Joseph:
That's not exactly true. There is buying and selling between factories in
your system, but because it takes place as accounting entries kept track of
by a central authority, you think it isn't real buying and selling. But in
fact, since the economic accounts kept with these accounting entries
determine what happens to the factories, including whether they stay open, it
has the same economic effect as real buying and selling. You rename the
economic entities as "projects", and you merge the central bank with the
planning authorities, but real buying and selling exists.
------------------
Paul:
The key difference is that we assume that labour credits are non transferable
and can only be held by workers, not by any other institutions, and the credits
are cancelled when they are redeemed for goods from the state retail
network.

Since the labour credits do not circulate they are not money, and without
money circulating there would be no commodity circulation.

The projects we propose  are not able to hold credits, and have no account into 
which
credits are transfered when goods are produced.

Statistics would certainly need to be held on the amount of labour used by each
product, and the amount of produced inputs and other resources used by the
project to produce its output. On the basis of these statistics, the planning
agency would be able to decide whether it was advantageous to the economy as
a whole to continue the project, expand it, or close it down.

In a capitalist economy the results of sales and purchases determine the 
solvency
of firms and ultimately whether they expand or close. A socialist economy also 
has
to make analogous decisions about expanding and contracting activities.
It needs some form of cybernetic feedback.
But just because a cybernetic mechanism exists, and that one effect of buying
and selling in a capitalist economy is to provide a cybernetic mechanism, does
not make all cybernetic mechanisms the same as buying and selling.

We suggest that the basic control mechanism of a planned socialist economy
should, with respect to internal flows, be a system of in natura calculation.
This could either use Kantorovich's algorithm, the Simplex algorithm or some
interior point algorithm like our 'Harmony' algorithm.
But all of these depend on the existence of the 'plan ray', or a Leontief final
consumption function which the planners optimise subject to constraints.
The question then is how does the planned economy arrive at its plan ray.
We say that in par this can be done by political decisions, with the public
as a whole voting on it. I presented a paper with Karen Renaud last year
to the BCS conference showing how such voting could be arranged 
(http://eprints.gla.ac.uk/25753/1/25753.pdf).
But such votes can only handle the broadest outline of resource allocation
within the economy : how many working hours per week do you want to allocate
socially to education, health, defence, scientific research, providing for the 
retired etc.
If one is to arrive at a detailed consumer goods bundle, we think that 
information
on the purchases of different types of consumer goods is an essential source of 
information,
and it is vital that the mix of consumer goods production is adjusted in line 
with
actual purchases.
But this only requires a single agency able to perform sales - a state 
retailing network,
and a single class of agency able to make final purchases : the individual 
consumer.
------------------

 Jospeh:

In the latter part of your book, there is a "proposed communalist model"
which you and Cottrell believe to be a preferable form of socialism. This
model is supposed to overcome buying and selling and the possibility of
factories going bankrupt, thus eliminating "khozraschet", the self-financing
principle. But when one looks closely, it turns out that everything has
simply been renamed. There is still buying and selling, but it takes place
with different names.

In the system in your book, enterprises supposedly no longer exist, but there
are "projects". The "projects" don't engage in buying and selling, but
instead notify the "planning authorities" of how much of their "budget" goes
to labor, and how much "on power, equipment and maintenance". (The example
given on p. 182 is of a "local leisure center", so it has no raw materials.
Presumably, in general, one should add "raw materials" to this list.) The
planning authority then deducts from the project's budget the appropriate
sums for labor, power, equipment, etc.

The worker does not get paid directly from the project. The payment comes
from the planning authority. But the worker only receives this payment if the
project has notified the planning authority that the worker is employed
there.

And similarly, the system will presumably provide that the projects do not
get paid for their goods directly from other projects. But if the project
produces something, whether it is a material product or a service, it will
presumably be entered into the planning authority's books. (If it doesn't,
the whole system collapses.)

The planning authority then makes use of some "rational system of economic
calculation" to decide whether to close the project or allow it to continue.

So it turns out that buying and selling takes place, pretty much as it did in
the Soviet system, only money doesn't directly go to or from the project.
Instead the planning authorities keep the books for the project, just as a
bank might. The project's account is continually deducted from and added to,
the same way the financial balance of an enterprise is repeatedly deducted
from or added to as it buys and sells. The project will get closed if it
fails according to "a rational system of economic calculation" -- i.e. if it
goes bankrupt according to calculation by the economic authorities according
to labor values.

So the whole system runs as if there were buying and selling. But enterprises
are called projects. The financial balance is called the budget.  Money is
called an accounting entry in the books of the planning authority. And
bankruptcy is called closing the project due to rational calculation.
-----------------------------------
Paul:
The use of the word accounting is wrong here. There would be economic
calculation but not accounting. Accounting practice is built on the modelling
of mutual liabilities between agents and on the conservation of monetary
value, these rules are built into double entry book keeping.
If one takes either Dorfman or Kantorovich's algorithms or any of the
interior point methods they follow quite different algebraic laws and
abstract data types from those represented by double entry book keeping.

The units to which optimisation is applied in the case of Kantorovich's work
are techiques of production rather than firms or enterprises. If a project,
by which we mean a social grouping or collective, only uses one technique
of production, then the planning decisions to expand or contract that
technique may directly have the effect of expanding or contracting the
project, but it is likely that the project will have multiple possible 
techniques
of production which can be used - that was certainly Kantorovichs assumption
in the practical examples he gives from the plywood industry, earth moving,
etc. In this case the rational calculation says which techniques are the best
to use to a achieve a given plan.
----------------------
Joseph
> We are quite clear that labour certificates are just a means of distributing 
> consumer goods.

I'm sorry, but that's an evasion. The problem is that labor values, in your
system, aren't used only for labor certificates, but for the "rational system
of economic calculation" that determines whether factories stay open, and
that is key to the proposed mechanism of central planning

The labor certificates are measured in labor values. But they are not the
only thing measured in labor values. So are the budgets of the projects, and
the accounting entries which the planning authorities calculate with. So the
labor values are actually used to determine how goods flow between
enterprises, whether the enterprises go bankrupt, and so forth. It doesn't
matter if you call some things labor certificates, and other things
accounting entries. If they are both labor values, then the supposed labor
certificate is, in effect,  being used for transactions between factories.
Sometimes the labor value is denominated in a labor certificate; sometimes it
is just an entry in a budget. But labor value = labor certificate =
accounting entry.
----------------
Paul:
I think you are confusing two levels of decision making:
1) what should be the net social output of a given product - for this
    the ability to sell it to final consumers at labour values does play a
    decisive role, this sets the plan ray for final consumer goods.
2) The decision as to what techniques to use to produce this. Labour values
    can be a first approximation used locally to guide product engineers, but
    the final decision is made by the planning algorithms that are able to
    integrate in a synoptic way all the in-natura constraints. This is
   not reducible to monetary calculation.

> We also say that environmental constraints have to be imposed
> from the outside by politics and can not be encoded in a system
> of prices or valuations.
Joseph:
Thank you, Paul. It is an important point that a system of prices cannot take
account of environmental concerns, and I'm very glad that you have repeated
this point here. I wish it were more widely realized.

However, it  is one of the fatal defects of the system you propose that it is
unable to take account of environmental factors *economically*. This follows
from the fact that the labor value of the environment is zero. It means that
economically, without political correction, a system which relied on labor
values as a "rational system of economic calculation" would devastate the
environment, just as Western capitalism does, and just as the state-
capitalist Soviet Union did.

In my opinion, it is entirely wrong to say that economic calculation can't
take account of the environment. It's economic calculation based on a single
index that can't do so. That's not the only type of economic calculation. But
economic calculation based on a single index, whether the labor value or
money or whatever, will tend to reproduce the festishism of commodities and
various other ills of capitalism.


You think that you can around this by supplementing the system of labor
values with political decisions. But this implies that the environmental
decisions only affect a small section of the economy. In reality, these
concerns cut through the entire economy, affecting energy production,
agricultural production, the building of the infrastructure, etc. In fact,
environmental concerns affect every factory and every enterprise -- in the
choice of methods of production, as well as in designing their product. The
Australian naturalist Prof. Timothy Flannery was right in noting that
environmental regulation, once begun, would have to take in a wider and wider
field (but he was wrong in regarding that this would inevitably be an
oppressive "carbon dictatorship").
-------------------------------------
Paul:
This relates to the problem of optimisation in the presence of constraints.
In a captialist economy monetary return is the target of optimisation. In
Kantorovich's case it is distance along the plan ray. In our Harmony algorithm
it is roughly hyperbolic function derived from the partial derivative of each
product with respect to the plan ray. But in order to optimise you need some
objective function otherwise optimisation is undefined. The difference is that
any of the in-natura optimisation techniques allow for an arbitrary number of
constraints to be added :
for example
  co2 emissions + 43*ch4 emissons < 4 * 10^9 kilos
  total forested area > 3.4 * 10^5 sq km
 etc

These constraints themselves can only be set by essentially political decisions.

For efficiency an economy has to optimise, but for environmental survival
it has to be able to add an arbitrary number of additional constraints.
Linear optimisers are actually pretty good at solving these constraint
satisfaction problems - Pat Prosser has found that they outperform
many other constraint satisfaction techniques in some surprising areas.
------------
Joseph quoting me
> But this does not mean that you can get away without some index of
>  social cost to guide choices of technologies.
> We advocate that labour time be the basic mechanism for this,

Thus here the labor value does become the "basic mechanism", not just for
distribution of consumer goods, but for the direction of individual factores
and the for the direction of the  entire economy. Your system does not
restrict labor values to helping with the distribution of consumer goods.
They are the "index of social cost" which every factory and enterprise must
orient themselves towarfs.

> but say that any proposal by a particular plant or industry to
> use a particular technology be subject to the over-riding imperative of the 
> plan,
> since the planning system has additional
> information that can not be effectively encoded into prices or values.

Actually, if I understand it properly, in the planning system you describe,
every proposal by a particular plant or industry to do anything is "subject
to the over-riding imperative of the plan". It must go to the central
planning agency which will evaluate it according to a massive computer
program that tracks every single item produced in society. It is a really
rigid and top-heavy plan. It is not my idea of socialist central planning.

Present-day capitalism restricts environmental concerns to some political
decisions that try to go against the basic economic logic of the system. This
doesn't work very well, does it? We will have to fight, even under
capitalism, to have environmental planning and regulation, but this will
involve a constant fight against the basic tendencies of capitalism. Isn't it
a bit disconcerting that,  in a model of a planned socialist economy, there
is the same contradiction between the basic economic logic of the system, and
 environmental regulation?
--------------------------------
Paul:
I think this is an inherent conflict between local information and global
optima. A local agent in the system can never be presented with all
the complexity of the constraints required to globally optimise - though
of course with the advance of computing techniques they could now be
given 'black box' copies of the entire planning algorithm and database snapshot
and could try out for themselves the global implications of their local
decisions.


> The planning algorithms can run a constraint satisfaction
> system when selecting a feasible combination of technologies.
>These constraints can include environmental ones.
>I give an example of how to do this for different forms of power production
> ( hydro and wind) using Kantorovich's technique in (Cockshott, W.P.C.
>(2010) Von Mises, Kantorovich and in-natura calculation. Intervention :
>European Journal of Economics and Economic Policies, 7 (1). pp. 167-200.
>ISSN 1613-0960).
> A shortened account is given in http://mltoday.com/images/pdf/cockshott.pdf.

The Kantorovich system described there depends essentially on a system of
prices. It is not a system of "planning in kind", but depends crucially on
having a system of valuating things according to price or labor-value or some
financial index.  For example, on p. 14 of your paper "Mises, Kantorovich and
Socialist Calculation", you point out that "If a refrigerator designer was
deciding on what components to use in a planned new model, she would need
some way of telling which components would, from a social standpoint, have
been the most economical, which implies a system of valuations." So here you
refer to a system of valuations -- i.e. of pricing. You eve refer to a
controversy over which system of pricing should be used: should it be
"marginal cost" or "average cost pricing".

The system of equations used by Kantorovich required the use of such a system
of valuation. He himself referred to this, as you point out, as maximizing
profit, which, as you point out in your article, is the maximizing of value.
You try to brush this point aside by saying that a more capable planning
agency might tell factories exactly how much of each item to produce, rather
than telling them to produce exactly so much value. You write that this
concern with maximizing value "reflected the limited ability of GOSPLAN to
specify detailed targets in kind as described by [Nove (1983)]."

 But this is an evasion, because even if this planning agency decided these
detailed targets, if it decided them via Kantorovich's equations, it would
need to use a system of pricing. It would reproduce in its planning what the
factory would have decided in its quest to earn a profit.
-----------------------
Paul:
It is true that Kantorovich's algorithm used discriminating multipliers which
Western commentatiors call shadow prices, but these were just numbers
used within an algorithm, they do not necessarily involve any purchases
and sales. Furthermore, interior point algorithms replace these with a
more general potential function.

 ------------
Joseph:

A key point behind much of what you write is the belief that there is no real
economic calculation aside from calculating with a single index, preferably
labor values. Since you claim that empirically labor values are close to
prices, this amounts to a belief that financial calculation is eternal. And
it seems so reasonable, because we are accustomed to calculating things via a
single index in this society. Commodity fetishishm, evaluating everything as
an exchange value, is beaten into us by daily experience, by the necessity of
buying things in the marketplace, by the necessity to fight the capitalists
for a proper wage, by the necessity of basic financial planning for our
families. It becomes difficult to see what the alternative is. And we have
become used to overlooking the irrational results of financial calculation.

 --------------
Paul:
I dont not believe that calculation with monetary or labour values is
is more than a first approximation to optaining the optimal plan, but
optimisation implies having a dimension or vector along which to optimise. The 
reason
why calculation directly or indirectly in labour time has to be at least
very close in terms of hyperspace angles to any other optimum vector
is that labour is the basic resource available to humanity.
Nature constrains us, but it is our efforts that enable us
to achieve things.





_______________________________________________
pen-l mailing list
[email protected]
https://lists.csuchico.edu/mailman/listinfo/pen-l

The University of Glasgow, charity number SC004401
_______________________________________________
pen-l mailing list
[email protected]
https://lists.csuchico.edu/mailman/listinfo/pen-l

Reply via email to