|
| |
| Miloudi Kobiyh and Ahmed Doghmi |
| |
| ''Learning Process and Emergence of Cooperation: The Prisoner''s
Dilemma Case'' |
|
| |
| |
| In recent years, the laboratory evidence confirms that people do care about other''s payoffs as well as their own (Cox et al. 2007). In experiments, one observes that people cooperate much more than predicted by standard economic theory. Charness and al. (2007) observe a reasonably high degree of cooperation in their games. Evidence from the laboratory and the field has begun to persuade
economists to develop specific models of how and when a person''s preferences depend on others'' material payoffs (Cox and al. 2008).
To explain these experimental results we which introduce learning in the game. Generally, two varieties have received the most scrutiny in experiments: belief learning models and reinforcement learning models. Both models can be combined in flexible hybrid models experience-weighted attraction (EWA)(Camerer and Ho, 1999). We attempt to explain how cooperation emerges using a model based on learning process under the EWA model. We present a learning model in the repeated prisoner''s dilemma game. Then, we examine the abilities of learning models to describe subject behavior in experiments.
In view of the importance of fairness and reciprocity (Fehr and Falk, 2002), we formalise this learning process in terms of a learning sequence, and we show that if players are motivated by fairness and reciprocity, then equilibrium behaviour accords well with the aforementioned stylized facts.
|
| |
| |
| Keywords: Learning, incentives, prisoner''s dilemma |
JEL: C7 - Game Theory and Bargaining Theory:General
C9 - Design of Experiments: General |
| |
| Manuscript Received : Feb 22 2008 | | Manuscript Accepted : Feb 22 2008 |
|
