Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
A lack of representative insights into the different dimensions of land rights, alongside ownership, can hamper effective targeting of land reforms, particularly in low-income countries where customary tenure systems and low levels of documented ownership persist. Using machine learning and nationally representative intrahousehold survey data on various types of land rights and control, we examine landowner profiles and the bundle of rights emerging in Ethiopia, Malawi, and Tanzania. Our algorithm reveals that profiles in those countries are distinguished by two key features: whether landowners hold transfer rights and whether they hold those rights jointly with other household members. Our newly constructed profiles allow us to explore ownership issues that are less understood in the literature, such as how bundles are distributed across gender and lineage traditions (a contentious issue between statutory and customary law), the rights of individuals who use land but do not claim ownership, and the rigidity of rights in a bundle.