Академический Документы
Профессиональный Документы
Культура Документы
Peace Research
http://jpr.sagepub.com/
Exploring Exploitation
Jon Elster
Journal of Peace Research 1978 15: 3
DOI: 10.1177/002234337801500102
The online version of this article can be found at:
http://jpr.sagepub.com/content/15/1/3
Published by:
http://www.sagepublications.com
On behalf of:
Peace Research Institute Oslo
Additional services and information for Journal of Peace Research can be found at:
Email Alerts: http://jpr.sagepub.com/cgi/alerts
Subscriptions: http://jpr.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://jpr.sagepub.com/content/15/1/3.refs.html
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
Exploring Exploitation
JON ELSTER
Department of History,
University
of Oslo
Both Marxist and neo-classical theories of income distribution agree that the worker is
exploited if he gets less than he produces. They differ over the imputation to the worker of
his product, Marxists using the average and neoclassical theorists the marginal product. In
order to bring out the deeper reasons behind this difference, the following notions and
theories are invoked: the fallacy of composition, permitting the capitalist to treat each
worker as if he were the marginal worker; the alienation of the worker from the capital
goods which are the product of past labour; the capital controversy and the doubts that
have been raised about the notion of aggregate capital; and Robert Nozicks distinction
between end-result principles and historical principles in ethical theory.
Pre-capitalist societies and advanced capitalist societies require a different conceptual framework than the exploitation theory developed by Marx for classical capitalism. In these
societies power relations and strategic interaction are more important than the anonymous
exploitation through the market mechanism. Kelvin Lancasters view of capitalism as a
differential game between workers and capitalists provides an insight into modern capitalism
that is more adequate than the Marxist model.
Different needs
To receive according to ones needs may
be interpreted in three ways: as satiation of
needs,
as
material affluence
or an
ascetic want-struc-
ture ; I state without argument that the foris unrealistic and the latter undesirable.
The second interpretation is based upon
the intuitively appealing idea that inequalities in consumption may be required for
equalization of welfare, so that persons with
low capacity for consumption, in the sense
of getting less enjoyment out of a given
amount of material goods and services,
should receive more of these goods and services. The snag here is that we must presuppose some interpersonal comparability of
welfare levels, the meaningfulness of which
remains in doubt despite the recent work of
Strasnick and others.
The third interpretation is the second turned upon its head: here we assume that larger amounts of goods and services should
be given to the persons most capable of turning them into enjoyment, by the kind of
Matthew effect that has long governed the
allocation of resources for investment.2 Perverse as the idea is from a substantial point
of view, it is perfectly acceptable as an interpretation of the phrase to each according to
his needs.
Now why did Marx think that distribution
according to needs would take place only in
the higher and definitive stage of communism, whereas the society immediately
mer
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
4
from the womb of capitalism
would have to base itself upon distribution
according to effort? Three distinct arguments
are more or less clearly stated by Marx and
his successors. In the first place there is the
problem of work motivation. Under capitalism hard work is forthcoming only by the
promise of a corresponding reward; if there
was no link between individual effort and
individual reward, a community of individuals imbued with the capitalist work ethos
would find itself in a Prisoners Dilemma
where everyone would opt for being a free
rider. A new work ethos can emerge only in
the higher stage of communism, where not
only property arrangements, but man himself has been transformed.
Amartya Sen3 has suggested that in this
stage the work ethos could take the form of
what he calls an Assurance Game: everyone
is willing to do his share of the hard work,
provided that everyone else does the same.
A more Utopian and (to my mind) less
appealing idea would be that hard work
could be a dominant strategy under communism, just as a low level of effort would
be a dominant strategy under capitalism if
there were no link between effort and reward.
In the second place, Marx suggests that
the principle of distribution according to
needs can come into its own only when a
material basis has been created, i. e. when
the springs of affluence are so abundant that
a complete satisfaction of needs is possible.
Unlike the first argument, this idea presupposes that distribution according to needs is
taken in the sense of satiation of needs. If
instead the principle is understood in the
sense of giving more to those less capable of
transforming consumption into enjoyment,
we arrive at the opposite conclusion, for it is
precisely in conditions of great scarcity that
such compensation becomes important.
In the third place, later Marxists have
often hinted at the idea that communist man
will have a lower level of aspiration, because
he will have rid himself of the artificial desires that are continuously generated under
emerging
capitalism
Equal needs
Now as stated above, in what follows I shall
abstract from all considerations of need.
I shall assume, for simplicity, that all persons are equally needy and equally capable
of work. In real societies this is not true, as
we all know. In a society governed by the
principle of distribution according to effort,
invalids, for example, would come out badly
on two counts: they produce (and therefore
get) less, and they are less efficient in transforming what little they get into enjoyment.
By abstracting from such difficulties we can
discuss in its conceptual purity a situation
where a class of workers-nonowners confronts a class of owners-nonworkers. We
shall then see that the principle of distribution according to effort lends itself to two
very different readings: one interpretation
where it excludes the possibility of income
accruing to nonworkers and one interpretation which permits this. Not surprisingly the
first notion is implied by the Marxist theory
of exploitation, whereas the second follows
from the neoclassical theory of exploitation.
The exposition and confrontation of the
Marxist and the neoclassical approach form
the core of the present article.
The relation between capitalists and workers in classical capitalism can be briefly
characterized by two features. ( 1 ) The
economic power of the capitalists over the
workers was reinforced by political domination.4 (2) The exploitation took place
through the impersonal mechanism of the
market. These features interacted so as to
generate a maximum of asymmetry in the
relation between the two classes. By contrast the first feature is lacking in modern
capitalist societies and the second in precapitalist economies. In advanced capitalist
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
5
economies the economically weak now have
the political power, which makes for complex game-theoretic aspects of the class
struggle. In pre-capitalist societies (of which
slave societies will be taken as a typical
example) the exploitation was mediated
through personal relationships, in which the
exploited had some power over the exploiters. I shall argue that the notion of exploitation developed by Marx has a less
universal relevance than is often thought;
while it fitted as hand in glove to the capitalism of Marxs time, it does not capture
the essential features of class conflict of earlier and later periods.
The Marxist theory
The Marxist notion of exploitation is a very
simple one: the worker is exploited if he
receives less than his average product. (II
here assume simple reproduction, i. e. no net
investment. This restriction is lifted below.)
Obviously if all workers get the equivalent
of their average product, nothing is left for
nonworkers.
When we take a closer look at the Marxist
idea as it can be worked out in concrete
analyses, two distinctions are useful. Firstly,
there is an interesting and rather strange
analytical difference between Marxs analysis
of exploitation in agriculture and the corresponding theory of industrial production.
Secondly, there are as many variants of the
theory as there are institutional property arrangements defining the rights and obligations of owners-nonworkers and workersnonowners. A few of the variants are explored below.
In his analysis of capitalist agriculture,
Marx assumed diminishing marginal returns
to labour. From this assumption, together
with the assumptions of perfect competition
and profit-maximizing tenants, we immediately deduce that the number of workers on
a given piece of land is so chosen as to
equalize wages and marginal product. The
important point of this ultrasimple piece of
analysis is that the idea of profit-maximization is here set to some work: for the conclu-
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
6
of not assigning a crucial role to the
subjective choice of the actors. Producers
have no choice between input combinations,
and workers no choice between commodity
bundles. If such choices are admitted, the
labour theory of value can no longer be
sustained. If producers have a choice between such input-combinations as mentioned
at the end of the preceding paragraph, they
can only be guided by the relative factor
prices, which means that prices, technology
and labour values are determined simultaneously, contrary to the (Hegelian) interpretation of the labour theory of value which requires that values should be determined independently of prices but not vice versa.
If consumers have a choice as to how they
will spend their wages, they may choose to
spend more money on goods with a high
labour value relative to prices, which will
have an impact upon the value of labour
power, the rate of surplus value, the rate of
profits and finally upon relative prices, contrary to the (Ricardian) interpretation of the
labour theory of value which requires that
prices should be independent of the composition of final demand.7
On the more or less conscious level, such
considerations must have been important in
Marxs choice of assumption concerning the
technical nature of production. In addition
to these intended (or at any rate acceptable)
consequences of the assumption, there are
also a number of rather disturbing conclusions that can be drawn. In the first place,
the assumption of profit-maximizing is virtually idling, at least at the micro-economic
level. (It does have a function at the macroeconomic level, viz. in determining the capital flows that bring about the equalization of
the rate of profits.) The use of factor combinations in industrial production will be the
same in any economic system, whatever the
motivations of the agents. This is a rather
paradoxical conclusion: on the one hand
Marx repeatedly stressed that capitalists are
motivated solely by hunger for profits, and
on the other hand it turns out that they
would have chosen identically whatever their
sense
motivation. There is
no formal inconsistency
but
on
a
here,
priori grounds one would exMarxists
to
pect
prefer (ceteris paribus) a
that
theory
brings the assumption of profit
maximization to bear upon the process of
production itself.
In the second
may appear to strengthen the Marxist position, but to my mind they do not. An important task for Marxist theory must be to
analyse why the neoclassical definition is
superficially and intuitively compelling,
even if false at a deeper level. This, obviously, cannot be done through a theory
that makes the neoclassical theory come out
as absurd even at the surface level.
The crucial idea is the following. To each
worker, the capitalist entrepreneur may say
with some appearance of truth: if you leave
the production process, production will be
curtailed by an amount X; it is impossible
for me, therefore, to pay you more than the
equivalent of X. This argument can be used
for all workers, and can therefore justify the
capitalist retaining for himself what is left
after all workers have been paid the amount
X. On the assumption of fixed coefficients,
exactly nothing would be left; on the alternative assumption of a diminishing and positive marginal product of labour there will
be something left for profits. This means
that i f we accept that each worker can be
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
7
treated as if he were the marginal worker,
then the existence of profits receives a
natural and compelling justification.
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
8
the number of tenants up to the point of zero
marginal productivity, but at the same time
the percentage must be adjusted so as to
permit the survival of the tenants. It then
follows that the rate of exploitation is smaller than under capitalism; more importantly
it follows that a larger number of labourers
will be employed at subsistence than under
capitalism. The transition from sharecropping to capitalism may therefore be a bad
bargain from the point of view of employment.
If to
we
as a
criterion
just income.
Actually the
according
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
9
warded. If there are decreasing returns to
scale, we have exploitation in the secondary
sense without exploitation in the primary
sense. If we assume constant returns to scale,
perfect competition and profit-maximizing
entrepreneurs, there can be no exploitation
in either sense.
This is a central idea of neoclassical
theory: exploitation can only occur in imperfect capitalism, most importantly because
of imperfect competition. On the Marxist
notion, on the other hand, exploitation takes
place in all capitalist economies, however
perfect. (This is even close to being a
definition of capitalism in Marxist theory,
but I shall argue below that even in the
absence of exploitation there may be features in the economy that would justify
the term capitalism.) It should be clear
enough that for purely analytical purposes
both the neoclassical and the Marxist notions
of exploitation may have their uses. When,
for example, workers revolt against the
system, we may say that they revolt against
exploitation in the Marxist sense; when they
revolt within the system against abuse of the
system, they revolt against exploitation in
the neoclassical sense. The controversial
issue is found along the normative dimension : should workers revolt against or within
the system?
Exploitation of workers may result from
monopoly (in the product market), monopsony (in the labour market) or cartellization. The last case may be briefly explored
in order to exhibit the multiplicity of notions
of exploitation that may apply in a given
situation. It is well known that the main
problem in cartel formation is the temptation, for each individual cartel member, to
be a free rider and exploit the high prices
created by the restriction of production. If
one cartel member maximizes individually,
whereas all the other members curtail production in order to maximize collective profits, then it is surely not an abuse of language
to say that the free rider is exploiting the
others, in a sense that differs both from the
Marxist one and either of the neoclassical
The free
Confrontation
In the preceding pages I have sketched the
bare bones of the Marxist and neoclassical theories of exploitation. Here I turn to
the task of staging a critical confrontation
between the two approaches, a task that
seems to be of some importance because
bourgeois economists so
often misunderstand
the Marxist notion of exploitation.ll As a
starting point for the discussion we may observe that if the apparent symmetry of capital and labour in the function f(K,L) is a
real one, then the neoclassical definition of
exploitation is the more appropriate one. If
capital and labour are on a par as factors of
production, then they also have an equal
right to rewards, so that we should reject
any theory implying that a just reward to one
factor excludes any reward at all to the
other. Or to put the matter positively, the
state of affairs where no factor is exploited
should at least be logically possible. These
objections to the Marxist theory will of
course be countered by denying that capital
and labour are symmetrical factors. Specifically Marx himself argued that capital really
is nothing but congealed labour, so that on
closer analysis the production function has
labour as its only argument. This idea can
be made more precise by stating that the
production process should be represented by
a functional Pet) = F(L(t) ), where P(t)
and L (t) are the time profiles of output and
of labour input respectively, and not by the
function P = f ( K,L) . If the production process is represented by this functional, all
justification for a reward to capital evaporates ; capital just does not have any role to
play at all.
Now I agree with this conclusion, but
submit that some care must be taken in
arguing for it. Four distinct issues are involved here: the so-called capital contro-
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
10
the problem of hysteresis in the
social sciences; the problem of historical
principles vs. end-result principles in ethical theory; and finally the ideological element in economic theory. Starting with the
first issue, it is by now generally accepted 12
that under many plausible circumstances the
process of production cannot analytically be
represented by a function f (K,L) . This is so
because (under these same circumstances) it
may be impossible to define any notion of
aggregate capital that behaves as a factor of
production should do, e. g. exhibit diminishing marginal productivity. From this fact
at least one author explictly,13 and many
implicitly, have drawn the following conclusions. Firstly, since the form f(K,L) is
inadequate, only the form F(L(t) ) can be
adequate. Secondly, since only this latter
form is adequate, capital cannot claim any
reward. In my opinion both of these deductions are fallacious. The second is fallacious on the elementary grounds that
normative conclusions cannot be deduced
from analytical premises; the first because
presently existing effects must have presently
versy ;
operating causes.
The point about presently operating causes
probably is the least familiar one, so that
some space is required to explain it. 14 If an
explanation of the value of some variable at
time t appeals to values of the same or of
other variables at time t-2 or earlier, then
we have a case of hysteresis. If, on the other
hand, no appeal is made to values earlier
than t-1, the explanation uses state variables
only. (For simplicity discrete time is asstate variable
in
form
of simulwould
be
the
explanations
taneous differential equations.) For general
metaphysical reasons hysteresis cannot be a
feature of the real world, only of our theories about it. If past causes could have an
impact upon present phenomena over and
above the influence that is mediated by the
traces left by the past in the present, then
we would have a case of temporal action at
a distance that is surely just as abhorrent to
the mind as spatial action at a distance. The
capital good).
prove is the possibility of
the process of production in a
manner where capital is absent, not the
indispensability of this representation. By
itself this possibility does not imply a conclusion to the effect that only labour has a
right to a reward, but if taken together with
the principle to each according to his work
such a conclusion does follow. Because the
working class as a whole, viewed as a sequence of generations of workers, generates
the entire product, viewed as a continuous
stream of output, it is entitled, on this principle, to the whole net product and nothing
is left for the capital owners.
The term of entitlement is used here deliberately, to show up the formal similarity
between the Marxist theory of exploitation
and Robert Nozicks theory of justice.15 In
Thus all
we can
representing
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
11
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
12
duction function
involving
the notion of
ally
that
net
(given
no
ex-
ploitation should be preferred to the neoclassical one. The appearance of a legitimate reward to capital owners is dispelled as an illusion once the process of production is seen in
a longitudinal or historical perspective or,
what amounts to the same thing, once working class solidarity is extended from the
intra-generational to the inter-generational
dimension.22 Once
we
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
13
control variable of the working class is the
working class consumption out of
production; the power over this
variable can be direct (through wage negotiations) or indirect (through taxes on profits). The control variable of the capitalist
class is the rate of investment out of profits,
i. e. out of what is left after working class
consumption. We shall set x~ for the control
variable of the working class, X2 for the control variable of the capitalist class, and a,
for the upper limit to xl. In modern capitalist economies, a, is very close to 100 per
cent; the workers have the power to consume all of the current output should they
so desire. The division of powers between
the classes may be described by saying that
the workers have the political power and
the capitalists the economic power, the workers the power over the present and the
capitalists the power over the future.
Before we go on to look at the solution to
this game, we may briefly characterize earlier
economic systems in the terms defined
above. Pre-capitalist economies had a low a,
and a low X2- Workers did not have the
power to raise their standard of living above
subsistence, and owners did not have any
desire to invest part of the surplus. Early
capitalism had a low a, and a high x2: workers were still at subsistence, but the reinvestment process had started up. One might
ask whether ul. or U2 is the most appropriate
measure of exploitation in this case. I believe that ul should be preferred, because the
investments did not have the effect of raising
the consumption level of the existing workers or their descendants, only of permitting
ever larger number of workers to live at
subsistence. From the point of view of the
individual worker, everything happened as if
a large part of his produce was taken from
him irrevocably; the fact that it was transformed into bread for more workers rather
than into caviar for the capitalists mattered
less than the fact that it was not transformed
into butter for himself.
In modern capitalism the situation is
rather characterized as one of high a1, low
rate of
current
xl, and
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
14
are
to
be
two
sary consumption.
I would not in any way try to draw an
idyllic picture of life on the plantations, nor
suggest that the slaves, because of their
hold on the masters, were in a better position
than the free workers who had no comparable hold on their employers; nor, finally,
am I suggesting that the reciprocal dependence relationship between master and slave
was in the least similar to the mutual power
relationship formalized by Lancaster in his
model of modern capitalism. The main point
of the brief remarks above was to substantiate the assertion that the quantitative
measure of exploitation is not very helpful
for the understanding of societies other than
those based on classical capitalism. For that
short period in human history class conflict
can be succinctly characterized by saying
that one class stole some of the goods produced by the other, this robbery gaining an
apearance of legitimacy from the alienated
state of mind in which the genesis of the
capital goods was hidden. In earlier and
later periods the class conflict has been
embedded in more ambiguous institutions
which cannot be reduced to the simple fact
of theft.
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
15
NOTES
1. Cp. Arrow (1977) and Sen (1976) for surveys
and discussions.
2. Sen (1973), p. 16-17. He assumes innate differences in the individual capacities for turning
consumption into enjoyment. A similar idea is
found in the writings of Leibniz (Elster 1975,
p. 152 ff.), who argues, however, that the differences stem from economies of scale in consumption. Giving more to those who are already better
(1974).
8. I
am
relying here
upon
Georgescu-Roegen
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
16
would use the undiscounted sum rather than the
discounted one; on the third because the marxist
measure of exploitation would divide the value by
the maintenance costs rather than by the gross
product. They seem quite unaware, however, of the
incongruity between the first element of their
derivation and the Marxist definition. In the
Marxist theory constructed by Fogel and Engerman, slaveowners would have an indisputable right
to profits on all other capital goods than slaves,
which is of course absurd.
12. A fairly good understanding of the issues
can be obtained by reading Harcourt (1973), Bliss
(1975) and the articles reprinted in Harcourt and
Laing (1971) and in Hunt and Schwartz (1972).
For a short, lucid and extremely instructive introsee Sen (1974).
13. Nuti (1970).
14. Cp. Elster (1976 a) for details.
15. Nozick (1974), especially p. 153 ff.
16. With unsurpassable clarity this is phrased as
follows by Nuti (1970): Attention is focused not
on past labour but on the present value of the
embodiment of past labour, and its current productiveness can be taken to provide a justification
for the attribution of the surplus of current output
over the wage bill to those who have appropriated
the embodiment of past labour, thereby providing
the current basis of future appropriation.
17. Cp. Luce and Raiffa (1957), Ch.6 for a
discussion of this and other solution-concepts for
duction,
cooperative
games.
ded
seem
to
me
(1976 b).
26. John Donne, The Prohibition; see also
Elster (forthcoming), Ch. 4.
27. Genovese (1974, p. 96) characterizes the
slaveowners in these terms: They were tough,
proud, and arrogant; liberal-spirited in all that did
not touch their honor; gracious and courteous;
generous and kind; quick to anger and extraordinarily cruel; attentive to duty and careless of
any time and effort that did not control their
direct interests.
28. Canarella and Tomaske (1975) suggest this
idea. I do not imply that it can be refuted a
priori; the choice between their conception and
Genoveses is a matter of empirical research.
29. Genovese (1974), p. 146.
REFERENCES
Arrow, K. 1977. Extended Sympathy and the Pos-
under
Semi-Feudalism.
Economic
Bliss,
History.
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014
17
Clark, J. M. 1899. The Distribution of Wealth.
New York: Macmillan.
Elster, J. 1972. Production et Reproduction. Essai
sur Marx. Unpublished doctoral dissertation for
the Doctorat dEtat. Paris: Université René Descartes.
Proceedings of the Fourth International Congress of the International Organization for the
Study of Human Development. Paris: Masson.
Elster, J. (forthcoming). Logic and Society. London : Wiley.
Fogel, R. & S. Engerman. 1974. Time on the Cross.
Boston. Little, Brown.
Genovese, E. 1965. The Political Economy of Slavery. New York: Pantheon.
Genovese, E. 1969. The World the Slaveholders
Made. New York: Pantheon.
Genoves, E. 1971. In Red and Black. New York:
Pantheon.
Genovese, E. 1974. Roll, Jordan, Roll. New York:
Pantheon.
Georgescu-Roegen, N. 1960. Economic Theory
and Agrarian Economics, Oxford Economic
Papers.
Georgescu-Roegen, N. 1976. Energy and Economic
Myths. New York: Pergamon.
Glantz, S. A. & N. V. Albers. 1974. Department
of Defense R & D in the University, Science
Maarek,
Sen,
Inequality. Oxford:
on Capital
Theory, Economica.
Sen, A. K. 1976. Welfare Inequalities and Rawlsian
Axiomatics, Theory and Decision.
Thompson, E. P. 1968. The Making of the English
Working Class. Harmondsworth: Penguin.
186.
Downloaded from jpr.sagepub.com at Serials Records, University of Minnesota Libraries on December 9, 2014