Purpose – This paper aims to study the relation between the employment of mothers and the activities of children with the aim of contributing to the understanding of child work in India. Design/methodology/approach – Multinomial logit specifications of children's activities and mothers' employment are estimated on survey data drawn from the National Family Health Survey 1998-1999 for all India. The joint specification combines four states of children aged 6-14 (studying, working in the market, working for the family or being inactive) with the employed/not employed status of mothers. Findings – The results show that the mother's preferred choice is not working and sending children to school. This is especially true for more educated mothers. Also the father's education is positively (negatively) related to child schooling (work), but the effect is smaller as compared to that of mothers. All specifications yield the result that the probability of children working increases if their mothers work. Higher levels of household wealth play a fundamental role in lowering the risk of child work. Research limitations/implications – This empirical model does not take into account the unobserved heterogeneity of two types – namely, the residual correlation among the outcomes of mothers and children, and the residual correlation among children of the same mother. Practical implications – The evidence that children of employed mothers have a higher risk of working suggests that the problem may be related to the low quality and pay of jobs accessible to women in India, especially in rural areas. The policy indication would then be to improve the condition of women in the labour market and also to improve the welfare of their children. Originality/value – While women's and children's time allocation has been studied in separate settings, in this paper these two aspects are analysed together for the first time. |