A WORLD TO WIN* The Political Economy of Neoclassical Marxism

c Peter King, unpublished
A WORLD TO WIN*
The Political Economy of Neoclassical Marxism
A new movement in Marxist thought has emerged in the past two decades.
Between the extremes of the purely mathematical treatment of Marxist eco-
nomics found in Morishima [1973] on the one hand, and the purely philo-
sophical treatment of historical materialism in Cohen [1978] on the other
hand, several thinkers—John Roemer, Jon Elster, Richard Miller, Alan
Buchanan, among others—who are technically sophisticated in both con-
temporary social science and contemporary philosophy have raised the level
of discussion of Marxist thought to new heights. Along with this new move-
ment has come a shift in emphasis, if not direction: topics such as alienation,
the dialectical method, commodity fetishism, and the role of the Party have
largely been put to one side in favor of topics such as the theory of explana-
tion, distributive justice, scientific methodology, exploitation, and economic
theory. Indeed, the distinctive use of neoclassical microeconomics has given
rise to the generic label ‘Neoclassical Marxism’ for the whole movement.
There is much that is valuable in this enterprise: its insistence on high
standards of rigor and conceptual clarity, too often absent from Marxism;
its integration of contemporary philosophical and economic thought into
areas once considered beyond the fringe; its informed critical stance, which
also manages to be sympathetic. Perhaps more valuable than anything else
has been its demonstration, by example, that interest in Marxism can be
intellectually respectable according to contemporary canons of respectabil-
ity. Or, to put the point colloquially, you don’t have to turn your back on a
hundred years of progress in the humanities and social sciences to do seri-
ous work on Marxism. This may be the lasting contribution of Neoclassical
Marxism, at least in anglo-american circles.
Roemer [1988] is a particularly good example of Neoclassical Marxism.
It summarizes and expands upon the technical results of Roemer [1981]
and Roemer [1982], while carefully paying attention to the historical and
philosophical context in which these results have their application. The
aim of Roemer [1988] is lucidly stated in the Preface (vii): “to trace the
connections between the economic concepts of Marxism and the ethical
ideas to which they are related.” Roemer’s method is also clearly stated
(op. cit.): “[t]he revised Marxism I present is shaped by the insights that
*
Presented at the Pacific Division APA Meetings, Los Angeles, 31 March 1990 as part
of a panel on the work of John Roemer.
– 1 –

---

2
A WORLD TO WIN
the tools of contemporary economics—that is, neoclassical economics—can
bring to bear.”1
Indeed, the main bulk of Roemer [1988] is given over to the attempt to
derive Marxist theses regarding classes and exploitation in a neoclassical
form, and rejecting those which cannot be so derived.
The results are rather surprising. To list only some of the contentions put
foward in Roemer [1988]: the labor theory of value is false (51); exploitation2
drops out the picture as not being of “fundamental interest” (89), and
is of derivative interest only as an index of inequality (103); under capital-
ism, “no one is coerced (in any normal sense of the word) to sell his labor
power” (32); social classes are not essentially related to exploitation, but
rather stem from the differential ownership of productive assets (76); agents
“choose their own class position” as a consequence of optimizing behavior
given their initial endowments (80); the existence of a labor market is not
a necessary feature of capitalism (96), and, for that matter, neither is a
capital market (103). On the positive side, having identified the differential
ownership of productive assets as the fundamental form of inequality at the
heart of capitalism, Roemer introduces a ‘general property-relations’ ap-
proach to exploitation (133), describes exploitation under socialism (140),
and concludes that the Marxist normative critique of capitalism is, essen-
tially, of a piece with contemporary theories of justice, differing primarily
in “the different counterfactuals that lie behind the moral judgments being
made” (137).
As a specimen of Marxist thinking, all this seems to be rather like Hamlet
without the Prince of Denmark. The conclusions that Roemer has reached
might provoke skepticism about his project—or, more exactly, skepticism
about the way in which he carries out his project: the project itself, namely
the careful analysis of the normative critique of capitalism offered by Marx-
ism, is an eminently sound and worthy one. And it is clear that the workman
is worthy of his hire: Roemer is extremely well-informed about Marxism and
neoclassical microeconomics. Nevertheless, it seems clear that something
1
See also Roemer [1988] 7: “My approach to Marxism is that of a contemporary student
of economics and political philosophy. I wish to study the logic of the ideas and the
internal coherence of the claims.”
2
Roemer is careful to state that he is speaking of what he calls exploitation in the
‘technical sense’: “Exploitation is said to exist if in a given economy some agents
must work more time than is socially necessary (longer than the socially necessary
labor time) to earn their consumption bundles and others work less time than is
necessary to earn their bundles” (Roemer [1988] 20). Note that this definition is more
limited than Marx’s; it does not include either the relational character of exploitation,
nor its coercive character.
c Peter King, unpublished

---

A WORLD TO WIN
3
has gone wrong.3
Now I do not want to object to Roemer’s conclusions simply because they
are at variance with traditional Marxist thought. Insofar as we hold on to
the hope of scientific socialism, we are free to claim that traditional Marxist
views are incorrect or confused about any number of points. Otherwise
there is only scholarship at best, and veneration at worst.
However, we can ask whether the analytical tools provided by neoclas-
sical microeconomics are the right tools for the job. There is certainly no
barrier in principle to the use of any given theoretical machinery to help
render socialism properly scientific, including neoclassical microeconomics.
Its absence from traditional Marxist thought is certainly no grounds for
rejection: compare the inappropriateness of rejecting the use of molecular
genetics in evolutionary biology on the grounds that traditional Darwinian
theory didn’t use it.
Yet even if there is no objection in principle, there might be specific
objections to the use of neoclassical microeconomics in clarifying Marxist
thought—that it does not provide the right tools for the job at hand. For,
as I shall argue, neoclassical microeconomics incorporates presuppositions
which are not normatively neutral, and indeed not compatible with Marxist
thought generally.4 Rational choice theory, in its particular incarnation as
neoclassical microeconomics, is not an adequate account of all aspects of
human activity—even of so-called ‘economic’ activity—though it may be
appropriate to more limited areas.5
The classical route to this conclusion is to endorse the traditional Marxist
account of ideology: ‘economics’ is not an autonomous and normatively
neutral discipline—which is perhaps why Marx himself is always careful to
refer to it as political economy—but rather is a series of ideas specific to a
particular social formation, capitalism, whose functional role is to reinforce
and justify the dominance of the capitalist class.
I see no reason to quarrel with this classical approach. But, as it stands,
3
That is, with respect to the conclusions described on the preceding page. There is
much else of value in Roemer [1988].
I have chosen to focus on his controversial
conclusions for obvious polemical purposes.
4
This is not to deny that rational-choice models can be applied to good effect to partic-
ular cases in the context of Marxist thought. Nevertheless, such rational-choice models
are more more at home in the analysis of the social thought of, say, Hobbes—see for
instance the works of David Gauthier and of Jean Hampton.
5
The traditional Marxist critism of rational choice theory, in its guise as neoclassical
economics, is that it is all too accurate a portrayal of a world gone badly awry—
the ‘inverted world’ of capitalism. That is, capitalism’s theoretical self-understanding
accurately portrays the degraded world of capitalism.
c Peter King, unpublished

---

4
A WORLD TO WIN
it provides at best only a sketch of what exactly is wrong with neoclassical
microeconomics; it is not itself an explanation why neoclassical microe-
conomics is inappropriate. (Indeed, it might be thought to be peculiarly
well-suited for analyzing capitalism, as a part of capitalism’s theoretical
self-understanding.) Labelling a particular set of ideas an ‘ideology’ is no
better than name-calling without an account of ideology, and why a given
set of ideas is an ideology, to back up the charge. So I shall put questions of
ideology to one side, and concentrate on directly identifying the question-
able presuppositions embodied in neoclassical microeconomics as a branch
of rational choice theory.
First, the relation between rationality and rational choice theory. In an
argument famously associated with Davidson [1980], the interpretation of
action in terms of belief and desire depends on the attribution of rationality.
Informally, an agent is rational when she acts to get what she wants given
what she believes. But this thin conception of rationality does not entail
the thicker conception of rationality used in neoclassical microeconomics,
which is identified with optimizing behavior in market interaction (an agent
is rational when she acts so as to maximize her expected utility). To see why
this should be so, let us take a closer look at the fundamental concepts of
neoclassical microeconomics, namely utility, indifference, rationality, market
interaction, private property rights, and mutual disinterest.
Given an agent’s individual preferences over a set of states of affairs,
utility is a measure-function defined over outcomes representing relations of
preference: if there is a weak total ordering of such outcomes which is tran-
sitive, then utility can be defined as an ordinal measure, which is sufficient
for most purposes in neoclassical microeconomics. Further assumptions,
namely monotonicity and continuity, are required for utility to be a cardi-
nal measure. Since the ordering-relation is weak—ties are permitted—an
agent’s choice allows for indifference among alternatives, which is a non-
explanatory description of the agent’s behavior. Clearly, the presumption
underlying the use of these concepts is that an agent’s desires are all com-
mensurable; an agent is represented as having a single fundamental need,
the need for utility, which may be equally well satisfied with different ‘con-
sumption bundles’ of various goods, with regard to which bundles the agent
is indifferent.
Given this presumption, rationality is identified with the
maximization of (expected) utility. These concepts are sufficient to define
individual rational choice under conditions of certainty, or so-called ‘para-
metric’ rationality.
The remaining three foundational concepts of neoclassical microeconomic
thought describe the case of interaction on a purely competitive market. For
c Peter King, unpublished

---

A WORLD TO WIN
5
market interaction, “[e]ach person is assumed to want to consume as many
products as possible, and as much of each as possible, although desire is
assumed to diminish as consumption increases. Each also wishes to provide
as few factor services [used in production], and as little of each, as possible,
and here the desire not to provide such services is assumed to diminish
as one’s supply of those services diminishes.”6 Three asssociated activities
take place in market interaction: production, consumption, and exchange.
More exactly, productive assets are bought or hired, and goods for (private)
consumption are acquired; these simply are the kinds of exchange which may
take place on the market. Yet in order for such exchanges to take place,
there must be private property rights: an agent must have exclusive control
over what is owned, reaping the benefits from its possession and managing
it without interference, as well as having the power to dispose of what is
owned at will (the power to alienate what is owned). This includes private
property in all productive assets, and hence in the agent’s natural abilities,
such as the capacity to perform labor. For the operation of a free market,
the initial distribution of such productive assets, prior to exchange, may be
taken as fixed. Finally, agents are characterized by mutual disinterest: an
agent takes no interest in the interests of those with whom she exchanges.
This allows the utility functions of each agent to be strictly independent,
determined only by the goods she consumes and the factor services she
provides.
Under these conditions, interaction on a purely competitive free market
has three features worthy of note. First, it is a case of rational choice under
conditions of certainty, for each agent treats the actions of other agents as
fixed circumstances. Second, and more importantly, optimality and ratio-
nality coincide: if each agent acts rationally, i. e. maximizes her utility, then
the equilibrium outcome of all actions will be optimal. This sets the free
market sharply apart from most real-world situations, in which rationality
and optimality do not conincide (e. g. in prisoner’s dilemma cases). Third,
it should be clear that the commonsense notion of ‘rationality’ described
above does not entail the sophisticated sense of ‘rationality’ as it applies to
market behavior, which involves many more assumtions about agents and
the circumstances of their interaction.
Not surprisingly, Roemer’s use of neoclassical microeconomics leads him
to focus on circumstances ‘external’ to the operations of the market to
provide grounds for the normative critique of capitalism. The differential
ownership of productive assets is such an external circumstance, and Roemer
6
Gauthier [1986] 85.
c Peter King, unpublished

---

6
A WORLD TO WIN
is certainly correct to draw our attention to its importantance. I do not want
to minimize his point. Yet at the same time it is also important to point out
that the fundamental concepts of neoclassical microeconomics are free from
criticism, precisely because they are part of the very method Roemer uses.
This is justifiable only to the extent that these concepts do not incorporate
any substantive normative presuppositions. Yet they do.7
To begin with the most obvious case: why should we adopt the require-
ment of mutual disinterest? It seems to be plainly false, since people do
take an interest in one another’s interests; altruism and malevolence are
both common phenomena. But there is a deeper point at stake here. The
requirement of mutual disinterest not only rules out one agent’s utility being
directly dependent upon another’s, which we might find to be an unrealistic
but acceptable idealization, but also rules out any interests agents might
have about the circumstances in which such exchanges take place. Stated
bluntly, it rules out political interests: placing any sort of value on the condi-
tions of interaction. And while we might rule out bonds of personal affection
or disaffection, there is no good reason to rule out political interests.
This point naturally leads to another. How should we understand the
neoclassical microeconomic models Roemer puts forward?
In standard
economic theory, models are often taken to be descriptive in character—
approximations to real-world situations. On this score, mutual disinterest
simply does not match the actual interests of actual agents. Nor can Roe-
mer’s models be taken in a purely instrumental fashion, ignoring the the-
oretical machinery they incorporate, where their justification is given by
their overall success in allowing predication and control. His models are
simply not that kind of model. Furthermore, Roemer explicitly states that
7
In a famous passage, Marx directly addresses—and derides—the normative presuppo-
sitions of the purely competitive free market: “This sphere we are deserting, within
whose boundaries the sale and purchase of labor-power goes on, is in fact a very Eden
of the innate rights of man. There alone rule Freedom, Equality, Property, and Ben-
tham. Freedom, because both buyer and seller of a commodity, say of labor-power,
are constrained only by their own free will. They contract as free agents, and the
agreement they come to, is but the form in which they give legal expression to their
common will. Equality, because each enters into relation with the other, as with a
simple owner of commodities, and they exchange equivalent for equivalent. Property,
because each disposes only of what is his own. And Bentham, because each looks only
to himself. The only force that brings them together and puts them in relation with
each other, is the selfishness, the gain and the private interests of each. Each looks to
himself only, and no one troubles himself about the rest, and just because they do so,
do they all, in accordance with the pre-established harmony of things, or under the
auspices of an all-shrewd providence, work together for their mutual advantage, for
the common weal and in the interest of all” (Capital I.ii.6).
c Peter King, unpublished

---

A WORLD TO WIN
7
his approach is not “an empirical one” (7). But if his approach is prescrip-
tive, then any normative presuppositions stand in need of justification.8
Mutual disinterest is hardly a postulate of pure reason, and an unlikely
candidate for a postulate of practical reason. Roemer dismisses traditional
Marxist criticism of capitalism as productive of needs and desires which
are detrimental to human self-realization, for the reason that such criticism
“is nonliberal, because it passes judgements on what kinds of preferences
are good for people, independent of what people themselves may think”
(153).9 But if mutual disinterest is held up as a normative ideal, then it
too is ‘nonliberal,’ since it directly legislates the content of preferences. To
put the point sharply: why think it part of admirable, much less actual,
human behavior to exclude preferences whose content is other-directed or
communal in character?
Consider another feature of interaction on the free market, namely its
characterization of rational human activity in exchange as purely paramet-
ric. There are two fundamental things wrong with this view, according
to traditional Marxist thought. First, it is the very phenomenon stigma-
tized by Marx as ‘commodity fetishism’: the representation of a collective
creation, namely the conditions of market interaction and supply/demand
phenomena, as something which has to be dealt with as if it were a natural
force like the weather (Capital I.i.4).10 The parametric rationality of mar-
ket behavior systematically obscures the fact that market phenomena, and
the very conditions of market interaction, are social rather than natural.
Insofar as such parametric rationality is held up as a normative model, it
covers over an important fact about market interaction itself: that it is, in
the broadest sense, optional.
Second, the form of interpersonal interaction on the market, in the sphere
of exchange, is purely instrumental: each person treats those with whom
she interacts as so many means towards maximizing her own expected
utility, nothing more, regardless of class position. Indeed, this is one of
the fundamental normative criticisms of capitalism Marx puts forward.11
8
It would beg the question to claim that the tools of neoclassical microeconomics are
validated by their successful employment elsewhere, since that is precisely what Marx-
ist thinking has called into question.
9
Roemer is right on target here: the traditional Marxist criticism of capitalism, based
on a conception of human flourishing, is nonliberal. But I am at a loss to see why
that should be a reason for dismissing it—quite the contrary, I should think.
10 See Ripstein [1987] on commodity fetishism.
11 Marx describes the merely instrumental recognition of others as one of the failings of
capitalism (Hoare 274–275): “In itself the result of my production has just as little
c Peter King, unpublished

---

8
A WORLD TO WIN
For capitalism incorporates a ‘systematic inversion of ends and means,’ as
Agnes Heller calls it.12 Since capitalist production is for the sake of profit-
maximization, it is an inherent feature of the economic system that it treats
individuals as mere means to the maximizing of profit, recognizing the in-
dividual only as a depository of produced goods. What is distinctive of
human beings—the ability to develop in their needs (and thereby to de-
velop in their capacities)—is treated as a means only, and not as an end
in itself. Yet even if we dismiss this ‘nonliberal’ criticism of capitalism in
general, the point remains that neoclassical microeconomics countenances
the instrumental use of persons, which is clearly a substantive normative
assumption ‘built into’ neoclassical microeconomics itself.
(The claim that market exchange interaction is instrumental in character
could be denied, on the grounds that each party to the exchange enters into
it voluntarily, and thereby treats the other not as a means only, but also
as an end in herself. But the point above is unaffected by this move: it
merely proves that there is a substantive normative point at issue here, one
requiring defense.)
Now let us consider the activity of consumption as represented by neo-
classical microeconomics. I have already noted that there is at least a theo-
retical reduction of an agent’s needs to a single fundamental need, utility—
or, as Marx puts it, the need for the means to maximize expected utility,
namely money.13 Since the need for money is intrinsically without limit
and equally intrinsically ‘object-less,’ the only desires which are system-
atically reinforced under market interaction are acquisitiveness and greed.
Furthermore, the satisfaction of all needs comes to be a matter of simply
spending the money, of gaining private property in the object bought—of
direct relation to you as the result of your production has to me. That is to say, our
production is not man’s production for man as man, i. e. it is not social production.
As men none of us has a claim to enjoy the product of another. . . Hence our exchange
cannot be the mediating movement which confirms that my product is for you because
it is an objectification of your own nature, of your own need.”
12 See Heller [1976] 50: “The numerical growth of needs will never be able to become
true wealth, because it is merely a means serving an alienated force, alien to individual
human beings, i.e. the expansion of capitalist production.”
13 See Marx’s early essay, ‘The Meaning of Human Requirements” (Tucker 93): “The
need for money is therefore the true need produced by the modern economic system,
and it is the only need which the latter produces. The quantity of money becomes
to an ever greater degree its sole effective attribute: just as it reduces everything to
an abstract form, so it reduces itself in the course of its own movement to something
merely quantitative.”
c Peter King, unpublished

---

A WORLD TO WIN
9
mere ‘having.’14 Hence the commensurability of needs systematically dis-
courages the recognition of the differences among the objects of needs as
well as the cultivation of needs themselves. There is a parallel situation
in the case of production. Because under capitalism all needs are repre-
sented as the need for money, needs which are only capable of other kinds
of satisfaction are not developed—in particular, the need for meaningful
activity and satisfaction through work. Rather, the requirements of pro-
duction for exchange, i. e. sale and consumption, lead to the creation and
manipulation of needs in other agents, reinforcing simple consumption, as
a consequence of a social system in which production is geared towards ex-
changeable, rather than useful, objects.15 Since it is an open question how
the conditions of human interaction should be set up, it is a non-trivial
normative claim that they should be directed toward exchange as the medi-
ating mechanism—rather than, say, dispensing with the market mechanism
for the allocation of goods and structuring production directly for needs, as
traditional Marxist thought proposes. The market mechanism itself stands
in need of defense, rather than simply being presupposed.
Indeed, the fundamental theoretical division of production and consump-
tion presupposed by neoclassical microeconomics itself incorporates presup-
positions which are by no means neutral. For each agent is assumed to
want to consume as much as possible, which is the arena for pleasant activ-
ity, and desires to produce—to provide factor services—as little as possible,
which is the arena for unpleasant activity. (This is indeed how neoclassical
microeconomics purports to explain wage levels.) But there is no neces-
sary correlation between work and unpleasantness, and any postulated by
neoclassical microeconomics stands in need of justification.
Yet there is a deeper point at stake here. The segmentation of human
activity into the spheres of production, exchange, and consumption typical
of neoclassical microeconomics has another, less obvious, effect. If homo
economicus maximizes expected utility through the consumption of bun-
14 See “Notes on James Mill” (Hoare 274): “Man—this is the fundamental premise of
private property—produces only in order to have. Having is the aim of production.
Furthermore, production not only has such a useful aim; it has a selfish aim. Man
produces to have something for himself. The object of his production is the objectifica-
tion of his own immediate selfish needs.”See also “Private Property and Communism”
(Tucker 87): “Private property has made us so stupid and one-sided that an object is
only ours when we have it. . . ”
15 See “The Meaning of Human Requirements” (Tucker 93): “Under private property
the significance [of human needs] is reversed: every person speculates upon creating
a new need in another, so as to drive him to a fresh sacrifice, to place him in a new
dependence.”
c Peter King, unpublished

---

10
A WORLD TO WIN
dles of goods, then needs are only fulfilled through consumption, not in
production. In particular, this means that work itself is construed solely
as an instrumental activity, one which is always subordinate to a further
end—under capitalism, it is the means by which one obtains the money to
engage in consumption. Hence productive activities are mere means to an
extrinsic end, and not in themselves meaningful or valuable (or only acci-
dentally so). Hence the deep presupposition incorporated in neoclassical
microeconomics is that human action is only motivated by, and hence only
makes sense in terms of, consequences which are realizable independently
of the action.16 At best, this is a questionable claim: we certainly seem to
engage in activities for their own sake, and often identify ourselves in terms
of them. It is descriptively false, and normatively stands in need of further
justification.
If we focus on the case of capitalism in the real world, this point emerges
clearly: the market in labor-power renders labor, just like any other com-
modity, commensurable (‘abstract’). To facilitate such commensurability,
production is developed to the point at which individual characteristics,
such as talent, skill, or ability, are not relevant to the performance of the
work at hand. That is, the jobs available on the labor-market become sim-
plified, repetitive, and do not of themselves contribute to the development
of any individual talents or abilities, hence intrinsically undesirable, rein-
forcing the purely instrumental character of work.
Let us turn now to the private property rights incorporated in neoclas-
sical microeconomics. Traditional Marxist thought—beginning with Marx
himself—have rejected the apparatus of natural rights wholesale, on a vari-
ety of grounds, most frequently nowadays because the apparatus of natural
rights seems metaphysically incoherent.17 Yet it is by no means clear that
neoclassical microeconomics is committed, in any sense, to a strong theory
of ‘natural’ rights; all that matters for market interaction is that the in-
16 See Arthur Ripstein’s “Rationality and Alienation,” forthcoming in the supplementary
volume on Analytical Marxism published by the Canadian Journal of Philosophy.
17 Marx offers at least two independent reasons for rejecting natural rights. First, in his
early essay “On the Jewish Question” (Tucker 42), Marx criticizes those natural rights
which are liberties for being inherently separative—they denote a sphere of ‘free’ ac-
tivity, the exercise of which is constrained by the Harm Principle; hence natural rights
represent other people as limitations of one’s freedom, rather than as the realization
of it. Second, in the late comments known as the “Critique of the Gotha Programme”
(Tucker 530), Marx criticizes those natural rights which are entitlements for being
inherently unequal in their content, that is, for overlooking relevant individual differ-
ences; hence such entitlement-claims legitimize inequality, particularly with respect to
distribution.
c Peter King, unpublished

---