Working Paper Number
Access to land and particularly its distribution has reemerged as an important part of both academic and policy discussions in the last decade, leading to the resuscitation of the debate on the relationship between size of holdings and output per land unit. Across the world, studies have suggested the existence of a decreasing relationship between land size and output per unit of land. The most-widely accepted explanation for this relationship is that households with smaller holdings tend to be labor rich relative to land, and therefore can achieve higher output through the increased application of labor. Despite the rich literature on this topic there has been little work on whether this relationship is valid for female-headed households, particularly in the case of African countries. Past African studies have found female-headed households to be smaller by close to one adult in comparison to male-headed households. Given this difference one would expect there to be a difference in the outcome of land redistribution for different types of households, ceteris paribus. Additionally, the aggregate impact in African countries could be substantial, as female-headed households comprise in several cases up to 30 percent of the rural households. In this paper we will examine empirically whether the inverse size and output relationship is different between female and male headed households in the case of Kenya, using the Kenya Integrated Household Budget and Expenditure Survey of 2006, which includes modules on agricultural holdings and agricultural output in addition to the standard demographic characteristics. By controlling for the endogeneity of crop choice and fertilizer use we are able to find that cash crop production and human capital, and not differences in household size, determines the differences in male and female headed land productivity. Hence, our study goes beyond the simple discussion of the inverse relationship between land size and output per unit and the potential impact of redistribution. Specifically we will be able to address the kind of broad rural development policies in addition to land redistribution that would allow female headed households to do at least as well as (if not better than) male headed households.