Game Theory, Politics and Society

In what is very convenient timing, MarkCC has an introductory post about game theory here, and it looks like part of a series he's starting. Very good. I'm going to be reading a couple books on game theory and politics very soon (they are sitting on my nightstand), but I wanted to order my thoughts before being educated. What I'm interested in is how to establish both economic and political systems that encourage and reward people for acting in a globally optimal way instead of locally optimal. At the same time, I value personal liberty above almost all else. Are both of these goals possible or forever contradictory? This is a common theme for me: balancing personal freedoms against enabling long-term thinking. For this post, I'll talk about the simplest game I know of, and then apply two other games to politics and society.

Cake Cutting

One of the simplest games in GT is that of cake cutting, or fair division of resources. In this scenario, two participants (easily generalized to N) are presented with a cake (some resource) and tasked with dividing it up fairly. The fairest way to divide the cake is to have one person cut the cake, and the other person gets first choice of the pieces. There are many variations, and in fact, on the books I'm planning to read is by Steven Brams who developed an envy-free solution for this game. Which is another post. The point I want to make here is that I do not think this game is appropriate for true economic situations -- economic resources, i.e., wealth, are not a fixed quantity. This is the major flaw in socialism; one person getting richer does not necessarily make another person poorer. I may have more on this scenario later. But there are at least two easy to understand games that

Tragedy of the Commons and Prisoner's Dilemma

Anyone who is familiar with game theory probably already knows what these two are, so please feel free to skip this section. The tragedy of the commons (TotC) is an expression of how common resources are exploited and maintained (more likely not maintained). With a common resource that all parties can use, who is responsible for maintaining the resource? With no clearly designated responsible party, what usually happens is that the resource is not maintained. I'll have some concrete examples later.

Prisoner's Dilemma (PD) embodies the following situation:

Two suspects, A and B, are arrested by the police. The police have insufficient evidence for a conviction, and, having separated both prisoners, visit each of them to offer the same deal: if one testifies for the prosecution against the other and the other remains silent, the betrayer goes free and the silent accomplice receives the full 10-year sentence. If both remain silent, both prisoners are sentenced to only six months in jail for a minor charge. If each betrays the other, each receives a five-year sentence. Each prisoner must make the choice of whether to betray the other or to remain silent. However, neither prisoner knows for sure what choice the other prisoner will make. So this dilemma poses the question: How should the prisoners act?

It's easy to see that, without communicating, it's very hard for both of the prisoners to not choose betray, which gives them a lose-lose situation. There are some interesting applications and variations which I'll get to below.

GT and Politics

How does game theory apply to politics? One of my favorite harping points is how entrenched the two-party system is in our government; PD captures this problem very well. Third-party candidates have a very hard time convincing people to vote for them because the established party candidates use the refrain "You'd be throwing your vote away." (Incidentally, this is captured exquisitely in a Simpsons episode about Kang and Kodos). Let's call the candidates X, Y, and Z, where Z represents an independent party. If a single (or, given the large numbers), a few people decide to vote for Z, they risk allowing whichever they view as the worse of X and Y being elected. In essence, unless everyone opts for 'cooperate' (voting for a candidate they like, instead of voting against candidates they dislike), the status quo is maintained. While there may be communication ahead of time -- a third party candidate can implore people to make the right choice, that communication is rendered useless without trust.

TotC can also help analyze our political landscape. Consider problems with Congress such as pork barrelling, earmarking, and special interests. In this situation, the 'commons' is our tax money. What should be done with it? Who is responsible for maintaining it, and who gets to use it? To apply TotC, assume that "maintenance" means "being fiscally responsible". Here, one would hope that each member of the Congress would be responsible for maintaining the budget, and that the budget is continually being used for various projects. Each member of Congress wants to use some portion of the budget to placate their constituents (or more likely, lobbyists representing a minor fraction of the constituency). Yet no single member is responsible for creating a balanced budget. Hence, we end up with rampant spending and no one holding the reins on it. I don't intend to lay all the blame for this tragedy at the feet of the Congress, as clearly it's the voters' fault -- people want their elected representatives to funnel money to them with no concern for the budget as a whole.

GT and Society

My favorite trivial (in the larger sense of what is important, not in a simplistic sense) application of PD is from Kim Stanley Robinson's main character in his Science in the Capital series, Frank Vanderwal. Frank points out that people face PD every day of their lives in traffic. In many cases traffic jams are exacerbated by people acting short-sightedly; cutting each other off and so forth. This makes traffic worse for everyone even though it appears to help in a local fashion. Even so, we have it good here in the U.S. -- people are much more inclined to follow the rules (which exist to make things safe and efficient). Even if it's not completely optimal, compare this to driving in a third world country where there is little respect for traffic laws. Traffic jams are much worse. In a more serious vein, the "rule of law" is often very dependent on individuals trusting that other individuals will obey the same laws and not try to cheat the system. This conscious choice of every driver shows that people are able to solve PD in some fashion.

The traffic example illustrates a problem with a common (and more generally applicable) extension to PD: iterative prisoner's dilemma. In this scenario, the game is played multiple times and the players are able to base their decision on past history of how other players have behaved. This allows for much more interesting player strategies, such as a 'tit-for-tat' strategy where a betrayal in one round is repaid in a future round. A naive analysis seems to indicate that adaptive players will end up trusting each other and never betraying, but, alas, it's not true. Truly honest players are still vulnerable to sporadic betrayers. I find this a disheartening commentary on society

Making the prognosis worse is that iterative PD isn't always applicable to real-world situations. Consider traffic again: the number of players is very large and hard to track. It's just not possible to track that kind of information and thus it defaults to a single round of PD, which engenders much worse outcomes.

Solutions

In a broader sense, our current economic and political systems suffer greatly from an inability to see beyond the short-term. The current problem of global warming is an extreme example but very informative. Most reasonable people are willing to admit we have a problem, but no one bears responsibility because doing so would lead to bearing an unfair burden while others would profit. Capitalism in general suffers from various manifestations of this; CEOs get paid millions of dollars for increasing profits in the short term with no thought for the long-range consequences.

I want to reiterate that I value personal and economic liberty. Many people who share that view are hardcore capitalists and very pro-democracy. But capitalism and democracy are not the end-all of ecopolitical systems; to paraphrase Churchill, they are the worst forms possible, except for everything else that's been tried. Let's find a better way without losing our liberty. That's probably a pipe-dream, but dreams are good; they keep me entertained and hopeful.

I freely admit I'm short on solutions to these problems. But I strongly believe that if we can create social and economic structures that either provide more upside to the "correct" choices or more downside to the wrong choices, we will be much better off as a society. These kinds of questions are very common themes in Kim Stanley Robinson's works (in addition to variations on eco-economics) and I'll be blogging more about this in the future. Encouraging long-range thinking in individuals is a hard problem; doing so without being coercive is harder.

Given that long-range thinking must come from the individual, we need to encourage people to be selfish, but to do so in an enlightened manner. It's not about altruism, it's about convincing everyone, in toto, that we'd be much better off. But how do you stop people from turning around and exploiting the system as what happens in iterative PD? Is there a just and objective way to penalize people for this? Can we change the game?