solve (5) - (6) simultaneously, determining the k-1 relative prices, the K wage
differentials, and the remaining distributive variable. As above, the system (6)
could be solved independently of (5) only if le° for i=a,...,k and for e=a,...,K.
In general we should find that wSp e (wage differentials different from the coeff-
icients of reduction; but in general we will also have pi/pj Xi/Xj (relative
prices different from relative labour values), the two problems are no different
in principle. The difference between the relative prices of two commodities and
the ratio of the quantity of labour contained in them depends on the manner in
which the surplus-value is distributed between the different capitalists: the
"postulate" of a uniform rate of surplus value is contravened (save in the special
case of uniform organic composition of capital) by the tendency to equality of
the profit rates in different industries, for the capitalists are interested in
the rate of profit on capital advanced, not on the rate of surplus value of labour
directly employed in production. In the same manner the inequality of the coeff-
icients of reduction and wage differentials corresponding to them depends on the
fact that skilled labour as well, following the assumptions given above, is inter-
ested in returns on the costs they have incurred in their own instruction in the
sense of a rate of profit and not a rate of surplus value. With such motives
the rate of exploitation will not be uniform in the different sectors if there is
not a uniform organic composition in the various 'processes of training' (and in
the sectors that produce the means of production for the process of training);
this even when the wage differentials are determined by the mechanism outlined
above which is particularly unrealistic. In fact, as we have seen above in re-
lation to Smith (paragraph 2) certain special conditions are necessary for the
operation of such assumptions (free competition, equilibrium, etc.) which are
far from being found in reality, and will be predominated by the forces of custom
and tradition such that there will be natural, legal or administrative barriers
between jobs or sectors of the labour force, as well as differences of a
structural nature (public or private schooling, competitiveness of the market,
etc )
On the other hand, it should be kept in mind that Marx did not intend his explan-
ation of the determination of the coefficients of reduction of complex to simple
labour to serve in the determination of wage differentials, but simply to fur-
nish a method of aggregation for the various types of labour-power. In the pre-
ceding pages we have seen how such a method was fully adequate to the completion
of his theory of value.

NOTES

1. Ricardo, like Smith, also presented the other method, cf. Works and Corres-
pondence, vol.II p226.
. 2. In recent times different empirical investigations have at least partially
confirmed Smith's intuition, and emphasise the importance of tradition in
the determination of wage differentials as well as their relative stability
in the long-period, although recently there seems to be a tendency towards
equalisation. Cf, for example, H.A. Turner, 'Trade Unions, Differentials
and the Levelling of Wages', Manchester School, vol.20, 1952, pp227-282 but
in particular pp244-258.
3. Bbhm-Bawerk comments: '...the standard of reduction is determined solely by
the actual exchange relations themselves'. Karl Marx and the Close of his
System, Sweezy edition, p.83. This critique is taken up in V below.
Roncaglia 12

4. "The reader must note that we are not speaking here of the wages or value
that the labourer gets for a given labour-time, but of the value of the
commodity in which that labour-time is materialised. Wages is a category
that, as yet, has no existence at the present stage of our investigation.
p44,n.2.
5. Or a simple mental construction (em n Gedankenbild) as Bernstein would say. On
the particular significance of abstract labour in Marx cf. L. Coletti,
Ideologia e Societa, Bari, 1969, pp103-124.
6. This interpretation is also supported by R. Meek in Studies in the Labour
Theory of Value, London 1956, pp169-70.
7. A description of the economic system as a flow of goods and a flow of people
meeting in the process of production which yields a flow of persons as well
as a flow of goods is developed in a manner similar to that used here by
A. Breglia in Reddito sociale, ed, P. Sylos Labini, Rome 1965. Cf. espec-
ially chapter X, pp61-65.
8. Similar formal solutions to the problem of reduction have been given by N.
Okishio, 'A Mathematical Note on Marxian Theorems', Weltwirtschaftliches
Archiv, vol. 91 1963, p289; A. Brody, Proportions, Prices and Planning,
Amsterdam, London and Budapest 1970, pp86-8; M. Morishima, Marx's Economics
Cambridge U.P. 1973, ppI91-2; R. Rowthorn, 'The reduction of Skilled to
Unskilled Labour', CSEB 8. Morishima shows that the simultaneous determin-
ation of an economically significant (i.e. positive) solution for the values
and the coefficients of reduction satisfies the assumption of the validity
of the system.
9. S. Grabski, in Deutsche Worte, vol.XV, 1895, p155.
10. After Marx this method has been proposed by others as well as Grabski, e.g.
H. Dietzel, Theoretische Sozialdkonommik, Leipzig 1885, vol.I, pp.248-61
and R. Hilferding, Bdhm-Bawerk's Criticism of Marx, included in the Sweezy
edition of Bbhm-Bawerk's Karl Marx, etc., pp121-99.
11. Cf. for example, C. Adler, Die Grundlagen der Karl Marxschen Kritik der beste-
henden Vokwirtschaft, TUbingen 1887, pp81-5; T. Masaryk, Die philosophischen
und sociologischen Grundlagen des Marxismus, Vienna, 1899; V. Pareto, Intro-
duction to Karl Marx, Le Capital, Extraits, ed. P. Lafargue, Paris 1893
(reprinted in V. Pareto, Marxisme et economie pure, ed. G. Busino, Geneva
1966, pp52-3); V. Pareto, Les Systemes socialistes, Paris 1926, vol.II, pp
379-81; H. Joseph, The Labour Theory of Value in Karl Marx, London 1963,
p41,64, 87-96.
12. W. Liebknecht, Werttheorie in England, 1902, pp99-103.
13. In fact the only possible alternative would have been the distinction between
whites and negroes: there is not a great deal of difference between racism
and male chauvinism.
14. Sraffa, op.cit., p13. For a similar system of the simultaneous determination
of prices and wage differentials cf. J.T. Schwartz, Lectures on the Mathe-
matical Method in Analytical Economics, New York 1961, p39; A,Brody, op.cit.
p87; and R. Rowthorn, op.cit.
15. 1,4'w e = w(w e -1) 4 W.