[OPE-L:3253] Re: TSS and value added

John Ernst (ernst@nyc.pipeline.com)
Thu, 3 Oct 1996 08:02:48 -0700 (PDT)

[ show plain text ]

I think it is a good idea to repeat what Duncan
sees as the points of agreement between his view
and that of TSS. In that same post he raises a
darn good problem which deserves further comment.

Duncan wrote in OPE 3236 (snip):

Just for the record, let me go over my points of agreement and disagreement

with Andrew's examples. I agree that the LTV can be interpreted as the
statement that the price value added is proportional to the living labor
expended, the factor of proportionality being the "value of money" or,
equivalently, the "monetary expression of value". I also think it is
fruitful to begin by assuming a constant value of money for expository
purposes, despite the fact that in real capitalist life the value of money
is changing. I also agree that a constant value of money does not imply a
constant price level under conditions of technical change. I agree with
Andrew's basic method of determining prices on the hypothesis of an
unchanging value of money by equating a concept of price value added to the

living labor expended. What I question is the specific concept of price
value added that enters this equation, since it is not net of the IVA. The
effect of this in the context of the technical change scenarios that the
examples address is to produce a price path which is sharply divergent from

the path that would hold if price value added is defined net of IVA. This
seems important, because it addresses Alan's concerns about the
appropriateness of using stationary approximations in model situations
where the dynamic solution asymptotically diverges from the path of the
stationary solutions, which is what started the discussion in the first

Referring to Fred's comments, Duncan wrote:

Let me also record my agreement with your earlier posts that pointed out
how strange the price and profit rate paths in Andrew's examples are when
we interpret them in the light of real capitalist experience. It is hard to

believe that a capitalist economy experiencing steady labor productivity
increases and constant capital productivity would be suffering a
catastrophic and irreversible fall in the rate of profit to zero.

John responds:

If we see "Andrew's prices" as those used by capitalists to make
investment decisions, then there certainly is a problem here.
Indeed, Marx noted in Vol. 3 that "moral depreciation" could be
so great that investments might not be made even though at the
prices prevailing prior to that investment profitability would
increase. Clearly, the price path proposed by Andrew requires

We should also note that price path implicit in the alternative
view is not without problems as well. There we see even more
dramatic falls in the historic rate of profit.

Put simply, both views of our example need to be clearer about
what the capitalists see as prices and profits as they


From ope-l@anthrax.ecst.csuchico.edu Thu Oct 3 10:39:21 1996
Received: from anthrax.ecst.csuchico.edu (anthrax.ecst.csuchico.edu []) by mail.st.rim.or.jp (8.7.5/3.4W3-rim1.1) with ESMTP id CAA27384; Fri, 4 Oct 1996 02:47:49 +0900 (JST)
Received: from anthrax (localhost []) by anthrax.ecst.csuchico.edu (8.7.5/8.7.3) with SMTP id KAA22122; Thu, 3 Oct 1996 10:39:21 -0700 (PDT)
Date: Thu, 3 Oct 1996 10:39:21 -0700 (PDT)
Message-Id: <UPMAIL01.199610031714490152@msn.com>
Errors-To: glevy@acnet.pratt.edu
Reply-To: ope-l@anthrax.ecst.csuchico.edu
Originator: ope-l@anthrax.ecst.csuchico.edu
Sender: ope-l@anthrax.ecst.csuchico.edu
Precedence: bulk
From: "andrew kliman" <Andrew_Kliman@msn.com>
To: Multiple recipients of list <ope-l@anthrax.ecst.csuchico.edu>
Subject: [OPE-L:3254] Contradiction between Value and Use-Value
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
X-Comment: Outline on Political Economy
Status: RO

This replies to Bruce's latest two posts (ope-l's 3163, 3202), to Duncan's
ope-l 3236, and to the whole value added discussion. I hope to respond in the
near future to some aspects of Bruce's post not addressed below.

In ope-l 3236, Duncan wrote (in part):

"Let me also record my agreement with your [Fred's] earlier posts that pointed
how strange the price and profit rate paths in Andrew's examples are when we
interpret them in the light of real capitalist experience. It is hard to
believe that a capitalist economy experiencing steady labor productivity
increases and constant capital productivity would be suffering a catastrophic
and irreversible fall in the rate of profit to zero."

Although by "strange" Duncan at first appears to mean contrary to "real
capitalist experience," the next sentence doesn't seem to refer to empirical
evidence. Instead, "strange" now appears to mean "hard to believe," i.e.,
contrary to theoretical expectations.

I cheerfully acknowledge that, in this 2nd sense, my example is "strange" or,
as I called it, "extreme." It was purposely designed to yield results that
are counter-intuitive from the standpoint of Sraffian corn model thinking and
of what amounts in practice to the same thing, the simultaneist "labor theory
of value."

To what Alan Freeman likes to call the "na´ve" reader of _Capital_, however,
whose "reading is literal but not simple-minded" ("The Psychopathology of
Walrasian Marxism," p. 4, in _Marx and Non-equilibrium Economics_), the
results are not strange in the least, but intuitively obvious. Duncan
presents the scenario in use-value terms, which is fine. But to see how
obvious the results are to the na´ve reader, please put the use-value figures
aside, just for the moment, and consider the exact same scenario in value
terms, as measured in labor-time:

* Surplus-value is a stagnant 100 labor-hours.
* None of the surplus-value is used for capitalists' or workers' consumption.
* Therefore, all surplus-value is invested, along with the original
capital-value advanced.
* Therefore, the value of capital advanced increases (by 100 labor-hours) from
period to period.
* Therefore, the rate of profit falls monotonically and approaches 0.

Na´ve. Literal. And exactly in conformity with Marx's law of the falling
tendency of the profit rate.

I will also cheerfully acknowledge that this is not the last word on the
matter. As the value of the corn falls (in labor-time terms), the existing
capital becomes devalued. Marx argues that this leads to crises, which
manifest the devaluation of capital. So the fall in the profit rate is not
necessarily "catastrophic," nor is it "irreversible," because the foregoing
scenario doesn't take crises into account. Falling values raise the rate of
profit calculated on replacement costs, so that when enough old capital-value
is destroyed, the system can begin again with the replacement cost rate
momentarily reigning (this is a highly overstylized account, of course).

I am incapable of modeling the concrete dynamics of capital accumulation.
Without doing so, some simpler problems concerning capital devaluation can
nonetheless be addressed. For instance, proponents of simultaneous valuation
would have us believe that the "rate of profit" is a constant 25 0espite ---
or because of --- the devaluation of capital. But the devaluation of capital
doesn't raise the rate of self-expansion of capital, or, what is basically the
same, the internal rate of return on investment. In my example, the internal
rate of return, r, is determined by:

capital-value invested at time t = (value realized at time t+1)/(1+r[t,t+1])

The TSS calculations give

400 + 100*t = (500 + 100*t)/(1+r[t,t+1])


r[t,t+1] = 100/(400 + 100*t).

So, as t increases, the IRR falls and approaches 0%.

But isn't this just a figment of the temporalist imagination, expressed in an
incorrect measurement of value added and/or constant capital?

Well, let's see. According to the *simultaneist* calculations, again in
labor-time terms, the capital actually advanced at the beginning of every
period (except period 0) is 500 labor-hours. (This is what everyone but Bruce
seems to call the historical value of the capital.) Note that the capital
advanced has to be 500 labor-hours if Alan's condition is met --- the value
received by the seller equals the value paid by the buyer ---since the value
received by the sellers always totals 500 labor-hours.

The simultaneist IRR is thus determined by

500 = 500/(1+r[t,t+1])

so that

r[t,t+1] = 0%

except in period 0.

Thus, simultaneism hides the contradiction between value and use-value by
measuring the "profit rate" as a constant 25%, but it does not make the
contradiction disappear. According to SIMULTANEISM ITSELF, the economy does
not grow in (labor-time) value terms. The value of gross output is a stagnant
500 labor-hours throughout all time. The historical cost of the capital
advanced is a stagnant 500 labor-hours (except in period 0). Ah, but the
proponents of simultaneous valuation want capital to be measured at
replacement cost. Well, replacement cost measurement also yields a stagnant
value of capital advanced, 400 labor-hours. Thus, the rate of accumulation of
value is 0%, even though, as Bruce says, "everything available is being
accumulated in every period." And again, except in the initial period,
capitalists plough 500 labor-hours back into production at time t and get back
500 labor-hours at time t+1. I grant anyone the right to say that the "profit
rate" is a constant 25%, but not to say that this 25 0s measuring the
self-expansion of capital or the rate of return on investment. Simultaneous
valuation itself shows that this falls from 25 0n the initial period to 0%
forever after!

I thus disagree with Bruce when he writes:

"In this example, as near as I can tell, nothing much happens to the situation
of capitalists. Capital grows at a steady rate (25%), and since everything
available is being accumulated in every period, it makes sense to me to say
(as simultaneous calculation does) that capital earns its profit at that same
steady rate of 25%."

The amount of CORN planted surely grows at the steady rate of 25%, the amount
of CORN produced surely grows at the steady rate of 25%, and the excess of
CORN output to CORN input is always a steady 25%. I am happy to accept these
figures as "value" magnitudes in the context of the Dmitriev-Sraffa theory, a
theory in which corn, or the standard corn-modity, is the substance of value,
such that a unit of corn (numeraire) always has a constant value. But in the
context of Marx's value theory, in which labor-time is the substance of value,
and the value of corn (numeraire) is determined by the labor-time needed to
produce it, one simply cannot say that CAPITAL grows at a steady rate.
CAPITAL is a contradictory unity of use-value *and* value, and even according
to the simultaneist calculations themselves, nay, especially according to
them, there is no self-expansion of capital in VALUE terms.

So, no, I do not think Bruce has satisfactorily shown that his theory is "just
as internally consistent and consistent with the elementary reality of
exchange as is
your own [temporalist] approach taken on its own terms." Once it is
stipulated that the value received is equal to the value paid --- in terms of
the immanent measure of value, labor-time --- then simultaneism yields
internally contradictory results: the "rate of profit" is 25 0n every
period, everything available is accumulated, yet the value of capital fails to
self-expand and the rate of return on investment in value terms, not corn, is
zero. The results make sense if one confines one's attention to a given
period, but once one takes the whole expanse of time into account, the results
no longer imply what simultaneist doctrine wants them to imply.

This contradiction is, however, not a curiosum produced by simultaneist
miscalculation. It is present in somewhat different (and consistent) form in
the temporalist calculations as well, because the contradiction is a
contradiction within the commodity itself, within capital itself, the
contradiction between value and use-value.

One can recognize this contradiction and take it to its perhaps
counterintuitive conclusion, as I would submit that Marx and the TSS
interpretation have done. One can deny it, and come up with an internally
consistent theory in which use-value *is* value and the numeraire commodity is
the substance of value, as Walrasianism and Sraffianism have done, and reach
very different conclusions.

But one cannot mix-and-match the two theories and come up with anything
consistent. Value is determined either by labor-time or by the amount of corn
produced in a span of labor-time. Either an hour of labor embodied is the
same amount of value in all periods, or a bushel of corn has the same value in
all periods.

Bruce wants to say that, in terms of my example, an hour of labor embodied in
the last period has only 800f the value of an hour newly embodied in the
present period. Why might this be? BECAUSE IT ONLY PRODUCES 80% AS MUCH
CORN. So we have value being measured in labor-time, but labor-time being
revalued such that, when all is said and done, value is being measured in
corn. The corn calculations make perfect sense, as corn calculations. But
the interpretation of them as labor-time magnitudes does not. Bruce arrives
at what he considers internally consistent and meaningful labor-time figures
*within* each period, but *across* periods, the figures don't make sense.

The reason is that the unit of account, labor-time, is unstable. Bruce, and
the simultaneist "labor theory of value" in general, are measuring with a
rubber ruler. Generally accepted accounting principles, and indeed the very
concept of value (however measured) require a stable unit of account.

In essence, Bruce acknowledges this when he emphasizes that he is doing
"current" labor-time accounting. But any viable dynamics requires continuity,
not a discontinuous sequence of "current" moments that cannot be compared. It
is worth noting that Philip Mirowski, certainly no TSSer, recognized this
several years ago:

"The crystalized-labor method can construct a viable dynamics beased upon an
invariant unit because 'an hour is an hour is an hour'; the real-cost method,
devoid of explicit invariants, can only calculate a sequence of static
equilibria in which the labor-value unit is not comparable from one
calculation to the next" (_More Heat than Light_, 1989, p. 184).

(The "crystalized-labor method" is akin to, but not exactly the same as, my
own; the "real-cost method" is the simultaneist "labor theory of value.")

It is also worth recalling Marx's critique of Bailey. Bailey had written that
"Value is a relation between *contemporary* commodities ..." so that values of
different periods cannot be compared. Now, for Bruce, value is not a relation
between contemporary commodities but between contemporary labor-time and
contemporary commodites. However, Marx's critique of Bailey centers on the
allegedly contemporary character of value: "But what a fool he is! Is it not
a fact that, in the process of circulation or the process of reproduction of
capital, the value of one period is constantly compared with that of another
period, an operation upon which production itself is based?" [TSV III, p. 154,
emphasis in original quote of Bailey]

Andrew Kliman