Model Economy

CSGnet Archive CSG2000
1

[From Rick Marken (2000.03.13.1620)]

I am attaching a GIF flow diagram of my model economy. There are
two controlling entities in this model; I call them the "aggrgate
producer" (top) and the "aggregate producer/consumer" (bottom).
These two aggregate entities are shown as being seperate in the
diagram but, in fact, the aggregate producer is just a subset of
the population that makes up the aggregate consumer/producer. For
present purposes, we can consider that population to be the entire
population of the USA.

Each aggregate entity controls a variable that is a function of
variables in the environment (middle of the diagram). The variables
in the environment are PQ', which is the cost of producing Q' (all
the goods and services produced by the aggregate producer/consumer)
and P'Q', which is the income from selling Q'. So P is the amount
paid (in wages and profits) per unit Q' and P' is the selling
price per unit Q'.

The aggregate producer controls PQ'-P'Q', trying to keep this
difference at zero; the reference for this perception (rPQ'-P'Q')
is thus fixed at 0. This makes sense because the aggregate producer
must get back exactly the amount spent to produce Q', no more
(profit is already included in PQ') and no less (lest inventories
of unsold goods accumulate). The aggregate producer controls
PQ'-P'Q' by varying selling price per unit, P' (and, implicitly,
cost per unit, P, in current dollars; P is just P'one time
increment, dt, later).

The aggregate producer/consumer controls P'Q', which represents
all the goods and services available in units of P'. The aggregate
producer/consumer controls P'Q' by actually producing goods and
services (Q'). The aggregate producer/consumer's reference for
P'Q' is continuously increasing; this is what drives growth in
this economy. Part of this growth in the reference for P'Q' can be
thought of as being a result of the fact that the size of the
aggregate producer/consumer (in terms of population) is always
growing. The reference for P'Q' can also be thought to be
increasing because the people who make up the aggregate producer/
consumer always want a higher standard of living; the rate at
which this increase occurs would then be a "cultural" variable;
some nations expect (set a reference for) a 5% /year increase
while others expect (and set a reference for) only a 1% /year
increase.

There are two important disturbance variables that act
independently of the two aggregate controllers and make action
by these controllers necessary. The first disturbance is
"leakage", which influences only P'Q'. Leakage is the proportion
of P'Q' that is not used to purchase Q'.The amount of leakage
changes over time; the rate of change in leakage at any time
instant, dt, is the rate of leakage (dl/dt); it's this variable
(dl/dt), rather than the value of leakage itself at any instant,
that determines the rate of economic growth (dGNP/dt).

The second disturbance is wage increases due to collective
bargaining; this disturbance influences only PQ'.

I'll be happy to answer any questions about this model. For
now, I'll just say that the model seems to work pretty well.
Both aggregate entities successfully control the variables
they are set up to control (PQ'-P'Q' and P'Q'). Leakage has the
expected effect on inflation rate and relative output (columns
4 and 2, respectively, of Table 3-6 in TCP's "Leagage"); that is,
leakage increases inflation (via the auto-inflation effect) and
decreases relative output (Q' is less than what the aggregate
producer/consumer is cabable of producing). Leakage rate (dl/dt)
has the effect on rate of growth that leakage is _assumed_ to
have (columns 1 and 3 in Table 3-6).

Wage increases have almost the same effect on inflation and
relative output as leakage. This makes sense; wage increases
should be inflationary and they are in the model. Moreover,
wage increases depress relative output, just as leakage does;
the aggregate producer/consumer does not produce as much
(Q') as it is capable of producing when it has to increase the
cost to itself of production. Moreover, the effect of wage
increases and leakage on inflation and relative output are
independent and additive; increasing wages and leagage has
twice the effect on inflation as either one alone. So wage
increases do not solve the leakage problem created by
unequal distribution of wealth.

Best

Rick

···

---
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken/

Content-Type: image/gif; x-mac-type="47494666"; x-mac-creator="47436F6E"; name="EconModel.GIF"
Content-Description: GIFConverter 2.3.7 Document
Content-Disposition: inline; filename="EconModel.GIF"

Attachment converted: Betelgeuse:EconModel.GIF (GIFf/GCon) (0000D2A8)

2

[From Bruce Nevin (2000.03.13.2204)]

Rick Marken (2000.03.13.1620)--

These two aggregate entities are shown as being seperate in the
diagram but, in fact, the aggregate producer is just a subset of
the population that makes up the aggregate consumer/producer. For
present purposes, we can consider that population to be the entire
population of the USA.

Perhaps you could say that Producer and Consumer are roles in the economy
that can be filled (with various degrees of influence and sometimes even at
the same time) by each individual in the population. At any given instant,
the aggregate Producer is a subset and the aggregate Consumer is a subset
of the population, with possibly some individuals playing both roles and
some playing neither at that moment.

Thanks for the diagram. I'll study it.

        Bruce Nevin

···

At 04:17 PM 03/13/2000 -0800, Rick Marken wrote:

3

[From Rick Marken (2000.03.13.2200)]

Bruce Nevin (2000.03.13.2204)--

Perhaps you could say that Producer and Consumer are roles
in the economy that can be filled (with various degrees of
influence and sometimes even at the same time) by each
individual in the population. At any given instant,
the aggregate Producer is a subset and the aggregate
Consumer is a subset of the population, with possibly
some individuals playing both roles and some playing
neither at that moment.

Good point. I'll have more to say about it tomorrow. For now,
I'll just quickly say that I think the biggest mistake
economists have made is imagining that an economy can be
divided into two seperate kinds of people: producers and
consumers. I think an economy consists of people who are
_both_ producers and consumers at the same time, always. The
aggregate producer/consumer is really a collection of one
kind of person: a _controller_. A controller is an individual
who continuously acts (produces) in order to maintain (consume)
its inputs at desired levels.

Thanks for the diagram. I'll study it.

Great. But before you get too serious about it I should point
out one possibly confusing error. There is an arrow that appears
to be pointing up from the producer/consumer's input function
to the variable P'Q'. So it looks like the aggregate producer/
consumer has a direct influence on P'Q' via its _input_ funtion.
This is obviously wrong. This arrow line should have it's origin
at the line leaving Q', the one that points to the left, toward PQ'.
The arrow going up to P'Q' should, thus, be much shorter; it
should show that the aggregate producer/consumer has an influence
on both PQ' and P'Q' via its production of goods and services (Q').

I imagine there are other mistakes. I'll wait a couple days
and post an improved version then, with better names for the
systems called "aggregate producer" and "aggregate producer/
consumer". Right now, I think the "aggregate producer/consumer"
should be called the "aggregate controller" and the "aggregate
producer" should be called the "aggregate bean counter" (or,
even better, the "aggregegate comptroller")!

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken/

4

[From Bruce Nevin (2000.03.14.1028 EST)]

Rick Marken (2000.03.13.2200)--
10:01 PM 03/13/2000 -0800

Thanks for the clarification. One less wild goose.

This may be a domestic duck: Wouldn't an aggregate comptroller look at
things like maximizing the profit margin? Roughly, P'Q'-PQ' if P includes
all costs of producing Q'. This would set a reference for minimizing
P=wages and other costs of producing Q'=goods and services a reference for
maximizing P'=prices.

For money, there is widely held belief that more is better. (Bateson, among
others, says somewhere that this is the root of all evil, or something very
like that.) The zero sum reference for the upper control system in your
diagram does not reflect this mandate to maximize profit. So whatever it
is, it doesn't seem to walk like a comptroller or talk like a comptroller.
I'm not sure what human economic/social role it does correspond to. Perhaps
not any.

The answer is probably that I should reread the opening parts of _Leakage_
with this diagram and prior email in hand until I know what you're getting at.

        Bruce Nevin

5

[From Rick Marken (2000.03.15.0820)]

Bruce Nevin (2000.03.14.1028 EST)--

Wouldn't an aggregate comptroller look at things like maximizing
the profit margin?

Remember that profit is part of the cost of production; profit is
just wages paid to the owners of the business. So PQ' (cost of
production) includes the cost of profit.

The zero sum reference for the upper control system in your diagram
does not reflect this mandate to maximize profit.

Any increase in PQ' due to an aggregate increase in profit would
act as a disturbance that adds to the wage increase disturbance.
Of course, in the case of a profit increase, the cause of the
disturbance is not collective bargaining; the cause is "individual
bargaining" (eg. CEO with board of directors) which results in
bigger bonuses, stock options, etc. These profits (as well as wages
and capital expenditure) make up the cost of production (PQ') and
must be covered by receipts (P'Q').

So whatever it is, it doesn't seem to walk like a comptroller
or talk like a comptroller.

It seems to me that it is exactly like a comptroller; money
managers have to make sure that what a company gets paid equals
what it spends. That's just what my aggregate comptroller does;
it makes sure that what the aggregate controller gets paid
(from sales) matches it costs to produce goods and services (Q').

The aggregate comptroller doesn't care why it cost $1 zillion
to produce Q' (PQ' = $1 zillion); it just has to make sure
it receives $1 zillion in payment for Q' (P'Q' = $1 zillion).
If the books don't balance, all the aggregate comptroller can
do is increase the price (P') of Q' (at least, that's how
the model works; the aggregate comptroller only indirectly
influences wages -- and profits-- by varying P').

The structure of my model (shown again, slightly improved, in
a GIF below) emerged from my attempts to build a working
version of the TCP model of the economy. TCP's model is not
explicitly a control model but it does _imply_ that control
is involved. For example, the TCP diagram of the economy shows
a circular flow of dollars that implies a closed loop control
process. Similarly, TCP's equations say that PQ = P'Q'; this
implies a control process that adjusts P and Q (TCP's notation
for the cost and quantity of goods and services produced _with_
leakage) so that the product PQ equals P'Q' (the cost and quantity
of goods and services that would be produced if there were no
leakage).

So my model of the economy is not just a simple mapping of TCP's
model into a working model. I had to do some interpretation
and guesswork. The result is the model you see. It may not be
the only way to map TCP's ideas into a working model; but this
model does have the virtue of producing some of the TCP's main
results. For example, it produces quantitatively the same
reduction in Q' that TCP predicts when there is leakage. It
also produces quantitatively the same autoinflation effect that
TCP predicts when there is leakage. It even produces quantitatively
the same effect of leakage on economic growth that TCP predicts;
it's just that the effect in my model is produced by d leakage/dt
instead if leakage itself, as TCP predicts.

I want to discuss how I think my economic model (GIF below) maps
to the real world. But this post is already long enough (and I
have some things to do) so I'll discuss that in a later post.

Best

Rick

···

---
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates mailto: rmarken@earthlink.net
http://home.earthlink.net/~rmarken
Content-Type: image/gif; x-mac-type="47494666"; x-mac-creator="47436F6E";
name="EconDiagram.GIF"
Content-Description: Unknown Document
Content-Disposition: inline;
filename="EconDiagram.GIF"

Attachment converted: Betelgeuse:EconDiagram.GIF (GIFf/GCon) (0000D2B4)

6

[From Bruce Nevin (2000.03.15.1328 EST)]

Rick Marken (2000.03.15.0820)--

···

At 08:20 AM 03/15/2000 -0800, Richard S. Marken wrote:

Thanks, that is helpful. I believe I understand the Aggregate Comptroller
better now.

In the Aggregate (Producer/Consumer) Controller, what is the basis for
setting rP'Q' the reference signal presumably for both producing and
consuming goods? In the model, the Comptroller sets wages P and prices P'
but only the latter (prices) is overtly available to the Controller as a
means of communication from Comptroller to Controller. In the social world,
the wage or at least cash available for purchases are important variables
to the Consumer. And the Comptroller or an agent for the Comptroller
typically sets production targets -- indeed, the communication from
Comptroller to Producer is very direct when they are both in the same
entrepreneurial head.

I'm only looking for social correlates, not evaluating TCP's model or your
working representation of it from within the model as I ought to, but maybe
these questions have some relevance.

        Bruce Nevin

7

[From Rick Marken (2000.03.15.1500)]

Bruce Nevin (2000.03.15.1328 EST)--

In the Aggregate (Producer/Consumer) Controller, what is
the basis for setting rP'Q' the reference signal presumably
for both producing and consuming goods?

The reference signal (which I denote rP'Q') represents the desired
numerical value of the perceptual variable P'Q'. So the aggregate
controller is controlling a number (the product P'Q') that corresponds
to P' ($'s per unit goods and services) times Q' (the number of units
of goods and services produced).

P'Q' is a variable that has the dimension dollars ($). So what the
Aggregate Controller is controlling for in the model is _money_ ($).
This makes sense; it assumes that the Aggregate Controller is a
collection of people working to control for money. Of course, they
are working for the money in order to consume their portion of Q';
people ultimately consume goods and services. But people are
controlling for these goods and services in the form of money.

This makes sense to me. I do control for money. I want the money so
I can buy goods and services. But I don't always control for any
_particular_ set of goods and services; that is, I don't work just
to get a particular meal or a particular movie. I am really
working for the symbolic representation ($) of goods and services
that I can potentially consume.

So the Aggregate Controller in my model is controlling for _income_;
Q' is what this income can actually get the Aggregate Controller
at any instant, in terms of goods and services.

The Aggregate Controller controls for $ (P'Q') by producing Q' (goods
and services). This also make sense; we control for money by
producing goods and services. We don't control for money by producing
money -- at least, we're not supposed to. So, according to the model,
the Aggregate Controller is controlling for income (P'Q') by producing
Q' at a rate that depends on the difference between the current reference
specification for P'Q' (rP'Q'), which might be, say ,$ 1,000,000, and
P'Q', which might be $ 999,000.

The Aggregate Comptroller is an aspect of the aggregate economy that
exists because we have a money based economy. There would no need for
an Aggregate Comptroller if we each produced our share of Q' for
ourselves. But money exists so that we can each specialize at making
part of Q' and get all those bits of Q' we need. So I can contribute
to Q' by producing a particular service, get money for providing that
service and then use the money to buy all the goods and services I need.

The _market_, which is not represented in the model, determines what
people should be paid for producing their share of Q' and what each
piece of Q' should cost. The model assumes that the market does it's
work perfectly (at the individual level, which is why it's not
represented in the model) so when we talk about P' and P we are
talking about the average price (and wage) of all the appropriately
priced components of Q'.

The Aggregate Comptroller would also not be needed if there were no
disturbances to PQ' and P'Q'; that is, the Aggregate Comptroller would
not be necessary if there were no leakage or wage/profit disruption.
But these disturbances (which are really disturbances to the monetary
system) exist and the economy would fail completely as a result of
them if there were no agents somewhere in the system making sure that
the monetary balance sheets are balanced.

The Aggregate Comptroller is not really part of the productive economy;
it is literally the bean counters in government and industry. it is a
part of the "monetary system" and it is important. The Aggregate
Comptroller makes the productive aspect of the economy even more
productive because it allows _specialization_. Everyone doesn't
have to make everything for themselves. Money is the basis of
cooperative control. But like all cooperative arrangements it can
be abused; and it apparently is being abused by those who, through
talent, luck or sheer ruthlessness, manage to get, as income (profit
mainly) far more money than they can ever manage to use. Such money
is lost from the economy as leakage and the Aggregate Comptroller
has to make up for it by putting more money into the economy
artificially (because it's money that _doesn't_ represent a quantity
of Q'), depressing production and creating inflation.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates mailto: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

8

[From Bill (much better, thanks) Powers (2000.03.16.1348 MST)]

I'm enjoying what you're going with the economics model, especially your
very straightforward and simple discussions of what is going on. One
question: you have an arrow going from "wage increase" into PQ'. That makes
it seem that the wage increase comes from outside the Aggregate Producer
rather than being provided by the aggregate producer. The wage increase is
really a change in the Aggregate Producer's own payments into PQ', as
mandated by the collective bargaining agreement, not so?

Best, Bill P.

···

At 02:57 PM 03/15/2000 -0800, you wrote:

[From Rick Marken (2000.03.15.1500)]

Bruce Nevin (2000.03.15.1328 EST)--

In the Aggregate (Producer/Consumer) Controller, what is
the basis for setting rP'Q' the reference signal presumably
for both producing and consuming goods?

The reference signal (which I denote rP'Q') represents the desired
numerical value of the perceptual variable P'Q'. So the aggregate
controller is controlling a number (the product P'Q') that corresponds
to P' ($'s per unit goods and services) times Q' (the number of units
of goods and services produced).

P'Q' is a variable that has the dimension dollars ($). So what the
Aggregate Controller is controlling for in the model is _money_ ($).
This makes sense; it assumes that the Aggregate Controller is a
collection of people working to control for money. Of course, they
are working for the money in order to consume their portion of Q';
people ultimately consume goods and services. But people are
controlling for these goods and services in the form of money.

This makes sense to me. I do control for money. I want the money so
I can buy goods and services. But I don't always control for any
_particular_ set of goods and services; that is, I don't work just
to get a particular meal or a particular movie. I am really
working for the symbolic representation ($) of goods and services
that I can potentially consume.

So the Aggregate Controller in my model is controlling for _income_;
Q' is what this income can actually get the Aggregate Controller
at any instant, in terms of goods and services.

The Aggregate Controller controls for $ (P'Q') by producing Q' (goods
and services). This also make sense; we control for money by
producing goods and services. We don't control for money by producing
money -- at least, we're not supposed to. So, according to the model,
the Aggregate Controller is controlling for income (P'Q') by producing
Q' at a rate that depends on the difference between the current reference
specification for P'Q' (rP'Q'), which might be, say ,$ 1,000,000, and
P'Q', which might be $ 999,000.

The Aggregate Comptroller is an aspect of the aggregate economy that
exists because we have a money based economy. There would no need for
an Aggregate Comptroller if we each produced our share of Q' for
ourselves. But money exists so that we can each specialize at making
part of Q' and get all those bits of Q' we need. So I can contribute
to Q' by producing a particular service, get money for providing that
service and then use the money to buy all the goods and services I need.

The _market_, which is not represented in the model, determines what
people should be paid for producing their share of Q' and what each
piece of Q' should cost. The model assumes that the market does it's
work perfectly (at the individual level, which is why it's not
represented in the model) so when we talk about P' and P we are
talking about the average price (and wage) of all the appropriately
priced components of Q'.

The Aggregate Comptroller would also not be needed if there were no
disturbances to PQ' and P'Q'; that is, the Aggregate Comptroller would
not be necessary if there were no leakage or wage/profit disruption.
But these disturbances (which are really disturbances to the monetary
system) exist and the economy would fail completely as a result of
them if there were no agents somewhere in the system making sure that
the monetary balance sheets are balanced.

The Aggregate Comptroller is not really part of the productive economy;
it is literally the bean counters in government and industry. it is a
part of the "monetary system" and it is important. The Aggregate
Comptroller makes the productive aspect of the economy even more
productive because it allows _specialization_. Everyone doesn't
have to make everything for themselves. Money is the basis of
cooperative control. But like all cooperative arrangements it can
be abused; and it apparently is being abused by those who, through
talent, luck or sheer ruthlessness, manage to get, as income (profit
mainly) far more money than they can ever manage to use. Such money
is lost from the economy as leakage and the Aggregate Comptroller
has to make up for it by putting more money into the economy
artificially (because it's money that _doesn't_ represent a quantity
of Q'), depressing production and creating inflation.

Best

Rick
--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates mailto: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

9

[From Rick Marken (2000.03.16.2250)]

Bill (much better, thanks) Powers (2000.03.16.1348 MST) --

you have an arrow going from "wage increase" into PQ'.
That makes it seem that the wage increase comes from
outside the Aggregate Producer rather than being provided
by the aggregate producer.

I shouldn't have labeled the arrow "wage increase". The
arrow actually represents the effect of all variables that
affect the cost of production (PQ'). So I should have labeled
it "cost drivers", which can include anything that drives up
the cost of producing Q'; wage, profit, capital expenses, etc.
The effect of these cost drivers on the variable (PQ'-P'Q')
controlled by what I call the Aggregate Comptroller is
independent of the effects of the Aggregate Controller on
this variable; so they function as disturbances in the
Aggregate Comptroller control loop.

Note that these cost drivers are also independent of the
effects of the Aggregate Controller (the closest thing to
an "aggregate producer" in the model). The "cost driver" arrow
is kind of a place holder; it says that something other than
the Aggregate Controller and the Aggregate Comptroller --
something not currently represented in the model -- is
contributing to the cost of production (PQ'). The same is true
for the disturbance to P'Q' called "leakage". The model doesn't
say what is causing some of P'Q' to be held back from the Aggregate
Comptroller; the model just says that there is an effect on
P'Q' independent of those produced by the Aggregate Controller
(variation in Q') and the Aggregate Comptroller (variation in P').

I am starting to look at the model in terms of the kind of
yearly aggregate data (like that reported in the Statistical
Index) we would expect to see if the model is correct. I
am making some _very_ interesting discoveries. I'll start
reporting them as soon as I'm reasonably confident that I've
got things right.

Best

Rick (very glad to hear you're feeling better) Marken

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken/

10

[From Bill Powers (2000.03.17.1155 MST)]

Rick Marken (2000.03.16.2250)--

Bill (much better, thanks) Powers (2000.03.16.1348 MST) --

you have an arrow going from "wage increase" into PQ'.
That makes it seem that the wage increase comes from
outside the Aggregate Producer rather than being provided
by the aggregate producer.

I shouldn't have labeled the arrow "wage increase". The
arrow actually represents the effect of all variables that
affect the cost of production (PQ'). So I should have labeled
it "cost drivers", which can include anything that drives up
the cost of producing Q'; wage, profit, capital expenses, etc.
The effect of these cost drivers on the variable (PQ'-P'Q')
controlled by what I call the Aggregate Comptroller is
independent of the effects of the Aggregate Controller on
this variable; so they function as disturbances in the
Aggregate Comptroller control loop.

OK, I get it. I slipped a cog. I was thinking of the diagram of the
environment as if it represented the money and goods being moved around (as
in the Circular Flow diagram), rather than as information about the
environment being perceived and controlled by the two controllers. I'm not
quite sure how to say that, but maybe you see what I mean.

Best,

Bill P.

11

[From Rick Marken (2000.03.20.1100)]

Bill Powers (2000.03.17.1155 MST) --

I was thinking of the diagram of the environment as if it
represented the money and goods being moved around (as in
the Circular Flow diagram), rather than as information about
the environment being perceived and controlled by the two
controllers.

I have revised my diagram _again_ and attached it as a GIF.
I have put the two control systems (now called the _Aggregate
Manager_ and the _Aggregate Producer/Consumer_) on the same
side of the environment to indicate that both systems are part
of the same aggregate economic controller, operating on the
aggregate variables (PQ', P'Q' and Q') in the environment.

Basically, I have reorganized TCP's model (see "Leakage"
Figure 2-5, p. 98) so that it works as two control systems,
with the controlled variables made explicit. My Aggregate
Manager does part of what TCP's Composite Producer does. The
Aggregate Manager produces purchasing power (PQ', which
corresponds to B in TCP's Figure 2-5) by paying people (the
cost of production, which corresponds to W+K in Figure 2-5).

In TCP's Figure 2-5, buying power (B, which enters the
Composite Producer from the right) is somehow made to
equal the dollar cost of production (W+K, which leaves the
Composite Producer from the top). In my model, this matching
of W+K (=PQ') to B (= P'Q') is done by the Aggregate Manager,
who controls PQ'-P'Q' (which is the same as controlling W+K-B),
trying to keep this difference equal to zero.

So the Aggregate Manager does the main job of TCP's Composite
Producer; it keeps dollar income matching dollar outflow. It
explicitly treats W+K-B as a controlled variable and it also
shows (as TCP does, also) that leakage is a _disturbance_ to
this variable.

My Aggregate Producer/Consumer controls a variable that is
not explicitly controlled in TCP's model; it controls Q',
represented in terms of it's dollar value (P'Q'). In TCP's
model, the production of goods and services, Q (which
corresponds to Q' in my model) is uncontrolled; Q leaves
the Composite Producer to the left in Figure 2-5. So Q, in
the TCP model, is an uncontrolled side effect of the
circular flow of money (and the implicit abilities and
productive capabilities of the Composite Producer, which
are not explicitly represented in the model).

In my model, the Aggregate Producer/Consumer is _working_
to produce goods and services (Q') as the means of
controlling its income (P'Q') and, hence, the consumption
of goods and services. So while Q' is not explicitly controlled,
it is part of a control loop. The Aggregate Producer/Consumer's
control of P'Q' (in dollars) can be thought of as being
equivalent to the consumption (purchase) of all of Q'.

A major difference between my model and TCP's model is
that, in my model, the driving force behind economic
growth is in the Aggregate Producer/Consumer's increasing
reference for P'Q'. In other words, I am hypothesizing
that what TCP calls "intrinsic growth rate" -- the growth
rate that would be achieved without leakage -- is the
net _reference_ for growth in the Aggregate Producer/Consumer.
In contrast, TCP makes growth rate an "intrinsic" aspect
of the Composite Producer; the American composite producer,
according to TCP, because of its resources and skills,
would grow at 12%/year without leakage.

I like my approach better because 1) it makes more sense
to me from a control theory perspective; growth depends on
the capabilities _and) wants of the people who make up the
economy and 2) it means that a no-growth economy (which we
must eventually achieve since we live on the surface of a
sphere) can be achieved if people simply adopt the goal
of no growth ("think small"). In my model, a no growth
economy would still provide people what they want and
need (Q'), it's just that their standard of living (in
terms of amount of Q'/person) would level off.

The environmental variables in my model (PQ', P'Q' and
Q') represent real entities. PQ' and P'Q' represent real
dollars, either in the familiar form of treasury notes
or numerical balances in a computer. Q' is also a
real entity; it's the "stuff" that is produced by the
economy, from cars to food to lectures on PCT.

Now let me quickly describe one of the main findings from
this modeling effort. First, some data. I calculated the
correlation between leakage (alpha only) and growth rate
(d GNP/dt) for the years 1952-1998. The correlation
between alpha and d GNP/dt is -0.157; the correlation
between change in alpha (d alpha/dt) and d GNP/dt -0.2291.

TCP says that the low correlation between alpha and d GNP/dt
results from not taking rho leakage into account. Be that
as it may, at least you can see that the correlation between
d alpha/dt (the actual cause of economic growth in my model)
and d GNP/dt is at least as high as the one between alpha
and d GNP/dt.

Now the finding. I produced a series of leakage (alpha) and
GNP (Q') values using my model to see what the correlations
between alpha and d GNP/dt and between d alpha/dt and
d GNP/dt would be when produced by the model. It turns out
that the correlations depend _strongly_ on the time interval
between measurements of leakage (alpha) and GNP (Q').

The model currently updates the variables 10 times per year
(dt =.01 year). If you take the measurements of leakage and
GNP only at the end of each year (as is done in the Statistical
Index) the correlations are remarkably low _and_ very similar
to the correlations found for the real data! For example, on
one run, I found a correlation between alpha and d GNP/dt
of -0.11 (compared to the empirical -0.157) and a correlation
between d alpha/dt and d GNP/dt of -0.37 (compared to the
empirical -0.2291). The exact correlations obtained from the
model depend on the nature of the temporal variations in leakage
(alpha) and cost drivers.

If, on the other hand, you take the measurements of leakage and
GNP ten times per year, the correlation between d alpha/dt and
d GNP/dt is quite high (eg., -.98), as it should be (because
d alpha/dt is what determines growth rate in the model) and
the correlation between alpha and d GNP/dt is quite low
(eg. -.07).

These results suggest that one way to test the model is to
obtain higher time resolution economic data. Quarterly reports
of GNP and alpha leakage should do the trick. If I can get
about 4 or 5 consecutive years of quarterly reports of GNP and
alpha I will be able to test the model. If my model is right,
the correlation between quarterly values of d alpha/dt and
d GNP/dt should be much higher than what was observed for
the yearly values (-0.2291) and the correlation between
quarterly values of alpha and d GNP/dt should be the same of
lower than what was observed for the yearly values (-0.157).

Best

Rick ("Adam Smith") Marken

···

---
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates mailto: rmarken@earthlink.net
http://home.earthlink.net/~rmarken
Content-Type: image/gif; x-mac-type="47494666"; x-mac-creator="47436F6E";
name="EconDiagram2.GIF"
Content-Description: Unknown Document
Content-Disposition: inline;
filename="EconDiagram2.GIF"

Attachment converted: Betelgeuse:EconDiagram2.GIF (GIFf/GCon) (0000D2C2)

12

[From Rick Marken (2000.03.24.0830)]

Is there no more interest in PCT or is it the "dismal"
economics thread that is keeping people from posting to
CSGNet? If there is anyone out there who is willing to
discuss it, I would appreciate getting comments on my last
version of the economics model (Rick Marken (2000.03.20.1100)).
Does it make sense? I have found some quarterly and monthly
GNP and savings data but there are scaling problems that I
don't understand (some is in chained 1996 dollars, some in
fixed 1996 dollars and some just seasonally adjusted) and
I don't have monthly and quarterly data for all the variables
of interest. The model still needs a lot of work. But right
now I think I have to learn what the publicly available
aggregate economic data represents.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates mailto: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

13

[From Rick Marken (2000.03.24.0830)]

I would appreciate getting comments on my last
version of the economics model (Rick Marken (2000.03.20.1100)).
Does it make sense?

Sorry, I've been wrapped up with putting my share of Q' out into the
economic environment to keep up with my family's share of taking P'Q' out
of the economic environment.

It seems wrong to treat the output of the "Aggregate Manager" as a single
variable that some unrepresented and unlabelled entity in the environment
splits into P (Producer cost, i.e. wages) and P' (Consumer cost, i.e.
prices). But to have two distinct outputs, you need more complexity in the
control loop for the Aggregate Manager.

I am not competent to advise about your statistical data. But I am
confident that you are doing is important and useful.

        Bruce Nevin

···

At 08:25 AM 03/24/2000 -0800, Richard S. Marken wrote:

14

[From Rick Marken (2000.03.25.1020)]

Bruce Nevin (2000.03.24) --

Sorry, I've been wrapped up with putting my share of
Q' out into the economic environment to keep up with
my family's share of taking P'Q' out of the economic
environment.

Good for you!

It seems wrong to treat the output of the "Aggregate Manager"
as a single variable that some unrepresented and unlabelled
entity in the environment splits into P (Producer cost, i.e.
wages) and P' (Consumer cost, i.e. prices).

Good observation. Actually, in the model, the Aggregate Manager
just sets the increase in the cost (P') of P'Q' to the Aggregate
Producer/Consumer. The change in P is a delayed side effect of
the Aggregate Manager's change in P'. I should figure out a way
to show this more clearly in the diagram.

This all makes (some) sense when you recall that the Aggregate
Manager and the Aggregate Producer/ Consumer are the _same
aggregate entity_. When the price (P') of Q' increases (due to the
actions of the Aggregate Manager), the Aggregate Producer/Consumer
must have more money to purchase it. So the Aggregate Producer/
Consumer (as Aggregate Manager) must pay itself more (in wages
and profits, PQ') so it can purchase all of the Q' that is being
produced. If P is not increased along with P' (and I can make
P a constant in the model) then the model still works but
inventories of unbought Q' increase to what seem to me to be
unrealistic proportions of GNP. Also, fixing P seems to make
no sematic sense; it says that the Aggregate Manager will
continue to increase prices without proportionate increases
in its own income (and, hence, it's cost of producing Q').
It's true that real individual managers may try to keep workers
pay raises at zero to control production costs, even while
prices are rising, but these managers (CEOs, etc) certainly
increase their own incomes (as executive salary, stock options,
etc) during these same price increase periods.

So I suspect that P (cost of production) does increase in
proportion to cost of goods and services (P'). Even if the
labor wage segment of P doesn't increase by much, the executive
salary segment will make up for it. Also, it's only when the
model is set up this way (so that P changes in proportion to
the actively produced changes in P' made by the Agggregate
Manager) that you get the expected effect of leakage on
production and economic growth rate.

Thanks for the sharp-eyed comments.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken/

15

[From Bill Powers (2000.03.25.1055 MST)]

Rick Marken (2000.03.20.1100)--

I have revised my diagram _again_ and attached it as a GIF.
I have put the two control systems (now called the _Aggregate
Manager_ and the _Aggregate Producer/Consumer_) on the same
side of the environment to indicate that both systems are part
of the same aggregate economic controller, operating on the
aggregate variables (PQ', P'Q' and Q') in the environment.

...

In my model, the Aggregate Producer/Consumer is _working_
to produce goods and services (Q') as the means of
controlling its income (P'Q') and, hence, the consumption
of goods and services. So while Q' is not explicitly controlled,
it is part of a control loop. The Aggregate Producer/Consumer's
control of P'Q' (in dollars) can be thought of as being
equivalent to the consumption (purchase) of all of Q'.

I wonder if it might not be a good idea to separate good/services from
dollars. As consumers we have no inherent interest in the dollars; we work
for dollars only in order to get a share of the total production of goods
and services.

There are really two flows: one goes from raw materials through production
through sales through consumption to trash, and the other is the circular
monetary flow.

One link between the two flows is productivity, in quantity produced per
unit time per dollar spent per unit time: 1/(cost of production). The other
link is quantity consumed per dollar spent (which is 1/price). The consumer
has a reference level for quantity consumed per unit time (an average over
all people, all goods, and all services), but there is no inherent
reference level for money. The need for money is dependent on the need for
goods and services.

Assume for a moment that we do away with the money. We now have the
aggregate consumer devoting a certain amount of effort to producing Q,
which is exactly what is needed to satisfy the needs of the consumer. The
amount of work will increase or decrease until the error between amount
consumed (Q) and amount desired to be consumed (rQ) is as small as
possible. A certain amount of Q will prove useless and will be discarded
without helping to correct error; this disturbance will simply reduce the
net amount produced and increase the error a little.

In this idealized setup, "people" will produce as much as they want to
consume by working as hard as they want to work. Given managers whose job
is to vary production and workforce to keep inventories constant in all
different sectors of the economy, eventually this system should settle down
to a condition in which exactly as much of any particular good or service
is produced as there is consumption of it. Increases in productivity will
result in reducing hours worked and increasing goods/services available (if
there is some remaining error). This kind of model should be quite easy to
set up.

HOWEVER! We now introduce individuals who want more than their share -- in
fact, who want an unlimited amount, as much as they can get. Furthermore,
these individuals want to maximize their consumption of Q while at the same
time minimizing their personal contribution to production of Q.

It would be interesting to see what happens when we add a population of
such individuals to the moneyless economy, especially if we put limits on
production (say, limit the time per week a person can or will work). One
thing I predict is that the total hours worked by employed people will
increase in order to satisfy the increased demand for goods by non-workers.
Non-workers taking goods is a disturbance to workers; it means workers must
work more to get the same input of Q for themselves. The total consumption
by workers will decrease and the error (rQ - Q), will increase.

If we now add money back into the setup, what changes? I'm sure there's a
long list, but some things will not change. Managers will still want to
keep inventories constant at a low level. People, as consumers, will still
work as much as they need and want to, now to get the money they need to
purchase the Q they want (want = rQ), rather than just picking up what they
need/want without paying for it. Rather than being paid in goods and
services, people are paid in money which represents their claim on goods
and services.

The presence of money is a great help to the people who want to get as much
Q as possible while working as little as possible to produce Q. Vast
accumulations of many (perishable) goods are more of a liability than an
asset, whereas vast accumulations of money can be spent on either goods or
services as they are needed. For people whose needs and wants are finite,
money is not a critical tool, because what is desired is obtained and
consumed on the spot. But when the wants are unlimited, money is essential.

Enough for now. Just some musings -- no suggestion that you should alter
what you're doing.

Best,

Bill P>

16

[From Bruce Nevin (2000.03.25 2236 EST)]
Rick Marken (2000.03.25.1020)

···

At 10:18 AM 03/25/2000 -0800, Rick Marken wrote:

This all makes (some) sense when you recall that the Aggregate
Manager and the Aggregate Producer/ Consumer are the _same
aggregate entity_.

Rick Marken the typist and Rick Marken the reader are the same aggregate
entity, but one would want to represent his internal control structure in
at least enough detail to show two distinct but perhaps interrelated
control loops for these activities.

        Bruce Nevin

17

[From Rick Marken (2000.03.26.1050)]

Bill Powers (2000.03.25.1055 MST)

I wonder if it might not be a good idea to separate
good/services from dollars.

That's basically what is accomplished by the two loops in my
model. The Aggregate Producer/Controller turns raw materials
into goods and services for itself; the Aggregate Manager
maps these goods and services into money that is paid to and
received from the Aggregate Producer/Controller.

There are really two flows: one goes from raw materials
through production through sales through consumption to
trash, and the other is the circular monetary flow.

Yes. The first flow coresponds to the Aggregate Producer/
Controller loop; the second to the Aggregate Manager loop.

In this idealized setup, "people" will produce as much as
they want to consume by working as hard as they want to work.

Yes. The Aggregate Producer/Controller works this way. Even
though it is controlling P'Q' rather than Q', it works as you
say if P' is made a constant (1.0). In the model, the Aggregate
Producer/Controller's reference (rQ') for input (Q'; with
P' = 1) constantly increases to account for the fact that the
size of the Aggregate Producer/Controller is always growing
(population increase) so the absolute desired amount of Q'
must also grow. In reality, GNP often grows faster than
population;, To the extent that this increase in GNP is not
exclusively due to increases in P' (and I'm finding that the
extent to which GNP reflects changes in Q' versus changes in P'
is not at all clear) the model assumes that there is a cultural
desire to continually have more Q', above and beyond that
needed to support a growing population.

HOWEVER! We now introduce individuals who want more than
their share...

It would be interesting to see what happens when we add
a population of such individuals to the moneyless economy,
especially if we put limits on production

This is similar to what my little "Cooperation" presentation
was about at the 1999 CSG Meeting (also available on the web
at http://home.earthlink.net/~rmarken/ConfEcon.html). It
was done in terms of population segments rather than individuals
but it shows that, if one segment can get more than its fair
proportion of income ("fair" being income nearly proportional to
that segment's proportion of the population) then members of
the other segments can't get the levels of input (combinations
of the products produced by all productive segments of the
economy) they want.

So if the control model of the economy is correct, people
who make a disproportionatly large amount of money are, contrary
to capitalist mythology, not good for the economy; they are
what keeps the economy from working for all it's members. The
very rich are simply being uncooperative. Unfortunately, current
capitalist mythology celebrates the rich as the "driving force"
of the economy; the rich entrepreneur has become an object of near
worship in the US. This myth, based on an illusion as profound as
the behaviroal illusion at the individual level, is going to be
_very_ tough to dispel.

The presence of money is a great help to the people who want
to get as much Q as possible while working as little as possible
to produce Q.

Yes. See my comments above. Money is absolutely essential to
an economy based on division of labor. Money is _much_ more
convenient than barter; but it makes it very easy for even well
intentioned people to screw up the system by hoarding an amount
of money that represents far more goods and services than they
could ever actually consume.

I continue to make very interesting discoveries with my model.
I am also learning a great deal about economic statistics. I
will give a report soon.

Thanks a ton for the comments.

Best

Rick

···

---
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken/

18

[From Rick Marken (2000.03.27.1400)]

Me:

This all makes (some) sense when you recall that the Aggregate
Manager and the Aggregate Producer/ Consumer are the _same
aggregate entity_.

Bruce Nevin (2000.03.25 2236 EST)--

Rick Marken the typist and Rick Marken the reader are the
same aggregate entity, but one would want to represent his
internal control structure in at least enough detail to show
two distinct but perhaps interrelated control loops for
these activities.

Yes, of course. That's what the two loops (Aggregate Manager
and Aggregate Producer/Consumer) are about.

Perhaps I should have said that the Aggregate Manager and
the Aggregate Producer/ Consumer are _components of_ the same
aggregate economy (which includes the aggregate environment).
My point was that all the money involved in purchasing goods
and services (Q') must be produced by the economic system
itself (consisting of the Aggregate Manager, Aggregate Producer/
Consumer and the environment). When the price of goods and
services is increased (by the Aggregate Manager in my model)
the extra money needed to _pay_ for these now more expensive
goods and services must come from the same economic system
that increased the prices. I attribute this increase to the
environment. In fact, it is probably done by the Aggregate
Manager itself; the environment just puts in a delay (pay
raises follow price increases) and includes currently
unspecified entities (the government) that produce the
money needed to cover these increased payments.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates mailto: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

19

[From Bruce Nevin (2000.03.27 2246 EST)]

Rick Marken (2000.03.27.1400)]

···

At 01:50 PM 03/27/2000 -0800, Richard S. Marken wrote:

Yes, of course. That's what the two loops (Aggregate Manager
and Aggregate Producer/Consumer) are about.

Is it possible that the Aggregate Producer/Consumer itself also "contains"
two loops, one the Producer, the other the Consumer? Then you wouldn't have
that funky unlabelled branch out in the environment that we were talking about.

        Bruce Nevin

20

[From Rick Marken (2000.03.28.1000)]

Bruce Nevin (2000.03.27 2246 EST)--

Is it possible that the Aggregate Producer/Consumer
itself also "contains" two loops, one the Producer,
the other the Consumer?

I think what you're asking is whether it's possible
to change the model so that the control loop called
the "Aggregate Producer/Consumer" is broken into two
control loops. And, of course, it is possible to
change the model in this way. But it's not immediately
apparent to me what would be gained from doing this.
What variables would be controlled by each of these
loops? How would this remove the "funky unlabelled
branch out in the environment" that connects the
output of the Aggregate Manager to both PQ' and P'Q'?
Perhaps you could send me a diagram of what you have
in mind?

Personally, I like the idea of having a single loop
represent the Aggregate Producer/Consumer. This captures
the fact that, at the aggregate level, "economic man" is
producing and consuming _at the same time_. That is,
economic man is a _controller_ of it's own input.

Even at the individual level we can see that people are
producing _while_ they are consuming. The worker in a car
production plant is producing a car _while_ consuming
capital (machinery) that was produced by others. If you look
only at the fact that this individual is producing a car,
then it looks like production and consumption are sequential;
the car must be produced by the worker before it can be
consumed by the worker. But if we look at the aggregate
of what any individual produces and consumes and then look
at the aggregate of what all individuals are producing and
consuming we get quite a different picture.

At the aggregate level it is clear (to me, anyway) that
economic man (the aspect represented by the Aggregate
Producer/Consumer in my model) is producing _while_ it is
consuming. Aggregate economic man is driving home while it
is producing cars; it is eating while it is growing and
packaging food; it is teaching computer science while it
is applying this knowledge to internet innovation.

We can look at economic man in terms of separate production
systems (eg. factories) and consumption systems (eg. households)
just as we can look at individual living systems in terms
of separate output devices (eg. muscles) and input devices
(eg. eyes and ears). But when these systems are functionally
linked in a closed negative feedback loop, so that outputs
affect inputs _while_ inputs affect outputs, then we have
to look at the system as a control of input system.

In my model, economic man is a control of input system; the
input controlled is income -- P'Q'. The Aggregate Producer/
Consumer controls perceived (consumed) income, P'Q', by varying
production (Q') to compensate for disturbances to P'Q'
created by leakage.

Then you wouldn't have that funky unlabelled branch
out in the environment that we were talking about.

Again, I think looking at this from the aggregate perspective
may make the branch look a bit less funky. All that the
branch says is that the output of the Aggregate Manager loop
affects both the price of goods and the cost of production
simultaneously. This seems may funky because it suggests
that a person can give himself a pay raise (increasing the
cost of production, PQ') whenever the price of what he
produces is increased. Of course, this makes no sense at the
individual level; A car production worker can't raise his
salary by $1000 when the retail price of the car he makes
goes up $1000.

But things work differently at the aggregate level because
the Aggregate Manager is raising prices to pay itself. A
price increase (increase in P' to cover the costs of
production) are immediately returned to the Aggregate
Manager as wage, salary and profit (PQ') that is used to
cover the cost of production (PQ'), which is, simultaneously,
the income (P'Q') used to buy what is produced. Obviously,
this process requires that money be put into the system from
somewhere in the environment. This is the part of the system
that is not currently shown in my diagram (or represented in my
model). The diagram (and the model) assumes that the Aggregate
Producer/Consumer can get the P'Q' dollars it needs to buy Q' at
the marked up price, P'. These dollars must come form somewhere
and my guess is that they come from _borrowing_.

In a real economy, borrowing is needed to bridge the time
gap (at the individual _and_, to some extent, at the aggregate
level) between a price increase and the wage/profit increase
that covers it. This borrowing is done by the entire consumption
side of the aggregate economy; it is done by household-type
consumers who buy the non-capital component of P'Q' and it
is done by corporate-type consumers who buy the capital
component of P'Q'.

So what is missing from the diagram is, possibly, a third
Aggregate Control System that controls a reservoir of extra
money that can be used for borrowing. Borrowing makes up the
dollar difference between PQ' and P'Q' that exists when there
is a _change_ in leakage. This dollar difference remains
small (when leakage is not changing too quickly) because
the Aggregate Producer/Consumer quickly adjusts its output
(Q') to compensate for the leakage change. If there is no
change in leakage, then PQ' is always kept close to P'Q'
(production costs always nearly match consumption costs)
and borrowing becomes almost unnecessary.

Anyway, the model does work as is and produces results
that are a least consistent with the behavior of many
aspects of the aggregate economy. But suggested improvements
are welcome as long as I can see (via a functional flow diagram,
preferably) how those suggested improvements are to be incorporated
into the model.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates mailto: rmarken@earthlink.net
http://home.earthlink.net/~rmarken