 |
           |
Information Resources"Health
Matters" | Facts
and Figures | Reports
and Articles | Why Increasing
Contraceptive Use Doesn't Always June 2001 Vera Zlidar, MHS, Research Writer,
Population Reports It might be expected that increases in contraceptive use would always result in a concurrent drop in fertility. However, because other factors affect fertility, the expected association does not always take place. Such factors include the measures that are used, population structure, and the timing of childbearing. Fertility determinants other than contraceptive use also affect changes in fertility. Demographic and Health Survey data for a number of countries are used below to illustrate several seeming incongruities for the commonly expected association between contraception and fertility. The Total Fertility Rate. Although it is not the only measure of fertility, the Total Fertility Rate (TFR) is perhaps the most commonly used, at least partly because of its intuitive appeal. However, it is important to understand how the TFR is calculated, the assumptions behind it, and its correct interpretation. The TFR is a "synthetic rate" or a "hypothetical measure" derived from age-specific fertility rates (ASFRs, which are births per woman of a given age) at one point in time. Adding these rates together for all ages, we have a figure that can be interpreted as the number of children a woman would have, assuming no mortality before age 50, if she experienced those particular rates at each age throughout her childbearing years. Although it is often described as measuring "family size" or "number of children," the TFR does not represent the family size of any group of real women, but rather the number of children that would be born if those rates pertained throughout. Thus, the TFR is not a direct measure of something that is happening today. The TFR makes the assumption that the ASFRs from a given period remain constant throughout a hypothetical woman’s lifetime. However, in reality fertility rates are always changing, so the TFR is hypothetical, not real. It is possible to calculate a completed fertility rate by looking at the childbearing experience of women who have finished childbearing, but this is a look backwards for periods of 35 years and more, and thus is not able to provide information on what is taking place today. Effects of Age Structure. The age structure of the population in a country strongly affects summary contraceptive use measures and some fertility measures, but not the TFR, which is derived from ASFRs and hence free of the effect of the age structure. If the majority of the population is young, as in many countries that have had high levels of fertility in the recent past, their behavior can dominate summary measures that do not take age into account. Take a hypothetical example. If 60% of the women of childbearing years are ages 15-29, their behavior, e.g., contraceptive use rates, will weigh the average contraceptive use rate more heavily than will the behavior of the 40% of women ages 30-49. The usually observed pattern of contraceptive use is that use is low among the youngest and oldest women, increasing and then decreasing in a curvilinear pattern. The pattern can vary greatly and is not aptly represented by the average level of contraceptive use among all women. Fertility measures, such as the Crude Birth Rate (CBR) and General Fertility Rate (GFR) that are not adjusted or controlled for age are similarly affected by the age structure. More detailed analyses of fertility will always take age structure into account, generally by age standardization or by using age-specific measures. Effects of Timing. The timing of births among women in a population—called "tempo"—can distort Total Fertility Rate calculations, even when women are ultimately having no more or no fewer children. The essence of the effect of the timing of births on TFR is this: TFRs are temporarily deflated during periods where women delay childbearing, and temporarily inflated in times when childbearing is accelerated among younger ages. There are many examples in the literature of the effects of tempo on Total Fertility Rates and proposed ways for adjusting TFRs in order to remove timing effects [see Bongaarts, 1999 for a detailed discussion on fertility distortions due to tempo]. Other Fertility Determinants. The CPR is not the only behavioral factor that affects fertility levels. Other "proximate factors" such as age at marriage, level of postpartum insusceptibility, coital frequency, and abortion also affect fertility and are included in detailed mathematical discussions of the determinants of fertility. To provide a complete explanation of changes in the TFR, one needs to break down the whole and examine the parts (age-specific fertility rates, age-specific contraceptive use rates, age at marriage, age at first birth, etc.) in order to understand better what factors are affecting fertility and among which sub-groups. While contraceptive use is one of the strongest factors affecting fertility, it is not the only one. How often does one find seemingly discordant changes in fertility and contraceptive use? Table 1 provides data on TFR and CPR in countries with more than one survey since 1990. One can see incongruities of four types. 1. Considerable increases in contraceptive use (50% or more users in the subsequent survey) were not always accompanied with considerable declines in the TFR (at least 10% decline in the subsequent survey), as in Haiti, Malawi, Senegal, and Tanzania (between the 1991-92 and 1996 surveys). 2. Increases in contraceptive use were accompanied by increases in total fertility, as in Bangladesh, Colombia (between 1990 and 1995 surveys) Niger, and Turkey. Also, despite increased contraceptive use in Paraguay and Peru, there was no change in fertility. These countries are showing the effect of changes in tempo. 3. In Ghana and Kazakhstan, there were substantial decreases in the TFR (10% or greater), but increases in contraceptive use less than 10%. 4. Lastly, in Rwanda fertility fell by .4 children per woman but the CPR actually decreased by 44 percent. Table 1.
Two main conclusions emerge from this discussion. First, because other direct factors also affect fertility, the CPR is not the only predictor of what will happen to fertility levels. Second, the limitations of summary statistics must be understood, and interpretations of such statistics should be made with their strengths and weaknesses in mind. The level of contraceptive use is one of the strongest factors affecting the level of fertility. A linear regression of 105 countries comparing fertility levels and contraceptive use levels finds that 77% of the variation in fertility is explained by variation in contraceptive use [Gardner et al., forthcoming]. As the data above indicate, the factors that comprise the remaining 23% of the variation in total fertility are also important, as are the measurements used to examine such associations. Bongaarts, J. The fertility impact of changes in the timing of childbearing in the developing world. Policy Research Division Working Paper No. 120. Population Council. New York, NY. 1999. 33 p. Gardner, R., Zlidar, V. M., Rustein, S., and Morris, L. Survey Update: The reproductive revolution continues. Series M, No. 16. Baltimore, Johns Hopkins School of Public Health, Population Information Program. Forthcoming. Click here to
view this document in Microsoft Word format
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
