Re: poly: Modeling Economic Singularities

From: Perry E. Metzger <>
Date: Mon Apr 13 1998 - 10:39:08 PDT

Robin Hanson writes:
> As I've told you before, I'm open to questioning specific points, but
> I can't effectively respond to unfocused requests to justify everything.

Hey, you made an extraordinary claim: that a short set of very well
behaved (largely linear, as I recall) equations could tell us
considerable useful information about the state of the economy into
the far future. Extraordinary claims require extraordinary evidence.

You may say "geesh, Metzger, you're being a jerk. This stuff is believed
by tens of thousands of my academic peers."

Well, big deal. Millions of people believe in god, too. Doctors around
the world used to believe in bleeding patients. Doctors only a
generation ago believed in routine tonsilectomies. Who cares what
people believe, or what your peers believe, or whether you've phrased
things in clean academic language.

Extraordinary claims require extraordinary evidence. What is the
evidence you have that your model produces any accurate predictions?
How do we know that it isn't just something you pulled from your
navel? Do you have any experiments that show it is accurate? Do you
have any tests we can conduct to falsify these equations?

Although I think they are *way* too extreme, I think the Austrian
economists had a point when they derided mathematical economic

For the most part, it has no predictive value. If you conduct
experiments with it, the falsifiable bits of the theory are rarely
confirmed by experiment or time. Econometrics is generally as
scientific as the study of chicken entrails or the bleeding of
patients. You are not advocating anything as silly as the econometric
"weather forecasts", but what you are doing is just as wrong.

You are telling us that a set of equations have predictive power. You
are asking us to believe that based on nothing more than you sitting
back and thinking about them. I would have thought that our experience
with Aristotle's Physics and five centuries of the scientific method
would have gotten people past this sort of thing.

Robin, don't get me wrong. I think you're brilliant. In fact, based on
our conversations, I suspect that in terms of raw intelligence you're
far smarter than me, and you certainly are far more comfortable with
the math (which I don't use much any more).

However, that's not the point here. Whether you are smarter than me or
not, you still have to prove your points. You're asking us to accept
some conclusions you have made based on some equations you have that
have no actual testing behind them and which might (or might not) be
deeply flawed. Your conclusions might or might not be correct, but in
the sense of science (that is, in the sense of hyptothesize, TEST,
REPEAT TESTS), they have no basis.

Sure, I could dispute all sorts of individual portions, but when you
get right down to it, the world I know is highly nonlinear and highly
resistant to certain classes of prediction. I'm going to need to be
convinced of the truth of your statements, and frankly, I find the
comments on your part that "all this is well understood in the
literature and I'm above teaching the ignorant" to be rather akin to
the comments of those who would cite the authority of the ancients on
the purpose of the various biles and humors.

Received on Mon Apr 13 17:40:24 1998

This archive was generated by hypermail 2.1.8 : Tue Mar 07 2006 - 14:45:30 PST