Sections 2.2 and 2.3 revealed an important difference between the outcome of a cooperative game and the outcome of a non-cooperative game. If the institutional setting is such that a game is cooperative, then the agents will fully exploit the gains from trade and reach a Pareto optimal equilibrium. If, however, the game is noncooperative, each agent’s pursuit of his self interest results in an equilibrium which is usually Pareto inferior. The games in Sections 2.2 and 2.3 were static. This section shows that the distinction between cooperative-game equilibria and non-cooperative-game equilibria is blurred in dynamic games: in a dynamic noncooperative game, there may be equilibria which are Pareto optimal. For this to happen a specific type of dynamics must be introduced: the agents adopt strategies in which their current actions depend upon the past actions of the other agents. Strategies of this form allow the possibility of retaliation, and so cooperative behavior can be enforced in a noncooperative context. This insight must now be given precise content.