Gender and land productivity on rented land in Ethiopia

This paper seeks to find explanations to why land productivity is lower on land rented out by female landlord households than on land rented out by male landlord households. First it is demonstrated that this productivity differential is highly significant after controlling for land productivity using nearest neighbour and kernel plot matching methods. Then we test alternative hypotheses that possibly may explain the productivity differential. These include transaction costs in the land rental market, larger problem with Marshallian inefficiency on land rented out by female landlords, poorer ability to screen good tenants and evict bad tenant by female landlords, and larger proportion of pure landlords (with poorer monitoring capacity) among female landlords. Using GLS and controlling for sample selection we found evidence of Marshallian inefficiency in the analysis but this could not explain the gender productivity differential. Female landlords have tenants who are older, own less oxen, are more related, and under longer-term contracts. In the parametric regressions land productivity was significantly lower on plots rented out to in-law tenants and this appeared to explain the gender productivity differential. Female landlords may be forced to rent their land to their inlaws and they are unable to evict them even if they are inefficient land users. Land reforms that strengthen women’s land rights may therefore be good for efficiency as well as equity. Recent land regulations restricting renting out of land to maximum half of the farm area may contribute to weaken women’s land rights as a large share of the poor female landlords currently rent out more than half of their land because they lack the necessary nonland resources to farm efficiently.