1
‘
Differentials in the fertility transition in vulnerable settings
– analyses in selected Afghan areas in the 2010s.
Resumo:
Initial analyses of data from a first round of six provinces using Socio Demographic and Economic Survey (SDES) data, carried out over the first quinquenuin of the 2010s, indicate that fertility level was indeed high before entering the second century decade, with TFR frequently above 6 children per women. This paper uses SDES data from a second round of provinces carried out in 2015-2017. Addresses fertility levels and patterns applying indirect demographic methods. The main objective is to estimate the level and age pattern of fertility. In an effort to identify socioeconomic determinants of fertility, the place of residence, education and wealth are considered. Results indicate that in general, fertility transition has started in Afghanistan. Marital fertility suggests a regime near to natural fertility, where fertility control depending on the number of children the women have is inexistent. Marital fertility at ages 15-19 are at the most elevated levels, the risk of having a live child is very often above 30 percent. At the following age groups (20-30), the risk remains high; there are provinces where almost every other married woman has, annually, a live birth, letting aside induced abortions, miscarriages and stillbirths. Social determinants, apart from the urban/rural condition, suggest an ambiguous relationship with the fertility behaviour. Firstly, education does not present, in general terms, significant influence when women have no education or have entered basic education. It is only among
2 most educated women (7 or more years of schooling) that results show TFR below 5,0. Secondly, classification of fertility levels according to either household head’s education or the highest educational level attained in the household suggests that an initial improve in education is associated with no change or even increase in the TFR. It is only when the highest educational level is reached that TFR goes down. Thirdly, family wealth suggests similar association: initial improvements in material welfare associate to either no changes or a slight increase in the TFR. Deep research is needed for a better understanding of this unusual association found in a contemporaneous population like Afghanistan.
3
INTRODUCTION
Whilst the greatest fertility decline in the developing world has taken place over the past six decades, still there is an important number of countries that are lagging behind this global trend. Evidences about the current Afghan fertility agree –despite some ambiguity about the precise level in the few sources available– that the country is among those registering the highest fertility levels in the world. The evidences also point to an initial declining trend that –if it consolidates– will witness a "demographic bonus" in a few years. On the meantime, though, and according to recent UN estimates, Afghanistan bears a growth rate above 3 per cent for the period 2010-2015 (UN-DESA; 2010-2015). The rate implies that the country –with a total population around 25-30 millions and a fragile socioeconomic environment– sees the arrival of more than a million newborns citizens yearly, as result of the current high fertility levels. This is the typical scenario where fertility plays one of the most important roles, thus justifying the relevance of this study.
Objectives, data and methods
Initial analyses of data from a first round of six provinces using Socio Demographic and Economic Survey (SDES) data, carried out over the first quinquenuin of the 2010s, indicate that fertility level was indeed high before entering the second century decade, with TFR frequently above 6 children per women (SDES thematic report on Fertility and Nuptiality, 2016)1. This paper uses SDES data from a second round of provinces (Bagdhis, Baghlan, Balkh, Samangan, Nimroz, Herat and Takhar), addresses fertility levels and patterns. Relevant analysis of differentials by marital status, education, areas of residence and socio economic categories are performed. After adequate estimation of the above fertility parameters and comparing indicators of current and cumulate fertility by birth order will be possible to unveil the prospective trend implicit in those indicators.
1
For simplicity of reading, the report on Fertility and Nuptiality relative to the first round of provinces publication is referred as F&N SDES/2016.
4 This paper analyses the fertility situation applying a number of indirect techniques to data collected over the period 2015-2017, i.e., almost a quinquenium after the first round of fieldwork covering Kabul plus other five provinces2.
The main objective is to estimate the level and age pattern of fertility. Social norms in Afghanistan only accept conception within marriage; hence, marital fertility is an important component in these analyses. In an effort to identify socioeconomic determinants of fertility, the place of residence, education and wealth are considered. The study conducted here uses William Brass's P/F approach (Brass, 1985); it relies on information about total children ever born (CEB) and live births occurred the 12 months previous to the survey data.
It happens, however, that the series of Pi/Fi, which is particularly robust to reporting errors, is also particularly important for establishing trends in fertility. (Brass, 1985, page 71). The pattern of the ratios with age may reveal data errors or fertility trends (UN, 1983). However, evolution of the series by age also indicates how recent fertility correlates with past fertility, thus it gives valuable insights about fertility behaviour over a recent period.
Therefore, the exploration of the trends reflected in both fertility measures allows, firstly, evaluating data quality, to define the adequate adjustment factor to correct the fertility level and to estimate the age pattern. Secondly, as Pi and Fi compare past behaviour of different cohorts the series also allows to observe possible changes by age and so to delineate a trend.
1. Contextualizing the Afghan reproductive process – how national fertility patterns correlates with results from the seven provinces
Most recent national surveys for 2010 and 2015 present a TFR of 5,1 and 5,3 for the previous three years to the survey fieldwork (AMS/2010; AfDHS/2015). The UN-DESA estimates that fertility in Afghanistan –whose level was over 7 children per woman at the beginning of the present century– may have entered into an impressive transition toward lower levels. Forecasts produce a figure of 3.7 children per woman by the period 20202025 (UNDESA, 2015). It would be a nearly 50 percent decline -or 3.5 children per woman less - in a 20-year period implying a relatively sharp
2
Provinces anaysed in the first round were Bamiyan, Daykundi, Ghor, Kabul, Kapisa and Parwan (See F&N SDES/2016).
5 decline thus the country would not be among those with the most high fertility levels by then.
A number of social programmes pursuing improvements in the social conditions of the population in general and women in particular are on the way; more coverage of school enrolment for boys and, particularly girls, is a good example (UNESCO, 2015). Because these actions are taking place in a context where gender relations are known to be very unequal, it is reasonable to assume that both better living conditions and women empowerment may determine a faster fertility decline; this assumption is based on the Bryant study whose evidence points to the interaction between fertility decline and development.
Figures from reproductive health surveys (AMS 2010 and AfDHS 2015) shows for the more recent period, quite similar levels and patterns of fertility suggesting no changes over the period 2006/2010 and 2011/2015. Analyses using a cohort’s follow up, however, present inconsistencies that impede comparison of these two sources before precede a careful evaluation. In any case, Pi and Fi using DHS data can be compared in each period separately and compare past and recent fertility.
Figure 1 shows the Pi/Fi series built from birth histories data for both surveys. The ratio by age equals 1,0 when –in the absence of errors– the cumulate current fertility equals the parity (or mean number of children); i.e., there are no changes in fertility throughout time. In a falling fertility context, Pi/Fi is expected to increase be age replicating path of the fertility decline and meaning that past fertility is higher than current fertility.
Very often, however, errors occur and Pi is usually underreported regardless of the fieldwork instructions and supervision. Thus, in a context of high and constant fertility levels it may be found that Pi/Fi tends to decrease with age, being the main explanation, the failure to remember the actual number of children ever born in the past. To a lesser extent, this happens also when parity and current fertility are recollected in DHS’ researches type via birth histories.
P/F for 2010 and 2015 Afghan surveys, illustrated in Figure 1, indicates, on the one hand, that they differentiate in the series level. While Pi/Fi2010 oscillates near 1, 5, in 2015, values approximate to 1, 0. On the other hand, despite the difference, both generate a near stable series by age that tends to slightly increase as age increases and somewhat intensifies for 2015 data. Pi/Fi2010 suggests that current fertility may have been underreported in view of the trend presented by same cohorts at the same
6 periods using both surveys. Pi/Fi2015, indicates, firstly, that at age 20-24, both current and cumulated fertility at this young ages, coincide; note that P2/F2 is virtually 1,0; if there were no errors in the data, they would indicate no recent fertility changes at these ages. Secondly, related to the ratios at older ages, the historical experience indicates that women tend, indeed, to underreport cumulated fertility at later ages. The prevalent Afghan social context considering the women low educational level and her relegate role in the gender relationships reinforces this expectation. I.e., it is accepted that children ever born at age of mothers, say, older than 30 are underreported in the AfDHS 2015
Figure 1. Ratio Pi/Fi by age - Afghanistan, 2010 and 2015
Source: AMS/2010 (Afghan Public Health Institute/MoPH - 2011) and Central Statistics Organization (CSO), Ministry of Public Health (MoPH), and ICF. (2017) AfDHS/2015
In any case, if there were no underreporting, Pi/Fi2015 plotted in Figure 1, indicates that over the last 20 or 25 years there has been a slow but rather determined process of declining fertility in Afghanistan. If parity was, indeed, underreported –and evidence indicates that it always is–, fertility decline may be experiencing a rather faster path in the national average fertility than that indicated in Pi/Fi2015.
It is in a national fertility decline context that SDES data for the provinces of Bagdish, Baghlan, Balkh, Herat, Nimroz, Samangan and Takhar are developed.
2. RESULTS
The recent fertility trend in the seven provinces
Before estimating the fertility level for the seven SDES provinces, evaluation of the information on live births occurred over the 12 months prior to the Survey is
1,42 1,41 1,39 1,36 1,41 1,43 1,20 1,18 1,16 1,13 1,19 0,99 0,5 1,0 1,5 2,0 2,5 20- 24 25-29 30-34 35-39 40-44 45-49 Age group i Pi / Fi 2010 2015
7 necessary. The procedure is done using the Brass’ P/F technique, that aside to indicate the adjustment factor gives important insights on the fertility recent trend. In order to contextualize those adjustment factors and any trend that data may reveal, Figure 2 presents four sets of Pi/Fi series. First set corresponds the first round of provinces whose fieldwork was done around 2011/2012 (Bamiyan, Daykundi, Ghor) and 2013/2014 (Kabul, Kapisa and Parwan). The second set of provinces refers to the second round and fieldwork took place over 2015-2017; the date of reference then differs and expands almost along the 2010 decade. In addition, the provinces are classified according to the Pi/Fi trends. Those with the series decreasing by age and those with a stable or rather increasing pattern by age. Three provinces from the first round present Pi/Fi with a clear decreasing trend by age (Bamiyan, Daykundi and Ghor). A decrease trend by age indicates either that the older the women, the lower the parity than the current fertility or increasing omission in the parity (or number of CEB). According to experience in the interpretation of these series on historical evidences, provinces with this profile were not yet signing fertility decline at the beginning of the decade. Socio economic characteristics of these provinces known to be rather unfavourable plus their isolated geo-localization, endorse the hypotheses that no changes were apparent in these provinces by the beginning of current decade.
Differently, there is Kabul, the capital city and its frontier provinces Kapisa and Parwan, where Pi and Fi (1b). Although showing significantly level differences of Pi/Fi when Kabul is compared to the other two provinces, the three series do not decline in the previous set of provinces. Pi/Fi, in general terms, increases with age while in Kapisa and Parwan it is basically constant. Considering that there is always the memory effect, these three provinces, were, indeed, experiencing fertility downward changes well before 2013/2014. As it was the case in the national average in 2015. Firstly, there is the clear case of Badghis and Takhar (2017 and 2016, respectively), whose patterns, although with more moderate falling (Figure 2a), are similar to those from Bamiyan, Daykundi and Ghor. Secondly, there are Baghlan, Samangan and Nimroz (Figure 2b), with a constant pattern by age and important difference between Pi and Fi. Finally, there is Herat, similar to Kabul three years before, whose Pi/FI series increases by age, indicating that fertility was significantly higher in the past. The constant Pi/Fi values in Balk –always considering the memory factor– also strongly suggest that fertility is moderately descending.
8
Figure 2. Provinces included in the first and second SDES round (2011-2014 and (2015-2017) Pi/Fi ratios
1. Provinces from the first round (2011- 2014)
1a) Bamiyan, Daykundi and Ghor 1b) Kabul, Kapisa and Parwan
2. Provinces from the second round (2015- 2017)
2a) Badghis and Takhar 2b) Baghlan, Samangan and Nimroz 2c) Balkh and Herat
Source: Provinces from the:
First round: SDES 2011-2014, UNFPA-Afghanistan and CSO of Afghanistan (Micro data) Second round: SDES 2015-2017, UNFPA-Afghanistan and CSO of Afghanistan (Micro data)
In short, once detected that national figures point towards a moderate falling in the fertility levels, SDES results indicate that this phenomenon is present in most of the surveyed provinces. Kabul and adjacent small provinces Kapisa and Parwan probably started the declining process during the first decade of the century. There is no evidence that less developed provinces, being Ghor, an example, were following this trend. The previous report (F&N SDES/2016) arrived to the same conclusion. Recent data provided by the second round surveys clearly reinforce the conclusion that fertility has been declining in several Afghan provinces. By 2015, clearer cases are Balk and Herat, not by chance, are the provinces with the better socio-economic indicators. They also, like Kabul, are highly exposed to international migration flows, including return migration. In similar situation are also Baghlan, Samangan and Nimroz. Provinces that apparently do not signal changes in the fertility trend are
1,0 1,5 2,0
20- 24 25-29 30-34 35-39 40-44 45-49
Age group (i)
Pi / Fi Bamiyan -2011 Daykundi - 2012 Ghor - 2012 1,0 1,5 2,0 20- 24 25-29 30-34 35-39 40-44 45-49
Age group (i)
Pi / Fi Kabul - 2013 Kapisa- 2014 Parwan - 2014 1,0 1,5 2,0 20-24 25-29 30-34 35-39 40-44 45-49
Age group (i)
Pi / Fi Badghis - 2017 Takhar - 2016 1,0 1,5 2,0 20-24 25-29 30-34 35-39 40-44 45-49
Age group (i)
Pi / Fi Balkh - 2015 Herat - 2015 1,0 1,5 2,0 20-24 25-29 30-34 35-39 40-44 45-49
Age group (i)
Pi / Fi Baglhan - 2016 Samangan 2015 Nimroz 2016
9 Takhar and Badghis, having the latter, out of this set of seven provinces, the poorest socio-economic indicators, on average.
Adjustments of the number of live birth and estimates of the Total Fertility Rate
Pi/Fi series in the second round provinces present, in general, smooth trends by age, indicating in all cases, important differences between declared current and past fertility. Due to the nature of the questions, the method suggest that differences are probably due to erroneous report of the live births occurred over the twelve months previous to the survey and that a reliable adjustment factor will be given by the report of total children ever born to young women, say at ages 20-24 (P2/F2).
Previous report on the first round of provinces used as adjustment factor information from ages 20-24 and 25-29, averaging P2/F2 and P3/F3. Because there was indication of fertility decline, estimates were referred to five years before the surveys and further procedures were done to update the estimates (F&N SDES/2016).
Adjustment factors, in this case, are similar among the provinces and are, on average, smaller than those defined for the first round of surveys, where average of P2/F2 and P3/F3 resulted in values near 2,0. In this report, with data collected after 2015 and considering inputs of the AfDHS 2015, where a moderate falling fertility is detected, P2/F2 is defined as the more adequate factor to adjust number of life births and fertility measures. P2/F2 adjusts the observed ASFR in about 60 percent in five out of the seven provinces (P2/F2=1,6). Factor for the provinces of Nimroz and Samangan are 1, 8 (See Table 1).
An important technical comment about the time location -or period of reference- for estimates produced using SDES data is due here: on the one hand, trends in Pi and Fi (cumulate fertility and recent fertility) indicate that fertility levels for older women were, in general, higher than that of younger women, suggesting that fertility is currently declining. On the other hand, adjustment factors obtained by comparison of cumulated fertility from young women (parity at ages 20-24) to current fertility at the same ages, indicate that levels and patterns defined in Table 1 should refer to average fertility experience, approximately to the five-year period previous to the survey’s date. Second line in Table 1 shows the dates to which each province fertility is related to.
10 Table 1. Badghis, Baghlan, Balkh, Herat, Nimroz, Samangan, Takhar (2015-2017) Total
fertility rate (TFR), Age Specific fertility rates (‰), adjustment factor, fertility age distribution (percent) and mean age of fertility.
Age groups Badghis Baghlan Balkh Herat Nimroz Samangan Takhar Period of reference 2012-2017 2011-2016 2015-2015 2010-2015 2011-2016 2010- 2015 2011-2016 TFR 7,00 6,45 6,45 6,38 8,18 6,53 7,89 15-19 106,0 63,1 48,6 88,0 130,1 66,8 57,6 20-24 306,7 281,1 266,6 286,0 345,4 266,1 321,0 25-29 320,1 342,4 340,2 309,9 363,1 342,5 399,8 30-34 268,8 269,4 291,6 264,9 338,2 293,7 359,7 35-39 201,6 202,7 208,4 192,8 256,3 203,7 276,1 40-44 124,4 87,5 93,2 88,6 135,5 88,6 121,0 45-49 73,2 43,5 41,3 45,4 67,9 43,8 43,7 Adjustment factor* 1,6 1,6 1,6 1,8 1,8 1,6 1,6 Adjusted Number of live births 24,798 56,862 74,710 114,892 12,771 21,080 65,862 Relative contribution of selected age groups to the total fertility ( per cent)
15 – 19 7,6 4,9 3,8 6,9 7,9 5,1 3,6
20 – 34 63,9 69,2 69,7 67,5 64,0 69,1 68,4
35 or more 28,5 25,9 26,6 25,6 28,1 25,8 27,9 Total 100,0 100,0 100,0 100,0 100,0 100,0 100,0 Mean age (years) 30,4 30,2 30,6 30,0 30,3 30,3 30,7 * P2/F2 from Brass’ technique
Source: SDES- 2015-2017, UNFPA-Afghanistan and CSO of Afghanistan (Micro data).
The Total Fertility Rate and the Age Specific Fertility Rates
In general, for most provinces the TFR is around 6,5 children per woman (See Table 1). Badghis’ TFR is 7,0 and the provinces with the highest TFR are Nimroz and Takhar with TFR around 8,0. The average TFT for these provinces is, in general terms, lower than average TFT from the first round of Surveys; TFR for most of them, with the exception of Kabul that greatly differs from the rest of the country were around 8 CPW.
These sets of provinces in either round of the SDES do not represent a specific population and except Kabul, they also do not differ significantly from one another in terms of life conditions. Then the average difference in the TFT fertility in the two rounds, being the estimates separated by nearly a quinquenium, turns out to be one more evidence that national fertility may have entered into the transitional phase. Adolescent fertility rates (at age 15-19) have a relatively high variation, ranking to over 100,0 in two provinces (Badghis and Nimroz) to less than 50 in Balkh that holds the minimum risk of having a live births among adolescents.
11 Considering the Afghan high fertility regime, ASFR15-19 above 100 would be expected for major part of the provinces; this feature should be certainly attributed to the national efforts in the last decade for implementing strategic interventions targeting young people. These strategies include incentives and obligations to increase female school attendance and legal rules to avoid child marriages. An important consequence of such policy interventions has been keeping young women in school and out of early marriage3. Data collected in the SDES may have captured the effects of such policy interventions, as they have registered lower risks of having a child among young women in most of these provinces; those effects also may explain the national low adolescent fertility rate register in the AfDHS (87,0). The consistency of the estimates for the seven provinces can be evaluated through the ASFR by province; they are have, in general, similar values according to age. At ages 25029 or 30-34, prime ages of the reproductive period, apart from Nimroz and Takhar, values are similar between provinces.
It is worth to emphasize that the probability of having a live birth at the central age of the reproductive period (25-29 and 30-34) are around 300 per thousand women, meaning that either, about a third of these women have a child every year or that, women at these ages, have children every two/three years.
Note, also the almost coincidental ASFR values among the last four provinces (Herat, Samangan, Baghlan and Balk) suggesting a rather homogeneous fertility behaviour in these provinces.
The age fertility pattern
The relative contribution to the total fertility by age (last panel in Table 1) indicates that the share of the very young women in the TFR is below 10 per cent in all provinces, even in the provinces where the ASFR at ages 15-19 is above 100 live births per thousand girls (Badghis and Nimroz). Other provinces with adolescent fertility risks around or below 60.0 per thousand contribute with 5 or less per cent. Similar pattern was found in the first round of Surveys. The type of Afghan social organization where live births only occur inside marriage explains the adolescents’ small contribution to the total fertility; incentives to delay female age of entrance at marriage after age 18 or 20 years is also a contributing factor. This relative small
12 participation diverges to those observed in developing countries in Regions other than South-Central Asia, where the contribution of adolescents to total fertility can be as high as 15 per cent or even near 20 per cent.4
Most of the fertility risks take place between ages 24 to 34, where women accomplish for nearly 70 per cent of the total fertility.
The relative distribution of ASFR is plotted in Figure 3.The shapes they reproduces are, in general, similar to each other and, up to age group 35-39 suggest very little fertility control, which is consistent with the relatively high TRF. As it happened in the previous SDES round, any decrease or increase in fertility takes place without considering the number of children the woman already has, thus causing a concave shape, opposite to the convex curves that are typical in populations with a significant prevalence of family planning or modern contraception, which is not the case of Afghanistan. By 2015, modern contraception prevalence among women at reproductive age was near 20 per cent; married women below age 20, notwithstanding presented a 6 percent prevalence. Contraception practice did not vary over the quinquenium 2010/2010 (AMS, 2010 and AfDHS, 2015).
Lastly, note that the age pattern of Nimroz shows a less concentrated shape, corresponds coincidentally (and consistently) to the highest TFR in this set of surveys, corroborating the hypotheses of a quite little contraceptive practice.
As already mentioned, in the Afghan society, children are born only inside marriage and contraceptive practices are rather scarce. Hence, probably the recent fertility changes may be mostly associated to changes in nuptiality –particularly, reduction in the proportion of ever-married women in young ages. In that case, the fertility decline may not affect the age pattern, except in the very young ages where delays in marriage occur.
The mean age of the fertility distribution (M), a synthetic indicator of the fertility pattern, is around 30 for all the provinces and consistent with a high TFR level. Furthermore, M from the first round of provinces was a year higher and corresponding TFR was, also, slightly higher, on average. The association of M and TFR and the equivalence within two rounds would signal the internal consistency of data on fertility and comparison with Historic M values.They corroborate the direct association between M and the TFR and contextualize SDES values. In general, M
13 from the first and second round of surveys locates within the range of the historic values although above the general trend.
Figure 3.Bagdish, Baghlan, Balkh, Herat, Nimroz, Samangan, Takhar 2015-2017 Total Fertility Rate (TFR) and relative distribution of the Age Specific Fertility
Rates (percent)
Sources:SDES- (2015-2017) UNFPA-Afghanistan and Afghanistan CSO (Micro data) and United Nations, Department of Economic and Social Affairs, Population Division (2015). World Population Prospects: The 2015
Revision, DVD Edition.
Finally, considering the rather homogenous fertility pattern prevalent at either set of provinces, as well as their socio-economic homogeneity, some liberties can be taken and consider them as representative of the country. In addition, as there is a time lag of a quinquenium, or wider, between the two rounds, it is possible to hypothesize that fertility may have, indeed, declined in the country over the period covered by these surveys.
Birth order patterns
Data collected make possible further analyses of the fertility age pattern. The fact that number of children ever born (CEB) is available allows us, also, to consider the fertility by birth order, an important issue in the Afghan context, whose high fertility signals a declining trend. Knowledge of the fertility pattern by birth order and age of woman can be incorporated in social programmes aiming to improve women and children wellbeing. Next lines briefly refer to fertility disaggregated this way in order to illustrate potentiality of SDES data.
Fertility rates by birth order (ASFRn) are calculated similarly to the ASFR, and consider the risk of having a live birth of order N, among women that have already N-1 CEB. As previous findings indicate that fertility is similar among the seven provinces, results for only two provinces are given: Balk and Badghis that have the extreme social development degrees in the set of this seven provinces.
0,0 5,0 10,0 15,0 20,0 25,0 15 20 25 30 35 40 45 Age groups A SF R ( %) Badghis Baghlan Balkh Herat Nimroz Samangan Takhar
14 Age patterns follows the expected shape with young/old age distribution at lower/higher birth orders (Figure 4). While ASFRn levels differ in each province, the shape according to birth order is similar, being a noticeable exception, the ASFR1. Badghis’ ASFR1
among adolescents is about three times higher than in Balkh. Because the rates by order depends on the parity order, ASFR1 at age group 20-24, is, this time, lower in Badghis than in Balkh. Women in Balk have similar risks of having their first or second child, while in Badghis, women at the same age, have a very much higher risk of having their second birth, because most of them have already her first child.
Figure 4.Balkh and Baghis, 2015-2017: ASFR by birth order
Source: SDES- (2015-2017) UNFPA-Afghanistan and Afghanistan CSO (Micro data).
Complete version of the paper will include deep analyses of fertility by birth order and will discuss importance of this data in the planning of social policies.
Figure 5, in any case, shows how fertility operates in both provinces up to the end of the reproductive period and how many times a women at a given age, has been exposed to deliver a live birth .
The Marital Fertility
The synthetic measure of the marital fertility is the total marital fertility rate (TMFR) indicating the number of children a married woman has. The concept assumes that women enter marriage at age 15 and remain married up to the end of their reproductive period; because this is not the case in most of current human populations, the TMFR needs to consider the marriage age pattern where the reproductive process occurs.
Marriage in Afghanistan – the seven provinces included– is universal and proportion of currently married women (CMW) reaches the maximum around age 25-29 and
0,0 15,0 30,0 45,0 60,0 75,0 15 20 25 30 35 40 45 50 A S F R n ( p /t h ou san d ) Age group BALKH 1 2 3 4 5 6 7 8 9 0,0 15,0 30,0 45,0 60,0 75,0 15 20 25 30 35 40 45 50 A S F R n ( p /t h ou san d ) Age group BADGHIS 1 2 3 4 5 6 7 8 9
15 starts to decline well before the end of the reproductive period (50). Although the synthetic measure TMFR does not represent exactly the level of marital fertility, the age specific marital fertility rate (ASMFR) reveals the intensity of fertility among CMW. (See panel a in table 2). There are age groups where most of the CMW are giving birth almost every other year which is, very often the case of women aged 20-24 and 25-29; the three provinces with the highest TMFR (Balkh, Nimroz and Takhar) present an annual risk of having a live birth equal or higher than 450 every thousand women.
Adolescent marital fertility (ASMFR15-19) presents, also significant values ranking from 237 to more than 400 children per 1000 women.
The shape, or age pattern, defined by the ASMFRs is consistent with the high levels detected (See Figure 5); compared to the ASFR from previous table, the contribution of every age group to fertility is similar, with only one important difference in the adolescent fertility. ASMFR15-19 contribution is greater than in the case of the ASFR. Furthermore, note that Herat and Nimroz ASMFR15-19 is even higher than at age 25-29, indicating the high exposure to pregnancies since the very beginning of the reproductive period.
Table 2. Bagdhis, Baghlan, Balkh, Herat, Nimroz, Samangan, Takhar (Circa 2010-2017)Total Marital Fertility Rate (TMFR), Age Specific Marital Fertility Rates
(ASMFR)
Age groups Badghis Baghlan Balkh Herat Nimroz Samangan Takhar
Period of reference 2012-2017 2011-2016 2015-2015 2010-2015 2011-2016 2010- 2015 2011-2016 TMFR* 8,6 9,0 9,3 8,7 10,7 8,3 10,4 a) ASMFR (‰) 15-19 286,9 353,0 346,8 356,8 427,8 237,4 340,9 20-24 386,1 429,6 449,3 403,6 452,1 381,2 468,0 25-29 338,8 375,6 387,9 339,7 397,6 369,7 434,8 30-34 277,9 281,1 308,8 280,3 358,5 304,4 373,2 35-39 211,3 210,2 218,6 204,4 271,1 214,3 286,7 40-44 137,1 94,3 101,9 99,0 150,7 98,2 131,8 45-49 84,2 49,4 47,6 53,8 76,7 51,9 50,2
b) Relative contribution of selected age groups to the total marital fertility ( per cent)
15 – 19 16,7 19,7 18,6 20,5 20,0 14,3 16,3
20 – 34 58,2 60,6 61,6 58,9 56,6 63,7 61,2
35 or more 25,1 19,7 19,8 20,6 23,4 22,0 22,5
Total 100,0 100,0 100,0 100,0 100,0 100,0 100,0
c) Mean number of children per marriage according age at entrance at marriage **
22,5 7,2 7,2 7,6 6,9 8,5 7,1 8,7
27,5 5,2 5,1 5,3 4,9 6,3 5,2 6,4
16
* Adjustment factor for the rates are the same used P2/F2 from Brass’ technique
** ** TMFR assuming woman’s entrance at marriage at the average age indicated in the first column: Source: SDES- 2015-2017, UNFPA-Afghanistan and CSO of Afghanistan (Micro data).
The shape of these age patterns appears very similar to the iconic shape of the Hutterites that practiced the called natural fertility. The shape is, also, similar to the patterns described for the first SDS round of provinces.
The complete version of this paper will include a more detailed analyses and discussion of the marital fertility and their similarity with a natural fertility shape.
The main difference between these Afghan provinces and the Hutterites is the higher relative participation of young women already mentioned which indicates a start of the reproduction process as soon a girl goes into marriage. The phenomenon is more accentuated in the cases of Baghlan, Herat and Nimroz where share of the teenagers marital fertility reaches 20 per cent (panel b in Table 2). In most of the cases, as it is among the Hutterites, major contribution to marital fertility (around 60 per cent) corresponds to women between ages 20-34.
Figure 5. Bagdish, Baghlan, Balkh, Herat, Nimroz, Samangan, Takhar (2015-2017), and Hutterites (1921-30) - Relative Distribution of the Age Specific Marital Fertility Rates –
ASMFR (percent)
Source: SDES- 2015-2017, UNFPA-Afghanistan and CSO of Afghanistan (Micro data) and Coale & Trussell (1978).
The similarity with the Hutterites' pattern indicates the absence of fertility control among CMW, which is indirectly measured comparing ASMFRs at central ages of the reproductive period (primarily, ages 20 to 29). In a different way, older women’s lower contribution suggests the incipient presence of fertility control, which is behaviour not detected in the first round survey (F&N SDES/2016).
0,0 5,0 10,0 15,0 20,0 25,0 15 20 25 30 35 40 45 Age group A S M F R ( pe r c ent ) Badghis Baghlan Balkh Herat Nimroz Samangan Takhar Hutterites
17 As marital fertility has still extremely high levels, this translates into high fertility in the total population fertility, as long as marriage is universal and happens at young ages. However, as mentioned in the previous report on fertility (F&N SDES/2016), “Since young women are delaying marriage, fertility in the total population appears to be relatively low in the youngest group. Yet, as soon as women marry childbearing starts with little child spacing, reaching a very high TFR”. A better procedure to estimate marital fertility must consider the marriage duration, which requires specific data not available now; instead, a simulation assuming different ages of entrance at marriage, gives an idea of the impact of early entrance at marriage on the marital fertility. Panel c in Table 2 indicates what would be the final family size for a woman entering marriage at a given age assuming that she would have the same fertility risks of the CMW at any age since her entrance at marriage up to the end of the reproductive period. That is, if a woman enter marriage at age group 20-24 (an average age of 22,5) her final mean number of children compared to the TMFR reduces as much as the equivalent to the adolescent marital fertility. At average age 27,5, her final mean number of children or average descendant would be around five children per woman5.
Findings about the factors influencing fertility changes
Differentials in the fertility behaviour are evaluated trough three dimensions as a first approach to identify the differentials in fertility and their correlates: place of residence (urban/rural), education and wealth. The findings related to those factors reveal ambiguous relationship in the light of contemporaneous experiences.
Aside results according rural/urban residence, whose expected differentials are rather small, results according education and other indicators of welfare are presented.
Educational levels, known as a powerful predictor of reproductive behaviour
changes indicate different degrees of association depending on the educational criteria adopted (women’s education, highest educational level attained in the household and household head’s education)6
.
5 A didactic explanation of fertility and marriage duration can be seen in Pressat (1961) or Pressat (1972) 6 In this short version of the paper, only fertility according to the highest education level attained in the
household is presented. Other results according to the other educational criteria will discussed in the complete version of the paper.
18 Highest educational attainment in the household suggest different sort of relationship with female fertility. Let aside the provinces with constant TFR across education of the household head, the overview of differentials in the TFR according education suggests that when the no-schooling is surpassed, fertility tends to increase (See figure 6). Findings are similar for some provinces analysed in the first SDES round. Namely, Ghor is one of them. Again, the coincidental pattern in the seven provinces validates the results.
Figure 6 -
Bagdish, Baghlan, Balkh, Herat, Nimroz, Samangan, Takhar (2015-2017) Total Fertility Rate (TFR) according to the household head educational levels
Source: SDES- (Circa 2012) UNFPA/Afghanistan and CSO of Afghanistan (Microdata) and Annex 7.
Household wealth presents fertility differential patterns similar to those found
according to educational levels. In a general way, woman’s number of children may, remain the same or increase, be she inside a household of any richness level. Or, may it decrease, only after the poorest conditions are surpassed.
The positive or negative relationship between fertility and determinants like education and wealth has been registered in other. Positive relationship is more often associated to pre-transitional stages, where better development conditions stimulates higher number of children regardless of any educational level. Stulp and Barret (2016) in a study about wealth, fertility and adaptive behaviour in industrial populations, mention evolutionary theories, where, if individuals are attempting to
maximize their fitness, then more resources should translate into a large number of offspring, as seen in a range of a pre-industrial populations. In a context of extremely
poor conditions, as it is very often the case in rural Afghanistan, slight improvements in material conditions benefiting the whole family would translate into higher fecundity and so fertility. It will only after reaching much better living conditions that family size
b) Household head education 4,5
6,5 8,5
No educ 1 to 6 7+
Education (in years)
T F R Baghlan Takhar Samangan Balkh Badghis Herat Nimroz a) Woman's education 4,5 6,5 8,5 No educ 1 to 6 7+
Education (in years)
T F R Baghlan Takhar Samangan Balkh Badghis Herat Nimroz
c) Highest education attained in the household 4,5
6,5 8,5
No educ 1 to 6 7+
Education (in years)
T FR Baghlan Takhar Samangan Balkh Badghis Herat Nimroz
19 and composition is balanced against other costs and opportunities and the onset of fertility decline would start.
Similar positive relationships has been found analysing patterns of marriage for these provinces and in the first SDES round, where it is found in Ghor –the most vulnerable in the set of the six provinces analysed– that TFR increased as education of the household head increases.
3. Conclusion
Analyses developed in this this paper depart from the premise that fertility transition has started in Afghanistan since several years now. Evidence presented by 2010 and 2015 demographic and health surveys suggests that there is a national process of fertility change, although some regions may remain yet outside. Main results are discussed next.
• Initial comparison of past and current fertility in both SDES rounds as in the case of the national surveys indicates that, for the first round, Kabul, clearer than its neighbors’ provinces of Kapisa and Parwan had experienced declines in their fertility level. That was not the case of the other provinces, where fertility level looked stable, at least. For the second round, almost a quinquennium later, with the exception of Badghis and Takhar, the other provinces’ data indicate that past fertility has been higher. The particularly case of Herat, with an increasing Pi/Fi by age –if all error patterns are similar– definitively signals that fertility is lower than before. As both Kabul and Herat has undergone heavy international return migration, a question to consider is the role of this phenomenon in the reproductive behaviour of each province.
• An additional evidence that fertility is declining in Afghanistan comes through the overview of the fertility mean age. SDES results show that, on average, the mean age of the fertility pattern has decreased between the two rounds. If each SDES round tells, in general terms, what is happening with the fertility behaviour in the country, results show that fertility is actually declining.
• A remarkable fact is the relatively low risk of having a baby among very young women. In most cases, that risk is less than 100 per thousand, and even, near 50 per thousand among women aged 15-19. Those rates are, in general terms, lower than expected for a TFR above, say,7.0/6.0. Whether the decline trend is a result of recent successful strategies, aiming to improve reproductive health, it should be kept in mind that fertility is still very high. Current ASFR at the foremost ages of the
20 reproductive behaviour are well above by world standards. A 25 or 30 year-old woman has, on average, a baby every two or three years in most of the thirteen provinces
• Marital fertility suggests a regime near to natural fertility, where fertility control depending on the number of children the women have is inexistent. This pattern was more accentuated in the first SDES round. Marital fertility at ages 15-19 are at the most elevated levels, the risk of having a live child is very often above 30 percent. At the following age groups (20-30), the risk remains high; there are provinces where the ASFR is above 40 percent, which means that almost every other married woman has, annually, a live birth, letting aside induced abortions, miscarriages and stillbirths. • Social determinants, apart from the urban/rural condition, suggest an ambiguous relationship with the fertility behaviour. Firstly, education does not present, in general terms, significant influence when women have no education or have entered basic education. It is only among most educated women (7 or more years of schooling) that results show TFR below 5,0. Secondly, classification of fertility levels according to either household head’s education or the highest educational level attained in the household suggests that an initial improve in education is associated with no change or even increase in the TFR. It is only when the highest educational level is reached that TFR goes down. Thirdly, family wealth suggests similar association: initial improvements in material welfare associate to either no changes or a slight increase in the TFR.
• Prospective results indicates (not included here) that if the difference between current and cumulated fertility remains, current young women will probably end up their reproductive period with a smaller number of children than women currently finishing this life phase. The majority of women aged 45-49 years (over 50 per cent in general) has, at least, seven children today. The equivalent proportion among current women aged 25-29 could be just around a third or less by the end of their reproductive period.
REFERENCES
APHI, CSO; IIHMR, ICF. Macro, EMRO: Afghanistan Mortality Survey 2010. Calverton, Maryland, USA: Afghan Public Health Institute, Ministry of Public Health, Central Statistics Organization, ICF Macro, Indian Institute of Health Management Research. World Health Organization/EMRO, 2011.
21
BASU, A. Why does education lead to lower fertility? A critical review of some of the possibilities. World Development, v. 30, n. 10, p. 1779-1790, 2002.
FERNÁNDEZ, R. The reproductive behaviour in Afghanistan–Signals of the onset of the fertility transition in Kabul and selected provinces at the 2 nd decade of the XXI century. 2017.
BONGAARTS, John. The fertility-inhibiting effects of the intermediate fertility variables. Studies in
family planning, p. 179-189, 1982.
BRASS, W. Advances in methods for estimating fertility and mortality from limited and
defective data. 1985.
BRESCHI, M et al. Fertility transition and social stratification in the town of Alghero, Sardinia (1866-1935). Demographic Research, v. 30, p. 823, 2014.
BRYANT, J. Theories of fertility decline and the evidence from development indicators. Population
and development review, v. 33, n. 1, p. 101-127, 2007.
CARLSSON, G. The decline of fertility: innovation or adjustment process. Population Studies, v. 20, n. 2, p. 149-174, 1966.
CENTRAL STATISTICS ORGANIZATION (CSO), Ministry of Public Health (MoPH), and ICF. 2017. Afghanistan - Demographic and Health Survey (AfDHS) 2015. Afghanistan: Central Statistics Organization, 2016.
COALE, A. The decline of fertility in Europe. Princeton University Press, 2017.
CUMMINS, Neil. Why did fertility decline?: An analysis of the individual level economics
correlates of the nineteenth century fertility transition in England and France. 2009. Tese de
Doutorado. The London School of Economics and Political Science (LSE).
http://etheses.lse.ac.uk/39/1/Cummins_Why_did_fertility_decline.pdf - In: 2/2/2018 6:16:06 PM
HAIDER, S. Childbearing and contraceptive decision making amongst Afghan men and women: a qualitative analysis. Health care for women international, v. 30, n. 10, p. 935-953, 2009.
HAYES, C. Risking the future: Adolescent sexuality, pregnancy, and childbearing. National Academies, 1987.
KELLEY, Allen C.; SCHMIDT, Robert M. Evolution of recent economic-demographic modeling: A synthesis. Journal of Population Economics, v. 18, n. 2, p. 275-300, 2005.
KLEPINGER, D.; LUNDBERG, Shelly; PLOTNICK, Robert D. Adolescent fertility and the educational attainment of young women. Family planning perspectives, p. 23-28, 1995.
(http://www.jstor.org/stable/2135973) - In 25-01-2018 21:07
BACCI, M. A history of Italian fertility during the last two centuries. Princeton University Press, 2015.
PRESSAT, R. L'analyse démographique: méthodes, résultats, applications. Population (french
edition), p. 505-508, 1961.
SINGH, S; DARROCH, J. Adolescent pregnancy and childbearing: levels and trends in developed countries. Family planning perspectives, p. 14-23, 2000.
STULP, G; BARRETT, L. Wealth, fertility and adaptive behaviour in industrial populations. Phil.
Trans. R. Soc. B, v. 371, n. 1692, p. 20150153, 2016.
DESA, U. N. et al. World population prospects: the 2015 revision. Population division of the
department of economic and social affairs of the United Nations Secretariat, New York, 2015.
https://sustainabledevelopment.un.org/post2015/transformingourworld/publication (in: 18/01/2018
19:47:46)
UNITED NATIONS FOUND FOR POPULATION ACTIVITIES/ CENTRAL STATISTICS OFFICE OF GoIRA. Fertility and Nuptiality - Provinces of Kabul, Bamiyan, Daykundi, Ghor, Kapisa and
