Children and maternal migration: evidence from exogenous variations in family size

Abstract

Both theoretically and empirically, childbearing decreases female labour supply, but few papers examine the effects of children on whether women emigrate to work. Using exogenous variations in family size induced by parents’ preferences for mixed sibling–sex composition in instrumental variable estimations, we find that in Sri Lanka where most migrants are women and mothers, children decrease maternal labour supply in the domestic market but they increase maternal labour supply abroad.

Keywords:

• maternal migration
• childbearing
• Sri Lanka

JEL Classification:

• F22
• J13
• J22

Notes

¹ We could have also used twins as an instrument, but the number of twins in our data is small.

² We use more than two children, not the number of children, as the dependent variable in Equation 1 because we want to emphasize the fertility increase that the instrument, same sex, induces, that is a change from two to more than two children. In any case, using the number of children as the endogenous variable produces similar results; we do not present these results here for brevity.

³ The instrument, same sex, may be invalid if parents prefer boys over girls. However, in Sri Lanka, unlike in other South Asian countries, parents have little or no son preference (Arnold, 1992; Abeykoon, 1995). In fact, girls in Sri Lanka have better child mortality rates and secondary school enrolment rates (World Bank, 2014).