The All-Seeing Eye

Musings from the central tower…

Panoptic Power and Competition

It is fairly uncontroversial in classic economic theory that free and fair competition is often vastly more productive than limited competition or no competition.  Many economists view a monopoly as a market failure and believe that anti-trust laws must be created and enforced in order to preserve competition.  So-called “no-bid contracts,” in which firms are granted lucrative government contracts based on cronyism rather than competition, are slammed, correctly, for costing a great deal more money than competitive contracts would cost.  As a general rule, when agents compete on the market, the goods or services that the agents are selling become more productive and/or less expensive – in other words, competition allows buyers to get more for less.

How is this related to panoptic power?  Economic competition conforms closely to the panoptic model of power.  Let us compare economic competition to the two hallmarks of the panopticon:  self-surveillance, and isolation.

In the panopticon, self-surveillance is produced within a subject by causing that subject to behave as though at any moment she might be under surveillance by a central observer.  Who is the central observer of the competitive market?  The consumer.  At any time, the consumer might evaluate the quality of the products offered up for sale by the competitors.  Competitors earn reputations based on the quality of their products, and these reputations greatly affect the profits of the competitors.  The consumer is also somewhat unpredictable, in that one never knows exactly what a consumer’s preferences might be.  Perhaps your innovative new product might become the next iPod – or perhaps it might become the next Betamax. Competitors must strive towards innovation and invention and reinvention, and must also master marketing, and still success is not guaranteed.  The point here is that competitors are always being evaluated, and they may live or die based on the results of these evaluations.  This is a powerful incentive towards self-surveillance.

In the panopticon, isolation is caused by physically separating prisoners in individual cells.  In a competitive market economy, isolation comes in the form of patents, trade secrets, and the information asymmetries that arise when competing agents each try to find and maintain a competitive edge.  If you own a restaurant, you might make the best marinara sauce in the county, but if you give away your secret recipe, that will not be the case for long.  Isolation also comes from the fact that individual agents may earn more profits through competition than through cooperation – because there will be fewer people to share the wealth.  Laws against cartelization and other forms of corporate cooperation can produce isolation effects.  In a situation where workers are competing for jobs, the workers may become isolated from each other because some wish to go on strike for higher wages while others wish to take over their jobs.

If the productive power of a competitive market is related to the productive power of the panopticon, then do the same downsides exist?  Sure.  One of these is that PD-like situations may arise in which competitors end up reaching a suboptimal equilibrium state because of their isolation.  An example of this is an industry in which advertising costs comprise a significant percentage of the industry’s income but do not effect a significant redistribution of market share for any one firm nor attract a significant number of new buyers to the market.  Each firm would be better off if no firm advertised, but if any firm advertises, they all must in order to avoid losses.  In the end every firm advertises, and the entire industry essentially throws money away.  The tobacco industry is one such example (although you won’t hear me mourning their suboptimal profits.)  There are also cases like railroads or utilities where, without collusion or intervention, redundant services may be established (imagine the case of two competing rail lines running parallel to each other).

Because the effects of panoptic power are generally experienced as difficult and unpleasant, free markets tend toward a mix of competition and cooperation.  Many agents would like to collude with each other in order to avoid the panoptic effects – in other words, break isolation to resist the power of the panopticon.  A cartel is a good example of this – competitors get together and decide that rather than compete with each other, they’ll fix prices and production at a certain level so they can all profit equally.  These cartels often result in higher prices and lower quality and quantity for goods produced – they are less productive, but they make things easier for those involved.  Labor unions are a form of cartel for workers, who agree to band together to achieve higher labor prices (wages) and shorter working hours (less production).  Such cartels and unions suffer from the risk that one agent will defect or a new agent will enter the market, thus destroying the cartel or union, and as a result an equilibrium can be reached.

The free market, which depends on competition for its functioning, is thus an example of panoptic power at work.  This is an important insight because many experience panoptic power as something which imprisons them, which calls into question how much “freedom” agents in the free market actually have and provides a theoretical framework to contrast, rather than conflate, liberty and productivity.

October 25, 2008 Posted by | Economics, Power | , , , , , , , | Leave a comment

The Prisoner’s Dilemma: A Semiotic Analysis

One of the many fascinating aspects of the Prisoner’s Dilemma is the way that it is framed. Take, for instance, this concise description from the Stanford Encyclopedia of Philosophy:

Tanya and Cinque have been arrested for robbing the Hibernia Savings Bank and placed in separate isolation cells. Both care much more about their personal freedom than about the welfare of their accomplice. A clever prosecutor makes the following offer to each. “You may choose to confess or remain silent. If you confess and your accomplice remains silent I will drop all charges against you and use your testimony to ensure that your accomplice does serious time. Likewise, if your accomplice confesses while you remain silent, they will go free while you do the time. If you both confess I get two convictions, but I’ll see to it that you both get early parole. If you both remain silent, I’ll have to settle for token sentences on firearms possession charges. If you wish to confess, you must leave a note with the jailer before my return tomorrow morning.”

First of all, what’s in a name? “The Prisoner’s Dilemma” tells us about a dilemma faced by a prisoner – note the placement of the apostrophe. In this example, though, there are two prisoners, and they both face the same dilemma. Why, then, do we not see this dilemma called “The Prisoners’ Dilemma?”

Well, the dilemma only comes about as a result of the separation of the two prisoners into individuals. If there were one player controlling both prisoners and trying to maximize her score, she’d have no dilemma. The Prisoner’s Dilemma, however, is faced by one individual, alone, isolated from contact with her fellow prisoner.

That leads to another question: why a prisoner? There are plenty of ways to propose the same paradigm – for example, two students who turned in identical examinations. Or an invading army offering a reward for whoever will open the town’s gates at midnight. The point is there are any number of anecdotes that match the payoff matrix of the PD, and perhaps infinitely many could be invented. Prisoners were chosen – why?

On some level, the prisoner’s dilemma applied to prisoners is well-understood. Plenty of people enjoy the legal drama as a story – plenty of people have been exposed, in our time, to Law and Order, Homicide, or CSI, and these modern shows have predecessors, and those predecessors use tropes set up in literature.

I would argue, however, that the choice of prisoners goes deeper than a simple familiarity. There is a match in situation between a prisoner and one caught in the prisoner’s dilemma. In other words, if this same problem were explained using students caught cheating instead of criminals caught robbing a bank, the students would still feel trapped. They would still feel like prisoners. And the person who reads the dilemma sympathizes with the subjects of the dilemma and feels their sense of being trapped. The situation that we call the Prisoner’s Dilemma works because of panoptic power, and panoptic power makes its subjects into prisoners, much more than it makes them into students, or frightened villagers.

When I say that people sympathize with the subjects of the dilemma, I mean, the reaction intended by the framing of the dilemma is that the reader puts herself in the place of the prisoner. The reader must ask “What would I do if offered such a deal,” and not “What would I do if I were the prosecutor and I had two prisoners?”

And that brings up another quite interesting point. We are told explicitly what the preferences of the prisoners are: they want to maximize their freedom, even to the detriment of their partner in crime. We are never told the goal of the prosecutor, although the prosecutor’s actions seem to speak for themselves. The prosecutor’s goal is, simply, to get convictions – to maximize jail time for the two prisoners. In our legal system, prosecutors build their careers by putting people in jail. If the goal of the prosecutor were to find out the truth, the Prisoner’s Dilemma would not be an effective tool. Imagine that one prisoner committed a crime and the other prisoner just happened to be in the wrong place at the wrong time. Who would be more likely to confess a crime, thus putting the other person in jail, and who would be more likely to maintain her innocence, thus going to jail?

If the panoptic model helps to maximize power, and the Prisoner’s Dilemma uses the panoptic model on behalf of authority, then it is a blind authority whose power is maximized. The authority does not care about the well-being of the society it has authority over; it is merely concerned with maintaining its own power. And indeed, this relationship is reflected by the names of the moves in the Prisoner’s Dilemma: defect and cooperate.

To defect is to cooperate with the authorities. To cooperate is to be defective from the point of view of the authority – a defective, criminal citizen, a defector, a saboteur of civil authority. The terms “defect” and “cooperate” refer to a presumed partnership between the two prisoners. I say presumed because nowhere in the description of the Dilemma is it stated or implied that the two prisoners have made any kind of agreement about how to handle such an eventuality as being placed into the Dilemma. The two prisoners are presumed to be partners only in the sense of (allegedly) participating in a criminal activity together; since nothing is stated about their guilt or innocence it may well be that neither of them has even met the other. The two are described as “accomplices,” but accomplices who prioritize each other’s well-being much lower than their own – in other words, hardly friends, or compatriots, or long-term partners in any sense.

However, the term “defect” implies something defected from, some country or alliance. Again, the term cooperate is ambiguous – the prisoners can cooperate with each other or with the authorities. We’ve already established that the authorities do not have the best interests of the prisoners in mind, and now we also establish that if the prisoners are to do any cooperating, it will be with each other, against the authorities. The authorities want to gain power by destroying the bonds of cooperation and causing someone to defect from society; the prisoners want to remain free by holding together as a society.

The choices could have been named differently. Confession could have been called cooperation – and certainly if we view the game from the perspective of the prosecutor, a confession would be a way for a prisoner to cooperate. However, again, the PD puts us in the place of the Prisoner, who is depicted as being in society with other prisoners but not with the prosecutor or the authorities.

Again, the PD does not have to be expressed in these terms that suggest that authority is opposed to, rather than part of, society. It doesn’t have to, but it is, and I consider this highly significant. The PD is not a dilemma about how we get justice – it is a dilemma about how we get freedom. And the freedom of the reduced sentences is not simply a freedom from jail, but a freedom from the power that would turn us against each other in pursuit of its own anti-social goals.

March 9, 2008 Posted by | Game Theory, Power | , , , | Leave a comment

The Prisoner’s Dilemma and the Panopticon

I’ll start this post with a brief recap:

The Prisoner’s Dilemma (PD) is a concept in game theory that describes the situation of two suspects who have been apprehended by the authorities. In the PD, the authorities need a confession in order to get the conviction they want, so they come up with a scenario to try to convince each suspect to confess. They offer each prisoner a reduced sentence in exchange for a confession that incriminates the other prisoner. If both prisoners stay silent – a play that is conventionally called “cooperate” – they both get a short sentence. If one prisoner chooses to “cooperate” but the other prisoner makes a confession – a play called “defect” – the defector goes free and the cooperator gets a full, long sentence. If both “defect” they both get a medium sentence.

Like the Traveler’s Dilemma, it is better in the Prisoner’s Dilemma for both players to cooperate – choosing (100) or choosing to stay silent. Also like the TD, in the PD if one player cooperates, the other player can increase his payoff by defecting – choosing (99), or choosing to confess. And finally, if one player defects – by choosing (2), or confessing – the other player can mitigate the harm done by also defecting.

The Panopticon is a philosophical concept that describes the situation of prisoners in a more general sense. The original panopticon was a design for a physical structure that would house prisoners in such a way as to maximize the number of inmates who could be supervised by one warden. This design consisted of a central tower where an observer could remain unseen by the inmates but from which all of the inmates could be seen. The inmates were situated in individual cells surrounding the central tower, separate from each other.

The idea of the panopticon is that this situation – isolation and the perpetual possibility of surveillance, would produce within each prisoner a sort of self-surveillance. Each prisoner would know at all times that he could be under supervision, and so each prisoner will act at all times as though he were under supervision.

The difference between self-surveillance and regular surveillance, though, is that self-surveillance can be much more intrusive. After all, an outside observer can only see certain physical manifestations of our actions – in other words, can only see what our actions look like. We, on the other hand, can, in a sense, see what our actions are. We form the intent that turns a motion into a gesture, an activity into an action, a sound into a word. We can read our own minds.

This paves the way for what I like to call the panoptic model of power. The panoptic model of power says that power is constituted and magnified by the effects of isolation and self-surveillance. Isolation and self-surveillance are interlocking, mutually reinforcing forces – in other words, isolation helps constitute self-surveillance and self-surveillance helps constitute isolation. A good example of how this works is the Prisoner’s Dilemma.

The most obvious intersection of the PD and the panoptic model of power is isolation. Without isolation, the PD would not be a dilemma. Imagine the PD with both prisoners in the same room. They can talk to each other, they can see each other, and they know what the other one is doing at all times. In other words, you’ve removed the hope that one player can defect without the other player defecting, and so now the options are only (defect, defect) or (cooperate, cooperate). Between those two options, one is strictly better, and it’s the one that benefits both players the most – so there’s no dilemma.

The self-surveillance part of the PD may not be as obvious. First we can look at the effects: The expected effect of the PD is that both prisoners confess. Is not confession a form of self-surveillance? It’s self-incrimination, certainly. One might expect the prisoners to provide additional information to the authorities in the course of their confession – details of the crime, perhaps the location of weapons used in the crime, perhaps details about other accomplices, or motives, or planning. In other words, the PD goes a lot deeper than the surveillance the authorities were able to place upon the prisoners without the PD.

To find the cause, we need only locate the central observer. In the panopticon, the prisoner exercises self-surveillance because the prisoner might be under surveillance. In the PD, the prisoner confesses because the other prisoner might confess. In the panopticon, the possibility of being watched leads the prisoner to watch himself. In the PD, the possibility of being incriminated leads the prisoner to incriminate himself.

The Prisoner’s Dilemma, thus, provides both an example of the panoptic model of power at work, and an insight into one of the mechanisms of the panoptic model of power.

February 17, 2008 Posted by | Game Theory, Power | , , , , , , | 2 Comments

Free Will, Determinism, and Motivation

It is possible to look at the universe like a giant computer. If you know the software a computer is running and all of its inputs you can predict the result. Similarly, one might think that if you knew all the rules of the universe – that is, if you understood physics perfectly and accurately – and if you knew the position and velocity of each particle in the universe, you would be able to predict the results – that is, how everything would turn out. Such a view is called determinism. When Newton first proposed that all matter obeyed certain laws, he was accused of atheism, because the obvious implication of his theories was determinism, which is a theory that leaves no place for God and no place for free will.

The question of free will vs. predestination or determinism is, of course, older than Newtonian physics, but physics is the way in which I first conceived of the question. One might ask, if God is all powerful, how can anyone act in a way God does not want? One might ask, if God knows all, then isn’t destiny written – isn’t there no way to change things? I have never been particularly into theology, but physics always fascinated and frightened me.

Some time ago I read Isaac Asimov‘s Foundation trilogy, which despite its name consisted of approximately thirteen thousand books. This series is based upon the idea that there was a mathematician named Hari Seldon who was able to predict the course of history using mathematical models and a deep understanding of historical trends. I did not find this idea credible. People, after all, are far too complicated to reduce to a mathematical model. Aren’t they?

Can we predict what people will do in a given situation? If we can, what does that say about free will? If we can’t, how can we enact social change?

In Freakonomics, authors Levitt and Dubner describe a scenario in which parents were charged a small fee (I think it was $3) for being late to pick their children up from daycare. The result of this fee was that lateness increased dramatically. According to Levitt and Dubner, the fee was too low, and parents felt as though paying $3 justified their lateness. In other words, when no provision is made for lateness, the parents have to pick their kids up on time or risk their kids being scared and alone. When the daycare center charges for lateness, watching the kids for a few more minutes becomes just another service that the parent can buy, and buy they do. The point of this story is that incentives don’t necessarily work the way we think they will. There are complicated issues at stake even in something as simple as daycare. As we saw in the Traveler’s Dilemma, it’s not a simple task to predict how people will make their decisions, and sometimes rational behavior isn’t what theorists think is rational.

However, what both of these scenarios show is that despite the difficulty, despite the complications, it is possible to develop models and predictions for how people will behave. It is possible to find, with experimentation, the fee amount at which parents will begin picking their children up on time to avoid the fee. It is possible to find, with experimentation, the punishment amount at which people will begin picking the low number rather than the high number in the Traveler’s Dilemma. In other words, people’s behavior may be more complicated than we think, but it is not unreasonable. People act based on motivations, and although these motivations are often not obvious, they are there and they can be found.

Of course there will always be exceptions. There’s always room for free will. There will always be people on the far ends of the bell curve, people who defy expectations and act inexplicably. But in order to effect positive change in the world, we have to believe that we can predict behaviors for most people. We have to believe that there’s a number of dollars that will decrease the amount of late pickups from daycare. After all, isn’t this how we determine prison sentences? Isn’t there a number of years of incarceration that we believe will make the commission of murder unattractive to most potential criminals? Isn’t there a number of dollars that we believe will deter people from speeding and thus decrease the number of traffic accident fatalities?

I’ve never really believed in free will. I’ve always thought that everything is already determined by particle vectors, that everything I do is explainable by something that happened to me in childhood or by a set of circumstances that outlined my choice to such an extent that I didn’t really have a choice. And that’s why, for me, I think it is important for us to search for these motivations, search for these incentives, to build and discredit and rebuild these mathematical models to predict behavior. Because I want to set things up so that people have no choice but to make the right choices. I want a society full of people who pick 100 on the Traveler’s dilemma and pick their children up on time from daycare, and if we’re going to have that we have to pick the right game.

And that, in turn, is why it’s worth looking at something like the Traveler’s Dilemma and finding out that people will cooperate with each other as long as the risk for doing so isn’t too high. It’s why it’s worth looking at the daycare paradox to find out how much guilt is worth. It’s why it’s worth asking why people follow their king or their president against their best interests. We need to find out what motivates people. And in exploring incentives and economics, game theory and modeling, philosophy and psychoanalysis, that’s exactly what I hope to do. I hope to find a solution, a way to set up society so that we’re all playing a game that everyone can win.

In closing, right now I feel that most people are not playing a game that everyone can win. There’s a game called the Prisoner’s Dilemma. In this game, two prisoners are in the custody of law enforcement, but the police don’t have enough evidence to convict them of a serious crime. Each of them is told that they are both suspects and given the following options. If one prisoner gives up the other, that prisoner will go free and the other will go to jail for a long time. If neither of them confesses they will both serve a short sentence for whatever smaller crimes the police can put on them. If they both confess they’ll both serve a short sentence. The implications of the game are that it is better for each player, no matter what the other player does, to confess their crime. Unlike the Traveler’s Dilemma, the Prisoner’s Dilemma tends to lead to uncooperative behavior – in other words, it is much better for each player to screw the other player over, unlike in the TD in which screwing the other player leads to a greater loss.

The Prisoner’s Dilemma game describes many situations in modern life – situations in which people have a great incentive to hurt other people. If there is some way to change the rules of the game so that, like in the Traveler’s Dilemma, or many other games, people have an incentive to help other people, then everyone could benefit immensely. Changing the rules of the game is what I’m aiming for, but it’s going to take a lot of searching to find the right game and a lot of convincing to get people to play it.

February 10, 2008 Posted by | About, Economics, Game Theory | , , , , , , , , | 2 Comments