Fertility transition in selected sub-Saharan African countries: the role of family planning programs

Analytical model

Using all the trend data from Measure DHS gathered between 1986 and 2016 across sub-Saharan African countries (listed in Table 1), we apply² a reformulation of^8–10 a conceptual scheme in which the variable ‘fertility’ is measured by the total fertility rate (F₀), a function of the supply of births (natural fertility), the demand for births (wanted fertility) and the degree of preference implementation index⁹. The latter in turn is dependent on cost of fertility regulation and cost of unwanted childbearing. The degree of preference implementation is the net result of a decision-making process in which couples weigh the cost of fertility regulation and the cost of unwanted pregnancy. Figure 1 shows the diagrammatic presentation.

Figure 1. Variants in the supply-demand model.

Relationship between variables

According to Bongaarts (1993):

C is an index ranging from 0 to 1 measuring the proportional reduction in F_n attributed to deliberate birth control mechanisms. Birth control not only confined to contraception also encompass induced abortion practices though always ignored in studies. F₀ data is always available and hence the only additional task required is to compute F_n in (i) with an estimate of C. Bongaarts further provided a procedure for deducing C, with an approach though the limitation was the unavailability of data. Hence:

Where U is the fraction of women in marriage practicing all forms of birth regulation except during the post-partum in-fecundity period. The error associated with this is negligible hence ignored sometimes. By substituting C in Equation 1 above yields the anticipated approximation of natural total fertility.

F_w computation: According to Bongaarts, the favoured approach is dependent on the equation below:

Where F^w is the proportion of women who want more children, equaling the resulting total fertility after deleting all births to women who want no more children at the time of the survey and Wm (40–49) is the proportion of women in union aged from 40 to 49 who want no more births.

These two equations helped normalize the biases in order to compute the respective F_n and F_w trends across the regions. With most of the erratic curves expected, a normalization process using the natural logarithms of the equation was applied to give meaning to the various trend curves.

I_p is derived from a synthesis of past studies. It begins from the fact that all social and economic factors of fertility operate through a unified pack of proximate factors to exert an impact on fertility⁸. Easterlin’s economic approach is a model of behavioural and biological factors affecting fertility in developing countries. The model consists of three central concepts: demand for children; the potential supply of children, and the momentary and psychic costs of contraception. According to the model, women whose potential supply of births exceeds demand would consider contraception, taking into consideration the costs involved while choosing suitable family planning methods^8,9.

The model is simple and attractive; however, it cannot address dynamic issues and has not succeeded in quantifying these factors in acceptable manner². Emerging from the model is the fact that fertility (measured using F₀) is a function of three determinants namely: supply of births (F_n), demand for births (F_w) and the degree of I_p (Figure 1).

Supply of births (F_n) is measured as natural total fertility. F_n infers the rate of birthing likely to prevail minus the premeditated attempts by spouses to limit their number of children. Demand for births (F_w) is the wanted total fertility defined as the rate of prevailing childbearing after eliminating all unwanted births. Under normal circumstances, it is simply calculated as F₀ while eliminating the unwanted births from the numerator. Unwanted births are births occurring after an achievement of the ideal family size. Any births that are mistimed though occurring before the achievement of the desired family size are considered wanted births as well.

The degree of I_p is an index from zero value to unity. Its level of implementation implies the net result of decision-making process. This is the state in which a spouse ponders the cost of fertility regulation as they consider costs of bearing an unwanted child to its end. In general, the index has an inverse variation to the cost of fertility regulation as well as a reverse correlation to the unwanted births. If couples fully implement their fertility preference, the index is equal to unity. This signifies that no unwanted births occur as actual fertility corresponds to F_w. Conversely, if the index is equal to zero, the observed fertility equals F_n, that is, fertility in the absence of any deliberate fertility control assuming women remain sexually active over their reproductive cycles. The value of the index at play stipulates the position where actual fertility falls as dictated by the range set between wanted and F_n parameter levels.

F₀ gives the estimate of the number of children a woman would have by the end of childbearing if she were to pass through her reproductive cycle at the customary age specific birth rates. The model shows that the operation of these variables determines the level of fertility in a community or households. In this variant of the original Easterlin and Crimmins (1985) model, infant and child mortality dynamics affects the desired fertility rather than F_n. Women are deemed to possess precise desired fertility size translated and actualized into numbers through subsequent births after considering past child losses and risks related to future child deaths as well.

According to this variant, as development occurs, the trend in prevailing fertility transforms to become a function of the equilibrium between the F_w, F_n and the degree of fertility I_p. F_w is expected to decline over time, as a result of the changes associated with the costs and benefits of child bearing⁹, as well as reductions in the infant-child mortality. I_p rises as fertility regulation costs decline; with the benefit of fertility regulation focusing on the elimination of any unwanted births⁸. According to 8, the relationship between these variables under discussions and fertility can be expressed in statistical form as follows:

Where F_u is the unwanted fertility (which can simply be expressed as F₀ – F_w).

Where I_p has a range of 0 to 1. With full I_p, I_p = 1 (which implies that F_u = 0 and F₀ = F_w) and I_p = 0 with no prefer I_p (This implies a substantial level of unwanted childbearing and F₀ = F_n). Noted here is that as defined by Bongaarts, F_n here is not the same as in total fecundity as in the Bongaarts proximate determinants but taken to mean fertility level achieved in absence of contraception⁸.

F_u is a function of the difference between supply and demand, and the degree of I_p.

Substitution of Equation 5 in Equation 6 yields:

Noting that F_n is given by:

Where C implies an index ranging from 0 to 1 measuring the reduction in proportional of F_n attributable to deliberate birth control is estimated as:

Where U represents the proportion of married women who were practicing contraception at the time of survey. It is measured as the number of married women using contraceptive method to the total number of married women. The values for U and C can be used to estimate F_n

Rearranging Equation 6 gives:

Equation 7 can now be used to estimate the degree of I_p once F_n (fertility in absence of contraception), actual fertility and F_w are known. One thing to note is that the estimation of F_w, as previously done overtime contain traits of upward bias as per the recent observation. An alternative estimation of F_w from Bongaarts model detail that the average F_w derived from the wanted status of births as reported by women was 2.8, indicating that this measure of wanted fertility contains an average upward bias of 0.4 birth.

The analysis consists of two stages. First, we computed the degree of I_p within the variable categories overtime from F_n, F_w and F₀. F₀ and F_w as a series of indicators are provided by the various country specific DHS reports. This involves compiling all the components of the index within the equation so as to come with the actual figures per the subsequent time intervals. The component variables are F_n, F_w computed to form the degree of I_p. Further correlation analysis between the degree of I_p and the unmet need to contraception, conducted using SPSS v18, was run.

Decomposition of fertility trends

According to Bongaarts (1993), the core objective of the demand framework lies in the identification of the causes of fertility decline in a population, with proceeding comparative analysis providing worthwhile insights yet not achieving its sole objective. Turning to the issue at hand, the decomposition of the variations in fertility and to abridge the methodological exposition, trends therefore should inform the basis of focus between two points in time, i.e. T₁ and T₂ running up to the determinants. The derivation of the decomposition equation also warrants the introduction of the variables listed in Table 2.

The decline in fertility between the two periods is F₁-F₂, conveyed by substitution as

The above equation therefore can be written as below

In Equation 8: ∆F, ∆F_w, ∆F_n and ∆I_p are the change within F, F_w, F_n and I_p respectively.

In Equation 9: Ѓ_w, Ѓn and Ī_p are the mean values of correspondingly, F_w, F_n and I_p. For example, the mean of the degree of implementation index (Ī_p) is: - [0.5(I_p1 + I_p2)]

Ī_p implies the average of the Degree of Fertility Implementation Index (I_p). The influence of change in wanted (∆F_w ) as well as the natural (∆F_n) fertility to prevailing fertility change hinge on the average extent or degree of implementation. Consequently, the outcome of fertility from every shift registered on the degree of fertility implementation index is determined by the corresponding mean change between F_n and F_w (Ѓ_w-Ѓ_n). This function requires two successive points in the estimates of the parameter measurers i.e. F₀, natural and F_w including the implementation index as well within the population under consideration. It is this function that is used to determine the extent to which implementation of fertility desires contributes to fertility transition (Table 4).

Decomposition of fertility change and the contribution of F_w and I_p

As Table 1 shows the trend change in fertility parameters measurers, Table 3 further shows the decomposition of fertility changes among countries with two or more surveys. Results reveal there are indeed substantial variations between countries in terms of fertility preference parameter measurers as well as the implementation indices by countries. These results clearly indicate the important role played by the changes in I_p, F_w and F_n. Converse to the expected, eight countries actually increased their F₀ over the period 1986–2016. In six out of the eight countries where fertility increased, there was a decline in degree of I_p. Subsequently, in five of the eight countries there was an increase in F_w. The largest decline in fertility rate occurred in Rwanda, Malawi, Kenya and Ethiopia. The four countries subsequently had the greatest contribution of the degree of I_p to fertility decline. On the same note the greatest contribution of F_w decline to fertility change occurred in Malawi, Rwanda and Kenya.

In absolute values, Rwanda, Malawi and Kenya experienced the highest fertility changes as well, while Niger, Mozambique and DRC experienced an increase in fertility rate within the periods 1986–2016. Looking at the contributions made by each of the fertility parameters, the fertility preference (F_w) and the degree of I_p are the reasons for the variations in the changes in fertility. Rwanda registered a 37% decrease in its average wanted fertility desires, with a corresponding degree of I_p of 97% (Table 2). Malawi and Kenya on the same note registered a reduction in F_w of 51% and 54% and corresponding implementation indices of 82% and 84%, respectively.

Figure 2 highlights the graphical correlation between I_p and unmet need for family planning. There is an inverse correlation between I_p and the unmet need for family planning. High unmet need for contraception leads to a low implementation index, since contraception is the sole contributing factor to fertility regulation (also referred to as the extent of I_p). This is because the extent of contraceptive availability and subsequent utilization of contraceptives is what defines I_p level. The absence of these essential birth control commodities leads to non-implementation of family planning, thereby failing to restrain F_n.

Figure 2. Correlation between implementation index and unmet need for family planning.

Figure 3 echoes the performance of I_p in facilitating the reduction of fertility by each of the countries within the periods under study. Ethiopia, Rwanda and Sierra Leone are the three countries where I_p has most contributed to the fertility decline. However, in some countries, the limited or non-implementation of fertility led to the index not facilitating any declines in F₀, thereby allowing the natural increase to take its course. These countries were Mozambique, Chad, Cameroun, Democratic Republic of Congo, Togo, Benin, and Congo Brazzaville.

Figure 3. Percent change in preference implementation since 2000.