M. Niaz Asadullah is a prolific researcher in the realm of education and social issues research. Coming from the Centre for Poverty and Development Studies, University of Malaya, Kuala Lumpur, his research mostly relates to educational and social developments in developing countries including Bangladesh. He has recently made a comparative study on the different modes of delivery of primary education based on location choice. His paper titled 'Do Pro-Poor Schools Reach Out to the Poor? Location Choice of BRAC and ROSC Schools in Bangladesh' has been published in The Australian Economic Review (vol. 49, no. 4). Salient features of the article, albeit paraphrased occasionally, follows:
Asaduallah focuses on the location choice of 'one teacher, one class-room' non-formal primary schools pioneered by the BRAC. Then the performance of this system is pitted against its replication under the government-managed Reaching-Out-of-School (ROSC) project. Research on location-specific dimensions of schools is rare, if not unknown. The reasons are possibly not far to seek. Large spatial datasets containing information on schools by location, alongside other school characteristics, are not available in public domain. By way of studying the empirical determinants of geographical location of BRAC vis-a-vis ROSC schools in Bangladesh, the author of the paper claims to have filled an important gap in the international literature on the strategic choice of NGO-run development programmes in low-income countries.
Non-formal schools around the world have emerged as a popular option for education service delivery in poor communities that are generally devoid of better physical and human capital. As various studies tend to reveal, these schools effectively provide an alternative to formal primary education and succeed in maintaining and mainstreaming children into post-primary education. Niaz Asadullah mentions that BRAC, one of the world's largest non-governmental organisations (NGOs), runs a network of about 40,000 non-formal schools in Afghanistan, Pakistan, the Philippines, Uganda, South Sudan and Sierra Leone.
"The finding that location decision of BRAC schools is negatively related with low infrastructure runs contrary to the popular notion that BRAC schools charge no fee and manage with little to no facilities. However, operational considerations may discourage BRAC to set up schools in under-provided areas. In the context of Bangladesh, the major input to education, female teachers, is the main constraint to delivering educational services in rural areas. It is possible that low-infrastructure areas also have lower proportion of potential teachers. In case of ROSC schools, the coefficient on the literacy variable is negative and significant. But, it is positive and significant in case of BRAC schools".
One interesting observation is that ROSC schools, overall, have higher presence in sub-districts with poor infrastructure, implying that they are better targeted, while BRAC schools concentrate in better-developed parts of a poor district. This finding is consistent with the existing literature on NGO branch placement in Bangladesh. Some researchers find that NGOs in Bangladesh operate in locations with good physical infrastructure and better productive and marketing opportunities and they do so to minimise operating costs. Using data from northern Bangladesh, argues Asadullah, they find that NGO coverage in a village (measured as number of working NGOs) decreases with distance of the village from marketplace and increases with adoption of modern irrigation methods and soil quality. The NGOs do not consider poverty incidence in the village. "The findings on BRAC school location are also consistent with the fact that these schools are run primarily by locally recruited female teachers and are therefore set up in areas with adequate supply of female graduates of secondary schools. This pattern of location is similar to that of for-profit private schools in Pakistan that have emerged in locations that have previously benefited from public investment in female secondary schools".
In contrast to ROSC and BRAC schools, the findings on madrashas are somewhat mixed. Madrasha concentration is positively associated with a number of proxies of poverty and infrastructure development such as higher proportion of households with an illiterate head, smaller proportion of households with good toilets and large number of bank branches per 1,000 people. The association is negative and significant only in case of one indicator of poverty - vulnerability to natural disaster. In case of two other indicators, travel time to the nearest major city and distance to the nearest road, the effects are not significant even though both indicators are significantly associated with expected counts of ROSC and BRAC schools (the association being positive and negative, respectively). Nonetheless, the poverty connection still holds for madrashas - even after detailed control for infrastructure effects, the coefficient on locale remains significant in madrasha regression. In other words, poor households mostly send their children to madrashas.
And finally, there are a few important tips from the study. First, BRAC, ROSC and madrasha schools are all found to have a significantly high concentration in poor areas although ROSC schools have higher concentration in poor sub-districts, compared to BRAC (and madrasha) schools. Second, ROSC schools have a greater presence in regions that are under-served by the government and government-supported formal primary schools. On the other hand, the supply of BRAC schools (and madrashas) is significantly and positively associated with the number of formal primary schools. This analysis shows that the positive association between BRAC school counts and infrastructure is partly explained by the fact that well-provided areas also have higher female literacy rate, making it easier for BRAC to recruit female teachers. However, no such evidence is present in case of ROSC schools. These findings suggest that the BRAC schools, contrary to popular commentaries, behave strategically, compared to the ROSC schools, when catering to the educationally underprivileged population.
The latter does a better job in targeting regions where socio-economic circumstances are most likely to keep children out of school. This study therefore adds to the existing evidence on the entry of NGOs in poor pockets of relatively well-developed areas, rather than in remoter, less-developed ones and into communities that are already being catered for by other service providers. The findings imply that evaluation of the performance of NGO programmes is likely to suffer from positive selection bias. The evidence also cautions against the strategy where developing country governments and aid agencies consider NGO-run pro-poor schools as the most effective means of bringing children from isolated communities into mainstream education. Further, low-cost innovation in NGO service delivery models should be emphasised to ensure that children in difficult-to-reach communities are equally served by NGO programmes.
Abdul Bayes is a former Professor of Economics at Jahangirnagar University.
© 2024 - All Rights with The Financial Express
