While 80 percent of all HIV-positive women in the world live in sub-Saharan Africa, there is substantial variation in infection rates across the continent. This paper traces current variation in female HIV rates within Sub-Saharan Africa to women's position in pre-colonial ethnic group's kinship organizations. In matrilineal kinship systems, lineage and inheritance are traced through female members. The structure of matrilineal kinship systems implies that, relative to patrilineal kinship systems, women have greater support from their own kin groups, and husbands have less authority over their wives. Using within-country variation across about 280,000 individuals in their ethnic group's ancestral kinship organization, in 18 countries, I show that females originating from ancestrally matrilineal ethnic groups are today more likely to be infected by HIV. This result is robust to the inclusion of subnational fixed e˙ects, as well as a large set of cultural, historical, geographical, and environmental factors that could be confounding the estimates. I further show that my result is not driven by di˙erential selection into HIV testing, nor by dif-ferences in general health status, and that omitted variable bias is very unlikely. Computing distance from DHS villages to nearest ancestral matrilineal ethnic boundary, I use a number of alternative estimation strategies, including and instrumental variable (IV) strategy and a geographic regression discontinuity design (RDD) at the ancestral matrilineal ethnic boundary, and find robust estimates. I provide several mechanisms to this finding: first, benefiting from more sexual autonomy, matrilineal women adopt riskier sexual practices which are more conducive to HIV and suggest channels of transmission of the virus alter-native to simple infection by the husband. Second, despite a better ability to impose safe sexual practices to their husbands, matrilineal women are less likely to use protective contraception methods such as male condom. I provide evidences that behavioural mechanisms are driving this counterintuitive result: I find that, relative to their patrilineal counterparts, matrilineal women are more likely to (1) perceive condom as an ex-post protective method against HIV transmission rather than an ex-ante preventive method against HIV infection; (2) underestimate their husband's propensity to be unfaithful. I conjecture that this latter result is an adverse e˙ect of matrilineal women's overestimation of the eÿciency of their intra-household threatening strategy against male infidelity.