Scrutinizing Punishment Institutions: A Lab Experiment on Alternative Punishment Regimes

This experimental paper provides evidence on the incidence of different third-party punishment institutions in the context of a game designed to represent simplified versions of specific, real-life situations. We begin with a benchmark scenario (which we will call "Baseline"), that involves the possibility of altruistic third-party punishment after playing a taking game. The "vertical control" treatment adds a second possible punisher who can confirm or overturn the initial decision. The "giving reasons" treatment requires the third party to provide reasons to motivate her/his choice. A fourth treatment combines both instruments. Thus, we manipulate our interest variables ("vertical control" and "giving reasons") both one at a time and, eventually, in a combined format. The main result is that both instruments have a significant positive impact on the incidence of punishment. In contrast, the hybrid scenario does not further increase the frequency of punishment, seemingly due to a "mis-use" of the "giving reasons" requirement by arguing vis-& agrave;-vis the second instance to uphold the first-instance, non-punishment decision.