On 31 Dec 2014 00:57:14 -0400, Mike Spencer
wrote:
Ed Huntress writes:
On 30 Dec 2014 19:14:05 -0400, Mike Spencer
wrote:
Ed Huntress writes:
Now he has to spend the rest of the day running his
latest predictive econometric model on his company's cloud-networked
SPSS system. He has a master's degree in math and a degree in
economics, and does econometric research and analysis for a top
consulting firm.
Ask him what relevance, if any, Stuart Kauffman's Origins of Order
might have to to his research and analyses. And report back here
with his answer.
His reaction: "Stuart who?" g
He doesn't know about Kauffman.
Dang. Well, no shame in that. Kauffman isn't an economist. :-)
Just ask him if complexity catastrophe or complexity theory or
anything along those lines enters into his econometric models.
Not his, which are mostly business applications, but he says there is
a lot of academic research in applying complexity theory to
econometric models. He touched on it in grad school but he doesn't use
it.
Equilibrium models are the basis for business applications, and Real
Analysis is the highest-level math that's ordinarily called for.
Sometimes they're trying to find converging series' when building
models:
http://en.wikipedia.org/wiki/Real_analysis
It's too exotic for me. I never got that far in math.
I'm no math wizard but it strikes me much of the apparently baffling
impenetrability of economics and finance derives from the fact
(alleged, by me :-) that we've reached a level of complexity that
defeats statistics or model that don't account for it. And part of
the problem is that "models" probably *can't* account for it.
Well, they never could, perfectly. What they're seeking is models that
*work* to a useful or acceptable degree.
It's like statistical sampling, in which you're looking for
plus-and-minus values at a certain level of confidence. Or
engineering: you may never know when an aircraft wing will fail from
fatigue, but you can produce a good enough answer to keep your plane
from crashing.
--
Ed Huntress