Re: Interesting Take on Modeling

[Jim Devine <jdevine03@xxxxxxxxx>]

the two views:

1) there are _no_ limits to modeling as a strategy for understanding,
explaining, and/or predicting phenomena in the real world.

2) there are _a lot of_ limits to this strategy, but it might have
some positive effects in clarifying our understanding of the real
world.

On 2/22/07, Tom wrote:
> ... Even
> agreeing with Jim's characterization that most serious economists
> would take the second view (if confronted directly with the question),
> they often operate within a discursive frame that tacitly accepts (and
> imposes) the second view.

> Three examples I'll mention are the
> aftermath of the Cambridge capital controversy, the posthumous
> flourishing of methodological individualism and the canonical
> income/leisure labor supply model. Each of these are examples of
> "loser wins" -- the clearly discredited argument has nevertheless
> prevailed in the discipline because the implications of the other
> would be catastrophic for model building.

on the second & third: I agree, though perhaps out of ignorance.

But the Cambridge Capital Controversy is a completely different
animal. The English Cantabrigians (Sraffa et al.) had a _better_ model
of aggregate production than did the US ones (Solow, Swann, et al.)
They were based on fewer unreasonable assumptions (e.g., that the
organic composition of capital or some other measure of technology is
the same in all industries).

The problem was not the ideological importance of models. Rather, it
was the dominance of neoclassical ideology. It's not that the English
models weren't good for promoting model-building. Rather, they weren't
directly useful for the kind of predictions and policy prescriptions
that the neoclassicals wanted to do, along with undermining the view
that the "factors of production" are rewarded according to their
contribution.
--
Jim Devine / "The truth is more important than the facts." -- Frank Lloyd Wright