March 2009 - Volume 3, Issue 2

Fertility Decline - the Role of Non-Government Organizations (NGOs) in Bangladesh: A Micro Data Analysis

	Dr. Md. Abdul Goni, Associate Professor
	Department of Population Science and HRD University of Rajshahi, Rajshahi-6205, Bangladesh Email: magoni_popsrubd@yahoo.com

ABSTRACT Bangladesh is one of the most populated countries of the world. In the last two decades despite pervasive poverty and under-development Bangladesh has achieved considerable fertility decline. Unfortunately, recent statistics suggest that despite continued increase in contraceptive use and the functioning of NGO activities the fertility decline has stalled. Thus Bangladesh becomes an interesting case study for exploring the question, of what factors are necessary to bring about further fertility decline. In this paper an attempt will be made to highlight women's status, socio-economic and demographic factors, which are affecting the decline in fertility. It is found that the desire for an additional child is lower among working women who have better access to mass media regularly and are involved in NGO activities. In order to achieve further decline in fertility, therefore, the policy makers should emphasize those policies that actively enhance women's status through education, more access to mass media as well as getting them involved in the workforce in the country and need to encourage other NGOs to include family planning and education programs like BRAC and Grameen Bank. Key Words: Fertility, NGOs, Women's Status, Micro-Data, Logistic Regression and Bangladesh.

INTRODUCTION

Women's status and their role in decision-making in reproduction have an important bearing on the success of family planning and the long-term reduction in fertility of a country. In a society in which women are mainly concerned with domestic affairs and raising children, fertility would be higher. The United Nations International Conference on Population and Development in Cairo, 1994, and the Fourth World Conference on Women in Beijing, 1995, outlined factors considered critical to "the empowerment of women". At these meetings, 179 countries agreed on a 20 year plan to stabilize the world's population, premised on the notions that population, development and the environment are integrally linked and the empowerment of women is required to make this vision a reality. Empowerment of women was defined- to include providing women with access to employment, education and reproductive health care, free from discrimination, coercion and violence. These same factors are also linked with fertility decline, a global phenomenon that causes growing concern for governments planning for future workforce and social security needs. The 1994 International Conference on Population and Development in Cairo have stressed the integration of population policies with development policy.

Bangladesh is a nation with an estimated population of 130 million (BBS-2001). Except some Island States, Bangladesh has the highest population density in the World; resource scarcity and subsistence-level economic conditions characterize the Bangladesh economy (Khuda, 1991). Per capita income is as low as US$ 444 and more than one-third of the population lives below the absolute poverty line (UNDP, 2003). In Asia, Bangladesh is one of the few countries where women's life expectancy remains lower than that of men. This is due in part to multiple high-risk pregnancies. The status of women appears to have changed somewhat over the past few decades, largely to their disadvantage. The shift is evidenced, for example, by a major trend away from bride wealth toward dowry and a general reduction of women's control over property transferred with them at marriage. The status of women, contemporarily at least, appears to be very low: women are powerless and depend upon men.

However, the past three decades saw a substantial decline in total fertility rate, for example, from 6.3 in 1975 to 3.3 in 2000 despite pervasive poverty and under-development, although recent statistics suggest that despite a continued increase in contraceptive use and NGO activities the fertility decline in Bangladesh has stalled.

It is interesting to note that the period saw various types of NGOs working in the country for the development of our society, particularly in villages where more than 80% of women live. Besides Micro credit (MC) activities most of the NGOs are working to improve health and education and trying to develop a new lifestyle among the Grameen (village) Society. The availability of various income-generating organizations such as Bangladesh Rural Development Board (BRDB), Bangladesh Rural Advancement Committee (BRAC), Grameen Bank and Mothers' club/Mothers' association may well have influenced women's reproductive behavior and consequently influenced fertility. Some NGOs have collaborated with the government to immunize children and pregnant women. Under the essential health care program, the NGO workers provide services to pregnant women for improving their health and nutritional status. The reproductive health needs of the community in general, with particular focus on NGOs members, are addressed through education on family life and awareness of some killer diseases and contraceptive methods. Under the BRAC's health, Nutrition and Population Programme, trained health workers and volunteers work to raise awareness among the rural poor about health issues that directly impact their lives and families. The NGO sector is the second largest source of contraceptive supplies for family planning users in the country (BDHS 2000).

In this paper, I would like to show by making use of micro data how different dimensions of women's status are related to different socio-economic and demographic factors. One factor that I will give particular emphasis in this paper is NGOs, which are believed to have positive interactions with some of the aspects of women's status. The structure of the paper is as follows. In section 2, I first discuss women's status, fertility, their employment and education situations, and possible relationships between female employment, fertility and related variables from a demographic point of view. Statistical methods will be applied to a Bangladesh Demographic and Health Survey sample (Section 3), whose results are presented in Section 4. Section 5 discusses regional variation, while concluding remarks and policy implications are set out in Section 6.

ON THE STATUS OF WOMEN

Literally "status" means a position in relation to others. Women's status is relative. A measure of women's status may implicity compare the relative position of women to men or women to other women. The status of women in a society is deeply rooted in culture, religious beliefs, traditions and economic environments. Traditionally, while women perform the major roles of reproduction, the household and so on, in recent times, many new roles have been added. The term status of women then would denote not only a conjunction of the rights and duties but also the degree of her subordination in the home, education, economic status, role in decision making in family affairs, and her self-perceived status in the home and in the community. According to the United Nations (UN), the status of women in society can be determined by her composite status which can be ascertained by the extent of control that she has over her own life derived from access to knowledge, economic resources and the degree of autonomy enjoyed in the process of decision making and choice at crucial points in her life cycle.

In recent years, despite increasing attention to the concept of women's status, the meaning of this concept has remained unclear. Among the terms used in the social demographic literature are not only "status of women"(e.g., Dixon, 1978), but also "women's autonomy" (Dyson and Moor, 1983). To measure women's status/autonomy more directly, the BDHS (2000) survey asked about women's roles in household decision making and their freedom of movement. Such information provides insight into women's control over their environment and their attitudes towards gender roles, both of which are relevant to understanding women's demographic and health behavior. Education, exposure to media, employment status and control over earnings are some of the means by which women gain status/autonomy, important aspects of their empowerment.

In the present study, I introduce the unconventional new variables NGOs (such as Bangladesh Rural Advancement Committee (BRAC), Grameen Bank, Bangladesh Rural Development Board (BRDB), Mother's club, etc.) involvement which can increase women's status. In addition included the variables related to women's status which are: residence, employment, education, religion and freedom of access to mass media etc. After independence in 1971, the BRAC was established and is the biggest NGO in terms of development and micro-credit activities. BRAC's other development activities include free informal education (set up 1985, about 49,000 schools in 2004, accounting for 11% of the total primary school in Bangladesh) and health and medical facilities for low-income people in rural areas. Among the employees of BRAC, 61% are women and working with the twin objectives of poverty alleviation and empowerment of the poor women. BRAC outreach covers all 64 districts of Bangladesh and 78% of the total number of villages in Bangladesh. BRAC's has collaborated with the government to immunize children and pregnant women. Under the essential health care program, with the help of Shashtho Shebikas (Health Volunteers) and Shastho Kormis (Health Workers) immunization coverage of the population is 80% (BRAC, 2004). The program also provides services to pregnant women for improving their health and nutrition status. The reproductive health needs of the community in general, with particular focus on BRAC members, are addressed through education on family life, contraception, and awareness of HIV/AIDS. Therefore, due to the BRAC women's education and financial independence, they themselves come forward to learn about family planning and the women become determined to have fewer children. The Grameen Bank (GB) was established 1976 (Founded by the 2006 Nobel Laureate Professor Muhammad Yunus) with the objectives of extending banking facilities to poor men and women; eliminate the exploitation of the poor by money lenders; and create opportunities for self-employment for the vast multitude of unemployed people in rural Bangladesh. As of July 2004, it has 3.7 million borrowers, 96% of whom are women. GB provides services in 46,000 villages, covering 68% of the total villages in Bangladesh. GB has introduced higher education loans for all students from Grameen families who can enter into the higher educational institutions (medical schools, engineering, universities etc.). Students are responsible for repaying the loans when they start earning. Half of the numbers of scholarships are reserved for female students. They believe that education leads to much better jobs and then much better income. BRAC is directly involved in education and health sectors and GB is indirectly involved in education sectors in Bangladesh.

Fertility: Fertility is a key demographic process determining population change. The fertility transition in Bangladesh by observing time series estimates of TFR 1 over the last 25 years, beginning with the 1975 BFS, indicates a decline of 48 percent in TFR; a decline of 1.9 percent per year (Table 1) mainly because of rising contraceptive prevalence, due to effective immunization and reduction of child mortality. The pace of fertility decline has slowed down recently compared to the exceptionally rapid decline in the late 1980s and the early 1990s and since then it has remained almost constant. For example, within a period of five years, starting from 1989 to 1993-94, fertility has declined by 33 percent or 6.6 percent per year, while for the next six year period 1993-94 to 1999-2000, it is only 4% or less than 1% per year. The fertility rates by age group shows that starting from 1975 to 1993/94 fertility declined steadily in 1993/94 in all age groups with the exceptions of age group 15-19. The age specific fertility rates in 1999-2000 also shows decline in all age groups since 1993-94, except the age groups 15-19 and 25-29. The decline is steeper among women aged 35 and above. This indicates that there is a shift in fertility towards younger age groups in recent years and fertility has declined substantially among older age groups.

Note: NA = Not available, * Khulna and Barishal division is combined, ** Chittagong and Sylhet division is combined.
Source: Bangladesh fertility survey 1975 and 1989, Contraceptive prevalence survey 1991.
Bangladesh demographic and health survey 1993-94, 1996-97 and 1999-2000.

Notes: TFR 1: Total fertility rate, expressed per women. Rates are for the period 1-36 months preceding the survey. Rates for age group 45-49 may be slightly biased due to truncation.

The rate of fertility decline in Bangladesh has not been uniform in different regions. Fertility is higher in rural (3.54) than in urban areas (2.45). For urban women, increased participation in the labour force, starting at adolescence, contributes significantly to controlling fertility. The difference is especially large at younger ages, which probably reflects longer education and later marriage of women in urban areas (BDHS 2000). Again Table 1 shows that during late 1980s to early 1990s all the divisions in Bangladesh experienced more or less uniform decline in fertility. Even in Chittagong division having highest fertility in the mid 1980s, which is still continuing, the rate of decline was almost identical in magnitude as compared to that of the other divisions. However, since 1993-94 the decline in fertility is not uniform across the regions. During the period 1993-94 to 1996-97, Chittagong division shows an increase in fertility level by about 3%, while all other divisions show some decline in fertility. The magnitude of decline was higher in Khulna (17%) followed by Rajshahi (8%) and Barishal (5%). On the other hand, during the period 1996/97 to 1999/00, fertility has increased in low performing regions of Khulna and Rajshahi and also in Dhaka, but declined in Barishal and Sylhet divisions.

There are so many NGOs currently working on health and family planning as well as other problem areas in different regions of Bangladesh that their activities may have encouraged people to adopt contraception in the late 1980s, precipitating a change in reproductive behavior. BRAC is directly working in family planning sectors while the Grameen Bank (GB) is in economic and financial sectors. But their activities are not uniform between the regions. It appears that they are more active in regions, such as Khulna and Rajshahi, where fertility decline has been substantial, suggesting a link between services they have created for women and fertility decline.

Women's Employment: Traditionally, women were involved largely in the non-monetised sector and in subsistence activities. In recent decades, women have experienced some major changes in the labour market, primarily in terms of the nature of jobs and opportunities. Table 2 shows that the rate was gradually decreasing till 1981 but there is an increase in 1991 and again a decrease in 2001 except for women's.

Source: Bangladesh Population Census 2001
Notes: 2 Under the extended definition of labour force, persons who are engaged in some major household activities such as food processing, threshing, cleaning, and livestock and poultry raising are considered as part of the economically active population.

Table 2 shows that the activity rate was highest in1961 for both males and females, which is undoubtedly due to the inclusion of household work in the economic activity during that time period (BBS 1981). From 1981 there has been an increasing trend in the female activity rate, which tremendously increased between 1991 and 2001. The urban female activity rates observed from different census years are higher than that of rural areas since 1974. Increases in female participation rate in urban areas from 6% in 1974 to 8% in 1981 and 10% in 1991 and 14% in 2001 display more and more involvement of females in economic activity outside the home. Also the proportion of females economically active in the non-agriculture sector increased substantially from 1961 to 1991. In 1961, only 8% of the total female workers were engaged in non-agriculture pursuits, which increased to 82% in 1991 although the percentage decreased to 56% in 2001. It is worth mentioning that a significant number of women also work as teachers (primary and BRAC School), lawyers, journalists, government employees and for non-government organizations (NGOs). Women's increasing labor force participation, on the one hand, and public awareness and efforts by NGOs and mass media, on the other, have played an important role in encouraging this.

In spite of these achievements, however, the majority of women in Bangladesh have yet to be empowered to participate actively in the social, cultural, economic and political life of the country. The policies and programs of the Government, some NGOs and other institutions do not sufficiently address the need for women's empowerment.

Women's Education and Training: Education is a key determinant of lifestyle and status, in order for an individual to be able to enjoy life in a society. It affects almost all aspects of human life, including demographic and health behavior. In Bangladeshi society, as boys are perceived to take care of parents, sending them to school and investing in their education is more common than for girls. Nevertheless, the primary level enrolment rates for girls have gone up remarkably (Table 3). The gap between male and female enrollment rates, which stood at 22 percent in 1985, declined to 3 percent over the past decade and a half. Although noteworthy achievements have been made in female enrollment at the primary level, progress has been very slow in secondary level education. At this level male and female enrollment rates improved only by 5 and 10 percentage points, respectively, between 1985 and 1995. At the primary level, girls' enrollment is increasing. In 1990 45% of girl's aged 6-10 were enrolled but in 2000 it was 49%. However, the dropout rate of female students who went on to secondary school reaches half, 10 percentage points higher than their male counterparts. Very few women continue their education up to the tertiary level. This negatively affects the overall rate of return from education and women's labour markets.

n.a. = not available
Source: Sample Vital Registration, BBS; Multiple Indicator Cluster Survey, BBS; Statistical Yearbook, 1997, BBS; Population Census (2001).

Notes: Secondary level covers VI to XII and age 13-16 group. 1 The adult literacy rate is defined as the proportion of literate population of age 15 years and above to the total population of the same age groups. 2 The primary school enrolment ratio is defined as the ratio of enrolled students belong to age group 6-10 years in the primary school to the total population of that age and expressed in percentages. 3 The secondary school enrolment ratio is defined as the ratio of enrolled students of age 13-16 years and above in the secondary school to the total population of that age groups and expressed in percentages.

Lower access to technical education and secondary and higher education, gender-biased curriculum and curriculum without job prospects are critical concerns for women's education that must be addressed through coordinated efforts. Adult education for women has also been emphasized. There is also provision for increasing teachers' training of women for both the primary and secondary levels. Government stipend schemes for girls up to grade 8 are some of the important factors behind the rise in girl's enrollment at the primary level. Non-formal education programs run by Bangladeshi NGOs such as the Bangladesh Rural Advancement Committee (BRAC) have been internationally acclaimed. Some large NGOs run thousands of schools providing education to children who are not able to avail themselves of Government-sponsored education. The Government has recognized this contribution of NGOs and there is significant collaboration between them. There is special emphasis on girls' education as education and eradication of illiteracy are considered to be the first steps toward the empowerment of women.

POPULATION AND FERTILITY DATA

This study utilized mainly the 1999-2000 Bangladesh Demographic and Health Survey (BDHS) based on a nationally representative, two-stage sample that was selected from the master sample maintained by the Bangladesh Bureau of Statistics (BBS) for the implementation of surveys before the census (2001). Other information comes from the Population census 2001. The 1999-2000 BDHS collected information on a respondent's background characteristics (age, residence, education, religion, etc), employment history and occupation, contraceptive use history, marriage and fertility preferences. The master sample consists of 500 primary sampling units (PSUs) with PSUs in each stratum except for the urban strata of the Barishal and Sylhet divisions. In the rural areas, the primary sampling unit was the mauza, while in urban areas, it was the mahalla. Mitra and associates conducted a household listing operation in all the sample points from September to December 1999. A total of 10,268 households were selected for the sample of which 9,854 were successfully interviewed. In these households, 10,855 women were identified as eligible for the individual interview (i.e. ever-married and age 10-49) and interviews were completed for 10,544 (97%) of them, with 9,720 currently married women. But my analysis covered only 9502 women who are able to bear children. Infecund, divorced, widowed etc. women were not involved in the analysis.

In BDHS, currently married women were asked "would you like to have (a/another) child or would you prefer not have any more children"? Interviewers were instructed to use the words in parenthesis depending on whether the respondent had children or not. Pregnant respondents were asked if they wanted another child then asked how long they would like to wait before the birth of the next child. Almost 52 percent of currently married women age 10-49 in Bangladesh said that they wanted no more children and an additional 7 percent had been sterilized. 37 percent of women wanted to have a child at some time in the future; however the vast majority of these women answered that they would like to wait two or more years before having their next birth (Table-4). The desire for additional children declined noticeably in Bangladesh over the past decade. In 1991, 45 percent of married women with two children wanted to have another child in the future (Mitra et al., 1983:84); in the 1999-2000 BDHS survey, the proportion was only 30 percent. Conversely, the percentage of women with two children who wanted no more children or who had been sterilized rose from 48 percent in 1991 to 66 percent in 1999-2000. There was little change in overall fertility preferences since 1996-97, with the proportion of women who either wanted no more children or who had been sterilized increasing from 58 to 59 percent.

For analytical purposes, I divide the above data into two categories: one is women who wanted another child and the other is those did not want another child.

METHODOLOGY AND ANALYSIS

In this section, we examine what affects fertility outcomes, given stated preferences for more children or for no more children. In this paper for analytical purpose, logistic regression models are used. The logistic regression model has become the standard method of analyzing data in which the dependency of a binary response variable is being tested on a number of explanatory variables. A response (dependent) variable (Y) is binary. In our analysis, it can take "1" or "0" depending on whether a respondent (an ever-married women in the age group 10-49) wants another child or not. Variables of this type are often called binary or dichotomous variables. The SPSS for windows version 11.5 is utilized for the logistic regression analysis.

The multivariate analysis controls for various characteristics that may have affected a woman's propensity to bear children. The probability of having a child generally decreases with the woman's advancing age owing to growing infecundity and less frequent sexual activity. The number of living children, controlling for the woman's age may reflect some measure of the woman's fecundity (in that she has a higher probability of conceiving and thus a greater potential for more births in her reproductive period) in addition to preferences for more children or to an infrequent use of contraception. A woman with higher education may be better able to control family size because she tends to be an effective user of contraception (Mamlouk, 1982; Rodriguez, 1979). In addition, some evidence indicates that with increasing education and greater participation of women in the labour force, domestic labour becomes more equally divided between husband and wife. Although women still perform the most domestic work, this change may precipitate a shift to a smaller family size because men want to do less work at home after returning from outside activity (UN, 1985).

In this paper, most of the independent or explanatory variables are quantitative and for the purpose of comparison we converted then into qualitative variables of interval scale and the respondent belongs to a particular interval scale that has the value '1' otherwise '0'. The independent variables include the woman's residential status (rural, urban and administrative division), employment status (employed, not employed), occupational status (agriculture, non-agriculture, working for cash only, working for kind only), household head (male or female), educational status (no education, primary level, secondary level and higher education), religion (Muslim, Hindu, Christian and others), age group (10-14, 15-19, 20-24, 25-29, 30-39 and 40-49), exposure to mass media (listening to radio irregularly/regularly, watching TV irregularly/ regularly), number of living children (below 2, exactly 2 and above 2 i.e. 3 or 4), pregnancy status (currently pregnant, not pregnant), involvement in NGOs (Grameen Bank, BRAC, BRDB, Mothers club/Mothers' association and Others), and her husband's occupational status (agriculture, non-agriculture) and educational status (no education, primary level, secondary level and higher education). We also include interaction terms between some of these independent variables. Table 5 reports the results of this multiple binary logistic regression analysis. In order to measure the effect of each explanatory variable, we calculate an odds ratio, which is the anti-log of the co-efficient for each variable.

r = Reference category, *Significant at P<0.001, **Significant at P<0.01, ***Significant at P<0.05
****Significant at P<0.10
Notes: Not employed = housewife/housework. Agriculture = Own land, Rental land and other's land worker/ Cultivator. Non-agriculture= Professional, Technical, Clerical and Managerial occupations.

The results show, first, that women residing in urban areas are generally less likely to want more children than those in rural areas and the results are statistically significant. For individual administrative divisions, the regression coefficients corresponding to the Chittagong and Sylhet division are positive but the Rajshahi, Khulna and Barisal division the coefficients are negative in sign and statistically significant except Barishal division. The effect in the Chittagong and Sylhet divisions is 1.7 and 1.84 times higher than that of the Dhaka division. On the contrary, in Khulna, Rajshahi and Barisal it is (1-0.62) x 100 =) 38%, 25% and 6% less than that of Dhaka.

With respect to employed women in comparison with non-employed women (housewives), the results show that the desire for another child among working women is 9% less than that of non-working women and that the women engaged in the non-agriculture sector want fewer children than those in the agriculture sector. The category "working for cash" is a better indicator of women's working status, for which we found that the regression coefficient is -0.09, which is greater than that for the umbrella of non-agriculture. The odds ratio corresponding to the women working for cash is 0.91, indicating that the desire for another child is 9% less than that of women whose earnings are in kind.

The effect of age is important. Table 5 shows that the ever-married women under age 30 are likely to want more children and above age 30 are less likely to want any more children and the results are statistically significant. Female headship also turns out to be statistically significant. The female-headed family is less likely to want more children than the male headed.

Education is another important characteristic of the women. Increased education is often suggested as one way to raise women's status and to prompt their participation in the labour force. The regression coefficients of the primary level, secondary level and higher education are 0.05, -0.11 and -0.19 respectively, which are negative in sign and statistically significant except for primary education. The women's husbands extend similar effects on the demand for children through their employment and education.

Religion has a significant relevance in demography. Muslim communities consistently tend to show higher fertility than many other non-Muslim communities. In our analysis, 88 percent of the respondents practice Islam and it is evident from the data of BDHS 1996/97 that the proportion of women currently using any contraceptive method is 49%, lower than for non-Muslim 59% women. Foe example, the desire for another child among Hindu women is 69% less than that of the Muslim women.

According to1996/97 BDHS data, the percentage of women exposed to Television has increased sharply (from 29 to 35 percent), while the percentage who listen to the radio has declined (from 39 to 29 percent). The regression coefficient of mass media is 0.14 for radio and -0.03 for TV, and it is for TV only that the coefficient has a right sign and is statistically significant. This result indicates that the demand for additional children among the women who watch TV regularly is 3% less than that for women who do not watch TV regularly.

The desire for an additional child also varies with the number of existing children. The regression coefficient is 3.0 if the number is below 2; if it is 3 the coefficient becomes -0.47 and -1.21 when it is 4. All the results are statistically significant. The odds ratios indicate that the women having less than two children are 21 times more likely and the women having 3 and 4 children are 38% and 70% less likely to want additional children than women having two children. It is interesting to note that the regression co-efficient of currently pregnant women is -2.24 and statistically significant. The odd ratio indicates that the desire for another child is 89% lower among the currently pregnant women than among non-pregnant women.

The next important variable is NGOs that are widespread in Bangladesh. In my study, a total of 2405 respondents are involved in different NGOs and out of them 1731 (72%) of respondents do not demand any more children. This proportion is highest in BRAC (21%) and next GB (17%). For measuring the association between fertility preference and involvement in different NGOs I would like to use the following hypothesis. H0: There is no association between fertility preference and involvement in NGOs Vs H1: There is some association between the two attributes. To test the hypothesis we construct the following 2×2 contingency table.

* Significant at p< 0.001, **Significant at P<0.01
Source: Goni, M. A. (2008).

Through x2 as well as Mantel- Hansel test it is observed that there is significant association between fertility preference and involvement in NGOs and the direction of the association is negative. Here we pay particular attention to the Bangladesh Rural Advancement Committee (BRAC) and Grameen Bank (GB). Their coefficients are negative, which suggests that those NGOs are instrumental in reducing fertility. According to Table 5, it is only BRAC that shows a statistically significant result. The odd ratio indicates that the women under the activities of BRAC have 26% lower demand for another child than that obtained from the women under the activities of Bangladesh Rural Development Board (BRDB). But Table 6 shows that all the NGOs have highly significant effects on fertility except Mothers club.

REGIONAL VARIATION

In the previous sections it was discussed that the rate of fertility decline in Bangladesh was not uniform across regions (Table 1). It is evident that the fertility level has declined rapidly in Khulna and Rajshahi Divisions. On the other hand, Sylhet and Chittagong Divisions are lagging behind, while the performance is in-between in Dhaka and Barisal Divisions. According to the DHS findings on the age at marriage and fertility patterns in Rajshahi, Khulna and Chittagong (Islam and Zaman, 1995 a,b), it is also noteworthy that the age at marriage is lowest in the high performing regions of Khulna and Rajshahi. Marriage age and the cumulative number of births are related. The women of Khulna and Rajshahi get married earlier so that they tend to start childbearing at an early age (high teenage fertility). However, it is interesting to note that the use of contraception is highest in these divisions. In contrast, women living in Chittagong and Sylhet Divisions get married at relatively higher ages but they tend to have more children within a shorter span of life, since the use of contraception is relatively lower there. This is clear from the data of BDHS 1999-2000: the percentage of women currently using contraceptive methods for Dhaka, Chittagong, Rajshahi, Khulna, Barisal and Sylhet is 54, 44, 59, 64, 59 and 34 respectively. It is also evident from the data that the percentage of working women is significantly lower in Chittagong and Sylhet as compared with Rajshahi and Khulna, while contraceptive use is generally higher among working women than among non-working women (Mitra and Associates, 2000). Studies have shown that Dhaka, Chittagong and Khulna are more urbanized while Rajshahi, Sylhet and Barisal are less urbanized (ranked according to per capita GDP and socio-economic index. See Islam 2003). More than 80% of women live in areas classified as 'rural'. BDHS 2000 data indicate that 56% of ever-married women in Bangladesh live in places where Grameen Bank, BRAC and Mother's club are active, and 60% live in areas where cooperative societies are functioning. It is also worth noting that BRAC and Grameen Bank are more active in rural divisions such as Rajshahi, Khulna and Sylhet than in others divisions. And it is these rural divisions where higher levels of contraceptive use are observed.

CONCLUSION AND POLICY IMPLICATIONS

The following points can be made by way of conclusion:

1. As expected, women's employment exerts a discouraging effect on the desire for another child. This effect is particularly marked for those engaged in non-agriculture activities and working for cash.
2. Another important factor reducing fertility is education. This operates through both men's and women's education. However, it is more important for women to receive secondary and higher education: they tend to want fewer additional children than non-educated women.
3. It is observed that mass media, especially television (TV), also plays a role in reducing the desire for more children.
4. It is also found that NGOs are instrumental in affecting fertility preference. Since the NGOs increase women's awareness and educational levels while creating employment opportunities, women involved in those activities can increase autonomy and gain empowerment. Especially, BRAC and Grameen Bank are likely to have been exerting an impact on the desire for an additional child.

The first three of these points have been shown in many studies, both in Bangladesh and elsewhere. But the fourth point is a new finding, and has important policy implications. The government cannot successfully intervene in all the areas commented above. For example, it is not easy for the administration to take a direct measure to increase the employment of women except in the state sector.

It is evident that the government should invest more in education for girls. Setting up more girls' schools is one of such measures, and parents should be encouraged to send more daughters, not just sons, to secondary and higher education. Therefore it should be considered that NGOs are partners in campaigns for women's empowerment and also for fertility decline. Many NGOs have their own programmes to increase the use of contraceptive measures, but what our research suggests is that women's involvement in these activities is likely to have interactions with the effect of education of fertility preference, so that the overall impact on fertility decline is expected to become more effective.

Acknowledgements: The author wishes to express his deep thanks to Professor Osamu Saito, Hitotsubashi University, Tokyo, Japan for reading carefully this paper and making his valuable comments and suggestions which was essential for the improvements of this manuscript.

REFERENCES

Alan, G., 1986. Rising Women's Status in Developing countries Lowers Fertility Rates, International Family Planning Perspectives, Vol. 12, No. 4, 136-137.

Asian Development Bank (ADB), 2001. Women in Bangladesh. Country Briefing Paper, Programs Department (West).

Bangladesh Bureau of Statistics, 2003. Population Census-2001, National Report (Provisional). Planning Division, Ministry of Planning, Dhaka, Bangladesh.

__________________, 1984. Bangladesh Population Census 1981, Analytical Findings and National Tables, Statistics Division, Dhaka, Bangladesh.

Bloom, David E. and Jeffrey G. Williamson, 1998. "Demographic Transitions and Economic Miracles in Asia," World Bank Economic Review 12(3): 419-55.

Balk, D., 1994. Individual and Community Aspects of Women's Status and Fertility in Rural Bangladesh, Population Studies, Vol. 48, No. 1, 21-45.

Bongaarts, J. and Robert G. P., 1983. Fertility, Biology and Behavior: An Analysis of the Proximate Determinants. New York: Academic Press.

BRAC (Bangladesh Rural Advancement Committee), 1990. A Tale of Two Wings: Health and Family Planning Programs in an Upazila in Northern Bangladesh, (Dhaka, BRAC Printers).

--------, 2004. About BRAC, See the Websites: http://www.brac.net/ and http://banglapedia.search.com.bd/

Cain, M., 1982. "Perspectives on family and fertility in developing countries", Population Studies, 36(2): 159-175.

Clealand, J., 1994. "The Determinant of Reproductive Change in Bangladesh, Success in Challenging Environment Washington, D.C. World Bank.

Chudhury, H. R., 1978. Female Status and Fertility Behaviour in a Metropoliton Urban Area of Bangladesh, Population Studies, Vol. 32, No. 2, 262-273.

Chudhary, G., 1982. Social Aspect of Fertility, Vikas, New Delhi

Cain, Mead, 1984. Women's Status and Fertility in Developing Countries, World Bank Staff Working Papers, Number 682.

Chowdhury, A.M.R. and Huda, K.S., 1990. "Lessons from successful experiences", in: South Asia Study of Population Policy and Programmes, M.B. Duza (ed.) (Dhaka, United Nations Population Fund).

Chowdhury, A.N., 1990. Let Grassroots Speak: People's Participation, Self-help Groups, and NGOs in Bangladesh, (Dhaka, University Press Limited).

Cain, M. Khanam, S. R. and Nahar, S., 1979. "Class, patriarchy and women's work in Bangladesh". Population and Development Review, 5(3): 405-438.

Dyson, T. and Moore, M., 1983. "On kinship structure, female autonomy, and demographic behavior in India, Population and Development Review, vol. 9, No. 1, pp.35-60.

Dixon, Ruth, B., 1978. Rural Women at Work: Strategies for Development in South Asia. Baltimore: Johns Hopkins University Press.

Epstein, T. S., 1982. "A social anthropological approach to women's roles and status in developing countries: The domestic cycle. London: Croom Helm.

ESCAP, 1987, Female Autonomy and Fertility: An Overview of the Situation in South Asia. Asia-Pacific Population Journal, Vol. 2. No. 4 pp. 43-52.

Gajigo, O., 2003. The Interaction Between Health and Fertility: Evidence from the Ivory Coast, Journal of Health and Population in Developing Countries, Vol. 3, No. 3.

Groenen, P. J. F. and Koning, A. J., 2004. A New Model Visualizing Interactions in Analysis of Variance. Econometric Institute, Erasmus University, Rotterdam, Netherland.

Goni, M. A. and Imon, A.H. M. R., 2005. "On the Decline in Fertility: A Comparative Study among SAARC Countries", Man in India, Vol. 85, Nos. 1& 2, 93-102.

Goni, M. A., (2008). Fertility Decline, Women's Status and Non-Government Organizations (NGOs) Evidence from Bangladesh. Unpublished Ph.D. dissertation, Graduate School of Economics, Hitotsubashi University, Tokyo, Japan.

Hosmer, D. W. and S. Lemeshow, 2000. Applied Logistic Regression, Second Edition, John Wiley and Sons, Inc. New York.

Islam, M. A. and Zaman, A. K.M.R., 1995. "Achievements and Challenges of the family Planning Program in the Khulna and Barishal Division". Paper Presented at a Regional Workshop in Khulna in 1995 sponsored by PDEU and RTI.

Islam, N., 2003. Urbanisation, Migration and Development in Bangladesh: Recent Trends and Emerging Issues, Demographic Dynamics in Bangladesh, 125-146.

Khuda, B., 1991. Fertility decline amidst unfavorable socio-economic conditions: Bangladesh a test case, Paper presented at Workshop on the Development of long-term strategic plan for FPAB, Cox's Bazar.

Khuda, B. and M. B. Hossain, 1996. Fertility decline in Bangladesh: towards an understanding of major causes, Health Transition Review, Supplement 6, 155-167.

Maharatna, Arup, 2000b. 'Fertility, Mortality and Gender Bias among Tribal Population: An Indian perspective', Social Science and Medicine, 50: 1333-1351.

Mamlouk, M., 1982. Knowledge and use of contraception in twenty developing countries, reports of the World Fertility Survey, Population reference bureau, Inc.

Morduch, M., 1999. The role of subsidies in microfinance: evidence from the Grameen Bank. Journal of Development Economics, Vol. 60, 229-248. Princeton University, USA.

Mitra and Associates, 2001. Bangladesh Demographic and Health Survey 1999-2000. Dhaka, Bangladesh and ORC Macro Calverton, Maryland USA.
_______________, 1997. Bangladesh Demographic and Health Survey 1996-1997. Dhaka, Bangladesh and ORC Macro Calverton, Maryland USA.

Mitra, S. N. and Kamal, G. M., 1985. Bangladesh Contraceptive Prevalence Survey-1983: Final report. Dhaka: Mitra and Associates.

Rodriguez, G., 1979. "Family planning availability and contraceptive practice", Family Planning Perspectives, Vol. 11, No. 1, pp. 51-56, 58-63, 67-70.

Sarkar, K. S. and Mian, M. A. B., 2004. Demand for a child in Bangladesh: A Multivariate Statistical Analysis. International Journal of Statistical Sciences, Vol. 3, pp. 103-118.

Sather, Z., Nigel, C., Callum, C., and Kazi, S., 1988. Women's Status and Fertility Change in Pakistan, Population and Development Review, Vol. 14, No. 3, 415-432.

United Nations, 1985. Women's Employment and Fertility: A Comparative Analysis of World Fertility Survey Results for 38 Developing Countries. Population Studies, No. 96, New York.

United Nations, 2003. Economic and Social Commission for Asia and the Pacific (ESCAP) 2003, Population Data Sheet, Bangkok, Thailand.

USAID (U.S. Agency for International Development), 1991. Briefing Package on the Family Planning and Health Services Projects' NGO Activities, (Dhaka, Bangladesh).