Re: [MLUG - DISCUSSION] interesting quotations

["Jonathan King" <EMAIL:PROTECTED>]

On 10/2/07, Mike Miller <EMAIL:PROTECTED> wrote:
> On Tue, 2 Oct 2007, Jonathan King wrote:
>
> > On 10/2/07, Mike Miller <EMAIL:PROTECTED> wrote:
> >> Below are some interesting quotations that I found on this page:
> >>
> >> http://www.slate.com/id/2561/
> >>
> >> The author of that article, Herbert Stein, died in 1999.  In
> >> addition to many professional achievements, he was Ben
> >> Stein's father:
> >
> > The most famous of these quotations is, however, Stein's Law:
> >
> > # If something cannot go on forever, it will stop.
> >
> > The thing I like about this one is that it is obviously true, and
> > unexpectedly deep. So I find it intersting to see Stein's own take on
> > this aphorism (there's no rush to end something that is doomed to end
> > anyway), which is a deeply conservative notion. My take on it has
> > always been much less about whether one should intervene or not, and
> > more about _sustainability_, which I would argue is also a deeply
> > conservative notion, albeit one that has fewer fans among
> > conservatives these days.
> >
> > Really, if Herbert Stein had done nothing else in his life that was
> > worthwhile, I think he would (or should) be remembered for Stein's
> > Law, which is probably even more important than Little's Law, and
> > that's saying a lot.
>
> I didn't mention Stein's Law partly because it seemed trivial to me and
> not very practical.

I guess tastes differ. I think the simple and systematic application
of Stein's law would have prevented almost every financial bubble in
the last 20 years. Really, it's the ultimate boundary condition.

> If something cannot go on forever, we don't know that
> for certain but that is our guess, then it will stop eventually, but we
> don't know when.

But in some cases, it's easy to see what cannot go on forever, and we
can high confidence. And then the simple, trivial fact that it won't
go on forever at least gets you to ask the right questions, such as
"when might it stop", "what will happen when it stops", and "how might
I plan for it stopping". In any case:

> Communism "cannot go on forever" does that really
> mean that we should make no attempt to stop it because it will stop on its
> own eventually?

That's an interesting strawman (seriously; I think it's a good
example). A statement like communism "cannot go on forever" is a bit
vacuous until you can state it in a form that makes it more obvious
what you really mean, and which suggests a reasonable answer. I mean,
there's nothing to prevent the PRC from declaring whatever political
system they have going now as "communism", but that's probably not
what you mean. So let's get more specific. Suppose I say something
like:

North Korea cannot go on forever.

I  may or may not get any buy-in, until I point out why this is. North
Korea's per capita long term economic growth rate is approximately
zero, maybe a little bit less. South Korea's is, I believe, around 6%.
Fifty years ago, they were about even. By now, South Korea's economy
is at least 16 times larger than North Korea's, and maybe even more.
In another dozen years or so, it will likely be over 30 times larger.
At some point, the political North Korea truly will not be able to
resist this difference no matter what they do, and some people high up
in the power structure will decide that they would do much better if
owned or controlled even 5% of a nation with a 6% growth rate than the
whole of what they've got, which just sits there. In the long run, we
know what is going to happen to North Korea: it either becomes a lot
more like South Korea, or it ceases to exist as an independent state.
The only thing we have to do, then, is make sure that nobody does
something stupid in the mean time and gets half of the Korean
peninsula destroyed in a nuclear war. I mean, sure, it would be great
if we could give everybody in North Korea their freedom tomorrow. That
just isn't a reasonable planning scenario, however.

> If it is good for something that cannot go on forever to
> end soon, and we can bring a quicker end by intervening, then we should
> intervene.

Sure, but we should also be aware of the costs of intervention, etc.

> I didn't know Little's Law until you brought it up, but here is the
> Wikipedia entry:
>
> http://en.wikipedia.org/wiki/Little's_law
>
> I think they might have been wrong to add "over some time interval" in the
> definition because it seems to me that the Law should apply at any
> instant.

I agree that's clumsily phrased, but I think what they meant here is
that the system only has to be stable over some time interval, and it
*does* have to be stable.

> It is a mathematical statement from probability theory.  It
> looks like something I could actually use.  Why is Stein's Law more
> important than Little's Law?  I just don't see how Stein's Law can be
> used.

I might have been guilty of some hyperbole there. Little's Law really
is the bomb. You're at a restaurant that has 20 tables, you figure
people take an hour to eat and move on, and you're 20th in line. *If*
you joined the line at a truly random time and not (say) five minutes
after the place opened, you can estimate it will take you an hour to
be seated. You can also note that if your asumptions are true, a table
will clear out every 3 minutes. This is totally cool because it means
that if you've been waiting 15 minutes and nobody has left, it means
that either (a) people are either taking more than 5 hours to eat, or
(more likely) (b) the system wasn't stable when you got there.

That is just really cool, so maybe I'm overselling Stein's Law. But
Stein's Law is a powerful counter-belief to unbridled growth. In 2005,
housing prices in the DC area went up by like 24%, and a bunch of
people got loans predicated on the fact that real estate was making
double digit gains. But, in the absence of inflation, or gains in
wages, or even gains in population, it was pretty clear that this
could not keep going for ever, and maybe not for very long (although
there you're moving beyond Stein's Law). But a bank that had Stein's
Law in mind would not today be having big problems with foreclosures
due to people not being able to refinance.

Or so it seems to me.

jking

_______________________________________________
discussion mailing list
EMAIL:PROTECTED
http://mlug.missouri.edu/mailman/listinfo/discussion