Economics: The interaction of Control systems
Unedited posts from archives of CSG-L (see INTROCSG.NET):
Date: Tue Feb 11, 1992 2:24 pm PST
Subject: CT Economics
[From Bill Powers (920211.1500)]
Here's a kickoff for a thread we might call CT Economics.
----------------------------------------------------------
It's hard to get a feel for the antiquity of Adam Smith's writings in _TheWealth of Nations_*. But there's passage in my Penguin edition thathelps:
"In the province of New York, common labourers earn three shillings andsixpence currency, or two shillings sterling, a day..."
It isn't the wages or the strange dual monetary units to which I refer, butthe footnote attached to "earn":
"1. This was written in 1773, before the commencement of the latedisturbances."
This is the book, as old as the United States, in which the "invisiblehand," the Law of Supply and Demand, was invented. Economists have taken thislaw as a basic given of economics ever since. It's been taken as the magicformula that sets a free-enterprisesystem straight when it goes astray, no matter what mistakes are made. Leavethe system alone and the invisible hand will make sure it flourishes. Meddlewith it, regulate it, insert human intentions into it, and only disaster canfollow. The Law of Supply and Demand is greater than any of us and operates inmajestic disregard of mere human desires and purposes --or so we are told by theoreticians in economics.
This, however, was not Adam Smith's conception of this law. For Smith, itwas simply the natural consequence of the fact that people want things and arewilling to devote some labor (or labour) to getting them. Here is what he saidabout demand.
"The market price of every particular commodity is regulated by theproportion between the quantity which is actually brought to market, and thedemand of those who are willing to pay the natural price of the commodity ...to bring it thither. Such people may be called the effectual demanders, andtheir demand the effectual demand; since it may be sufficient to effectuate thebringing of the commodity to market."
" ... When the quantity of any commodity which is brought to market fallsshort of the effectual demand, all those who are willing to pay the whole valueof the rent, wages, and profit, which must be paid in order to bring itthither, cannot be supplied with the quantity which they want. Rather than wantit [lack it] altogether, some of them will be willing to give more. Acompetition will immediately begin among them, and the market price will risemore or less above the natural price, according as either the greatness of thedeficiency, or the wealth and wanton luxury of the competitors, happen toanimate more or less the eagerness of the competition. ... Hence the exorbitantprice of the necessaries of life during the blockade of a town or afamine."
Similarly, he says, when there is an excess of supply over effectualdemand, some of the goods must be sold to those who are willing only to payless, so the market price will sink lower than the "natural" price.
"Effectual demand" is one side of a purely psychological explanation of theLaw of Supply and Demand (the other side would describe the producer-seller's adjustments of the price, given the need to pay the costs ofproduction and distribution and the desire to have a profit left over).
The payment of money for goods, Smith pointed out, misrepresents what isactually traded. "The real price of everything, what everything really costs tothe man who wants to acquire it," he said, "is the toil and trouble ofacquiring it."
So we have the basis for a model of the demand side of economictransactions. Human beings desire goods and services, some being desiredbecause human beings can't live without them and the rest because human beingshave learned to value them. People are willing to exert a certain amount oftoil and trouble in order to bring the goods and services thither --a great deal of toil and trouble, if they require the goods and services to sayalive. If the shortfall is great, they will put out more effort to get whatthey want and need; if there is an excess, they will not make any effort to getmore than they want or need.
On the supply side, there are people who also want goods and services, andare willing to go through toil and trouble to get them. Their labor is not onthe production line, but is used to buy materials, rent facilities, borrow,organize, and otherwise manage production in order to create the goods andservices that they and others want. They give receipts for goods and servicesto laborers who transform raw materials into production machinery andcommodities, and they adjust prices (which are paid to them using the samereceipts they gave out) for those commodities, so that the whole product can besold and the costs of production can be repaid. The receipts returned to theproducer-sellerare used in part to pay the laborers who harvest the raw materials, build themachinery, and produce the goods and services; in part to repay the borrowedreceipts with interest; and in part (the part they are most concerned about) tospend on goods and services for their own consumption.
So all human beings produce toil and endure trouble in order to bring thequantity of goods and services made available to them by that toil and troubleto the level that they need or want. This is the great engine that drives allhuman interactions, including those that come under the heading of economics.The invisible hand is the hand of a human being working to control whathappens, not to an abstract economy, but to that human being. There are noother forces at work.
Theoretical economics has dropped the human being from these equations andhas tried to explain the workings of an economic system with no human beings init. This is why theoretical economics has so little relationship to whatactually happens in this economy. In order to build a clear picture of economicinteractions, we must understand that they result from the basic nature ofliving control systems, human beings.
Best Bill P.
Date: Mon Jun 01, 1992 7:23 pm PST
Subject: Economics
[From Bill Powers (920601.2000)] to Greg Williams (920601)
Good point about the reference levels. But I think there's a deeper glitchin the economy than just oil prices.
The problem is that there's a basic conflict between consumers andproducers --the same one that communism tried and failed to resolve. It hasn't gone away.The split between wage income and capital income in the for-profitsector (government is not-for-profit)is about 40/60 --40 percent for labor, 60 percent for owners, stockholders, debtholders, etc.This has been pretty close to the ratio since 1930, with the capital-incomeshare having risen slowly from about 53 percent in 1930 to today'sapproximately 60 percent. The conflict is that receivers of capital income wanttheir share to increase, while wage-earnerswant it to decrease.
The composite consumer (not the producer) has the reference level ofimproving the standard of living. This means working fewer hours to obtain ever-bettergoods and services, or even just to be working and eating instead of notworking and not eating. The idea is that technology or ingenuity --increased productivity --should be rewarded by obtaining a better life with less prolonged, unpleasant,boring, dangerous, unremunerative, or mind-numbinglabor.
The composite producer (with bean-countersin charge) has the reference level of maximizing the return on investment forthe owners of the means of production, or those who have invested in it. Thismeans cutting costs wherever possible and charging the most the market willbear for the lowest quality goods or services that can consistently be sold.Cutting costs means, in large part, reducing the cost of labor. When youreflect that cutting material costs is also cutting costs of labor (on someoneelse's part), it all comes down to cutting labor costs --if capital income isn't to decrease.
The kicker is that the wage-earnerswho produce the products have no way of buying the products except with themoney they are paid in wages. So if costs are cut by laying people off,substituting cheaper overseas labor, or reducing domestic wages, the result inall cases is that the buying power of the consumers is reduced --so the goods and services can't be sold at higher or even the same prices, inthe same volume. This is where the conflict comes to a focus.
Unfortunately, this system doesn't have any natural reference levels in themiddle of its range of operation --it just has limits. It always tends toward the state where some large number ofpeople is existing at a subsistence level. The only thing that keeps thecomposite producer from reducing labor costs any further is the fact that a lotmore people would begin dying of starvation or untreated illness or would havetheir physical living conditions reduced to an intolerable state. The resultwould be an explosion of crime, or revolution. So a balance is reached wherethe deleterious effects of further reductions in consumer buying power willincrease costs (through taxes for welfare) and reduce sales (through loss ofbuying power) unacceptably. Government tries to alleviate this situationthrough redistribution --spending tax money in ways that increases the slice of the wage-earneror dependent. But the composite producer has no such motive, except when somany people become impoverished that the market begins to fall off.
The government and private philanthropies together manage to acquire enoughmoney from the composite consumer to bring the fraction of capital income downto about 40 percent by redistributing income. Evidently, this is the fractionat which the wage-earningor seeking population has to be maintained even to keep the economy in itscurrent state. If there were no redistribution, there is no way that capitalincome could remain at 60 percent of the total without creating a violentrebellion by starving people.
People talk in the same breath about our prosperity reaching new highs, ifmore slowly nowadays, and about the increasing split between high-incomepeople and low-incomepeople. The high-incomepeople are also the chief recipients of capital income. They form the high endof the market. So companies who see sales falling off try to aim for the peoplewho have the money: they produce luxury services, labor-savingitems and toys, high-techor disposable goodies, that will attract the small fraction of the populationthat has the most money to spend. The result, of course, is that the people atthe low end find fewer and fewer items they can afford to buy. The people whoCAN maintain their 1970 standard of living work like hell to do so (to get toyour point). But just in working like hell to do so, they've sunk below thatstandard of living. And of course, there are far more people who can't get orhandle two jobs, who work less than they used to or at lower wages, and arehaving a more miserable time than ever.
I think that the owners and managers of this economy need a visit from EdFord. Somebody has to ask them, "Is it working?" The problem is that theiranswer is really "yes" --so far, it's working for them. A CEO earning $3 million per year plus perkscan't really complain. But the SYSTEM CONCEPT isn't working for the people whoactually make the system go. It's only working for those who own the system orhold its debts.
There is something drastically missing from the hallowed concept of freeenterprise. It's keeping the people whom the economy is supposed to serve inthe condition of Skinner's rats. This is something that I think controltheorists need to be talking about.
Best, Bill P.
Date: Wed Jan 27, 1993 7:52 pm PST
Subject: Economics; misc
[From Bill Powers (930127.2030)] Bob Clark (930127) --
I have a lot of things to say about economics, Bob, but I'd better postponethem a while. I'm going away for 10 days starting Feb. 7, and Wolfgang Zocheris about to spring a very nice version 4 of Simcon on the world which I want tohelp get distributed, and I'm working on the 7-dfarm model, and the chatter on the net never ceases.
Just a brief word. My father has done a lot of research on the factualrecord of economic phenomena, comparing it with basic theories, and has foundthat the basic theories are almost total nonsense. Examples: when money is madetight, the rate of inflation goes UP, not down. The analogy of national savingsto savings of families is completely wrong. What economists use as an indirectmeasure of national savings really represents a loss of buying power out of thenational economy. Industry has spent 20 +/-2 percent of its total income on capital investment percent every year for thelast 100 years; you can't jazz up the economy by increasing investment. Andother goodies. His approach has been the macroeconomic one, and I've becomevery interested in it. Economists pay very little attention to thefacts.
From the macroeconomic point of view, taxation has essentially no effect onthe economy. Those who object so strongly to taxes seem to forget that thegovernment spends its tax income AT LEAST as fast as it receives it. Most ofthe money taken in is handed right back, although redistributed, becomingbuying power in the hands of the public and industry again. What changes is whogets the money, and that's a microeconomic problem, or a problem of socialjustice, or politics. Maybe taxing one industry discourages that industry, butanother one gets the money and is encouraged. The best place to put tax moneyis in the hands of poor people, because they are sure to spend it all rightaway, and not hoard it or leak it away in bad investments outside our economy.It's the people who don't spend all of their income on goods and services inthis economy who are responsible for most of our problems. Those tend to be thepeople and organizations who make so much money that they can't possible spendit all on goods and services.
Control theory gets into macroeconomics by explaining what keeps thecircular flow going.
Later, Bill P.
Date: Sat May 01, 1993 8:14 am PST
Subject: Economics: the Williams Effect
[From Bill Powers (930430.1100)]
Six or eight years ago, an economist named Bill Williams came to visit me.He thought that control theory might be able to explain a phenomenon called theGiffen Paradox. We worked nonstop through the weekend, and indeed came up witha control-systemmodel that reproduced this effect (no longer a "paradox"). I've recently lookedinto it again, and have found a simplified version of it. I think that theGiffen Effect (or perhaps it should now be known as the Williams Effect) canexplain a lot of economic problems and perhaps give us a meaningful workingdefinition of poverty.
The Giffen Paradox has been known (and ignored) for a long time. The effectis called paradoxical because it results in a reversal of the normally-acceptedlaw of supply and demand. In a situation where people are on a limited budget,it can happen that when the price of a good increases, people are forced to buymore of it (the normal law of supply and demand requires that an increase inprice result in a decrease in sales).
The representative case that Bill Williams started with was one in which aperson has a choice of buying meat or bread. Meat and bread can provide aboutthe same number of calories per pound, but meat costs much more per pound thanbread. Bill also introduced a "prestige" factor, in which there was a built-inpreference for meat over bread regardless of cost of calories.
The model turned out to consist of three control systems:
1. The calorie control system had a reference level for calories needed. Ifthe amount being obtained was less than the reference level, purchases of breadand meat would be increased equally, as either one can provide calories. Thiscontrol system worked only when the obtained calories were less than thereference amount. Excesses of calories were not resisted.
2. The "budget" control system had a reference level for amount of moneyspent. This control system became active only when the total being spentexceeded the reference level. An error (excess of spending) was turned into adecrease in purchases of meat alone (the more expensive commodity), with noeffect on purchases of bread.
3. The "prestige" control system gave a high weight to perceptions of meatbeing consumed, and a low or somewhat negative weight to perceptions of breadbeing consumed. Deficiencies of prestige led to increases in purchases of meatAND decreases in purchases of bread.
These systems operated independently and in parallel. By adjusting the gainfactors and the weightings of the various perceptions, it was possible toreproduce the Giffen effect. Raising the cost of bread resulted in an increaseof bread consumption and a decrease in meat consumption, but only if the totalallowable budget was below a certain level.
On returning to this model, I realized that the prestige factor wasunnecessary except for producing a preference for meat when the budgetarylimits were removed. If the output weights of the calorie control system areequal, equal amounts of bread and meat will be purchased when no budgetaryconstraint exists. If a preference for meat is wanted in the model, it can beput in simply as an increased output weight for meat purchases in the caloriecontrol system.
When the budget is reduced, the total cost of providing the needed numberof calories tends to rise above the budgetary reference level, and purchases ofmeat are reduced. Since this reduces the number of calories consumed, thecalorie control system raises the tendency to buy BOTH bread and meat. But atendency to increase meat purchases is offset by the budget control systemwhich forces meat purchases down, leading to a net increase in bread purchasesand a decrease in meat purchases. The essence of the Williams Effect is thusrecreated without any need for a third factor.
Increasing the cost of bread has the same effect as reducing the budgetaryreference level: it drives the total cost above the budgetary reference level.The two control systems respond as before, increasing bread purchases anddecreasing meat purchases, keeping the calories the same and reducingexpenditures to the budgetary reference level. So raising the price of thecheaper commodity results in an increase in consumption of the cheapercommodity.
Actually, raising the price of EITHER bread or meat will result inconsumption of more bread and less meat, which makes sense. It's only when theprice of bread increases and more bread is purchased, however, that anythingparadoxical (in terms of conventional economic theory) appears to occur.
It's easy now to extend the Williams Effect to a large assortment of goodsthat provide alternate means of supplying a specific want, but at differentcosts. Whatever the mix of purchases without budgetary constraints, an increasein the price of any item that tends to cause spending over the total budgetwill depress the purchases of at least some items. If the excess spending iscorrected by reducing purchases of the more expensive items, the WilliamsEffect will be observed for all the less expensive items; increasing the priceof the less expensive items (one or more of them) will result in an increasedconsumption of those items, and less consumption of the more expensive items.This is the result of the control system controlling for the non-budgetaryeffect of purchasing all these items whether expensive or inexpensive --calories, in the above example.
The Williams Effect may have a close connection with the well-knownphenomenon of the rich getting richer while the poor get poorer. Richness andpoorness can be measured in part by what people are able to buy. High-qualityand luxury goods tend to be more expensive than low-qualityordinary goods that satisfy the same basic need such as clothing,transportation, or health care. If manufacturers continually probe the marketto see what prices it will bear, there will be a tendency to raise the price oneverything until resistance appears in the form of lower sales. At that point,those with the lower budgets run into the Williams Effect first. They mustdecrease their purchases of high-pricedgoods, but to maintain the same level of the needed good or service they mustincrease their purchases of the low-priced(and low-quality)goods. So the manufacturers find that they can raise the price on the low-pricedgoods disproportionately to the price of high-pricedgoods, and still get a net gain in profit.
The only equilibrium condition would seem to be the one where people withthe lowest budgets lose entirely the ability to buy any goods or services ofhigh quality. People in the poorest neighborhoods find themselves paying highprices for ordinary or low-qualityfood; they live in dilapidated housing and pay exorbitant prices for it; theydrive used cars of greater and greater age, or take public transportation, theprice of which keeps going up. They do without health insurance altogether, andseldom see a doctor, a dentist, an optometrist, or a counselor. They can'tafford lawyers or bail. And as they are forced more and more toward the poor-qualitylow-costend of the market, those supplying the lower markets find that they canincrease prices even further without losing sales --and indeed, even increasing sales.
So it appears that courtesy of the Williams Effect, the free market systemis organized to create a wide gulf between people without budgetary limits andpeople with them, and to keep this gulf increasing, limited only by thecondition in which too many people can't afford to live at all, a non-economicconsideration. The law of supply and demand works only for those withoutbudgetary limits --who can afford to choose what they buy on the basis of aesthetic objections tohigh prices, rather than being forced by necessity to adjust their purchases toavoid going into debt. For all those who must spend essentially all that theymake, the Williams Effect dominates and the road leads only downward.
I think this is an example of a way in which control theory can explainsituations that are unexplained under the assumptions of conventionaltheories.
Best to all, Bill P.
Date: Thu Sep 16, 1993 7:20 pm PST
Subject: Marcos: Economic PCT Proposal
[From Marcos Rodrigues (930915.1600 BST)]
The Economic and Social Research Council in the UK has just launched a newresearch programme entitled "Economic Beliefs and Behaviour":
"The purpose of the programme is to investigate the economic beliefs andbehaviour of people at a time of rapid economic change and to fosterinterdisciplinary research in this area. In particular, it will address thethemes of personal financial and economic behaviour; economic socialization;and the interaction of social and cultural norms and economic beliefs andbehaviour."
Since I know nothing of economics, I'm trying to convince a Professor ofAccounting at Manchester University to collaborate on a PCT-basedresearch project. I'm going to meet him over this weekend. Meanwhile, I'mwriting a draft outline proposal (a 3-pagedocument). If we get this through, then we will be invited to submit a fullproposal by next April.
I would be grateful to anyone who is prepared to spare some time readingand commenting on this. The deadline for submission is in around two weekstime. Since this is an outline, details are not normally required.
I've tried to describe PCT in one paragraph, and have failed. I thendecided to use Bill's description as in the Intro to CSG document. I hope Billdoesn't mind.
Best regards to all, Marcos.
-----------------------------------------
The objective of this proposal is to model human behaviour during dramaticeconomic changes. Since the market environment is by definition dynamic, ourview is that it is essential that we describe the factors that influencepeople's behaviour as a function of time. Those factors, or control variables,can have their relative importance dramatically changed (inside people) as newinputs are taken into consideration at different times.
We will dedicate especial attention to the change rate of some selectedvariables, that is, their derivative with respect to time, since we believethat people's perception of "transitions" is a determinant factor on behaviourregarding economic matters.
We propose to use a new approach for describing behaviour in livingorganisms called Perceptual Control Theory (PCT) whose development has startedsince the 1930's by William Powers.
PCT explains how organisms control what happens to them. It explains what agoal is, how goals relate to action, how action affects perceptions and howperceptions define the reality in which we live and move and have our being.Perceptual Control Theory is the first scientific theory that can handle allthese phenomena within a single, testable concept of how living systemswork.
The development of this project, within a PCT framework, will be based onthree planks: "identification", "modeling", and " simulation". Since we areworking in the realm of microeconomics, these stages will be constantly checkedagainst the behaviour of human subjects with suitably designed test procedures."Identification" refers to the identification of a set of perceptual variablescontrolled by people, that is, variables people are trying to bring to somespecified reference state, often through variable means. The correct level ofabstraction of such variables is essential to the success of the followingstages. " Modeling" is the indication of the interdependencies between thosevariables and the environment (internal and external). "Simulation" refers tobuilding and running computer simulations of the representation of controlvariables and their interdependencies, that is, to simulate the phenomenon ofperceptual control. We expect this project to give us a deeper understanding ofeconomic perceptual control variables, their interdependency, and their varyingimportance as a function of time, so that we will be better equipped to predictbehaviour under a certain level of abstraction.
In economics, the intersection point between supply and demand curves (thecanonical representation of market behaviour) is normally referred to as themarket equilibrium point for a product or service. The economist view is thatmarkets are not always in equilibrium but, if they are not interfered with,there are good reasons to believe that they are normally moving towards anequilibrium point. The problem with this representation and view is that it maytell economists what happens with a product, that is, it may give someinformation about an object being produced.
It does not give, however, any information about individual behaviour ofpeople producing or consuming a product.
It does not give information about the goals, desires and wants of bothproducer and consumer.
In PCT, the intersection point between supply and demand is not in anysense an equilibrium point; in fact, markets will oscillate forever as long aspeople have enough degrees of freedom regarding variables they can control inthe process of production and consumption of economic goods.
From a PCT perspective, the economist representation is not meaningfulbecause it fails to include people in it. In addition, PCT does not see themarket as a social control system that punishes anyone who tries to interferewith its normal working.
The failure to include people in economic analysis is similar to making adetailed economic analysis of milk production, distribution and consumptionwith a model that does not include cows in the picture. Adam Smith's "TheWealth of Nations" is quite illuminating in both aspects concerning people'sbehaviour and market outcomes: it attributes self-interestas the instrument of the ``invisible hand'' whose actions economists see as anoutcome, namely the market resistance to interference.
This tells us two things: first, the "invisible hand" are individual peoplecontrolling their perceptions of economic factors such as profits, budgetaryconstraints, dietary intake, and so on. Second, it follows that the apparentmove towards market equilibrium is nothing but a side-effectof individual control of perceptual variables that are so diverse to enumerate.Markets do not resist disturbance; individual people do, and they do it byexercising choice within given constraints.
PCT implies that there is a great number of control variables that bothsupply and demand sides of the economy are steering close to a desired states.For instance, from the supply side perhaps arguably the most important controlvariable is the perception of profits. A management will almost invariably takedecisions in order to maximize profits given existing economic constraints. Itmay be argued that not everyone during all the time go for decisions thatmaximize profits. Sometimes management make decisions that are just good forthe employees, or safe, irrespective of being maximizing or not the companyprofits. Such behaviour is fully accounted for in perceptual controltheory.
A decision that is just good is explained by PCT as the control of someperceptual variable (for instance, in the relation between capital and labour)that, at certain "moment in time", is more important to the management for anygiven reason than the immediate maximization of profits. The decision sometimesmight function as a constraint in maximizing revenues or it may incur in extraexpenditure. However, as soon as the decision is made, the management task isimmediately to maximize profits given the new constraints imposed by thedecision.
Controlling for the perception of maximum profits is a complex process thatis achieved through the control of a number of potentially complex variablesand interdependencies. In this project, we propose to undertake a systematicapproach to the identification and modeling of economic control variables, thatis, perceptual control variables inside people, and build computer simulationsshowing interdependencies and interactions between people within a dynamicenvironment of economics.
The modeled control variables should account for people's behaviour underdramatic economic changes. We expect that such studies will allow us a betterunderstanding of people's behaviour in relation to important (to people)economic factors.
It is essential that we use the right kind of approach to modeling andsimulation and that the final model shows a high correlation between predictedand expected behaviour. Once that is achieved, the approach can lead topolicy.
The first aspect we would like to consider is the identification andrepresentation of control variables. Since people interact within a dynamicenvironment, it does not make sense to simulate behaviour within a staticenvironment. Therefore, it is essential that all control variables berepresented as a function of time. The time factor properly taken into accountwill help us to understand managerial decisions and consumer behaviour. Webelieve that this approach will bring a broad qualitative understanding of someof the intricacies of economic interactions.
We have stated above that the maximization of profits is the single mostimportant control variable in the supply side of the economy. It follows thatfor the demand side the minimization of costs is the single most important one.The following is an outline of our approach in an attempt to show how theproject will be conducted in relation to the identification, modeling, andsimulation of a number of control variables. We cannot possibly enumerate thenumber of control variables we will try to model or the level of abstractionrequired for a reliable simulation. Therefore, the identification and modelingpresented in the following example is not claimed to be complete, it is just anindication of our methodology.
We define profit-quantity(PQ) as the perceptual control variable on the supply side of the market whosemathematical modeling is given by the following equation:
"$ PQ = \integral{(price -cost)*quantity*dt}$,"
which, in graphical representation, gives the area below the curve inFigure~1.(a). In the above definition, cost is the total cost (includinglabour) to produce the quantity in the consider time interval. By no means therelation $(price -cost)*quantity*dt$ is a linear relationship. PQ is a non linear,multidimensional control variable that depends, among other things, on theprice of the inputs, fixed costs, marginal costs for increase in amountproduced, etc. Modeling such a non-linearfunction of interdependent variables in a way that is ease to visualize, changeparameters, increase or decrease the relative importance of various inputs isone of the objectives of the present project. We propose to build a model inwhich such dependencies are directly indicated in a diagram, facilitatingchanges and the understanding of the system's dynamic behaviour within economicinteractions.
^
Profit | ____________________ PQ
| __________/ |
|_________/ |
| | (a)
| |
| |
|------------------------------------------------>
time
^
Costs | _________
| \ __________
| \
| \______________ CQ (b)
| |
| |
|------------------------------------------------>
time
Figure 1. Examples of time-dependentperceptual control variables. (a) represents the supply side of the market and(b) the demand. While suppliers control for zero or positive slope in curve PQ,consumers control for zero or negative slope in curve CQ.
Analogously to the supply side, we identify in the demand side of themarket a control variable which we define as cost-quantity(CQ). Taken the consumer's point of view, it is realistic to assume a limitedbudget so that within this constraint, the consumer will control for low costsspread over a period of time. Therefore, CQ is also a function of time, definedas
$CQ = \integral{(\frac{cost}{q})*quantity*dt}$,
where $q$ is a utility function ($0<q<1$) which may depends onquality and/or use, among other things. Again, the function$(cost/q)*quantity*dt$ is a non-linear,multidimensional control variable. Graphically, $CQ$ is represented inFigure~1.(b). The consumer will control this variable so as to minimize thearea below the curve $CQ$.
A rigorous PCT approach will give us some invaluable insights on people'sbehaviour under economic changes. A continuous changing environment does notrepresent a handicap in a PCT model; in fact, it is in these situations thatthe implications of the model can be fully appreciated. For instance, if $CQ$has an ascending slope over a period, this means that the consumer isoverspending, and s/he then may decide to cut expenses by selecting someproducts of which the amount purchased will be decreased. Note that it does notnecessarily follow that an increase in price of one product will reduce itsdemand. It depends on the relative importance of the consequences for theconsumer on incurring extra costs. The consumer may choose to selected aproduct --whose price is falling --that plays a large proportion in expenditure so that the quantity purchased ofthat product is decreased. Looking at the market outcome, it may be that afalling price product may see its demand decreased while a rising price productwill have its demand increased. This seemingly anomalous market behaviour is aside-effectof consumer preferences and tastes (sometimes for religious, cultural, or someother reason) that can be fully explained and understood within a perceptualcontrol approach to behaviour.
The modelling and computer simulations of complex interactions betweenpeople and their economic environment, their beliefs, wants, and desires willbe implemented through a visual programming environment called PROX (PictorialProgramming for Control Systems and Simulation) developed within the Departmentof Computer Science of the University of Wales. PROX generates software codefor complex systems with interdependent variables directly from a user'sdiagram representation of the state variables and their interconnections. PROXhas been conceived as a tool for modelling control mechanisms in cognitiveprocesses and implies a very wide range of applications for modelling dynamicsystems behaviour. Moreover, the development of PROX is grounded in theassumption that dynamic behaviour is generated by the workings of closed-loop,interdependent processes. There are two kinds of such processes: negative andpositive feedback. Negative feedback loops generate converging behaviour to aspecified state, while positive loops are typical of runaway systems, that is,systems with no converging behaviour. Feedback is then just a means to control --the process of producing predictable results in an unpredictable environment.The purpose of a control process is then to maintain the input signal (thecontrolled variable) at or near to the value specified by its referencesignal.
As with the identification of control variables $PQ$ and $CQ$ above,Figure~2 depicts an oversimplified model of how the variable $PQ$ would berepresented and interconnected with other economic variables. The model islargely incomplete and must be seen only as an indicative of our approach. Oneaspect that comes clear from this simple model is the role of transitions onthe determination of people's behaviour. In controlling for maximum profits(maximum $PQ$), it is clear that the quantity produced must be dependent on theway stocks are varying. The change rate in which stocks are coming up or downis represented by the derivative $dQ/dt$. The perception of this transitionwill influence production. Suppose that $Dq/dt$ is positive. That means thatstocks are growing, and this might bring production down. If $Dq/dt$ isnegative or zero, there is a demand for the product, so that the amountproduced might be increased.
.
. Ref=PQmax
. | P*Integral(Dq/dt)
. v |
.---------------->Compare----------->Compare<----
| . ^ |
PQ<---quantity. PQ<---quantity<------'
^ . ^ | Adjust
| . | |
Price-Cost . Price-Cost |
^ ^ . ^ ^ |
/ \ . / \ |
Price Cost .Price Cost |
^^^^^ ^^^^^ .^^^^^ ^^^^^ |
||||| ||||| .||||| ||||| |
inputs inputs .inputs inputs <-'
.
.
Next best . Actual product
alternative .
.
Figure 2. A simplified control model for variable $PQ$ in the supply sideof the market.
Date: Fri Sep 17, 1993 2:48 pm PST
Subject: Marcos & economics
[From Bill Powers (930917.1600 MDT)] Marcos Rodrigues (930915.1600BST)
The economics project sounds wonderful. Economics has been in the back ofmy mind since a CSG member and economist, Bill Williams, enlisted me eight orten years ago in tackling the "Giffen Paradox" as a control-systemproblem, and since my father started writing, in his extreme old age, abouteconomics.
The Giffen Paradox is so-calledbecause it concerns a situation (quite common) in which the "law" of supply anddemand works backward. In this situation there is a budgetary constraintlimiting the total amount of money that can be spent on a given set of goods.The representative goods Williams used were "bread" and "meat." Meat is morecostly per pound than bread, but contains more calories per pound than bread.Also, eating meat is more "prestigious" than eating bread (i.e., there is apreference for eating meat on some grounds other than calories).
We set this situation up as three control systems operating inparallel:
Calorie system: perceives total calories represented by meat and breadpurchases, acts by raising meat and bread purchases equally to bring caloriesto reference level. Purchases = consumption.
Prestige system: perceives positive prestige from meat purchases and zeroor negative prestige from bread purchases. Acts to raise total perceivedprestige toward reference level by increasing meat purchases.
Budget system: perceives total cost of meat and bread; if cost exceedsreference level (one-waycontrol system), reduces purchases of meat.
When relative loop gains were properly adjusted, this set of controlsystems reproduced the Giffen Paradox, to wit: WHEN THE PRICE OF BREAD ISINCREASED THE SYSTEM PURCHASES MORE BREAD. That is the "paradoxical" violationof the law of supply and demand, easily explained with a PCT model. So theGiffen Paradox should now really be called the Williams Effect.
I think this solution of the Giffen Paradox has enormous implicationsconcerning poverty, in fact providing a clear definition of what constitutespoverty. You are in poverty if your income is so low that when the price on thecheapest goods you buy is raised, you are forced to buy more of the cheapestgoods and less of the more expensive and higher-qualitygoods.
---------------------------------------
My father, a retired scientist, has been interested in macroeconomic theoryfor about 15 years (he is now 93), and has been writing about it. The mainpoint of his study is that he has been comparing existing economic theorieswith the historical record in the Statistical Abstracts (the record of what hashappened in the US economy). He finds that the predictions of economic theoryhave almost nothing to do with the facts. In place of standard economicassumptions, he has come up with a set of relationships that DO fit the facts,and some rather startling conclusions about what makes the US economy run, andfail.
One interesting fact is that for the past 100 years, the expenditures bythe "composite producer" on capital costs --i.e., investment --has remained constant at 20 +/-2 percent of total income, this range not been exceeded in any year. There isno relationship between amount of investment expressed as a fraction of totalincome and the Gross National Product: the same ratio appears in good years,bad years, and all other years. There is a fixed market for investment: theeconomy can't be made to grow faster by increasing investment. And it never hasbeen made to grow faster in that way.
Another interesting fact. In the US, the Federal Reserve has tried toreduce inflation by tightening the money supply. Tightening the money supplyhas had three main effects: it slows growth; it increases unemployment, and iteither has no effect on inflation or it INCREASES inflation. The only cleareffects of tightening the money supply have been to make the economyworse.
And now the biggest shocker. It is standard economic practice to calculatesavings (consumer and producer) by subtracting total expenditures from totalincome (NOT by adding up actual savings). This simple approach was suggested byKeynes' analogy of the whole economy to a single family, scaled up. To Keynes,it was obvious that a family does not spend all it earns, saving the differencefor future needs. It was equally obvious that this was how the whole nationworks: the difference between earnings and expenditures must represent thesavings of the whole nation.
What Keynes overlooked, and my father saw, was that this analogy isinvalid. At any given time, the nation represents individuals or families inALL stages of the economic cycle. Some are spending all they earn; others alittle farther along are putting money aside; still others are withdrawingtheir money, because their future has arrived. So the net savings rate of thewhole nation must be very much less than the savings rate for a family. Theapparent savings rate calculated from income and expenditures, which averagesabout 7% of yearly income, can't possibly be the real savings rate, for thereal savings rate must average around zero.
So what IS this observed difference between total income and totalexpenditures? It is a leakage of buying power out of the economy. There aremany sources of leakage, including bad foreign loans and investments, imbalanceof trade, and the use of overseas labor in place of domestic labor. One of themain reasons for leakage is the fact that some people or institutions have suchhuge incomes that they can't possibly spend them all on goods and serviceswithin the economy. Because they do not spend all of what they earn inside theeconomy, they do not return to the composite producer enough to pay forproducing the whole economic product. This creates an automatic annual markupof prices: one of the main contributors to inflation. The leakage rate alsodirectly subtracts from the exponent in the expression for growth rate of theeconomy.
Finally, inflation itself. It turns out that the primary culprit, otherthan leakage, is industry-widecollective bargaining. The historical record shows that increases in wages dueto collective bargaining have, for 100 years, been completely offset byincreases in prices. Organized labor has not gained one cent in purchasingpower through wage increases. There may have been other kinds of gains, butwage negotiations have not produced any increase in real wages. They haveproduced, instead, inflation. When leakage is added to wage increases,inflation is almost entirely accounted for.
-----------------
I haven't been able to talk my father into merging his analysis with a PCTanalysis ("What individuals want has nothing to do with economics!!!").However, PCT naturally enters into macroeconomics by providing the missingexplanation for what keeps the circular counterflows of goods and money going.What keeps them going is the difference between what people want and what theyget.
I think it would be possible to derive a macroeconomic demand curve fromthe composite behavior of individuals who want specific amounts of specificgoods, neither less nor more. Over a population, for any particular good, therewould be a range of amount wanted, with very few people wanting none at all,and very few people wanting enormous amounts (for example, of oatmeal or movievideos). The result would be that as the available amount of a good increases,there would be less and less effort on the part of the population to obtain it,producing a concave demand curve that is high at low supply and low or zero athigh supply.
This amounts to a new definition of economic man: not as a maximizer whocan be driven to indefinitely large outputs of labor when given indefinitelyincreasing rewards, but as a controller aiming to obtain specific amounts ofgoods and services. This change of assumptions, it seems to me, would lenditself to surveys of people's actual economic behavior. If you simply askedpeople to list how much or how many of a wide variety of goods and servicesthey would like to consume, I will bet that for most of them the number wouldbe quite finite and reasonable. How many pencils would you like to have? Howmany cars? How much food would you like to eat? How many clothes would you liketo have in your closet? And so forth.
In your proposal for modeling how managers work, you could try out the sameidea. Do managers really try to maximize profits? According to Newell (or wasit Simon), they do not: they pick what seems a reasonable goal and adjust theirefforts accordingly, a process he called "satisficing." He got a Nobel prize ineconomics for that. While he didn't advertise this idea as following from PCT,it most certainly does.
------------------------------
You can use any of my writings that will do you any good: feel free.
Best, Bill P.