More Closed Loop Economics (was Mumbo Jumbo...)

CSGnet Archive CSG2004
1

[From Rick Marken (2004.03.03.0920)]

Peter Small (2004.03.03)--

Although PCT is a neat concept to visualize the way humans make
decisions,

Actually, it's a model that explains how people control.

its weakness is that it can make predictions only if all
the variables are known. This is always highly unlikely.

I agree that if you don't know the variables involved in behavior then you
can't make predictions about behavior. But in the example you gave
(regarding Keynes ideas about the relationship between profit and
employment) the variables and their relationship to each other are known and
understood. Profit is one variable and it's the difference between expenses
and income. Employment is another variable and, for simplicity, I assumed
that employment was a producer's only expense. Income is equivalent to what
I called the disturbance, d. So in my equation, p = k1 (d - e) I was
basically defining profit as being proportional to income (d) less expenses
(e), which is certainly what profit (and loss, depending on the relative
values of d and e) is from my experience.

In this instance, PCT assumes that profit is the control variable and
provides the motivation for employers to hire more employees.

According to PCT, profit is a _controlled_, not a control variable. The
verbal distinction is very important. While a control variable controls the
actions of a system, a controlled variable is controlled _by_ the actions of
the system. A controlled variable doesn't provide motivation; it is the
object of control.

But, in a real life situation, profit is unpredictable.

This is true in my analysis as well because d (income) is assumed to vary
unpredictably over time. I can see how you wouldn't notice this if you are
unfamiliar with PCT. But in PCT, as applied to human behavior, disturbances
(d) are typically assumed to be unpredictable (and often detectable) time
varying influences on a controlled variable, things like the the direction
of a crosswind when controlling the direction in which a car is traveling.

The driving force is usually a perception of demand.

PCT shows that perceptions don't "drive" actions, they are controlled by
actions. Demand is not itself a controllable perception but a variable that
contributes to a perception that is controllable (profit). Demand
corresponds to d in my analysis. Variations in demand have an influence on
profit (a perception of the "bottom line") as do variations in expenses
(employment). The producer varies expenses to compensate for variations in
demand which are unpredictable. The point of my analysis was to show that
demand -- the disturbance to profit -- does appear to "drive" variations in
employment (expenses). But it "drives" employment via its effect on the
variable that is actually under control -- profit.

Employers aim to satisfy demand and
if demand appears to be greater than can be satisfied by a current
work force they will employ more people. This will continue to happen
as long as this maintains a profit. This is what Keynes' model is
telling us.

If this conclusion is based on the understanding that profit is a controlled
variable and that expenses are driven by demand because demand is one of the
main disturbances to that variable, then Keynes had it right.

Having spent some time in the fashion business, I've had plenty of
experience of this. When a fashion trend appears, all the
manufacturers respond by concentrating their resources upon it. There
is often a shortage of a particular skill and the cost of this skill
goes up.

Just as quickly as fashion trends appear, they can wane and die out.
Then resources are taken away and the employees specializing in that
area become redundant.

I have no doubt that this is what is observed. This is a _fact_: basically
it's the observation that e = d (employment varies directly with demand).
The behavioral illusion refers to the _interpretation_ of this fact as a
direct causal (or S-R) relationship where demand causes the producer to
increase employment. Control theory shows that this apparent S-R
relationship is an illusion. PCT shows that the e = d relationship is
actually a side effect of the operation of a control system (the producer)
that is acting (by varying expenses) to protect a controlled variable
(profit) from disturbance. The reason for the illusion is failure to notice
the controlled variable (profit in this case). PCT shows that the controlled
variable _must_ be taken into account in explanations of human controlling
(like that done by producers).

This is a stimulus-response situation and PCT
is only useful at a current, micro level and is ho help in being able
to control or predict the overall situation.

As I noted above, the apparent S-R relationship between demand and
employment (expenses) is an _illusion_. PCT is not only useful but essential
to understanding what is going on here. What is going on is _not_ a response
(varying expenses) to a stimulus (demand). It is an action aimed at
protecting a controlled variable (profit) from disturbance.

Best regards

Rick

ยทยทยท

--
Richard S. Marken
MindReadings.com
Home: 310 474 0313
Cell: 310 729 1400

2

From[Bill Williams 3 March 2004 3:30 PM CST]

I am comment on a discussion by Rick Marken of

some ideas introduced by Peter Small.

[From Rick Marken (2004.03.03.0920)]

Peter Small (2004.03.03)--

Rick says,

.... in the example you gave

(regarding Keynes ideas about the relationship

between profit and employment) the variables

and their relationship to each other are known

and understood.

The question, however, remains as to who understands what. It is well to
remember that Rick's attempts to do economic modeling have in the past been
described by Bill Powers as "giant leaps in the wrong direction."

Rick defines profit :

Profit is one variable and it's the difference

between >expenses and income.

So far so good. The matter becomes interesting when

the nature of a transaction-- that sales for one

party are equal to purchases for the other party, is considered.

But, Rick and Bill are too busy revolutionizing

economic theory to take time out to consider

fundamental assumptions.

However, when transactions are aggregated, then

sales or income and costs, or expenses are going to

be equal. This is true by definition. But you can

test the identity by actually adding the transactions

up.

Rick goes on to explain that,

profit as being proportional to income (d) less

expenses (e),

When considered in terms of the definition of a

transaction and an understanding of balance sheets,

profits ( in the aggregate ) are always equal to zero.

Peter says,

> In this instance, PCT assumes that profit is the

control variable and provides the motivation for

employers to hire more employees.

Since profits are equal to zero, the issue of the

motivation of capitalists to hire labor is as they say,

problematic.

Now, Rick goes on to say,

According to PCT, profit is a _controlled_, not a

control variable. As I have pointed out, and Powers

Rick is correct in regard to correct speech, but the

language "According to PCT, ..." is pretentious. "PCT"

never actually says anything. It is really Rick talking.

Rick goes on, and to this I can agree.

The

verbal distinction is very important. While a control

variable controls the

actions of a system, a controlled variable is

controlled _by_ the actions of

the system. A controlled variable doesn't provide

motivation; it is the object of control.

Good. Well put!

Then Peter says,

> > The driving force is usually a perception of demand.

Peter is using PCT incorrect language in his assertion, but

I think if we neglect PCT correctness for the moment,

what he says is worth paying attention to. Capitalists in

their attempts to maximize profit may use indirect measures

and indicators to guide their activity-- such as maintaining

what they call market share. And I think Peter's perception

is more likely than not, correct. Control theorist ought

to be able to convert what Peter says into a PCT theory

correct expression.

Rick goes on for a bit and eventually says, that

[demand drives employment ] via its effect on the

variable that is actually under control -- profit.

How a "variable"??? that is always equal to zero can be

controlled is a matter that Rick has never explained.

Peter states that employment is determined by profit maximization.

This is what Keynes' model is telling us.

This is not what Keynes says, but , Rick comments that,

If this conclusion is based on the understanding that

profit is a controlled

variable and that expenses are driven by demand because

demand is one of the

main disturbances to that variable, then Keynes had it right.

But, both Peter and Rick have it wrong. Profit

is not, in the aggregate a controlled variable. Control of a "variable"???
that is always zero is meaningless, unless an absurdity is itself

a meaningful condition.

After some statements by Rick that appear correct, Rick says,

PCT shows that the e = d relationship is

actually a side effect of the operation of a control system (the producer)

that is acting (by varying expenses) to protect a controlled variable

(profit) from disturbance.

In the aggregate context, which is what this discussion is ostensibly

about, profit equals zero, and Rick's discussion is meaningless--

except in the sense that it is one more hop in the wrong direction.

Rick asserts that control theory is essential to understand what is

going on. Actually people have gained a fairly reasonable level of

understanding without control theory. However, I would agree that,

What is going on is _not_ a response

(varying expenses) to a stimulus (demand).

However, when Rick goes on to say that, what is going on is

"... an action aimed at protecting a controlled variable

(profit) from disturbance.

We are back to the mode in which hops are once again being

taken in the wrong direction. In the meantime, some of what

Peter says that, despite a PCT incorrect mode of expression,

Has to say is being neglected.

Bill Williams