Motivations for Indirect Reciprocity

Abstract

We test the leading models of reciprocity in a new experimental framework that introduces strategic considerations. In an augmented rendition of the classic helping game, participants (referred to as Helpers) are asked to help a Recipient, and this choice is observed by a third-party Observer with either high or low probability (pseudo-private). The Observer is then given the opportunity to help the Helper. We study the strategic response of Helpers to the changing probability of being observed, and the extent to which Observers condition their indirect reciprocity (i.e. reciprocity towards someone who helped someone else) on those strategic incentives. Then, in a further modification to the helping game, Helpers are asked to help two Recipients: one necessarily observable, the other in pseudo-private. Afterwards, an Observer has the discretion to help the Helper. In this variation, there is now the possibility to directly witness guile, which we define as prosocial behaviour done in public but not when observation is uncertain. Our results demonstrate patterns of behaviour that cannot be explained by the existing models of reciprocity. Furthermore, we find strong evidence that Observers punish guileful behaviour while accounting for helpful intentions, altruism, and distributional preferences.

About the presenter's meeting

If you would like to meet with Kieran, please contact Professor Lionel Page (lionel.page@uq.edu.au)