[OPE-L:5179] Reproduction Costs vs. Replacement Costs

andrew kliman (Andrew_Kliman@msn.com)
Wed, 4 Jun 1997 14:00:21 -0700 (PDT)

[ show plain text ]

A reply to Duncan's ope-l 5164. I thank him for responding to the textual
evidence I presented.

Duncan: "Isn't the issue of replacement cost quite prominent right in the
first chapter of Volume I?"

No. To my knowledge Marx discusses replacement costs only in connection with
reproduction, never in connection with the determination of value.

Marx does continually say that a commodity's value is not determined by the
labor-time that went into producing it originally (historical cost), but by
the labor-time needed to *reproduce* it. It is quite a leap from this to the
*replacement* cost interpretation. That interpretation maintains that the
value of a shirt produced today is the sum of the value added by living labor
plus what it would cost, tomorrow, to replace the cloth, buttons, etc., not
what it does cost, today, to use them to produce new shirts. Marx never says
anything like this.

At times, he explicitly says the opposite. For instance, in MECW 33, p.
91(emphasis in original), he writes:

"Profit, however, expresses the excess of the *value of the product over the
value of the whole of the costs of production*; hence it expresses in fact the
increment of value which the total capital receives at the end of the
processes of production and circulation, over and above the value it possessed
before this process of production, when it entered into it."

Note that this implies that costs of production are understood to be given by
the value the capital possessED BEFORE production, in fact precisely the value
it had when it entered into production. The temporalist reasoning is
unmistakable. Note that this passage conforms exactly to what I wrote in
ope-l 5045 (before I happened across it): "the value transferred is
determined by the cost of reproducing the means of production when they enter
production, i.e., when their use-value is destroyed. In other words, the
value transferred is determined neither by historical cost, nor by
post-production replacement cost, but by the current pre-production cost of
the means of production."

If there's a way of reconciling this passage with the replacement cost
interpretation -- other than by claiming that Marx didn't mean what he wrote
-- I'd like to hear it.

BTW, Marx is certainly not alone in this. In his work on intertemporal
general equilibrium, Malinvaud also defines profit as the difference between
revenues and pre-production costs, and the rate of profit is likewise computed
on the basis of pre-production costs.

I wrote:

(1) Assume that workers harvest some wild corn, having no value, plant it,
and harvest new corn. The new corn has value, due to the value added by the
living labor. According to the replacement cost interpretation, then, a
positive amount of value is 'preserved' and 'transferred' from the wild corn
to the new corn. How do you reconcile this with the following from Capital
I, p. 314 of the Vintage ed. (emphasis added):

"'Its value is determined not by the labour process into which it [a means of
production] enters as a means of production, but by that out of which it has
issued as a product. In the labour process it serves only as a use-value, a
thing with useful properties, and cannot therefore transfer any value to the
product unless it possessed value BEFORE its entry into the process'?"

Duncan replied:

"If wild corn is a free good, then I don't think it would play a role in the
pricing of the new corn, any more than the air would have in pricing the
output of a steam-powered mill.

"I don't see the difficulty. Since the wild corn did not have any value (on
the assumption that it's a free good) it wouldn't transfer any value to the

I'm very glad to hear this! I'm hopeful that this example may help clarify
matters. I especially appreciate the recognition that the actual prices of
inputs "play a role" in the pricing of outputs. This is in contrast with the
replacement cost interpretation, according to which it is irrelevant what the
value of the means of production was when they entered production. The value
of free goods is "preserved" -- even though they had no value -- and they
"transfer" to the product this sum of value that they didn't have.

Hence, if free goods do not transfer value, the replacement cost
interpretation must be rejected. Moreover, if free goods do not transfer
value, the monetary expression of labor-time (MELT) cannot be computed on the
basis of the "value of the net product." Let me show this by putting some
numbers on my example:

Wild corn input: 4 bu.
Living labor added: 1 labor-hour
Corn output: 5 bu.
Output price of corn: $x/bu.
"value of net product": [$x/bu.]*[5 bu. - 4 bu.] = $x

According to the "net product"-based measure of the MELT, it equals the "value
of the net product" divided by living labor added, $x/(1 labor-hour) =
$x/labor-hour. This means that the 1 hour of living labor extracted "added a
money value" of $x. But the 5 bu. of corn output are worth a total of
[$x/bu]*[5 bu.] = $5x. So the other $4x is the "value transferred" from the
free wild corn.

This "net product"-based measure of the MELT also implies that, measured in
labor-time, the value of the corn output is $5x/[$x/labor-hour] = 5
labor-hours. Since value added by living labor is 1 labor-hour, the free
seed-corn "transfers" a value of 5 - 1 = 4 labor-hours.

The actual situation, I think we agree, is this. The entire value of the corn
output is the value added by living labor. The 5 bu. of corn output thus have
a total value of 1 labor-hour, expressed monetarily as $5x, so the MELT at
time of output equals $5x/(1 labor-hour) = $5x/labor-hour. No value is
transferred from the free seed-corn.

It should be clear that this discrepancy between the figures based on the
"value of the net product" and the actual figures is just a particular
instance of the discrepancy that arises whenever the cost of inputs at the
time they enter production differs from the replacement cost of these inputs.
It should also be clear that the second set of figures ("actual situation")
follows straightforwardly from the TSS interpretation.

I do not think that it is possible to reconcile the two sets of figures, or to
discount the latter, by referring to the revaluation of inventories or stocks.
In this example, the 4 bu. of seed-corn no longer exist as commodities once
they are planted, so there are no unused stocks to revalue at the end of the

I also wrote:

"(2) The replacement cost interpretation implies that the rate of profit (in
production) obtained in production of luxury goods has no influence on the
general rate of profit, as Bortkiewicz demonstrated. Yet Marx held that the
opposite. How do you reconcile his conclusion with the replacement cost

Duncan replied: "Actually, this is one point where the New Interpretation
differs from Bortkiewicz. Since we define the monetary expression of labor
time in terms of the value added in the whole mass of commodities, that
includes luxury goods."

I don't think this matters. If we consider a 3-dept. scheme, without fixed
capital, but with inputs being costed at their replacement cost, so that the
relevant input and output prices must be equal, the equal-profit-rate
equations are:

(1) (p1*a1 + p2*b1)(1+r) = p1

(2) (p1*a2 + p2*b2)(1+r) = p2

(3) (p1*a3 + p2*b3)(1+r) = p3

where the ai are the means of production per unit of output i and the bi are
the wage goods per unit of output i. Good 3 is the luxury good.

Now, simultaneously solving (1) and (2) for (1+r), we get

(a1*b2 - a2*b1)(1+r)^2 - (a1 + b2)(1+r) + 1 = 0

so that the profit rate, r, is a function solely of the coefficients a1, a2,
b1, and b2. It does not depend on the units in which prices are measured, and
it does not depend on the coefficients a3 and b3, or p3.

Thus, the definition of the MELT has no effect on the profit rate, and
production conditions in luxury sectors continue to be irrelevant to the
determination of the profit rate, as long as input and output prices are
constrained to be equal, as the replacement cost interpretation requires.

I also wrote:

"(3) Torrens had written: 'The farmer ... expends one hundred quarters of
corn in cultivating his fields, and obtains in return one hundred and twenty
quarters. In this case, twenty quarters, being the excess of produce above
expenditure, constitutes the farmer's profit ...' Marx objected that
'*profit* ... is applicable solely to exchange-value,' and that 'As far as
exchange-value is concerned, there is no need to explain further that the
value of 90 quarters of corn can be equal to (or greater than) the value of
100 quarters, that the value of 100 quarters can be greater than that of 120
quarters, and that of 100 quarters greater than that of 500' TSV, 3, p. 77,
pp. 78-79. How do you reconcile Marx's response with the replacement cost

Duncan replied that Marx was "trying to illuminate the point that profit, as a
value category, is
not coextensive with the physical surplus in production"; that "I haven't
looked this up in context, but as written it seems incoherent. It seems to
need some stipulation about the time at which the comparisons are being made.
In the Torrens example the production period intervenes; and that "Since the
quote isn't very clear to me, I don't see necessarily that it's
in conflict with the replacement cost interpretation."

Please do look it up, so you can see the whole context, which is very
important. The whole thing in context is far too long for me to reproduce
here. The last passage I quoted is indeed incoherent, or at least has nothing
to do with the issue at hand, *unless* the first of each pair of figures
refers to inputs and the second refers to outputs. This seems pretty obvious,
because that is what the passage from Torrens stipulated: 100 qrs. are
expended, 120 qrs. are produced. In saying that "the value of 100 quarters
can be greater than that of 120 quarters," Marx is saying that profit can be
negative, even though the "surplus product" is 20 qrs., because the *value* of
the 100 qrs. of input, at the time they entered production, is greater than
the *value* of the 120 qrs. of output, at the time when they emerged.

This is the only way of making sense of this passage, and when you read it in
context, I think you'll agree. For additional context, let me refer to a
similar passage, in Vol. III, Ch. XLVII (Progress Pub. ed., p. 788), that I
posted about six months ago on this list: "We have previously shown that
although surplus-value is manifested in a surplus-product THE CONVERSE DOES
VALUE" (emphases added).

According to Ryu, the original text reads: "Wir haben schon frueher gezeigt,
dass, obgleich der Mehrwert sich in einem
MINUS VON WERT DARSTELLEN."(MEW, vol.25, p,796 : emphasis added).

Andrew Kliman