[OPE-L:3426] Zero repression costs?

andrew kliman (Andrew_Kliman@msn.com)
Tue, 15 Oct 1996 10:32:33 -0700 (PDT)

[ show plain text ]

This replies to Paul C's ope-l 3394. I wish to thank Paul for returning to
the issues we were discussing.

To get some minor points out of the way first:

Paul: "I will repeat the question, why, in capitalism is human labour rather
than something else, such as equine labour or sunlight, the source of value?

"To reply as Andrew does :'Because in Marx's value theory, human labor, not
horse labor, is the source of value.' is surely to avoid the point. Sure we
know Marx said that, but it is not true just because Marx said it, there has
to be some underlying real cause."

This is not a repetition of the question but a different question. Paul's
original question (ope-l 3350) was "Why should the wages of a slave overseer
count as constant capital, when the wages of teamster supervising the labour
of horses counted as variable capital?" I stand by my answer to that

A common mistake among empiricists is to confuse theoretical categories and
characteristics of the world, lumping them together as "facts." It is because
Paul makes this mistake that he thinks I avoided his point, and appealed to
Marx to justify a claim concerning characteristics of the world. I did
neither. "Constant capital" and "variable capital" are theoretical
categories, *Marx's* theoretical categories, and the original question asked
why one event is included in *his* theory under one category and a second
event under the other category. I answered that clearly and sufficiently.

Paul's *new* question concerns not theoretical categories, but characteristics
of the world. To ask a "why" question in this context is besides the point
and probably unanswerable. Just as the law of gravity is meant to explain
phenomena, so is the law of value. "Why is there gravity" is besides the
point and probably unanswerable. What is important here is to account for
what takes place, not "why." Marx certainly recognized this; I've never
found a single sentence in his works that attempts to answer Paul's question.

In any case, Paul's own answer "because wages are the single most prevalent
cost of production," is a regression to a na´ve cost of production, adding-up,
labor commanded theory of value.

Paul: "When Andrew says:
>Machines are used to
>bring an ahuman, "objective" discipline to the labor process, and the
>subjection of the worker to this discipline.

he takes a side effect of some machines and makes it the main purpose of
machinery in general. Machines are employed to cheapen the cost of production,
generally by enabling tasks to be achieved with less living labour, but
sometimes to enable raw materials and other inputs to be saved. It is to this,
rather than the intensification of labour, that the great progress of
capitalist civilization is largely due."

I agree that the "main purpose" of mechanization is to cheapen the cost of
production, whatever component of cost is being reduced. (BTW, this is a
rather strange statement coming from an opponent of "idealist" teleology and
Hegelianism. Steve Cullenberg is much more consistent in this regard.) But
if that is its *main purpose*, then, if mechanization does bring about the
elimination of the human being from production, or his/her domination by the
means of production, this is certainly no "side effect." It is the *main
purpose*, as Paul admits, of those innovations that "enabl[e] tasks to be
achieved with less living labour."

Note that the statement of mine that Paul quotes does not refer narrowly to
intensity of labor, which is what he reduces it to. It is costly to pump
labor out of labor-power for a variety of reasons. One is that no human being
can ever match the intensity of a machine, so what is involved in
mechanization is not merely a speed-up of the remaining human workers. Others
are that the workers resist and fight back --- and sabotage, slowdowns,
strikes, sitdowns and other S'es raise the unit cost of output, so it is
rational for capitalists to get rid of them by mechanizing, try to break their
resistance by mechanizing, use machines to monitor the workers' effort. The
key to everything is that the "specifically capitalist mode of production,"
Marx's term for machinofacture, overthrows the subjective principle on which
production had been based and replaces it with a strictly objective
organization of production. (This, BTW, was precisely Marx's analysis.)

On to the main issue. Paul writes: "It is difficult to imagine any situation
where labour could be had for free, the only one I could think of at the time
being the working of capitives to death. Andrew argued that this would still
involve accumulation of constant capital in the form of guards, guns and
barbed wire.

"This point falls to two objection, one of which has already been made by

"a) it involves treating the costs of supervision as capital, whereas
in Marx they are treated as an unproductive overhead deducted
from surplus value."

First, it is worth savoring the irony of this appeal to Marx from someone who
has claimed to think that the mere discussion of his ideas is scholasticism.
And not only does Paul mention him, but he uses Marx as "authority" to justify
his own claim concerning the operation of capitalism (whether mechanization
would be rational were v to = 0), something which I think we all agree is
invalid and dogmatic.

So this objection is irrelevant even if Marx treated these costs in the way
Paul says he did, which I think is not the case. But I don't want to argue
that point now, because the objection is irrelevant (both when Paul makes it
and when Jerry makes it) for another reason. If capitalists are concerned
with maximizing their *net* profit, it simply doesn't matter whether these
costs are counted as a deduction from gross profit (surplus-value) or as an
advance of capital. Call P the price of output, M the cost of materials and
machines, W wages, and R repression costs. Assume all other costs and
deductions from surplus-value = 0. Then, if repression costs are counted as

net profit = gross profit = P - capital advanced = P - (M + W + R)

whereas, if R is counted as a deduction from gross profit instead,

net profit = gross profit - R = P - capital advanced - R = P = (M + W) - R = P
- (M + W + R).

So it makes no difference. If either Paul or Jerry wants to argue that
capitalists care about their gross profit instead of their net profit, I'm
willing to listen. Otherwise, arguing about the fund that R comes from is
irrelevant here. This also means that Jerry's statement in ope-l 3372:

"to not explicitly include R is different from negating the existence of a
category, such as v, which has already been
developed and is crucial for the subject matter in question -- accumulation
and the general r."

is wrong --- net profit and thus R is crucial for understanding accumulation
and the general r --- and thus the assumption that R = 0 is impermissible
according to Jerry's methodology. He thus has a double standard, as I have

Paul's second objection is:
"b) it assumed that costs of supervision were a constant fraction of the
amount of labour employed. The history of mass forced labour indicates
that this is not the case, that as the number of forced labourers
rises, the cost of supervision falls. It is not 1000 times as difficult
to guard 1000 slaves as it is to guard 1. Thus under conditions of
forced labour an increase in the scale of production leads to lower
costs of supervision so that even if we accept Andrews doubtful accounting
whereby costs of supervision/repression count as capital, then
we would expect the organic composition of such 'capital' to fall
as the scale of production rose."

Actually, my specific example assumed diminishing, not constant, returns to
repression costs. But it doesn't matter. As I also noted, all that matters
is that labor extraction increases monotonically with R, so the increasing
returns to R that Paul here assumes do not change the conclusion. Assume a
capitalist firm is deciding whether a new technology will raise its net
profits, N, for a *given* level of output, and will adopt it iff that is the
case. Then, using D to indicate "change in," the viability condition is:

DN > 0, which implies that

DP - DM - DW -DR > 0.

Now, rewrite W, total wages, as wL, where L is the amount of living labor and
w is the wage per unit of living labor, and hold P, price, and w constant,
since the technological change doesn't affect them directly. Then, the
viability condition is:

-DM - wDL - DR > 0


-wDL - DR > DM


(w + DR/DL)(-DL) > DM

Now, assume the new technology reduces the employment of living labor per unit
of output, so that (-DL) > 0. Then the viability condition is

(w + DR/DL) > DM/(-DL)

DM/(-DL) is a measure of the change in the organic composition of capital.

Let's assume that w = 0. Also, let's assume R = 0 and thus DR = 0, although
this is surely unrealistic and indeed impossible under capitalism, and it
therefore negates the object it is supposedly studying and makes an
understanding of both capitalist accumulation and Marx impossible. Then, but
only then, is it the case that an increase in N requires a reduction in
machine and materials costs relative to living labor [i.e., DM/(-DL) < 0].
Then but only then is it always irrational to use more M relative to L. Then
but only then would I predict that a profit maximizing firm would forego
investment in machines and use handicraft labor alone.

This case, but only this case, conforms to Paul's notion, which he has
repeated as: "if wages were zero there would be no incentive to accumulate
constant capital, free handicraft labour would always out compete it. More
generally, the cheaper labour is to be had, the less likely there is to be
accumulation of capital in the form of more machinery per worker."

Let us, however, examine the case in which w = 0 (YES!) but R > 0 (YES!) *and*
DR/DL > 0. Note that DR/DL > 0 only implies that total repression costs and
total extraction of living labor move up and down together; it doesn't imply
increasing, or constant, or diminishing returns to repression. So Paul may
assume DR/DL falls as L increases.

In this case the new technology will be viable iff (DR/DL) > DM/(-DL). Hence,
even if the machines and materials replace living labor --- DM/(-DL) > 0 ---
under the new technology, it is rational to invest in it iff the rise in M is
more than offset by the fall in R that is brought about by not having to
exploit as much living labor. So Paul is wrong when he writes: "Machinery
only becomes profitable to use when wages are high."

Note again that none of the above conclusions depend on whether R is counted
as a capital advance or a deduction from gross profit.

As I've stated before, my conclusion *also* carries over to the case in which
w > 0 as well, though it is seen most clearly by assuming w = 0, as I just
did. If we let w be > 0, then it may appear that the incentive to invest in M
and reduce L is due only to the firm being able to reduce wage costs. But by
setting w = 0, we see that this is totally incorrect; there is an incentive to
increase M relative to L in order to reduce repression costs, *regardless* of
the level of wages.

Thus, again, the conclusion is clear. I have demonstrated, and demonstrated
again, that Paul's statement in ope-l 3317:

"If v were 0 then there would be no incentive for capitalists to accumulate
constant capital. It would
obviously be cheaper to throw away their machines and have all the work done
by hand."

is false. As I wrote in ope-l 3362: "In ope-l 3340, I then demonstrated
conclusively that (a) even if v were 0, there would be an incentive for
capitalists to accumulate constant capital, and (b) it would not necessarily
be cheaper to throw away their machines and have all the work done by hand. I
explained that the basic reasons for these conclusions are that investment in
constant capital is needed to extract labor from labor-power when workers
don't want to do the work and that it can be
cheaper to produce with machines rather than having to bear these non-wage
'labor costs' to put down or keep down recalcitrant workers."

I have made the latest demonstration as general as possible so that specific
assumptions won't (I hope) be picked at. We are dealing with an elementary
cost minimization problem. If capitalists buy labor-power instead of labor,
so that they must incur costs in order to get the workers actually to work,
and these costs increase with the amount of work done, then one must conclude
that I've been right all along and, at the risk of annoying our neutral
disinterested moderator, conclude that my critics have been wrong.

Andrew Kliman