Recent work criticises both the logic and relevance of the theoretical basis of the approach to estimating the costs of raising children adopted in much of the economics literature. This tends to be restricted purely to models in which the household members consume market goods with given household income. The "costs of children" are perceived essentially as market consumption costs. This ignores the fact that an important, possibly preponderant element of child costs takes the form of parental time, which must be diverted from alternative uses such as market work, other house-hold production activities and leisure, to care for children. The studies also ignore the question of the distribution of income among adults and, in particular, the differential incidence of child costs on adult members of the household. In this paper we first of all argue that a satisfactory theoretical approach to modelling child costs must simultaneously incorporate an "individualistic" formulation of the household (as in Apps and Rees, 1988, 99) and a formal treatment of household production (as suggested by Becker 1965, and adapted in Apps and Rees, 1988, 99). We then provide such a model. Using data from a recent Time Use Survey, we estimate specialized versions of the model for families with two children and use the results to derive the intra-family distribution of resources and implied child-rearing costs.