With the current dissertation, I aimed to shed light on antecedents of abusive supervision from a leader perspective. First, I investigated whether leader narcissism is associated with abusive supervision (Research Question 1). Building on a two-dimensional approach to narcissism (Back et al., 2013), I differentiated between leaders’ narcissistic rivalry and admiration. Second, building on threatened egotism theory (Baumeister, Smart, & Boden, 1996; Bushman & Baumeister, 1998), I examined which underlying cognitive processes could explain the relationship between leader narcissism and abusive supervision (Research Question 2). Third, I tested how follower behavior influences narcissistic leaders and their underlying cognitive processes, thus evoking abusive supervision (Research Question 3). The present dissertation includes three manuscripts composed of two empirical studies each (see Chapters 2 to 4) and a mini meta-analysis corroborating some of the research findings (see Chapter 5). In the first manuscript (see Chapter 2), a direct positive effect of leaders’ narcissistic rivalry on abusive supervision was proposed along with a moderated mediation suggesting that supervisor-directed deviance will moderate the indirect effect of leaders’ narcissistic rivalry on abusive supervision via perceived self-esteem threat. Hypotheses were tested in two studies: a field study with leader-follower dyads and an experimental vignette study with a leader sample. Across both studies, I found that leaders high in narcissistic rivalry were more likely to show abusive supervision. However, this effect was independent of followers’ supervisor-directed deviance and leaders’ perceived self-esteem threats could only in part explain why leaders high in narcissistic rivalry had abusive supervision intentions. In the second manuscript (see Chapter 3), I hypothesized that leaders’ narcissistic rivalry would be positively associated with abusive supervision. Furthermore, I proposed that leaders high in narcissistic rivalry would be particularly prone to show abusive supervision in response to followers’ organization-directed deviance, but to a lesser degree in response to followers’ supervisor-directed or coworker-directed deviance. Finally, I hypothesized that leaders’ injury initiation motives, but not their performance promotion motives, would explain why these leaders showed abusive supervision in reaction to followers’ organization-directed deviance. I conducted an experimental vignette study and a mixed-methods study with leader samples to test the hypotheses. Across both studies, leaders’ narcissistic rivalry was positively associated with abusive supervision. Furthermore, only in Study 1 (but not in Study 2) there were differences in the effect sizes of leaders’ narcissistic rivalry on abusive supervision depending on the type of follower behavior, but not in the expected direction. Finally, leaders’ injury initiation motives, but not their performance promotion motives, explained why these leaders engaged in abusive supervision in response to followers’ organization-directed deviance. In the third manuscript (see Chapter 4), I proposed that leaders’ narcissistic rivalry, but not admiration, would be positively associated with abusive supervision. Furthermore, I proposed an indirect effect via leaders’ injury initiation motives and a moderation of this indirect effect by type of follower behavior (differentiating between counterproductive work behavior [CWB], organizational citizenship behavior [OCB], and task performance [TP]). Two experimental vignette studies with samples of working adults were conducted. Results revealed that only leaders’ narcissistic rivalry, but not their admiration, was positively related to abusive supervision. Furthermore, leaders showed abusive supervision because they experienced injury initiation motives. The indirect effect was significant in all conditions of follower behavior, but significantly stronger when followers showed CWB than when they showed TP. Finally, I conducted a mini meta-analysis (see Chapter 5) to obtain a more precise estimate of the relationship between leaders’ narcissistic rivalry and abusive supervision. More specifically, I conducted a meta-analysis of the effects of leaders’ narcissistic rivalry on abusive supervision from this dissertation’s primary studies. Results indicated that the association between leaders’ narcissistic rivalry and abusive supervision was moderately positive in size, which again stresses the idea that leaders’ narcissistic rivalry is an important precursor of abusive supervision. Overall, the findings of this dissertation underline the idea that abusive supervision results from a complex interplay between leaders’ personality, underlying cognitive processes, and follower behaviors. These findings expand the understanding of abusive supervision from a leader perspective and offer fruitful directions for future research. Limitations (e.g., in terms of theoretical and methodological considerations) are discussed along with practical implications for practitioners and organizations.