Proposals for post apartheid agrarian reconstruction in South Africa are founded on assumptions about gender relations and households that obscure the underlying basis of gender differentiated access to productive resources. Based on their assumptions, policy makers believe that legalistic approaches in themselves will be sufficient to redress existing gender inequalities in resource access. However even when such policy intends to benefit women they may end up reinforcing existing gender imbalances precisely because these very policies get reshaped by existing gender and household relations on the ground. A more nuanced understanding of processes of resource allocation stems from the realization that access to resources is shaped by a persons membership and status within a household as well as by the positions a person holds in non-domestic spheres. Household dynamics are diverse and evolving. A household is not a bounded, static unit. Instead its boundaries are permeable, and it is linked to other households as well as to wider social and economic structures. Relations within the household are characterized by co-operation and conflict. Macro-historical processes of state formation and capital accumulation in dynamic tension with interand intra-household relations shaped access to resources over time. Within these processes gender ideologies play a vital role in reproducing and maintaining gender hierarchies at the level of the state, the economy, and households. Women's disadvantaged position in resource access is a result of a complex combination of factors. Access to resources, power, and authority is hence not shaped by law alone. Women's ability to take advantage of legal rights is linked to questions of wider authority and power in society. This suggests that while legal solutions are necessary they are not sufficient to redress gender inequities in resource allocation. Hence policy will have to rely on empowerment strategies that challenge the basic power relations in society. Nuanced understandings of local situations, rather than broad generalizations and idealized constructs will better inform such strategies.