Re: Incomes and Exchange rates

["John B. O'Donnell" <jackodonnell@xxxxxxxx>]

My apologies for the prior incomplete posting. I'm not much of a typist,
and I hit the wrong button and sent instead of saving.

Bruce R. McFarling wrote:
>
> On Fri, 12 Sep 1997 07:07:42 -0700, "John B. O'Donnell"
> <jackodonnell@xxxxxxxx> wrote:
> >Leigh Harkness wrote:
> >> ...
> >> You are still denying that it is possible to have
> >> international and domestic stability in the value of
> >> money.
>
> >No, I'm denying that a single variable can be controlled
> >to attain values proportional or identical to two
> >independent measures. It makes no difference if the
> >variable is the value of money or the direction of the
> >car you drive. You can no more drive the same car east
> >and west at the same time as you can control the value
> >of money to be the same as it was domestically in a
> >chosen base year and at the same time control it to be
> >the same in terms of another independent value standard.
>
> Isn't this assuming that these are mechanical systems that
> we are talking about?

No, but I am more familiar with control of physical systems / devices.
This does not imply or require that the things being controlled must be
physical. It is just as impossible to control an abstract variable to
meet two diverging measures as it is to control physical entities.

> Its been awhile, but it seems to that that classical systems
> control theory, or perhaps most control models, assumes that
> the systems under discussion are out of control if the
> control variables under discussion are left free.

Uncontrolled (free) variables are somewhat likely to not be controlled.
Want to try saying what you mean. [Preferably without starting a flame
war. ;-)]

> Certainly, looking at many *models* of economic systems
> and systems of economic systems would lead us to think of
> them as mechanical models.  But if there is a degree of
> homeostasis in the two systems, it is entirely conceivable
> that when a control has contradictory effects in the
> context of the two systems, one of the systems can
> successfully maintain stability in spite of the
> destabilising effects of the variable; further since the
> two systems are (by definition) in direct contact, it is
> possible that certain stable states of one system may
> actually aid in stabilising the other.

I gather that you are saying here that a single control can affect more
than one [i.e. - two systems.] variable. This is, of course true in both
physical and abstract control efforts. It is also why I mentioned this
possibility [The talk circuit bit of congress creating separate domestic
and foreign use currencies.] to Leigh. This does present the possibility
of using a second control [If one is available] or letting one of the
variables go free.

> How else (to pretend to more Ozzieness than I in fact
> posses) can one sheep dog move a flock of sheep? Get the
> flock going in the right direction, run over, get a stray
> member or three back with the flock (meanwhile giving
> the flock its head because it'll keep going more or less
> in the right direction), run back, get the movement of
> the flock corrected again, run over, get some more sheep
> with the flock, etc.

Aside from my unfamiliarity with Ozzieness [Is that Roo language?] what
you are describing is how a control may perform. It may only indirectly
affect the variable you want to control, but ultimately it is known to
cause the thing you want to happen happen. In your example it is the
behavior of sheep to follow a leader that allows the influence on one to
move the flock. In my example of the car it is even more indirect. Your
foot moves the position of a throttle, the change of throttle position
allows the air pressure to increase within the manifold which changes
the power output of the engine which causes the car to accelerate or
decelerate.

> Now, in a mechanical system, one control might not be
> able to maintain both the trim and the compactness of
> the flock.  But the sheep dog is not trying to move a
> pride of cats (which would require, as near as a I can
> figure, either a number of sheep dogs increasing as an
> exponential function of the number of cats, or
> knowledge by the cats that there is a friendly
> individual with an ample supply of bowls of cream in
> the direction they are heading -- and in either case,
> each cat would still pretend that is was entirely his
> or her own unique idea to head in that direction).  A
> flock of sheep is a self-maintaining system, and the
> sheep dog is simply reinforcing what is a natural
> tendency.

Yes, some things are easier to control than others. But, again, this is
_not_ a difference between concrete and abstract controls. Does a sheep
dog do precisely the same action to control trim as it does to control
compactness? Or, does it do one thing for trim and another for
compactness? [I might assume chasing the stray involves correcting
compactness and affecting the leader is for trim.]

> >...
> >> The reason appears to be that you consider that
> >> only variable you have to control either is the
> >> money supply.
>
> >Almost true. You might also be able to control the
> >opposite variable -- the supply of goods and services
> >available for exchange -- but why would one want to?
> >And, that does not alter the fact that whatever
> >control is used, the measure of a variable in a
> >particular dimension can only be one thing at a time.
> >And it doesn't matter if that dimension is direction,
> >mass, velocity or economic value.

> So I wonder whether John is inferring the contradiction
> between measures in the same dimension of a variable
> from an assumption on the nature of these systems
> that Leigh does not share.

As I gather it, Leigh is saying that a single variable [the value of
money] can be held concurrently to two standards -- a domestic measure
of value like the CPI and a second standard based on a measure of
exchange rates. I say that simply can not be done except in the event of
a serendipitous occurrence by chance, which by any reasonable definition
is not control.

> And, while I am not sure whether Leigh would describe
> the situation in these terms, it might not be a
> significant distinction.  Leigh is, after all, trying
> to work out models directly in the terms of the problem
> domain, so that if the systems in the problem domain
> are not mechanical systems, Leigh might adopt that as
> an explicit assumption in his models, or he might adopt
> it implicitly, simply due to the fact that the
> regularities that he is trying to account for are the
> regularities of non-mechanical systems.

As I state above, an assumption that the system is not mechanical or any
other physical form does not matter. Howsoever abstract it is, a single
variable _cannot_ be controlled to two standards of the same dimension
[Value, in this case.] at the same time.

> And, BTW, if someone either disputes or misunderstands
> what I am trying to say here, I *promise* I will not
> accuse them of not understanding 'plain english'!  Not
> only would I consider that rude to list participants
> for whom English is a second language (this is, after
> all, taking place in an international forum), but I
> would predict, as an empirical regularity of discussion
> lists, that it would fail to move the discussion ahead
> in any useful direction.

Do you mean you actually don't expect all these foreigners to not be
capable of speaking American / Australian / Canadian / British / etc.
[Sorry, I'm not familiar with all the commonwealth countries that
approximate the Queens language.] english? How _______ of you.

			-- jbod
___________________________________________________
Come visit and see a new economic perspective --
       http://www.geocities.com/CapitolHill/1067
           Comments/arguments welcome.
..