The assumption that household income is strongly and positively correlated with a household's real standard of living provides the basis for the joint taxation of families, which has the effect of discriminating against married women as second earners. This paper shows, in the context of a model of the household with young children present, that this assumption is not tenable. The fact that there is considerable heterogeneity in female labour supply which cannot be explained by wage rates and the number and ages of children requires us to look for other explanations, and we argue that these can be found in the variation of child care costs and productivities across households. When these are taken into account, we show, by theoretical modelling and numerical simulations based on survey data, that household income is a poor indicator of household well-being.