Evidence from individualistic cultures suggests that power corrupts. Using a goals-based perspective, here we argue that power and culture jointly predict corrupt attitudes and behavior. Four studies ( = 447) and one meta-analysis were conducted to test these hypotheses. Study 1 investigated the joint effects of power and individuals' cultural orientations on corruption proclivity. Studies 2 and 3 assessed if power and cultural orientations affect actual corrupt behaviors (i.e. abuse of discretion in Study 2 and bribe-taking in Study 3). Study 4 tested the hypothesis at a national level, using monocultural samples both in the UK and China. The results consistently showed that the effects of power on corruption depend on culture: for collectivistic individual orientations and cultures, holding power predicts less corruption than lacking power; in contrast, holding power predicts more corruption for individualist orientations and cultures. Our findings represent the first direct experimental and correlational evidence regarding the links between power, culture, and corruption.