Subject: Re: Metastable oscillators
From: "Stephen J. Turnbull" <stephen@xemacs.org>
Date: Mon, 03 Mar 2003 13:06:37 +0900

>>>>> "Benjamin" == Benjamin J Tilly <" <ben_tilly@operamail.com>> writes:

    Benjamin> Actually engineers deal with the notion of design
    Benjamin> tolerances all of the time.

Evidently mathematicians don't.  An 0.1 sigma interval means 99% of
the time you fall outside of it (normal distribution).  As long as
it's an _open_ interval, a mathematician will happily declare
stability, and go home.  The engineer still has a long night at the
office ahead, and will be giving you _dirty_ looks as you head for
"Miller Time".

    Benjamin> I see your point (about the problems of controlling an
    Benjamin> optimized process), but don't think that it is always
    Benjamin> that clear in practice.

Not always, but often enough in practice that people make the
artificial system vs natural system distinction that you objected to
by appealing to Le Chatelier.  I want to make the point that that
distinction is _intuitively_ justified.  Mathematics is supposed to
help people who want to do better than intuition, not confuse them.

    Benjamin> In that case then it is fine to assert the plausibility
    Benjamin> of one when waving your hands about really general
    Benjamin> stuff.

Not really.  Except in mathematics.

    Benjamin> I don't think that it is wrong in principle.  Public
    Benjamin> policy is based on the _need_ to have minimum
    Benjamin> quantities.

Are you kidding?  Public policy in agriculture is based on price
maintenance for the farmers, and occasionally for consumers.  Serious
famines don't, and won't, happen in the advanced market economies.
The U.S. is amazingly well-endowed for agriculture, but none of the
OECD countries needs worry about food self-sufficiency.  On the
contrary, when push comes to shove we regularly see price control
programs that guarantee shortfalls if not quickly relaxed or
counteractive policies introduced: food, rent, gasoline.

The proponents of intervention usually claim that if left to itself
the market will fail to provide sufficient quantities, but I've never
seen a case where that was anything but FUD.  True, markets do
sometimes fail to provide sufficient quantities in extreme cases, but
there are very simple, fairly cheap solutions to that problem (public
stockpiles targeted to consumer risk rather than producer profit, for
one) that don't involve distorting the market and covering up for the
incompetence of those demanding subsidy.

    Benjamin> I am aware of the situation in India right now.  That is
    Benjamin> only the second incident of famine in a democracy that I
    Benjamin> know of since WW II.  The other being a famine among the

You have perhaps heard of Bihar?  That's only the most famous of many.
Cause: politicians buying votes with low prices for bread and
circuses, leading to underproduction of bread.  The same thing is
happening in India today.  Several state governments got a bad case of
hubris over the success of the "Green Revolution" and decided they
could get away with tinkering with market pricing.  They got badly
screwed because the farmers are now more mobile than they realized,
and many tenants went to (or stayed in, having temporarily migrated
there) states where they got a better deal.  That's what a recent
Ph.D. candidate from India told us in his oral examination, anyway.
India is a democracy, rough and ready and so far acceptably stable,
but as an economy it is far more dirigiste than any OECD economy.

    Benjamin> inuit in Canada that was resolved by the drastic effort
    Benjamin> of directly subsidising the whole population.

The Inuit problem is extremely complex.  I'm not surprised it
sometimes manifests as famine, but AFAIK it wasn't a failure of
agricultural markets; the Inuit aren't farmers, I believe.

    Benjamin> I knew both things, hence my comment about higher order
    Benjamin> effects that kick in when inflation is near 0.

This time you were lacking precision.  You said nothing about near
zero.  Nor is it a "higher order effect".  The effect of deflation is
not the same as the negative of the effect of inflation, because of
the non-negativity constraint on nominal interest rates.  This is
first-order.

    Benjamin> (Deflation is merely a negative interest rate.)

No, it is not.  It is a general decrease in the price level, and
cannot lead to negative nominal interest rates.  Furthermore, although
real interest rates can be, and within recent memory have been,
negative, in deflation the real interest rate is necessarily positive.

    Benjamin> The problem [in the great Irish potato famine] was a
    Benjamin> blight that wiped out the potato crop several years
    Benjamin> running.

Uh-huh.  Monoculture is a famine waiting to happen.  In other words,
what happened was not a market failure, but a natural disaster.  What
you are implicitly advocating is simply a large income transfer from
people who were unaffected by the potato blight to those who were.  I
certainly agree that that would have been the nice thing to do, but I
fail to see that need as evidence for instability of the potato market.

    Benjamin> Famines like the one that happened in Ireland were
    Benjamin> regarded as inevitable.

They were--and are.  They still happen today in monoculture economies.

Just as catastrophes happen regularly to monoproduct companies.  The
difference between bankruptcy and famine is that bankruptcy is
deliberately engineered so that the irreplaceable resource---an
intelligent human being with the guts to bet her livelihood that she
can produce something others will find useful---is left unencumbered
to try it again.  In famine, of course, the human is debilitated or
worse.

    Benjamin> Also remember that this was the era which was hesitant
    Benjamin> to take action to guarantee clean water supplies because
    Benjamin> people should be free to decide whether they wanted to
    Benjamin> pay for that.  (A policy that suddenly reversed after
    Benjamin> Prince Albert died of typhoid...)

Right.  And today we know why clean water supplies are a public
problem in a way that the supply of potatoes is not.  The economics of
this were not at all well understood until the middle of the twentieth
century, I remind you.

    Benjamin> I am glad that he is not as smug as he comes across in
    Benjamin> his writing.

Oh, I didn't mean to imply that he's not smug enough that I
occasionally wished I had a Motie's gripping hand to do a little
facial reconstruction with.[1]  Just not about others' misfortunes,
and how he feels public policy should deal with them.  He has thought
carefully and generously about that, and come to the dismal conclusion
that to "think of it as evolution in action" is the _least harmful_
thing to do.[2]

    Benjamin> I disagree with him on the value of "doing
    Benjamin> good" if only for the personal psychological benefits...

I don't know, but I would not be surprised to find that Friedman made
substantial charitable contributions---and sat on the boards of those
charities.  In any case, that would be in no way contrary to his
philosophy.  Personally doing good is a very important aspect of human
behavior, and one that I value myself, despite having few useful ways
to model it economically.[3]

The "doing good" that is probably not a good idea is using "public
policy," ie, tax revenues exacted from other people, and restrictions
on their freedom to choose.  That not only surely harms those who are
taxed, but is of unclear long-run benefit to those who are subsidized.


Footnotes: 
[1]  He's a Chicago School economist, he can't help it.

[2]  Two in one paragraph.  I'll stop now; that's N-P complete.

[3]  The "why hackers do it" thesis I mentioned in my reply to Jean
Camp is one way, though rather unflavorful and less than satisfying.

-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.