We examined the moderating role of narcissism – a personality variable associated with overconfidence – on the reciprocal relationship between self-efficacy and performance. Participants (N = 87) completed ten experimental trials on a driving simulator and we recorded participants’ performance and self-efficacy beliefs across trials. Hierarchical linear modelling demonstrated that performance had a positive relationship with self-efficacy (supporting the majority of self-efficacy research). However, narcissism moderated this relationship. Specifically, when narcissism was high, performance had no relationship with subsequent self-efficacy. Conversely, self-efficacy had a significant negative relationship with performance and narcissism did not moderate this relationship. A secondary purpose of the study was to examine the role of narcissism in the relationship between self-efficacy and effort, and between effort and performance. Narcissism significantly moderated the relationship between self-efficacy and self-reported effort, and between self-reported effort and performance.