The welfare effects of GM (genetic modification)-led productivity growth for cassava producers are partly affected by the characteristics of individual cassava producing households. Those household characteristics include the elasticity of production and home consumption of cassava. Some studies assume the inelastic home consumption when conducting ex-ante welfare effects analysis for subsistence crops. This study modifies the estimation methods used in the past literature to estimate both elasticities using the dataset from Benin. Several assumptions are also tested regarding the heterogeneity of cassava producers. On estimation of elasticities, the paper tests the hypothesis that on-farm sellers are characteristically different from off-farm sellers by employing the double hurdle model. The findings contribute to the literatures analyzing the distributional effects of welfare effects from GM-led productivity growth for cassava, which are gaining importance in the context of the policy impacts on poverty reduction.