Re: [OPE-L] price of production/supply price/value

From: Ian Wright (wrighti@ACM.ORG)
Date: Tue Feb 07 2006 - 13:33:23 EST

Hi Fred

I appreciate you taking the time to respond to my comments. I will try
to respond to your emails over the coming days.

> But my point is that the Sraffian framework in terms of physical
> quantities is entirely different from Marx's framework of the circulation
> of money capital.

I agree that the Sraffian framework omits the circulation of money-capital.

> The key difference is the different analytical frameworks, with
> different initial givens, as I have been emphasizing.  Marx's
> analytical framework is the circulation of money capital,
> and takes the initial M as given (along with the quantity of
> current labor and the MELT), and Sraffa's analytical framework
> is the production of commodities by means of use-values,
> with the physical quantities of inputs and outputs
> (or the technical coefficients) as given.

I agree that Marx and Sraffa have different starting points. For
example, it seems to me that for Marx the rate of exploitation and the
value rate of profit are the key distributional variables.

However, the neo-Ricardian critique (e.g. Steedman), rejects that
Marx's labour-value distributional variables can possibly explain the
corresponding price phenomena. That is one aspect of the
transformation problem (in its simultaneous form). Marx *thought* that
the rate of exploitation is a key distributional variable that
*explains* the price rate of profit, but the modern formulation of the
TP demonstrates that it *cannot*. (I should point out that I do *not*
agree with the neo-Ricardian critique -- but I do appreciate its
logical force).

> > Are you rejecting simultaneous determination, for example?
> Yes, I reject simultaneous determination.  That is, I argue that Marx's
> theory is not based on the simultaneous determination of input prices and
> output prices.  My rejection of simultaneous determination is not the same
> as TSS (see below), but they have been pioneers in calling simultaneous
> determination into question.

I agree that Marx's theory is irreducibly dynamic. For instance, his
distinction between constant and variable captial, in which variable
capital is precisely that part of capital that does not regularly and
systematically transfer value during the production process, is a
"dynamic" distinction, which is bound to be effaced in static
equilibrium models.

But, there still needs to be a good reason or explanation for why
Marx's labour-value accounting cannot determine the price rate of
profit (and other price phenomena) in Bortkiewicz's special case. We
shouldn't we expect Marx's claims to hold in this special case?

By rejecting simultaneous determination I think you are more
accurately developing Marx's theory, but you are not directly
addressing the modern form of the transformation problem. That is ok
-- it might not interest you, or you might think it unimportant etc.

> I argue that the analytical framework for Marx's theory is the circulation
> of money capital:
>         M - C ... P ... C' - M'
> M is equal to the price of the inputs, and is taken as given, as existing
> prior to production and the determination of the price of the outputs
> (M'), and is used to determine M' and dM.  This is the sequential
> determination of input prices and output prices, not simultaneous
> determination.

OK, by taking the price of inputs as given then it seems to me we are,
theoretically, in a dynamic world of the interaction between
labour-value and price over time.

> This interpretation does not mean that the means of production are always
> valued at their original "historical" price, as in some versions of TSS.
> If the price of a given means of production (e.g. cotton in the production
> of yarn) changes after it has been purchased in a prior circuit, then the
> price of the cotton that is taken as given in the determination of the
> price of the yarn is the NEW CURRENT price of the cotton, not the original
> historical price.  Even though the price of the yarn is the current price
> is its current price, the current price is still taken as given in the
> determination of the price of the yarn; it is not determined
> simultaneously with the price of the yarn.  Once means of production have
> been purchased at the new price, then the price of all other cotton still
> in the circulation of capital in the yarn industry (in the process of
> production or produced but not yet sold) will be revalued according to the
> new current price, and this new current price of the cotton is taken as
> given in the determination of the price of the yarn.

OK, I take this to mean that cotton has one price at any one time.

> Therefore, according to my interpretation of Marx's theory, input prices =
> output prices, even though they are not determined simultaneously (derived
> from given physical quantities).

OK, in your model, prices can change during the process of production.
Any consequent price re-evaluations are then immediately applied to
all commodities of the same type. No historical cost lags in

> Ian, some of your comments seem to suggest that input prices can = output
> prices ONLY if there is simultaneous determination of input prices and
> output prices.  But this is not true.

OK. But the difficulty is in precisely specifying dynamic change. I
would need to look in detail at how you are managing price
re-evaluation in a non-simultaneous framework.

> I think that part of the difficulty here is that linear production theory
> makes the totally unrealistic assumptions that all industries have the
> same turnover period, and furthermore that at the beginning of the period,
> all capital is in the form of input commodities, then all these inputs
> move together into the production process, and then all outputs are
> completed at the same time, and finally (and most importantly) all outputs
> are sold at the same time.  Sraffa called this method the "annual
> harvest".  This totally unrealistic "harvest" method is necessary if all
> prices are to be determined simultaneously, from given physical
> quantities.

Yes the Sraffian framework is in many ways unrealistic. (Although they
can deal with different turnover periods). But this doesn't imply that
your non-simultaneous framework is any more realistic. There is a
reason why people have made such simplifying assumptions -- it is
difficult to think about more complex, dynamic situations.

> Sraffa's question is:  how can exchange of all goods take
> place at the end of the period (after the "harvest"), so that the initial
> physical quantities can be exactly reproduced, and the same process can
> start all over again the next period.

Yes, but some Sraffians reject the self-replacing interpretation of
Sraffa's theory -- they view his surplus equations as being more
general than that. But overall I think this is fair comment.

> This unrealistic scenario puts the revaluation of the prices of the means
> of production into a misleading straightjacket.  It makes it look like the
> revaluation of the prices of means of production can take place only at
> "harvest time".  Either the "harvest time" of the current period, or, if
> simultaneous determination is rejected, then the "harvest time" of the
> previous period.  Revaluation of prices of the means of production cannot
> take place between the time the means of production are purchased and the
> output produced is sold, because it is assumed that no exchanges take
> place during this interval of time.

Paul mentioned that the equations can be interpreted in terms of
simultaneous flows. So it is more accurate to view the Sraffian
framework as assuming that production and circulation take place
simultaneously. Exchange, then, is taking place at the "same time" as
production. But certainly no re-evaluation of prices occurs -- at
least if we view the Sraffian system as describing a state of
self-replacing equilibrium. All you are really saying here is that the
standard Sraffian model does not take change into account. It is a
simpler case than the one that concerns you.

> But in reality, different industries have different turnover periods, and
> in most industries production and sales are continuous, occurring daily,
> not all together once a year.  Also in reality, capital is in all of its
> forms simultaneously - part of the capital is money capital at the
> beginning of the circuit, part of it is commodity capital as inputs, part
> is productive capital, and part of it is commodity capital as outputs.
> Therefore, if the price of a given means of production changes, due to
> changes in the industry that produces that means of production, then as
> soon as the means of production are purchased with the new price, then all
> similar means of production in use in other industries, and at any stage
> of the circuit of capital in these other industries, are revalued to equal
> the new price of the means of production.

Yes, the Sraffian model abstracts from many of these things. So what?
Bortkiewicz pointed out quite some time ago that if Marx's theory
doesn't hold in a simple, special case, then there is no reason to
think it will scale-up to more complex situations. (Unless there is a
good reason to think otherwise).

So Fred, I agree about the different starting points between Marx and
Sraffa, and I agree that the Sraffian framework is in many ways
unrealistic. But what have these points got to do with the
neo-Ricardian critique of Marx's value theory?


This archive was generated by hypermail 2.1.5 : Sun Feb 12 2006 - 00:00:02 EST