![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
Abstract
This paper investigates the long-term impact of Bangladesh's primary education stipend (PES) programme on a range of individual and household welfare measures using a unique longitudinal study spanning the years 2000–2006. Using covariate matching and difference-in-difference methods, the programme is shown to have negligible impacts on school enrolments, household expenditures, calorie consumption, and protein consumption. At the individual level, the PES has a negative impact on grade progression, especially among boys from poor households who are ineligible to receive stipends at the secondary level. The programme does, however, lead to improvements in height-for-age among girls and body mass index among boys. Nonetheless, the impacts of the PES are remarkably small for a programme of its size. Poor targeting, particularly limited coverage and lack of geographical targeting, plus the declining real value of the stipend, are the most plausible reasons for this lack of impact.
Acknowledgements
This research was funded by the UK Economic and Social Research Council and Department for International Development under their Joint Research Scheme (Award Number RES 167-25-0361). The author thanks Akhter Ahmed, Peter Davis, Neha Kumar, Agnes Quisumbing and two anonymous reviewers for helpful comments on earlier versions of this paper. Wahid Quabili provided invaluable assistance preparing and understanding the data.
Notes
1. See the companion paper by Ahmed et al. in this issue for a detailed account of the development of both the FFE and PES programmes.
2. In addition, primary schools attached to higher-level madrasahs (Dakhil) are excluded from the PES programme for administrative reasons (Al-Samarrai Citation2009).
3. This calculation was made using the domestic wholesale market prices in Dhaka for November/December, quoted in the Food Policy Monitoring Unit's (various dates) quarterly Bangladesh Food Situation Reports.
4. Note that household splits were not followed between 2000 and 2003.
5. For consistency with previous International Food Policy Research Institute reports we use six to 12 years old as the age range for calculating primary enrolment rates, although the official age range for primary education is six to 10 years old. The World Bank's lower estimates of primary net enrolment rates are based on using the official age range.
6. This suggests that Bangladesh has achieved gender parity in education, at least at the primary level. These trends are consistent with those that have been observed in nationally representative household surveys, such as the Household Income and Expenditure Surveys of 2000 and 2005 (World Bank Citation2008).
7. Most of these households have several children of different ages.
8. One hundred decimals equals one acre of land. This variable was also logged with one decimal added to all values to avoid taking the log of zero for landless households.
9. Stunting is defined as height for age less than −2 standard deviations below the World Health Organization (WHO) reference standard, and low body mass index is defined as a score less than 18.5 (thin) using the WHO reference standards for five to 19 year olds (De Onis et al. Citation2007).
10. Note that grade progression relates to children who were of primary school age (six to 12 years) at some point during 2000 and 2006. Thus, mean grade progression is reduced by children who were less than six years old in 2000 and entered primary school in later years.
11. The new WHO reference groups for children five to 18 years old are used for calculating these z-scores.
12. This approach is at the heart of the randomised control trial approach to evaluating development interventions popularised by the Jamil Latif Poverty Action Laboratory at the Massachusetts Institute of Technology. See White (Citation2011) for an introduction and careful assessment of the pros and cons of randomised control trials.
13. Note that the PES eligibility criteria were unchanged between 2002 and 2006.
14. For example, a near-landless household whose youngest child attended Grade One of primary school in the 2000/01 academic year would no longer be eligible to receive PES in 2006/07 if the child had completed primary school on track in 2004/05.
15. The weak targeting of the PES that such ‘adjustments’ imply is confirmed in Section 5.
16. Following Cornia and Stewart (Citation1995), they are also known as E and F errors, where the ‘E’ stands for excessive coverage and ‘F’ for failure to reach the target population.
17. An evaluation of the PES programme in 2005 found that attendance figures were exaggerated for one-third of eligible stipend holders. In addition, 7 per cent of beneficiaries received payments despite failing the last annual examination (Al-Samarrai Citation2009).
18. This finding is consistent with the findings of a recent International Food Policy Research Institute study of the impact of the Productive Safety Net Programme in Ethiopia (Gilligan et al. Citation2008).
19. Note that anthropometric indicators were collected from boys younger than 13 years old, plus females younger than 19 years old and all mothers in the 2000 survey. Anthropometry on all available household members was collected in the 2006 survey.
20. Another policy option would be to replace cash transfers with food transfers. Although a return to FFE has certain attractive features (including its counter-cyclical and self-targeting nature), the high and variable costs of food transfers, plus the substantial leakages associated with their distribution, make it an unattractive option to most policy-makers.