An important branch of economic research on emotions has used power-to-take game
experiments to study the impact of negative emotions, such as anger, irritation and contempt,
on the decision to punish. We investigate experimentally the role that the specific
punishment technology adopted plays in this context, and test to what extent punishing
behavior can be truly attributed to negative emotions. We find that a large part (around
70%) of the punishment behavior observed in previous PTTG studies is explained by the
technology of punishment adopted instead of negative emotions. Once this effect is
removed, negative emotions do still play an important role, but the efficiency costs associated
to them are much smaller.