The primary factor leading to the success of any group, hierarchy, or institution is something that resonates through the whole of economics: game theory, development, organizational economics, industrial organization, public choice, and economic imperialism have found this principle necessary to the survivorship and effectiveness of group goals. This spills over into the real world when we begin talking about policies and the make up of organizational structures.
Group actions/policies must be individually rational for each member of the group.
You see, groups don’t make decisions, individuals do. In some sense, there is no such thing as collective action, but merely as an aggregate of the multitude of individual actions made by it members. Even then, there are multitudes of actions that happen outside the the control of said group, and these can also be taken into account as affecting the structure and perseverance of the group because of the transfer of information amongst individuals. Back to our point: Group actions must be individually rational. This is important, if the actions are not individually rational, many actions that are considered undesirable (from the group’s point of view) are more likely to occur.
In the example of providing public goods, it is individually rational for a farmer to allow more of his sheep to graze in the commons. No one is able to refuse him entry because the space is owned by no one and is open to all. (Demsetz has an excellent paper explaining why collective ownership is effectively the same as no ownership.) Since he has no incentive to allow less of his sheep into the common pasture and in fact is rewarded by the ability to have more sheep in lieu of lower costs from not having to purchase additional feed, he will continue bringing more sheep into the field. Eventually, more people will enter the commons, and each, having the same individual rationality as the first, will engage in the overgrazing of the commons. At which point, the public good the commons have provided will diminish to nothing and the farmers will have to turn elsewhere to feed their livestock. At no time until this point has it been rational for any one individual farmer to bring less livestock in the commons or cordon off specific plots to allow for regrowth of vegetation.
How do you incentivize individual action to make group action possible? In the case of the farmers and the commons, there are few ways to handle this. The first is for an entity to buy the commons and then ration out grazing via prices; the private entity can then take care of the commons and earn a profit (incentive) by doing so. The second is for the group of farmers once using the field as a commons to enter a collective agreement in which certain farmers graze on certain days, fees are taken up to pay for group entry and maintenance of the field, and to ensure cooperation, articles of punishment may be established. In larger forms, these rules of interactions are institutions. As long as there is a credible threat of being punished for defection, a collective agreement might be sustained. Other opportunities may arise, however, that are profitable for any one individual. In that case, it may be rational for an individual to switch to another group or collective agreement. In any event, it is the actions of the individuals that make up the actions of groups.
It can also be noted, that variations in the types of collective agreements and rules are the result of initial conditions and ongoing constraints. No one type of rule system is inherently better than any other unless you can first control for the constraints and then provide well-defined outcomes.
People engage in a specific group as long as it remains rational to do so.