Discussion Paper
No. 2013-6 | January 15, 2013
Prachi Gupta, Upasak Das and Ashish Singh
Child Disability and Maternal Work Participation: New Evidence from India


Using data from the India Human Development Survey, this paper analyses the relationship between child disability and maternal work participation for India. The authors’ findings suggest a significant positive relationship between child disability and the work participation of the urban mothers who are wives of household heads. These mothers are 1.27 times as likely to participate in labour market as mothers (wives in urban areas) without a disabled child. However, for the same mothers, child disability significantly affects the weekly work hours of those participating in the labour market in a negative manner with presence of a disabled child reducing the weekly work hours by 3.6 hours. For the rural mothers and the mothers in urban areas who are household heads, our findings do not suggest any significant association between child disabilities and their work participation (or weekly work hours).

JEL Classification:

J22, J10

Cite As

Prachi Gupta, Upasak Das, and Ashish Singh (2013). Child Disability and Maternal Work Participation: New Evidence from India. Economics Discussion Papers, No 2013-6, Kiel Institute for the World Economy. http://www.economics-ejournal.org/economics/discussionpapers/2013-6

Comments and Questions

Anonymous - Referee Report 1
March 19, 2013 - 10:24
see attached file

Anonymous - Referee Report 2
May 02, 2013 - 12:44
• Many of the diseases causing disability (cited on page 5) are related to the status of the mother, that is, poverty plays a crucial role in capturing one of these diseases. Disability of the child is thus not an exogenous random event but correlated to unobserved characteristics of the mother which renders the variable capturing child disability endogenous thus biasing the results. A possible solution would be a two stage approach, at least for the OLS based equations. • Building upon the comment made above, one could in an extreme interpretation even argue that having a disabled child is causing a self selection into the sample as unobserved characteristics of the mother/ father may prevent medical treatment (such as vaccinations against polio). • There is a clear difference shown by the sample statistics between rural and urban female headed households that call for explanation: labor force participation is much lower in urban areas compared to rural areas (51 percent vs. 80 percent). If the reason for working for female household heads is the same for rural and urban residents as explained by the authors – namely poverty and the need to work – why then do so many female household heads do not work at all in urban areas? • Why is it, that female household heads are more likely to have a disabled child as shown in the sample statistics? Beyond problems implied for estimating parameters (have fathers left the mother because of the disabled child (non-random sample)?), this has implications for policy, too. • The distinction between rural and urban households by using separate estimation equations can be problematic, too. Families may move either into the city in order to find a potentially better paid job or move back from urban to rural areas in order to be closer to the extended family (grandparents, siblings of the parents of the disabled child) which may support the mother. Other disability related causes for rural/urban – urban/ rural migration may well be possible. I don’t how likely such migration is, but as a reader I’d like to know from the authors. • The authors use way too often the word “maybe” for explaining their findings. This is dissatisfying as particularly with respect to policy implications one would need more clarity on the topic. More insights can be supplied by either getting deeper into the literature or by doing field visits. Otherwise the results appear too much as correlations accompanied by ad hoc explanations that somehow fit into the picture. • The authors should provide much more test statistics in order to provide some more confidence on the robustness of the results.