Oye: Explaining cooperation under anarchy

Oye. 1986. Explaining cooperation under anarchy: Hypotheses and strategies. In Cooperation under Anarchy, ed. Oye, Princeton: Princeton University Press.

See Fearon 1998 for a formal contradiction to two points made here. First, Fearon claims that simplistic models like prisoner's dilemma are not the appropriate model to understand bargaining situations; they analyze how states might enforce cooperation without explaining how they decide on the terms of cooperation in the first place. He also claims that the "shadow of the future" (iteration, in Oye's article) is not necessarily a good thing, as it may increase incentives to hold out in negotiations for a better deal.

[from handout]

Three important factors affecting cooperation:

1. Payoff structure
   • The concept of "cooperation" is not applicable to some games (Deadlock and Harmony)
   • In games where the concept of "cooperation" is applicable (like Prisoners' Dilemma, Stag Hunt, and Chicken), two payoff-related factors affect the prospects for cooperation
     • Changes in payoffs can change the game (i.e. from Prisoners' Dilemma to Stag Hunt)
     • For iterated games, the magnitude of the payoffs is important for determining how long the "shadow of the future is" (i.e. if payoffs increase over time, that is good for cooperation)
   • Ways to change payoff structure
     • Unilateral strategies (like building defensive weapons instead of offensive ones to overcome the security dilemma)
     • Bilateral bargaining (like issue linkage)
     • Multilateral negotiation (like using institutions to create norms that appeal to domestic constituencies and to share information among states)
2. Iteration ("the shadow of the future")
   • Iteration improves the prospects for cooperation for Prisoners' Dilemma and Stag Hunt but may not do so for Chicken
   • Effectiveness of strategies like Tit-for-Tat depends on two factors
     • Ability to distinguish reliably between cooperation and defection (i.e. could fail due to poor definition of concepts, to lack of transparent information, lack of control over sub-national agents, etc.)
     • Ability to punish defectors (i.e. could fail due to collective action problems)
3. Number of players
   • "As the number of players increases, transaction and information costs rise."
   • As the number of players rises, discount rates for future payoffs are more likely to be heterogenous, which increases the risk that one state will rationally defect which could send the whole game into a defection spiral
   • As the number of players rises, punishing becomes a collective action problem
   • Multilateral institutions can reduce information and transaction costs and can have clauses for collective punishment in the event of defection

Opinions: Nice overview for a symposium. Maybe in 1985, not that many political scientists were as familiar with these concepts as they are today. This paper/chapter does not really try to make an original contribution to our understanding of IR, so it is perhaps not fair to criticize it for failing to do so. But I think most people who have taken one semester of game theory would find most of Oye's points obvious. [from handout]