Standard homo economicus lives in a world of complete markets and maximizes utility which is a function of his personal consumption. This approximation cannot account for parents making transfers to adult children, children taking care of old parents, nor for gifts, inheritance and many other services exchanged within families. Such behavior can be derived from three main mechanisms. Firstly, in the so-called pure altruism model, the parent's utility is augmented by the utility of his child. This leads to transfers from the parent to his child. An important feature of this model is the strong property of redistributive neutrality: since parents and child pool their income, any government transfer to one will be undone by the other adjusting his transfer. In a second model, altruism is impure as the parents want the child to behave in a certain way: exchange and strategic considerations enter the picture, as both parents' and child's income become endogenous. Thirdly, in a non-altruistic setting, with imperfect credit market, transfers to children and to old parents correspond to a reciprocity contract and are an investment for old age. Families embody long term and widespread commitments: born as a needy child, one becomes a parent and ultimately a (perhaps) needy grandparent. Moreover for much of what is exchanged within families, there is no market substitute. These features explain why the network of reciprocities can be large both in time and space, why those transfers change but do not disappear as market or public insurance develop, and why displacing them can have perverse side effects. Family transfers influence intra- and inter-generational inequality, hence the importance to assess their motivation. Tests usually conclude that the income pooling predicted by pure altruism is not observed, but family transfers are also far from being entirely motivated by direct exchange considerations.
