A double-hurdle analysis of the gender earnings gap for working children in Bangladesh

Abstract

Child labor continues to be pervasive throughout the developing world. Although the number of child workers has declined, the ILO’s 2010 estimate remained at 215 million children. An estimated 100 million girls are involved in child labor, often for low pay. Girls are often engaged in some of the worst forms of child labor, sometimes in illegal activities such as prostitution. Studies regarding the gender earnings gap among adults are abundant, but there is very little economic research on girls’ unequal treatment in the labor market in terms of pay. In this paper, we measure and analyze the gender earning gap between working boys and girls in Bangladesh, those who are paid as well as those who receive no pay. In order to investigate the gender gap in the earnings of children, we use the 2002-03 National Child Labour Survey of Bangladesh and limit our analysis to children ages 5 through 14, yielding 50,263 cases. Of these children, 6,668 performed market work; 2,283 were paid for their work and 4,385 were unpaid workers. We present descriptive statistics and then apply the Double-Hurdle model to adjust for selectivity bias and to include the working children who do not earn a wage. We then estimate Mincerian earnings equations for a pooled sample of children who perform market work, using four alternative models. We find evidence of a significant gender earnings gap in the overall child labor market in Bangladesh with a raw earnings gap of 36 percent. We specify four alternative Double-Hurdle regression models in which we include alternative sets of independent variables. All four models show a statistically significant wage gap ranging in magnitude from 31 percent to 29 percent. The magnitude of the gap is large, due mainly to a greater proportion of girls in unpaid work. A child’s own characteristics, family characteristics, as well as industrial segregation play important roles in explaining the gap. Girls’ earnings remain well below boys’ earnings even when all measured characteristics are included in the analysis. Our results show that the concentration of girls’ in unpaid work contributes to the earnings gap, yet we understand the importance of part-time (often unpaid) work in allowing girls to pursue their education. Policies are needed that encourage both educational and employment opportunities for girls in safe environments.