Gender quotas to increase women’s representation are often motivated by the assumption that men and women have different policy preferences. In Africa – where gender quotas have been particularly widespread – we find that gender differences in preferences are quite small on average, but vary significantly across both policy domains and countries. We propose a theoretical framework for differentiating policy domains where preference divergence indicates increased gender parity from those where it signifies growing inequality. We then demonstrate that favourable gender gaps increase with female labour-force participation, while unfavourable gaps are more likely where women are most vulnerable. We show that these gender gaps in preferences are related to gender gaps in both political participation and representation.