Researchers often assume household structure is exogenous, but child fostering, the institution in which parents send their biological children to live with another family, is widespread in sub-Saharan Africa and provides evidence against this assumption. Using data I collected in Burkina Faso, I analyze a household's decision to adjust its size and composition through fostering. A household fosters children as a risk-coping mechanism in response to exogenous income shocks, if it has a good social network, and to satisfy labor demands within the household. Increases of one standard deviation in a household's agricultural shock, percentage of good network members, or number of older girls increase the probability of sending a child above the current fostering level by 29.1, 30.0, and 34.5 percent, respectively. Testing whether factors influencing the sending decision have an opposite impact on the receiving decision leads to a rejection of the symmetric, theoretical model for child fostering.