07-02: Farm size & productivity in Africa
Revisiting the farm size-productivity relationship based on a relatively wide range of farm sizes: evidence from Kenya
Michigan State University, United States of America
This paper revisits the inverse farm size-productivity relationship in Kenya. The study makes two contributions. First, the relationship is examined over a much wider range of farm sizes than most studies, which is particularly relevant in Africa given the recent rise of medium- and large-scale farms. Second, we test the inverse relationship hypothesis using three different measures of productivity including profits per hectare and total factor productivity, which are arguably more meaningful than standard measures of productivity such as yield or gross output per hectare. We find a U-shaped relationship between farm size and all three measures of farm productivity. The inverse relationship hypothesis holds on farms between zero and 3 hectares. The relationship between farm size and productivity is relatively flat between 3 and 5 hectares. A strong positive relationship between farm size and productivity emerges within the 5 to 70 hectare range of farm sizes.
Does mechanization reverse the farm-size productivity relationship? Evidence from Ethiopia
World Bank, United States of America
Although inverse farm-size productivity relationship is a recurring evidence in the literature of agricultural production in sub-Saharan Africa, almost all the empirical applications use data from smallholder agriculture. This paper aims at contributing to the ongoing debate by investigating the relationship over a large variation of farm sizes for major crops commonly grown by small, medium and large farms in Ethiopia. The analysis is further expanded to value of output and some measure of “profit” depending on availability of input prices (e.g., for family labor). The wide variation in farm size allows us to assess the effect of farm heterogeneity arising from level of mechanization which requires contiguous piece of land. We use data from two rounds (2014 and 2016) of smallholder panel household survey (less than 10 ha) and two rounds (2014 and 2015) large and medium panel commercial farm survey (10 ha and above).
Can large farm spillovers foster smallholders structural transformation? Evidence from Zambia
1World Bank, United States of America; 2IAPRI
A decade after the global commodity price boom led to a wave of land acquisition in developing countries (and foreigners have given way to locals), the extent to which such investment can act as a catalyst for structural transformation (and by implication policies to maximize such effects) remain poorly understood. Combining a smallholder survey with data on ‘emergent’ farmers shows that mechanization and substitution of purchased inputs for labor allows the latter to outperform small producers. Smallholder farmers located close to emergent farmers who provide traction services to neighbors benefit from spillovers in terms of access to traction, fertilizer use and land productivity. Policy implications of the fact that positive spillovers remain local and limited to ‘small’ emergent farmers are drawn out.
Does sample truncation affect assess the inverse farm size-productivity relationship? Evidence from Malawi
1World Bank, United States of America; 2University of International Business and Economics, Beijing, China, People's Republic of
To explore if the focus on household-based samples characteristic of most studies of the inverse farm-size productivity relationship (IR) affects results and policy recommendations, we complement household survey data from Malawi with a representative survey of estates. For a wide range of specifications, a strong IR between area operated and yield disappears if profits valuing family labor at market rates are used. An IR at farm level holds irrespectively of the sample but disappears at plot level for estates, supporting the notion of imperfections in labor market that affect smallholders disproportionately being at the root of the relationship. For corporate estates, the IR for yield disappears and a significant negative relationship between farm size and labor use per ha emerges.