Income redistribution may be performed both publicly and privately. Ever since Mirrlees (1971), static optimal income taxation problems have typically been concerned with redistribution by the State between independent individuals. I depart from this independence assumption and account for the fact that individuals have familial and social links that induce them to make redistributive transfers. This paper is the first to consider the existence of such networks of private transfers in an optimal income taxation framework. Individuals are part of networks of altruism when they value the private utilities of their network’s members. As a result, private transfers between agents arise. More specifically, in this paper, individuals belong to groups and differ in their productivity levels. They thus have two dimensions of choice: their productive effort and the transfers they make to other members of their group. My main results are the following. First, I derive the structure of the endogenous income distribution and transfers when there is no governmental intervention. Second, I characterize theoretically the optimal linear and non-linear income tax schemes in this context. I provide sufficient statistics formulas for both, that highlight how tax design is affected by the existence of altruistic private transfers. Public and private redistribution are substitutes but at the optimum, private transfers are only partially crowded-out: the government may rely on these transfers to achieve its equity objective. Taking into account private redistribution unequivocally increases welfare, and depending on the structure of groups, may induce an efficiency gain. Finally, I illustrate these results using numerical simulations. These novel insights can then be applied to a variety of settings, such as those of family taxation and of taxation in developing countries, for which I document the structure and nature of informal transfers using the World Bank's Living Standard Measurement Surveys.