behavior in field situations, mainly measured by attitudinal and self-report indicators in surveys. In Stern et al. (1999), it is shown, for example, that people who accept a movement’s values, who believe that things important to those values are threatened and who believe that their actions may help to alleviate the threat, experience a personal norm to support the movement.

The latter explanations as well as the fairness models presented in this paper assume different prosocial motives that mitigate free-rider incentives as suggested by the empirical findings. The theories do not, however, ask about possible evolutionary roots of such prosocial behavior. This important question is addressed in Richerson et al. (this volume:Chapter 12).20

Yet another remark is in place. We have emphasized the importance of reciprocity and inequity aversion but have not mentioned the impact of reputation and repeated game effects. Many of the real life common-pool resource or public goods problems are in fact “played” repeatedly. In such repeated interaction, players usually can condition their behavior on past behavior of others. This allows players to build up reputations and to ensure cooperative outcomes, even among selfish players. In the parlance of game theory, this kind of cooperation may be supported as an equilibrium in infinitely repeated games (folk theorems) or in finitely repeated games with incomplete information (see Kreps et al., 1982).21 Many experiments have demonstrated the efficiency-enhancing effect of repeated versus one-shot interactions. Moreover, it has been shown that reciprocity and repeated game effects interact in a complementary way (Gächter and Falk, in press). In Gächter and Falk’s experimental study of a bilateral labor relation, the reciprocal relationship between workers and firms is increased significantly in a repeated interaction compared to one-shot encounters. The driving force behind this “crowding in” of reciprocal behavior is the fact that people who behave selfishly in the one-shot game have an incentive to imitate reciprocity in the repeated game.22 Thus, in the presence of repeated game incentives, the prospects for cooperative outcomes are expected to be better than according to the one-shot analysis undertaken in this paper.



We also refer to the book by Ostrom et al. (1994) that summarizes and discusses the experimental findings. For an overview on experimental results, see also Kopelman et al. (this volume:Chapter 4).


The importance of reciprocity has been established in dozens, if not hundreds, of experiments. For rewarding behavior in response to kind acts, see Fehr et al. (1993) or Berg et al. (1995). For punishing behavior in response to hostile acts, see Güth et al. (1982). Recent overviews are provided in Ostrom (1998) and Fehr and Gächter (2000b).


For a first attempt to endogenize the choice of refernce agents or standards in a formal model, see Falk and Knell (2000).


All propositions are proved in the Appendix to this chapter.


The derivation of the social optimum is given in the Appendix to this chapter.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement