economics

[Bill_Powers1]

[From Bill Powers (2009.09.05.2315 MDT)]

I ended my last post by saying
"I wish we could just start working on the model, not worrying about where it will lead but just trying to get it right. That's all we really need to do."

At the moment, having had an extra cup of coffee at lunchtime which for me is a mistake, I'm wide awake and thinking even more firmly that building the PCT model of the economy is the only way we are ever going to understand the economy, and that understanding it is the only way we will ever learn to manage it competently.

I hope that others will become engaged in this. But I also realize that it's difficult to guess what this Powers guy has in mind as a model, and that I do have more practice than most at designing working models. So maybe I can do a little guiding and teaching just to get things moving.

One thing I'd like to do as a starter is to catch that horse pushing a cart along in front of it, and get the cart back behind the horse where it belongs. Economists like to generalize from economics to life as a whole, and in some ways I agree with them, but not to the extent of confusing the part with the whole. The only reason there is a subject called economics is that human beings have invented some cultural systems for dealing with the production and exchange of goods and services, including the mechanics of producing, buying, and selling as well as a financial superstructure invented largely for the purpose of extracting some of the vital fluids from the system without doing any actual manual labor. Bernie Madoff is a natural product of the latter part of the economic system. Bernie Made Off with a large chunk of other people's buying power. But people had working economies long before this present system came into being and before such depredations were possible.

The system of rules governing the transactions and the superstructure results from the interactions among people trying to live their lives as best they can. But there is a lot more to living a life than engaging in economic interactions. The cart is economics; the horse is the whole human hierarchy of control systems and the body they live in.

I suggest that we could start just by modeling a consumer who is trying to get along in the economic system as it exists today, not trying to model the whole system but just this one part of it. Economists have done this only in a sort of back-door way, making assertions about what people want and do and why they act that way, but not really trying to produce a bona fide model of a person that could be tested.

PCT gives us a way to start. Bill Williams had the right idea when he tackled the Giffen Paradox. He came from Boulder, Colorado to Northbrook, Illinois one weekend (in a suit and tie -- he had never met me before) equipped with some graphs he had drawn showing how the Giffen Effect arises, and wanting me to help him write a computer program using control systems to make his idea work. During that weekend we worked out the program and got it to run. A good deal of the effort was devoted to eliminating the abstractions of economics and boiling the problem down to simple concrete interactions arising from controlling variables that interact with each other. Meat and bread as objects can be acquired by trading money for them. Both are a means to a non-negotiable higher-level end: obtaining calories to stay alive. However, meat provides more of something else at a higher level than bread does -- pleasure, prestige, protein, taste, or whatever you want to propose that leads to preferring meat to bread. Add to that a limit on the availability of money, and the Giffen Effect pops right out. There is no Giffen "Paradox." It's simply a result of how a particular set of interacting control systems behaves under specified conditions.

We didn't try to model the production of meat and bread or account for the prices that are charged for each. We just took those as givens of the current setup, and looked at some of the consequences.

What we can do now is to set up a collection of control systems inside a consumer that try to achieve various goals, and treat the rest of the economy as a big set of Environmental Feedback Functions. Consumers can act on this external world to get money from it, and to get goods and services by using some of the money so obtained. Later, of course, we might look at the producers/managers who treat consumers as an EFF by the use of which the producers/managers control whatever variables are of concern to them, but we don't have to do that yet. All this is done in the most mundane, most direct, and least abstract way we can think of.

Eventually we will have to introduce more complications, especially when we start letting the human agents in the two main sectors of the economy, producers and consumers, interact with each other more directly (for example, through advertising and the manipulation of prices and through wage negotiations). We will also need to introduce sex in the form of an expanding population, and violence in the form of laws and law enforcement, which will lead to banks (to adjust the money supply as required by growth) and government (to resolve conflicts by force, alas). But by that time we will have expanded the model beyond economics to the whole huge arena of Life. The principles we will use will not be economic principles but principles that are inherent in everything human beings do. But always composed of Perception; Comparison; Action. Maybe even Reorganization.

We can start with Goods and Services and reference conditions that pertain to them, and Wages and Capital Income and their related reference conditions. Both of those variables involve what Jay Forrester (System Dynamics) called "Stocks and Flows." At any given time there is a Stock of goods and services and another Stock of money. These stocks grow and shrink as the items they contain are acquired and used. The means of acquiring money is to work for some length of time at a given wage or to receive capital income from a given amount of investments; the income goes into the Stock of money. The means of acquiring goods and services is to purchase them at various prices, with the cost coming out of the Stock of money, and the goods and services going into the Stock of possessions (some very briefly). The Stock of possessions is depleted as they are used or as they depreciate.

These Stocks are integrators; the total amount of Stock is the cumulative sum over time of inputs (positive) and outputs (negative). This is very handy for control systems, because a control loop with all relationships but one being simple proportionalities, and the one exception being a time integration, is unconditionally stable. You can freely adjust its loop gain without worrying about oscillations or runaway (unless you get the sign of feedback wrong, which is easily fixed).

Finally I'm sleepy so that will have to do for now. I really want to know who is aboard for this. Let me know.

Best,

Bill P.

[Gary_Padover]

To Bill. as usual i am fascinated by your presentations of concepts and their applications to people motivations. Would like to correspond more about pct applications to economics. you have a conceptual collaborator in me at the very least Bill. it is refreshing and inspiring to continue to read your thoughts and ideas as you formulate them. Respectfully gary padover.

[rsmarken]

[From Rick Marken (2009.09.06.1100)]

Bill Powers (2009.09.05.2315 MDT)–

Again, missing from this (for me) is any description of the data against which the behavior of the model will be evaluated.

Best

Rick

···

I ended my last post by saying

“I wish we could just start working on the model, not worrying about where it will lead but just trying to get it right. That’s all we really need to do.”

At the moment, having had an extra cup of coffee at lunchtime which for me is a mistake, I’m wide awake and thinking even more firmly that building the PCT model of the economy is the only way we are ever going to understand the economy, and that understanding it is the only way we will ever learn to manage it competently.

I hope that others will become engaged in this. But I also realize that it’s difficult to guess what this Powers guy has in mind as a model, and that I do have more practice than most at designing working models. So maybe I can do a little guiding and teaching just to get things moving.

One thing I’d like to do as a starter is to catch that horse pushing a cart along in front of it, and get the cart back behind the horse where it belongs. Economists like to generalize from economics to life as a whole, and in some ways I agree with them, but not to the extent of confusing the part with the whole. The only reason there is a subject called economics is that human beings have invented some cultural systems for dealing with the production and exchange of goods and services, including the mechanics of producing, buying, and selling as well as a financial superstructure invented largely for the purpose of extracting some of the vital fluids from the system without doing any actual manual labor. Bernie Madoff is a natural product of the latter part of the economic system. Bernie Made Off with a large chunk of other people’s buying power. But people had working economies long before this present system came into being and before such depredations were possible.

The system of rules governing the transactions and the superstructure results from the interactions among people trying to live their lives as best they can. But there is a lot more to living a life than engaging in economic interactions. The cart is economics; the horse is the whole human hierarchy of control systems and the body they live in.

I suggest that we could start just by modeling a consumer who is trying to get along in the economic system as it exists today, not trying to model the whole system but just this one part of it. Economists have done this only in a sort of back-door way, making assertions about what people want and do and why they act that way, but not really trying to produce a bona fide model of a person that could be tested.

PCT gives us a way to start. Bill Williams had the right idea when he tackled the Giffen Paradox. He came from Boulder, Colorado to Northbrook, Illinois one weekend (in a suit and tie – he had never met me before) equipped with some graphs he had drawn showing how the Giffen Effect arises, and wanting me to help him write a computer program using control systems to make his idea work. During that weekend we worked out the program and got it to run. A good deal of the effort was devoted to eliminating the abstractions of economics and boiling the problem down to simple concrete interactions arising from controlling variables that interact with each other. Meat and bread as objects can be acquired by trading money for them. Both are a means to a non-negotiable higher-level end: obtaining calories to stay alive. However, meat provides more of something else at a higher level than bread does – pleasure, prestige, protein, taste, or whatever you want to propose that leads to preferring meat to bread. Add to that a limit on the availability of money, and the Giffen Effect pops right out. There is no Giffen “Paradox.” It’s simply a result of how a particular set of interacting control systems behaves under specified conditions.

We didn’t try to model the production of meat and bread or account for the prices that are charged for each. We just took those as givens of the current setup, and looked at some of the consequences.

What we can do now is to set up a collection of control systems inside a consumer that try to achieve various goals, and treat the rest of the economy as a big set of Environmental Feedback Functions. Consumers can act on this external world to get money from it, and to get goods and services by using some of the money so obtained. Later, of course, we might look at the producers/managers who treat consumers as an EFF by the use of which the producers/managers control whatever variables are of concern to them, but we don’t have to do that yet. All this is done in the most mundane, most direct, and least abstract way we can think of.

Eventually we will have to introduce more complications, especially when we start letting the human agents in the two main sectors of the economy, producers and consumers, interact with each other more directly (for example, through advertising and the manipulation of prices and through wage negotiations). We will also need to introduce sex in the form of an expanding population, and violence in the form of laws and law enforcement, which will lead to banks (to adjust the money supply as required by growth) and government (to resolve conflicts by force, alas). But by that time we will have expanded the model beyond economics to the whole huge arena of Life. The principles we will use will not be economic principles but principles that are inherent in everything human beings do. But always composed of Perception; Comparison; Action. Maybe even Reorganization.

We can start with Goods and Services and reference conditions that pertain to them, and Wages and Capital Income and their related reference conditions. Both of those variables involve what Jay Forrester (System Dynamics) called “Stocks and Flows.” At any given time there is a Stock of goods and services and another Stock of money. These stocks grow and shrink as the items they contain are acquired and used. The means of acquiring money is to work for some length of time at a given wage or to receive capital income from a given amount of investments; the income goes into the Stock of money. The means of acquiring goods and services is to purchase them at various prices, with the cost coming out of the Stock of money, and the goods and services going into the Stock of possessions (some very briefly). The Stock of possessions is depleted as they are used or as they depreciate.

These Stocks are integrators; the total amount of Stock is the cumulative sum over time of inputs (positive) and outputs (negative). This is very handy for control systems, because a control loop with all relationships but one being simple proportionalities, and the one exception being a time integration, is unconditionally stable. You can freely adjust its loop gain without worrying about oscillations or runaway (unless you get the sign of feedback wrong, which is easily fixed).

Finally I’m sleepy so that will have to do for now. I really want to know who is aboard for this. Let me know.

Best,

Bill P.

–
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[Bill_Powers1]

[From Bill Powers (2009.09.06.1229 MDT)]

Rick Marken (2009.09.06.1100)

···

–

[re Bill Powers (2009.09.05.2315 MDT)]–Again, missing from this (for me) is any description of the data
against which the behavior of the model will be evaluated.

Yes, it’s missing in a formal sense. However, as we put the model
together we’ll use informal understandings or placeholders at first just
to get a preliminary model that will run at all. Then I expect the
revisions will start and the model will be related more to real
data.

Here’s one way it could begin.

People need food, clothing, shelter, and other things. Since we can’t say
offhand what the list really is, we can just say that people need G,
where G is an array of goods and services. Money is a good like any
other.

type GoodsType = array[1…maxgoods] of record

NAME: string;

Perception, Reference, Error, Gain, Output,

UseRate, DepreciationRate,

Stock: double;

end;

Each item in the array is a record, which in turn is an arbitrary
collection of variables of different types, one collection per item in
the array. The above type definition says that there are
“maxgoods” items in the array, each item referring to one good.
The constant “maxgoods”, defined and given a value somewhere
else in the program, merely tells the compiler how much memory to reserve
to hold the largest array of records we plan to use. I’m not worried: I
just bought a one-terabyte external disk drive for my desktop computer,
for a shade under $100. One terabyte is a thousand gigabytes or a million
megabytes.

Let’s say that the NAME entry in G[1] is “Peanut Butter” and in
G[2] it is “Bank Account Balance.” Those are “string”
type variables, meaning words that you can read.

The next line in the definition of GoodsType is used to store the current
values of the control system variables and parameters. There is one
control system per good.

The next line lists effects of disturbances (per day) on the stock of
goods.

The final line before the end of the record, Stock, is the current
balance on hand of whatever the item is. We might also want to add a
parameter indicating how far negative the good can go. Money can go
negative (if you have credit) up to the credit limit. Material goods
can’t go negative – though perhaps you could say “I ow Jon Hancoc 3
pig”. You might also want to add interest rates, one that makes
G[2].stock increase (interest received), and another that makes the same
stock of money, if negative, go more negative (interest paid). The last
three lines of entries in the record list variables of type
“double,” meaning a double precision floating point
number.

If a consumer buys n units of G[1], this will increase G[1].stock by n,
and decrease the cash reserve G[2].stock by nG[1].price. In a model that
includes a producer, the same transaction would decrease the goods stock
of the particular producer who sold the item by n units, and increase the
cash stock (i.e., bank account) of that producer by nG[1].price units.
For those not familiar with programming symbols, * means multiply. We
don’t use x or X because that looks like the name of a variable and is
often used that way. “X times Y” would come out as “X X
Y,” which wouldn’t do. Note also that G[1] means the first record in
the array, and .Name (for example) indicates one of the sub-items that is
stored in that record. An array is just a list, a record is an arbitrary
collection of different types of variables.

Every time G[1].NAME is used by eating some of it, G[1].stock is reduced
by the amount eaten. If the item depreciates, its usable quantity is
reduced by G[1].stock * G[1].DepreciationRate every day, whether it’s
being used or not.

If the consumer works H hours per PayPeriod at wage W, every PayPeriod
days of iterations of the model (where one iteration represents one day)
the stock of money will increase by H*W dollars. For those unfamiliar
with this way of talking, when I say “PayPeriod” in that odd
way, I’m referring to a variable that gets specified somewhere else in
the program before this part of it is run; “PayPeriod” is the
name of the variable, and its magnitude would be set to 7 (days) if the
pay is distributed every week, or 28, 29, 30, or 31 if distributed every
month. The program would be adding a day to Daycount on every iteration,
and when DayCount equals PayPeriod, the pay would be calculated and
transferred, and Daycount would be reset to zero. PayPeriod can be
changed, too, if appropriate.

Very tedious details, but now we’re close to being able actually to run
this model and see how it behaves. We’d have to supply arbitrary starting
values of the stocks, the reference levels for the stocks, the gain, the
depreciation rates, the interest rates, the use rates, and so forth, but
the variables would begin to change right away and would come to some
equilibrium value after a few dozen iterations. Some auxiliary functions
would be needed to implement the control systems:

error := reference - perception;

output := gain*error;

Behind all this is the rest of the program, which does the iterations and
calculations over and over, perhaps plotting the values of the variables
on the screen so we can watch what is happening. I can suggest ways to
handle these details too for anyone trying to write the actual program
and run it. Rick Marken, Bruce Abbott, Martin Taylor, Richard Kennaway,
and perhaps some others who have been hiding to avoid being shanghaied
can help, too.

Note that we could also define an array of consumers, with the above
array of records existing within each consumer. It would be possible to
set up groups of consumers who have different reference levels for the
same goods, difference goods preferences, different incomes, and so
forth. This takes us closer to matching the model up with
reality.

This approach to modeling is very different from the way any economic
models I have seen are organized. It’s really a simulation like the
simulations of physical systems that engineers use all the time. The
model itself is simply an attempt to represent, quantitatively, the
characteristics of the system that really exists. There is no economic
theory in it. Any economic theory will emerge as a description of how the
model behaves, not as an explanation for why it behaves that way. Any
explanation runs in the other direction: the organization of the model
explains why it behaves economically as it does.

This is Sunday, tomorrow is Labor Day in the USA. I hope I hear from
participants, whether programmers, non-programmers, or spectators, by the
end of the coming week. I’m not going to continue this if this doesn’t
turn into a real project.

[rsmarken]

[From Rick Marken (2009.09.06.2135)]

Bill Powers (2009.09.06.1229 MDT)–

Rick Marken (2009.09.06.1100)

Again, missing from this (for me) is any description of the data
against which the behavior of the model will be evaluated.

Yes, it’s missing in a formal sense. However, as we put the model
together we’ll use informal understandings or placeholders at first just
to get a preliminary model that will run at all. Then I expect the
revisions will start and the model will be related more to real
data.

I would rather start with the data that’s available and then start thinking about what kind of model to build. I personally think the kind of model you describe below is way too detailed. I think your Dad’s model was closer to the right level, because it was framed in terms of variables that are, indeed, available as data. I would prefer to go in that direction but include things that were not in your Dad’s model (or my working implementation of it), like government spending, employment levels, etc. Maybe we can meet somewhere in the middle.

Best

Rick

···

Here’s one way it could begin.

People need food, clothing, shelter, and other things. Since we can’t say
offhand what the list really is, we can just say that people need G,
where G is an array of goods and services. Money is a good like any
other.

type GoodsType = array[1…maxgoods] of record

NAME: string;

Perception, Reference, Error, Gain, Output,

UseRate, DepreciationRate,

Stock: double;

end;

Each item in the array is a record, which in turn is an arbitrary
collection of variables of different types, one collection per item in
the array. The above type definition says that there are
“maxgoods” items in the array, each item referring to one good.
The constant “maxgoods”, defined and given a value somewhere
else in the program, merely tells the compiler how much memory to reserve
to hold the largest array of records we plan to use. I’m not worried: I
just bought a one-terabyte external disk drive for my desktop computer,
for a shade under $100. One terabyte is a thousand gigabytes or a million
megabytes.

Let’s say that the NAME entry in G[1] is “Peanut Butter” and in
G[2] it is “Bank Account Balance.” Those are “string”
type variables, meaning words that you can read.

The next line in the definition of GoodsType is used to store the current
values of the control system variables and parameters. There is one
control system per good.

The next line lists effects of disturbances (per day) on the stock of
goods.

The final line before the end of the record, Stock, is the current
balance on hand of whatever the item is. We might also want to add a
parameter indicating how far negative the good can go. Money can go
negative (if you have credit) up to the credit limit. Material goods
can’t go negative – though perhaps you could say “I ow Jon Hancoc 3
pig”. You might also want to add interest rates, one that makes
G[2].stock increase (interest received), and another that makes the same
stock of money, if negative, go more negative (interest paid). The last
three lines of entries in the record list variables of type
“double,” meaning a double precision floating point
number.

If a consumer buys n units of G[1], this will increase G[1].stock by n,
and decrease the cash reserve G[2].stock by nG[1].price. In a model that
includes a producer, the same transaction would decrease the goods stock
of the particular producer who sold the item by n units, and increase the
cash stock (i.e., bank account) of that producer by nG[1].price units.
For those not familiar with programming symbols, * means multiply. We
don’t use x or X because that looks like the name of a variable and is
often used that way. “X times Y” would come out as “X X
Y,” which wouldn’t do. Note also that G[1] means the first record in
the array, and .Name (for example) indicates one of the sub-items that is
stored in that record. An array is just a list, a record is an arbitrary
collection of different types of variables.

Every time G[1].NAME is used by eating some of it, G[1].stock is reduced
by the amount eaten. If the item depreciates, its usable quantity is
reduced by G[1].stock * G[1].DepreciationRate every day, whether it’s
being used or not.

If the consumer works H hours per PayPeriod at wage W, every PayPeriod
days of iterations of the model (where one iteration represents one day)
the stock of money will increase by H*W dollars. For those unfamiliar
with this way of talking, when I say “PayPeriod” in that odd
way, I’m referring to a variable that gets specified somewhere else in
the program before this part of it is run; “PayPeriod” is the
name of the variable, and its magnitude would be set to 7 (days) if the
pay is distributed every week, or 28, 29, 30, or 31 if distributed every
month. The program would be adding a day to Daycount on every iteration,
and when DayCount equals PayPeriod, the pay would be calculated and
transferred, and Daycount would be reset to zero. PayPeriod can be
changed, too, if appropriate.

Very tedious details, but now we’re close to being able actually to run
this model and see how it behaves. We’d have to supply arbitrary starting
values of the stocks, the reference levels for the stocks, the gain, the
depreciation rates, the interest rates, the use rates, and so forth, but
the variables would begin to change right away and would come to some
equilibrium value after a few dozen iterations. Some auxiliary functions
would be needed to implement the control systems:

error := reference - perception;

output := gain*error;

Behind all this is the rest of the program, which does the iterations and
calculations over and over, perhaps plotting the values of the variables
on the screen so we can watch what is happening. I can suggest ways to
handle these details too for anyone trying to write the actual program
and run it. Rick Marken, Bruce Abbott, Martin Taylor, Richard Kennaway,
and perhaps some others who have been hiding to avoid being shanghaied
can help, too.

Note that we could also define an array of consumers, with the above
array of records existing within each consumer. It would be possible to
set up groups of consumers who have different reference levels for the
same goods, difference goods preferences, different incomes, and so
forth. This takes us closer to matching the model up with
reality.

This approach to modeling is very different from the way any economic
models I have seen are organized. It’s really a simulation like the
simulations of physical systems that engineers use all the time. The
model itself is simply an attempt to represent, quantitatively, the
characteristics of the system that really exists. There is no economic
theory in it. Any economic theory will emerge as a description of how the
model behaves, not as an explanation for why it behaves that way. Any
explanation runs in the other direction: the organization of the model
explains why it behaves economically as it does.

This is Sunday, tomorrow is Labor Day in the USA. I hope I hear from
participants, whether programmers, non-programmers, or spectators, by the
end of the coming week. I’m not going to continue this if this doesn’t
turn into a real project.

–
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[Frank_Lenk]

[From Frank Lenk (2009.09.06.20:32 CDT]

Well, Bill, since you’ve thrown down the gauntlet, I guess
I need to reply.

Of course, I am interested in pursuing this project, but I am
approaching it differently. I am attaching a draft of my
dissertation proposal for you (and the list) to peruse. Frankly, I wasn’t
quite ready to do this – I have only just given it to my dissertation
chair for comment – but at this point in the project it’s probably
best to put all the ideas on the table where we can discuss them.
WARNING: About 2/3 of the proposal is actually a review of the urban
sociology literature. The problem I (ultimately) want to solve is concentrated
poverty among minorities in urban areas (affecting mostly African-Americans in
Kansas City, but also Latinos). The model I am hoping develop is merely a tool
for understanding more clearly the nature of problem and the effectiveness of
proposed solutions. I think PCT and agent-based modeling give me the best
chance of developing such a tool. I attempt to use the literature review to
derive model specifications for an “adequate” model of concentrated
urban minority poverty. These requirements are summarized beginning on page 46,
and it is from this point on that most of the discussion pertaining to modeling
can be found.

While I have thought a lot about this, it seems that every time
I read an email from Bill P. I learn something new that somewhat alters my
thinking, and so please read the draft with the idea that this is a work in
progress. However, I welcome all criticism – the point is to
develop a working prototype model by the end of my dissertation, one that I can
continue as part of my job after the dissertation is complete (you will see how
it is related to my work if you read the Problem Statement section). So
anything that gets me a better working model faster will be greatly
appreciated.

I should also say that, while I did not know him, Bill Williams
was a professor at my school, the University of Missouri-Kansas City (UMKC).
I was introduced to PCT by a colleague of his, Dr. Jim Sturgeon. The UMKC econ
department is one of the few heterodox economics schools in the U.S., which
means that it is critical of mainstream neoclassical economic theory (and like
PCT, has trouble getting recognized as legitimate by the keepers of orthodox
viewpoints). In being critical of its theory, many heterodox economists
are also very critical of its tools – especially mathematical and econometric
modeling. However, I learned my economic theory and econometrics as an
undergrad and masters student at Stanford and believe math and statistics
remain useful tools, but that sometimes we let what is easy to do with the tool
dictate our analysis, bending the analysis to the tool rather than the tool to real-world
observations. So we economists make unrealistic assumptions about human
behavior so we can get the math to be solvable, or we model things as a single
regression equation when there are actually multiple things happening
simultaneously that should be modeled as a simultaneous system of equations (Rick,
I suspect this is one reason why the sign on your correlation between taxes and
growth is positive). PCT and agent-based modeling seem to me to be a way out of
the box of having to trade accuracy of representation for tractability. From
what I understand from Dr. Sturgeon, Bill Williams’ views on the
usefulness of models were similar to mine. If so, I would be grateful if I am
able to continue his work in some small way.

The major difference in my approach from Bill P.’s, at
least as I see it, is that I am doing the modeling at a couple of levels below the
economy itself. I understand the need for simplicity – In trying to
explain what I am doing to others, I have often to said that I would be
satisfied if I produced decent model of a family, let alone an entire economy
or society. But in our search for starting simple, Bill and I are starting
in different places. I would like to model at the level of, not just
individuals, but individuals who interact with each other within families and
communities. I want a model that starts from their basic (intrinsic)
needs and has the economy emerge from their social interactions and the interactions
with the available technology (which I would have to specify).

As I said in my earlier post, one can view the economy as simply
a set of behaviors we use to control the amount of food in our bellies, subject
to certain environmental constraints. My conception of the environment includes
the social as well as the natural and the technological, however. I want
my agents to figure out how to organize themselves to live in the virtual world
inside my computer, much as we have had to figure out how to organize ourselves
to live in this one.

There are several reasons for my wanting to model at this
level. First, I think that the only human behavior we truly understand is
that which we experience, which is that of ourselves and small groups of others
with whom we regularly associate. This is the level at which we have any
expertise to specify our model. Our common, every-day experience provides
the data we need to begin, though ideally this would be checked against results
from psychology, anthropology and sociology.

Second, human behavior is always in reference to others. Under
all but the rarest of circumstances, there is no such thing as an isolated
individual living alone. We depend upon others for our most basic needs - for
food, for sex and parenting of offspring to reproduce. A
model that begins with its “atom” an isolated individual is likely to
produce a different outcome from one that specifies interactions between
individuals from the beginning (perhaps a mother-child relationship is the atom
of social models). Obviously, to get interacting individuals one has to
start by identifying, specifying and debugging the control systems inside a
single individual, but those control systems should be built to perceive others
and let others influence the individual’s references (learning by
emulation or mimicry, for example) and perceptions (or at least what is important
to perceive).

Third, and this is especially important given my problem of
concentrated minority poverty, once we allow others to influence either our
references or our perceptions (and probably both), the notion of Power rears
its head. Who influences whom and how? There is much evidence that, at
least partially, we set our references according to the behavior of those more
powerful than we. This is certainly true of a child. It also probably
accounts (again, at least partially) for the transmission of stereotypes, roles
and norms.

So in my proposed approach, the economy is behavior we engage in
to control the food, shelter and water we need to survive while we engage in
other behavior to control for the other things that are important (perhaps more
important) to us – mating, parenting, gaining social position or esteem.
Clearly, these purposes are inter-related – one who controls the
resources for food and shelter also becomes powerful and desirable as a mate
and a parent. Some (many?) will argue that gaining esteem or social
position is really all about maximizing potential for mating, so in this sense
is only a behavior to control for mating probability or frequency rather than a
reference. I tend to think that while this may have originally been the
case, humans and their antecedents have been social creatures long enough that
some hardwiring of an intrinsic need for belonging to a group or being valued
by the group is likely.

Reading through Bill’s approach again, I think it is clear
that it is simpler and more achievable more quickly than mine. But one of
the things I uncover in my survey of the agent-based modeling literature is
that the models are very sensitive to assumptions, especially about what the
rules are that the agents follow. In PCT terms, where do our references actually
come from? If they are simply assumptions made by the modeler, they will be
subject to criticism. Some are intrinsic, yes. But the rest? We must learn
them. We likely learn them as kids, both from experience and from others. So
let’s model the learning process in families and groups of families and
see what kind of economy we can derive from members’ efforts to satisfy
their needs.

Waiting for the “fun” to begin . . .

Frank

Frank Lenk

Director of Research Services

Mid-America Regional Council

600 Broadway, Suite 200

Kansas City, MO 64105

www.marc.org

816.474.4240

flenk@marc.org

816.701.8237

Lenk - Draft Dissertation Proposal - formatted.docx (796 KB)

···

From: Control Systems
Group Network (CSGnet) [mailto:CSGNET@LISTSERV.ILLINOIS.EDU] On Behalf Of Bill
Powers
Sent: Sunday, September 06, 2009 3:39 PM
To: CSGNET@LISTSERV.ILLINOIS.EDU
Subject: Re: [CSGNET] economics

[From Bill Powers (2009.09.06.1229 MDT)]

Rick Marken (2009.09.06.1100) –

[re Bill Powers (2009.09.05.2315
MDT)]–

Again, missing from this (for me) is any description of the
data against which the behavior of the model will be evaluated.

Yes, it’s missing in a formal sense. However, as we put the model together
we’ll use informal understandings or placeholders at first just to get a
preliminary model that will run at all. Then I expect the revisions will start
and the model will be related more to real data.

Here’s one way it could begin.

People need food, clothing, shelter, and other things. Since we can’t say offhand
what the list really is, we can just say that people need G, where G is an
array of goods and services. Money is a good like any other.

type GoodsType = array[1…maxgoods] of record

NAME: string;

Perception, Reference, Error, Gain, Output,

UseRate, DepreciationRate,

Stock: double;

end;

Each item in the array is a record, which in turn is an arbitrary collection of
variables of different types, one collection per item in the array. The above
type definition says that there are “maxgoods” items in the array,
each item referring to one good. The constant “maxgoods”, defined and
given a value somewhere else in the program, merely tells the compiler how much
memory to reserve to hold the largest array of records we plan to use. I’m not
worried: I just bought a one-terabyte external disk drive for my desktop
computer, for a shade under $100. One terabyte is a thousand gigabytes or a
million megabytes.

Let’s say that the NAME entry in G[1] is “Peanut Butter” and in G[2]
it is “Bank Account Balance.” Those are “string” type
variables, meaning words that you can read.

The next line in the definition of GoodsType is used to store the current
values of the control system variables and parameters. There is one control
system per good.

The next line lists effects of disturbances (per day) on the stock of goods.

The final line before the end of the record, Stock, is the current balance on
hand of whatever the item is. We might also want to add a parameter indicating
how far negative the good can go. Money can go negative (if you have credit) up
to the credit limit. Material goods can’t go negative – though perhaps you
could say “I ow Jon Hancoc 3 pig”. You might also want to add
interest rates, one that makes G[2].stock increase (interest received), and
another that makes the same stock of money, if negative, go more negative
(interest paid). The last three lines of entries in the record list variables
of type “double,” meaning a double precision floating point number.

If a consumer buys n units of G[1], this will increase G[1].stock by n, and
decrease the cash reserve G[2].stock by nG[1].price. In a model that includes
a producer, the same transaction would decrease the goods stock of the
particular producer who sold the item by n units, and increase the cash stock
(i.e., bank account) of that producer by nG[1].price units. For those not
familiar with programming symbols, * means multiply. We don’t use x or X because
that looks like the name of a variable and is often used that way. “X
times Y” would come out as “X X Y,” which wouldn’t do. Note also
that G[1] means the first record in the array, and .Name (for example)
indicates one of the sub-items that is stored in that record. An array is just
a list, a record is an arbitrary collection of different types of variables.

Every time G[1].NAME is used by eating some of it, G[1].stock is reduced by the
amount eaten. If the item depreciates, its usable quantity is reduced by
G[1].stock * G[1].DepreciationRate every day, whether it’s being used or not.

If the consumer works H hours per PayPeriod at wage W, every PayPeriod days of
iterations of the model (where one iteration represents one day) the stock of
money will increase by H*W dollars. For those unfamiliar with this way of
talking, when I say “PayPeriod” in that odd way, I’m referring to a
variable that gets specified somewhere else in the program before this part of
it is run; “PayPeriod” is the name of the variable, and its magnitude
would be set to 7 (days) if the pay is distributed every week, or 28, 29, 30,
or 31 if distributed every month. The program would be adding a day to Daycount
on every iteration, and when DayCount equals PayPeriod, the pay would be calculated
and transferred, and Daycount would be reset to zero. PayPeriod can be changed,
too, if appropriate.

Very tedious details, but now we’re close to being able actually to run this
model and see how it behaves. We’d have to supply arbitrary starting values of
the stocks, the reference levels for the stocks, the gain, the depreciation
rates, the interest rates, the use rates, and so forth, but the variables would
begin to change right away and would come to some equilibrium value after a few
dozen iterations. Some auxiliary functions would be needed to implement the
control systems:

error := reference - perception;
output := gain*error;

Behind all this is the rest of the program, which does the iterations and
calculations over and over, perhaps plotting the values of the variables on the
screen so we can watch what is happening. I can suggest ways to handle these
details too for anyone trying to write the actual program and run it. Rick
Marken, Bruce Abbott, Martin Taylor, Richard Kennaway, and perhaps some others
who have been hiding to avoid being shanghaied can help, too.

Note that we could also define an array of consumers, with the above array of
records existing within each consumer. It would be possible to set up groups of
consumers who have different reference levels for the same goods, difference
goods preferences, different incomes, and so forth. This takes us closer to
matching the model up with reality.

This approach to modeling is very different from the way any economic models I
have seen are organized. It’s really a simulation like the simulations of
physical systems that engineers use all the time. The model itself is simply an
attempt to represent, quantitatively, the characteristics of the system that
really exists. There is no economic theory in it. Any economic theory will
emerge as a description of how the model behaves, not as an explanation for why
it behaves that way. Any explanation runs in the other direction: the
organization of the model explains why it behaves economically as it does.

This is Sunday, tomorrow is Labor Day in the USA. I hope I hear from
participants, whether programmers, non-programmers, or spectators, by the end
of the coming week. I’m not going to continue this if this doesn’t turn into a
real project.

[Bill_Powers1]

[From Bill Powers (2009.09.07.0841 MDT)]

Rick Marken (2009.09.06.2135) –

RM: I would rather start with
the data that’s available and then start thinking about what kind of
model to build. I personally think the kind of model you describe below
is way too detailed.

BP: I have no objection to anyone’s thinking about higher levels of
control. My own preference is to lay the foundation first, then build on
it and with it.

RM: I think your Dad’s model was
closer to the right level, because it was framed in terms of variables
that are, indeed, available as data. I would prefer to go in that
direction but include things that were not in your Dad’s model (or my
working implementation of it), like government spending, employment
levels, etc. Maybe we can meet somewhere in the middle.

BP: I once proposed to my Dad that we join forces in looking at the
control phenomena in individual living systems in economics. His reaction
was to say rather energetically that individuals have nothing to do with
macroeconomics. I remember Bill Williams’ saying something similar on
many occasions; when he uttered the word “individual” it was
with a sneer. So you’re not alone in thinking my approach is way too
detailed.

My own opinion is that things like government spending and unemployment
levels are emergent phenomena, caused by what individual people want and
do. A simple example is what happens when two control systems, each
intent on achieving its own goals, come into conflict with each other. If
either system were a stimulus-response system or a passive system like a
rock, the result would not be conflict of the kind PCT is concerned with,
in which each side actively tries to overcome the other side. So to
understand human conflict, it’s necessary to understand how each
individual works. The social interaction called conflict grows out of the
characteristics of the individuals.

When we have a reasonably good model of individual economic behavior, we
can start to expand the economic model to include more than one
individual. For a long time I have had in the back of a dusty shelf a
model of bargaining, which looks as it it might work; the offers and
counteroffers either converge to a realistic-looking deal, or it becomes
apparent that no deal is possible. It would be interesting to set up
bargaining experiments to see how well this model fits real behavior.
Also, it’s possible to make predictions about what might happen if the
managers of businesses tended to adopt similar policies such as
outsourcing or laying off each other’s customers. Or we could ask what
will happen when it becomes fashionable among consumers to run negative
cash reserves.

A good model is not just a way of representing or describing observable
facts. It proposes an underlying set of entities and relationships that
predicts a certain set of observations, against which real observations
can be compared. Given that electrons exist and have certain masses,
charges, and spins, it follows that if you do things that should generate
ballistic electrons and magnetically deflect them as they fly by in a
vacuum, you should see a spot of light on the face of a cathode-ray tube
in a predictable position – which is all you ever see of the electrons.
If you start only with the spot of light observed on the screen, it’s
hard to think up a model at that level of description that would seem to
make the observation inevitable; you almost have to begin by positing
particles with particular individual properties. In fact, models build
out of nothing but observations were what gave rise to behaviorism, and
the conviction that there is no need to ask how a behaving system
works inside.

Even in your own models, we’ve have some disagreements about what should
be represented as a control system. I believe you proposed a reference
level for GDP, using the idea of a reference signal as something a social
system could have, as opposed to my idea that reference signals exist
only in individuals, with very few individuals in a position to control
GDP. In my modeling approach, I would treat GDP as a side-effect of what
people are really controlling. Would the result be very different? I
don’t know, but my sense of reality about the model is pretty strongly
affected by which interpretation is used.

I take it, then, that you are not aboard for the modeling effort I’m
proposing. We shall see who reaches the finish line first, the tortoise
or the hare.

Best,

Bill P.

···

[Bill_Powers1]

[From Bill Powers (2009.09.07.0945 MDT)]

Frank Lenk (2009.09.06.20:32 CDT] –

FL: Well, Bill, since you’ve
thrown down the gauntlet, I guess I need to reply.

WARNING: About 2/3 of the
proposal is actually a review of the urban sociology literature. The
problem I (ultimately) want to solve is concentrated poverty among
minorities in urban areas (affecting mostly African-Americans in Kansas
City, but also Latinos). The model I am hoping develop is merely a tool
for understanding more clearly the nature of problem and the
effectiveness of proposed solutions. I think PCT and agent-based modeling
give me the best chance of developing such a tool.

BP: I think so, too. I hope you’re aware of the recent book by McClelland
and Ferraro, “Purpose, Meaning, and Action,” which is about
using PCT in sociology.

[
http://www.amazon.com/Purpose-Meaning-Action-Theories-Sociology/dp/1403967989

](http://www.amazon.com/Purpose-Meaning-Action-Theories-Sociology/dp/1403967989)I’ll go over the draft dissertation in more detail, but there are a
couple of points to make.

The main one for now is the idea that reference levels originate in the
social system. I think this is a simple mistake, and one reason why the
model of individuals has to be kept in mind. A reference level is simply
a particular state of a perception that has been selected as a target or
goal. A cruise control works that way: when the driver pushes a button,
the current speedometer reading is stored as a reference signal, which
then serves to specify the speedometer reading that is to be maintained.
The driver, a higher-order control system, pushes the button when
satisfied with the current speed.

Each person has perceptions of various aspects of society. Society exists
only as the set of all individual perceptions of it (where else could it
be?). People may learn to use similar words to refer to social variables
(like “respect”), but those words often point to very different
perceptions.

From interacting with people, individuals determine which results of
social interactions they like or don’t like, and set their reference
levels for the social variables they perceive accordingly. They control
those social variables (or try to) according to the way they see them,
not the way other people see them. The mistake old-time sociologists made
was to assume that the society they saw was the same one that the
individuals in it saw. I think your model might benefit from
acknowledging that what a suburban researcher sees as “urban core
decline” might be quite different from what a resident of the urban
core sees. This might make the behavior of such a resident more
comprehensible to the researcher. It might be that the resident sees the
“decline” as part of someone’s effort – maybe his own – to
make life better.

FL: I should also say that,
while I did not know him, Bill Williams was a professor at my school, the
University of Missouri-Kansas City (UMKC). I was introduced to PCT by a
colleague of his, Dr. Jim Sturgeon.

BP: I remember Bill’s talking about Jim Sturgeon, who helped Bill achieve
recognition that he despaired of ever getting. It’s good to know that
Sturgeon was spreading the word about PCT, too.

FL: The major difference in my
approach from Bill P.’s, at least as I see it, is that I am doing the
modeling at a couple of levels below the economy itself. I understand the
need for simplicity – In trying to explain what I am doing to others, I
have often to said that I would be satisfied if I produced decent model
of a family, let alone an entire economy or society. But in our
search for starting simple, Bill and I are starting in different
places. I would like to model at the level of, not just
individuals, but individuals who interact with each other within families
and communities. I want a model that starts from their basic
(intrinsic) needs and has the economy emerge from their social
interactions and the interactions with the available technology (which I
would have to specify).

That’s actually the same level I want to get to, but I’m starting with a
model that initially takes into account only the mechanics of the
economy: buying and selling, working and earning. The next step is a
similarly low-level model of the producers and their plants. All
agent-based, of course – nothing happens just because it happens. Then I
would start getting into questions of why the reference levels for the
lower-level perceptions are set one way rather than another. Because I am
not an economist, I have to take small steps, and would expect that at
some stage real economists would get impatient and say “Move over,
let me do that.” I would happily comply.

As I said in my earlier post,
one can view the economy as simply a set of behaviors we use to control
the amount of food in our bellies, subject to certain environmental
constraints. My conception of the environment includes the social as well
as the natural and the technological, however. I want my agents to
figure out how to organize themselves to live in the virtual world inside
my computer, much as we have had to figure out how to organize ourselves
to live in this one.

You’re just talking about higher-level organizations than I am starting
with. To me, social aspects of the economy boil down to what individuals
think the society is and should be, which in turn come from their
experiences with it (including what others tell them about it). The only
way these higher-level systems can work, however, is through adjusting
the reference levels for the systems I am trying to model first.

There are several reasons for my
wanting to model at this level. First, I think that the only human
behavior we truly understand is that which we experience, which is that
of ourselves and small groups of others with whom we regularly associate.
This is the level at which we have any expertise to specify our
model. Our common, every-day experience provides the data we need
to begin, though ideally this would be checked against results from
psychology, anthropology and sociology.

OK, I agree that families are a good place to start, being a finite unit
with not too many interactions. In my approach families would enter
initially only as dependents – drains on the store of goods and services
which breadwinners work to pay for. I don’t mean they’re resented for
that – that’s a matter for higher levels of control. I just mean they
have that effect.

Second, human behavior is always
in reference to others. Under all but the rarest of circumstances, there
is no such thing as an isolated individual living alone. We depend upon
others for our most basic needs - for food, for sex and parenting
of offspring to reproduce. A model that begins with its
“atom” an isolated individual is likely to produce a different outcome
from one that specifies interactions between individuals from the
beginning (perhaps a mother-child relationship is the atom of social
models).

Don’t worry about that part of it, it will be added once the foundation
is laid. At the level where I’m considering the model now, reference
signals are simply variables and we can look at the consequences of
adjusting them. We can introduce more people, and see the consequences of
different combinations of reference levels being sought in a common
environment. This doesn’t commit us to any particular reference settings;
it just says that if you set them this way, that is what
will happen. Then, of course, we have to try to test that prediction with
data.

Obviously, to get
interacting individuals one has to start by identifying, specifying and
debugging the control systems inside a single individual, but those
control systems should be built to perceive others and let others
influence the individual’s references (learning by emulation or mimicry,
for example) and perceptions (or at least what is important to
perceive).

Yes, and this is my intention once the lower levels are taken care of.
Though I think references are acquired in a way somewhat different from
what you assume (I maintain that children learn perceptions and
references mainly by watching what adults do and interacting with them,
not by listening to what they say). And anyway, we don’t just “learn
references.” We learn to adjust references in real time for lower
systems as a means of controlling higher-order variables. Reference
signals are normally variable according to actions initiated at higher
levels and disturbances that tend to alter higher-level perceptions. We
don’t just “set and forget” references.

Third, and this is especially
important given my problem of concentrated minority poverty, once we
allow others to influence either our references or our perceptions (and
probably both), the notion of Power rears its head. Who influences whom
and how? There is much evidence that, at least partially, we set
our references according to the behavior of those more powerful than
we. This is certainly true of a child. It also probably accounts
(again, at least partially) for the transmission of stereotypes, roles
and norms.

I think our references and perceptions grow out of social intereactions
not by imitation so much as by resisting disturbances. What we take away
from social interactions is mainly what we learned to do to maintain
control, not what we saw the other person doing. In some cases, of
course, a child sees something another person is doing, and says “I
could do that!” But the actual learning comes from the doing, not
the watching. A great deal of what is learned (for example, from being
abused) is what one has to learn to maintain control of one’s own
life.

So in my proposed approach, the
economy is behavior we engage in to control the food, shelter and water
we need to survive while we engage in other behavior to control for the
other things that are important (perhaps more important) to us – mating,
parenting, gaining social position or esteem. Clearly, these
purposes are inter-related – one who controls the resources for food and
shelter also becomes powerful and desirable as a mate and a parent.
Some (many?) will argue that gaining esteem or social position is really
all about maximizing potential for mating, so in this sense is only a
behavior to control for mating probability or frequency rather than a
reference. I tend to think that while this may have originally been
the case, humans and their antecedents have been social creatures long
enough that some hardwiring of an intrinsic need for belonging to a group
or being valued by the group is likely.

Isn’t it simpler to assume that what is hardwired is a desire for certain
consequences of belonging or being valued? Those desires are satisfied in
different people by different behaviors, sometimes by opposite behaviors
– which accounts for the low correlations one gets in any study aimed at
finding universal behavior patterns. Some people like being with other
people all the time; some people like a lot of time alone. It’s not the
behaviors we care about, it’s the consequences they produce. Some might
say that people learn to make fires in order to imitate others who made
fires. I say they learned to make fires because they want to stay
warm.

Reading through Bill’s approach
again, I think it is clear that it is simpler and more achievable more
quickly than mine. But one of the things I uncover in my survey of
the agent-based modeling literature is that the models are very sensitive
to assumptions, especially about what the rules are that the agents
follow.

Yes, indeed. And those rules are variables or adjustable features in my
model. I check out rules to see how they work, but anyone can plug any
rules they like into the same model to see what happens. For example, in
the model I sent you, a plant manager keeps inventories at a fixed level
by adjusting prices. That works quite well and gives a reasonable law of
supply and demand, but maybe there are other ways of doing the same
thing. You can substitute the other ways, or you can add more plants and
managers and see what happens when different managers work according to
different rules. Maybe some rules lead to going out of business. Others
may generate conflicts.

In PCT terms, where do our
references actually come from? If they are simply assumptions made by the
modeler, they will be subject to criticism. Some are intrinsic, yes. But
the rest? We must learn them. We likely learn them as kids, both from
experience and from others. So let’s model the learning process in
families and groups of families and see what kind of economy we can
derive from members’ efforts to satisfy their needs.

Fine, the reorganization algorithm has been checked out in systems of up
to 500 interacting control systems. It ought to work with four or five
people controlling 10 or 20 variables. But you do need a starting point
– an underlying architecture of control systems to provide parameters
that can be reorganized.

My strategy in general is to model control systems and leave their
detailed properties to be determined. Where I assume reference levels to
test the model, I want to substitute higher-order control systems to
adjust them once the lower level is working right. Where I assume
particular controlled variables, anyone is free to propose other
controlled variables. Ditto for means of action. The only reason for
making particular assumptions about the lower-level systems is to make
sure they will work, and to learn what they can do.

Best,

Bill P.

[rsmarken]

[From Rick Marken (2009.09.07.1210)]

Bill Powers (2009.09.07.0841 MDT)

Rick Marken (2009.09.06.2135) –

BP: I once proposed to my Dad that we join forces in looking at the
control phenomena in individual living systems in economics. His reaction
was to say rather energetically that individuals have nothing to do with
macroeconomics. I remember Bill Williams’ saying something similar on
many occasions; when he uttered the word “individual” it was
with a sneer. So you’re not alone in thinking my approach is way too
detailed.

Not me. My analysis of economics starts with the individual.

My own opinion is that things like government spending and unemployment
levels are emergent phenomena, caused by what individual people want and
do.

Yes, I feel the same way.

A good model is not just a way of representing or describing observable
facts. It proposes an underlying set of entities and relationships that
predicts a certain set of observations, against which real observations
can be compared.

I’m right there with you!

In fact, models build
out of nothing but observations were what gave rise to behaviorism, and
the conviction that there is no need to ask how a behaving system
works inside.

Yes, I’m into models built out of something.

Even in your own models, we’ve have some disagreements about what should
be represented as a control system. I believe you proposed a reference
level for GDP, using the idea of a reference signal as something a social
system could have, as opposed to my idea that reference signals exist
only in individuals, with very few individuals in a position to control
GDP. In my modeling approach, I would treat GDP as a side-effect of what
people are really controlling. Would the result be very different? I
don’t know, but my sense of reality about the model is pretty strongly
affected by which interpretation is used.

I think it’s more a question of level of resolution than of interpretation. There is always a question about resolution when we build models. Even in our simple tracking models we are making compromises about resolution. We, for example, do not try to write physiologically plausible models of the perceptual and output functions (particularly the former) in our models. In our models of tracking, for example, we often write p = c, reducing the perceptual function to a multiplication by 1. In terms of the economic modeling, we will both have to compromise on resolution. Looking at GDP as a controlled variable may be too much of a compromise for you; I see it as a virtual controlled variable with a virtual reference representing the average reference for all goods and services for the entire population. You may want a higher degree of resolution, looking for example, at the references for several different goods and services for a few individuals, but even then you are reducing resolution; for example, you are not going to consider all goods and services in the real economy and you will not have the references for the gods/service in the model vary differently over time, as they certainly do in reality. So it’s really, I think, a question of what is a sensible level of resolution; the reason I prefer my level of resolution is because there are aggregate data measures available against which to test the behavior of the model.

I take it, then, that you are not aboard for the modeling effort I’m
proposing. We shall see who reaches the finish line first, the tortoise
or the hare.

We shall see. Yes. Frankly, I hope you win because I currently don’t even know what winning would look like. I think of winning as a model that can predict data. I think you think of winning as a model that can just “work” when agents controlling in the context of known rules of exchange of goods and money are implemented as a computer program. Anyway, I wish you and your team the very best of luck because a “win” for you would certainly be a win for me.

Best

Rick

···

–
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[Shannon_Williams]

(Shannon Williams 2009.09.07 15:45 CDT)

Several times in the text below Bill focuses on behavior that is chosen to achieve a well-defined goal. (for example - making fire to keep warm, or ‘selecting references’ for some conscious objective) But the vast majority of references are not consciously chosen. They develop simply because the neurons are exposed to those perceptions. If you want to know why certain people ‘chose’ to remain in an environment or continuously recreate the same environments wherever they go then realize that the vast majority of references serve simply to do this.

Thanks,

Shannon

···

Sent from my iPhone

On Sep 7, 2009, at 1:17 PM, Bill Powers powers_w@FRONTIER.NET wrote:

[From Bill Powers (2009.09.07.0945 MDT)]

Frank Lenk (2009.09.06.20:32 CDT] –

FL: Well, Bill, since you’ve
thrown down the gauntlet, I guess I need to reply.

WARNING: About 2/3 of the
proposal is actually a review of the urban sociology literature. The
problem I (ultimately) want to solve is concentrated poverty among
minorities in urban areas (affecting mostly African-Americans in Kansas
City, but also Latinos). The model I am hoping develop is merely a tool
for understanding more clearly the nature of problem and the
effectiveness of proposed solutions. I think PCT and agent-based modeling
give me the best chance of developing such a tool.

BP: I think so, too. I hope you’re aware of the recent book by McClelland
and Ferraro, “Purpose, Meaning, and Action,” which is about
using PCT in sociology.

http://www.amazon.com/Purpose-Meaning-Action-Theories-Sociology/dp/1403967989

I’ll go over the draft dissertation in more detail, but there are a
couple of points to make.

The main one for now is the idea that reference levels originate in the
social system. I think this is a simple mistake, and one reason why the
model of individuals has to be kept in mind. A reference level is simply
a particular state of a perception that has been selected as a target or
goal. A cruise control works that way: when the driver pushes a button,
the current speedometer reading is stored as a reference signal, which
then serves to specify the speedometer reading that is to be maintained.
The driver, a higher-order control system, pushes the button when
satisfied with the current speed.

Each person has perceptions of various aspects of society. Society exists
only as the set of all individual perceptions of it (where else could it
be?). People may learn to use similar words to refer to social variables
(like “respect”), but those words often point to very different
perceptions.

From interacting with people, individuals determine which results of
social interactions they like or don’t like, and set their reference
levels for the social variables they perceive accordingly. They control
those social variables (or try to) according to the way they see them,
not the way other people see them. The mistake old-time sociologists made
was to assume that the society they saw was the same one that the
individuals in it saw. I think your model might benefit from
acknowledging that what a suburban researcher sees as “urban core
decline” might be quite different from what a resident of the urban
core sees. This might make the behavior of such a resident more
comprehensible to the researcher. It might be that the resident sees the
“decline” as part of someone’s effort – maybe his own – to
make life better.

FL: I should also say that,
while I did not know him, Bill Williams was a professor at my school, the
University of Missouri-Kansas City (UMKC). I was introduced to PCT by a
colleague of his, Dr. Jim Sturgeon.

BP: I remember Bill’s talking about Jim Sturgeon, who helped Bill achieve
recognition that he despaired of ever getting. It’s good to know that
Sturgeon was spreading the word about PCT, too.

FL: The major difference in my
approach from Bill P.’s, at least as I see it, is that I am doing the
modeling at a couple of levels below the economy itself. I understand the
need for simplicity – In trying to explain what I am doing to others, I
have often to said that I would be satisfied if I produced decent model
of a family, let alone an entire economy or society. But in our
search for starting simple, Bill and I are starting in different
places. I would like to model at the level of, not just
individuals, but individuals who interact with each other within families
and communities. I want a model that starts from their basic
(intrinsic) needs and has the economy emerge from their social
interactions and the interactions with the available technology (which I
would have to specify).

That’s actually the same level I want to get to, but I’m starting with a
model that initially takes into account only the mechanics of the
economy: buying and selling, working and earning. The next step is a
similarly low-level model of the producers and their plants. All
agent-based, of course – nothing happens just because it happens. Then I
would start getting into questions of why the reference levels for the
lower-level perceptions are set one way rather than another. Because I am
not an economist, I have to take small steps, and would expect that at
some stage real economists would get impatient and say “Move over,
let me do that.” I would happily comply.

As I said in my earlier post,
one can view the economy as simply a set of behaviors we use to control
the amount of food in our bellies, subject to certain environmental
constraints. My conception of the environment includes the social as well
as the natural and the technological, however. I want my agents to
figure out how to organize themselves to live in the virtual world inside
my computer, much as we have had to figure out how to organize ourselves
to live in this one.

You’re just talking about higher-level organizations than I am starting
with. To me, social aspects of the economy boil down to what individuals
think the society is and should be, which in turn come from their
experiences with it (including what others tell them about it). The only
way these higher-level systems can work, however, is through adjusting
the reference levels for the systems I am trying to model first.

There are several reasons for my
wanting to model at this level. First, I think that the only human
behavior we truly understand is that which we experience, which is that
of ourselves and small groups of others with whom we regularly associate.
This is the level at which we have any expertise to specify our
model. Our common, every-day experience provides the data we need
to begin, though ideally this would be checked against results from
psychology, anthropology and sociology.

OK, I agree that families are a good place to start, being a finite unit
with not too many interactions. In my approach families would enter
initially only as dependents – drains on the store of goods and services
which breadwinners work to pay for. I don’t mean they’re resented for
that – that’s a matter for higher levels of control. I just mean they
have that effect.

Second, human behavior is always
in reference to others. Under all but the rarest of circumstances, there
is no such thing as an isolated individual living alone. We depend upon
others for our most basic needs - for food, for sex and parenting
of offspring to reproduce. A model that begins with its
“atomâ€? an isolated individual is likely to produce a different outcome
from one that specifies interactions between individuals from the
beginning (perhaps a mother-child relationship is the atom of social
models).

Don’t worry about that part of it, it will be added once the foundation
is laid. At the level where I’m considering the model now, reference
signals are simply variables and we can look at the consequences of
adjusting them. We can introduce more people, and see the consequences of
different combinations of reference levels being sought in a common
environment. This doesn’t commit us to any particular reference settings;
it just says that if you set them this way, that is what
will happen. Then, of course, we have to try to test that prediction with
data.

Obviously, to get
interacting individuals one has to start by identifying, specifying and
debugging the control systems inside a single individual, but those
control systems should be built to perceive others and let others
influence the individual’s references (learning by emulation or mimicry,
for example) and perceptions (or at least what is important to
perceive).

Yes, and this is my intention once the lower levels are taken care of.
Though I think references are acquired in a way somewhat different from
what you assume (I maintain that children learn perceptions and
references mainly by watching what adults do and interacting with them,
not by listening to what they say). And anyway, we don’t just “learn
references.” We learn to adjust references in real time for lower
systems as a means of controlling higher-order variables. Reference
signals are normally variable according to actions initiated at higher
levels and disturbances that tend to alter higher-level perceptions. We
don’t just “set and forget” references.

Third, and this is especially
important given my problem of concentrated minority poverty, once we
allow others to influence either our references or our perceptions (and
probably both), the notion of Power rears its head. Who influences whom
and how? There is much evidence that, at least partially, we set
our references according to the behavior of those more powerful than
we. This is certainly true of a child. It also probably accounts
(again, at least partially) for the transmission of stereotypes, roles
and norms.

I think our references and perceptions grow out of social intereactions
not by imitation so much as by resisting disturbances. What we take away
from social interactions is mainly what we learned to do to maintain
control, not what we saw the other person doing. In some cases, of
course, a child sees something another person is doing, and says “I
could do that!” But the actual learning comes from the doing, not
the watching. A great deal of what is learned (for example, from being
abused) is what one has to learn to maintain control of one’s own
life.

So in my proposed approach, the
economy is behavior we engage in to control the food, shelter and water
we need to survive while we engage in other behavior to control for the
other things that are important (perhaps more important) to us – mating,
parenting, gaining social position or esteem. Clearly, these
purposes are inter-related – one who controls the resources for food and
shelter also becomes powerful and desirable as a mate and a parent.
Some (many?) will argue that gaining esteem or social position is really
all about maximizing potential for mating, so in this sense is only a
behavior to control for mating probability or frequency rather than a
reference. I tend to think that while this may have originally been
the case, humans and their antecedents have been social creatures long
enough that some hardwiring of an intrinsic need for belonging to a group
or being valued by the group is likely.

Isn’t it simpler to assume that what is hardwired is a desire for certain
consequences of belonging or being valued? Those desires are satisfied in
different people by different behaviors, sometimes by opposite behaviors
– which accounts for the low correlations one gets in any study aimed at
finding universal behavior patterns. Some people like being with other
people all the time; some people like a lot of time alone. It’s not the
behaviors we care about, it’s the consequences they produce. Some might
say that people learn to make fires in order to imitate others who made
fires. I say they learned to make fires because they want to stay
warm.

Reading through Bill’s approach
again, I think it is clear that it is simpler and more achievable more
quickly than mine. But one of the things I uncover in my survey of
the agent-based modeling literature is that the models are very sensitive
to assumptions, especially about what the rules are that the agents
follow.

Yes, indeed. And those rules are variables or adjustable features in my
model. I check out rules to see how they work, but anyone can plug any
rules they like into the same model to see what happens. For example, in
the model I sent you, a plant manager keeps inventories at a fixed level
by adjusting prices. That works quite well and gives a reasonable law of
supply and demand, but maybe there are other ways of doing the same
thing. You can substitute the other ways, or you can add more plants and
managers and see what happens when different managers work according to
different rules. Maybe some rules lead to going out of business. Others
may generate conflicts.

In PCT terms, where do our
references actually come from? If they are simply assumptions made by the
modeler, they will be subject to criticism. Some are intrinsic, yes. But
the rest? We must learn them. We likely learn them as kids, both from
experience and from others. So let’s model the learning process in
families and groups of families and see what kind of economy we can
derive from members’ efforts to satisfy their needs.

Fine, the reorganization algorithm has been checked out in systems of up
to 500 interacting control systems. It ought to work with four or five
people controlling 10 or 20 variables. But you do need a starting point
– an underlying architecture of control systems to provide parameters
that can be reorganized.

My strategy in general is to model control systems and leave their
detailed properties to be determined. Where I assume reference levels to
test the model, I want to substitute higher-order control systems to
adjust them once the lower level is working right. Where I assume
particular controlled variables, anyone is free to propose other
controlled variables. Ditto for means of action. The only reason for
making particular assumptions about the lower-level systems is to make
sure they will work, and to learn what they can do.

Best,

Bill P.

[Frank_Lenk]

[From Frank Lenk (2009.09.08.2121 CDT)]

Bill – thank you for the very thoughtful reply. A
complete response will have to wait a while as I have attend to several
projects at work. Thanks especially for the McClelland and Ferraro
reference. I will have to get it. I appreciate your attempts to get me to
start simpler, as I am learning that perfect is the enemy of good, and much of
what we learn at the “good” stage will probably invalidate much
what we thought was required for perfect. In other words, getting
something running quickly will lead to quicker debugging of ideas as well as
software.

I will think on what you have written, and reply more fully when
I get the chance.

Frank

···

From: Control Systems
Group Network (CSGnet) [mailto:CSGNET@LISTSERV.ILLINOIS.EDU] On Behalf Of Bill
Powers
Sent: Monday, September 07, 2009 1:18 PM
To: CSGNET@LISTSERV.ILLINOIS.EDU
Subject: Re: [CSGNET] economics

[From Bill Powers (2009.09.07.0945 MDT)]

Frank Lenk (2009.09.06.20:32 CDT] –

FL: Well, Bill, since you’ve thrown down the gauntlet,
I guess I need to reply.

WARNING: About 2/3 of the proposal is actually a
review of the urban sociology literature. The problem I (ultimately) want to
solve is concentrated poverty among minorities in urban areas (affecting mostly
African-Americans in Kansas City, but also Latinos). The model I am hoping develop
is merely a tool for understanding more clearly the nature of problem and the
effectiveness of proposed solutions. I think PCT and agent-based modeling give
me the best chance of developing such a tool.

BP: I think so, too. I hope you’re aware of the recent book by McClelland and
Ferraro, “Purpose, Meaning, and Action,” which is about using PCT in
sociology.

[http://www.amazon.com/Purpose-Meaning-Action-Theories-Sociology/dp/1403967989

](http://www.amazon.com/Purpose-Meaning-Action-Theories-Sociology/dp/1403967989)I’ll go over the draft dissertation in more detail, but there are a couple
of points to make.

The main one for now is the idea that reference levels originate in the social
system. I think this is a simple mistake, and one reason why the model of
individuals has to be kept in mind. A reference level is simply a particular
state of a perception that has been selected as a target or goal. A cruise
control works that way: when the driver pushes a button, the current
speedometer reading is stored as a reference signal, which then serves to
specify the speedometer reading that is to be maintained. The driver, a
higher-order control system, pushes the button when satisfied with the current
speed.

Each person has perceptions of various aspects of society. Society exists only
as the set of all individual perceptions of it (where else could it be?).
People may learn to use similar words to refer to social variables (like
“respect”), but those words often point to very different
perceptions.

from interacting with people, individuals determine which results of social
interactions they like or don’t like, and set their reference levels for the
social variables they perceive accordingly. They control those social variables
(or try to) according to the way they see them, not the way other people see
them. The mistake old-time sociologists made was to assume that the society
they saw was the same one that the individuals in it saw. I think your model
might benefit from acknowledging that what a suburban researcher sees as
“urban core decline” might be quite different from what a resident of
the urban core sees. This might make the behavior of such a resident more
comprehensible to the researcher. It might be that the resident sees the
“decline” as part of someone’s effort – maybe his own – to make
life better.

FL: I should also say that, while I did not know him, Bill
Williams was a professor at my school, the University of Missouri-Kansas City
(UMKC). I was introduced to PCT by a colleague of his, Dr. Jim Sturgeon.

BP: I remember Bill’s talking about Jim Sturgeon, who helped Bill achieve
recognition that he despaired of ever getting. It’s good to know that Sturgeon
was spreading the word about PCT, too.

FL: The major difference in my approach from Bill
P.’s, at least as I see it, is that I am doing the modeling at a couple
of levels below the economy itself. I understand the need for simplicity
– In trying to explain what I am doing to others, I have often to said
that I would be satisfied if I produced decent model of a family, let alone an
entire economy or society. But in our search for starting simple, Bill
and I are starting in different places. I would like to model at the level
of, not just individuals, but individuals who interact with each other within
families and communities. I want a model that starts from their basic
(intrinsic) needs and has the economy emerge from their social interactions and
the interactions with the available technology (which I would have to specify).

That’s actually the same level I want to get to, but I’m starting with a model
that initially takes into account only the mechanics of the economy: buying and
selling, working and earning. The next step is a similarly low-level model of
the producers and their plants. All agent-based, of course – nothing happens
just because it happens. Then I would start getting into questions of why the
reference levels for the lower-level perceptions are set one way rather than
another. Because I am not an economist, I have to take small steps, and would
expect that at some stage real economists would get impatient and say
“Move over, let me do that.” I would happily comply.

As I said in my earlier post, one can view the economy as
simply a set of behaviors we use to control the amount of food in our bellies,
subject to certain environmental constraints. My conception of the environment
includes the social as well as the natural and the technological, however.
I want my agents to figure out how to organize themselves to live in the
virtual world inside my computer, much as we have had to figure out how to
organize ourselves to live in this one.

You’re just talking about higher-level organizations than I am starting with.
To me, social aspects of the economy boil down to what individuals think the
society is and should be, which in turn come from their experiences with it
(including what others tell them about it). The only way these higher-level
systems can work, however, is through adjusting the reference levels for the
systems I am trying to model first.

There are several reasons for my wanting to model at this
level. First, I think that the only human behavior we truly understand is
that which we experience, which is that of ourselves and small groups of others
with whom we regularly associate. This is the level at which we have any
expertise to specify our model. Our common, every-day experience provides
the data we need to begin, though ideally this would be checked against results
from psychology, anthropology and sociology.

OK, I agree that families are a good place to start, being a finite unit with
not too many interactions. In my approach families would enter initially only
as dependents – drains on the store of goods and services which breadwinners
work to pay for. I don’t mean they’re resented for that – that’s a matter for
higher levels of control. I just mean they have that effect.

Second, human behavior is always in reference to others.
Under all but the rarest of circumstances, there is no such thing as an
isolated individual living alone. We depend upon others for our most basic
needs - for food, for sex and parenting of offspring to
reproduce. A model that begins with its “atom” an
isolated individual is likely to produce a different outcome from one that
specifies interactions between individuals from the beginning (perhaps a
mother-child relationship is the atom of social models).

Don’t worry about that part of it, it will be added once the foundation is
laid. At the level where I’m considering the model now, reference signals are
simply variables and we can look at the consequences of adjusting them. We can
introduce more people, and see the consequences of different combinations of
reference levels being sought in a common environment. This doesn’t commit us
to any particular reference settings; it just says that if you set them this
way, that is what will happen. Then, of course, we have to try to test
that prediction with data.

Obviously, to get interacting individuals one has to
start by identifying, specifying and debugging the control systems inside a
single individual, but those control systems should be built to perceive others
and let others influence the individual’s references (learning by
emulation or mimicry, for example) and perceptions (or at least what is
important to perceive).

Yes, and this is my intention once the lower levels are taken care of. Though I
think references are acquired in a way somewhat different from what you assume
(I maintain that children learn perceptions and references mainly by watching
what adults do and interacting with them, not by listening to what they say).
And anyway, we don’t just “learn references.” We learn to adjust
references in real time for lower systems as a means of controlling
higher-order variables. Reference signals are normally variable according to
actions initiated at higher levels and disturbances that tend to alter
higher-level perceptions. We don’t just “set and forget” references.

Third, and this is especially important given my problem of
concentrated minority poverty, once we allow others to influence either our
references or our perceptions (and probably both), the notion of Power rears
its head. Who influences whom and how? There is much evidence that, at
least partially, we set our references according to the behavior of those more
powerful than we. This is certainly true of a child. It also probably
accounts (again, at least partially) for the transmission of stereotypes, roles
and norms.

I think our references and perceptions grow out of social intereactions not by
imitation so much as by resisting disturbances. What we take away from social
interactions is mainly what we learned to do to maintain control, not what we
saw the other person doing. In some cases, of course, a child sees something
another person is doing, and says “I could do that!” But the actual
learning comes from the doing, not the watching. A great deal of what is learned
(for example, from being abused) is what one has to learn to maintain control
of one’s own life.

So in my proposed approach, the economy is behavior we
engage in to control the food, shelter and water we need to survive while we
engage in other behavior to control for the other things that are important
(perhaps more important) to us – mating, parenting, gaining social
position or esteem. Clearly, these purposes are inter-related – one
who controls the resources for food and shelter also becomes powerful and desirable
as a mate and a parent. Some (many?) will argue that gaining esteem or
social position is really all about maximizing potential for mating, so in this
sense is only a behavior to control for mating probability or frequency rather
than a reference. I tend to think that while this may have originally
been the case, humans and their antecedents have been social creatures long
enough that some hardwiring of an intrinsic need for belonging to a group or
being valued by the group is likely.

Isn’t it simpler to assume that what is hardwired is a desire for certain
consequences of belonging or being valued? Those desires are satisfied in
different people by different behaviors, sometimes by opposite behaviors –
which accounts for the low correlations one gets in any study aimed at finding
universal behavior patterns. Some people like being with other people all the
time; some people like a lot of time alone. It’s not the behaviors we care
about, it’s the consequences they produce. Some might say that people learn to
make fires in order to imitate others who made fires. I say they learned to
make fires because they want to stay warm.

Reading through Bill’s approach again, I think it is
clear that it is simpler and more achievable more quickly than mine. But
one of the things I uncover in my survey of the agent-based modeling literature
is that the models are very sensitive to assumptions, especially about what the
rules are that the agents follow.

Yes, indeed. And those rules are variables or adjustable features in my model.
I check out rules to see how they work, but anyone can plug any rules they like
into the same model to see what happens. For example, in the model I sent you,
a plant manager keeps inventories at a fixed level by adjusting prices. That
works quite well and gives a reasonable law of supply and demand, but maybe
there are other ways of doing the same thing. You can substitute the other
ways, or you can add more plants and managers and see what happens when
different managers work according to different rules. Maybe some rules lead to
going out of business. Others may generate conflicts.

In PCT terms, where do our references actually come from? If
they are simply assumptions made by the modeler, they will be subject to
criticism. Some are intrinsic, yes. But the rest? We must learn them. We likely
learn them as kids, both from experience and from others. So let’s model
the learning process in families and groups of families and see what kind of
economy we can derive from members’ efforts to satisfy their needs.

Fine, the reorganization algorithm has been checked out in systems of up to 500
interacting control systems. It ought to work with four or five people
controlling 10 or 20 variables. But you do need a starting point – an underlying
architecture of control systems to provide parameters that can be reorganized.

My strategy in general is to model control systems and leave their detailed
properties to be determined. Where I assume reference levels to test the model,
I want to substitute higher-order control systems to adjust them once the lower
level is working right. Where I assume particular controlled variables, anyone
is free to propose other controlled variables. Ditto for means of action. The
only reason for making particular assumptions about the lower-level systems is
to make sure they will work, and to learn what they can do.

Best,

Bill P.