A Little Help May Be No Help at All: Size of Scholarships and Child Labour in Nepal

South Asia

Nepal

Education

Vouchers / Scholarships / School Subsidies

G Datt, L Uhe

2016

Datt, Gaurav, and Leah Uhe. “A Little Help May Be No Help at All: Size of Scholarships and Child Labour in Nepal.” The Journal of Development Studies 55.6 (2019): 1158–1181.
View publication

Nepal Living Standards Survey. Available at: https://microdata.cbs.gov.np/index.php/catalog

2019

N.A.

Since the 1970s (with variations)

Improve equity and access to education, especially for girls and disadvantaged children.

There are eight different types of scholarships: (i) scholarships for girls, which aim to support gender equality in education, (ii) scholarships for Dalits, which target caste discrimination, (iii) scholarships for the ‘poor and talented’, (iv) scholarships for children affected by conflict, (v) scholarship for children with disability, (vi) HIV-AIDS scholarships, (vii) scholarships for children living in Karnali zone, and (viii) a generic category for ‘other scholarships’. However, except for the first three, the other types are minor.

The authors investigate the context of the schooling scholarship program in Nepal. To identify the impact of scholarships on child work hours, we employ the relatively new method of coarsened exact matching (CEM) to estimate average treatment effects.

• Size of scholarship-based transfers
• Total hours of work per week, and hours of work per week for the three categories of work, namely economic, extended economic, and domestic chores.

Propensity Score Matching

Study design/Identification Strategy: in developing an empirical strategy to investigate scholarship effects on child labor, two implications

of the foregoing discussion are arguably relevant. First, there are significant differences in the intensity of child labor across the three main categories of work (economic, extended-economic, and domestic) which is suggestive of the importance of distinguishing these categories in empirical analysis. Second, given significant differences between girls and boys both in their hours of work (total as well as across types of work) and in their receipt of scholarships, it is important to estimate scholarships’ effects on work hours separately for girls and boys. The empirical methodology allows for both these elements of heterogeneity.

To identify the (causal) effect of scholarships on child labor, the authors use the method of coarsened exact matching (CEM).
Sample: the paper uses data from the Nepal Living Standards Survey, the NLSS3, undertaken in 2010. The survey has a large nationally representative sample of 5,988 households and 28,484 individuals including 9282 children between the ages of five and 17 years. The NLSS3 data have been meticulously collected using methods of in-field data entry and in-field validation. PSM sample: girls = 3337 (810 treatment, 2527 control), boys = 3220 (315 treatment, 2905 control).
Survey Instrument: The survey contains detailed information on the time-use of all household members above the age of five. This includes the number of hours per week that each child spends working in either wage or self-employment, as well as the hours spent working around the home.

The authors find significant effects in the impact of scholarship-based transfers on child labor in Nepal. High-value scholarships decrease 8–16-year-old girls’ total work hours by one-third, largely reducing their hours in paid and unpaid economic activities with little impact on domestic chores.

Child Labour

Households

National

There are significant effects in the impact of scholarship-based transfers on child labour in Nepal. High-value scholarships decrease 8–16-year-old girls’ total work hours by one-third, largely reducing their hours in paid and unpaid economic activities with little impact on domestic chores.

Low-value scholarships will not bring any impact on child labour outcomes.

To identify the impact of scholarships on child work hours, we employ the relatively new method of coarsened exact matching (CEM) to estimate average treatment effects. The CEM estimator has some strengths relative to several of the commonly used matching estimators, as discussed later in the paper. The main findings however come from a relaxation of the assumption of constant average treatment effects. Allowing for size-dependent effects, we find that higher-value scholarships lead to significant reductions in the hours of child labor per week for girls.