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

From: Paul Cockshott (wpc@DCS.GLA.AC.UK)
Date: Fri Feb 03 2006 - 04:58:05 EST


Your argument is not clear and an example of how formal systems, here
that of probabilistic statistics, can 'sneak' in assumptions (something
you accused dialectics of not so long ago!). You make it look as if all
you have done is 'assumed' a MELT ["All that one is assuming here is
that there is some MELT that equates total values to total prices."]
but a MELT is not an assumption at all to anyone who agrees that both
prices and labour-times exist!
A MELT is just a conversion operator analogous to the constants
of physics which map between different dimensions.
Thus "c" maps between time and distance.
If one wants to discuss homomorphism between different 
domains - say values and prices, then one needs a mapping
operator that goes between them.
The only way I can interpret what you are saying lies in focusing on
what you originally called 'costs', and discussed as a 'price/costs
ratio'. I presume you have in mind that money costs (money measures of
constant capital, c, and variable capital, v) for any given firm are
close to being equal (directly proportional) to labour-time measures of
the same. Why? Because the money measure of 'c' is determined by
aggregation over the different means of production purchased by a firm.
The prce/value ratio is random, so deviations in it will begin to cancel
out when aggregating over a number of commodities. Hence the labour-time
measure of 'c' is likley to be close to the money measure of 'c', even
at firm level. A similar argumet can be applied to 'v' if we think of it
as determined by a basket of consumption goods.  When we consider that
costs are the sum of c + v then this further reinforces the point. This
would be enough to make sense of your argument because it leaves firm
level deviations in surplus value ('s', measured in labour-time) from
profit (price measure) as the main cause in deviations of price from
value. And, sure enough, in that case the coefficient of variation of
deviations of price from value can't be very large for any plausible
rate of exploitation since otherwise too many firms produce at too great
a loss.
That is exactly the argument.

The point about c being derived from a large number of different
commodities and hence the deviation of prices/values being small
for c, was originally made by Shaik in his contribution to the book
Freeman edited ( I think it was called something like Ricardo, Marx,
Sraffa ).
As stated, there are problems with the argument (relating to my earlier
remark about the aggregate equalities) which I won't elaborate since the
problems must mean that I have not grasped your argument properly.
Please enlighten me!
I am not sure where the problem with the aggregate equalities lies?

The aggregate equality of prices/values is a definitional assumption.
The theory's predictions would not be altered in any 
significant way if one assumed
that aggregate prices were exactly double aggregate values, or 
Pi times aggregate values. What is important is the the correlation
between the two, or perhaps the mutual information between the two:
if I know the relative values of two commodities, how much does that
tell me about the relative prices?

This mutual information is preserved whether the mean price is the same
as or a scalar multiple of the mean value.
On the other hand I do not expect that, subject to total price=total
then total profit=total surplus value. I would expect them to be
related but only loosely.

When dealing with total surplus value, one has to ask how you will
estimate it.

One approach would be to use i/o tables and sum the row corresponding
to Other Value Added which is close to profits, but to get total
surplus value one has then to add the rows corresponding to interest tax
and some unproductive expenditures. At the end of this one would have
a monetary sum corresponding to the broad definition of surplus value.
Using the MELT one could convert this into an anticipated amount of

However this is working backwards from monetary accruals, it is not a 
direct computation of value flows. To get the surplus value directly
in labour one would have to look at the columns to the right of the
table showing things like gross fixed investment, net exports, 
government expenditure etc. These columns vectors have elements
corresponding to the gross fixed investment that takes the form of
all the major industry sectors : how much of the fixed investment is
in things like
 New construction                                                    
 Maintenance and repair construction                                 
 Ordnance and accessories                                            
 Food and kindred products           
   Since we can work out the price/labour ratios for the outputs
of each of these industries, we can impute a figure for the labour
required to produce net fixed investment for the whole economy.
Similarly for govt expenditure, net exports etc.

However there is a big hole in the available data in that there
is no separate breakdown of personal expenditure on a class basis.
It all appears in one column. SO we don't know  form the 
personal expenditure of capitalists takes, ie, its distribution
across different commodities. Without knowing this we do not
know the total labour embodied in the commodities they personally

This makes it hard to estimate the actual deviation of total surplus
value from the total of profits, interest etc, that one can
get from summing along the lower columns of the table.


        -----Original Message----- 
        From: OPE-L on behalf of Paul Cockshott 
        Sent: Wed 01/02/2006 16:03 
        Subject: Re: [OPE-L] price of production/supply price/value

        Hi Paul
        You write:
        "I was making very parsimonious assumptions in my post:
        a) Assume that the selling prices of firms are a random function
of the
        value of their products."
        You seem to actually be specifying this function such that
prices are
        proportional to values with a random disturbance (i.e. prices
        around values with zero mean fluctuation). If so then you are
        the famous aggregate equalities hold (with random disturbance).
        isn't this assuming just what is at issue?
        All that one is assuming here is that there is some MELT that
        total values to total prices. But this is a necessity in any
        the two are in principle in different units.
        One could similarly set total actual prices to total of
         prices of production and see what the dispersion of prices of
        around the mean would be.
        What I am concerned with is the constraints that reproduction
        on the dispersion of the price/value ratio as a random variable.

This archive was generated by hypermail 2.1.5 : Sat Feb 04 2006 - 00:00:01 EST