Purpose: The purpose of this paper is to build and test an integrative model of leader identity as an important mechanism explaining why reactions to leadership training associate with leader effectiveness. It is proposed that this mediation relationship is conditional on leadership experience (i.e. time in a formal managerial role), such that it will be weaker for more experienced leaders because they already possess complex leadership-related knowledge and skills. Design/methodology/approach: Hypotheses were tested using a sample of German managers (n=196) in formal leadership positions (i.e. with direct subordinates) across a range of industries. Data were collected using online questionnaires. The proposed first-stage mediation model was tested using the structural equation approach. Findings: Leader identity was found to mediate the relationship between reactions to leadership training and leader effectiveness. This mediation was conditional upon leadership experience, such that the indirect effect only held for less, but not for more, experienced leaders. Research limitations/implications: The findings should be interpreted with caution because all data are self-report and cross-sectional. Practical implications: Leadership training for senior leaders should qualitatively differ (in terms of content and length) from that for novice leaders. Originality/value: Leadership training can substantially improve managers’ ability to lead effectively. The present study is the first to establish leader identity as a motivational mechanism that explains this relationship. This is also the first study to test for the role of leadership experience in leader development.