[OPE-L:1036] Discussions on the labor Theory of Value

Duncan K Foley (dkf2@columbia.edu)
Sat, 10 Feb 1996 09:31:54 -0800

[ show plain text ]

While I've been on jury duty, I've been following the threads in OPE-L
but haven't had time to respond. There are a couple of points on a couple
of the issues raised I'd like to make. They are stimulated more by the
whole line of discussion than any particular contribution.

First, on the labor theory of value and the status of value and price in
Volume I of Capital. It seems to me that the key notion for Marx is that
newly produced value is the form living labor takes in a commodity
economy, and that surplus value (which includes profits, rents interest,
and other profit-type income flows) represents the labor exploited by
capitalists through the wage-labor relation. This newly produced value
can be measured or expressed either in money or in labor time (corrected
where necessary for differences in skill levels.)

In any ongoing economy there are also stocks of commodities produced in
the past. These are consumed more or less rapidly in the production of
new commodities and reproduction of the labor force. The effect of a
commodity economy is to impute value to these stocks. This imputed value
may change because of technical change that cheapens the commodities in
the stocks and thus lowers their value, or because of changes in the
value of money, but it is vital not to confuse these changes with the
production of value. When an existing stock of commodities is revalued no
new value is either created or destroyed. This changed imputed value
changes the claims agents believe they potentially can make on the flow
of new value, but not the flow of new value itself. The accounting
treatment of devaluation due to technical change should be analogous to
the destruction of part of the stock of commodities by fire, for example.

In any accounting system (whether it is based on money prices, or money
prices corrected for inflation, that is, the change in the money price of
use-values, or on labor-input coefficients) some account has to be taken
of these changes in the valuation of stocks of commodities over time.
Marx argued, correctly, in my opinion, that value-added accounting was
the appropriate framework in which to understand the production of new
value; we should deduct from the prices of commodities produced and sold
in a period the value of the input commodities used up in their
production, in order to calculate the value newly produced which we
associate with the expenditure of living labor. Capitalist accounting
practice does precisely this. Unfortunately this accounting is inherently
ambiguous, a fact which is reflected in the coexistence of various
conventions for valuing intermediate inputs drawn down from stocks or the
gradual exhaustion of the value of long-lived fixed capital: first-in,
first-out, last-in, first-out, straight-line and various types of
accelerated depreciation. I doubt that there is any scientific way to
determine which of these conventions is "right"; the important thing is
that we choose one or another in a given circumstance and stick with it
consistently. The actual magnitude of value added will vary with these
conventions (whether measured in labor time or money). I myself find it
much easier to resolve these issues by agreeing first what question one
wants the resulting accounting measure to answer.

When capitalists anticipate a devaluation of productive stocks due to
technical change, the effect is to lower the value imputed to the stocks
and to impose a capital loss on the capitalists holding them. If all
capitalists agree in anticipating this, they will take it into account in
deciding whether or not to invest in a sector, and compete accordingly.

I don't see that these points have a lot of bearing on the labor theory
of value as I stated it above. If we adopt the labor theory of value we
regard value added as the form living labor takes. Different accounting
definitions will lead to different concrete measures of value added and
as a result different values of money in the sense of the ratio of living
labor time to value added, but I'm not sure very many important political
economic questions are sensitive to these accounting differences.

There has been some discussion of the appropriateness of imputing labor
times to individual commodities through the v = vA + L method. Notice
that this method will not work in the presence of joint production, which
always leaves a range of possible imputed labor coefficients for
individual commodities. This is one of the reasons it seems to me more
robust to regard the labor theory of value as operating at the level of
the aggregate value added and the aggregate living labor time.

On this ground I think it makes most sense to view Volume I of Capital as
an attempt to work out the consequences of the labor theory of value at
the level of the aggregate commodity (or, if you prefer, the average
commodity). The results are supposed to hold good whether or not prices
are proportional to embodied labor coefficients (if they exist). For
essentially pedagogical reasons Marx often works through examples on the
assumption that prices are proportional to embodied labor times, but I
don't read the text as limiting the conclusions to that assumption.