Review of the Bangladesh female secondary school stipend project using a social exclusion framework.

The Female Secondary School Stipend Project in Bangladesh was established to increase the enrollment of girls in secondary schools, thereby delaying marriage and childbearing. This analysis examined the existing data using the social exclusion framework to clarify the primary exclusionary factors that have kept girls from education: harassment, poverty, and the primacy of marriage and childbirth and explored the extent to which the project has diminished such barriers. While causality is difficult to establish, data suggest that the stipend programme has contributed to the rise in enrollment of girls in secondary schools. Questions remain as to the impact of the stipend programme on delaying marriage, empowerment of girls and women, and enhancing employment opportunities. A thorough assessment of the impact is required. The case study suggests that, if the programme design had focused on the quality and content of education and the broader economic and social context, more opportunities would have been created for social and economic participation of girls.

INTRODUCTION

Secondary education is an important goal in development, providing opportunities for active participation in the global knowledge economy, civic skills, and social cohesion (1). As an intervention, increasing access to secondary education has great potential to counter social exclusion for girls, whose traditional gender responsibilities have kept them from full economic and social participation. The Female Secondary School Stipend Project (FSP) in Bangladesh was established in 1982 to increase the enrollment of girls in secondary schools, thereby delaying marriage and childbearing. The Bangladesh Association for Community Education (BACE), a national non-governmental organization (NGO), initiated and implemented the project, and it was then scaled up with technical and financial support of international actors.

This analysis examined the intervention through the lens of social exclusion to assess whether the scheme has increased girls’ capabilities and their levels of participation in society. The social exclusion framework entails an examination of barriers to inclusion, and an analysis of the extent to which the policy in question has overcome these barriers (2). This paper argues that the programme is a partial success at best and suggests a stronger programmatic emphasis on capabilities rather than just enrollment figures. Such an emphasis, achieved through transformation of the curriculum, would improve the ability of the programme to achieve its goals of delayed fertility and greater female participation in development. A focus on capabilities would mean a stronger contribution of girls’ school enrollment (target 4) to the greater Millennium Development Goal (MDG): promote gender equality and empower women (goal 3). The absence of such a focus represents a missed opportunity and signals inefficiency and inefficacy within the programme.

BACKGROUND

Bangladesh has made impressive strides in a relatively short time in terms of achieving gender parity at both primary and secondary levels. The country has now surpassed target 4: Eliminate gender disparity in primary and secondary school of the gender-related MDG 3 (Promote gender equality and empower women). It should be noted here that the broader goal of promoting gender equality and empowering women has certainly not been achieved, and the target 4 does not necessarily indicate progress towards the larger goal—especially as it focuses on enrollment rather than educational outcomes.

Educational attainment for women in Bangladesh was once among the lowest in the world. The 1991 census indicated that only 20% of women were literate, with a rate of 14% in rural areas. The enrollment rate in primary schools for girls was 64% in 1990 compared to 74% for boys. The gender disparity was even more stark in secondary education, in which only 33% of enrolled students and 29% of graduates were girls (Table 1).

Table 1.

Secondary education, Grade 6–12, 1970 and 2002 (see Appendix for more data)

Year	Institution			Students in 000's				Teachers
	Government	Non-government	Total	Boys	Girls	Total	% of girls	Men	Women	Total	% of females
1970	180	6,008	6,188	1,266	266	1,532	17.4	52,436	4,119	56,555	7.2
2002	327	17,794	18,121	3,967	4,531	8,499	53.3	178,829	37,163	215,992	17.2

Source: BANBEIS, November 2002, December 2003

Appendix.

Aggregate secondary enrollments, 1995–2005, by grade and gender

Year	Sex	Grade 6	Grade 7	Grade 8	Grade 9	Grade 10	Total
1995	Total	1,426,399	1,209,619	1,009,321	782,371	630,899	5058,610
	Girls	699,939	591,840	475,374	341,361	264,328	2,372,842
	% of girls	49.07	48.93	47.10	43.63	41.90	46.91
1996	Total	1,575,237	1,335,684	1,114,419	863,862	696,604	5,585,806
	Girls	775,021	655,252	526,307	377,890	292,604	2,627,073
	% of girls	49.20	49.06	47.23	43.74	42.00	47.03
1997	Total	1,727,103	1,464,455	1,221,858	947,146	763,763	6,124,325
	Girls	861,899	728,703	585,304	420,250	325,404	2,921,560
	% of girls	49.90	49.76	47.90	44.37	42.61	47.70
1998	Total	1,841,802	1,559,274	1,308,464	1,102,025	957,513	6,769,078
	Girls	951,782	812,855	685,637	557,078	457,390	3,464,742
	% of girls	51.68	52.13	52.40	50.55	47.77	51.18
1999	Total	1,931,857	1,706,516	1,408,810	1,139,710	1,050,046	7,236,939
	Girls	1,003,007	891,146	744,220	597,816	522,634	3,758,823
	% of girls	51.92	52.22	52.83	52.45	49.77	51.94
2000	Total	1,957,398	1,722,863	1,524,492	1,348,770	1,093,362	7,646,885
	Girls	1,023,700	914,050	843,088	699,164	540,235	4,020,237
	% of girls	52.30	53.05	55.30	51.84	49.41	52.57
2001	Total	2,008,565	1,767,903	1,564,350	1,406,242	1,139,950	7,887,010
	Girls	1,064,312	950,312	876,536	736,135	568,802	4,196,097
	% of girls	52.99	53.75	56.03	52.35	49.90	53.20
2002	Total	2,079,712	1,831,573	1,618,924	1,445,296	1,186,629	8,162,134
	Girls	1,106,329	987,708	910,537	765,425	590,479	4,360,778
	% of girls	53.2	53.93	56.24	52.96	49.76	53.43
2003	Total	2,052,219	1,832,685	1,616,291	1,440,158	1,185,009	8,126,362
	Girls	1,052,498	974,348	899,735	765,272	630,715	4,322,568
	% of girls	51.3	53.2	55.7	53.1	53.2	53.2
2004	Total	1,926,409	1,700,118	1,517,666	1,364,296	994,758	7,503,247
	Girls	1,001,007	904,147	812,256	706,788	500,917	3,925,110
	% of girls	51.96	53.20	53.52	51.81	50.36	52.31
2005	Total	1,976,729	1,685,056	1,481,110	1,294,110	961,547	7,398,552
	Girls	1,030,127	881,506	779,147	675,859	501,375	3,868,014
	% of girls	52.11	52.31	52.61	52.23	52.14	52.28

Source: BANBEIS, 2006,

Note: The figures may be inflated by ghost enrollments—many girls entered into class registers never enroll for examinations, for example

The figures for the first decades of the 21st century are vastly different: in the primary sector, gender parity in enrollment has been achieved, and in 2002, the secondary sector enrollment of girls exceeded the enrollment of boys at 53% (Table 1) (3,4). A number of policy interventions are credited for this dramatic change, including the Food for Education Programme sponsored by World Food Programme, the rise in non-formal education pushed by NGOs, and the increase in formal sector employment opportunities for women that require secondary education, especially the garment sector. In addition, one of the most accredited drivers of this change is the Female Secondary School Stipend Project (FSP). Launched nationally in 1994 and funded by the World Bank, Asian Development Bank, and the governments of Norway and Bangladesh, the project pays tuition-fees and provides monthly stipends for unmarried rural girls up to class 10 who attend recognized institutions, remain unmarried, maintain at least 75% attendance, and secure at least 45% marks in the annual examinations (a pass requires 35%). In 2005, nearly 2.3 million girls were enrolled in the programme (Table 2).

Table 2.

Number of girls receiving stipends, 1999–2005

Year	No. of schools (madrasah and government, but not NGO schools)	No. of girls receiving stipends	Amount distributed (million US$, at exchange rate of specific year)
1999	18,788	3,564,404	55
2000	19,919	3,961,194	54
2001	21,027	4,191,058	55
2002	22,893	4,193,352	54
2003	23,719	3,467,123	43
2004	24,950	2,356,856	23
2005	25,425	2,270,343	25

Source: BANBEIS, 2006; NGO=Non-governmental organization

A primary aim of the programme since its inception has been to delay marriage and childbearing to reduce fertility, through increasing school enrollment and retention. As married girls are excluded from the stipend, there is a clear incentive for parents to delay the marriages of girls. This intervention was part of a greater push to curb unsustainable population growth. Bangladesh has a population of 144.3 million, and the highest population density in the world (5). The Government of Bangladesh made family planning a national priority from the time of Independence in 1971, with the support of large donors, such as World Bank and United States Agency for International Development (USAID). Unreserved political commitment resulted in a dramatic decline in total fertility from nearly seven children per woman in the 1970s to 3.3 children per woman in the early 1990s (5). While the fertility transition is hailed as a success, population growth in Bangladesh is still not sustainable, and reducing population growth is still an important development goal. Both fertility transition and achievement of gender parity in primary and secondary schooling in Bangladesh are important (and inseparable) stories in the development discourse. However, it is not yet clear if the stipend programme had made a significant contribution to achieving these goals.

What had previously kept girls from school?

Traditional exclusion from education is part of a larger pattern of female exclusion, where girls are married young, kept from the formal job sector, denied expenditure for health, and sequestered to the domestic sphere through purdah. [Purdah is the practice of keeping females secluded and well-covered by clothing from all men, except those related by blood]. Early marriage is a common practice in Bangladesh, with girls traditionally being considered eligible for marriage from menarche. While the legal age for marriage for girls is 18 years, inconsistent birth registration and a lack of enforcement means compliance is negligible. The 2004 Bangladesh Demographic and Health Survey data found that women aged 20–24 years reported an average age at marriage of 16 years, with the average age of 14.4 years for girls with no education, 14.9 years for girls with a primary education, and 17.9 years for girls with a secondary education (6). Poverty and economic insecurity lead to even younger girls being married as they require a lesser dowry. Dowry poses a huge financial burden to families, leaving little funds available for education or health. Any financial benefit accrued from education of a girl would ultimately be reaped by the husband's household, not by the girl's natal household. For these reasons, education for girls has not traditionally been a financial priority for most Bangladeshi families. Not only were girls rarely sent to school, they were rarely out of the domestic sphere. Women in Bangladesh have been systematically excluded from the public sphere due to traditional restrictions on mobility and the imperatives of motherhood and family. The 1999/2000 Labour Force Survey found that three-quarters of employed women aged over 15 years work as unpaid family labourers in subsistence agriculture (7).

Benefits of educating girls

Low levels of education are not just a negative indicator of women's low level of resources and participation in the public and productive sector. Low educational attainment also correlates with high rates of child marriage, early pregnancies, and high fertility and also poor health and nutrition for mothers and their children. For example, in a historical cohort study, Hurt showed that education was the most important determinant of adult female mortality in Bangladesh. Women who had attended any formal education had significantly lower levels of mortality than those with no education (8). Using data from the Matlab area of Bangladesh, Bhuiya and Streatfield demonstrated that education of mothers positively affected child survival, although not equally for both the sexes (9). Cleland and Van Ginneken (1989) and Hobcraft (1993) used a meta-analysis of available global data to find that education is important for maternal and child survival independent of other socioeconomic markers (10–12), although the pathways are not always clear or consistent between different countries. Furthermore, educated mothers are more likely to send their daughters to school and also delay marriages of their daughters. Daughters of educated mothers and mothers-in-law also initiate childbearing later (13).

Furthermore, should education succeed at delaying marriage, as the theory supporting the FSP predicts, the impact of delayed marriage on health and fertility are important considerations. Ambrus and Field used data from rural Bangladesh to indicate that delays in marriage are associated with increased use of preventative healthcare services (14). A global meta-analysis by Jensen and Thornton suggests that women who marry later have more reproductive control and a decreased rate of maternal mortality (15).

Some evidence indicates that traditional values are waning in rural Bangladesh, along with a move away from the agricultural lifestyle and the increased economic importance of non-farm activities (16). This provides an opportunity for policy interventions, like the FSP, to assist this transition in a way that will benefit girls and women and increase their economic and social participation. Still, exclusionary mechanisms remain and must be understood to properly inform appropriate interventions.

SOCIAL EXCLUSION FRAMEWORK: BARRING GIRLS FROM EDUCATION

Social exclusion can be understood as a cumulation of structural processes contributing to social isolation and the stripping of different domains of social participation (17). The concept has been used for explaining why some groups do not bene-fit equally from state service provision, or remain chronically deprived. The social exclusion framework requires an account of these mechanisms of exclusion and an analysis of the extent to which specific policies or interventions have mitigated or removed such barriers.

Sen traces social exclusion to the Aristotelian concept of ‘capability deprivation,’ and Adam's Smith's relational understanding of deprivation (18). Understanding deprivation to be relational rather than absolute recognizes that the cost of living and the cost of participating in society are not universally consistent and may be better understood as the lack of capability to live a decent life. An inability to interact with others is an important deprivation in itself but it may also lead to other deprivations, such as lack of employment opportunities or lack of access to credit that may lead to impoverishment, which may, in turn, lead to other deprivations (18). Sen's work on capabilities and entitlements emphasizes that it is not what the poor possess, or how much money they have but what this enables them to do. Capabilities for participation in the public sphere may include literacy (to access information and engage in news media), good health, and access to spaces and institutions where decisions are made. Therefore, education is not an asset in its own right. It is only meaningful to the extent to which it increases the ability of students (in this case women and girls) to participate in society and the economy.

Mechanisms of exclusion in Bangladesh

Various mechanisms work to exclude girls from secondary education. A review of the available literature indicates three important mechanisms: purdah and harassment, the importance of early marriage to women's livelihood, and poverty. Each of these factors is likely to be considered in household decision-making around educating girls. These mechanisms are described to examine how and to what extent the stipend programme was able to overcome them. While these factors are explored separately, deprivation is always multidimensional, and these factors overlap and compound each other (19,20).

Social capabilities: purdah and harassment

One of the primary mechanisms of exclusion is the imperative of purdah, the practice of female exclusion (21). Purdah norms govern and restrict movement of women between the domestic and the public sphere and are enforced by harassment (or ‘eve-teasing’) of girls, which occurs to and from school and in the classroom. A young Bangladeshi woman in the public or private sphere faces many perils, and her main asset is her respectability. This is easily threatened—if she ‘holds her head up high’ or dresses nicely, she is often taunted or teased, or ‘put to shame’. Tactics like these are used by both male students and teachers to harass or discipline girls (22). This kind of harassment in the classroom prepares young women to accept subordination as part of everyday life (23) both in the home and in public.

An ActionAid nation-wide study on gendered violence in education, which included responses (n=856) from adolescent boys, girls, teachers, parents, and civil society groups, found that 86% of girls reported that they faced harassments on the way to school, and 97% of boys reported that they harassed girls on the way to school. While this is obviously intimidating for girls, 78% said that they felt intimidated while 54% stated that it would not stop them from studying. Parents, however, did not necessarily agree. With the burden of marrying off their daughters firmly in mind, 73% said that ‘eve-teasing’ was the reason enough to stop education of their daughters (22). Household data from Education Watch 2005 found that 4% of drop-outs among rural girls are due to security concerns. Moreover, if the distance to school is too far, or the route too public, a girl may not be allowed to attend.

Cultural capabilities: primacy of early marriage and childbearing

A second mechanism of exclusion of girls from education and employment is the primacy of marriage in girls’ lives. Marriage is vital in rural Bangladeshi society and, along with bearing sons, is the main determinant of quality of life for many women. Traditionally, a woman in Bangladesh derives her status from her family; her role includes maintaining her family as a social institution and as an economic entity. In Bangladesh, while the legal age for marriage is 18 years for girls and 21 years for boys, child marriage is common: a 1996 meta-analysis of demographic and health survey data identified Bangladesh as the country with the second highest rate of child marriage in the world, with an estimated 75% married before 18 years of age (24). Bangladesh Bureau of Statistics reports that nearly 50% of adolescent girls are married by the age of 15 years, and 60% become mothers by the age of 20 years (14). The 1996 Matlab Health and Socioeconomic Survey data from rural Bangladesh confirmed this, indicating that 70% of marriages occur within two years of menarche (14). With the high risk to reputations of girls through unwanted attention and their perceived vulnerability in the public sphere, marriage is seen as a safe alternative to sexual attention and unsanctioned sexual activity (25–27). Early marriage, in almost every instance, ends education of girls and can act as a barrier to inclusion in broader society.

In the previously-mentioned ActionAid study (22), all respondents admitted that the main goal of a girl's life is a good marriage—education is only useful to the extent that it facilitates this. Indeed, in Mahmud's panel survey of nearly 3,000 girls in rural Bangladesh, marriage was found to be the single most frequent reason for dropping out of school (28). Parents who can afford education often see it as an investment in daughters’ marriage capital, not their human capital (16). Once girls reach adolescence, parents are anxious to keep them out of the public sphere—to reduce their contact with boys and keep their reputation and marriageabi-lity intact. Del Franco found in her 2003 anthropological study that many girls are sent to madrasah (Islamic religious schools) so that they will receive a good religious education, be strictly supervised, and not be tempted to have prohibited relationships with boys to facilitate better marriages (29). One mother interviewed said, “My daughters are poor and dark, if they learn at least to do namaj and to behave as good Muslims, it would be easier to marry them off.” (29). So, not only does marriage itself curtail girls’ pursuing a quality education, the concern with marriageability also can be an inhibiting factor.

Economic capabilities and resources: poverty

The high cost of schooling is a significant factor that excludes the poor from education in Bangladesh. Expense is almost always cited as one of the main factors that prevent children from attending school—not just the cost of tuition, stationery, and books but also the opportunity cost of lost income. Household data of Education Watch 2005 indicate that child work was more frequently a barrier to education for girls than for boys (1). Half of all households cite poverty as the main reason for children ending schooling.

The financial and non-financial costs of schooling often are perceived to outweigh the benefits in rural Bangladesh (30). The cost of schooling, including opportunity costs, uniforms, books, examination fees, and kerosene to enable studying at night, is substantial. Data from Education Watch 2005 suggest that private household spending for education was more than four times higher than public spending (1). On average, households spend 17% less on education of girls than they do for education of boys (Tk 8,874 per male child and Tk 7,411 per female child per year). Moreover, limited teacher interaction and low quality of instruction lead many families to enlist tutoring, further increasing expense. Often tutors are the child's own teacher, which may diminish teachers’ incentive to teach well during school hours. Expenditure for private tutoring accounts for half of all private household expenditure on education (1).

However, ‘poverty’ should not be taken at face value as a reason for not sending girls to school if disposable income is being allocated to other financial priorities, such as educating sons. Boys are considered an investment in the future economic security of parents and have much higher earning potential (4). Traditionally, girls will leave the parental home at marriage to go and live with their husband's family and will rarely be given a chance to work in the formal sector. In a survey, 93% of boys themselves agreed that they are ‘future investment’ of their parents (22). This gives boys a strong sense of entitlement, confidence, and purpose—factors that might help explain their higher educational achievement. Shafiq also suggests that it may explain the falling enrollment rate of boys—the higher earning abi-lity of boys means that they are withdrawn from school to earn for the family (4). Furthermore, irrespective of household food-security, or a rural or urban setting, families spend less on education for girls. Probably due to the FSP, which targets rural areas, gender variation in spending is wider in urban areas than in rural areas (31).

THE INTERVENTION: FEMALE SECONDARY STIPEND PROGRAMME

In response to the low ratio of enrollment of girls in secondary schools, the BACE implemented the FSP in 1982 at a local level, with the specific aim of reducing fertility (Table 4). The intervention was inspired by population literature that suggested that education of girls would delay marriage and increase contraceptive-use, reducing fertility. The partnerships with international agencies, first by the USAID and the Asia Foundation, then the Norwegian Agency for Development Cooperation (NORAD), and, finally, the World Bank and the Government of Bangladesh, helped the programme scale up (32).

Table 4.

FSP timeline

1977	Bangladesh Association for Community Education (BACE) set up by Dr. Ellen Sattar (education specialist) and Dr. Mohammed Abdus Satter (former Secretary for Population Control and Family Planning in the Ministry of Health and Population Control, GoB). The stipend programme for girls and boys from poor families at the primary level established
1982	First form of the Female Secondary Stipend Programme piloted in Shahrasti upazila and Kaharole upazila by BACE—supported by USAID and the Asia Foundation
1990	Primary education becomes compulsory and free
1990	Free tuition introduced for girls in class VI-VIII
1992	NORAD takes over funding the FSP—increased coverage to 7 upazilas
1994	Nation-wide FSP launched, funded by NORAD, the World Bank and GoB, ADB, and GoB for students in class VI and IX
	30 metropolitan upazilas excluded in Khulna, Dhaka, Chittagong, and Rajshahi because they have better facilities
	270 upazilas covered by GoB: FSSP
	118 upazilas covered by World Bank and GoB: FSSAP
	53 upazilas covered by ADB: SEDIP/SESIP
1996	The stipend programme extended to girls in class VII and VIII
1997	NORAD takes on another 12 upazilas: coverage 282 upazilas
2000	Enrollment of girls reported at over 50%
2000	Programme received a World Bank gold medal for excellence
2004	Conference in Shanghai co-hosted by the Chinese Government at the World Bank's ‘Reducing poverty, sustaining growth’—FSP was held up as an example of a successful scale-up of an intervention
2004	All programmes renewed for a five-year period

Adapted from Raynor, 2004 (32); ADB=Asian Development Bank; FSP=Female Secondary Stipend Project; FSSAP=Female Secondary School Assistance Project; GoB=Government of Bangladesh; NORAD=Norwegian Agency for Development Cooperation; SEDIP=Secondary Education Development and Improvement Project; SESIP=Secondary Education Sector Improvement Program; USAID=United States Agency for International Development

The 1990 Bangladesh Country Assessment by the World Bank emphasized the importance of increasing access of women to social services and investing in women as stakeholders of development. In partnership with the Government, the World Bank decided to scale up the FSP, which was at that time implemented in selected upazilas (a subdistrict, the lowest level of administrative government) by the Asia Foundation and funded by the USAID. Initially, the intention was to target the stipend to the poorest families. However, corruption and the cost of targeting meant that this was not feasible. Also, targeting the poor would mean that key support from the rural elite would be lost. As a result, the programme became available to all rural girls who met the criteria. The project also expanded according to demand. When the project was launched in selected upazilas in 1994, nearly twice as many girls as estimated at appraisal were recorded as enrolled in the programme, encouraging the Government to immediately expand the project to a national level. The unexpected enlistment into the project was, in part, the cause of the dilution of resources and reduction in quality inevitable from the unplanned increase in the sizes of classes (33). The stipend project has now been in place at the national level for over a decade (Table 4). In June 2002, the Government extended the female stipend programme to include higher secondary education (class XI and XII). However, for higher secondary, the coverage is more selective.

The priorities of the FSP have evolved over its lifetime according to the different priorities of the funding agencies. The most enduring objectives have been to increase enrollment and retention, delay marriage, reduce fertility, and increase employment (Table 5). The quantitative targets (enrollment and pass rates) do not reflect a rights-based agendum, for example, to educate girls to increase their own capabilities, improve their quality of life, or because education is their right. Instead, girls are seen as important players in the development of the nation as potential members of the workforce and future mothers. Subsequently, the agendum for empowerment of girls is subsumed in the primary priorities of the programme, although empowerment may have propelled the success of other goals. The negligence of ‘empowerment’ as an indicator of programme success reflects the difficulty of defining and measuring ‘empowerment’ in culturally-relevant terms and the challenge girls’ and women's empowerment poses to traditional values.

Table 5.

Changing priorities’ of the FSP

Objective	Project dates (most projects are of 5 years’ duration)
	FESP, 1984	FESP, 1990	FESP, 1990	FESP, 1992	FSSAP 1993/1999	NWFSP, 1994	SEDP, 1995	NWFSP, 1994	FSSAP II, 2004	FESP, 2004	SESIP, 2004	FSSP, 2004	HSFSP, 2004
Enrollment	x	x	x	x	x	x	x	x	x	x	x	x	x
Retention	x	x	x	x	x	x	x	x	x	x	x	x	x
Fertility control	x		x	x		x		x		x	x	x	x
Delayed marriage		x				x		x					x
Employment/income		x	x		x	x		x		x		x
Enhanced status		x		x		x
Decision-making			x									x
Socioeconomic development				x	x	x		x		x		x	x
Poverty alleviation				x
Quality education					x			x	x		x
Further education					x	x		x
Female teachers					x
Public awareness					x				x
Health and security					x				x
Capacity-building					x				x	x		x
Physical infrastructure					x		x		x	x
Empowerment/equality									x	x			x
Study science													x

Adapted from Raynor, 2004 (32) FESP=Female Education Stipend Project; FSSAP=Female Secondary School Assistance Project; HSFSP=High School Female Stipend Project; NWFSP=Nation-Wide Female Stipend Program; SEDP=Secondary Education Development Project; SESIP=Secondary Education Sector Improvement Program

As the stipend is not targeted, it is a very expensive programme, accounting for 15% of the secondary education budget and 6% of the total education budget, causing critics to question whether it is sustainable in its current form (32). Alternative approaches, however, present other barriers. Targeting only the most excluded (for example, the ultra-poor, minorities, or most remote) would also be stigmatizing for recipients, expensive, in addition to being unpopular and otherwise difficult to administer. If girls from influential families were eliminated from the programme, strategic community support would likely be withdrawn.

IMPACT OF THE FEMALE SECONDARY SCHOOL STIPEND PROGRAMME

Unfortunately, data for adequately assessing the impact of the FSP are limited. The management information system of the Ministry of Education maintains few variables on students and schools. While FSP students are given identification numbers, they are not tracked; so, it is not known if specific students passed or even sat for their examinations (34). No comprehensive assessment of impact has been attempted, despite the widespread praise the programme has received. At this point, designing an assessment of impact would be challenging: the national scope of the programme eliminates the option of a counter factual, no baseline data were collected, socioeconomic data are not collected, and enrollment and pass figures are too often forged at the school level (32). Furthermore, the full impact of education can take time to manifest and may only be demonstrated after a generation, if education leads to delayed marriage, lower fertility, increased employment, increased mobility, etc. In the absence of rigorous programme-specific data, this section examines the available literature and data about the school system overall to draw conclusions about the effects of the stipend and also highlights areas that need more research.

Quality of education and empowerment of women and girls

Increased enrollment has put a huge strain on the secondary education infrastructure, making it a victim of its own success. The secondary education sector has grown at twice the rate of the primary sector. This is not only due to the FSP (which implies a large influx of first-generation learners) but population growth and a move away from a subsistence agrarian economy (increasing the need for formal employment) are also contributing factors (35). Predictably, the increased enrollment has had an adverse effect on quality (33). Data from Education Watch 2005 indicate that learning achievement is low, teachers are under- and misqualified, and infrastructure is strained; for example, one-quarter of schools had no toilets. Numbers of teachers are also low but this is offset by student absenteeism.

Enrollment in the FSP requires a 45% pass mark but the extent to which this serves to ensure high-quality educational experiences is unclear. The emphasis the FSP projects put on quantitative targets, such as enrollment and pass rates, may contribute to the neg-ligence of important issues of quality (32). Furthermore, school-level data indicate that School Management Committees are more interested in increasing enrollment to receive government grants rather than improving the quality of education (16).

At the end of 10th grade, students take the Secondary School Certificate (SSC) examination. The failure rate of this examination is usually very high; in 2003 and 2004, 64% of all students taking the examination failed. The pass rate appears to be improving: it was 53% in 2005 and 58% in 2006. Unfortunately, this improvement is slow. Based on the current trends for every 100 students newly arriving in grade 6 in 2005, only 35 will pass the Secondary School Certificate. Of those who then continue schooling, less than 10 will pass the Higher Secondary Certificate examination (36). An analysis of 2005 SSC data shows that there are gender gaps for those entered for the examinations and those passing it. As a percentage of the total, girls were only 46% of those entered, and less than 44% of those who passed (37). For boys and girls, the system creates a large percentage of failures.

The pass rate of this examination is an indicator of quality but the examination is not standardized and is run by nine different boards throughout the country—so, it is not a rigorous indicator. There are no standardized national indicators for quality, or curriculum content. Little is documented, for example, on what is taught in the quomi madrasahs, which account for 15% of all female enrollments (33).

There is evidence from other educational programmes that, while the quality of education is generally low, it is even lower for girls. A study in BRAC primary schools, which have explicit goals in terms of gender equality, indicates that girls are less likely to get the attention of teachers, less likely to have their experiences reflected in the national textbooks, less likely to be expected to do well in mathematics and are, therefore, less likely to succeed (32). A 2003 World Bank report noted that, despite increased enrollments, the girls:

… received little, if any, instructional support. Aside from general upazila targeting, FSSAP and its follow-on project have had no special activities to help very poor girls stay in school once they enrol. … Without satisfactory learning outcomes, the girls cannot become teachers or get employment that will significantly empower them and alleviate their poverty (34).

This area needs more research but suggested reasons for their poor outcomes would be a heavier domestic workload, discrimination in the classroom, and lower societal expectations of girls.

Low-quality education does not just waste time of students but also lowers their morale and provides a disincentive to maintain attendance or their studies and causes significant grade repetition (1), increasing the overall cost of education. Education Watch 2005 found that drop-out and repetition rates were higher for girls than boys. The gap was the widest in class X—the promotion rate was 76% for boys and 65% for girls from class IX to class X. Due to the sporadic quality of teaching, achieving a pass can mean a higher cost to families, usually by hiring the class teacher as a tutor. Lower rates of achievement by girls suggest that these high costs affect girls more than boys, with families perhaps less willing to hire tutors for girls.

Alongside the quality of education is the question of utility. Secondary education in Bangladesh is primarily a screening device for higher education; it has little utility of its own (33) other than increasing status and delaying marriage and childbearing. Critics claim that the curriculum has no technical or life-skills and does not prepare girls for employment (33,32,38). While just being in school is likely to increase girls’ extra-kin horizontal social networks and social capital, the programme is not designed to give girls a chance to enjoy their childhood and adolescence, gain empowerment, or enhance employment opportunities (despite these being written in programme priorities). Arends-Kuenning and Amin argue that the school curriculum in Bangladesh has no transformatory significance (16). There is some evidence that the FSP may increase parental perception of the value of girls but this perception is likely based primarily on the stipend attached to the girls’ participation, not the independent value of girls themselves (16). Finally, the popularity of the programme may hinge on its subsumation of explicit empowerment goals. Support of the community for the programme may be derived from the belief that education makes women better wives and mothers, not more empowered people (32). A number of critics argue that, if empowerment were the explicit objective, fewer families would keep their daughters in school (16,32). For the strategic reason of gaining and maintaining community support and political viability, the programme administrators restrain from disrupting gender norms.

Delaying marriage and first birth

One of the FSP objectives is to delay marriages. A 2002 World Bank report suggests that the impact of the programme on delaying marriage was immediate and significant, claiming that the overall proportion of females who married declined from 29% in 1992 and 14% in 1995, for 13–15 years old girls and from 72% to 64% for 16–19 years girls (study design and source of data not cited) (33). Findings are not reflected in the latest Demographic and Health Survey data, which has not found any change. The median age of marriage is still 16 years. Streatfield suggests that, despite this delay in impact, it is safe to predict that ongoing investment in education of girls will have an effect on age at marriage and, therefore, fertility (5). Qualitative data suggest that the stipend does enter into the parents’ decision-making and calculations of when to have their daughters married (39). However, it may take some time for the full effect of education (in terms of values, knowledge, expectations) to manifest, due to the high influx of first-generation learners. The true impact may only be captured over longer periods of time.

Poverty

Stipends are often included in programmes to make services more available to resource-poor families. The study by Khandker and Pitt, however, indicates that the programme disproportionately affects enrollment of girls from land-rich families (40). Furthermore, the poorest citizens of Bangladesh, characterized by low consumption, no education, high infant mortality, and lack of access to services (35), may receive no boost from the FSP. This is because limited household resources can prevent girls from completing primary schooling to be eligible for the secondary school programme; if enrolled in secondary schools, the stipend will not cover all educational costs incurred by families, and poor students may receive a lower quality of education than their peers if they cannot afford tutoring services. Teachers have an incentive to spend more time with students who are paying them to teach after hours, clearly putting girls with fewer resources at a further disadvantage (34). As a result, the FSP primarily benefits families who can afford to put their children through primary school and makes insufficient provision for the multiple structural disadvantages faced by the poor. In this way, the programme provides financial bonuses to the middle class and the wealthy for sending the daughter to school, instead of addressing poverty. Indeed, by not ensuring access for the ultra-poor, the intervention perpetuates intergenerational disadvantage.

Enrollment and retention

In periods following the implementation of the FSP, the school enrollment recorded for girls increased to overcome that for boys. Recorded enrollments in FSP schools increased by 105% between 1994 and 1999 while aggregate enrollments of girls increased by 111%. Unfortunately, accurate data to measure the specific impact of the programme on enrollment do not exist. The FSP was implemented among many other interventions pushing for increased enrollment of girls, including increased focus on non-formal education and tuition waivers, and it is difficult to extract the impact of the stipend from these other causal factors. It is unlikely that the stipend alone caused changes in enrollments, as the amount does not cover girls’ full expenses and often arrives late. Given that the FSP was implemented in rural areas where barriers to enrollment of girls are particularly high, the stipend may have had a uniquely positive effect for some population subgroups.

The extent to which FSP students remain in school following enrollment is even less clear. School-wide absenteeism and drop-outs are common. Education Watch 2005 found that the primary reasons for drop-outs were scarcity of money, children's dislike of school (quality), and the need for the child to work. Two of these reasons are wealth-based, emphasizing that the FSP is insufficient to keep children of poorer families enrolled in secondary schools. While there is a general paucity of data, there is a lack of specific quantitative data on programme uptake according to wealth quintiles, religion, and ethnicity and a dearth of qualitative data on why girls drop-out or stay enrolled.

The high enlistment in the project, despite the inadequacy of the stipend in covering the household costs for secondary education of girls, may indicate a demand for female secondary education independent of the stipend programme. It may also indicate widespread forging of enrollment numbers. It is likely that, for some families, any amount of extra income is an incentive, even if overheads are higher than expected. However, it needs to be established if the demand would be maintained even if the stipend is withdrawn.

Replicability

Cash-transfer type of interventions reportedly have been successful in other countries, for example, Bolsa Escola project in Brazil, Progresa in Mexico, and Chile Solidario in Chile (41–43). While the Bangladesh FSP focuses on girls, these other projects are targeted to those in most need and are for both boys and girls. In terms of the gender focus, the World Bank has replicated the FSP programme in Punjab region of Pakistan since 2003, a region with particularly poor gender indicators (44). In addition, female school stipend projects are implemented in sub-Saharan Africa (Rwanda, Congo, and Ghana), with the additional objective of reducing the incidence of HIV/AIDS (45).

The non-targeted nature of the FSP is unlikely to be replicated as it is expensive and inefficient. However, the Government of Bangladesh kept it untargeted for the popularity and electoral support it provided. Also, while the project is replicable as is, it may be wiser to emphasize quality of education concurrently with enrollment and retention rates for long-term sustainability. A higher quality of educational supply may attract and retain students more efficiently than increasing demand through the stipend (33).

CONCLUSIONS

It is clear that macro-level economic changes, with a move away from an agricultural economy, are providing some limited opportunities for women to play a more active role in broader Bangladeshi society and the economy. The FSP programmes are timely interventions to assist such economic and social transitions. The social exclusion framework, in hand with Sen's capabilities approach, provides an external lens to analyze the efficacy of this long-established, highly-acclaimed intervention. Drawing upon the existing data, this analysis finds little evidence that the stipend project actually improves girls’ and women's quality of life and opportunities. There is no evidence that girls are receiving a quality education that better prepares them to be participants in social, political and economic life. There is also little evidence that the status of girls, as a group, has changed as a result of the programme.

Often, in developing countries and elsewhere, interventions have to be made on the basis of faith, i.e. education for girls is good; however, such a large-scale and well-funded intervention should have invested financial and technical resources for a monitoring and evaluation plan.

While an overt focus on empowerment for girls may not be widely acceptable in Bangladesh, a focus on improving the capabilities of girls through providing life and livelihood skills (financial management, agricultural techniques) would help girls get a terminal secondary education that is relevant to their income-earning opportunities. Nation-wide indicators need to be established to assess quality across different schools, and ‘quality’ needs to include the quality of the curriculum in terms of relevance and utility. Performance appraisals, based on improved indicators of quality assessment, should be tied to grants or other incentives to improve school performance. To best understand the impact of the stipend on the most excluded groups, data specific to programme participants, including socioeconomic background and school achievement indicators, should be maintained on a national level. While costly, this will better establish the impact of the programme.

Although specific data are limited, the programme is likely responsible for increasing enrollment of girls and at least partially responsible for delaying marriage, thereby delaying fertility among programme participants. The results of social exclusion analysis suggest that, throughout much of its existence, the FSP has lacked a capabilities-development emphasis and, through its ‘non-targeted’ approach, may actually be benefitting the rural rich and middle class the most. However, if the concepts of social exclusion and development of capabilities are explicitly addressed in programme design, the programme might have a greater impact as a means to enable girls and women to participate more fully in society. While the social exclusion framework is an external framework used here for analyzing an established intervention, the findings suggest a means to increase the impact of the intervention according to its own priorities: reducing fertility, and generating empowerment and positive economic outcomes. As is, the project seems to do little to increase the social, cultural and economic capabilities of girls, representing a missed opportunity to transform girls’ and women's lives.

ACKNOWLEDGEMENTS

This work was funded by the World Health Organization (WHO) and undertaken as work for the Social Exclusion Knowledge Network established as part of the WHO Commission on the Social Determinants of Health. The views presented in this publication are those of the author and do not necessarily represent the decisions, policy, or views of WHO or Commissioners.

The author thanks the following people for review and feedback of previous drafts of this paper: Heidi Johnston, Shelley Golden, Anne Martin Staple, Kim Streatfield, Janet Raynor, Rosemary Fernholz, Micah Risher, Liz Miller and anonymous reviewers.