One way of getting round a CollectiveActionProblem is to change the structure of the problem by introducing individual incentives for group members.

One way to do this is to have an 'Authority' that makes threats and offers to citizens to encourage them to act in a certain way - eg. according to a group plan of action that will help bring about a desired group outcome. In a public goods problem, the citizen’s decision was – act socially or free-ride. It now becomes - act as authority instructs or be punished.

There can be different kinds of authority roles, eg.

  • 'Experts' who decide what the group plan is
  • 'Monitors' who check that agents follow the plan
  • 'Enforcers' who punish agents who don't

The question is - what is the decision problem for the authority?

The authority is also made up of of rational agents, who will also have incentives and decisions to make. (It may often make sense to think of the 'authority' as a 'sub-group' of the whole group of agents involved in the CAP.) Will authority members have incentives to do their jobs? If the authority group has more than one member, will there be a new CAP created within the authority group? Will authority members abuse their positions?

I want to explore some of these issues using GoatPolice.

OK, so the argument is that hierarchy arises when you need to control people's behaviour to prevent CommonsProblems? I guess the other sort of policing is a distributed ImmuneSystem. Does that automatically form itself into a hierarchy? – PhilJones