Content uploaded by Russell Kabir
Author content
All content in this area was uploaded by Russell Kabir on Oct 17, 2018
Content may be subject to copyright.
What Went Wrong with the Achievement of Replacement Fertility
in Bangladesh and Its Consequences on the Demographic
Dividend: The Role of Proximate Determinants
Ahbab Mohammad Fazle Rabbi*, Mohammed Kabir
*
*, Russell Kabir***
*
Department of Statistical Sciences, University of Padua, Italy, fazlerabbi@stat.unipd.it
**
Department of Statistics, Jahangirnagar University, Bangladesh, kabirm46@yahoo.co.uk
***
Department of Medical Science & Public Health, Anglia Ruskin University, UK,
russell.kabir@anglia.ac.uk
Abstract: According to 2010 World Population Prospects (WPP), Bangladesh is
passing the second phase of fertility transition. The recent fertility level (TFR) of
Bangladesh is 2.3 births per woman. The Bangladesh Demographic & Health
Survey (BDHS)-2014 data showed that fertility is stalled again since BDHS-2011
unexpectedly. This stagnation raises questions about the prospect of reaching
replacement fertility which was supposed to be achieved in Bangladesh by 2015
but failed again. This also has implications on the demographic window and
consequently on the demographic dividend. Using the data of BDHS-2014 and
applying the Bongaarts framework of the proximate determinants of fertility, this
study attempts to identify the factors responsible for not achieving replacement
fertility yet. The results demonstrate that contraception still dominates the fertility
reduction in Bangladesh, followed by lactational infecundability, marriage and
induced abortion. The change in the level of the proximate determinants and
other key factors showed that fertility did not change much since BDHS-2011
which possibly caused this stagnation. The important factor which may have
contributed to this is the high proportion of adolescent marriage in Bangladesh.
This factor still creates a higher value of index of marriage which in turn affects
fertility rates. Simulation on proportion married at adolescent age group suggests
policy implications for achieving replacement fertility in Bangladesh can be
achieved. In addition, the role of abortion and its measurement problem are also
discussed
Keywords: Bangladesh, proximate determinants of fertility, nuptiality, fertility
stagnation
https://doi.org/10.24193/RJPS.2018.1.06
© Centre for Population Studies
104 • Romanian Journal of Population Studies • Vol. XII, No. 1
1. Introduction
Bangladesh is among the most densely populated countries in the world.
Throughout the past century, the population of Bangladesh has increased
exponentially. Between 2001 and 2011, 2 million individuals increased
population figures yearly. According to 2010 World Population Prospects
(WPP), fertility transition of a country is modelled in three Phases and they are-
(I) a high fertility, pre-transition phase (II) a fertility, transition phase and (III)
a low fertility, post transition phase. These phases are based on level of Total
Fertility Rate (TFR) of a country. The first phase is labelled as high fertility l;
on the other hand, a sharp fall from TFR of 7.0 to below replacement level
(TFR=2.1) is defined as fertility transition. Bangladesh is now in the third
phase of demographic transition. Once a country had reached convergence
level, it would stabilize and continue at this point for certain period (UN,
Department of Economic & Social Affairs, Population Division, 2017). Thus,
fertility level is one of the core components of population dynamics that
determine the size, structure, and composition of the population in any
country. Graphically, the fertility transition is described in WPP 2010 (Alkema
et al. 2011),shown below (Figure1).
Figure 1. Different stages of fertility transition
Phase I: Fertility is high and the fertility transition has not yet started
Phase II: Fertility transition
Phase III: Sub-replacement recovery
Source: Alkema et al. 2011: 819
Contemporary Population • 105
Due to fall in fertility, a country typically enjoys demographic dividend, a span
of time in which the relative number of older people rises and younger
population decreases. The results are dramatic increase in the share of working
population. The result is a dramatic increase in the share of the population and
opens the scope of the demographic window and leads to a “demographic
dividend”. It will turn into a demographic dividend only if the country invests
heavily now in health, education, skills development, and employment
generation, especially for the large number of youths.
The current total fertility rate (TFR) of Bangladesh is 2.3 births per
woman which has stalled since last BDHS (BDHS-2014). Bangladesh aimed to
achieve replacement TFR of 2.1 births per woman by 2015 through improved
access to health and nutrition services for the poor and geographically
marginalized population (BDHS-2011). The TFR declined from 6.3 births per
woman in 1971-1975 to 5.1 births per woman in 1984-1988, followed by
another rapid decline in the next decade of 1.8 births per woman to reach 3.3
births per woman in 1994-1996. TFR remained plateaued during the 1990s for
about a decade, at around 3.3 births per woman. Since 2004 TFR has again
begun to decline. The TFR declined further by one child per woman during the
current decade to reach 2.3 births per woman in 2011 which continues till
2014. The trend of TFR from 1975 to 2014 is shown in the following graph
(Figure 2).
Figure 2. Fertility trend in Bangladesh (1974-2014)
Note: The data are obtained from two World Fertility Surveys (BFS-1975&1989), five
Contraceptive prevalence surveys (during 1989-1991), six BDHSs (during 1993-2011) and
two Maternal Mortality surveys (BMMS-2001&2010).
106 • Romanian Journal of Population Studies • Vol. XII, No. 1
For the 1975 and 1989 BFS surveys, the rates refer to the 5-year period
preceding the survey; for the other surveys, the rates refer to the 3-year period
preceding the survey. The BFS and BDHS surveys utilized full birth histories,
while the 1991 CPS used an 8-year truncated birth history. Source: 1975 BFS
(MOHPC 1978: 73); 1989 BFS (Huq and Cleland 1990: 103); 1991 CPS (Mitra
et al. 1993:34); BDHS 2014 (NIPORT).
The fertility projections are done using exponential regression of TFR
against time, as Exponential model fitted the data with maximum goodness of
fit compared to other kinds of regression approach. The fitted model is TFR =
1E+26e-0.029 * Year with R2 = 0.9567.
In Figure 2, the green line represents the projected fertility level up to
2025, and red line presents the replacement level of fertility (TFR=2.1). The
projection suggests that Bangladesh was supposed to achieve replacement level
of fertility by 2015 if the trend continued. All subsequent governments that
have come into power have identified population control as the top priority for
government’s population policy planning. This political commitment played a
crucial role in the fertility decline in Bangladesh. Since 1980 the family planning
program has emphasized the importance of integrating health and family
planning services. The goal is to provide an essential integrated package of high
quality, client-centered reproductive and child health care, family planning,
communicable disease control, and curative services at a one-stop service
point.
Many developing countries in Asia and Africa experienced the second
phase of fertility transition without much socioeconomic development, on the
other hand, countries at similar levels of economic development are often seen
to show very different patterns of fertility pattern (Bongaarts and Watkins
1996). Past declines in the level of fertility across these countries were mostly
attributed to a strong family planning program, in the absence of any
remarkable change in socioeconomic status in a fundamentally traditional and
impoverished society (Cleland et al. 1994). This statement is not consistent as
many exceptions were held in 2000s. Also, the argument exists that change in
other sectors of society and the economy might also play an important role on
declining fertility (Caldwell et al. 1999). Few contradictory scenarios may be
seen in Bangladesh, too. Though increase is observed in the level of using
family planning method in the recent DHSs, still the relation between fertility
level and contraceptive use prevalence are not well synchronized. The
contraceptive prevalence rate (CPR) of 62 percent reflects a small rise from
previous BDHS (BDHS-2014).
Contemporary Population • 107
As the relationship between fertility and its determinants is very complex and
studying the determinants of fertility at aggregate level is not an easy task,
because human reproduction is an outcome of both biological and behavioural
factors, along with cultural and socioeconomic factors (Bongaarts and Potter
1983). The principal characteristic of biological and behavioural factors is that
they can influence fertility directly, while socioeconomic and environmental
factors affect fertility through modification of one or more biological or
behavioural factors (Bongaarts 1978). Using data from 41 developed and
developing countries, Bongaarts and Potter (1983) further observed that 96
percent of the variance in the total fertility rates of these populations could be
explained by four principal proximate determinants: specifically, marriage,
contraception, locational infecundability, and induced abortion. Because of
these findings, the analysis of the determinants of fertility becomes more
simplified since contribution of each factor can be estimated and policy
strategy can be considered accordingly.
As these proximate determinants are well recognized for explaining
aggregate fertility level, many researches exist till now for Bangladesh along
with other developing and developed countries (Rabbi 2015; Islam et al. 2011;
Mahjabeen and Khan 2011; Erfani and McQuillan 2008; Kabir and Chowdhury
2004; Islam et al. 1998, Islam and Islam 1993; Kabir and Uddin 1987; Wang et
al. 1987). Most of these studies are done to explain current fertility level of
specific country or countries, though many of them estimated the required
level of contraceptive prevalence rate to achieve replacement level of fertility
(Mahjabeen and Khan 2011). However, these approaches overlooked policy
options regarding other proximate determinants which have perceived
influence on fertility decline. Furthermore, with an inconsistent relationship
between aggregate fertility levels and contraceptive prevalence rate, further in-
depth analysis on other proximate determinants is crucial for Bangladesh to
ensure that Bangladesh is on right track to achieve the replacement level of
fertility.
This unexpected stagnation near the replacement level of fertility draws
attention to in-depth analysis to understand the phenomena of past decline as
well as quality of birth history data. Since 2010, poverty declined considerably
(BBS 2010) and per capita income along with females’ enrolment at all level
increased considerably. Female’s participation in the labour market has also
increased to a noteworthy degree in formal and informal sectors (BBS, 2010).
Hence, to improve our understanding of the causes of fertility decline in
Bangladesh, it is necessary to analyse how proximate determinants influence
fertility. Keeping this in view, this paper investigates the levels and trends of
108 • Romanian Journal of Population Studies • Vol. XII, No. 1
the proximate determinants along with their key fertility intention behaviours
in Bangladesh. This study provides a critical review of the major proximate
determinants of fertility, and estimates their fertility-inhibiting effects using the
Bongaarts (1978) model. Based on the above argument, the main objective of
this paper is to assess why fertility remained stagnant from 2011 to 2014 again
in Bangladesh.. We attempted to explore the reason behind this stagnation of
fertility and as a result of which Bangladesh has failed to achieve replacement
fertility in BDHS-2014. The purpose is also to identify which are the factors
that policy makers should target for immediate intervention to reach the
demographic goal. We also raise question regarding the possible effects of
reducing adolescent marriages are reduced, and induced abortion data are used
indirectly for determining future fertility change. The paper also investigates
the implications on the demographic dividend.
2. Methodology
The data utilized for this research is a secondary data extracted from the
Bangladesh Demographic and Health Survey conducted in 2014 under the
authority of the National Institute for population Research and Training
(NIPORT) of the Ministry of Health and Family Welfare and funded by
USAID. The 2014 Bangladesh Demographic and Health Survey (BDHS) is a
nationally representative sample survey designed to provide information on
basic national indicators like fertility, childhood mortality, contraceptive
knowledge and use, maternal and child health, nutritional status of mothers
and children, and so on. BDHS-2014 contains 17,863 ever-married women of
child bearing age of all the regions of Bangladesh.
To measure the fertility inhibiting effects of the four principal
proximate determinants of fertility in a given population, the aggregate fertility
model of Bongaarts (1978) and Bongaarts and Potter (1983) has been used in
the current study. This model is an aggregate model and it assumes that the
natural reproductive capacity, i.e. total fecundity rate (TF) of women is nearly
the same for all women, but their actual reproductive performance is modified
by four major proximate determinants. A population’s actual level of fertility is
measured by the total fertility rate (TFR), while in the absence of inhibiting
effects of the proximate determinants, the fertility level of the population could
reach a hypothetical maximum level, called total fecundity rate (TF). The
observed level of fertility in a given population reflects the extent to which the
proximate determinants reduce the TF. TF for Bangladesh is taken as 15.3 like
the previous studies (Islam et al 1998, Islam and Islam 1993). The fertility-
inhibiting effects of the four principal proximate determinants are proportion
Contemporary Population • 109
married, contraception, induced abortion, and postpartum infecundability and
they are measured in the model by four indices: Cm= index of marriage, Cc=
index of contraception, Ca= index of induced abortion, and Ci= index of post-
partum infecundability. The value of each index lies between 0 and 1; 0
signifying complete fertility inhibition and 1 meaning no fertility inhibition.
Symbolically, the relationship between the actual level of fertility in a
population, as measured by total fertility rate (TFR) and the biological
maximum TF is,
TFR = Cm × Cc × Ca × Ci × TF (1)
The complement of the value of an index is the proportionate reduction in
fertility due to the inhibiting effect of that proximate variable.
2.1 Estimation of the model Indices
2.1.1. Estimation of Index of Marriage, Cm
The index of marriage is determined by the age-specific proportions of
currently married among females. Cm is estimated as the weighted average of
the age-specific proportions of females currently married m(a), with weights
provided by the age-specific marital fertility rates g(a). Symbolically,
)(
)()(
ag
agam
TMFR
TFR
C
m
(2)
2.1.2. Estimation of Index of Contraception, Cc
The effect of contraception on the risk of conception is measured by the index
Cc. For current contraceptive prevalence rate u and average use effectiveness e,
the value of the index is calculated as
Cc = 1 − 1.08 × u × e (3)
Where, u is the current contraceptive prevalence rate (CPR) and e is the average
use-effectiveness of contraception, which is calculated as the weighted average
of the following method-specific use-effectiveness levels (Bongaarts 1982).
Contraception method umUse-effectiveness* em
Pill 0.90
Condom 0.62
IUD & implants 0.95
Injection 0.99
Sterilization 1.00
Others 0.70
* Source: Bongaarts 1982
110 • Romanian Journal of Population Studies • Vol. XII, No. 1
This implies,
m
uu
and
u
eu
e
mm
The value 1.08 is the adjustment factor for sterilizing, on an assumption that all
the contraceptive users may not be fecund at the time of using contraceptives
and a small proportion of sterile women may use contraceptive without
knowing their fecundity (Bongaarts and Potter 1983).
2.1.3. Estimation of Index of lactational infecundability Ci
The index of postpartum infecundability Ci measures the effects of postpartum
amenorrhoea and lactation (breastfeeding) on fertility. In the presence of
breastfeeding and postpartum abstinence the average birth interval equals
approximately 18.5 months plus the duration of postpartum infecundability
(Bongaarts and Potter 1983). Thus, Ci is estimated as,
i
C
i
5.18
20
(4)
Where, i is the average duration of post-partum infecundability. In current
study, i is considered as median. For average (median) duration of
breastfeeding B, i can be estimated from the following fitted model of
Bongaarts and Potter (1983),
;753.1
2
001872.01396.0
BB
ei
Where, R2 = 0.96 (5)
2.1.4. Estimation of index of abortion, Ca
Abortion is not well reported in Bangladesh, so the index of abortion does not
provide good results. The index of abortion Ca is estimated as,
TAuTFR
TFR
C
a
)1(4.0
(6)
Where TA is the total abortion rate, estimated as the number of abortions in
survey preceding three years divided by the number. of currently married
women at that time. A termination of pregnancy after 8 weeks is considered as
an abortion in current study (Johnston and Hill 1996).
2.1.5. Stover’s Review: inclusion of Index of Sterility, Cp
This index is recommended later by Stover (1998) in his review of proximate
determinants. This index is not suggested by the aggregate model of Bongaarts
and Potter (1983), though the effect of sterility on fertility was discussed
(Bongaarts and Potter 1983). The index of pathological sterility is intended to
estimate the fertility-inhibiting effects of primary and secondary sterility. Since
Contemporary Population • 111
data on sterility were scarce at the time, Bongaarts later developed an equation
to estimate the index as a function of primary sterility (Bongaarts 1984). The
index is:
3.7
11.063.7
s
C
p
(7)
Where, s is the percentage of women aged 45-49 who have had no live births.
This index is equal to 1.0 when 3 percent of women are childless at age 45-49.
Anything above this level is assumed to be the effect of pathological sterility.
The reason is, about 3 percent couples are sterile from the beginning of the
reproductive period and consequently remain childless (Bongaarts and Potter
1983, Conception 1981). For Bangladesh, this index is omitted from analysis,
since only 1.4 percent of women aged 45-49 were childless at survey preceding
three years of BDHS-2014.
2.2. Fertility-inhibiting effect
The difference between the total fecundity (TF, taken as 15.3) and the
predicted or model-estimated TFR demonstrate the resultant inhibitory effect
of each determinant while, the fertility controlling effect is prorated by the
product of difference between TF and model TFR to the proportion of the
logarithm of each index to the sum of the logarithms of all indices (Wang et al.
1987). For example, the fertility inhibiting effect of marriage can be expressed
symbolically as,
aicm
m
CCCC
C
estimatedTFRTF
loglogloglog
log
)]([
3. Results
3.1. Role of major proximate determinants in fertility decline
The summary measures required for the application of Bongaarts (1978) model
and corresponding reproductive indicators for Bangladesh (BDHS-2014) are
presented in the following table (Table 1). Multiplying all the indices together
by the total fecundity rate of 15.3 produces the predicted TFR for the
population. The predicted TFR typically differs from the observed TFR
because of underreporting of births, measurement errors of the proximate
determinants, or the omission of any other potential proximate determinants
that are influential in determining fertility levels in that population under study
(Islam et al. 2011).
Clearly, the lowest impact of marriage may be seen on recent fertility
decline in Bangladesh; having a value of 0.888, Cm may reduce only 11 percent
112 • Romanian Journal of Population Studies • Vol. XII, No. 1
of the fertility in Bangladesh. The impact of family planning on current fertility
is still a dominant factor for Bangladesh, as 59 percent of the fertility decline is
attributed to use of contraception. Besides contraception, the highest fertility
decrease is occasioned by post-partum infecundability, which reduces almost
51 percent of fertility for that index. The lowest effect of abortion is seen in
fertility level; around 1 percent of fertility is reduced by abortion.
Table 1. Reproductive indicators and derived indices of proximate determinants of fertility for
Bangladesh (BDHS-2014)
A. General Reproductive Indicators
TFR 2.3
TMFR 2.59
Median age at first marriage (25-49) 15.8 years
CPR (u) 62.4 percent
Contraceptive use effectiveness (e) 0.87
Median duration of breastfeeding 31 months
Median duration of postpartum
infecundability
21.98 months
Total Abortion rate (TA) 0.043
B. Model indices
Cm0.888
Cc0.41
Ca0.988
Ci0.494
Combined effect of four determinants
(Cm× Cc × Ca × Ci)
0.1777
Total fecundity (TF) 15.3
Predicted TFR 2.72
Table 2 exhibits the magnitude of the total inhibiting effect being accounted
for by each proximate determinant at BDHS-2014. For Bangladesh, out of
12.58 (=15.3—2.72) births being inhibited, 1.01 births (or 8 percent) were due
to the marriage variable, 6.33 births (or 50 percent) were due to contraception,
0.07 births (or 0.5 percent) were due to abortion and 5.16 births (or 41 percent
of total inhibiting effects) were because of post-partum infecundability.
The trends of four indices are summarized in the following table (Table
3). The estimated indices for BDHS 1993-94, 1996-97, 1999-2000, 2004 and
2007 are taken from the previous research (Islam et al. 2002; Mahjabeen and
Khan 2011; Rabbi 2015). The differences between original and predicted TFR
Contemporary Population • 113
were high in earlier BDHSs, the lowest gap being observed for BDHS-2014.
Cm is unusually high in Bangladesh; lowest fertility decline occurs by marriage
over the time. On the other hand, values of Cc decreased sharply in all the
BDHSs, which indicate the increasing trend of the use of family planning
methods in Bangladesh. Generally the longer duration of breast feeding is also
common in Bangladesh which is supported by Ci index.
Table 2. Magnitude of the total fertility-inhibiting effect being accounted for each proximate
fertility determinants for Bangladesh (BDHS-2014)
Proximate
determinants
Value of Index Fertility-inhibiting effect
Reduction of births
per woman
Percentage
reduction
Cm0.888 0.86 6.83
Cc0.41 6.49 51.58
Ca0.988 0.09 0.72
Ci0.494 5.13 40.77
Total [TF-TFR(est)] 12.58 12.57 100.0
Table 3. Trends of proximate determinants of fertility in Bangladesh (1993-2014)
BDHS TFR
C
m
C
c
C
i
C
a
Predicted
TFR
TFR-TFR
(est)
1993-94 3.4 0.878 0.575 0.660 1.0 5.10 1.70
1996-97 3.3 0.858 0.531 0.680 1.0 4.74 1.44
1999-2000 3.3 0.843 0.495 0.714 1.0 4.56 1.26
2004 3.0 0.743 0.454 0.813 0.963 4.04 1.04
2007 2.7 0.750 0.489 0.823 0.959 4.43 1.73
2011 2.3 0.871 0.42 0.493 0.991 2.73 0.43
2014 203 0.888 0.41 0.494 0.988 2.72 0.42
Note: Values of proximate determinants for first three BDHSs are taken from Islam et al
(2002); BDHS-2004 and 2007 are taken from Mahjabeen and Khan (2011) and BDHS-
2011 are taken from Rabbi (2015). For BDHS-2004 and 2007 an index of abortion is
constructed using data on abortion from the Matlab study (ICDDR,B 1996), which is used
as a proxy (Mahjabeen and Khan 2011).
To determine the reason of this stagnation in fertility level, we compared all the
fertility indicators with that of BDHS-2011 as well. The summary measures of
Bongaarts (1978) model and corresponding reproductive indicators for
114 • Romanian Journal of Population Studies • Vol. XII, No. 1
Bangladesh during BDHS 2011 and BDHS 2014 are summarized in the
following table (Table 4).
As evident from the various proximate determinants, the change in the
level of the reproductive indicators did not occur much in the BDHS-2014
compared to the BDHS-2011 which may be the reason for not achieving
replacement level of fertility. We explored each of these indicators with
associated formal proximate determinants in the following sections, and by
creating simulation we tried to show what should be prioritized in future
Table 4. Comparison of reproductive indicators and derived indices of proximate
determinants of fertility for Bangladesh (BDHS 2014 and BDHS 2011)
General Reproductive Indicators BDHS-2014 BDHS-2011
TFR 2.3 2.3
TMFR 2.59 2.64
Median age at first marriage (25-49) 15.8 years 15.5 years
CPR (u) 62.4 percent 61.2 percent
Contraceptive use effectiveness (e) 0.87 0.88
Median duration of breastfeeding 31 months 31.2 months
Median duration of Post-partum infecundability 21.98 months 22.08 months
Total Abortion rate (TA) 0.043 0.028
B. Model indices
Cm0.888 0.871
Cc0.41 0.42
Ca0.988 0.991
Ci0.494 0.493
Combined effect of four determinants
(Cm× Cc × Ca × Ci)
0.1777 0.1787
Total fecundity (TF) 15.3 15.3
Predicted TFR 2.72 2.73
3.2. Marriage
It has already been noted that impact of marriage on fertility decline is not
significant which is due to the high prevalence of adolescent marriages. The
estimated value of the index suggests that there is a little gap for younger ages
and consequently the difference between TFR and TMFR (marital fertility rate)
is also low. This high value of index of marriage occurs due to higher
proportion of adolescents’ marriage in Bangladesh. The age specific fertility
rate (ASFR), age specific marital fertility rate (ASMFR) and the proportion
Contemporary Population • 115
married for BDHS 2014 in all age groups are presented in the following table
(Table 5).
Table 5. ASFR, ASMFR and proportion married for Bangladeshi women (BDHS-2014)
Age group Proportion
married*
ASFR ASMFR Median age
at marriage**
15-19 0.442 113 0.442 -
20-24 0.830 143 0.830 17.2
25-29 0.920 110 0.920 16.4
30-34 0.946 57 0.946 16.0
35-39 0.923 24 0.923 15.6
40-44 0.890 4 0.890 15.3
45-49 0.855 5 0.855 15.3
Total 0.798 TFR = 2.3 TMFR = 2.59 15.5
Notes: *Bongaarts multiplier is applied. ASMFR for 15-19 is computed as 0.75 × ASMFR
for women 20-24 (Bongaarts and Potter 1983).
** The age at first marriage is defined as the age at which the respondent began living with
her first spouse/partner. Median age at marriage for women aged 25-49 is 15.5 years. For
women aged 20-49, the median age at marriage is 15.8 years. Median age at marriage is not
applicable for age group 15-19 due to censoring.
Age at first marriage has a major effect on childbearing because the risk of
pregnancy depends primarily on the age at which women first marry (Islam et
al 1998). Women who marry early, on average, are more likely to have their
first child at a young age and give births to more children overall, contributing
to higher fertility (Islam and Islam 1993). For BDHS 2014, 80 percent of
women aged 15-49 were married at the time of survey. Among them, about 44
percent of women aged 15-19 were married, which affects age specific marital
fertility rate at adolescent ages. A few important indicators of marriage in
Bangladesh are presented in the following Table (Table 6). Compared to early
BDHSs, median age at first marriage increased slightly in Bangladesh in the last
twenty years with highest value at BDHS 2014. The rate did not increase with a
secular trend, as a decrease may be observed in 1996-97 and 2004. It should be
noted that, the lowest median age at first marriage is observed in Bangladesh
compared to the neighbouring South Asian countries during BDHS 2011.
Age specific fertility rates for women aged 15-19 have a sharp decline
since BDHS-1996-97. After an increase in BDHS-1996-97, it falls steadily to at
113 at BDHS-2014. Despite this the contribution of adolescent age specific
116 • Romanian Journal of Population Studies • Vol. XII, No. 1
fertility to the total fertility is about 25 percent. To achieve replacement level of
fertility, adolescent marriage should get priority in the policy if we want to
achieve our demographic objective.
Table 6. Trends of few marriage indicators of Bangladesh (BDHS-1993-94 to 2014)
BDHS Proportion
married (all)
Proportion
Married (15-
19)
ASFR
(15-19)
Median age at
first marriage
(20-49)
1993-94 79.4 47.7 140 14.4
1996-97 77.6 48.3 147 14.2
1999-2000 76.2 46.6 144 15.0
2004 77.1 46.0 137 14.8
2007 78.0 45.6 126 15.3
2011 80.0 44.7 118 15.8
2014 79.8 44.2 113 15.8
3.3. Contraception
BDHS surveys defined current use of contraception as the proportion of
currently married women who report that they are using a family planning
method at the time of the survey (BDHS 2014). The contraceptive prevalence
rate is high in Bangladesh (BDHS 2014). With 62.4 percent contraceptive
prevalence rate (CPR), more than half of the currently married women aged
15-49 use a modern method of contraception (52 percent). Use effectiveness
of 87 percentages is seen for contraceptives in the current study (Table 1),
which is high enough for a developing country (Simmons 1985). Also the value
of Cc is 0.41 along with reduction of 6.49 births per woman (Table 2).
Increasing the family planning method more effectively will reduced the
fertility level of Bangladesh, as has also been suggested in previous studies
(Islam et al. 1998). The trends of using modern family planning methods for
the seven BDHSs are summarized in the following table (Table 7).
CPR of 62 percent reflects a slight increase in FP utilization compared
to previous BDHS. Between 1993 and 2011 the use of female sterilization
among currently married women declined from 8.2 to 4.6 percent. At the same
time, two methods gained popularity; the pill is being used by 27 percent of
women (BDHS 2011), compared to 17 percent at 1993 and injectables (4.6
percent in 1993 to 11 percent in 2011).
Contemporary Population • 117
Table 7. Trends of modern family planning methods in Bangladesh (1993-2014)
BDHS Any
modern
method
Pill IUD Inject
ions
Con-
dom
Female
sterilia-
tion
Male
steriliza
-tion
Impl
ants
CPR
1993-94 36.6 17.5 2.2 4.6 3 8.2 1.1 - 44.9
1996-97 42.1 21.1 1.8 6.3 3.9 7.7 1.1 0.1 49.8
1999-00 44 23.3 1.3 7.3 4.3 6.8 0.5 0.5 54.3
2004 47.6 26.4 0.6 9.8 4.2 5.3 0.6 0.8 58.5
2007 47.5 28.5 0.9 7 4.5 5 0.7 0.7 55.8
2011 52.1 27.2 0.7 11.2 5.5 5 1.2 1.1 61.2
2014 54.1 27.0 0.6 12.4 6.4 4.6 1.2 1.7 62.4
3.4. Lactational infecundability
5.13 births were being inhibited (or almost 41 percent of total inhibiting
effects) due to the effect of post-partum infecundability (Table 2). Information
on breastfeeding in the BDHS 2014 was collected on all children born during
the last three years preceding the survey date. This data includes children living
and deceased at the time of the survey. The median duration of any
breastfeeding among Bangladeshi children in BDHS-2014 is 31 months which
consequently implies almost 22 months of median duration at post-partum
infecundability. The median duration of exclusive breastfeeding is estimated at
2.8 months in BDHS 2014. The median duration of exclusive breastfeeding has
decreased since 2011 (BDHS 2014). Generally longer breastfeeding is common
in Bangladesh. The median duration of breastfeeding in Bangladesh was
extraordinarily long during BDHS-1993-94. It was so long, in fact, that it was
not possible to be calculated exactly from BDHS 1993-94 data. This is because
breastfeeding status was asked only for children age 35 months or less and 60
percent of the children ages 34-35 months were still being breastfed.
3.5. Induced abortion
Abortion is illegal and strictly prohibited by law in Bangladesh, unless
otherwise recommended by registered doctors (Islam et al. 1998). Fertility
reduced by abortion is very low Bangladesh (approximately 0.1 percent),
possibly due to misreported abortion rates. In earlier research, the index of
abortion is assumed to be 1.0 due to the very low number of reported
abortions, though several studies on abortion suggest that it is not rare in
Bangladesh (ICDDR,B 1996). The gap between observed and predicted TFR
may also arise due to abortion, as further analysis on abortion rates may
conclude more precisely (Johnston and Hill 1996). In most cases, abortion is
done under the name of menstrual regulation, a procedure which is approved
118 • Romanian Journal of Population Studies • Vol. XII, No. 1
by the government's health and family planning program (Islam et al. 1998).
Distribution of terminated pregnancies during the last three years preceding
BDHS 2014 is summarized in the following table (Table 8).
Table 8. Distribution of terminated pregnancies (both MR and Abortions) in Bangladesh
(BDHS 2011)
Month pregnancy
terminated
No. of observations Percentage
153 6.1
286 10.0
375 8.7
463 7.2
564 7.4
688 10.1
776 8.8
874 8.6
9287 33.1
Total 866 100.0
During BDHS 2014, a total of 866 terminated pregnancies occurred during
three years preceding survey date. Among them 727 terminations (84 percent)
occurred after 8 weeks of conception, which gave a total abortion rate of 0.043.
Ca is 0.988, which means approximately 1 percent fertility may be reduced by
induced abortions. Indirect techniques are suggested in many studies to
estimate abortion rate for countries where abortions are misreported; i.e. to
estimate Ca from the values of other proximate determinants of fertility
(Johnston and Hill 1996). If we consider Ca index is 0. 95 then reduction of
fertility contributed by the induced abortion will be higher than the current
estimate suggests. Due to social, cultural and religious stigma, induced abortion
is heavily underestimated.
3.6. Role of proximate determinants to achieve replacement level fertility
The sudden stagnation at 2.3 births in BDHS 2014 is supposed to be
unexpected as the fertility was falling exponentially during the previous surveys
(Figure 2). From the findings of the current study, it is clear that, to achieve
replacement level of fertility, further decline is required for proximate
determinants. The required level of index of contraception and lactational
infecundability are estimated in the following sections along with simulation
for marriage. We are not considering the induced abortion because the
Contemporary Population • 119
abortion is severely misreported in BDHSs data, which seriously affect the
abortion rates used in Bongaarts framework
3.7. Contraception
For reaching replacement level of fertility without any change in other
proximate determinants rather than family planning, the required level of
contraceptive prevalence rate (CPR) is obtained using equation (3). Suppose
that TFR1 and TFR2 be the observed total fertility rate and 2.1 (TFR for
replacement level fertility) respectively and let the corresponding levels of
contraceptive prevalence and use-effectiveness be u1and e1, and u2 and e2
respectively. If we assume that the indices for all other proximate determinants
except for contraception remain constant, i.e., Cm1 = Cm2, Ca1 = Ca2, Ci1 = Ci2,
TF1= TF2 then obviously
11
22
1
2
1
2
08.11
08.11
eu
eu
C
C
TFR
TFR
c
c
Thus,
1
1
2
2
2
1
)08.1(
1
c
C
TFR
TFR
e
u
For BDHS 2014, Cc was 0.41 and TFR was 2.3. Assuming use-effectiveness is
unchanged (e1 = e2
=0.873), we have u2 as 0.6636. Therefore, 66.36 percent
CPR is required to achieve replacement level of fertility.
3.8. Lactational infecundability
The required duration of post-partum infecundability may be estimated using
equation (4). Continuing with symbols of contraception and assuming that the
indices for all other proximate determinants except for breastfeeding remain
constant, i.e., Cm1 = Cm2, Ca1 = Ca2, Cc1 = Cc2, TF1= TF2 then obviously
2
2
1
2
1
2
5.18
20
5.18
20
i
i
C
C
TFR
TFR
i
i
Here, i1and i2 are corresponding duration of post-partum infecundability for
TFR1 and TFR2. Then the required duration of post-partum infecundability to
achieve replacement level of fertility will be,
120 • Romanian Journal of Population Studies • Vol. XII, No. 1
5.18
20
1
1
2
2
i
C
TFR
TFR
i
For BDHS-2014, estimated Ci1 was 0.494, which implies that a median
duration of post-partum infecundability of 25.84 months is required to achieve
replacement level of fertility.
3.9. Marriage
It was already stated that, among all the indices, reduction in fertility
contributed to by the proportion of married is at a minimum, which is due to
the high prevalence of adolescent marriage in Bangladesh. In Bangladesh still
one fifth of the total fertility is attributed to the adolescents despite the fact
legal age at marriage is 18 years. Unlike previous two indicators, the required
level of marriage can’t be determined from estimated Cm. It will also not imply
anything for the policy-makers. A simulation is performed to assess what
would be necessary to reach the desired goal of replacement fertility, and the
following Table shows that if the proportion of married individuals aged 15 to
19 is brought down from the current 44% to 36% Bangladesh can easily
achieve replacement level of fertility.
Suppose with 0.442 proportion married at age group 15-19, we have
ASMFR of 129 while ASFR was 113 for age 15-19. Here ASMFR was
estimated using Bongaarts multiplier on ASMFR of the next age group
(Bongaarts and Potter 1983). Due to the short average of marital duration, the
marital fertility rates for women aged 15-19 years, therefore, do not represent
the potential fertility of the whole age group. In this case Bongaarts
recommends that the marital fertility rate for women aged 15-19 be taken as
0.75 of the rate for women aged 20-24 (Bongaarts and Potter 1983). In this
simulation, Bongaarts multiplier is omitted to check the basic contribution of
adolescents on fertility. Instead of Bongaarts multiplier, the adjusted ASMFR,
TMFR, Cm and model TFR are 255, 3.25, 0.714 and 2.18 respectfully. The
results of the simulated ASMFR, TMFR, Cm and model TFR for various level
of proportion married at age 15-19 are summarized in the following table
(Table 9).
The simulation shows that the aggregate fertility level will decline along
with the decrease in the level of adolescent marriages. The fall in ASMFR,
TMFR, Cm and model TFR almost followed a linear trend with the fall in
Contemporary Population • 121
proportion of individuals married at age 15-19. These findings may help the
policy makers to take decisions regarding shifts in the age at marriage of
women precisely. A decline in the proportion married from 0.45 to 0.40 will
help Bangladesh to achieve replacement level of fertility, while a decline to 0.3
will help Bangladesh to gain third phase of fertility transition (WPP 2010). On
the other hand, if we consider the abortion index from 0.99 to 0.95, then
replacement fertility would have been much earlier than expected.
Table 9. Simulation for proportion married at age 15-19 and its impact on fertility of
Bangladesh (BDHS 2014)
Proportion
married at 15-
19
ASMFR TMFR
C
m
Model
TFR
0.44 255 3.22 0.714 2.18
0.43 263 3.26 0.706 2.16
0.42 269 3.29 0.699 2.14
0.41 276 3.32 0.692 2.12
0.40 283 3.35 0.685 2.09
0.39 290 3.39 0.678 2.07
0.38 297 3.43 0.670 2.05
0.37 305 3.47 0.662 2.03
0.36 314 3.51 0.654 2.00
0.35 323 3.56 0.646 1.98
0.34 332 3.61 0.637 1.95
0.33 342 3.66 0.629 1.93
0.32 353 3.71 0.619 1.89
0.31 364 3.77 0.610 1.86
0.30 376 3.82 0.601 1.84
4. Discussion and conclusion
The purpose of this paper is to assess the current fertility level through use of
proximate determinants, but the analysis of proximate determinants shows a
discouraging level of TFR. This is puzzling in view of the fact that during the
last decade Bangladesh has made significant progress in all the MDGs
indicators with the exception of TFR. Even poverty level has declined from
30.5% to 24.7% present in 2015. Per capita income also increased. Female
enrolment in educational institutions from primary to tertiary level also
increased significantly. At secondary and higher secondary level, there is a
gender equity. At the tertiary level about 40% of the enrolled students are
122 • Romanian Journal of Population Studies • Vol. XII, No. 1
female students. There are about 90 universities in the private sector while in
government sector the number of universities is 37. About 40% of the female
labour force is in the labour market. Per capita income along with infant, under
five and maternal mortality also declined significantly. Despite all these positive
factors, TFR in Bangladesh has stalled again. A close analysis shows that the
proportion of individuals married at adolescent ages and abortion are the two
noisy factors, which may be responsible for this.
The application of the Bongaarts model (1978) suggests that most of
the fertility declines in recent era are attributable to family planning and
lactational infecundability. The present analysis suggests that though the
fertility transition of Bangladesh sharply declined up to period 1993-94,
thereafter TFR did not change as much as expected trends had forecast. Prior
to 2007 contraception played its role adequately as major fertility inhibiting
factor. According to our analysis lactational infecundity has been placed in the
second highest fertility inhibiting factor. The fertility-inhibiting effect of
postpartum infecundability as found in Bangladesh is similar to the pattern
found in most traditional societies (Erfani and McQuillan 2008). Lengthy
breastfeeding is common in Bangladesh along with presence of exclusive
breastfeeding. Both culture and the current health policy in Bangladesh favour
universal and prolonged lactation, which improve both the health condition of
children and the widening of birth intervals (Islam et al 2011). Modernization
and increased use of contraception might be related to the declining trends in
breastfeeding and post-partum amenorrhea; previous studies (Salway et.al.
1993) identified increased use of contraception as the most important
determinant for the declining trend in post-partum amenorrhea.
Due to legal and social constraints, national level data on induced
abortion are not available and its effects remain almost unknown.
Nevertheless, verification from hospital and clinic records and other sources
suggests that induced abortion is not rare, even though it is done under the
name of menstrual regulation in Bangladesh (Islam et al 1998).
However, the impact of nuptiality patterns on the aggregate fertility
level is not the expected one. While a slight increase has been observed in age
at first marriage during the last two decades, the proportion married at age 15-
19 is still alarming. There has been little decline in the proportion of individuals
married at age 15-19. Child marriage is still prevalent. The BDHSs show about
two thirds got married before reaching the age of 16 years (which is the legal
age at marriage). Another important finding of the study is the change in age-
specific fertility patterns, which indicate that childbearing is taking place at an
earlier age that had been found in previous studies. As the desired level of
Contemporary Population • 123
fertility is declining and there is little change in age at marriage, it appears that
couples tend to reach their desired number of children in quick succession
immediately after marriage and then regulate fertility at older ages with little
effect on the replacement fertility. Compared to neighbouring South Asian
countries, the proportion of married individuals, ASFR, ASMFR are highest in
Bangladesh for age group 15-19.
The simulation of proportion of married individuals in the adolescent
age group suggests new policy implications to achieve replacement fertility.
The prevailing cultural and social norms in Bangladesh are unlikely to permit a
change in the proportion of non-married individuals beyond a certain limit and
the prospect for an immediate rise in the age at marriage for females does not
seem to be very bright (Islam et al 1998). At the same it also raises questions
regarding the effects of the change in socio-economic conditions such as
females’ enrolment in education, entry into the labour market, increased per
capita income and high, gender equity in higher secondary education and about
40% enrolment of females in tertiary education, which should have produced
gradual changes in reproductive behaviour. The fact that these effects were
not visible is puzzling. Was this an artefact of data produced by the BDHSs? It
is therefore warranted to ask what went wrong with Bangladesh, how the
problem should be tackled, and what might happen next? The demographic
projection suggests that if Bangladesh would have achieved replacement
fertility in 2015 it would have a 25 years window of demographic dividend. If
replacement fertility is achieved later, then the window would be of 20 years.
The duration of the demographic dividend depends on when we achieve
replacement fertility, and a shorter window is depriving the country from the
benefit of replacement fertility.
This study has shed further investigation on the fertility transition in
Bangladesh. Nevertheless, the extent to which Bangladeshi women are making
their own decisions, the mechanisms of decision making among the couples,
the institutional framework by which the government implemented its policies,
and the question of achieving replacement fertility in Bangladesh and benefit of
demographic dividend are among the questions remain to be explored and
policy prescriptions should be followed accordingly in the coming years.
124 • Romanian Journal of Population Studies • Vol. XII, No. 1
References
Alkema, L., Raftery, A., Gerland, P., Clark, S., Pelletier, F., Buettner, T., and
Heilig, G. (2011). “Probabilistic Projections of the Total Fertility Rate
for All Countries”. Demography 48(3): 815-839.
Bongaarts, J. (1978) “A framework for the analysis of the proximate
determinants of fertility”. Population and Development Review 4(1): 105–132.
Bongaarts, J. (1982) “The fertility-inhibiting effects of the intermediate fertility
variables”. Studies in Family Planning 13(6/7): 179-189.
Bongaarts, J., Frank, O., and Lesthaeghe, R. (1984) “The proximate
determinants of fertility in sub-Saharan Africa”. Population and
Development Review 10(3): 511-537.
Bongaarts, J., and Potter, R. G. (1983). Fertility, biology and behavior: Analysis of the
proximate determinants of fertility. New York: Academic Press.
Bongaarts, J., and Watkins, S. C. (1996). “Social interactions and contemporary
fertility transitions”. Population and Development Review 22(4): 639–682.
Caldwell, J. C., Khuda, B. E., Caldwell, B. C., I. Pieris, I., and Caldwell, P.
(1999) The Bangladesh fertility decline: An interpretation. Population and
Development Review 25(1): 67–84.
International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B)
(1996) Fertility Intention and Subsequent Abortion in Matlab, Bangladesh.
ICDDR,B, Dhaka.
Islam, M. A., Islam, M. M., and Chakraborty, N. (2002) Achieving Replacement
Level Fertility in Bangladesh: Challenges and Prospects. Paper 18, CPD-UNFPA
Programme on Population and Sustainable Development, CPD, Dhaka.
Islam, M. M., Dorvlo A. S. S., and Al-Qasmi, A. M. (2011) “Proximate
determinants of declining fertility in Oman in the 1990s”. Canadian
Studies in Population 38(3-4): 133–152.
Islam, M. M., Mamun, A. A., and Bairagi, R. (1998) “The proximate
determinants of fertility in Bangladesh: Evidence from BDHS 1993-94”.
Asia-Pacific Population Journal 13 (3): 3-22.
Islam, M. N., and Islam, M. M. (1993) “Biological and behavioral determinants
of fertility in Bangladesh: 1975-1989”. Asia-Pacific Population Journal 8(1):
3-18.
Kabir, M., and Uddin, M. M. (1987) “Effect of nuptiality, contraception and
breast-feeding on fertility in Bangladesh”. Journal of Biosocial Science 19(3):
345-350.
Contemporary Population • 125
Johnston, H. B., and Hill, K. H. (1996) “Induced Abortion in the developing
world: Indirect estimates”. International Family Planning Perspectives 22(3):
108-114.
National Institute of Population Research and Training (NIPORT), Mitra and
Associates, and ICF International (2016). Bangladesh Demographic and
Health Survey 2014. Dhaka, Bangladesh and Rockville, Maryland, USA.
National Institute of Population Research and Training (NIPORT), Mitra and
Associates, and ICF International (2013). Bangladesh Demographic and
Health Survey 2011. Dhaka, Bangladesh and Calverton, Maryland, USA.
Rabbi, A. M. F. (2015). “Do increasing legal age at marriage and increased use
of contraception will matter to achieve demographic goal: Revisiting the
role of proximate determinants in Bangladesh”. Bangladesh Journal of
Scientific Research 28(1): 27-34.
Salway, S., Roy, N. C., and Koenig, M. A. (1993) Levels and Trends in Post-partum
Amenorrhea, Breastfeeding, and Birth: Results from WFS surveys. WFS
Comparative Studies 40. Voorburg (Netherlands): ISI.
Simmons, G. B. (1985). Theories of fertility, in Fertility in Developing Countries: An
Economic Perspective on Research and Policy Issues. London: Macmillan Press.
Stover, J. (1998). “Revising the Proximate Determinants of Fertility
Framework: What Have We Learned in the Past 20 Years?”. Studies in
Family Planning 29(3): 255-267.
United Nations, Department of Economic and Social Affairs, Population
Division (2011). World Population Prospects: The 2010 Revision, Volume I:
Comprehensive Tables. New York, NY: UN.
United Nations, Department of Economic and Social Affairs, Population
Division (2017). World Population Prospects: The 2017 Revision, Methodology of
the United Nations Population Estimates and Projections, Working Paper No.
ESA/P/WP.250. New York: United Nations.
Wang, S. X., Chen, Y. D., Charles, H. C., Chen, R. W., Rochat, L.P., and Rider,
R. V. (1987). “Proximate determinants of fertility and policy
implementation in Beijing”. Studies in Family Planning 18 (1): 222–228.