[OPE-L] the point of a dynamic model?

From: Jurriaan Bendien (adsl675281@TISCALI.NL)
Date: Sun Mar 25 2007 - 11:05:15 EDT

Hi Jerry,

Well, how do you model a chessgame (or football)? I don't exactly know how
they do it, but they do do it, and of course the game has one purpose only,
to win it, and there's a clear definition of winning. That makes it easier.
The chesspieces could be thought of as partially autonomous variables, to
the extent that they can move freely to some extent, although some moves are
ruled out, and each move incurs certain consequences.

The transformation of quantitative changes into qualitative changes can be
modelled I would think, through stipulating a rule that if a given
combination of quantitative relationships obtains, then a variable acquires
different properties or behavioural possibilities, just as e.g. possession
of a taken queen in chess can be re-obtained by moving a piece to the end of
the board on the opponent's side.

I agree that oversimplification is possible, but a model or theory is always
by nature a simplification. The question is whether you could integrate more
and more variables and information in the model, without destroying the
validity of the model's structure.

I don't necessarily agree that "It is the nature of the subject matter
itself that determines the level of complexity of the theory" because the
theory normally has a purpose, and if the purpose is satisfied, no
additional complexity is necessary. Reality is infinitely complex, and even
the simplest things can appear as highly complex if we study their
properties comprehensively.

The problem with Mandel's model is not so much, that it contains a large
number of variables which can all change with varying degrees of
independence of each other, but rather that it is not clear that their
interaction necessarily has to lead to any specific result, or at least
leads to one of a specifiable, limited number of possible outcomes. There is
no proof there, only a personal judgement based on experience.

Scientific statements always at least rule out some possibilities or
variants as a minimal condition, because otherwise there would be no
determinism at all, and no condition of fallibility; it usually involves a
statement of under what conditions a proposition would apply. This is true
even for stochastic models. But if the number of possibilities which aren't
ruled out remains very large, then the model has very little predictive
power, since all sorts of things could happen, within the realm of the
possible. It also has little explanatory power, insofar as many different
variables, or different configurations of variables, could be invoked to
explain a particular phenomenon. Therefore it cannot orient behaviour -
political or otherwise - much either.

Whereas Mandel's model might have great "heuristic" power, it is often not
clear why one sort of possible explanation should be preferred to another
possible sort of explanation. Mandel tells you that a phenomenon can be
explained, but he often doesn't tell you, or prove, why one explanation is
necessarily the correct one. The amount of data analysis that would be
necessary to prove it, often exceeds by far the data which he refers to.

Taking the example the average rate of profit on capital invested, Mandel
believed that it was the "synthetic indicator" of the performance of the
capitalist economy as a whole, rather than, say, sales volume or investment
ratios. But he frequently slipsides between regarding profitability as
explanandum (that which needs to be explained) and regarding it as explanans
(that which does the explaining), and beyond that, anyone knows that the
number of factors influencing profitability can be very great, since the
relationship between costs and revenues can be influenced by innumerable
conditions. Therefore, the level of profitability actually doesn't explain
very much at all - it describes, rather than explains.

Specifically, Mandel argues:

"It is correct that in the last analysis, capitalist crises of
overproduction result from a downslide of the average rate of profit. But
this does not represent a variant of the 'monocausal' explanation of crises.
It means that, under capitalism, the fluctuations of the average rate of
profit are in a sense the seismograph of what happens in the system as a
whole. So that formula just refers back to the sum-total of partially
independent variables, whose interplay causes the fluctuations of the
average rate of profit."

Since the "interplay of the sum-total of partially independent variables"
could take all sorts of forms, this implies that crises of
overproduction/overaccumulation are not necessarily the result of any
particular or recurrent causal sequence. So then at best the theory only
tells us what kinds of variables are candidates for an explanation, but it
does not tell us how they are specifically related, and how the chain of
causation should be understood. This is a heuristic theory, but not causal
theory. It tells you to what factors the explanation should refer, but not
what factors explain the phenomenon.

By contrast, the theories of Henryk Grossman, Louis Fraina and Anwar Shaikh
etc. are far simpler and monocausal. Aggregate economic growth is simply a
function of the relationship between the average rate of profit and the
total volume of profit from production. The rate of profit can gradually
decline more or less continuously (due to a steadily rising OCC) while the
volume of profit increases continuously (due to a larger output and sales
volume). At a certain point, however, the volume of profit stops growing
fast enough, causing a sharp fall in the rate of profit, sufficient to lower
investments in production absolutely, and causing rising unemployment.

If however the amount of constant (especially fixed) capital required to
obtain a certain volume of profit declines due to a technological revolution
(e.g. computers and information services), the amount of credit money
available is strongly expanded, and foreign trade in commodities, money and
capital strongly expands (globalisation), economic growth can nevertheless
continue despite a falling rate of profit in production. Moreover,
profitability may stay high, even although the additional profits are simply
not invested in industries. But that means the simpler theory might not
explain all that much either.


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