This dissertation is composed of two self-contained essays, which examine two
different factors that could affect human capital accumulation in a developing country.
Both essays utilize cross-sectional data from the second round (2003/04) of national level
household survey from Nepal.
In the first essay, I estimate the impact of remittances on school attendance of
children in Nepal. Over the last decade Nepal has experienced an increase in both
domestic and international migration and consequently, Nepal has also seen a large surge
in remittances from expatriates, growing from less than 3 percent of the GDP in 1995 to
about 17 percent in 2004, to 22 percent in 2008, becoming one of the top ten recipients in
terms of the share of remittance to GDP. In developing countries, investment in human
capital is often viewed as significantly constrained by household resources. The premise
of this essay is that remittances, by relaxing household resource constraints, can promote
investment in education of the children living in remittance-receiving household. I use
the proportion of households receiving remittances and the migrant’s age as instrumental
variables to identify remittance-receiving households and level of remittance flow. I find
that remittances increase the probability of school attendance for young girls (ages 6-10)
and for older boys (ages 11-18). But the positive effect does not extend to younger boys
(ages 6-10) and older girls (ages 11-18).
In the second essay, I estimate the causal effect of child's number hours worked
on school attendance and school attainment. Here, number of hours worked is defined
broadly to include hours worked in market and non-market activities within and outside
the household as well as hours worked on domestic chores within the household. The
central identification problem in estimating the causal effect of child labor on schooling is
that these two decisions are simultaneously driven by different confounding factors such
as household income, family preferences, child characteristics, availability and quality of
school, etc. All of these are likely to induce a negative (or a positive) relationship
between schooling and child labor. To abstract from these confounding factors, I use
community level average daily agricultural wage for children and the distance to water
source to provide variation in the demand for child labor. The results show that the effect
of hours worked on schooling outcomes differ by demographic subgroups. For girls, the
number of hours worked adversely affects both school attendance and grade attainment.
For boys, the results are significantly different. The results of this study suggest that
working up to 12.7 and 14.5 hours per week have no adverse effect on school attendance
of boys of ages 5-9 and ages 10-16, respectively. Whereas, working less than 15 hours a
week has no detrimental effect on grade attainment of older boys. I find no effect of the
number of hours worked on grade attainment of younger boys aged 5-9.