Kinship, Tenure Insecurity, Input Use, and Land Productivity: The Case of Sharecropping in Ethiopia

This chapter of the book assess the output levels and fertiliser input levels of kin and non-kin sharecropped tenants’ plots, using the sharecropping tenants’ owner-operated land as counterfactual.

There have been very many empirical studies of efficiency of sharecropping, especially in Asia (Otsuka and Hayami 1988). Most empirical studies compare efficiency of sharecroppers with owner-operators and fixed-rent tenants and have failed to control well for tenant and plot characteristics. Most studies have also failed to explain why specific sharecroppers are efficient or not. Exceptions include Sadoulet et al. (1997) who found sharecroppers with a kinship relationship with the landlord not to be affected by the Marshallian disincentive effect while other sharecroppers were. We follow up on this and assess the importance of kinship for sharecropping efficiency in our study in the Ethiopian highlands.

In this chapter we assess the output levels and fertilizer input levels of kin and non-kin sharecropped tenants’ plots, using the sharecropping tenants’ owner-operated land as counterfactual. In our empirical study in the Ethiopian highlands we were, unlike most other studies of sharecropping efficiency, able to control for self-selection due to unobservable household characteristics (by using of household fixed effects) and a range of plot level characteristics when assessing input use and output value on sharecropped plots of tenants with and without kin relationship with their landlords.We refer to Kassie and Holden (in press) for a complete representation of a theoretical model with kinship affecting land productivity directly and indirectly through its effect on probability of contract renewal. In this chapter we include new results on fertilizer input use on kin and non-kin sharecropped vs. owner-operated plots.

We found significantly higher output levels on sharecropped plots than on sharecropping tenants’ own plots. We also found significantly higher fertilizer input levels on non-kin sharecropped plots than on kin sharecropped plots and than on sharecroppers’ owner-operated plots. Both non-kin and kin sharecropped plots stochastically dominated sharecroppers’ owner-operated plots in terms of output value and non-kin sharecropped plots stochastically dominated kin sharecropped plots.