Improving educational attainment is important for achieving population targets and meeting economic development goals in low-income countries. Over the past few decades, targeted stipend programs have been used to improve the school attainment in poor populations in countries as diverse as Turkey, Mexico, and Brazil. The principal behind these programs is simple: cash or in-kind incentives are provided to targeted poor households conditional upon children’s school attendance. Evaluations of stipend programs show positive impacts across contexts: stipend programs improve enrollment and attainment and – in some cases – delay the start of marriage and childbearing. However, the targeted nature of many programs means that many people in poor places will not receive stipends. It is possible that targeted stipend programs do not always reach the poorest populations because the poorest parents lack the necessary resources, skills, or connections to access stipends. Even if stipend programs have positive impacts for beneficiaries, the question remains: do targeted stipend programs reduce inequalities in school attainment among poor populations?
I explore the relationship between targeted stipends, inequality, and school attainment using a case study from rural Bangladesh. The Bangladeshi government implemented an innovative set of targeted stipend programs between 1990 and early 2002. These programs were intended to improve the school attainment of the rural poor, with a special emphasis on improving female attainment. The 1993 Food for Education program (FFE) provided wheat and flour to targeted vulnerable households conditional on primary school attendance. The 1994 Female Secondary School Stipend Project (FSP) provided cash stipends to rural adolescent girls conditional on secondary school attendance. The 2002 Primary Education Stipend (PES) program replaced FFE and provided cash stipends to targeted vulnerable households conditional on primary school attendance. Both programs are still in place today.
Stipend programs should reduce inequalities between groups if there are high participation rates for disadvantaged groups and positive returns to participation. I use a longitudinal sample of 1,012 children from 46 rural Bangladeshi villages to investigate these issues (data source: the Chronic Poverty and Long Term Impact Study in Rural Bangladesh). First, I explore how participation varies by socioeconomic status (SES) and gender. I find that girls had a significantly higher probability of participating in stipend programs than boys. Children from lower SES households had a significantly lower probability of participating in stipend programs compared to those from households with higher SES. If stipend programs were correctly targeted, lower SES females should have a higher probability of participation. However, I find no significant differences in female participation by SES. All of this raises serious questions about whether these stipend programs successfully reached their intended target populations in this sample.
Next, I investigate whether there were positive returns to stipend participation. Stipend participation is positively associated with schooling attainment for females, but not males. However, the returns to stipend participation are not significantly different for children from lower SES or higher SES families. This means that the targeted stipend programs improved absolute levels of schooling attainment for the sample, but did not reduce relative achievement gaps between children of different socioeconomic backgrounds. In other words, stipend-related schooling improvements for lower-SES females were matched by comparable improvements for higher-SES females. Meanwhile, there was no significant relationship between stipend participation and schooling attainment for males, which suggests a limited impact of stipend programs for male children.
My research points to the importance of thinking about the ‘big picture’ when evaluating the impact of targeted stipend programs. It is possible that stipend programs can have positive impacts for beneficiaries, but fail to reduce socioeconomic inequalities in schooling attainment in the population. The results also point to a troubling trend: the poor performance of lower SES male children. Many recent policy initiatives focus on improving female educational attainment, which risks overlooking the educational needs of lower SES boys. This particular case study is limited by the small sample size and the lack of nationally representative data; further work is needed to explore whether similar trends hold up in other contexts.
For more information please the accompany article on the topic in Demography http://link.springer.com/article/10.1007/s13524-015-0435-9/fulltext.html