Recent decades have seen a worldwide diffusion of electoral gender quotas, an institution designed to enhance women's political representation. This has sparked scholars' interest in the actual effects of quotas, including their impact on female descriptive (numeric) representation in national legislatures. As with many other issues within the domain of comparative politics, however, establishing causality with respect to the effect of quota laws is intrinsically difficult. In this article, we attempt to overcome these difficulties. We apply the generalised synthetic control method to study the impact of legislated candidate quotas on the descriptive parliamentary representation of women in three European countries: Belgium, Greece and Poland. As these states elect their national parliaments by means of variants of a proportional representation system with preferential voting for candidates, we compare their levels of female descriptive representation to a set of corresponding figures observed for a 'synthetic control' unit, comprising 16 European countries that implement fundamentally similar electoral systems but do not have legislated candidate quotas in place. The quota effects that we obtain vary substantially among the three cases analysed. The apparent effect for Belgium is strong, while much weaker effects are observed for the two other states. Relying on the extant literature, we qualitatively interpret these results, putting forward a number of conjectures as to how contextual factors can contribute to the strength and persistence of quota effects. We conclude that the impact of legislated candidate quotas is essentially limited, albeit it may be magnified under favourable contextual conditions, such as those which occurred in Belgium.