Re: A very tired (despite the coffee) response

[ROSSERJB@xxxxxxxxxxxxxxxx id <01HPVJ5OHZ868ZGX91@xxxxxxxxxxxxxxxx>; Fri, 28 Apr 1995 14:43:14 -0500 (EST)]

Paul:
     Accept my apologies for accusing you in my
response to John Gelles of not responding.  That
went out before I saw your response.
     I congratulate you on making better forecasts
about the effects of dollar depreciation than did Fred
Bergston.  I have no idea what I would have said
under the circumstances.  I note that we have been
around the post before about your IMCU proposal, so
I shall not re-open that box, which I think is a
separate issue anyway (a specific proposal for a
specific problem rather than the general issue),
     Broadly I agree with you that one way to go
is to construct institutions which can snap in somehow
"when things go very wrong."  I note that this is not
too far from what David Colander was calling for in
NYC, although you did not seem terribly receptive then.
Of course sometimes things have to go very wrong before
anybody will be willing to do anything serious.
     With respect to exchange rates, I am cautious
about making any forecasts precisely because this is an
area where almost more than any other, complex dynamics
seem to be operative.  A lot of really weird stuff has
gone on.  As I said last summer, I do not have an answer.
I respect your IMCU proposal but fear it will not work,
for reasons I have stated before.
     Please do not identify Bergsten's "complex
econometric model" with complexity theory.  His model
did not work at least partly because, despite having lots
of equations, it was NOT really a complexity model.  One
of the complaints that a lot of us have had is that indeed
complexity theory can serve as a foundation for deriving
de facto non-ergodicity outcomes, although you insist that
non-ergodicity must be imposed as an axiom.  I agree with
your assessment of Bergsten's model, however, except for
your implicit labeling.
Barkley Rosser
James Madison University
PS:  Don't get TOO tired.  A good nap might help also :-).