While research on abusive supervision is thriving, we still know very little about the sustained nature of the phenomenon. Additionally, most papers focusing on the prolonged character of the detrimental relational dynamic take a within-dyad perspective, largely ignoring within-person, group or other external influences. Addressing these gaps in the literature, we introduce the Barriers Model of Abusive Supervision. This model posits a hierarchically organized set of obstacles that make it difficult for followers to escape the abusive supervisor, explaining why abuse can continue over long periods of time. Specifically, we present an onion-shaped model in which the follower has a central position with each subsequent layer representing a more external cluster of barriers to leaving the abusive supervisor. Ranging from external to internal, these layers are: Barriers in the larger societal context (Layer 1; e.g., ambiguous laws), barriers in the organizational context (Layer 2; e.g., unclear policies), barriers due to the abusive supervisor (Layer 3; e.g., isolating followers), and barriers within the abused follower (Layer 4; e.g., implicit leadership theories). We hope that our model inspires future research on the sustained nature of abusive supervision and provides practitioners with the necessary background information to help abused followers escape their supervisors.