Re: Prospects vs Forecasts.

["Dr. Bruce McFarling" <ecbm@xxxxxxxxxxxxxxxxxxx>]

Clifford Poirot <cpoirot@xxxxxxxxxxx> wrote,

on Thu, 22 Aug 2002 15:39:04 -0400,


[Quoth I]

> How can you assess prospect without the form of
> continuity that you seem to be throwing out the window
> as soon as true uncertainty is recognised.  If there is
> no correlation between current conditions and future conditions,
> there is no reason to invest in something based on its prospects.
> The lack of correlation would imply that a random guess would
> have as good (and as bad) result.


> My response:
> This is my problem with what I might call, the strong

> non-ergodicity hypothesis. I see a difference between
> uncertainty and the ability to use rules of thumb.

Now a new distinction enters the discussion ... the
"strong" non-ergodicity hypothesis -- againat what I
would supposed is the "normal" or "weak"
non-ergodicity hypothesis?

Maybe this is the problem.  Maybe when some people are
talking about the non-ergodicity of the economy, as a
critique of theories that REQUIRE ergodicity, then
others hear a "strong" non-ergodicity hypothesis.


> If one examines overall returns of the stock market

> for the last 70 or so years, one gets an average
> of about 11%. I have no guarantee that the next
> 70 years will be like the last 70 years, and probably

> some guarantees it will not be.


> However, barring environmental catastrophe, nuclear war,

> global meltdown into anarchy I have good reason to believe

> that the performance of capitalism a whole will be
> more or less what it has been over the last 70 years.

With the addition of another earth's worth of resources,
I am sure that it could be ... which might be available,
or might not, depending on how movement of production and
resource extraction into space proceeds or not.  Otherwise
environmental catastrophe would seem to be fairly certain,
and the inference would be that the performance of
capitalism as a whole will not be more or less what it has
been over the last 70 years.

If the system was ERGODIC, then there would be information
in the system now that would permit us to tell the
probabilities of either of those two directions for the
world economy.  Maybe we wouldn't have the capability
to extract that information, but the information would be
present, which would permit efficient market theory to
argue that the individual bits and pieces have been relied
upon by market participants when making their individual
decisions, and somehow the market equilibrium represents
a value that reflects that information.

BUT THAT INFORMATION IS NOT PRESENT.  It depends upon
conflicts between parties where in some cases the arena
of the conflict and the organisations through which
various will act have not even been constructed yet,
and the construction of those arenas of conflict and
collective parties to the conflict will depend on the
outcome of conflicts that are themselves non-ergodic.

It may be that you are confusing the strength with which
some people INSIST ON non-ergodicity with the strength
of non-ergodicity that they insist on.  Quite absurd
theories, for the non-ergodic world in which we OBVIOUSLY
reside, are developed and set forward as state of the art
theory in economics with high frequency.  The absurdity
is that they are theories that are only applicable to
ergodic systems.


> Thus I might reasonably guess an approximate 11% return

> on my stock portfolio over the next 25 years-knowing

> full well that there is no guarantee.


But in theories where ergodicity is assumed, then if the
basis for your guess is *effectively* sound, it will be
selected for by market outcomes, and if the basis for
your guess is *effectively* unsound, it will be selected
against ... whether you understand the reason for the
connection between the way you made your guess and the
actual outcomes or not.  This is Friedman's "pool-player"
argument.


> That said, this asessment is still better than random and

> rooted in intelligent assessments about how the economy

> works, how technological processes influence growth and so on.

But it is an assessment that recognises that the world is
NOT ergodic.  Therefore, it is also sensible to make
provision for contingencies to permit you to respond to
the impact of events that neither you nor anyone else
could conceivably have expected.

In an ERGODIC world, why hold precautionary reserves as
self-insurance against stochastic risk in the form of
money at all?  The only reason would be that you do not
have sufficient scale of holdings to make it worthwhile
to adequately diversify, and therefore there really
should be a mutual fund out there that gives you the
risk diversification that you need, and ALSO yields
a favourable return.

In a non-ergodic world, the shape of the pool table,
tilt of the table, weight of the balls, length of
cue and grip o fthe felt for English are all subject
to change between the time that the player practices
on the table and the time that the game is played,
which undermines the pool player analogy.


> Or as you put it:

> "Now obviously the further into the future you project,
> the less credence you give the projection.  At some point,
> you would give no credence whatsoever, and say, "I'll
> manage through to that point, and as it gets closer, I
> will take another look at conditions." "
>
> My response:
> but this is very different from random guessing. It allows

> me to distinguish between silly projections and sensible
> projections, just as I can distinguish between a sound
> business idea and a Ponzi scheme.

The observation that the economy is non-ergodic does not

mean an inability to reject foolish projections.  But it
does tell us that in addition to committing resources
on the basis of our projection, it is sensible to have
contingency reserves on hand to meet commitments if
projected income does not show up at the time or in
the amount expected.