Purpose – The aim of the paper is to understand whether children in home-worker (HW) households in Pakistan and Indonesia are more likely to work than other children, and, if so, how this impacts their capabilities. The paper also aims to outline some policy implications for the two countries. Design/methodology/approach – The data are drawn from two ad hoc surveys and country studies carried out in Pakistan and Indonesia in 2000/2001. The paper examines the incidence and reasons of child work and child schooling in home-worker households, the work conditions, and gender issues. A bivariate probit is applied to analyse the determinants of child activity status. Findings – Children from HW households have a higher probability of working. There is evidence of the feminisation of home work from childhood. This is dramatic in Pakistan while little evidence is found for Indonesia. In Pakistani urban slums the majority of children are working, but in Indonesia they are in school. The mother's education and per capita income/expenditure or assets in the household are important determinants of the child's activity status. Research limitations/implications – The model cannot use the control group for econometric analysis since the number of households and children interviewed (although randomly chosen) are not sufficient. Practical implications – Collective action plays a role in the reduction of children “only working”. The number of hours that children work in Pakistan suggests that their ability to do school-related activities is likely to be impacted. Originality/value – Although child labour is common in home-based manufacturing activities in the informal sector in most Asian developing countries research on child labour remains scarce. This paper contributes to this area of research.