In this study, I examine the extent to which child labour is explained by the main determinants proposed in the literature. I focus on three determinants: income shocks, household composition and parental preferences. In addition, I examine the role of child-specific characteristics, such as psycho-social abilities and nutritional endowment. While there is extensive literature on child labour determinants, it is predominantly theoretical and tends to treat each determinant in isolation from the others. I use two rounds of data from the Young Lives project on a cohort of children living in Andhra Pradesh, India who were interviewed at the ages of 7 and then 11 to examine the effects of inter- and intra-household factors - as well as child-specific characteristics - on child labour in a unified empirical framework. I find that children work more in response to income shocks; that child labour is sensitive to household composition in terms of the age and gender of the other children; and, in urban areas, that it varies according to the bargaining power of women in the household. Investigation of child-specific fixed effects further suggests that nutritional status, reading skills and ability to get along with peers may also be relevant to whether and how much children work.