Bitter Sugar: Slavery and the Black Family

Abstract: We empirically assess the effect of historical U.S. slavery on the African American family structure. The latter is proxied by the likelihood that a household is headed by a single woman. Our hypothesis is that the black family structure is more likely to emerge in association with the experience of slavery in sugarcane plantations. Indeed, across the Americas, sugarcane planting has been linked with extreme demographic outcomes within the slave population - involving a prevalence of men over women, low fertility, and high mortality - that may have impeded the formation of stable families adhering to the nuclear model. We test our hypotheses on U.S. Census data covering the period 1900-1940. By exploiting the exogenous variation in suitability to crops, we establish that higher sugar suitability is associated with a higher likelihood of a single female headship. We complement OLS estimates with a matching estimator and a RDD. The effect we uncover starts fading in 1930 with the Great Migration, and vanishes by 2000. By constructing a matched linked dataset, we also assess intergenerational and geographical persistence and, by merging data on Louisiana slaves' ethnic origin with ethnographic data, we exclude an alternative explanation based of African cultural traditions.