The transition from high fertility to low fertility has proceeded at very different rates in the major regions of the developing world. The diversity of experience in Asia ranges from Thailand and Sri Lanka and other East Asian countries, where fertility has fallen to replacement and even lower, to Nepal and Pakistan, where the transition is just beginning. Similarly, in Latin America, fertility has declined rapidly in Colombia, and more recently in Brazil and in many other South American countries, but it remains higher in much of Central America. In contrast to North Africa, where fertility decline is well underway, most of the countries in Sub-Saharan Africa remain at the threshold of the transition—although for some countries in southern and eastern Africa, there is convincing evidence of significant declines.1
Not only does the tempo of the fertility transition vary greatly in these different regions, but the pattern of change in Sub-Saharan Africa differs from that seen in other regions of the developing world, as well as from earlier European experience. The decline typically began when couples attempted to limit completed family size; the regulation of marital fertility would begin mainly when the desired number of children had been reached. In Sub-Saharan Africa, on the other hand, with few exceptions, contraceptive use has been adopted mainly to space births, motivated principally by considerations of maternal and child health, rather than by the socioeconomic forces that presumably motivated concerns about family size.2
Thus, whereas the adoption of contraception for family limitation has been the traditional and principal instrument of fertility reduction in most of the world and its use for spacing purposes a later development, the pattern seems reversed in much of Sub-Saharan Africa. Where the motivation has been to limit births, fertility reduction has typically been concentrated initially among women at older ages; in contrast, where the objective has been to space births, declines have occurred at all ages.3 Increases in age at marriage have also played an important role in general fertility reductions in all parts of the world.
In this article, we track trends in demand for fertility limitation in developing countries and document patterns of regional differences in the spread of the family planning norm. Using data spanning more than two decades, we examine trends in total potential demand for fertility limitation, in its component parts ("met" need and "unmet" need) and in the proportion of users who practice contraception to limit births.
Sources of Data and Measures
The data assembled for this article were collected mainly as part of the Demographic and Health Surveys (DHS) project and its predecessor, the World Fertility Survey (WFS). Contraceptive Prevalence Surveys (CPS) were also used in a few instances to provide comparative data. Our main criteria for selecting a country were that it had participated in at least two nationally representative surveys and that the surveys included the relevant measures of demand and contraceptive use. Altogether, 41 countries participated in a total of 108 surveys that currently satisfy these criteria. The range of time covered by the surveys extends from the mid-1970s to the late 1990s. The number of surveys and the interval between surveys in these countries have all varied greatly over this period.
We have grouped the 41 countries that provide the data for this analysis into three regional categories. The 18 Sub-Saharan African countries and the 11 Latin American countries are reasonably representative of their respective regions, although some important countries (such as South Africa) are not included. Central American nations are also underrepresented, with survey data only for Guatemala.
The third grouping is a much more heterogeneous residual category of 12 developing countries that consists of a number of Asian populations, as well as Egypt, Jordan, Morocco, Tunisia and Turkey. We included estimates for India taken from a published 1980 national family planning survey4 and from the 1991-1992 National Family Health Survey (which was modeled after the DHS). Recent estimates from Pakistan are from a 1994-1995 CPS. China is not included because comparable data are not available over time.
Our analysis focuses on the potential demand for limiting births, which consists of two groups: the proportion of married women who want no more children and who are practicing some method of contraception; and women who want no more children but who, for one reason or another, are not practicing contraception (although they presumably are fecund*). This latter component is commonly referred to as the unmet need for family planning; its use here only crudely represents that measure, however, and by design is confined to women who wish to limit or terminate their childbearing. These two groups together constitute an index of the total potential demand for limiting births—defined as the percentage of currently married fecund women who want no more children.
We concentrate here on birth limiting rather than on spacing not only because of its historical significance, but also because of its much greater impact on fertility rates. In addition, the proportion of women who want no more children has been shown to be a strong predictor of contraceptive prevalence and the total fertility rate.5 Finally, data on spacing intentions are not comparable across the different survey projects.
For most of the surveys, the estimates that we present are calculated from the data files. For the few surveys for which such data are unavailable, we obtained estimates from published information. We present country-specific percentages for Sub-Saharan Africa, Asia and North Africa, and Latin America and the Caribbean. The trends in the total potential demand for limiting are examined first, followed by the percentage of women wanting no more children and using contraceptives (the actual current demand, or "met need," for limiting) and trends in the percentage of women who want no more children and who are not using any contraceptive method (a crude measure of unmet need).
We also study in this analysis the subset of fecund women who want no more children, are not practicing contraception and are currently exposed to the risk of an unwanted pregnancy; this group excludes women who are currently pregnant or amenorrheic,† regardless of whether that pregnancy or birth was wanted or unwanted. These are the women who ostensibly are immediately in need of family planning. This is a very narrow indicator of unmet need, but across countries it is highly correlated (Pearson's correlation coefficient, r=.85) with the conventional DHS measure, which takes into account the planning status of the pregnancy or birth among those who are currently pregnant or amenorrheic.
In all of the measures described above, the denominator is the total number of currently married women. However, in the course of our examinations, we shift the denominator to all currently married women who are using contraceptives and focus on use for limiting births relative to use for spacing births. To highlight regional trends, as well as differences and similarities between regions in the various measures, we summarize the data for the 41 countries in terms of regional averages. To obtain fairly stable averages for the trend analysis, we divided the period covered by the study into three main time periods: 1975-1979, 1980-1989 and 1990-1998. For each region, we averaged the proportions for each measure over each of the three time periods.‡
Total Potential Demand for Limiting
In Sub-Saharan Africa, there is evidence of a slow but steady rise in the proportion of women who want no more children (Table 1). The increase has been dramatic in Kenya, where it rose from 15% in 1977 to 30% in 1984 and to 46% in 1993 and 1998. (A similar pattern might have been evident in Botswana and Rwanda if earlier data had been available.) Although all but one (Madagascar) of the 18 countries in this region show some increase in the demand for fertility limitation, levels remain far below those in other parts of the developing world. The absence of a change in potential demand in Kenya between 1993 and 1998 suggests, however, that this measure may be leveling off in that country, as has been the case in some countries in Asia and North Africa and in Latin America and the Caribbean.
The demand for fewer children has continued to increase in the countries included here in Asia and in North Africa (Table 2). In Morocco, for example, the proportion who wanted no more children rose from 32% in 1980 to 42% in 1987 and to 51% in 1995. In general, the level of demand for limiting births clusters in the 40-60% range. Compared with Sub-Saharan Africa, where the level of demand was below 20% in the 1980s in most countries, demand was well above 20% during the 1970s in the Asian and North African countries represented here.
The pattern of a high proportion of women wanting to stop childbearing seen in Asia and North Africa can also be found in Latin America (Table 3). However, there are more instances of a leveling of the trend in the Latin American countries than in Asia. In five of the 11 countries in the region, the proportions exceed 60%, and in Brazil the proportion had reached 75% by 1996.
Use of Contraceptives for Limiting
The percentage of married women who both want no more children and are using contraceptives (the "met need" for limiting) is the primary component of the total potential demand for limiting. Overall, while Asia and North Africa and Latin America have similar patterns and trends of contraceptive use for limiting, Sub-Saharan Africa lags behind these regions in this behavior.
Table 1 shows that several countries of Sub-Saharan Africa, notably Botswana, Kenya and Zimbabwe, are conspicuously in the vanguard of the take-off of the smaller family transition. For example, in Kenya, the proportion of currently married women who want no more children and are currently practicing contraception increased from only 3% in 1977 to 10% in 1984 and to 23% in 1993 and 25% in 1998. (When comparable data for South Africa become available,§; it will probably be in this category as well.)
A second group, comprising Ghana, Rwanda and Uganda, also shows an emerging momentum. Many of the countries in this region, however, remain below 10%, although for some of these the most recent survey data are nearly 10 years old. One has the sense that the norm of fertility limitation and smaller families could develop quickly in the region.
A continuing, steady increase in the proportion of limiters is evident in Asia and North Africa (Table 2), although there is considerable variance within the group. At the low end is Pakistan, where the proportion practicing contraception for limiting rose from 5% in 1975 to 14% in 1994-1995. At the high end, and continuing to increase, are Sri Lanka, Thailand and Turkey: In Turkey, the proportion rose from 27% in 1978 to 51% in 1993. These countries are probably now at about the same level as the United States (about 52%), as estimated from the 1995 National Survey of Family Growth.6
Again, the pattern of growth of limiters in the populations of Latin America (Table 3) is quite similar to that described above for Asia and North Africa. Only Haiti shows no sign of taking off: Twelve percent of married, fecund women wanted no more children and were actually using contraceptives both in 1977 and in 1994-1995. Guatemala and Paraguay appear to be on a slow upward trajectory. In the other countries of the region, in contrast, the proportion of limiters is increasing rapidly. For instance, in Peru, this proportion rose from 22% in 1977 to 35% in 1986 and to 46% in 1996.
Unmet Need for Limiting
The unmet need for means to limit births is defined simply as the proportion of all currently married, fecund women who want no more children and are not practicing any method of contraception. It represents the other component of the total potential demand for limiting. We focus here more on the trend in this measure than on the absolute level.
Unmet need for limiting has increased fairly uniformly for most of the countries in Sub-Saharan Africa (Table 1). A continuous rising trend is evident in Cameroon, Ghana and Senegal, where data are available for four time periods. This can probably be interpreted as a more rapid growth in the preference for fewer children than in the corresponding behavior that is required to implement that preference. It is more complex than simply a race between supply and demand, however, because of ambivalence and various social impediments to contraceptive use.7
The results for Kenya and Zimbabwe, where unmet need appears to have peaked at 26% and 15%, respectively, and then turned downward, suggests that the upward trend is not likely to continue for very long in most of the other countries in the region. Perhaps because of their more mature family planning programs, Kenya and Zimbabwe may well have started to experience a trend in unmet need similar to that observed for most of the countries in Asia and North Africa and Latin America.
Unmet need for limiting is declining as contraceptive practice is adopted in most Asian and North African countries (Table 2). In Bangladesh, the proportion of married women with an unmet need declined from 29% in 1979-1980 to 16% in 1996-1997, while in Indonesia it dropped from 17% in 1976 to 9% in 1997. The only exceptions among these countries are Pakistan, where an initial decline in the level of unmet need for limiting was followed by a rise, and Morocco, where unmet need for limiting has tended to remain constant.
Although less evident than in Asia and North Africa, the same pattern of decline in unmet need can be seen in the Latin American region (Table 3). For example, in Colombia, the proportion of married women with an unmet need for limiting declined from 30% in 1976 to 16% in 1986 and then to 10% in 1995. This same pattern is evident in Bolivia, the Dominican Republic and Peru. The one exception to the general trend is Haiti, where a small rise in unmet need for limiting is apparent.
Women at Risk of Unwanted Childbearing
The proportion of women actually exposed to the risk of unwanted childbearing is a narrower indicator of the total unmet need for limiting than the foregoing measure. It is not intended to reflect the total level of underlying unmet need in the population, since it excludes both currently pregnant and amenorrheic women.** Its purpose is merely to capture trends in the proportion of women who are immediately at risk of a pregnancy that would exceed their desired number of children.
As is evident from the countries with more than two surveys, there is no clear direction of change in this measure in Sub-Saharan Africa (Table 1). On the one hand, Ghana and Senegal show a tendency toward an increase in the proportion at risk of unwanted childbearing at the time of the survey. For example, in Ghana, the proportion of married women currently exposed to the risk of unwanted pregnancy and not practicing contraception increased from 5% in 1979 to 7% in 1988 and to 9% in 1993. On the other hand, estimates for Kenya and Zimbabwe show an initial increase, followed by a sharp decline. In Kenya, this measure increased from 6% in 1977 to 12% in 1989 before declining to 9% in 1993 and to 5% in 1998.
There is a pronounced decline in the proportion of women at risk of unwanted childbearing in most of the Asian and North African populations included here (Table 2). This pattern is especially precipitous in Bangladesh, which has had a very strong family planning program, with the proportion dropping from 22% in 1979-1980 to 6% in 1993-1994. In some of the countries included here (especially Egypt, Indonesia and Morocco), the measure seems to be leveling off following an initial decline. Only in Pakistan is an initial decline followed by a substantial increase.††
The countries in the Latin American region are also mostly on a downward trajectory (Table 3). This trend is clear in Bolivia, Colombia, the Dominican Republic and Peru: In both Colombia and Peru, the proportion of married women who were at risk of unwanted childbearing and were not using contraceptives decreased from about 16% in the middle of the 1970s to less than 5% in 1996. Three countries (Brazil, Guatemala and Haiti) showed little or no change over time.
Contraceptive Use for Limiting vs. Spacing
As noted earlier, most women in Sub-Saharan Africa who practice contraception do so to space births rather than to limit them. The trends depicted in Table 1, however, suggest that this is changing: Among users, the proportion using contraceptives for limiting has increased, although it is typically still below the halfway mark. In Cameroon, Ghana and Senegal, this proportion more than doubled between the late 1970s and the early 1990s. Kenya is the only country that has reached and surpassed the halfway mark: In 1977, the proportion of all users who were practicing contraception for limiting was 38%; by 1993, this proportion had increased to 70%, before decreasing to 65% in 1998.
Use of methods for birthspacing was the first type of use in most of Sub-Saharan Africa, while use for limiting was a more recent development. In contrast, in Asia, Latin America and the rest of the world, family planning was first adopted to avoid childbearing after a couple had reached its desired number of children, while use for spacing is only beginning to emerge.
Given that the initial adoption of contraception in parts of the world other than Sub-Saharan Africa was for limiting rather than for spacing, it is not surprising that little change in this balance has been evident in Asia and North Africa and in Latin America and the Caribbean. In both regions, the great majority of users are limiters (Tables 2 and 3).‡‡ It might be expected that as more of these countries move into the later stages of the fertility transition, the proportion using contraceptives for spacing purposes would begin to increase as a result of an interest in postponing the first birth, as has been evident in the developed countries. Of course, with postponement of marriage and increases in premarital sex, "spacing" behavior will increasingly occur before marriage and would not be apparent in the data presented here, which are confined to married women.
The Broad Picture
A striking feature of the family planning transition, as others have noted in observations of the general fertility transition,8 is the regional diversity in the tempo of the emergence of the family planning norm. The reasons for this diversity, which are beyond the scope of this article, surely include social and economic development, but also certainly transcend this level of explanation. Even without major improvements in economic status or in the education of women, the idea of fertility limitation can "catch on" and spread through social networks9 and population mobility and through the mass media,10 and can be significantly augmented by family planning programs that have strong commitment from governments.
In Figures 1 through 4, we summarize with crude averages the regional patterns and trends of the empirical data assembled here. The evidence documents the spread of the demand for fertility limitation in a large number of developing countries over the period from the 1970s to the 1990s. It is important to understand that unlike the preceding discussion, in which we traced trends over time in the same countries, the averages for the three time periods are based on different groupings of countries.
The total potential demand for limiting (the sum of contraceptive use and unmet need for limiting) increased virtually in all three regional groupings and shows unmistakable signs of taking off in Sub-Saharan Africa, while beginning to level off in the other two regions (Figure 1). Over the entire period covered in this article, the total potential demand is considerably higher in Asia and North Africa and in Latin America than in Sub-Saharan Africa. However, the relative percentage-point increase in demand is largest in Sub-Saharan Africa (the only region where both current use and unmet need are increasing simultaneously).
The actual use of contraceptives for limiting is increasing almost universally in all regions of the developing world (Figure 2). On the other hand, the unmet need for family planning for limiting purposes shows sharp increases in Sub-Saharan Africa and sharp declines in the other regions (Figure 3). If evidence prior to 1960 were available, Figure 3 would show an overall inverted U shape for Asia and Latin America, with unmet need then at the approximate level of that for Sub-Saharan Africa in the 1970s. Thus, unmet need is a moving target, rising in the early stages of the transition as interest in family limitation grows, and declining in the later stages when family planning use is adopted. With continuing population increase, however, even in these later stages the actual numbers of women in need may not be diminishing and in some countries may still be increasing. Furthermore, even in the developed countries of the world, the proportions in need do not seem to become negligible.11
Figure 4 summarizes trends in the reliance on contraception as used for limiting purposes compared with use for spacing. Starting at a low level in the 1970s, the proportion of contraceptive users who were using the method for limiting increased steadily over the entire period in Sub-Saharan Africa. In Latin America, this proportion was very high, and an initial period of increase (which occurred up until the 1980s) was followed by a period characterized by little or no change. In Asia and North Africa, this proportion was very high and remained relatively constant between the 1970s and 1990s.
It is clear from the foregoing that demand for contraception is increasing throughout the developing world. One major difference between the three regional groupings considered in this article is that while most of the demand is already being met in Asia and North Africa and in Latin America and the Caribbean, the reverse is the case in sub-Saharan Africa. However, the experiences of Kenya and Zimbabwe, two of the few countries at the forefront of the fertility transition in the region, suggest that this difference may not be expected to continue indefinitely. The trends in contraceptive demand and its components (met and unmet need) in both of these countries are similar to the patterns typical of the other two regions.
An implication of this finding is that policies and programs aimed at promoting contraceptive use need to be strengthened, particularly in Sub-Saharan Africa, where contraceptive use is currently low and where unmet need is rising. Women in that region have identified the opposition of others (especially husbands) to contraceptive use and fear of side effects as the main reasons for not practicing contraception.
Although it is now generally agreed that men ought to be involved in family planning programs, the means of accomplishing this are still a matter of contention. Policy and program efforts to increase contraceptive adoption and continuation must be directed toward eliminating or at least mitigating these obstacles. Such efforts should encourage men to desire fewer children and to support their wives' desire to use modern and effective contraceptives. Policies should promote understanding and cooperation between spouses and sexual partners, and should emphasize that family planning helps them to have only the number of children they desire and contributes to the health of women and children.
In addition, family planning programs should endeavor to provide women and men a more appropriate environment in which to seek services, to help clients choose methods that are suitable for them and to supply information that will improve users' confidence in their methods and in their ability to use them successfully. In Asia and in Latin America and the Caribbean, where contraceptive use is already high and unmet need is declining, service providers should pay more attention to helping users to continue use and to practice contraception more effectively.