GUARDING ENGAGEMENT AGAINST ABUSIVE LEADERSHIP: WHY AND WHEN BEHAVIOR-SPECIFIC STRESS AND EMOTIONAL EXHAUSTION OPERATE AS PARALLEL MECHANISMS, AND THE CONDITIONING ROLES OF PSYCHOLOGICAL SAFETY AND PERCEIVED ORGANIZATIONAL SUPPORT IN PAKISTANS IT INDUSTRY

Abstract

Purpose: The literature on abusive leadership typically focuses on one strain pathway to understanding how it affects employee engagement. The focus in this study is instead on two parallel mechanisms, namely on behavior-specific stress (ST) and on emotional exhaustion (EE), and on which conditions the latter is weaker and more pronounced. In addition to the boundary condition examined on the second stage – perceived organizational support (POS) – psychological safety is also explored as a boundary condition on the first stage in the model (abusive leadership → each mediator).Design/methodology/approach: A cross-sectional survey was administered to [N = 350] information-technology employees in Pakistan, recruited through referral sampling. Abusive leadership was specified as a higher-order construct formed by four behaviors: credit stealing, belittling, yelling and scapegoating. Data were analyzed by means of partial least squares structural equation modeling (PLS-SEM) in SmartPLS 4, with 5,000 bootstrap resamples, the two-stage approach for interaction terms and an index of moderated & mediation.

Findings: Abusive leadership is positively related to both mediators; both mediators are negatively related to engagement; and the indirect effects are weaker when psychological safety and (perceived organizational support) POS are high. Implications: Findings are interpreted through Affective Events Theory and Conservation of Resources theory. Practically, behavior-focused leader development, psychologically safe team climates and visible organizational support appear to work best in combination rather than separately. Originality/value: To our knowledge, this is the first study to test behavior-specific stress and emotional exhaustion as parallel mediators of a disaggregated abusive-leadership construct within one moderated mediation model that includes both a first-stage and a second-stage moderator.