Decentralized Peer Punishment in Social Dilemmas: Identification at the Individual Level, Type Heterogeneity, and Consistency across Games
Final Report Abstract
The possibility of peer punishment, i.e., mutual monitoring and sanctioning, is present in almost any form of social interaction. Given its prevalence, the project aimed at a comprehensive analysis of peer punishment behavior at the individual level. The main focus was on the elicitation and classification of individual peer-punishment behavior in cooperation problems. Data from controlled laboratory studies revealed that subjects’ individual punishment behavior can be broadly classified into three categories: i) non-punishers, who do not engage in costly social sanctions, ii) pro-social punishers, who use social sanctions to enforce norms of cooperation, and iii) anti-social punishers, who show patterns of punishment that are targeted towards cooperators. When combining the novel individual punishment measure with an established classification of individual cooperation behavior, we observed two behavioral archetypes: for about 65% of our subjects, there was an overlap in their pro-social behavioral patterns, suggesting that punishment and cooperation may be driven by a common altruistic motivation; but for the remaining ∼ 35%, punishment and cooperation were not ‘two sides of the same coin’. Given that such a significant part of the population displayed a divergence in punishment and cooperation patterns, we were able to explore how the interplay between these types shaped a group’s cooperation success in repeated interactions. It turned out that particular the presence of pro-social punishers was crucial for sustaining cooperation, especially in large, steadily alternating groups. Additional experiments showed the effect to be most pronounced w.r.t. strategic pro-social punishers; even though at the same time, the data confirmed that punishment was primarily, but not exclusively, compatible with retributive motives of punishment. Finally, another series of experiments showed that similar behavioral patters can be found in coordination problems. Moreover, the comparison of subjects’ punishment behavior between coordination and cooperation games suggested punishment types to be very stable at the individual level. Overall, our findings help to draw a more complete picture of cooperation and punishment behavior, but future studies should reassess the underlying motivations of peer punishment; ideally by complementing our findings with data from games that are more complex, of a different kind, or with different punishment consequences. This should lead to more precise micro-foundations of cooperative behavior and, ultimately, to rigorous theoretical models. Still, already our results should be of interest to both researchers and practitioners likewise; in particular for those who work on questions related to i) (the adaptation of) institutional designs for fostering cooperation, ii) group formation issues, or iii) optimal team composition.
Publications
-
(2015): “On Probation – An Experimental Analysis.” Journal of Empirical Legal Studies, Vol. 12 (2), pp. 252 - 288
Engel, Christoph, Heike Hennig-Schmidt, Bernd Irlenbusch and Sebastian Kube
-
(2018): “Cooperation and Norm Enforcement: The Individual-Level Perspective.” Journal of Public Economics, Vol. 165, pp. 1-16
Albrecht, Felix, Sebastian Kube and Christian Traxler
-
(2018): “Peer-Punishment in a Cooperation and a Coordination Game.” Games, Vol.9 (3), 54
Albrecht, Felix and Sebastian Kube
-
(2018): “The joint benefits of observed and unobserved social sanctions.” Journal of Behavioral and Experimental Economics, Vol. 75, pp. 105 - 116
Glöckner, Andreas, Sebastian Kube and Andreas Nicklisch
