Re: poly: Modeling Economic Singularities

Robin Hanson writes:
> Perry M. writes:
> >... I'm making the point that what you're doing is not based in the
> >methods of science.
>
> I am taking well-established well-tested equations and seeing what they
> imply in a new hard-to-test context. I claim this is standard practice
> in all science and engineering fields I am familiar with.

You shouldn't confuse SCIENCE and ENGINEERING.

In science, we take hypotheses and test them.

In engineering, we assume the truth of the science we know, and we
design something.

> For example, this is the essense of all engineering feasibility
> analyses.

You aren't doing an engineering feasibility analysis. You are making a
scientific claim.

Were you making an engineering feasibility analysis, btw, you would
insist on basing it on well demonstrated theories of science, and not
on the sort of unproven hypothesis you have given here.

> >Well, as an honest experimentalist in another field who, say, claimed
> >to have a model that predicted earthquakes, one might hold off on
> >announcing for a couple of years to document one's claim based on
> >future events.
>
> If one had theories that predicted earthquakes well in the U.S., one
> might reasonably look at what those theories predict for earthquakes in
> China. Sure there is no guarentee that China isn't different somehow,
> but one could easily publish such an analysis, and is would certainly
> be considered "science".

Indeed, one might.

You've done nothing like it, however. You haven't tried applying your
model to real data and showing if it works or not. This makes me
rather skeptical of your model.

> >You could also, conceivably, have tried to apply your model
> >retrospectively to reasonably accurate data in order to see if the
> >model at least fits past behavior.
>
> The equations I combined *have* been applied ad nausium to historical data,
> and to experiments.

Equations *similar* to the ones you have used have been *individually*
tested, some better than others.

> It is just that I don't have experimental results testing this exact
> combination of equations in the exact area I'm applying them to.

You don't have results applying this exact combination of equations in
any area, actually, if I'm not mistaken.

> >I hate to use vulgarity, but the claim that an economist can't be held
> >to the same standards as a medical researcher is pure bull.
>
> It seems to me that you are imposing different standards on economics than
> are usually applied elsewhere.

Hardly.

You haven't done one study on your model.

> >medical studies take twenty years. Five year studies are bloody well
> >routine. The scientists in question don't sit back and say "well, it
> >would take too long to determine if our model of heart disease is
> >correct, so we are going to just publish right now and pretend we are
> >exempt from the rules of science."
>
> But if that heart disease model is built from pieces that have strong
> support in other contexts, they *do* publish their analysis before that
> twenty year test is done.

They publish the TESTED portion of their analysis.

> The idea that they must never mention their theoretical model to
> anyone before the twenty year test is up is complete lunacy,

No one publishes theoretical models with as little backing as the soft
"sciences". I use "science" in quotes, as from what I can tell, social
"science" does not follow the conventional methods of science.

> >I've simply asked you for scientific evidence for your claims. ...
> >Sure, you are allowed to use a THEORY to exrapolate beyond your
> >current data. However, what we have here is a *hypothesis* -- a
> >potential theory. That is to say, you do not yet have a demonstrated
> >model -- you only have a model for which you have yet to gather
> >evidence of predictive power. ...
> >Simply noting that components of your hypothesis have some evidence
> >for them isn't sufficient. What if you completely ignored an important
> >factor? That is why people do EXPERIMENTS.
>
> I provided evidence, which you ignored.

So please tell us about the study you've conducted, and what it
showed.

If you haven't conducted the study on YOUR model, but someone else
has, that will also be fine. Tell us about what the results were, and
give us enough experimental information that we may attempt to repeat
the test.

> When a model is made of parts, evidence in support of a part is
> evidence in support of the whole, even when that evidence is from
> the context of some other whole.

Boy, you're resisting really hard.

You and your fellows claim that you possess excellent historical data
-- most of it goes back a good seventy years or more. (I frequently
claim that the economics community can't even measure the present
particularly well, let alone present solid retrospective data, but
that's another debate.)

You also claim that you have a mathematical model of economic growth.

It would seem a simple enough problem to apply the one to the
other. This would hardly be as solid as other sorts of experiments we
could conduct, but it would at least be a nice start.

Why are you resisting even trying this small bit?

> Yes, confidence in a model increases when it is more directly
> tested. But to insist that theorists be silent until there is
> evidence specific to a particular combination of parts is contrary
> to all standard practice.

I'm not insisting that you remain silent. I'm insisting that you quit
refering to what you have as anything more than an informed
speculation you came up with.

Standard practice is not relevant in any way, btw. That's another
logical fallacy. The notion that "everyone thinks this so it must be
true" is known as "Argumentum ad Numeram", btw. -- it is so old that
it appears in all the standard lists of logical fallacies.

> That standard would, for example, end the theoretical subspecialty
> of any field.

In physics, when people come up with a new hypothesis, they try to
publish suggestions on how to test said hypothesis. Are you even going
to *try* to have your model tested?

> >Now, as it turns out, I've spent quite a lot of time with people who
> >produce hypotheses about the future near-term price trends of various
> >securities and commodities. These individuals, in general, do not
> >attempt to risk money on their hypotheses until they have considerable
> >bodies of data backing them up. They collect time series data going
> >back a century or more, and carefully check their hypotheses against
> >historical data and make predictions and see whether or not these
> >predictions come true BEFORE risking capital on them.
>
> I'll bet they do not, however, always stay silent about their theories
> until tests back them up.

Actually, they generally remain silent even *after* tests back them
up, for obvious pecuniary reasons.

I am not, however, suggesting you be *silent*. I'm suggesting that you
and many of your colleagues produce reams of speculation, rarely if
ever test it, produce lovely papers based on towers of speculation,
and happily continue on, behaving as though you had no obligation to
even so much as label your speculations as speculations and not as
demonstrated theory.

> And do they always insist on years of experiments, rather than just
> tests on historical data, before risking capital?

Not necessarily years, but they generally require pretty hard
experimentation. Money people tend to be extremely tough on such
things.

Perry
Received on Tue May 5 00:15:51 1998