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

From: Anders Ekeland (anders.ekeland@ONLINE.NO)
Date: Sun Mar 18 2007 - 12:52:42 EDT

Hi Jerry

>The problem is analytical -- pertaining to the quantity of  variables --
>rather than simply a lack of data.  How could one, for instance,  construct
>a  formal dynamic model  of  the trade cycle in which there was:
>-- a changing but tendencially rising c/v,
>-- a changing but tendencially rising s/v,
>-- a  tendencially falling s/c+v,

The point with a dynamic model would be to show the that there is
counteracting tendencies to those "secular" rising tendencies. That
is why "the falling rate of profit" only can be seen as a tendency
that interplay with other tendencies, creating stagnation, corrections/crises.

My experience from simulation was precisely that the models where
much less stable than real existing capitalism, because everybody
reacted (herd behavior) to the same price signals, but the
heterogeneity of firms, the different/conflicting strategies makes
the system more stable.

>-- periodic changes in the intensity of labor,
>-- periodic depression of wages below their value,
>-- periodic cheapening of the elements of constant capital including
>-- the periodic release and tying-up of constant capital including
>-- moral depreciation,  including
>-- the forcible destruction of constant capital values,
>-- changing capital centralization and concentration,
>-- changes in the size of the relative surplus population and its relation
>       to changes in wages,
>-- changing terms of trade including unequal exchange,
>-- changes in the credit system and the growth of fictitious capital,
>-- how changes in state policy can impact the timing of the above,
>-- how class action by workers can affect the timing of the above,
>--  __________________  (feel free to add other important variables)?

>Such a model can not be constructed unless we make a lot of heroic
>assumptions, including assuming that a lot of  these variables are
>constant or their range of change can reasonably be stipulated in advance.

The choice of model, the constraint put on the variables etc. comes
from the analytical purpose. But I think a "Grand Model" modelling
the above is feasible - and would direct the efforts of economist
towards a real understanding of real-life mechanisms.

>But, that would in my view do an injustice to Marx's theory. And, more to
>the point, it would get in the way of studying  actual trade cycles.

I cannot follow you here. Why injustice, why get in the way?

>That's because the timing of a  particular cycle may be affected in ways
>which run against the assumptions regarding the variables and parameters
>of the model.

But to study such cases would be part of the modelling exercise.

>In general, I don't think that Marx liked to make empirical predictions
>about the  trade cycle because of the above.   His so-called  'predictions'
>were more along the lines of  asserting long-run tendencies. Some of  the
>Marxists, though, (Loren Goldner comes to mind) seem much more
>comfortable making predictions about crises and collapse.   Their success
>rate (i.e. the extent to which their predictions have materialized in the
>time  and manner asserted) has not been very good.  But, still they forge
>forward with their predictions ....

It follows very straight forwardly from dynamic non-linear theory
that it is very easy to get chaotic dynamics - like in weather
forecasting - a lot of data and measurement do not necessarily help
when you have non-linear systems - as one obviously have in the economy.

>With apologies to Dylan:
>           "How many times will Marxists make predictions
>                about 'impending' economic crises? [snip]

I share your skepticism towards crises predictions, but that is in my
case because I have come to emphasize more the creative aspect of
capitalism, than its destructive. But I am convinced that precisely
capitalisms "creative destruction" must be modelled by dynamic models
- not static ones. Static models just do not capture that contradiction.


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