Critique of Atomism

[Colin Danby <danby@xxxxxxxxxxxxxxx>]

Winslow's contribution, bringing up Keynes' critique of atomism, strikes me as
very useful in unravelling the differences over chaos theory/nonlinear models
and Post Keynesian Theory.  The issues come down to the ontology of the
economic world, and the relation of models to it.

The most important issue is ontology.  What JMK is getting at in his critique
of atomism, I would argue, is the embeddedness of an economy in a social and
historical world.  An obvious example: how wages change depends on unions'
strength which depends on their political fortunes, and cross-cutting social
movements and political forces.  A purely atomistic model of the labor market,
specifying employers and workers and setting them loose to interact,
necessarily misses these influences.  I would suggest that a reading of JMK's
theoretical work against his policy writings supports this interpretation.  To
cite the most obvious example again, downward rigidity in money wages in the
GT is clearly explained not as innate (let alone microfounded) but as a
product of contemporary political institutions.  JMK developed that argument
over a decade earlier in the course of debates over deflation to return to the
gold standard.

Given that feedbacks between the social, political, and "economic" are
inevitable even in the short run, macro modelling becomes a heuristic exercise
useful for policy discussion, but never a simulation of the macroeconomy with
genuine predictive value.  One can reach a similar position through Marx,
Kalecki, and/or some of the Austrians.

I belabor the above because it is so different from the views of uncertainty
and non-determinacy that have been put forward in a series of fascinating and
carefully reasoned posts by Fonseca, Chu, Rosser, and others.  What has
emerged is a view of uncertainty and/or nondeterminacy as finely-grained, if
it exists at all.  It is sometimes seen as a matter of individual
"consciousness," meaning a limited degree of indeterminacy inhering in the
individual's behavior.

The whole point of doing nonlinear modeling is NOT to generate chaos, but to
generate large-scale regularities of some kind despite the chaos.  Chaos
theory advocates appear to inhabit a world which is deterministic and
predictable in large changes -- no ice age next year -- but unpredictable
(though still deterministic) in small ones -- rain tomorrow.  Uncertainty is
tamed.

To return to Keynes, in the famous passage in his 1937 QJE article (after
stating that roulette is not subject to uncertainty and that *"Even the
weather is only moderately uncertain."*)  JMK says: "The sense in which I am
using the term is that in which the prospect for a European war is uncertain,
or the price of copper and the rate of interest twenty years hence, or the
obsolescence of a new invention, or the position of private wealth-holders in
the social system in the year 1970."  This is large-scale, historically-rooted
uncertainty, not indeterminacy resulting from small-scale randomness.

Colin Danby, U.Mass/Amherst danby@xxxxxxxxxxxxxxx