Conditional Cooperation and the Effect of Punishment

Abstract

We study how punishment influences conditional cooperation. We ask two questions: 1) how does conditional cooperation change if a subject can be punished and 2) how does conditional cooperation change if a subject has the power to punish others. In particular, we disentangle the decision to be a conditional cooperator at all from the strength of conditional cooperation. We find that the possibility of being punished increases the strength of conditional cooperation. At the same time the possibility of being punished increases the number of free riders. In our study the net effect on cooperation still is positive. The possibility of punishing others has two effects: Substitution and responsibility. Players substitute conditional cooperation with punishment which leads to a decrease in conditional cooperation. The power to punish means more responsibility which leads to an increase in conditional cooperation. In our design the overall effect of responsibility is stronger than the effect of substitution. We conclude that the threat of being punished and the power to punish changes conditional cooperation behavior in several, unexpected, ways.

Punishment: one tool, many uses

“…2015, 2016). Similar results have been obtained in the context of n -player games (de Melo and Piaggi 2015; Kirchkamp and Mill 2018; Rand et al . 2009; Sefton et al .…”

“…Relatedly, it is unclear to what extent punishment increases cooperation by converting defectors into co-operators (i.e. increasing the number of co-operators), or by increasing the contributions of conditional co-operators (see Kirchkamp and Mill 2018;Lergetporer et al 2014;Shinada and Yamagishi 2007;Yamagishi 1986). In a recent experimental study, the number of free-riders increased under the threat of punishment, relative to a condition where no punishment was possible (Kirchkamp and Mill 2018).…”

“…increasing the number of co-operators), or by increasing the contributions of conditional co-operators (see Kirchkamp and Mill 2018;Lergetporer et al 2014;Shinada and Yamagishi 2007;Yamagishi 1986). In a recent experimental study, the number of free-riders increased under the threat of punishment, relative to a condition where no punishment was possible (Kirchkamp and Mill 2018). By way of explanation for this counter-intuitive result, the authors posit either a possible crowding out effect (where any intrinsic motivation to cooperate might be extinguished by the extrinsic threat of punishment) or the possibility that people expected to be punished and thus attempted to minimise payoff losses ex ante by defecting.…”

“…Relatedly, it is unclear to what extent punishment increases cooperation by converting defectors into co-operators (i.e. increasing the number of co-operators), or by increasing the contributions of conditional co-operators (see Kirchkamp and Mill 2018; Lergetporer et al . 2014; Shinada and Yamagishi 2007; Yamagishi 1986).…”

False beliefs can bootstrap cooperative communities through social norms

“…Experimental research has shown that the threat of punishment can increase conditional cooperation (Fehr & Gachter, 2000; Lergetporer et al, 2014). However, punishment has also been shown to increase cheating, which might be due to defectors compensating for being punished by attempting to extract more from the public good (Kirchkamp & Mill, 2020). Players’ knowledge that cheaters can be punished can impact their behaviours.…”