Revisiting the old debate : on the relationship between size and productivity in Tanzania

The debate on the existence of an inverse relationship between farm size and productivity is probably one of the oldest debates in the development economics literature. While publication of more detailed agricultural data has pushed for an empirical revival of the topic, the concept of size is still problematic in these studies, as well as the limited attention given to existing varieties of farming practices. Using agricultural data on Tanzania, we introduce a crop/plot level of analysis which allows us to enquire whether an inverse relationship exists for crops grown on a given plots. In a context where intercropping is widespread, this level of analysis looks more appealing than the more traditional plot or farm levels. We control for the existing hypotheses in the literature that could explain the existence of the relationship. Further, we propose to control for a new set of hypotheses which have not received enough attention in the existing literature. Our results show that the inverse relationship is strikingly robust at this new level of analysis: yields are on average higher on smaller cultivated areas in all specifications and for all crops.