The purpose of the current study was to investigate the relationship between team sport coaches’ power and coaching effectiveness using French and Raven’s (1959) taxonomy of power bases as a theoretical framework. Coaching effectiveness (CE) was conceptualized as an umbrella concept and four different CE outcomes were used; athletes’ satisfaction with the coach, coaches’ general influence, adaptive training behaviours, and collective efficacy. Hypotheses were made on the specific relationships between the individual power bases and the effectiveness criteria. The total sample consisted of 820 athletes (47% females), representing 56 elite and nonelite teams from three team sports (soccer, floorball, and team handball). Data were analysed separately for adults and youths. Structural equations modelling showed that 30% (in the youth sample) and 55% (in the adult sample) of the proposed hypotheses was supported. Overall, coaches’ bases of power were strongly associated with coaching effectiveness, explaining between 13% and 59% of variance in the effectiveness outcomes used. Expert power was consistently positively related to coaching effectiveness; reward and coercive power had mixed relationships (positively, negatively, unrelated) as had legitimate power (negatively, unrelated) and reward power (positively, unrelated). The results are discussed in relation to coaching effectiveness, limitations, practical implications and future research.