[OPE-L] equilibrium and simultaneous vs. sequential determination

From: Jurriaan Bendien (adsl675281@TISCALI.NL)
Date: Sat Sep 15 2007 - 10:12:41 EDT

One thing I forgot to emphasise - the "ontological status" attached to
production prices (whether they are strictly theoretical (ideal) prices only
in a model, or whether they are in some way derived from actual prices
charged) obviously affects the question of what these prices can possibly

In the case of e.g. Anwar Shaikh or Makoto Itoh, the argument is clearly
that production prices are strictly theoretical prices, which do not exist
in reality. They function only to describe the movement of production
capital in its purest form, under simplifying assumptions, at quite some
distance from empirical reality. But the corrolary is that production prices
cannot explain actual empirical price levels, and that, if they are said to
perform a "regulative function" or describe "averages", this exists only
within the context of a theoretical model. Shaikh therefore moots the
concept of "regulating prices" which are prices distinct from production

In the case of e.g. Ernest Mandel, the production prices do have an
empirical reality, i.e. they can explain real price averages, at least by
specifying production constraints (an empirically given cost-structure for
inputs and an empirically given price-structure for outputs). The implicit
argument here is, that if production prices are only theoretical prices with
no bearing on empirical reality, then what is the point of the theory? The
theory is supposed to explain the facts of experience, and be disciplined by
observables. If it doesn't do that, there is no point in the theory.

If production prices are only theoretical prices with no reality, it would,
according to this argument, be analogous to saying that in order to explain
a moving object X, I have to assume that X is stationary, and that I cannot
explain X, if it is a moving object, only as a stationary object. But X is
in truth a moving object, not a stationary object, and therefore if I
abstract from its movement, the "explanation" ignores an essential feature
of what makes X what it is. In that case, we haven't really got any further
than perhaps specifying "some things which you would have to explain", if
you wanted to explain X.

I said previously "A theoretical price cannot really be a regulating price,
because it exists only in theory, and therefore it "determines" nothing
about actual prices".  Obviously the way some economists get around that
problem, is to postulate that actual prices will trend across time towards
the equilibrium price, and therefore, the equilibrium price-level can be
said to exercise a real regulative function, because actual prices trend
towards it. This is just another way of saying that prices, "being what they
are", have the tendency to approximate the equilibrium price more closely
across time, or that markets tend to move towards equilibrium.

The question is then, how you know that, and how you could prove it. And
basically you can prove it only by (1) devising a theory of the equilibrium
price, (2) predict what the equilibrium price-level will be, and then (3)
test if actual prices do tend to approximate it, across time. If it is found
that actual prices do not behave in a way that equilibrium theory predicts,
however, then you always have three possibilities: (a) the theory is wrong,
(b) the actual price movement is an exceptional case to the rule, or (3)
there is some problem in the methodology of measurement. And you may not be
able to establish which it is.

But such tests are anyway very rare, in part because of the difficulty of
specifying what the equilibrium price-level would be (the choice of
assumptions and what entitles you to those assumptions). Instead, mostly
what economists do, is simply to extrapolate future prices from past prices,
an extrapolation which usually has very little to do with equilibrium
theory. It is just established empirically, that if certain prices go up
some other prices go up, if certain prices go down, other prices go down
etc. and then you can infer stochastically that a certain constellation of
price-levels will tend to move in a certain direction, even quite
independently of any proof in causal terms of why they happen to move in
that direction.

Given that therefore equilibrium theory has little explanatory power, beyond
proving that an equilibrium price level "could" exist, it may be regarded as
odd that Marxists should devote so much importance to it.

Although this is often overlooked in the literature, Marx himself in his
abstract theory of capital distributions never talked about "inputs" and
"outputs", he talked about a quantity of capital which is transformed into a
larger quantity of capital, in abstraction of the specific assets which that
quantity of capital represents (except C, V and S, and a broad distinction
between three output-defined sectors) and in abstraction of the moment of
purchase or sale.

What is the difference? Inputs are defined in terms of a volume of purchases
during an interval, and outputs are defined in terms of a volume of sales
during an interval, and the net output or value added is defined in terms of
these costs less sales, by some accounting procedure involving grossing and
netting. The point however that Marx makes, is that the "value" of capital
advanced, capital consumed and capital earnings realised cannot be simply
expressed in those terms, from a "macro-economic" point of view.

Why? Because from that point of view, capital assets are in reality
constantly being bought, used and sold above or below their current socially
established value ("what they are really worth"). That is precisely why Marx
abstracts in the way that he does. If assets are worth simply what they
currently sell for (price=value), there is zero point in this procedure at

What is the real purpose of such a theory, in which capital values and
capital prices constantly deviate?

I don't really think the reason is to show "how the production system
reaches equilibrium, using an accounting theory". I think the reason is to
explain the real economic behaviour of the agents of production, and the
dynamics of the movement of capitals, i.e. to reveal the way production
activities will be adjusted and developed according to the movements and
valuations of capital, given that the production of output is conditional on
the accumulation of capital, and the competition over that capital. You
don't even need price-value deviations to devise an equilibrium theory, just
inputs,outputs and prices will do.

In the real world, however, management and capitalists do in fact use models
making similar assumptions to Marx's models. Group controller Gerard
Ruizendaal of Royal Philips Electronics said for instance that "The main
idea is to improve our economic value-added every year so our return of
capital is more than our cost of capital." (cited in "The value creation
equation", Corporate Finance Magazine, March 2004). The main idea does not
describe all they do, but it is the "main idea".

Why is value theory brought into the discussion by these managers and
capitalists? Ultimately, because you cannot even talk about price aggregates
and price relations which constantly change without making reference to
values. All accounting presupposes notions of value conserved, value
destroyed, value used up, value added, and comparable value, without which
you cannot analyse transactions at all. Beyond that, the accounting
information is imperfect, and we still need to theorize about what it means,
i.e. we have to interpret it, to know what to do.

We can dispute about how these values are in fact formed, or how they
specifically exist, but that you need them and do assume them, becomes
perfectly obvious when you examine what people actually do, when they "work
with prices", i.e. the phenomenology of prices.

(As I am now at last on holiday I am taking a break from OPE-L)


This archive was generated by hypermail 2.1.5 : Sun Sep 30 2007 - 00:00:05 EDT