This study investigates the effect of Vice-Chancellor (VC) narcissism on measures of university performance in the UK. We show large differences in narcissism levels across VCs and that Oxbridge graduate VCs are more likely to exhibit traits of narcissism. We argue that this finding points towards the direction of a strong link between membership of a “social elite” and narcissism. We then exploit the quasi-natural experiment of VC changes and employ a difference-in-differences approach; this approach, which alleviates concerns related to endogeneity and identification bias, aims to infer the causal relationship between VC narcissism and university performance. We show that the appointment of a high narcissist VC leads to a deterioration in research performance as well as in league table performance. Our findings show support of the destructive leadership theory and have practical implications for leadership recruitment and the monitoring of leadership practices in the higher education sector.