On Mon, 10 Jun 1996, Paul Cockshott wrote:
(among other things)
> >
> >> Paul C:
> >> To say that individual commodites do not sell exactly at their values
> >> is not the same thing as saying that commodities in aggregate sell
> >> at their value. The latter is a quite vacuous statement.
> >Duncan F:
> >I think it is a tautology, but like many tautologies, such as f = ma, I
> >don't think it is vacuous, since it plays a critical conceptual role in
> >the framework of the historical materialist theory of capitalism.
>
> Paul C:
> The relationship between force mass and acceleration is not one of
> tautology but one of mutual definition of the concepts. The concept
> of labour value is not defined in the process of saying that 'in aggregate'
> commodities exchange at their values. It is defined independently of
> and prior to exchange taking place.
There are two different approaches here. As in most scientific debates, it
is probably very difficult for us to settle this from an appeal to first
principles. We might have a discussion about which side of this divide
Marx actually came down on, but we know from our earlier experience on the
list and in the literature that this discussion often proves very
difficult to resolve. (If you're interested in my side of it, I tried to
put the case concisely in my book Understanding Capital.)
What intrigues me, however, is that despite really quite different
philosophical interpretations of the LTV, we converge in the implications
of our positions for actual empirical work. My concern was to return the
LTV to the position of an operational theory that could analyze capitalist
social relations and economic development at the phenomenal level of the
national income and asset accounts. The difficulty as I saw it in the '70s
was that the embodied labor coefficients interpretation either prevented
Marxists from saying anything directly about the evolution of the rate of
exploitation and so forth, or saying it only indirectly by recasting the
national income accounts in embodied labor coefficients prices (as, for
example, Ed Wolff did in his early papers on the rate of exploitation in
Puerto Rico). As a result, the historical materialist point of view,
despite what I took to be a rich set of insights into the actual progress
of capitalist social relations, put itself into a marginal position in the
ongoing scholarly debate.
Since your position is that embodied labor coefficients are very highly
correlated with market prices, and that is what justifies reference to
them in empirical work, it seems that you come out operationally pretty
much where I do: that we can, either exactly, in my view, or to a very
good approximation, in your view, use market data to measure the critical
Marxian parameters.
>
>
> Duncan:
> >I don't think Marx had any empirical basis for judging the actual
> >"signal/noise" ratio between values and price. But suppose for a moment
> >that changes in technology greatly diminished the correlation between
> >embodied labor coefficients and price. Would this lead us to reject the
> >labor theory of value in the sense of, for example, regarding surplus
> >value as an expression of unpaid labor?
> >
> >>
> >> Saying that commodities in aggregate exchange at their values is vacuous
> >> since it would still be true even if there was a zero correlation
> >> between prices and values.
> >
> >Or a negative one, I suppose. But isn't that exactly what we want in an
> >interpretation of the labor theory of value? Marx saw the idea that labor
> >is the substance of value as the way to link exploitation in capitalist
> >societies with exploitation in pre-capitalist societies. This idea should
> >not be conditional on any particular form of competition or configuration
> >of market prices, I think.
> >
> Paul C:
> He and the classical political economists may not have had access to
> modern statistical data, and were thus unable to give precise correlations,
> but from their knowledge of prices and working practices they certainly
> had a basis for making the judgement that prices were almost entirely
> determined by labour contents.
>
> If it were the case that in fact they were totally wrong, and there was
> in fact no connection between prices and labour contents in the real
> world, then we must unequivocally reject the Marxian theory of surplus
> value as an explanation of the origins of profit.
I wonder how many participants in the list would agree with this position
philosophically. I myself don't. It seems to me to be onesidedly
positivist.
>
> Duncan F:
> > From a
> >purely formal point of view there isn't any argument against taking energy
> >or land as the substance of value, as John Roemer argued at length. The
> >stipulation that we regard value added as the expression of living labor
> >time links the reading of capitalist value relations to the underlying
> >human relation of exploitation. I don't think it really makes sense to try
> >to ground this point of view purely in empirical correlations.
> >
>
> Paul C:
> Your disdain for the real world verges on the platonic, but that is notoriously
> the characteristic vice of theoretical economists.
Oh, well, win some lose some.
> Roemer may have argued
> at length that energy etc are equally valid as substances of value in some
> hypothetical universe. However, in this one, it is labour that determines
> exchange value not energy. The empirical evidence for this is very robust.
> I enclose a short passage from an article due to appear in the
> Cambridge Journal of Economics on this:
I read this paper with great interest, and would like to underline my
feeling that these empirical findings are important, even if I don't
regard them as relevant to the foundations of the LTV.
(I omit Paul's quote from his article here.)
>
> >> Paul C:
> >> The analysis of the production of surplus value is contingent upon there
> >> being a positive correlation between prices and values.
> >
> >Here we come to the nub of the issue, I think. Are you really content to
> >ground the analysis of surplus value on an empirical correlation that
> >might very well have turned out different, and as history progresses, may
> >diminish?
> >
>
> Paul C:
> If prices were not correlated with values then we would be living
> in a different world, one to which Marx's theories did not apply.
> Should the passage of time lead to a situation where prices were
> negatively correlated with values, something you seem happy to
> contemplate, then it would pay capitalists to reduce surplus labour
> time to a minimum, as each additional hour worked would reduce
> their proceeds.
I don't think this follows, for two reasons. First, the aggregate surplus
value still depends on the aggregate unpaid labor time, so that it would
certainly not pay capitalists in the aggregate to reduce unpaid labor time
to a minimum. Second, the correlation you find is not a direct reflection
of the decisions of individual capitalists, but reflects the operation of
broad competitive forces. An individual capitalist, taking market prices
as given, still has a strong incentive to maximize unpaid labor time.
Yours,
Duncan