Acta Juridica 13

JOBNAME: Acta Juridica 13 PAGE: 1 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

Changing patterns of marriage and cohabitation in South Africa* DORRIT POSEL† AND STEPHANIE RUDWICK‡

In this study, we investigate recent trends in marriage and cohabitation in South Africa. We use national micro-data to describe how marriage rates diverge sharply by race, with African women far less likely than White women to be ever-married and more likely to be never-married and not cohabiting with a partner. Large racial differences in marital status are evident also among women who are mothers, helping to explain why the majority of African children do not live in the same households as their fathers. We discuss these trends and patterns by reviewing particularly recent research, which suggests that there are economic constraints to marriage, and which explores possible links between widespread support for the custom of bridewealth and low marriage and cohabitation rates among African women, even in the context of childbirth.

I INTRODUCTION In this study we describe recent trends in marriage and cohabitation in South Africa using micro-data drawn from large, nationally representative surveys; we consider some of the implications of these trends; and we review how they have been explained, particularly in recent studies. The data, presented in section II, show that marriage rates are substantially lower among African women than White women, and that this gap in marriage rates has widened over the post-apartheid period. However, African women are as likely as White women to be mothers. Consequently, the marital status of mothers differs sharply by race: African women who are mothers are far less likely to be married than White women who are mothers. Moreover, among single (‘never-married’) mothers, cohabitation rates are lower among African women than White women. As a result, a much larger share of African mothers is nevermarried (and not cohabitating) compared to White mothers. These patterns of union formation have significant implications for the wellbeing of children, because children are considerably less vulnerable to poverty if their mother is married or cohabiting with a partner.

* We thank an anonymous reviewer for helpful comments received. † PhD (Economics) (Massachusetts); Professor and NRF/DST Chair, University of KwaZulu-Natal. ‡ PhD (Linguistics) (UKZN); Researcher, University of Leipzig.

169

JOBNAME: Acta Juridica 13 PAGE: 2 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

170

MARRIAGE, LAND AND CUSTOM

A common explanation for low and falling marriage rates among African women concerns the specific political and economic nature of South Africa’s past, and particularly the effects of apartheid policies on the structure of African families.1 In section III, we review more contemporary research, which associates low marriage rates among Africans with economic constraints to marriage. The economic status of African men has particular significance, given traditional marriage practices that involve the payment of bridewealth by the prospective husband to the family of the bride.2 Widespread support for the custom of bridewealth (lobolo), at least in Zulu society, also renders cohabitation a culturally unacceptable type of union formation even in the context of childbirth, unless lobolo payments are underway.3 II MARRIAGE AND COHABITATION TRENDS IN THE POST-APARTHEID PERIOD Although historically, non-marriage among Africans in South Africa seems to have been rare,4 falling marriage rates have been documented since at least the 1960s.5 During the post-apartheid period, racial differences in marriage rates have widened further. By 2010, there was a 40 percentage point differential between the shares of African and White women (20 years and older) who were ‘ever-married’6 (41 per cent of African women, compared to 81 per cent of White women).7 Figure 1 plots the percentage of women ever-married using micro-data derived from a number of nationally representative surveys for South Africa, undertaken from 1995 to 2010. These surveys include the household surveys regularly conducted by Statistics South Africa, which each sampled approximately 30 000 households or 100 000 individuals, and a 1 E Preston-Whyte ‘Women migrants and marriage’ in EJ Krige and JL Comaroff (eds) Essays on African marriage in Southern Africa (1981) 158–172; HL Moore ‘Households and gender in a South African Bantustan: a comment’ (1994) 53(1) African Studies 137–142. 2 M Hunter ‘Fathers without amandla: Zulu-speaking men and fatherhood’ in L Richter and R Morrell (eds) Baba: Men and Fatherhood in South Africa (2006) 99–117; M Hunter Love in the Time of AIDS (2010); V Hosegood, N McGrath and T Moultrie ‘Dispensing with marriage: marital and partnership trends in rural KwaZulu-Natal, South Africa 2000–2006’ (2009) 20(13) Demographic Research 279–312; D Posel, S Rudwick and D Casale ‘Is marriage a dying institution in South Africa? Exploring changes in marriage in the context of lobolo payments’ (2011) 25(1) Agenda 102–111; D Posel and D Casale ‘The relationship between sex ratios and marriage rates in South Africa’ (2013) 45 Applied Economics 663 – 676. 3 D Posel and S Rudwick ‘Ukukipita [cohabiting]: socio-cultural constraints in urban Zulu society’ (2013) Journal of Asian and African Studies. 4 Preston-Whyte (n 1). 5 Hunter (2010) (n 2). 6 Individuals are identified as ‘ever-married’ if they reported being currently married, or widowed, or divorced or separated. These response options did not distinguish between heterosexual or same-sex marriages, and it is therefore likely that women in same-sex marriages are included in the percentage of women ever-married. 7 Own calculations from Statistics South Africa General Household Survey for 2010 (2010).

JOBNAME: Acta Juridica 13 PAGE: 3 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

CHANGING PATTERNS OF MARRIAGE AND COHABITATION IN SOUTH AFRICA

171

2008 survey conducted by the Southern Africa Labour and Development Research Unit (SALDRU), which sampled approximately 7 300 households or 28 000 individuals.8 Marriage rates are presented for two groups of African women and White women – those aged 20 to 30 years; and those aged 40 to 50 years. The Figure shows that the percentage ever-married is substantially lower among young African women (20–30 years) than among young White women. Moreover, the share of young African women ever-married has steadily declined from 1995, falling by almost half. In 2010, only one out of every 10 young African women was ever-married. The percentage of older African women (40–50 years) ever-married has also decreased considerably (by 25 per cent), suggesting that the drop in marriage rates among young African women is not simply explained by a rising age at marriage. In contrast, the data for young White women are noisy (possibly because of the relatively small sample of Whites, who comprise less than 10 per cent of the total population) and it is not possible to identify a clear trend. For example, if 2008 is chosen as the end year, the micro-data would suggest that the percentage of ever-married young White women had returned to 1995 levels. Among older White women, the percentage ever-married shows a slow decline over the 15 year period. But by 2010, still almost 90 per cent of all older White women were ever-married. To the extent that there has been a decrease in marriage rates among young White women, therefore, this appears to reflect largely a delay in marriage rather than non-marriage.

8 All statistics presented are population estimates, derived using the weights provided with each dataset.

JOBNAME: Acta Juridica 13 PAGE: 4 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

MARRIAGE, LAND AND CUSTOM

172

Figure 1: Percentage of women ever married, South Africa, 1995–2010 100 90 80 Percentage

70 60 50 40 30 20 10 0 1995

1997

1999

2004

2005

2006

African (20-30) White (20-30)

2007

2008

2010

African (40-50) White (40-50)

Source: Own calculations from the 1995, 1997 and 1999 October Household Surveys, the September 2004–2007 Labour Force Surveys, the 2008 National Income Dynamics Study and the 2010 General Household Survey; see Statistics South Africa October Household Survey for 1995 (1995); Statistics South Africa October Household Survey for 1997 (1997); Statistics South Africa October Household Survey for 1999 (1999); Statistics South Africa Labour Force Survey (September 2004); Statistics South Africa Labour Force Survey (September 2005); Statistics South Africa Labour Force Survey (September 2006); Statistics South Africa Labour Force Survey (September 2007); Southern Africa Labour and Development Research Unit National Income Dynamics Study (2008); Statistics South Africa General Household Survey (2010). Note: The data have been weighted to represent population estimates.

Percentage

Figure 2: Percentage of women currently co-habiting, South Africa, 1995–2010 20 10 0 1995

1997

African (20-30)

1999

2004

African (40-50)

2005

2006

2007

White (20-30)

2008

2010

White (40-50)

Source: Own calculations from the 1995, 1997 and 1999 October Household Surveys, the September 2004–2007 Labour Force Surveys, the 2008 National Income Dynamics Study and the 2010 General Household Survey; see Statistics South Africa October Household Survey for 1995 (1995); Statistics South Africa October Household Survey for 1997 (1997); Statistics South Africa October Household Survey for 1999 (1999); Statistics South Africa Labour Force Survey (September 2004); Statistics South Africa Labour Force Survey (September 2005); Statistics South Africa Labour Force Survey (September 2006); Statistics South Africa Labour Force Survey (September 2007); Southern Africa Labour and Development Research Unit National Income Dynamics Study (2008); Statistics South Africa General Household Survey (2010). Note: The data have been weighted to represent population estimates.

The large fall in marriage among African women has been partially offset by an increase in cohabitation (see Figure 2). From 1995 to 2010, cohabitation

JOBNAME: Acta Juridica 13 PAGE: 5 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

CHANGING PATTERNS OF MARRIAGE AND COHABITATION IN SOUTH AFRICA

173

rates among younger and older African women grew almost three-fold, although from low bases of 5 and 4 per cent, respectively. In comparison to White women, cohabitation rates are higher among African women particularly for the age cohort of 40–50 years. However, cohabitation rates among White women increased by even more than among African women over the period, although from lower bases of between 2 and 3 per cent. Moreover, when measured in relation to women who are not married, then both younger and older White women are more likely than African women to be cohabiting.9 Consequently, notwithstanding higher cohabitation rates among African women overall, a far larger share of African women are single and not cohabiting compared to White women in the same age cohort. By 2010, 73 per cent of young African women, and 28 per cent of older African women, had never been married and were not currently cohabiting with a partner, compared to 52 per cent of young White women and only 8 per cent of older White women. Racial differences in the nature of union formation in South Africa are interlinked with racial differences in poverty and the economic status of children. Although marriage rates are substantially lower among African women, similar proportions of African and White women are mothers (see Table 1). The marital status of mothers, therefore, differs sharply by race. In 2008, for example, White mothers were more than twice as likely as African mothers to be ever-married (92 per cent compared to 39 per cent). A higher percentage of all African mothers were cohabiting with a partner (14 per cent), compared to all White mothers (5 per cent). However, among mothers who had never married, a far larger share of White women were living with a partner (59 per cent compared to 23 per cent of African never-married women). Table 1: Marital status of women (20 to 50) who are mothers, South Africa, 2008 Proportion of women who are mothers Proportion of mothers who are ever-married Proportion of mothers who cohabit Proportion of never-married mothers who cohabit

AFRICAN 0.827 (0.008) 0.391 (0.011) 0.140 (0.009) 0.229 (0.013)

WHITE 0.798 (0.036) 0.917 (0.025) 0.047 (0.020) 0.591 (0.151)

Source: Own calculations from the 2008 National Income Dynamics Study; see Southern Africa Labour and Development Research Unit National Income Dynamics Study (2008). ||Note: Population estimates are shown. Standard errors are in parentheses. ‘Ever-married’ includes currently married, divorced and widowed.

9 In 2010, among young unmarried women, 17.9 per cent of White women and 17.1 per cent of African women were currently cohabiting with a partner. Among older unmarried women, 41 per cent of White women and 30 per cent of African women were in cohabiting relationships.

JOBNAME: Acta Juridica 13 PAGE: 6 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

174

MARRIAGE, LAND AND CUSTOM

From the perspective of children, in 2008 only 30 per cent of all African children (14 years or younger) were living in a household where their father was co-resident (see Table 2). In contrast, over 70 per cent of White children were co-resident with their father. Higher mortality rates, particularly in the context of HIV/AIDS, are part of the explanation for this difference. Approximately 14 per cent of African children were reported as having fathers who were deceased, compared to 1 per cent of White children. However, lower marital rates among African mothers, and lower cohabitation rates among nevermarried African mothers, are also relevant. Half of all African children were identified as having fathers who were alive, but not part of the child’s household, and for a further 6 per cent, the status of the father (alive or dead) was not known. Table 2: Co-residence of fathers of children (14 years and younger), South Africa, 2008 Proportion of children whose: Father is resident in household Father is alive but not resident Father is dead Father’s status is unknown

AFRICAN 0.299 (0.008) 0.501 (0.008) 0.144 (0.005) 0.056 (0.004)

WHITE 0.710 (0.044) 0.201 (0.033) 0.013 (0.008) 0.076 (0.035)

Source: Own calculations from the 2008 National Income Dynamics Study; see Southern Africa Labour and Development Research Unit National Income Dynamics Study (2008). Note: Population estimates are shown. Standard errors are in parentheses.

The presence of fathers in the household is clearly related to the economic well-being of children. Although poverty rates among all African children are very high, they are significantly increased for children whose fathers are not resident in the household. Using a poverty line commonly adopted in poverty studies for South Africa (of R515 per capita per month in 2008 prices), between 68 and 77 per cent of African children with absent fathers10 were living in poverty in 2008, compared to 54 per cent of children with resident fathers (see Table 3).

10 There is no significant difference in poverty rates according to whether the father is alive but absent, deceased, or his status is unknown.

JOBNAME: Acta Juridica 13 PAGE: 7 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

175

CHANGING PATTERNS OF MARRIAGE AND COHABITATION IN SOUTH AFRICA

Table 3: Poverty rates among children, by residence status of fathers, 2008 Proportion of children living in a poor household where: Father is resident in household Father is alive but not resident Father is dead Father’s status is unknown

AFRICAN 0.544 (0.016) 0.763 (0.009) 0.771 (0.017) 0.684 (0.031)

WHITE 0.005 (0.004) 0.112 (0.066) 0 0

Source: Own calculations from the 2008 National Income Dynamics Study; see Southern Africa Labour and Development Research Unit National Income Dynamics Study (2008). Note: Population estimates are shown. Standard errors are in parentheses. The poverty line is R515 per capita per month in 2008 prices. A child lives in poverty if the child is part of a household where average per capita household income is below the poverty line.

III DISCUSSION A key explanation for racial differences in marriage rates concerns the effects of apartheid legislation, which prevented Africans from settling permanently, and with their families, in ‘White’ urban areas. Influx control regulations gave rise to patterns of individual temporary migration, where men in particular, would migrate to places of employment for an extended period of time. Long periods of separation in turn undermined gender relationships and reduced the possibilities of marriage.11 However, even with the elimination of all restrictions on migration and urbanisation, marriage rates among Africans have continued to fall throughout the post-apartheid period, and by considerably more than among Whites (as illustrated in Figure 1). Some studies suggest that changing attitudes to marriage, in the context of rising levels of education and increased employment opportunities for African women, contribute to low and falling marriage rates among Africans.12 A number of mostly recent studies, however, suggest that economic constraints may delay or inhibit the ability of African men to marry. Two sets of quantitative findings from national micro-data highlight a relationship between the economic status of African men and marriage outcomes in South Africa. First, African men who are married earn considerably more than African men who are not married. However, much of this marital earnings premium derives from the selection of

11

Preston-Whyte (n 1); Moore (n 1); Hunter (2006) (n 2). M Garenne, S Tollman, K Kahn, T Collins and S Ngwenya ‘Understanding marital and premarital fertility in rural South Africa’ (2001) 27(2) Journal of Southern African Studies 277–290; I Kalule-Sabiti, M Palamuleni, M Makiwane and AY Amoateng ‘Family formation and dissolution patterns’ in AY Amoateng and TB Heaton (eds) Families and Households in Post-Apartheid South Africa: Socio-demographic Perspectives (2007) 89–112. 12

JOBNAME: Acta Juridica 13 PAGE: 8 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

176

MARRIAGE, LAND AND CUSTOM

higher-earning men into marriage.13 Married African men earn significantly more than unmarried African men, because higher-earning African men are more likely to marry.14 Second, marriage rates among African women are more highly correlated with sex ratios, which take into account the economic status of unmarried African men, than with sex ratios that capture only the quantity of unmarried African men relative to African women in local marriage markets.15 One reason why the economic status of African men may be particularly important for marriage concerns the traditional practice of bridewealth; and several qualitative studies draw a link between changes in the practice of bridewealth, and a delay in marriage or an increase in non-marriage.16 The custom of bridewealth, which involves marriage payments from the prospective husband to the family of the bride, continues to be widely practised among Africans in South Africa.17 One of the key functions of the custom is to provide compensation to the parents for the loss of the productive and reproductive labour power of their daughter,18 with the exchange expressing a commitment to future reciprocal relationships between the families. With marriage, a husband and wife establish a separate household, or in rural contexts, the wife may move to the homestead of the husband’s family.19 In more recent years, the payment of bridewealth may have become a greater hurdle to marriage for three reasons. First, research suggests that the traditional practice has become increasingly commercialised, acquir13 D Casale and D Posel ‘The male marital earnings premium in the context of bride wealth payments: Evidence from South Africa’ (2010) 58(2) Economic Development and Cultural Change 211–230. 14 The study uses national panel data for South Africa and shows that among African employed men, a robust and positive premium to marriage in cross-sectional estimations is substantially reduced after controlling for individual fixed effects. Furthermore, African men with faster earnings growth in the initial periods of the panel were more likely than other African men to have married by the end of the panel (Casale and Posel (n 13)). 15 The study uses lagged district age-specific sex ratios and shows that among young African women, the probability of marriage is positively and significantly related to the local pool of unmarried African men, but that the relationship is strongest where these men are employed and earning incomes above a certain threshold. In contrast, among White women there is no evidence that economic-based sex ratios are better predictors of marriage outcomes than simple sex ratios; see Posel and Casale (n 2). 16 C Campbell ‘Learning to kill? Masculinity, the family and violence in Natal’ (1992) 18(3) Journal of Southern African Studies 614–628; Hunter (2006) (n 2); M Hunter ‘The changing political economy of sex in South Africa: the significance of unemployment and inequalities to the scale of the AIDS pandemic’ (2007) 64(3) Social Science and Medicine 689–700; Hosegood et al (n 2); Posel et al (n 2); D Posel and S Rudwick ‘Marriage and ilobolo [bridewealth] in contemporary Zulu society’ (2013) forthcoming African Studies Review. 17 Casale and Posel (n 13); Hunter (2010) (n 2). 18 J Guy ‘Gender oppression in Southern Africa’s precapitalist societies’ in C Walker (ed) Women and Gender in Southern Africa to 1945 (1990) 33–47. 19 Preston-Whyte (n 1).

JOBNAME: Acta Juridica 13 PAGE: 9 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

CHANGING PATTERNS OF MARRIAGE AND COHABITATION IN SOUTH AFRICA

177

ing ‘more of an economic imperative’.20 Although national micro-data on bridewealth payments are not available, data from a 1998 regional study suggest that in KwaZulu-Natal (one of nine provinces in the country), the average value of bridewealth payments between 1985 and 1998 was approximately R20 000 (in 2000 prices), representing about 13 times the average monthly real earnings of African men in 1998.21 Second, the payment has become more individualised. Rather than bridewealth payments being made in cattle drawn from a herd belonging to the father of the prospective husband, payments, particularly in urban areas, are made in cash and without assistance from the father.22 Third, and alongside this generational shift in the responsibility to pay bridewealth, unemployment rates among African men have remained stubbornly high in the post-apartheid period. For the period 1999 to 2007 (years for which comparable survey questions on employment and unemployment were asked in national labour force surveys), the official rate of unemployment among African men remained above 20 per cent (26 per cent in 1999 and 23 per cent in 2007); and among young African men (20 to 35 years), it was even higher (34 per cent in 1999 and 30 per cent in 2007). When unemployment rates include those who did not actively search for work in a recent period, then the rate of unemployment among young African men exceeded 40 per cent during this period (47 per cent in 1999 and 42 per cent in 2007).23 At the same time, studies document widespread support for the custom of bridewealth, which has been sustained during the post-apartheid period. Research from the 1980s found that the custom was seen increasingly as the ‘distinguishing mark of a black as opposed to a white marriage’,24 a finding that has been corroborated in more contemporary studies that document the importance of bridewealth as a source of African identity and pride.25 Recent qualitative research suggests that

20 JH Shope ‘ ‘‘Lobola is here to stay’’: rural black women and the contradictory meanings of lobolo in post-apartheid South Africa’ (2006) 6 Agenda 64–72 at 48; see also H Ngubane ‘The consequences for women of marriage payments in a society with patrilineal descent’ in D Parkin and D Nyamwaya (eds) Transformations of African Marriage (1987) 173–182; S Burman and N van der Werff ‘Rethinking customary law on bridewealth’ (1993) 19(2) Social Dynamics 111–127; Posel et al (n 2); Posel and Rudwick (n 16). 21 Casale and Posel (n 13). 22 M de Haas ‘Is there anything more to say about lobolo?’ (1987) 46(1) African Studies 33–55; Ngubane (n 20); Posel and Rudwick (n 3). 23 Own calculations from the 1999 October Household Survey and the September 2007 Labour Force Survey; see Statistics South Africa October Household Survey for 1999 (1999) and Statistics South Africa Labour Force Survey (September 2007). 24 De Haas (n 22) at 5. 25 C Walker ‘Attitudes to lobola: findings from students’ research projects’ (1992) 13 Agenda 57–58; Burman and Van der Werff (n 20); Posel and Rudwick (n 16).

JOBNAME: Acta Juridica 13 PAGE: 10 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

178

MARRIAGE, LAND AND CUSTOM

lobolo is a particularly salient Zulu custom.26 Although lobolo has not been a legal requirement for a Zulu customary wedding since 1932,27 Zulu socio-cultural etiquette still requires that bridewealth payments precede marriage, whether the wedding is based on a customary, church or civil union.28 Delayed marriage and low marriage rates among Africans, therefore, may reflect a tension between men’s inability to pay bridewealth29 and their respect for the custom as an integral part of the marriage process.30 However, if there are economic constraints to marriage, then it is curious that more African couples do not cohabit, particularly following childbirth. In countries where marriage rates have fallen, rates of cohabitation typically have increased and cohabitation has become a more acceptable form of partnership, both as an alternative to marriage and as a prelude to marriage.31 In South Africa, cohabitation rates have risen as marriage rates have declined, and a greater share of all African women currently cohabits with a partner, compared to White women. However, a particular feature of family formation in South Africa (as illustrated in Table 1) is that among single mothers, African women are far less likely to be in a cohabiting relationship than White women. Recent research suggests that, at least in Zulu society, cohabitation even in the context of childbirth is widely viewed as culturally unacceptable, unless there are marriage intentions and lobolo negotiations are underway.32 Traditionally, in the event of pregnancy before marriage, the father of the child would have been expected to marry the mother and the rights to the child would be transferred to the father’s descent group through the payment of bridewealth. However, a number of ethnographic studies describe a growing acceptance of non-marital childbirth. With a high value placed on childbearing, the role of fertility in defining womanhood, and low marriage rates, the links between marriage and

26 Marriage rates among young African women are also lowest in KwaZulu-Natal, where the large majority of Africans are Zulu-speakers; see Posel and Casale (n 2). 27 De Haas (n 22); CRM Dlamini A Juridical Analysis and Critical Evaluation of Lobolo in a Changing Zulu Society (1994). 28 Posel and Rudwick (n 16). 29 Hunter (2006) (n 2). 30 Posel et al (n 2); Posel and Rudwick (n 16). 31 A Thornton and L Young-DeMarco ‘Four decades of trends in attitudes toward family issues in the United States: the 1960s through the 1990s’ (2001) 63 Journal of Marriage and Family 1009–1037; P Heuveline and JM Timberlake ‘The role of cohabitation in family formation: the United States in comparative perspective’ (2004) 66(5) Journal of Marriage and Family 1214– 1230. 32 Posel and Rudwick (n 3). Also see a recent study by Budlender et al of a rural area in KwaZulu-Natal, which documents that, because of the stigma attached to cohabitation, this type of union formation was almost ‘not existent’; see D Budlender, S Mgweba, S Motsepe and L Williams Women, Land and Customary Law (2011) at 51.

JOBNAME: Acta Juridica 13 PAGE: 11 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

CHANGING PATTERNS OF MARRIAGE AND COHABITATION IN SOUTH AFRICA

179

reproduction have been dislodged,33 and childbearing may even take precedence over marriage.34 In contemporary settings, non-marital childbirth is accommodated by requiring the man who fathered the child to pay inhlawulo (‘damages’) to the woman’s family if he wants to claim rights to the child. However, the payment of these ‘damages’, which is typically considerably smaller than the bridewealth payment,35 does not also grant the man co-residence rights with the mother and his child. Cohabitation without the payment of bridewealth would undermine one of the original functions of bridewealth payments, which is to compensate parents for the loss of their daughter’s productive and reproductive labour power, and would signal profound disrespect for the custom of lobolo. In their study of Zulu urban adults, Posel and Rudwick found that cohabitation was widely viewed as a culturally unacceptable practice, even with childbirth, unless lobolo negotiations were underway.36 IV CONCLUSION In South Africa in 2010, the majority of African women were not, or had never been married. In contrast, four fifths of White women were ever-married. Low marriage rates among African women have been offset only partially by rising rates of cohabitation. Even among African women who are mothers, only a little over half were ever-married or cohabiting with a partner in 2010, compared to almost 97 per cent of White women who are mothers. In addition to the effects of HIV/AIDS and labour migration, trends in marriage and cohabitation are central in explaining why the majority of African children in South Africa grow up in households without their fathers. The factors underlying patterns of union formation in South Africa are likely to be varied and complex. This paper has highlighted findings particularly from more recent research, which suggest that low marriage rates among African women may result, at least in part, from economic constraints to marriage. These constraints may be heightened by tensions between a widespread support for the custom of bridewealth, and the high cost of bridewealth given high unemployment rates and a commercialisation of the custom. The practice of, and respect for, the custom of 33 C Walker ‘Conceptualising motherhood in twentieth-century South Africa’ (1995) 21 Journal of Southern African Studies 417–437; CE Kaufman, T de Wet and J Stadler ‘Adolescent pregnancy and parenthood in South Africa’ (2001) 32(2) Studies in Family Planning 147–160; Hunter (2010) (n 2). 34 E Preston-Whyte and M Zondi ‘Assessing illegitimacy in South Africa’ in S Burman and E Preston-Whyte (eds) Questionable Issue: Illegitimacy in South Africa (1992) 226–246; Hosegood et al (n 2). 35 Hunter (2010) (n 2). 36 Posel and Rudwick (n 3).

JOBNAME: Acta Juridica 13 PAGE: 12 SESS: 9 OUTPUT: Mon Nov 25 10:34:53 2013 /dtp22/juta/juta/acta−juridica13/ch08

180

MARRIAGE, LAND AND CUSTOM

bridewealth may also help to explain why more couples do not cohabit, even if they have had a child together. Most of the qualitative studies, however, focus only on Zulu society, where lobolo is a particularly salient custom, and there is considerable scope for further research that explores bridewealth, marriage and cohabitation among other ethnic groups in South Africa.