Can robust pro-female policies be identified when the true model of the household is unknown?

This paper seeks to identify parameter changes which are robust in the sense that they benefit women relative to men in a wide range of household models. The models considered are unitary, Nash- bargaining and non-cooperative with and without cash transfers. Reductions in the relative price of ‘female` consumer goods prices are robust; increases in relative wages are highly non-robust. Increases in the relative returns to domestic activities and transfers of financial, physical, and human assets to women are weakly robust in that they are unlikely to hurt women and benefit them in some cases though they make no difference in others.