Evidence on the performance of Zimbabwe's early land reform effort is mixed. In this paper we use a unique micro data set with information on land reform beneficiaries and a control group of nonbeneficiaries and employ propensity score matching to examine household and per capita benefits associated with land reform. Combining this information with estimates on the cost of land reform we find that the economic return is positive but modest. Its magnitude depends on assumptions about the opportunity costs of land and the reasons for the observed increase in the size of resettled households. © 2004 Elsevier Ltd. All rights reserved.