Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 117
Demographic Change in Sub-Saharan Africa 4 Recent Trends in Marriage Ages Etienne van de Walle INTRODUCTION The aim of this chapter is to review recent data on nuptiality in sub-Saharan Africa (Africa in short) and assess the demographic evidence of changes over time in the proportions married and the age at marriage. Demographers have been interested in nuptiality mostly because of its possible implications for fertility. In a classic article, Davis and Blake (1956) included the factors governing the formation and dissolution of unions in the reproductive period among the intermediate variables affecting exposure to intercourse. In their review of the proximate determinants of fertility, Bongaarts and Potter (1983:4) defined marriages as “relatively stable sexual unions” to which “socially sanctioned childbearing” is limited in most societies. As a consequence, fertility surveys that have featured so prominently in recent demographic research in Africa, particularly the World Fertility Surveys (WFS) and the Demographic and Health Surveys (DHS), have included questions on marriage. But the nature of the data sources and the actuality of the concern about fertility in Africa should not lead us to forget that marriage has long played a major role in the studies of anthropologists and Etienne van de Walle is a professor at the Population Studies Center, University of Pennsylvania.
OCR for page 118
Demographic Change in Sub-Saharan Africa sociologists because of its role in shaping descent systems and social organization.1 Much of the qualitative information collected by anthropologists, with its rich institutional and cultural context, is overlooked in the statistical studies pursued by the demographer. Marriage patterns in their own right nonetheless constitute an important topic of study for the student of population because they associate many socioeconomic, cultural, and demographic variables at the individual and societal levels. The evidence on nuptiality change could further the understanding of other social change. For example, the changing frequency of the types of unions that occur in a society (customary or civil marriages with full social recognition versus informal or temporary unions) may influence the prevalence of female-headed households and the economic environment of children. Even from the narrow perspective of the demographer, the type of union or the active involvement of a man in a household may affect infant mortality; a plausible mechanism is through the lesser access to resources by single mothers or wives of polygynists.2 Moreover, mating patterns play a major role in the transmission of the human immunodeficiency virus (HIV), a major killer that will influence the size, growth, and distribution of African populations. For better or worse, the nature of the sources will affect the analysis of these issues, and hence understanding of them. The vast body of ethnographic descriptions of African marriage presents a complexity of peoples, perspectives, and time periods, making interpretation difficult in a comparative context. Nuptiality data are most rich, detailed, and useful when they are provided on one particular subpopulation of a country—for example, the Yoruba of Nigeria, the Mbeere of Kenya, or even the Creoles of Freetown. Recent fertility surveys, however, have stressed international comparability and reduced the concepts to the simplest common denominator. Censuses, which represent a major source of information on marriage because of their wide coverage, have also used simple definitions (but not necessarily the same ones as the surveys). The task, then, is to look at the available material for the purpose of ascertaining the evolution of simple indices of nuptiality in recent times. The fertility perspective will dominate, but cannot be exclusive of other concerns. For example, when one examines the fertility implications of recorded changes in nuptiality, the conclusion is that age at marriage has risen in many countries of sub-Saharan Africa, but that this trend appears to 1 The comparative insignificance of childbearing in the eyes of most anthropologists is obvious in a recent publication on the evolution of marriage in Africa by Parkin and Nyamwaya (1987). 2 Adegbola (1987) has raised the issue of the relationship between premarital fertility and infant mortality; see also Lesthaeghe et al. (1989:329–330).
OCR for page 119
Demographic Change in Sub-Saharan Africa have little relationship to any change in fertility; the proportions single increase sharply in the age group 15–19, but the number of children ever born at the same ages often changes little, if at all. It is likely that the proportion of unmarried mothers is increasing—a change that may have important demographic and social consequences—but not in relation to total fertility. Excessive concentration on the fertility aspects of age at marriage would lead us to lose sight of the overall picture. Before reviewing the empirical evidence on nuptiality, it is important to understand some problems concerning the use of these data. The first section of this chapter deals with the consideration of three crucial issues, without which it would be impossible to proceed. These are (1) the definition of marriage, (2) problems of recall and of age reporting encountered in retrospective reports, and (3) the measurement of age at first marriage. These problems are conceptually distinct, but their effects on available data may be difficult to disentangle. The second section is devoted to a look at data sources and findings. The topic of marriage has received a great deal of attention in recent publications (e.g., Lesthaeghe, 1989; United Nations, 1990). Because the results of the 1990 round of censuses have not yet appeared fully, the main new data sets consist of results from the DHS. The section discusses the extent to which retrospective evidence from surveys on the date and age at which individuals report they were married (as contrasted with information on the current marital status of individuals) can be used to evaluate trends in nuptiality. In the third section, I consider the effects of nuptiality changes on fertility. It may be in order to list some of the topics that are not considered here. The discussion is limited to the marital status of women for two reasons. First, the issues raised by the nuptiality of men are even less well understood than those of women. Men marry later and spend a sizable portion of their adult life in the single state, although they probably lead in most cases a complex sexual life that does not appear in the statistical record. Moreover, in Africa, polygyny is a typical feature of men’s married life, which dilutes the connection between their nuptiality and reproduction.3 Second, the new sources on nuptiality (the WFS and DHS, discussed below) pay little attention to the marriage of men. Census data have provided the available information on men, and I have not tried to go beyond the available monographs on the subject (United Nations, 1988, 1990). 3 For what a detailed investigation of male nuptiality would entail, see a suggestive study of South-Benin by Donadjè and Tabutin (1991). With an average age at marriage of 28 (compared to less than 20 for women), men are reported to have a total fertility of 11 children (compared to between 6 and 7 for women) because more than half end up in polygynous unions.
OCR for page 120
Demographic Change in Sub-Saharan Africa The next topic to which I cannot do justice, despite its interest, is the topic of premarital sexual activities of women. Informal unions may be fruitful and may also represent an early stage in the contracting of a marriage. They complicate the retrospective definition of unions in demographic surveys, as I point out, but to study them in depth, the anthropological approach is essential. It may be that informal unions and premarital or even illegitimate births4 have become more frequent, but in the absence of reliable information, one confronts “the interpretive problem of sensing that marriage may be more informal than in the past without having any clear ‘data’ or clear chronology to support such a conclusion” (Guyer, 1988:1). Finally, in the absence of new data from the last round of censuses, there is little information that would allow review of the conclusions made in the authoritative review of Lesthaeghe et al. (1989) on the subject of polygyny.5 The institution remains alive and well, although its multiple forms, which owe some of their complexity to the ambiguity of the definition of what a stable union is, challenge the analyst. CONCEPTUAL AND MEASUREMENT ISSUES Definitions The subject of nuptiality is complex and is governed by different rules and practices in different countries. It has long been accepted that the description of marital status in censuses can be kept simple because people are reasonably certain about their own present status when an interviewer asks questions; being “married” corresponds to a social reality recognizable in almost any culture, so that there is no need for elaborate definitions. There are clearly diminishing returns to adding questions or attempting to narrow concepts. However, the particular reality that people recognize as the married state is by no means uniform across societies. A corollary is that one does not always know what reality is covered by conventional census categories (single, married, widowed, and divorced), and it is always possible that the changes apparent over time from the comparison of several surveys and censuses taken at different dates are more changes of implicit (the population’s) or explicit (the census taker’s) definitions than changes 4 By illegitimate births, we mean those births that have no socially recognized legitimacy. The term illegitimate is controversial (see Adegbola, 1987), but legitimacy of offspring is a recognized goal of marital unions. 5 Their observations spanned a period roughly between 1960 and 1980.
OCR for page 121
Demographic Change in Sub-Saharan Africa in the underlying reality. Also of concern is the inverse danger that an apparent stability of some indicators of nuptiality in fact hides deep changes in the sociological reality. Lesthaeghe et al. (1989:244) remark that …“outside wives” are more akin to concubines than to women in a polygynous union since they are external and often illicit. But demographically they are of significance as an alternative mode of reproduction in societies undergoing socioeconomic change. How they are defined in a census or survey will affect measures of trends and the impact of economic change on these trends. Ever since censuses have been taken in sub-Saharan Africa, they have included a simple typology of marital status (single, married, widowed, or divorced), based on self-definition by the respondents. In this rather coarse net, a variety of people are caught as “married,” including some who have not performed any ceremony of marriage or whose unions have not been formally recognized by society. It is likely, however, that what pass for marriages in a census are relatively stable unions, benefiting from a degree of public recognition; in many instances, cohabitation is also involved, although there are forms of marriages in Africa, particularly among matrilineal groups, in which cohabitation is not essential (see, for example, van de Walle and Meekers (1988) for Côte d’Ivoire). Generally, marriage in Africa is “a process,” and therefore there is some ambiguity in determining exactly when a couple is married. The ambiguity is less critical in a census than in a fertility survey where retrospective questions are asked about age at, or date of, marriage. A retrospective question of the type—At what age were you first married? —is different in nature from the assessment of the current marital status of an individual at the time of a survey or census. Even if perfect recall by the respondent is assumed, identification of the date (by month and year) when a woman’s first conjugal union started is conceptually difficult. A number of unions that have some features of marriage, and might have been reported as such at the time they were still extant, may retrospectively appear never to have taken place; also, some unions that turned out to be successful would not have been reported as “marriages” in their early stages, but with the benefit of hindsight appear to have started in earnest at a time when their status was actually quite uncertain. Such systematic misreporting introduces biases in the a posteriori reporting of age at marriage in surveys. The WFS and DHS surveys have introduced new concepts in the description and measurement of African marriage. Designed as fertility surveys, they have focused on one aspect of the marital state—exposure to sexual intercourse. For female respondents in surveys, being married was held by the designers of the WFS and the DHS to be synonymous with “living in union” or “living with a man”; the fact of cohabitation was the
OCR for page 122
Demographic Change in Sub-Saharan Africa single most important criterion distinguishing a marriage for these surveys.6 The concept is not exactly comparable with the commonly accepted (but somewhat imprecise) category of married in a census, and comparisons of these surveys with censuses taken at the same time usually reveal differences that must be due to the definitions used. Moreover, differential coverage may account for some of the difference between census and survey results; censuses may be better at capturing young adults than surveys. Yet, because censuses and surveys are usually not taken exactly in the same year, the differences between the two sources in the same countries have sometimes been attributed to changes in nuptiality. Table 4–1 presents the comparison of the percentage ever married at ages 15–19, 20–24, and 40–44 in the two kinds of sources. The expectation is that surveys should find more women living in union than there are married women in the census, and typically fewer widows and divorcees. Blanc and Rutenberg (1990:53) give the following rationale: It is expected that the retrospective estimates of the proportion of women ever married calculated from DHS data will be higher than the estimates from previous censuses or surveys for three reasons. First,…censuses often use a less inclusive definition of marriage than that of the DHS surveys…Second, information on marital status and date of marriage in DHS surveys usually comes from the individual questionnaire, for which the respondent is a woman, rather than from the household questionnaire, for which the respondent is often a male head of household. A third factor which might act to improve the validity of estimates from DHS surveys, relative to earlier censuses and surveys, is that the quality of reporting of …marriage dates may have improved in recent years… Surprisingly, Table 4–1 does not conform to these expectations when one compares direct estimates based on DHS and census data for each country. At age 15–19 (and where available, age 20–24), the proportions ever married were higher in the census than in the DHS for Botswana, Burundi, Kenya, Togo, Uganda, and Zimbabwe; it is only at 40–44 that the DHS always finds a higher proportion of ever-married women. Because all the DHS surveys in Table 4–1 have a later date than the census, part of the difference could be explained by a time trend, either in the age at marriage or in the accuracy 6 According to the DHS Interviewer’s Manual (Institute for Resource Development, 1987:59): “Lived with a man” means that they stayed together for some time, intending to have a lasting relationship, regardless of the formal status of the union…. For example, if a woman went to live with her boyfriend and his family, and stayed for several years, she would be considered as “living together”, whether or not the couple had any children. On the other hand, if a woman had a boyfriend for a year but never lived with him, she would not be considered as ever having married or lived with him.
OCR for page 123
Demographic Change in Sub-Saharan Africa TABLE 4–1 Comparison of Women Ever Married (percent) Age Country Source Date 15–19 20–24 40–44 Botswana DHS, Ra 1981 13.1 42.2 82.8 Census 1981 7.3 31.2 78.9 DHS 1988 6.1 30.3 81.5 Burundi DHS, Ra 1979 21.7 76.3 98.7 Census 1979 19.2 72.6 97.4 DHS 1987 6.8 66.7 99.1 Kenya Census 1979 28.8 – 97.9b DHS 1989 20.1 68.2 98.5 Liberia DHS, Ra 1984 38.1 77.0 98.4 Census 1984 35.7 70.9 96.4 DHS 1986 36.0 75.3 98.3 Mali DHS, Ra 1976 66.2 95.9 100.0 Census 1976 51.1 88.0 98.0 DHS 1987 75.4 98.0 99.7 Senegal DHS, Ra 1976 52.5 87.1 100.0 Census 1976 38.6 76.1 97.5 DHS 1986 43.5 77.4 100.0 Togo DHS, Ra 1981 40.1 81.8 100.0 Census 1981 43.3 81.8 97.2 DHS 1988 27.2 75.8 99.6 Uganda Census 1969 49.9 86.5 93.9b DHS 1988–1989 40.8 83.0 99.0 Zimbabwe DHS, Ra 1982 29.9 80.7 98.7 Census 1982 26.1 76.5 97.0 DHS 1988 19.8 71.5 99.1 NOTE: Data are from national-level census and DHS. aR: Reconstruction for census date. bAt age 50. SOURCE: Data from Blanc and Rutenberg (1990: Table 2.5) for reconstruction and census proportions; United Nations (1990) for census data on Kenya in 1979 and Uganda in 1969; otherwise, Lesetedi et al. (1989); Segamba et al. (1988); Kenya (1989); Chieh-Johnson et al. (1988); Traore et al. (1989); Ndiaye et al. (1988); Agouke et al. (1989); Kaijuka et al. (1989); and Zimbabwe (1989). of reporting. Blanc and Rutenberg tried to remove the effect of the change over time by reconstructing from DHS data the proportion ever married at the date of the previous census; in doing so, they had to rely heavily on the retrospective reporting of dates of union, which I argue here is incompatible with current status data. I show the results of their reconstruction in Table
OCR for page 124
Demographic Change in Sub-Saharan Africa 4–1 (labeled DHS, R). The reconstructed proportions ever married are usually substantially higher than those recorded in the census, but they are also much higher than those at the actual time of the survey. This disparity suggests a severe bias in the proposed reconstruction. A likely explanation of the difference between current marital status as measured at the times of surveys and censuses is that the criterion of cohabitation, which is part of the DHS definition of a union (as of the time of the interview), excludes many people who would have defined themselves to a census taker as married. It is true that surveys also may include people living in unions who are not viewed as married by the census, but their number is smaller than that of the noncohabiting married who are excluded. Other reporting biases in the retrospective reporting of dates or ages at the onset of the first union vitiate their use in the comparison by Rutenberg and Blanc. Despite the general principle of looking at unions (defined by the criterion of cohabitation) rather than at formalized marriages, there was a remarkable diversity in the phrasing of the WFS questions, and sometimes they were different enough to yield quite different results in neighboring countries with similar populations, such as Ghana and Côte d’Ivoire. Additional diversity must have resulted from the use of local languages in the interviewing. The DHS has standardized questions to a greater extent than the WFS, but here too, there is some variety in the phrasing, which may reflect diversity in the legal definition of marriage in particular countries, or even in the ideologies of survey takers. It is interesting, for example, that all the DHS surveys in French-speaking countries (with the exception of Togo) have more or less denied the existence of consensual unions, whereas all the English-speaking countries (with the exception of Zimbabwe and Sudan) recognized their importance by giving them a separate column (headed “living together”) in the published survey reports.7 Among the DHS reports for French-speaking countries, the one for Senegal stated explicitly that “marriage remains the only socially accepted framework for sexual links” (Ndiaye et al., 1988:13); in Burundi, “cases of concubinages are rarely declared as being marriages” (Segamba et al., 1988:17); in Mali (Traoré et al., 1989), the survey categories appeared “very ambiguous” to the authors of the report, and married women were classified in the report with women “living with someone.” In contrast, among English-speaking countries, the Liberia (Chieh-Johnson, et al., 1988) DHS seems to have elicited an extraordinary proportion of responses qualifying the union as consensual: 38 percent of women between 15 and 49 “lived together,” whereas only 29 7 Zimbabwe and Togo adopt a similar solution toward consensual unions: they classify them with marriage, not because they are unimportant but because they are difficult to distinguish.
OCR for page 125
Demographic Change in Sub-Saharan Africa TABLE 4–2 Never-Married Females in Botswana by Age (percent) —Successive Censuses and Surveys Census Family Health Survey Age 1971 1981 1984 1988 15–19 87 93 47 94 20–24 56 69 4 70 25–29 37 47 0.5 43 30–34 27 32 0.4 30 35–39 20 25 0.1 25 40–44 17 21 0.0 19 45–49 13 17 0.0 20 SOURCES: Botswana (1985); Lesetedi et al. (1989). percent were “married”; in Botswana, the report stated that “a union is not prerequisite to childbearing” and “current marital status…does not take into account the large proportion of women in stable relationships that do not involve cohabitation” (Lesetedi et al., 1989:11). Rather small differences in phrasing of the survey or census questions can yield extraordinary differences in the proportions ever married. For Botswana, there are two censuses and two Family Health Surveys (FHS) in the 1970s and 1980s, providing an opportunity to examine trends in the proportion married over time. From the 1971 and 1981 censuses, and the 1988 FHS (taken under the DHS program), the age at marriage seems to be relatively late in Botswana, the proportions who never marry are high for an African country, and there is a trend toward later marriage. However, the 1984 FHS breaks the trend, giving an opposite impression of early and universal marriage (see Table 4–2). In comparing the two FHS surveys, the main difference appears to be between those “in a consensual union” in 1984 (52.3 percent of women aged 15–49) and those “living together” in 1988 (10.8 percent). In 1984, the operative survey question on consensual unions was, Have you ever had a partner? In 1988, the criterion of cohabitation was introduced: Have you ever been married or lived with a man? Marriage in Botswana still involves the payment of bridewealth (Timæus and Graham, 1989:373), a prolonged and cumbersome process. Heads of cattle that are part of the bridewealth are an important factor of production in Tswana agriculture (Peters, 1986), and marriage is an index of the economic viability of a household. On the other hand, living with a man in a country where half of the women currently in union reported that their partner was not living with them at the time of the 1984 survey (Botswana, 1985:48) is probably not a good indication of the stability of a relationship or of its likelihood to bring forth
OCR for page 126
Demographic Change in Sub-Saharan Africa offspring. If the aim is to provide a denominator of “women in union” for the study of fertility, it could be argued that the census definition (or its 1988 FHS extension) is unduly restrictive. A convincing case could probably also be made that the 1984 FHS definition was too inclusive. Marriage, as contrasted with temporary and unstable free unions, remains an important concept in Botswana. Timæus and Graham argued that late marriage was an important feature of the demography of Botswana, and one that had been established for some time. They acknowledged, however, that the effect of late marriage on fertility was “less than one might expect” (Timæus and Graham, 1989:381). It is probable that marriage has important social consequences for women and men, and that the permanence of the institution of the bridewealth is connected more to some aspects of social organization than to the biological reproduction that keeps demographers busy and conditions survey design. A classic article by Comaroff and Roberts (1977; see also Caldwell et al., 1991) has examined some of the conflictual aspects of marriage in Botswana, and the different interests served for men and women. Informal relationships may be exploitative of women (Timæus and Graham, 1989:381), and their consequences for the welfare of children and the structure of society are worthy subjects of study. It is possible, however, that historical changes in the timing and prevalence of marriage had little influence on fertility. Botswana may appear to be an extreme case in the demography of marriage in Africa, and the very high proportions of never-married women recorded by the 1988 Family Health Survey (see Table 4–2) are not replicated elsewhere. But there are many indications from other countries that the criterion of cohabitation may not be the sine qua non identifier of a union. In Lomé, Togo, for example, the APEL (Arrivée du Prochain Enfant à Lomé) survey of 1983–1984 found that 31 percent of women in union were living in a different residence than their mates (Ekouevi, 1992). A preliminary conclusion should be that the haziness of the concepts will make comparisons over time and space hazardous. Problems of Recall and of Age Reporting In addition to the problem of definition discussed in the previous section, retrospective questions of the kind included in the WFS or the DHS— In what month and year did you start living with your first husband? — suffer from a fundamental weakness when asked in societies where people do not know their ages; moreover, it is possible that older women tend to “forget” the existence of earlier unions or to edit them out in their reports. The question continues to be asked in surveys, and in particular was asked in both the WFS and the DHS; it is therefore worthwhile to look at the
OCR for page 127
Demographic Change in Sub-Saharan Africa results. Taken at face value, the information may provide interesting information about trends (inferred from differences in the ages at first marriage reported by successive cohorts) and about differentials between ethnic or social groups. The raw material used here is drawn from the DHS.8 All of the country reports (except for Botswana) contain a table that presents reported age at first marriage or union, classified by age of the respondent at the time of the survey. Several of the DHS reports comment on the reliability of the evidence. The report for Burundi notes that women appear to have a relatively precise knowledge of their date of marriage: in the survey, 58 percent gave the month and year of occurrence. Similarly, in Zimbabwe, 77 percent of the women were able to do so; the report notes, however, “a tendency on the part of some women to report the date (age) when the marriage was officially registered rather than the date (age) when the couple first began living together” (Zimbabwe, 1989:18). In Ghana, only 29 percent of ever-married women reported both a month and a year of first marriage; the report notes: “In addition to the difficulty in correct dating of events, the formalization of marriage itself may span a number of years” (Ghana, 1989:11). These problems are akin to problems of definition addressed in the previous section. If reported dates indeed correspond to a stage of formalization or to a date of registration, the reporting of age at marriage might be biased upward. The staff of the DHS has addressed the issue of the quality of data on age at first marriage (Blanc and Rutenberg, 1990). Table 4–3 gives the percentage of ever-married women by age who reported their date of first union by year and month, and for whom no imputation was necessary. The information is particularly deficient in West African countries but has been improving in recent cohorts. In Mali, for example, less than 10 percent of women were able to provide month and year of first union, and it is hard to place too much reliance on the precision of the ages given and the reliability of trends. In Burundi, Uganda, and Zimbabwe, the percentages of unions with precise dates are much higher, perhaps because these correspond to recorded Christian ceremonies. Incidentally, the quality of reporting is related to the level of education of the respondents, which raises the question of whether apparent trends in age at marriage that characterize cohorts or educational groups are genuine or are at least partly the result of better reporting. The evidence on reported age at marriage in the DHS is most interest- 8 For a review of this evidence in the WFS, see Lesthaeghe et al. (1989:245). They express skepticism about the apparent trends in cohort comparisons based on a single retrospective survey.
OCR for page 142
Demographic Change in Sub-Saharan Africa FIGURE 4–1 Percentage married and children ever born (CEB) for women aged 20– 24 (DHS data). SOURCES: Botswana (1985); Segamba et al. (1988); Ghana (1989); Kenya (1989); Chieh-Johnson et al. (1988); Traoré et al. (1989); Nigeria (1992); Ndiaye et al. (1988); Sudan (1991); Agounké et al. (1989); Kaijuka et al. (1989); Zimbabwe (1989). FIGURE 4–2 Percentage married and childless for women aged 20–24 (DHS data). SOURCES: Botswana (1985); Segamba et al. (1988); Ghana (1989); Kenya (1989); Chieh-Johnson et al. (1988); Traoré et al. (1989); Nigeria (1992); Ndiaye et al. (1988); Sudan (1991); Agounké et al. (1989); Kaijuka et al. (1989); Zimbabwe (1989).
OCR for page 143
Demographic Change in Sub-Saharan Africa FIGURE 4–3 Never-married and childless women, ages 15–19 (in percent). SOURCES: Botswana (1985); Segamba et al. (1988); Ghana (1989); Kenya (1989); Chieh-Johnson et al. (1988); Traoré et al. (1989); Nigeria (1992); Ndiaye et al. (1988); Sudan (1991); Agounké et al. (1989); Kaijuka et al. (1989); Zimbabwe (1989). FIGURE 4–4 Singulate mean age at first marriage and nulliparate mean age at first birth (DHS data). SOURCES: Botswana (1985); Segamba et al. (1988); Ghana (1989); Kenya (1989); Chieh-Johnson et al. (1988); Traoré et al. (1989); Nigeria (1992); Ndiaye et al. (1988); Sudan (1991); Agounké et al. (1989); Kaijuka et al. (1989); Zimbabwe (1989).
OCR for page 144
Demographic Change in Sub-Saharan Africa approximately equal numbers of countries are found on either side of the 45-degree line that would describe the relationship if age at first marriage and age at first birth were the same; one would expect marriage to precede the first birth by a number of months if it truly marked the start of exposure. However, births occur in the absence of marriage, and marriages are not always followed immediately by births. Changes in Proportions Married and Children Ever Born A test of changes would be the comparison of proportions ever married and children ever born over time. If age at marriage reflected the beginning of exposure to the risk of childbearing, its changes should be related to those recorded in the number of children ever born (CEB) at young ages. Kenya offers the longest series, with a steady increase in the proportion single at the youngest ages, but little change in the number of children ever born (Table 4–11). Over time, the age at marriage has been going up in Kenya, but the exposure to the risk of childbearing appears to have remained almost constant (it is too early to say whether the drop in CEB in 1989 is a real change). CONCLUSION Several conclusions emerge from this review of recent data on age at marriage and the proportions married. The first is that the tools of investigation remain inadequate. Recent surveys have collected detailed and sophisticated information on nuptiality with the clear rationale of using it in a TABLE 4–11 Proportions Single and Children Ever Born, Kenya Proportion Single (%) Children Ever Born Data Source 15–19 20–24 15–19 20–24 1962 census 55 13 0.4 1.7 1969 census 64 18 0.4 1.9 1977 National Demographic Survey 71 22 0.3 1.8 1977–1978 Kenya Fertility Survey 72 21 0.4 1.8 1979 census 71 25 0.3 1.9 1984 Kenya Contraceptive Prevalence Survey 74 24 0.4 2.0 1989 Kenya Demographic and Health Survey 80 32 0.3 1.6 SOURCE: Kenya (1989:10, 25).
OCR for page 145
Demographic Change in Sub-Saharan Africa subordinate position in the study of fertility. But the links between fertility and nuptiality have proved tenuous and variable in sub-Saharan Africa. There are tantalizing suggestions that nuptiality has changed in recent years. Whether the change is purely structural, because more women live in cities and have gained schooling (factors that tend to delay marriage), or whether the changes represent a profound transformation of the patterns of early and universal marriage that affect the entire population is a question that cannot be settled easily with the data at hand. The changes are certainly linked with deep transformations in the African family and are accompanied by, or perhaps in part caused by, increasing female independence inside and outside of unions. (The extensive literature on this subject includes Little, 1973; Burnham, 1987; Guyer, 1988; Locoh, 1988; Obbo, 1981; and many others.) Various authors have discussed the recent evolution of African marriage in negative terms in sharp contrast with earlier interpretations based on convergence and modernization theory (Goode, 1970). Thus, Caldwell et al. (1991) talk about the “destabilization of the traditional sexual system”; Frank and McNicoll (1987) discuss the “caribbeanization” of African nuptiality. The same authors advance the hypothesis that new types of female-headed households that are emerging may offer promising opportunities for the deliberate control of reproduction. The debate on African nuptiality is likely to be increasingly cast in terms either of resilient adaptation by the African family to the forces of socioeconomic change or of social pathology; the HIV epidemic will feature prominently in this debate. There have been speculations on the possible consequences of the new nuptiality patterns for fertility through variables other than exposure to the risk of pregnancy. Although the direct effect of a later age at marriage on total fertility by curtailing the period of exposure is probably at work in such countries as Burundi or the Sudan, where rules against premarital sex have remained strong, the effect of a delay in age at marriage on fertility may be more subtle in most other countries, including Kenya, Botswana, and Togo. Van de Walle and Foster (1990) suggested that premarital sexual relations may constitute a training ground for the use of birth control, because young women want to avoid the pregnancies that would jeopardize their prospects of education and jobs. The acquired knowledge of techniques of contraception and abortion would later be carried over to marriage. Other mechanisms may link the types of sexual unions to the duration of sexual abstinence or breastfeeding, to infecundability and male sterility, and to the prevalence of sexually transmitted diseases. This chapter opens by stressing the fact that there are demographic topics other than fertility for which nuptiality patterns and intrahousehold relationships may be crucial: Infant mortality and the transmission of AIDS are obvious examples. There are good reasons to continue to investigate
OCR for page 146
Demographic Change in Sub-Saharan Africa nuptiality in surveys in view of its epidemiologic and social importance. More reflection will be needed on the best ways to ask questions about nuptiality and to interpret the answers. The next task will be to continue to analyze the nuptiality data from fertility surveys and see how they can be improved in future data collection efforts. Moreover, the publication of censuses from the 1990 round will provide fresh information to help document the changes that are thought to be taking place. It has been noted that the Tanzanian data suggest the old norm of universal female marriage may be changing. This finding will have to be confirmed in subsequent research, and its meaning ascertained. The simple questions used in conventional censuses (the “Are you married?” type) and the more sophisticated questions in fertility surveys (Have you ever been married or lived with a man?) cannot go very far when one tries to ascertain changes over time or to capture the qualitative diversity of the types of unions or sexual relationships that exist. The statistical record is weak or void when one attempts a typology of types of union or tries to distinguish free unions and informal relations from more durable marriages. To make sense of the evolution here, there is a need to resort to other methodologies or to the descriptions of other social scientists. Their observations do not range as wide as the representative samples characterizing the sociological survey; what they gain in depth, they also lose in breadth. More work on linking anthropological or ethnographic study findings with statistical data collection in large sample surveys could greatly enhance knowledge of marriage patterns and their consequences in sub-Saharan Africa. APPENDIX: MEDIAN AND MEAN AGE AT FIRST MARRIAGE Computation of the median age at marriage may, or may not, take into account those women who will never marry, or, for practical purposes, have not married before a given age, such as 50, after which very few marry for the first time. In the estimates published by the WFS or the DHS, the median is the age at which half of the women in a cohort are married. The difference is illustrated by considering Kenya and Botswana. In Kenya, almost all women go on to get married; such is not the case, however, in Botswana. The comparison between the two countries appears in Table 4– A.1, which uses the proportions of currently never-married women in the DHS. If one treats these proportions as a cohort of women aging through an unchanging nuptiality schedule, the respective median ages at marriage for Kenya and Botswana, obtained by interpolation of the proportions in Table 4–A.1, would be 20.6 and 26.2 years. If one limits the computation of the median, however, to those women who will ultimately marry (98.5 percent
OCR for page 147
Demographic Change in Sub-Saharan Africa TABLE 4–A.1 Proportions Never Married Women, Kenya and Botswana (percent) Age Kenya, 1989 Botswana, 1988 15–19 79.8 93.9 20–24 31.8 69.7 25–29 10.7 43.3 30–34 5.4 30.4 35–39 3.2 25.1 40–44 1.5 18.5 45–49 2.4 20.2 NOTE: Data are national-level DHS. SOURCES: Data from Lesetedi et al. (1989); Kenya (1989). in Kenya and 81.5 percent in Botswana), the question becomes: When would half of the 98.5 percent of Kenyans who ever get married (or 49.3 percent) and half of the 81.5 percent of those in Botswana (or 40.8 percent) be married? The medians become 20.5 years for Kenya and 24.5 for Botswana. This seems a more legitimate computation of the median (see Shryock and Siegel, 1976:166). It cannot be used, however, before one knows the proportion of women who are likely ever to marry in a cohort and is, therefore, not always practical for estimation of the age at first union of young women reporting in a survey. Limiting the discussion to Kenya, where most women marry, note that the two computations of the median yield very similar results and that these results in turn are close to the singulate mean age at marriage, 20.5, which is also computed on the basis of current marital status of women at the time of enumeration. The published WFS and DHS estimates of the median age of marriage by cohort are based on retrospective reporting of the ages at which women were first married, not on current marital status as in the previous comparison. In Table 4–A.2, I have systematically computed the mean age at marriage by cohort for the DHS, by assuming that women were all married in the middle of the age group at which they reported their marriage. The computation uses all the information available and is likely to be less affected than the median by the concentration of marriages at the younger ages; its value should therefore, on the whole, tend to be a bit higher. It is obvious, however, that estimates of the means are generally close to estimates of the median based on the same retrospective data. This exercise suggests that medians and means based on the same type of data tend to be similar, but that retrospective and current status reports differ systematically.
OCR for page 148
Demographic Change in Sub-Saharan Africa TABLE 4–A.2 Mean Age at First Marriage of Women Age Burundi Ghana Kenya Liberia Mali Senegal Togo Zimbabwe 20–24 19.1 17.9 18.3 17.3 16.5 16.9 17.7 18.3 25–29 19.5 18.6 18.3 18.1 16.8 17.4 18.4 18.8 30–34 19.5 18.5 18.1 17.9 16.4 17.6 18.3 18.7 35–39 20.1 18.4 18.1 19.8 16.4 17.3 18.8 19.8 40–44 19.8 18.1 17.8 17.3 16.6 17.2 18.5 18.4 45–49 19.9 18.6 18.7 18.3 16.9 17.0 19.0 19.0 Mean of the means 19.7 18.5 18.2 18.2 16.6 17.4 18.5 18.8 Mediana 19.5 18.1 18.2 17.2 15.7 16.4 18.5 18.7 NOTE: Data are national-level DHS. aMedian value of the medians for each group computed from Table 4–5. SOURCES: Data from Segamba et al. (1988); Ghana (1989); Kenya (1989); Chieh-Johnson et al. (1988); Traoré et al. (1989); Ndiaye et al. (1988); Agounké et al. (1989); and Zimbabwe (1989).
OCR for page 149
Demographic Change in Sub-Saharan Africa REFERENCES Adegbola, O. 1987 An appraisal of non-marital births in Nigeria. In The Cultural Roots of African Fertility Regimes. Ile-Ife: Department of Demography and Social Statistics; Philadelphia: Population Studies Center. Agounké, A., M.Assogba, and K.Anipah 1989 Enquète Démographique et de Santé au Togo 1988. Lomé: Unité de Recherche Démographique, Direction de la Statistique, Direction Générale de la Santé; Columbia, Md.: Demographic and Health Surveys, Institute for Resource Development/Macro Systems, Inc. Blanc, A.K., and N.Rutenberg 1990 Assessment of the quality of data on age at first sexual intercourse, age at first marriage, and age at first birth in the Demographic and Health Surveys. Pp. 41–79 in An Assessment of DHS-I Data Quality. DHS Methodological Reports, No. 1. Columbia, Md.: Institute for Resource Development/Macro Systems, Inc. Bleek, W. 1987 Lying informants: A fieldwork experience from Ghana. Population and Development Review 13(2):314–322. Bongaarts, J., and R.G.Potter 1983 Fertility, Biology, and Behavior, An Analysis of the Proximate Determinants. New York: Academic Press. Botswana 1985 Botswana Family Health Survey 1984. Gaborone: Family Health Division, Ministry of Health; Columbia, Md.: Westinghouse Public Applied Systems. Brass, W., A.J.Coale, P.Demeny, D.F.Heisel, F.Lorimer, A.Romaniuk, and E.van de Walle 1968 The Demography of Tropical Africa. Princeton, N.J.: Princeton University Press. Burnham, P. 1987 Changing themes in the analysis of African marriage. Pp. 37–53 in D.Parkin and D.Nyamwaya, eds., Transformations of African Marriage. Manchester: Manchester University Press for the International African Institute. Caldwell, J.C., I.O.Orubuloye, and P.Caldwell 1991 The destabilization of the traditional Yoruba sexual system. Population and Development Review 17(2):229–262. Chieh-Johnson, D., A.R.Cross, A.A.Way, and J.M.Sullivan 1988 Liberia Demographic and Health Survey 1986. Monrovia: Bureau of Statistics, Ministry of Planning and Economic Affairs; Columbia, Md.: Institute for Resource Development/Westinghouse. Comaroff, J.L., and S.Roberts 1977 Marriage and extra-marital sexuality: The dialectics of legal change among the Kgatla. Journal of African Law 21(1):97–123. Davis, K., and J.Blake 1956 Social structure and fertility: An analytic framework. Economic Development and Cultural Change 4:211–235. Donadje, F., and D.Tabutin 1991 Nuptialité et fécondité des hommes au Sud-Bénin. Paper presented at the International Union for the Scientific Study of Population Seminar on Fertility Transition in Sub-Saharan Africa, Harare, Zimabwe. Ekouevi, K. 1992 Family patterns and reproductive behavior in urban Togo. Ph.D. dissertation, Demography Group, University of Pennsylvania.
OCR for page 150
Demographic Change in Sub-Saharan Africa Frank, O., and G.McNicoll 1987 Fertility and population policy in Kenya. Population and Development Review 13(2):209–243. Ghana 1989 Demographic and Health Survey 1988. Accra: Ghana Statistical Service; Columbia, Md.: Institute for Resource Development/Macro Systems, Inc. Goode, W. 1970 World Revolution and Family Patterns. New York: The Free Press. Guyer, J. 1988 Changing nuptiality in a Nigerian community: Observations from the field. Paper presented at the International Union for the Scientific Study of Population Seminar on Nuptiality in Sub-Saharan Africa: Current Changes and Impact on Fertility. Centre Port-Royal, Saint-Lambert-des-Bois, Paris. Hajnal, J. 1953 Age at marriage and proportions marrying. Population Studies 7(2):111–136. Institute for Resource Development 1987 Interviewer’s Manual: For Use with Model “B” Questionnaire for Low Contraceptive Prevalence Countries. Columbia, Md: Institute for Resource Development, Inc. Kaijuka, E.M., E.Z.A.Kaija, A.R.Cross, and E.Loaiza 1989 Uganda Demographic and Health Survey 1988/1989. Entebbe: Ministry of Health; Columbia, Md.: Institute for Resource Development/Macro Systems, Inc. Kenya 1980 Kenya Fertility Survey 1977–1978. First Report, Vol. 1. Nairobi: Central Bureau of Statistics. 1989 Demographic and Health Survey 1989. Nairobi: National Council for Population and Development; Columbia, Md.: Institute for Resource Development/Macro Systems, Inc. Lesetedi, L.T., G.D.Mompati, P.Khulumani, G.N.Lesetedi, and N.Rutenberg 1989 Family Health Survey II, 1988. Gaborone: Central Statistics Office, Ministry of Finance and Development Planning, Family Health Division, Ministry of Health; Columbia, Md.: Institute for Resource Development/Macro Systems, Inc. Lesthaeghe, R.J., ed. 1989 Reproduction and Social Organization in Sub-Saharan Africa. Berkeley: University of California Press. Lesthaeghe, R.J., G.Kaufmann, and D.Meekers 1989 The nuptiality regimes in sub-Saharan Africa. Pp. 238–337 in R.J.Lesthaeghe, ed., Reproduction and Social Organization in Sub-Saharan Africa. Berkeley: University of California Press. Little, K. 1973 African Women in Towns. New York: Cambridge University Press. Locoh, T. 1988 Evolution of the family in Africa. In E.van de Walle, P.O.Ohadike, and M.D. Sala-Diakanda, eds., The State of African Demography. Liège: International Union for the Scientific Study of Population. Ndiaye, S., I.Sarr, and M.Ayad 1988 Enquête Démographique et de Santé au Sénégal 1986. Dakar: Ministère de l’Economie et des Finances, Direction de la Statistique, Division des Enquêtes et de la Démographie; Columbia, Md.: Institute for Resource Development/Westinghouse. Nigeria 1992 Nigeria Demographic and Health Survey 1990. Lagos: Federal Office of Statis
OCR for page 151
Demographic Change in Sub-Saharan Africa tics; Columbia, Md.: Institute for Resource Development/Macro International, Inc. Obbo, C. 1981 African Women. Their Struggle for Economic Independence. London: Zed Press. Parkin, D., and D.Nyamwaya, eds. 1987 Transformations of African Marriage. Manchester: Manchester University Press for the International African Institute. Peters, P.E. 1986 Household management in Botswana: Cattle, crops, and wage labor. Pp. 133–154 in J.Moock, ed., Understanding Africa’s Rural Households and Farming Systems. Boulder, Colo.: Westview Press. Segamba, L., V.Ndikumasabo, C.Makinson, and M.Ayad 1988 Enquête Dèmographique et de Santé au Burundi 1987. Gitega: Ministère de l’Intérieur, Département de la Population; Columbia, Md.: Institute for Resource Development/Westinghouse. Shryock, H.S., and J.S.Siegel 1976 The Methods and Materials of Demography. San Diego: Academic Press. Sudan 1982 The Sudan Fertility Survey, 1979. Principal Report, Volume II. Khartoum: Department of Statistics. 1991 Sudan Demographic and Health Survey 1989/1990. Khartoum: Department of Statistics, Ministry of Economic and National Planning; Columbia, Md.: Institute for Resource Development/Macro International, Inc. Tanzania 1982 1978 Census. Vol. VII: Basic Demographic and Socioeconomic Characteristics. Dar es Salaam: Bureau of Statistics, Ministry of Planning and Economic Affairs. Timæus, I., and W.Graham 1989 Labor circulation, marriage, and fertility in southern Africa. Pp. 365–400 in R. Lesthaeghe, ed., Reproduction and Social Organization in Sub-Saharan Africa. Berkeley: University of California Press. Traoré, B., M.Konaté, and C.Stanton 1989 Enquête Démographique et de Santé au Mali 1987. Bamako: Centre d’Etudes et de Recherches sur la Population et le Développement, Institut du Sahel; Columbia, Md.: Institute for Resource Development/Westinghouse. United Nations 1983 Manual X. Indirect Techniques for Demographic Estimation. (ST/ESA/SER.A/ 81). New York: United Nations Department of International Economics and Social Affairs. 1988 First Marriage: Patterns and Determinants. (ST/ESA/SER.R/76). New York: United Nations Department of International Economic and Social Affairs. 1990 Patterns of First Marriage: Timing and Prevalence. (ST/ESA/SER.R/111). New York: United Nations Department of International Economic and Social Affairs, van de Walle, E., and A.Foster 1990 Fertility Decline in Africa. Assessment and Prospects. World Bank Technical Paper Number 125, Africa Technical Department Series. Washington, D.C.: The World Bank. van de Walle, E., and D.Meekers 1988 Marriage drinks and kola nuts. Paper presented at the International Union for the Scientific Study of Population Seminar on Nuptiality in Sub-Saharan Africa: Current Changes and Impact on Fertility. November 14–17, Paris.
OCR for page 152
Demographic Change in Sub-Saharan Africa van de Walle, E., and F.van de Walle 1988 Woman’s autonomy and fertility. In Conference on Women’s Position and Demographic Change in the Course of Development, Oslo 1988. Liège: International Union for the Scientific Study of Population. Westoff, C.J. 1992 Age at Marriage, Age at First Birth, and Fertility in Africa. World Bank Technical Paper Number 169. Washington, D.C.: The World Bank. Zimbabwe 1989 Zimbabwe Demographic and Health Survey 1988. Harare: Central Statistical Office, Ministry of Finance, Economic Planning, and Development; Columbia, Md.: Institute for Resource Development/Macro Systems, Inc.
Representative terms from entire chapter: