Free-riding on altruistic punishment? an experimental comparison of third-party punishment in a stand-alone and in an in-group environment

While second-party punishment is suitable in small groups, third-party punishment is much more common in large societies, where it is generally recognized as a social norm enforcement device that may guarantee social stability. However, in large societies, the presence of a potential additional third-party punisher who observes the norm violation and decides to intervene becomes more probable. The question arises as to whether third-party punishment would be robust with respect to an enlargement of the pool of potential altruistic punishers, namely the introduction of a second potential punisher. The relevance of this question is evident because, should the case be that the presence of several potential third-party punishers activates free-riding attitudes, third-party punishment may decline or even collapse altogether. In our paper we compare, by means of an economic experiment, punishment by a single third party (the Stand-Alone case) with punishment by third parties (In-Group environment). Shifting punishment choices into this "enlarged environment" allows us to study, in a systematic way, the complex relationship between the punisher's expectations about her/his peer's punishment decisions and her/his own punishment choices. Our data suggest that individual punishers are heterogeneous as to their individual punishment characteristics and the presence of a second punisher affects their choices to a certain extent. Consequently, the implementation of voluntary punishment depends on the distribution of types within the population. This result allows both to put into discussion the extreme emphasis devoted to voluntary third-party punishment as the "golden cornerstone" of spontaneous social order and to explain why large developed societies need institutional legal systems as the root of stability.