Most of the literature on land tenure in developing countries asserts that agricultural performance is negatively affected by indigenous land tenure systems and farm fragmentation. This study is the first empirical test of these assertions for a North African country and is based on primary data collected from 841 plots on 131 farms in the Mnasra region of northwestern Morocco. The data analysis was designed to meet the following study objectives: (1) to analyze the prevailing indigenous and modern land tenure systems; (2) to determine the effects of tenure security on farm investment, input use, and agricultural productivity; and (3) to examine the impact of farm fragmentation on agricultural productivity. The results indicate that indigenous land tenure has evolved toward increasing privatization. Estimation of an OLS model revealed that farm fragmentation has no impact on agricultural productivity. In fact, farm fragmentation may be an appropriate response to the farmer's needs to achieve ecological diversity, handle labor shortages, and manage risk. Estimation of a 2SLS structural model indicated that land tenure security has a positive effect on long-term investments, but no direct impact on medium-term investment, input use, or agricultural productivity. However, tenure security does exert an indirect positive influence on input use and agricultural productivity through its effect on long-term investment. Policy implications of the study include the finding that consolidation and titling programs cannot be expected to automatically increase agricultural productivity.