[From Rick Marken (2004.05.20.0830)]
Bill Powers (92004.05.20.0711 MST)
Rick Marken (2004.05.19.2130) --
This is how TCP's model worked as well. He assumed that P would change
to make up for leakage. This was his autoinflation.Are you sure that's the same? TCP said that if there is leakage, the price
P will increase; your model of the GNP controller says that if there is
leakage, the quantity of goods Q will increase.
It may have been a silly thing to do, but my model is nothing more than an
attempt to implement TCP's closed loop description of the economy using
control agents that make the things happen that TCP describes analytically.
For example, my model includes an agent -- the composite manager -- that
raises prices when leakage reduces income to the composite producer. So in
my model, as in TCP's, leakage leads to an increase in P (autoinflation) and
a decrease in Q relative what could have been produced if there were no
leakage. This result is shown in Table 1. In TCP's analysis, there is no
explanation of _why_ autoinflation occurs; in my implementation the "why" of
autoinflation is explained by the existence of an agent that actively
adjusts prices in order to keep income equal to expenditure.
Price can be freely
manipulated, but production of goods costs money and can't simply be
increased without increasing costs. Since your model doesn't keep track of
costs it allows quantity of production to be raised as freely as prices,
but is that realistic?
If production is increased the cost of that production (PQ') is
automatically increased and must be paid for by the composite producer. The
composite manager is controlling for keeping the composite producer's income
matching production expenses. So the model does keep track of production
costs. It cannot raise production of Q freely because the composite manager
is there making sure that what is spent for production doesn't exceed what
is taken in as income. Indeed, it is this constraint that leads the
composite producer/controller to _reduce_ production of Q when there is
leakage, since neither Q nor P can be raised freely.
But it doesn't contain any relationship saying that the quantity of goods
produced depends on the prevailing wage, the number of workers willing to
work at that wage, and the productivity of the workers.
True. I should put these in explicitly. Right now they are as invisible as
the market forces that determine the costs of goods and labor. They are not
in there, however, because they were not part of TCP's circular flow model
(Figure 1) that I was implementing using control agents.
But that money has to be repaid -- it's not just printed and given away.
And the borrower must repay not only the principal, but a little more, the
interest. Where in your model, or in TCP's analysis, is the provision for
repaying the money that providentially shows up, not to mention the interest?
If it wasn't in TCP's analysis it isn't in my model. Again, my model (for
better or worse) is simply an implementation, using control agents, of TCP's
analysis.
I don't know how your model handles this, but TCP's analysis simply added
whatever amount of money was lost by leakage, from an unnamed source.
My model does the same.
Since
this money didn't cost anyone anything, its effect should simply have been
to cancel the effect of leakage, restoring the circular flow to exactly the
level that would have existed if there had been no leakage. He asserted
that the producer raised the price to make up for leakage, thus causing
autoinflation
Right. And that's exactly what my model does, because it's an implementation
of his analysis.
the modeling helped me understand that the effect of leakage on growth rate
was assumed rather than derived.As is the failure of leakage to affect growth rate in your model.
Exactly. My model is an implementation of TCP's analysis.
It is simply not realistic to use Q' as an
independent means of counteracting changes in price or leakage, the way you
have done in your model.
OK. But this is simply the way TCP's analysis worked. But I think it is
realistic. When there is leakage, production of Q is reduced as a means of
making up for lack of income. This would account for the observed fact that
the amount of goods and services that _could_ be produced by an economy is
almost always greater than the amount that is actually produced.
My highest-level objection to your model is that I simply don't believe
that there is any GNP-controlling entity.
I think of the GNP controller as a _virtual_ controller. There is no entity
controlling GNP. The GNP controller represents a large group of individuals,
each controlling for getting their reference amount of goods and services
measured (perceived) in terms of current dollars. But I'm not married to
this idea. I like your suggestion of having consumers that control Q and PQ
separately. Changing the H. economicus model in this way would be striking
out on my own, in a sense, since the model was originally developed simply
as an implementation of TCP's circular flow analysis using control agents. I
will get to it as soon as possible.
Best regards
Rick
PS. I saw the bumps/dents display only as bumps. And I went back and forth
on the shading quite a bit.
···
--
Richard S. Marken
MindReadings.com
Home: 310 474 0313
Cell: 310 729 1400