Electoral clientelism or vote buying has been regarded as undermining democratic institutions and weakening the accountability of the state towards its citizens, especially the poor. Social identity as a form of political mobilisation may contribute to this, enabling support to be won with clientelist transfers. This paper reports data from a novel laboratory experiment designed to examine whether clientelism can be sustained as a political strategy, and whether identity impacts the nature or efficacy of clientelism. Specifically, we design a voting and leadership game in order to examine whether individuals vote for clientelist allocations by a leader even at the expense of more efficient and egalitarian allocations. We find group identity does not significantly impact the prevalence of clientelist plans. Leaders are more likely, however, to choose allocations that provide fewer benefits (lower rents) to themselves when they are part of the majority in-group than when they are in the minority.