This paper investigates how the possibility to ostracise, which is a familiar punishment mechanism to subjects in the experiment, affects harvest in a common pool resource experiment. The experiment was framed as a fishing problem and the subjects were young fishers in Ghana. We find that the introduction of the possibility to ostracise other members of a group at a cost to the remaining members of the group decreased over-fishing significantly in comparison with a situation where ostracism was not possible. The ostracism was based on at least 50% voting rule. Moreover, the subjects demonstrated a strong desire to ostracise those who over-fished. Copyright 2012 , Oxford University Press.