Child labour and school attendance: Evidence from selection on observed and unobserved variables in Zambia

Although the determinants of child labour and school attendance are well established in the literature, the causal effect of child labour on school attendance is largely unexplored owing in part to econometric challenges. The difficult in finding a valid and strong instrument for child labour, which is argued to be interdependent with school attendance, is one such impediment. Recognizing this difficult and the fact that children in child labour differ from their counterparts who are not in child labour along an array of observed and unobserved characteristics, I proceed along an alternative path in this paper. I examine the causal effect of long run child labour on children current school attendance using a novel estimation method that assumes that the amount of selection on observed variables closely approximates the amount of selection on unobservables. Using data for children between 5 and 17 years in Zambia, empirical results show that child labour has a significant negative effect on the probability of school attendance. The point estimate is also robust to unobserved variables. Results also show that children who participate in child labour are relatively older, hardly live with their mothers, mostly live in rural areas and are found in households were parents or household heads are relatively older and spent less time in school.

Grantor dc:publisher.institution

School of Economics

Year dc:date.issued

2010

Author dc:creator

• Simumba, Joseph

Advisor dc:contributor.advisor

• Leibbrandt, Murray

Language dc:language.iso

eng

Handle dc:identifier.uri

http://hdl.handle.net/11427/5692

OAI identifier oai:identifier

oai:open.uct.ac.za:11427/5692