Most of the literature on sharecropping deals with the reasons for the adoption of an allegedly irrational contract. Empirical testing of theoretical hypotheses is comparatively under-developed and the data on the worldwide diffusion of sharecropping are very scarce. This article argues that sharecropping is much less common than is sometimes assumed and that its share of total land has been decreasing. These ‘stylized facts’ are the ‘real’ puzzle that economic historians and theorists should try to explain.