Re: [OPE-L] Dynamic model?

From: Anders Ekeland (anders.ekeland@ONLINE.NO)
Date: Mon Mar 12 2007 - 03:28:28 EDT


this is a vast topic - let me just throw in a few cents:

- first of all: the capitalist system is *never* in equilibrium. I am
deeply sceptical (as f.ex. the historian of Austrian Ec. thought
Rothbard) about the usefulness of studying equilibrium states,
"natural long run prices". This goes not only for Arrow-Debreu, but
also for Smith and Ricardo. Long run equilibrium prices is a
contradicto in adjectivo = impossible under capitalism.

- That the classics (Smith, Ricardo, Marx) had a fundamentally
different -  and dynamic model of capitalism is clear, but on can
discuss if Marx was successful in "proving" his dynamic mental model,
his main results in a static framework (= the transformation problem
as conventionally defined).

- Or to put it another way: Why is most economics obsessed by
studying perfect stagnation? (the real name of perfect competition or
long run equilibrium) - when it is growth (= increasing welfare) that
is clearly the nest interesting aspect of capitalism. Economics is
not about use of efficient and/or fair distribution of scarce
resources - it's all about how to overcome scarce ties - in the final
end - creating abundance. Accumulate, accumulate - M-C-M' that's
Moses and the prophets - not static equilibrium.

- The first conclusion is: No result from a static model ("Marxian",
Sraffian or neo-classical) can be accepted before it is shown that it
also holds in a dynamic setting, i.e. where producers can set/change
prices and technology. Despite serious attempts - the major results
(efficiency, Pareto-optimality etc. do *not* hold in a dynamic
setting, cf. Franklin M. Fisher "The disequilibrium foundations of
equilibrium economics", or Currie and Steedman "Wrestling with time".

- The fact is that perfect competition is no competition at all. The
most reasonable interpretation of Arrow-Debreu (or Sraffa) is a fully
monopolized (one producer of each good) in a total stagnant state.
That's certainly not a very attractive state, no "first best" at all.
Almost any non-equilibrium, labour productivity enhancing "fifth
best" is to be preferred to this state. (Cf. the Solow growth model,
without change in technology.)

- The fundamental point is that the same forces that drives towards
equalisation of profit rates - also creates disequilibrium. Firms are
forced to innovate in order to protect themselves against a situation
where they compete by producing homogenous goods with the same
technology to a given price. (Cf. Baumol's "The free market
innovation machine" 2002 for a non-Marxian (but Marx inspired)
description of this process.

 From the perspective of Marxian economics a dynamic model would be
able to handle:

- different and changing technologies - with increasing returns to
scale - where *demand* determines the SNLT

- where there are conflicting plans, not all plans get realised and
consequently learning, change of strategies

- where there is *waste*, that is labour time not being recognized as
socially necessary - modelled explicitly, that is bankruptcies etc.

In short - a dynamic model must not be an "Harmonilehre", that is a
model of imaginary harmonious state, but to show how the incomes of
the rich is actually based on the exploitation of other peoples work
- out of a non-existent equilibrium, to quantify the amount of waste
in order to show that a planned economy would be more rational and
efficient (and "just" and "fair").

A dynamic model would show *all* the transformations ("averaging")
from the private labour of each specific   instance of each specific
product until (labour)value is "frozen", socially accepted in the
form of a market price.

A dynamic  will be able to visualize the why the "simple" LTV model
is the best predictor of market prices and is true in this sense (=
grasping the essential), but that the simple LTV model is to simple
since it does not model technological change, does not model the
tension, the fight over the surplus.

It follows from this, that market prices are more real - and just as
important as "prices of production". Market prices have no inferior status.

A dynamic model would treat money - not as a commodity money - or a
commodity numeraire shares, but as the expression (a complex one) of
value. Money (and shares, bonds, debt, derivates, interest etc.)
would be seen as part of a "control" system that creates "order out
of disorder", i.e. socializes the individual labour, regulates the
conflicting claims to the goods and services produced (now and future).

Only in a dynamic model can you model business cycles, i.e. the macro
results of weakly coordinated micro behaviour.

A lot more could be said about this, but the important point is to
realize that static equilibrium models are very uninteresting for the
study of real capitalism. The "new solution" and even TSSI is still
too "static", or to "un-dialectic" - although on the right track. But
a Marxian model without technological change is not the final Marxian
model. Innovation = competition and competition = innovation, and
innovation is *the* most important aspect of capitalism.

 From a revolutionary point of view, the fact that capitalism
revolutionizes the productive forces more - and order of magnitude
more - than any know social mode of production - creates an enormous
potential for reformist policies - so a dynamic view of capitalism
also has implications for left-wing policies.

Neo-classical equilibrium theory is an ideology, not a scientific
theory. Sraffian and linear algebra Marxist models are very partial
models. Their main use has been ideological - against dynamic
economic theory. Economics will not become a science before
equilibrium states are seen as rather uninteresting special cases.
When that happens economics will (again) find it's place among the
other social sciences since they mainly are dynamic (that is have
learning, feed-back mechanisms, non-linearity) etc. The core of
"Economics imperialism" is the use of static equilibrium models.

When Jurriaan writes:

Wouldn't it be simpler to study the real history of capitalism using the
human brain as processor?

Of course, but I think that simulation has a role to play. But the
important thing is the kind of results you get. For example that a
compressed wage scale drives innovation drives welfare - and is more
efficient in creating growth than huge wage differences.

But the point here is to have a dynamic mental model to study a
highly dynamic system (capitalism). Let me end with an a quote from
Haavelmo on Keynes:

... In reality he [Keynes] came close to building a separate theory
for employment and investment, but since he in contrast to Wicksell
did not build an explicit dynamic theory, he was always in a
logically dangerous zone, and that created a lot of confusion.  (from
"The Keynesian revolution", 1977, Memorandum from the Institute of
Economics, Oslo, in Norwegian.)

... a lot of confusion is what characterizes economics since almost
99% of economists try to use static concepts to grasp a dynamic reality.

Anders Ekeland

At 00:00 11.03.2007, you wrote:
>What intrigues me is Ian Wright's and Anders Ekeland's references to
>"dynamic modelling" of the relationship between prices, values and
>labour-time worked.
>What would the dynamic model prove, exactly?
>If I read Ian correctly, the model would prove that, under the assumption of
>specified "purely capitalist conditions", a robust quantitative relationship
>exists such that not only will (i) market prices of products trend towards
>profit-equalising prices of production for those products, but also that
>(ii) those prices of production themselves are regulated, as Marx says, by
>current product-values which are (iii) constrained by, but not identical to,
>quantities of labour-time worked to produce those products (note 1: this is
>a different operation from the transformation of direct prices into prices
>of production; note 2: current product-values refer to current quantities of
>SNLT and not historical quantities of SNLT).
>The proof would take the form that, if the qualitative and quantitative
>assumptions of the model are accepted, then this logically entails the
>stated quantitative relationship.
>I think such a quantitative description is in principle possible to devise -
>for simplicity we need not even refer necessarily to values, but just talk
>only of prices and labour-hours - but even so the list of variables and
>relationships necessary to specify the purely capitalist conditions
>realistically is large. It exceeds by far anything we could empirically
>verify from available social statistics.
>Wouldn't it be simpler to study the real history of capitalism using the
>human brain as processor?
>The suggestion seems to be that the mentioned quantitative proof would make
>the theory more rigorous. But what if the pure model is only weakly
>reflected in verifiable reality? In that case, we would, in addition to the
>large number of variables we already have in the pure model, have to
>introduce all kinds of new qualifications to better approximate reality.
>But isn't it a scientific norm that the simplest theory with the greatest
>explanatory power is the best theory?
>The Arrow-Debreu type model states roughly that a set of prices can always
>be found at which supply and demand will balance, for all traded goods in
>the economy, if we define competition, demand and prices in a certain way.
>The objection to that model is that its assumptions are insufficiently
>realistic, and imply a near-tautology such that if all obstacles to trade
>are removed, one will be able to trade all goods. But a similar type of
>objection could be made against the suggested Marxian model of pure
>capitalism based on a general rate of profit.
>In Kozo Uno's view, it was sufficient to state the principles of political
>economy for pure capitalism, and develop more specific theories at lower
>levels of abstraction. The objection against this approach is that it is
>unclear how the pure theory abstracted from history is to be concretised
>consistently, and that Uno's pure capitalism is strangely a capitalism
>without any consumption or a political state. It is more a Weberian ideal
>type (Uno studied with Weber).
>What is preferable, a Weberian ideal type such as Marxists in reality use
>despite their odes to "dialectics" and suchlike, or a computational model
>based on counterfactual (oversimplified) assumptions? What kind of theory is
>really fruitful, and how is it best formed?

This archive was generated by hypermail 2.1.5 : Sat Mar 31 2007 - 01:00:12 EDT