Imagine a world in which everyone pursued the same goal. It would appear perfectly plausible that each person would put forth her individual efforts to achieve this goal. Suppose, for example, that everyone wanted, more than anything else in the world, a public granary. With such an overriding interest the granary would be built. Suppose, however, that each individual's goal was to have the granary but not contribute to its construction in any way. If this were everyone's goal, then no one would contribute to the construction of the granary and, therefore, the granary would not be built.

This simple example illustrates one of the many problems faced by a group of individuals, called in the literature collective action problems. Collective action problems range in nature from how a group of individuals becomes a cohesive society (e.g., Hobbes or Rousseau) to how a society provides itself with self-defense (Laver, 1985). Jon Elster, taking on tones of Hobbes, succinctly puts forth the research question behind the study of collective action problems. He asks, "What is it that glues societies together and prevents them from disintegrating into chaos and war?" (Elster, 1989) This demands that we look to the individual for our answer. But in looking to the individual, where do we start? It is easier to say where we should not start--namely with social norms. While Elster (1989) says that "social norms provide an important kind of motivation for action," one must remember that social norms arise from social interaction, even if the exact means by which this occurs is not yet known. We, however, are starting with the individual, who is as yet untouched by social interaction. I thus adopt the assumption of self-interest for our individual, in the hope that a social future can be arrived at from this primitive.

Of the many theories of how individuals do (or should) behave, none is more austere than the theory of self-interest. The theory of self-interest, as its name suggests, assumes that an individual does what she perceives to be in her best interest. As to the validity of this assumption, consider the regulation of society. Rules are considered necessary for the maintenance of society because without rules people would engage in socially destructive behavior more often, such as plundering private property. However, even with rules in place, much time and energy goes into attempting to thwart these regulations--consider, for example, tax evasion. Thus, it is plausible that people are indeed self-interested. And what would happen in a world where everyone is self-interested? The theory of self-interest, also known as rationality, paints a grim picture of this world. The grimness is associated primarily with two propositions of rationality--the Voters' Paradox and the aforementioned collective action problem. This investigation is only concerned with the collective action problem, in which each participant has a choice between giving her resources to a group project or consuming all of her resources for her own benefit. Even if each individual desires to see the group project completed, each individual could have preferences such that the group project is never started. That is, if each individual prefers that all other individuals contribute to the group project while she herself consumes her own resources, then no individual will contribute to the group project.

Formally, then, a collective action problem is any failure of a group of individuals to achieve an outcome everyone one likes at least as well as the outcome that results from everyone acting in her own self-interest. This problem of collective action can be divided into two general classes. First, individuals could fail to coordinate when coordination is preferred to not coordinating (Taylor, 1976, 1987). Consider, for example, driving. Each individual prefers that everyone else drive on the same side of the road (from the driver's perspective) as she does. However, there are two possible ways to accomplish this. We can assume that no one really cares which side is chosen as long as everyone chooses the same side. In a world which has not yet solved this problem through established rules, everyone is presented with quite a quandary. This class of collective action problem, while one of the most common problems faced by society, is only beginning to be thoroughly studied. Although not part of my principal research, we will touch on the importance of coordination problems later when dealing with larger games.

The second class of collective action problem involves a failure to achieve an outcome everyone prefers over the outcome arrived at because each individual wanted to achieve her most preferred outcome without, in essence, paying for it herself. For instance, everyone rationally withholds contributions from the group project, resulting in a socially inferior situation in which the group project is not obtained. This is much more serious than the first class of collective action problems in the sense that, while coordination can occur by chance, each individual in this class of collective action problems has a strictly dominant strategy to follow. That is, an individual is better-off not contributing to the group project no matter what the others do. With such a strict preference for the noncooperative strategy, rational individuals faced with a one-time collective action problem of this kind can never achieve the group project--not even by accident. Two well known examples of this class of collective action problem are the Tragedy of the Commons and the Prisoner's Dilemma (Taylor, 1976, 1987; Hardin, 1968, 1982; Axelrod, 1984). Of particular interest to me is the Prisoner's Dilemma, as it is the simplest of this second class of collective action problem. Through the use of the self-interest assumption and formal modeling (i.e., game theory), it is possible to create a representation of society as an equilibrium of the formal model. The Prisoner's Dilemma game is one building block of such a model.