Essays in economic development: education, child labour and wage inequality

This thesis presents four self-contained essays that explore issues that are crucial in improving human well-being in a developing country: improving health, minimising child labour and reducing gender inequality. The analysis is focused on Bangladesh where the prevalence of child labour and gender differences in several domains is still widespread. The first essay aims to examine the gender wage gap along the entire wage distribution into an endowment effect and a discrimination effect, taking into account possible selection into full-time employment. Applying a new decomposition approach to the Bangladesh Labour Force Survey (LFS) datasets of 1999 and 2005, we find that women are paid less than men everywhere on the wage distribution and the gap is higher at the lower end of the distribution. Discrimination against women is the primary determinant of the wage gap. We also find that this gap has widened between 1999 and 2005. The second essay examines whether gender differences in tertiary enrolment rates can be explained by wage premiums in returns from secondary to tertiary education levels. Using LFS data, we find that wage premiums do not have any significant effect on the gender gap in tertiary enrolment rates. We also note that wage premiums in returns from secondary to tertiary education significantly influence tertiary enrolment rates for males but not for females, once additional variables are added. We offer evidence that part of the explanation for low female enrolment in tertiary education is attributable to demographic factors. The third essay investigates whether there is any trade-off between child labour hours and schooling. By drawing on the 2002 dataset of the Bangladesh National Child Labour Survey (NCLS), we find that working hours adversely affect child schooling from the very first hour of work. However, the marginal impact of child labour hours weakens when working hours increase; yet, working hours always negatively affect schooling when we use a non-parametric approach. We find that parents do not have identical preferences towards schooling decisions concerning boys and girls. Both mother and father show a significant preference for educating a female child. The same incentive effect is not found for a male child. These conclusions persist, even after allowing for sample selection in child labour. The fourth essay tests the effect of child labour on child health outcomes in Bangladesh. We use self-reported injury or illness due to work as a general measure of health status. Using NCLS data, we find that child labour is positively and significantly associated with the probability of being injured or becoming ill, once the endogenous relationship between these factors is accounted for. These findings remain robust when we consider child labour hours and restrict our analysis to rural areas. Moreover, the intensity of injury or illness is significantly higher in construction and manufacturing than in other sectors.