Abstract
The theory of infinitely repeated games may lack predictive power due to its insensitivity to, e.g., changes in some game parameters, the timing of players' moves and communication possibilities. We propose a new approach by studying an infinitely repeated prisoner's dilemma game and its variants where preferences for cooperation are heterogeneous, and strategic risk arises from incomplete information about opponents' preferences. Our model generates a rich set of comparative static predictions in a variety of settings. We show that, unlike standard theory and other existing models, our approach organizes the findings of a host of experiments including our novel experiments.
Original language | English |
---|---|
Publisher | SSRN |
DOIs | |
Publication status | Published - Aug 2018 |
Keywords
- cooperation
- infinitely repeated games
- strategic risk
- game theory
- experiments
- prisoner's dilemma