Expectations Research Papers - Academia.edu (original) (raw)

In his famous article “An Essay on Bargaining” (1956), Thomas Schelling introduced the concept of commitment into the game-theoretical framework. Committed behaviour, following Schelling, is a kind of strategic action whose goal is to modify the other players' strategies, through the manipulation of their expectations (Schelling 1960, 122). The two paradigmatic cases of social commitment would be promises and threats. Thus, for a commitment (either a promise or a threat) to be effective, it has to be both believable and believed by the other players. Schelling argues that credibility can be attained by the agent by means of different mechanisms, such as voluntarily ruling out one or more of the available options (either by making some choices impossible or by raising its cost), the power of reputation, and the ability to bargain. Nevertheless, as Schelling points out, credible promises and threats seem to lead to a paradoxical situation: once the commitment has been stated, and having already manipulated the other agent's choices and actions, what's the reason to keep it? Let's imagine, for example, that two agents, Sally and Bob, are playing a public goods game; Sally promises to invest all her tokens into the public good if Bob does the same. Bob believes that Sally's promise is credible, and invests all his tokens. Then, why would Sally keep her promise? From the point of view of narrow strategic rationality, Sally's best option is to free-ride, not only before she made the promise, but also after (and probably she would be more tempted to do so after Bob invested all his tokens in the public good). This problem can be stated as follows: the incentives to cooperate are exactly the same than the incentives to keep a promise or a threat. Nonetheless, experimental results show a significant tendency to act according to one's commitments, as well as a tendency to consider the other's commitments credible (Ostrom, Walker, and Gardner 1992; Kerr and Kaufman-Gilliland 1994; Kerr et al. 1997; Kurzban et al. 2001; Boadway, Song, and Tremblay 2007; Balliet 2010). It would be unrealistic to state that everyone lives up to their commitments, but the puzzling results do not have to do with the amount of promises and threats that are kept or broken; they are puzzling because, everything else being equal, keeping a commitment is irrational if the behaviour promised is irrational as well—and if it is not irrational, because it is the agent's preferred choice, then the commitment is unnecessary.
This Chapter deals with the strategic problem of commitment, as stated by Schelling.