ARTICLE IN PRESS
Social Science & Medicine 64 (2007) 689–700
www.elsevier.com/locate/socscimed
The changing political economy of sex in South Africa: The
significance of unemployment and inequalities to the scale of the
AIDS pandemic
Mark HunterÃ
Department of Social Sciences/Geography, University of Toronto at Scarborough, 1265 Military Trail, Toronto, Ont., Canada M1C 1A4
Available online 9 November 2006
Abstract
Between 1990 and 2005, HIV prevalence rates in South Africa jumped from less than 1% to around 29%. Important
scholarship has demonstrated how racialized structures entrenched by colonialism and apartheid set the scene for the rapid
unfolding of the AIDS pandemic, like other causes of ill-health before it. Of particular relevance is the legacy of circular
male-migration, an institution that for much of the 20th century helped to propel the transmission of sexually transmitted
infections among black South Africans denied permanent urban residence. But while the deep-rooted antecedents of AIDS
have been noted, less attention has been given to more recent changes in the political economy of sex, including those
resulting from the post-apartheid government’s adoption of broadly neo-liberal policies. As an unintentional consequence,
male migration and apartheid can be seen as almost inevitably resulting in AIDS, a view that can disconnect the pandemic
from contemporary social and economic debates.
Combining ethnographic, historical, and demographic approaches, and focusing on sexuality in the late apartheid and
early post-apartheid periods, this article outlines three interlinked dynamics critical to understanding the scale of the AIDS
pandemic: (1) rising unemployment and social inequalities that leave some groups, especially poor women, extremely
vulnerable; (2) greatly reduced marital rates and the subsequent increase of one person households; and (3) rising levels of
women’s migration, especially through circular movements between rural areas and informal settlements/urban areas. As a
window into these changes, the article gives primary attention to the country’s burgeoning informal settlements—spaces in
which HIV rates are reported to be twice the national average—and to connections between poverty and money/sex
exchanges.
r 2006 Elsevier Ltd. All rights reserved.
Keywords: South Africa; AIDS; Health inequalities; Sexuality; Gender
Introduction and the argument
29%.1 Combining ethnographic, demographic and
historical insights, this article addresses the impor-
Between 1990 and 2005, HIV prevalence rates in
tant question posed nearly 5 years ago by prominent
South Africa jumped from less than 1% to around
South Africanist scholars: Was AIDS in South
ÃTel.: +1 6032988861.
1See Marais (2005, p. 25–43) for a thorough discussion of
E-mail address: mhunter@utsc.utoronto.ca.
South African AIDS statistics.
0277-9536/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.socscimed.2006.09.015

---

ARTICLE IN PRESS
690
M. Hunter / Social Science & Medicine 64 (2007) 689–700
Africa ‘an epidemic waiting to happen’?2 Ap-
ary South Africa. These social forces are perhaps
proaches that forefront the legacy of apartheid to
materialized most vividly in the country’s burgeon-
address this question have succeeded in providing
ing informal settlements where HIV rates are
important historical context and thereby in decisi-
reported to be almost twice as high as they are in
vely challenging models that reify an ‘African
rural and urban areas, attracting surprising little
system of sexuality’ supposedly characterized by
comment (HSRC, 2002, 2005; Pettifor et al., 2004).
sexual permissiveness (for instance as contained in
Characterized by the presence of one roomed
Caldwell, Caldwell, & Quiggin, 1989; for a direct
imijondolos (roughly ‘shacks’) and a population
critique see Heald, 1995). In particular, scholars
that is typically young, unmarried, and without
have brought to attention the ways in which racial
secure work, informal settlements are testimony not
segregation and male migration fuelled an earlier
only to the failure of the state to create viable jobs
epidemic of syphilis only partially quelled by the
and build adequate housing, but to a set of
introduction of penicillin in the 1950s; moreover,
dynamics that have been largely neglected in the
they note how the forces of urbanization, indus-
study of the AIDS pandemic.3
trialization, and Christianization have long been
A few caveats are necessary before approaching
argued to have destabilized African family struc-
such an exploratory discussion. First, the complex
tures (for instance Delius & Glaser, 2002; Horwitz,
interplay between race, class, and geography belie a
2001; Jochelson, Mthibeli, & Leger, 1991; Marks,
single political economy of sex. It is important to
2002; Phillips, 2004). These accounts of the social
state up front therefore that this article mainly
origins of AIDS built on earlier groundbreaking
considers poor South Africans and specifically those
work on the political economy of health in South
classified as ‘African’ under apartheid—the primary
Africa (for instance Andersson & Marks, 1988;
occupants of informal settlements.4 Second, the
Packard, 1989).
paper considers in greatest detail the spatial move-
But while political economy approaches destabi-
ments and livelihoods of women and not men (on
lize the ‘culture’ thesis in vital ways, there is a
related changes in masculinities as a consequence of
danger that they too may be interpreted as evidence
unemployment and inequalities see Hunter, 2005b).
that the severity of the AIDS pandemic was almost
Third, although these arguments evolve out of
inevitable—this time not because of culture but
extensive ethnography in an informal settlement,
apartheid and industrialization. In this article I do
the article draws mostly from secondary data and is
not question the inextricable link between apartheid
aimed at a broader level of analysis.5 Fourth, this
and AIDS. I do, however, draw attention to recent
article considers only one side of the political
changes in South Africa’s political economy of sex
that can help to explain the scale of the pandemic.
3In the last year, informal settlements have emerged as major
This leads us beyond the well-studied political
areas of social protest, especially in Durban with the founding of
the organization abahlali basemjondolo (shack dwellers move-
economy and geography of male-migrancy which,
ment). See Patel & Pithouse (2005).
since the 1940s, has become something of a cliche´ in
4Although the term ‘African’ is problematic, without recourse
explaining sexually transmitted diseases and in
to statistics that use the category it is very difficult to make
framing scholars’ understandings of the political
arguments about social change. That AIDS is not an essential
economy of sex—albeit with much merit. Instead,
African disease but rooted in socio-economic forces—although in
more complex ways than has been previously accepted—is the
this article gives attention to a series of relatively
main thrust of this article. It is possible that growing class
recent interconnected trends, namely rising unem-
divisions are making AIDS less concentrated in one ‘race’, and
ployment and social inequalities, dramatically re-
much more work needs to be done in this area, but the recent
duced marital rates, and the extensive geographical
HSRC (2005) study finds infection rates of 13% for ‘Africans’,
movement of women as well as men in contempor-
1.9% for ‘Coloureds’, 1.6% for ‘Indians’, and 0.5% for ‘Whites’.
If this is broadly correct, what requires explaining is not only the
historical reasons for this discrepancy but the socio-spatial
2The title of a keynote speech given by Shula Marks at the 2001
dynamics that underpin it today. For a particularly important
‘AIDS in Context’ conference in Johannesburg (subsequently
argument linking race, ‘premature death’, and geography at a
published in African Studies in 2002) and also the title of a book
more general level see Gilmore (2002).
(Waiting To Happen) based on the same conference (Reid,
5Ethnographic research was conducted in Mandeni, KwaZulu-
Walker, & Cornell, 2004). This article is indebted to these and
Natal, where I lived extensively with a family in Isithebe Informal
other analyses that emphasize the connection between apartheid
Settlement between 2000 and 2004. I have refrained from giving
and AIDS; the purpose of this piece is to take the political
extensive background to the study because of space limitations,
economy approach they represent forward.
but see Hunter (2005a).

---

ARTICLE IN PRESS
M. Hunter / Social Science & Medicine 64 (2007) 689–700
691
economy of sex, namely sexual relationships be-
apartheid South Africa (on state ‘denialism’ and the
tween men and women; it does not look at the
way this has hindered post-apartheid AIDS inter-
connection between political economy, same-sex
ventions see Robins, 2004). This article argues that
relationships and AIDS or, indeed, sexual violence
recognizing the shifting intersections and spatiality
and AIDS.6 But a final important caveat must be
of race, class, gender and sexuality might go some
made—sex and AIDS should not be too easily
way towards reconfiguring debates on AIDS in
equated. Starting with Packard and Epstein’s (1991)
South Africa around a more politically enabling
important piece, it has long been argued that racist
agenda—one that conceives of AIDS as a symptom
assumptions exaggerate the importance of sex to the
of ‘structural violence’ but does not foreclose sex as
spread of AIDS and exoticize sexuality in proble-
a mechanism for the transfer of HIV (for a broader
matic ways (see also Vaughan (1991), and McClin-
overview of structural violence and health in sub-
tock (1995), on racial tropes surrounding African
Saharan Africa see Schoepf, Schoepf, & Millen,
sexuality, and Stillwaggon (2006), for a recent
2000).
reassertion of the significance of nutrition, parasites
Over the last several decades, social scientists,
and other co-factors to the spread of AIDS in
especially within anthropology, have given in-
Africa). Yawning divides in health care provisions,
creased attention to the connection between poli-
for instance, can certainly go a long way to explain
tical economy, the body, and health (for instance
the uneven unfolding of AIDS in South Africa; 15
Bourgois, 2003; Scheper-Hughes, 1992; for a review
percent of the population has access to private
of new approaches to the body see Lock, 1993). One
health care facilities while the remainder, most of
branch of this work examines the changing social
whom are black, depend on an overburdened public
roots of sexually transmitted infections. Recent
health service (Health Systems Trust, 2004). The
ethnographies from the ‘Third World’, for instance,
social context of AIDS is, without doubt, much
demonstrate how global trade and the informaliza-
broader than the political economy of male/female
tion of work can propel women into the sexual
sex, the central focus of this article.
economy, a scenario ultimately driven in some
Nevertheless, there is strong evidence for a
instances by World Bank/IMF sponsored structural
changing political economy of sex relevant to the
adjustment programmes (Brennan, 2004; Kempa-
way we conceive of the social dynamics linking sex
doo & Doezema, 1998; Schoepf et al., 2000).
to HIV transmission. It is clear that most African
Ethnographic research in South Africa also shows
women today are no longer waiting in rural areas to
how unemployment, poverty and sex/money ex-
be infected by their migrant partners, the pattern of
changes can fuel multiple-sexual-partners, some-
infection described convincingly in the 1940s for
times across large age gaps (Hunter, 2002; LeClerc-
syphilis and often evoked uncritically in the
Madlala, 2003; Selikow, Zulu, & Cedras, 2002; for a
contemporary period (for the classic enunciation
review of the broader sub-Saharan literature see
of this view see Kark, 1949).7 If President Mbeki’s
Luke, 2003). Moreover, quantitative data seem to
questioning of sex as a mechanism for the transmis-
support ethnographic evidence of younger women
sion of the disease has to be seen in part as a
having sexual relationships with older, often richer,
reaction to longstanding racialized representations
men: one recent study in South Africa found that
of Africans as inherently diseased and promiscuous,
nearly four times as many women as men aged
it has not yet been challenged by scholarship that
20–24 were HIV positive (23.9% compared to 6%),
adequately explores the social context of sex in post-
and that a significant number of young women had
older partners (HSRC, 2005).8
6For an excellent recent history of same-sex relationships in
Zimbabwe see Epprecht (2004). On connections between sexual
8Some caution should be exercised, however, in interpreting
violence and HIV prevalence see Dunkle, Jewkes, Brown, Gray,
age/prevalence figures in this way since it is likely that biological
McIntryre et al. (2004).
factors may play a significant part in the susceptibility of young
7Indeed, in addition to the movement of women, to which I
women to HIV infection (see Glynn et al., 2001). On links
point, studies show that in many cases rural-based women infect
between ‘transactional sex’ and AIDS see Dunkle, Jewkes,
their husbands, and not the other way around. One study of
Brown, Gray, McIntryre et al. (2004). For the argument that
discordant couples (where only one partner is infected) in rural
the same number of overall partners within concurrent rather
KwaZulu-Natal showed that in four out of 10 cases it was
than serial relationships leads to a considerably more rapid
actually women and not their partners who were HIV positive
advance of STIs see Morris and Kretzschmar (1997) and
(Lurie et al., 2000).
Halperin and Epstein (2004). More generally, I am well aware

---

ARTICLE IN PRESS
692
M. Hunter / Social Science & Medicine 64 (2007) 689–700
What are the historical origins of the ‘sexual
incidence (the rate of new infections) finds that a
economy’? Especially over the last century, coloni-
disproportionate amount of HIV infection takes
alism and apartheid molded the contours of sexual
place within these spaces (HSRC, 2005). Research-
relations in South Africa in distinctive ways.
ers estimate annual incidence rates in urban
Migrant labor, an institution entrenched in the
informal areas of 7% as compared to 1.8% in
19th century following the discovery of gold and
urban formal areas, 2.7% in rural formal areas, and
diamonds, restricted Africans from settling in urban
2.7% in rural informal areas.9 This recent study, in
areas and forced men into long absences from their
addition to an earlier one from the same institution
rural homes. Men’s relationships with urban ‘pros-
(HSRC, 2002), also found that urban informal areas
titutes’ helped to fuel the syphilis epidemic that
contained the highest reported rates of multiple-
peaked in the first half of the century (Kark, 1949).
partnered-relationships. Of course, even if these
During the course of the 20th century, however,
figures are accurate (and we must maintain some
several major shifts took place that reworked
scepticism) sex can only partly explain such large
connections between wage labor, the rural house-
geographical variations in HIV prevalence; higher
hold, and the sexual economy (I look elsewhere at
infection rates in informal settlements compared to
urban areas, see Hunter, 2005a). First, land
richer areas are in part a consequence of inadequate
dispossession meant that the countryside shifted
water, nutrition, and sanitation and the general
from a position where it subsidized low urban wages
poor state of health in the former. Yet by
to a scenario where it depended on migrant men’s
developing a better understanding of the changing
remittances (Wolpe, 1972). But a second change was
political economy of sex scholars and policy
the dramatic rise in unemployment from the mid-
makers can untangle important new connections
1970s that undermined the ability of men to act as
between poverty, gender, migration, and AIDS.10
reliable providers. One noted consequence for the
The rest of the paper considers these themes in more
sexual economy was the redistribution of the wages
detail.
of male migrants still in employment (now an elite)
through liaisons with rural-based women (Spiegel,
Beyond the male migrant: towards a new political
1981). But it is to the greater movement of women,
economy of sex
especially to informal settlements, that I give
attention; in these spaces some women find access
The archetypal infection route for syphilis in the
to employment, and yet the sexual economy can
1940s, outlined famously by Sydney Kark (1949),
play an important role in everyday subsistence.
involved a male miner becoming infected by an
Instead of a geography characterized by circular
urban prostitute and then passing on the disease to
male migration and the building of a rural home,
his rural wife. I relay below very briefly the case of
therefore, I give consideration to how many rural
Fikile (a pseudonym) that captures a quite different
born women today are engaged in circular move-
political economy of sex outlined in this paper.
ments to informal settlements, spaces dominated by
Fikile was 26 years old when she was interviewed in
densely packed imijondolos.
2003. She grew up in Northern KwaZulu–Natal and
Informal settlements have long been part of
mothered two children from two different men. At
South Africa’s divided landscape but they are given
the time of the interview, one of these children had
attention here because of the way they provide a
died while the other stayed with her grandmother in
window into important recent economic and demo-
Fikile’s rural home. Throughout the 20th century a
graphic changes. These areas have very high HIV
growing number of women gave birth out of
prevalence rates (the proportion of people who are
wedlock. But Fikile, like most young women today,
HIV positive) and, furthermore, new data on HIV
had a very low prospect of marrying the biological
(footnote continued)
9Although the accuracy of the test used to measure HIV
of arguments that sex has either been always easily tradable in
incidence has been recently disputed. See /http://www.unAID-
Africa or thrown into exchange relations as a consequence of the
S.org/en/HIV_data/Epidemiology/default.aspS
accessed
20
forces of modernity. I believe, however, that sex has become
March 2006. Thanks to Hein Marais for this reference.
material in distinctive new ways over roughly the last three
10Suggestive of the greater risk attached to women’s migrancy,
decades. See Hunter (2005a) for a further discussion. For an
one of the few studies on women migrants to urban areas found a
excellent review of the often inappropriate use of ‘prostitution’ in
positive correlation between migrancy and HIV status (Zuma,
Africa see Standing (1992).
Gouws, Williams, & Lurie, 2003).

---

ARTICLE IN PRESS
M. Hunter / Social Science & Medicine 64 (2007) 689–700
693
father of a child. Fikile had not therefore moved
Table 1 draws out distinctions between social
into a husband’s umuzi (homestead) as most women
dynamics that can fuel AIDS and those that fuelled
of her parents’ generation would have done.
STIs in the 1940s/1950s.
Instead, like many rural women, Fikile migrated
The following section tentatively sketches out the
to an informal settlement, in her case to Isithebe in
contemporary political economy of sex, giving
KwaZulu–Natal. When interviewed, she said that
attention to significant changes that occurred since
she had two boyfriends who both gave her money
roughly the 1970s. Despite the election of the first
and were working. Her main boyfriend supported
non-racial government in 1994, the adoption of neo-
her with food and money, some of which she sent
liberal economic policies that favor market forces
home. She was thus reversing the longstanding
over state interventions has accentuated, and not
pattern of men acting as ‘providers’ for rural
diminished, the thrust of these trends. Three
households. The other, an umakhwapeni (secret
dimensions of the contemporary political economy
lover, literally ‘under the armpit’), gave her 50 or
of sex are important to understand: (1) rising
100 Rands irregularly (US$8–$16). Her main
unemployment and the marginalization of women;
boyfriend did not want to use condoms, and she
(2) rapidly declining marital rates; (3) the growth in
says that she trusted him but she uses condoms with
women’s movement, often in circular migration
the other. Typical of the way that sexual relation-
patterns that pivot around a rural ‘home’.
ships can be stretched from rural to urban areas by
women’s as well as men’s movement, Fikile said
Unemployment, new social inequalities and the
that she also slept with a boyfriend from home and
marginalization of poor women
did not use condoms in part because he claimed to
be her ‘indoda yami ngempela ngempela’ (my man,
Under apartheid, work for African men and
for real, for real) although he was not working and
women was typically dangerous, humiliating, and
could not support her. She did washing and ironing
insecure; what’s more, oppressive state policies such
1 day a week, earning R50. Under apartheid, many
as ‘forced removals’ and the hated pass laws made
women moved to towns and survived through the
even the most basic decisions, such as where to live,
informal sector and sometimes the sexual economy;
highly charged. Nevertheless, because of high
but today what Fikile’s case illustrates is the sheer
economic growth in the apartheid era (with steady
scale of women’s movement, the absence of
growth after 1948 and accelerated growth in the
marriage as a rural alternative, and the very poor
1960s) the labor of young African men and, to a
opportunities for income generation in the informal
lesser extent women, was in heavy demand. All of
sector.
this changed decisively from the mid-1970s as
Table 1
Summary of the changing political economy of sex affecting sexually transmitted diseases
Social context
1940s/1950s
Today
Economic
A heavy demand for the labor of African men, some
Chronic unemployment, especially among Africans.
demand for the labor of African women. Very low
Rising class divisions
wages
Household
Marriage is increasingly unstable but still common in
Growth in the number of households, many of which
rural and urban areas. Rural households are
are one-person households. Rural areas dependent
increasingly dependent on male remittances
on state pension and remittances from men and
women
Geography
Men in circular migration patterns, some women
Men and women in multiple migration patterns
moving to urban areas. Many informal settlements
including circular migration. Growth of informal
‘removed’ by apartheid planners
settlements typified by one roomed imijondolo
Sexuality
Some migrant men with multiple partners. Some
Many women dependent on men, sometimes
women dependent on men in extra-marital
multiple men, outside of marriage. Pre-marital
relationships. Premarital relationships not, on the
relations among the poor often characterized by sex/
whole, characterised by exchanges of sex for money
money exchanges

---

ARTICLE IN PRESS
694
M. Hunter / Social Science & Medicine 64 (2007) 689–700
economic and political crisis shook the country.
interventions. It is true that the state has substan-
From this time, positive per capita growth drifted
tially increased the value of the old age pension and
into negative growth and unemployment increased
introduced a child support grant aimed at assisting
rapidly. The first casualties of economic crisis were
children up to the age of 14.12 As we shall see, some
African men; a new class of men who had never
houses have been built and, importantly, a large
been formally employed quickly came into exis-
number have been allocated to single women (this
tence. But not all men lost out. Migrant men who
group is able to benefit from rules stating that
were lucky enough to hold secure work after the
household heads who are single must have depen-
1970s benefited from unionization and the stabiliza-
dents). But seen against the collapse of formal
tion of mining employment and returned to rural
employment in the last three decades these inter-
areas to become a rural elite (Sharp, 1994; Spiegel,
ventions are inadequate.
1981). Nearly two decades ago, Andersson & Marks
(1988) pointed to the implications of rising social
An unprecedented decline in marriage
inequalities for health.
Many argue that market-led economic policy in
African marriage has tended to be painted with
the post-apartheid period has accentuated social
broad brush strokes. Going back to the 1930s a
inequalities. A recent UNDP (2003) study con-
number of ethnographic studies noted the negative
cluded that ‘Income distribution remains highly
effects of ‘cultural contact’ on the African family
unequal and has deteriorated in recent years’.11
(e.g. Hellman, 1948; Krige, 1936; Longmore, 1959).
Seemingly shell-shocked by the perceived power of
Rich in their detail, and undoubtedly capturing a
‘globalisation’, the government retreated from an
sense of change, they fed perceptions that African
interventionist economic and social strategy to one
families were in slow but steady decline. Yet more
that stressed growth through the market. Rapid
recently, scholars have questioned teleological
trade liberalization, one element of this broadly
narratives of ‘family breakdown’ in Africa (Moore
neoliberal program, dramatically increased wage
& Vaughan, 1994; Thomas, 2003); certainly, in
competition and placed sectors such as clothing
many respects, apartheid was a system that con-
under great pressure, helping to nudge unemploy-
tained forces that both undermined and strength-
ment up to over 40% (Kenny & Webster, 1998;
ened the patriarchal home (Hunter, 2005a). Studies
Nattrass, 2003). In the 1980s, women were drawn
in urban and rural areas attest to the continued
into factory work in increasing numbers, but by the
centrality of the institution of marriage as late as the
1990s they were joining men in the ranks of the
early 1980s (De Haas, 1984; Mayer, 1980). A
unemployed. Between 1995 and 1999, the number of
recognition that there was no straightforward causal
economically active women (searching for or secur-
link between apartheid and family breakdown
ing informal or formal work within the labor
focuses our attention instead on the seismic changes
market) increased by two million, twice the increase
heralded by the deterioration of formal employ-
in the female population of working age (Casale &
ment. What is significant about the present genera-
Posel, 2002). As rural areas continued their decline,
tion of young South Africans is that they are
therefore, women were pushed into poorly remun-
experiencing a simultaneous collapse of agrarian
erated and highly unstable informal work; conse-
and wage livelihoods with very important conse-
quently, women’s median income fell sharply in the
quences for marriage, household formation and
post-apartheid period (Casale, 2004). Added to
sexuality.
these labor market changes was the AIDS pandemic
Ethnographic studies today reveal extremely low
itself, which accentuated social inequalities, includ-
marital rates among Africans, especially from the
ing gendered inequalities (Marais, 2005). These
1980s.13 Statistics on African marriage are more
forces drove a wedge between rich and poor and
have not been adequately countered by state
12Pensions were raised in February 2006 to R820 a month
(dollar;130). Their importance to rural homes, even at much
lower rate, was noted by Ardington and Lund (1995). Child
11The Gini-coefficient rose from 0.596 in 1995 to 0.635 in 2001
support grant is currently R 190 (dollar;30) a month. Both are
(UNDP, 2003, p. xvi). Terreblanche (2002) too argues that social
vital to survival in rural areas, especially for women and, indeed,
inequalities increased from the 1970s and were accentuated in the
for the continued support of the African National Congress
post-apartheid period. On the neo-liberalism in post-apartheid
(ANC) among the poor.
South Africa see Bond (2000) and Marais (2001).
13See Denis and Ntsimane (2004) and Hunter (2005a).

---

ARTICLE IN PRESS
M. Hunter / Social Science & Medicine 64 (2007) 689–700
695
Other
Never Marr
Married
100%
6
12
9
12
12
15
17
90%
80%
38
34
35
70%
39
47
48
60%
54
50%
Percentage
40%
30%
56
57
54
49
42
20%
38
30
10%
0%
1936
1951
1960
1970
1980
1991
2001
Year
Fig. 1. Marital status for Africans over 15 years of age, 1936–2001. Source: various census statistics. For more details see Hunter (2005a).
difficult to interpret: African marriage is a process
Among the poor, wedlock is being exposed as a
and not an event, different systems of civil and
decidedly inflexible institution through which to
customary marriage co-existed (with different re-
organize social alliances and the flow of resources
gional administrations), and apartheid statistics are
(cf. Niehaus, 1994). Marriage today is, in many
notoriously unreliable. Nevertheless, census data
respects, a middle-class institution (Fig. 1).
support the claim that there has been a quite
dramatic decline over the last four decades.14 The
Women’s increased movement: from migrating men
factors behind this decline are complex; they include
to moving women?
(until recently) women’s increased work prospects
and thus their growing economic independence
Male migrant labor is so dominant an institution
from men. But particularly from the mid-1970s,
in South Africa that it overwhelms almost all
when unemployment rose sharply, men’s inability to
discussions on migration. Yet for over a century
secure ilobolo (bridewealth) or act as dependable
Southern African women have moved to towns,
‘providers’ became additional brakes on marriage.
informal settlements, and white owned farms (Bon-
ner, 1990; Walker, 1990). The ending of influx
14
controls (that sought to restrict Africans from
The simplest way to track changes to marriage would be to
scrutinise marital rates. But such figures were collected only for
entering towns) in 1986 has typically been noted as
Whites, Indians and Coloured groups, leaving marital status,
the main driving force behind women’s recent
available from population census data, as the most reliable proxy
movement. But I want to suggest that this interacts
when considering African marriage. For a discussion of South
with a relatively new set of dynamics: the reorganiza-
African data regarding marriage see Budlender, Chobokoane,
tion of rural households into more geographically
and Simelane (2004). The above figures include civil and
customary marriages.
flexible institutions with an expectation that women

---

ARTICLE IN PRESS
696
M. Hunter / Social Science & Medicine 64 (2007) 689–700
as well as men will migrate in circular patterns to
resulting mainly from population movement within
urban/informal areas. The extent of these new
urban areas to settlements resulting largely from
patterns of movement has not been captured by
population migration from rural areas. The most
national statistics (although represented for some
common housing type in informal settlements,
time in micro-level accounts by anthropologists
however, is imijondolo, one roomed accommodation
(James, 1999; Spiegel, 1995)) in part because of a
sometimes translated into English as ‘shacks.’ These
male bias in migration data, outlined below.
structures are also widespread in formal townships,
The most common source of data on women’s
where they are sometimes called ‘backyard shacks.’
migration comes from census or household surveys.
Informal settlements have a long history in South
Together these studies show a rise in women’s
Africa. Rooted in the uneven provision of formal
migration from the seventies but indicate that men
housing for Africans under colonialism and apart-
still migrate more than women (see Kok, O’Dono-
heid, they mushroomed around urban areas from
van, Bouare, & Van Zyl, 2003; Posel & Casale,
the mid-1980s following the relaxation of influx
2003). The limitation of this data is that it defines
controls and rising unemployment (on the variety in
migration as either ‘absence’ from a home or a long
informal settlements across national and local
distance/long term change in residence. In contrast,
geographies see Harrison, 1992, and Crankshaw,
data collected by the Africa Centre for Population
1993). Since the early 20th century informal
and Health Studies in rural KwaZulu-Natal appears
settlements have been known as spaces of poverty
to show that rural women move more than men.15
and sex exchanges but also as places that allowed
The Africa Centre visits each household in a
women a certain independence (for instance Bon-
geographical area that it calls the Demographic
ner, 1990). The most lucrative informal activity
Surveillance Areas (DSA) every 4 months and
associated with women is probably the brewing of
counts migration as a change of residence for longer
traditional beer. But today the mass production of
than 3 months. Within the DSA’s roughly 90 000
beer and other consumer items undermines many
population, the Africa Centre is therefore able to
informal opportunities; what’s more the sheer
capture shorter-term ‘movement’ to and from a
number of women eking out a living in the informal
place, a pattern more followed by women, especially
sector drives down earnings opportunities (see
those who leave children in rural areas. Initial data
Rogerson, 1997).
shows that, at the peak age—when men and women
The provision of housing, perhaps more than any
are in their early 20s—slightly more women than
other policy, was heralded by the newly elected
men change residency or ‘move’. We can see
ANC in 1994 as having the ability to jump start
immediately that the emphasis on ‘movement’
radical economic and social redistribution. Some
rather than ‘migration’ forces us to challenge quite
scholars blame shortcomings in housing policy on
radically the way that we think about the gendered
the weakness of the chosen mechanism for delivery,
nature
of
migration/movement.
Longstanding
namely a market-driven one-off capital subsidy
methods of measuring migration therefore tend to
system (see Huchzermeyer, 2004). But others have
capture very well the movement of men for most of
argued that the ANC’s failure to meet its housing
a year into an urban area; they describe less well,
targets was partly due to significant increases in the
however, women’s more frequent movements back-
overall number of households—the government was
wards and forwards from rural areas.
thus dealing with a moving target (Hempson &
O’Donovan, 2005). Evidence of the proliferation of
smaller, single, households is strong: from 1995 to
Informal settlements, the informal economy, and the
2002 average household size was reduced from 4.3
sexual economy
to 3.8, driven by a rising share of single households
from 12.6% to 21% of all households (Pirouz,
There are many types of informal settlements in
2004). From 1996 to 2003 the number of informal
contemporary South Africa, from squatter camps
dwellings rose by 688 000 in South Africa, despite
the existence of house building projects funded by
15This section draws from my interaction with Africa Centre
the state (Mail Guardian, 2005). Informal settle-
linked scholars, particularly Carol Camlin, Caterina Hill, Kobus
ments today therefore are not only testimony to
Herbst, Vicky Hosegood, & Thembeka Mngomezulu. For further
high unemployment rates and an inadequate
details see Africa Centre Population Studies Group (2003) and
Hosegood and Timaeus (2005).
government housing strategy but to significant

---

ARTICLE IN PRESS
M. Hunter / Social Science & Medicine 64 (2007) 689–700
697
demographic trends, namely the rise of smaller
running taxis. It is important, however, to distin-
households not formed around a marital bond.
guish the sex/money exchanges described here from
The author’s ethnographic research was mostly
‘prostitution’—an activity that most residents say is
conducted at or nearby to Isithebe Informal
rare in Isithebe. These sexual networks serve to
Settlement on the North Coast of the province of
widen women’s ability to make claims on resources
KwaZulu-Natal. At the beginning of the 20th
and are not simply instrumental exchanges of sex
century, the area was designated as ‘tribal’ land
for money: some partners can co-habit, gifts are
where only Africans could reside. In 1971, the
often enacted in terms of men’s ‘provider’ role,
apartheid state established Isithebe Industrial Park
claims can be made through evoking ‘love’, and
in the area as part of its strategy to reduce migration
participants frequently discuss sexual pleasure and
to large urban ‘white’ towns. It was on tribal land
physical attraction.16 It is common to hear stories of
surrounding Isithebe’s factories that Isithebe infor-
women having material relationships—’one for
mal settlement grew (this is where the author
money, one for food, and one for clothes’—but
stayed). In the 1990s, many scholars felt that the
also common to hear about love letters and signs of
collapse of apartheid would herald a return to
affection. A final important point to recognize is
‘normal’ migration patterns—settlement patterns
that these sexual networks operate alongside—and
would be concentrated around (rising) employment
not in opposition to—social networks based on
opportunities and migrants would not be forced to
kinship, friendship groups, churches, and neighbors.
maintain such strong links with rural areas. But in
In many cases therefore, sex exchanges do not
the 1990s Isithebe informal settlement, like many
cause family breakdown, a fact that questions the
others, continued to mushroom despite a decline in
very long association between ‘prostitution’ and
employment, in the Isithebe from around 23 000 in
‘social degeneration’. On the contrary, remittances
1990 to roughly 15 000 by the end of the decade.
from sexual networks can help to foster kinship ties.
What’s more movement rates continued to be high,
There is an expectation that women will furnish
with migrants generally keeping a foothold in rural
money to a rural home, especially if a woman’s child
KwaZulu-Natal. Put simply many people who
is looked after by other family members.17 Earlier
migrated to Isithebe in the 1990s did not find work
scholarship showed how men’s urban wages were
and did not consider themselves permanent resi-
distributed through sexual networks in rural Le-
dents (see Hunter, 2005a). Women in Isithebe
sotho (Spiegel, 1981) and how rigid conjugal bonds
depend on a myriad of informal sector activities,
in South Africa were being superseded by more
from selling dagga (marijuana) to petty trading. But
flexible sibling bonds characterized by reciprocity
the sexual economy is also an increasingly impor-
(Niehaus, 1994). What is different today is the ways
tant mechanism for the redistribution of formal and
in which women’s migration further reflects and
informal earnings and the provision of shelter for
affects this changing household structure. In a
migrant women.
situation where marital bonds are no longer
Fikile’s case shows quite clearly a set of dynamics
common, rural women tend to pivot multiple
that can fuel AIDS. Yet ethnographic data from
movements around their rural home (sometimes
Isithebe Informal Settlement suggests that informal
where a child is left), a fairly flexible arrangement
settlements are not places of inherent HIV risk.
allowing for women’s frequent movement, the
Women with reasonably well-paid work can some-
times secure housing independent of men and enter
16I draw inspiration from economic anthropologists’ work on
relationships on their own terms. New female
‘the Gift’ that stems from Marcel Mauss’s (1925 [1989]) famous
migrants can look up to them as role models, seeing
ethnography of Melanesia of the same title. For an excellent
independent women as successfully challenging
discussion of gift exchanges and intimacy in Brazil see Rebhun
(1999). The nature of these relationships is important: while
patriarchal models of marriage. Moreover, not all
policy makers tend to see low condom use in narrow terms of
men are able to take part in the sexual economy.
‘male power’, it is often in affairs between ‘boyfriends’ and
Many men are desperately poor and complain
‘girlfriends’—positioned as being about love—where men and
bitterly about richer men who are able to secure
women are least likely to use condoms and in the most
multiple girlfriends; in Isithebe, men with disposable
commodified relationships, prostitution, where condoms are used
the most (for example Smith, 2004).
income are typically those who still work in union-
17One study found that 60% of temporary migrants had
ized factories or have invested redundancy pay-
communicated with the rural home in the 2 weeks prior to one
ments in lucrative ‘male’ informal activities such as
study (Collinson, Tollman, Kahn, & Clark, 2003).

---

ARTICLE IN PRESS
698
M. Hunter / Social Science & Medicine 64 (2007) 689–700
transfer of resources through sexual liaisons, and
the presence of co-factors, and high HIV rates. Key
the redistribution of state benefits, especially pen-
features of the post-apartheid landscape such as the
sions, often through the presence of a rural grand-
rise of overcrowded informal settlements must be
parents, usually a gogo (granny). As unemployment
closely examined for both what they reveal about
bites deeper into society, sexual exchanges and the
the geography of co-factors as well as what they say
household have been interwoven in new ways.
about the political economic context of sexual
Indeed, without sexual exchanges, many of the
relations. Although AIDS is sourced in colonialism
women-headed imijondolo households would simply
and apartheid, the scale of the impact of South
not exist.
Africa’s AIDS pandemic was not inevitable; argu-
ably its trajectory has been worsened by continued
Conclusions
social and geographical divisions in the post-apart-
heid period.
In the last decade political economists have
rightly stressed the deep social roots of AIDS. Yet
Acknowledgments
in the main, old models, typically male-migration,
have been projected forwarded to explain the scale
I would like to thank Deborah James, Jenna
of the pandemic and shed light on contemporary
Loyd, Lynn Thomas, and this journal’s anonymous
sexual practices. Stillwaggon’s (2006) powerful
reviewers for helpful comments on previous versions
argument that the social context of AIDS must
of this article. The research upon which this article
extend beyond sex is undoubtedly true. But so too
is based was assisted financially by fellowships from
must scholars re-assess what is meant by the social
the Wenner–Gren Foundation for Anthropological
context of sex. To illustrate important trends, the
Research and the International Dissertation Field
article has given attention to women’s movement,
Research Fellowship Program of the Social Science
especially into informal settlements, and to the role
Research Council with funds provided by the
of sex-money exchanges in fuelling AIDS; these are
Andrew W. Mellon foundation.
meant to stimulate discussion and not be taken as
all-encompassing alternative explanations. What the
paper does insist, however, is that the roots of AIDS
References
can no longer be simply conflated with apartheid.
Africa Centre Population Studies Group. (2003). Socio-demo-
This oversimplistic view sits rather neatly with many
graphic and health events 2000. In E. Chiweni, A. J.
in government and policy circles and is rarely
Chimbwete, & Herbst (Eds.), Socio-demographic and health
challenged by epidemiological research that gives
events 2000. Mtubatuba: Africa Centre.
priority to statistical analysis over historical and
Andersson, N., & Marks, S. (1988). Apartheid and health in the
ethnographic perspectives.
1980s. Social Science & Medicine, 27(7), 667–681.
Ardington, E., & Lund, F. (1995). Pensions and development:
In a democratic South Africa social equity is
Social security as complementary to programmes of recon-
important not only to create a more just society but
struction and development. Development Southern Africa,
to stem a shattering health crisis. The remarkable
12(4), 557–577.
success of the Treatment Action Campaign in
Bond, P. (2000). Elite transition: From apartheid to neoliberalism
forcing the state to introduce antiretrovirals has
in South Africa. London: Pluto Press.
Bonner, P. (1990). ‘Desirable or undesirable Basotho women?’
not been matched by a campaign that links housing,
liquor, prostitution and the migration of Basotho women to
employment, and social equality to AIDS—one that
the Rand, 1920–1945. In C. Walker (Ed.), Women and gender
considers the pandemic as a symptom of ‘structural
in southern Africa to 1945 (pp. 221–250). Cape Town: David
violence’ with longstanding but also contemporary
Philip.
roots. Of course, the links between economic policy
Bourgois, P. I. (2003). In search of respect: Selling crack in el
barrio (2nd ed.). Cambridge: Cambridge University Press.
and AIDS are complex. The article has touched on
Brennan, D. (2004). What’s love got to do with it?: Transnational
how employment, housing and state benefits can
desires and sex tourism in the Dominican Republic. Durham,
help to lift women out of the sexual economy. But
N.C.; London: Duke University Press.
these connections and others must be more fully
Budlender, D., Chobokoane, N., & Simelane, S. (2004). Marriage
teased out and emphasized if prevention campaigns
patterns in South Africa. Methodological and substantitive
issues. Southern African Journal of Demography, 9(1), 1–26.
are to resonate with the social conditions in which
Caldwell, J., Caldwell, P., & Quiggin, P. (1989). The social
many South Africans live. So too must scholars
context of AIDS in sub-Saharan Africa. Population and
more vigorously investigate links between poverty,
Development Review, 15(2), 185–234.

---

Document Outline

• The changing political economy of sex in South Africa: The significance of unemployment and inequalities to the scale of the AIDS pandemic
  • Introduction and the argument
  • Beyond the male migrant: towards a new political economy of sex
    • Unemployment, new social inequalities and the marginalization of poor women
    • An unprecedented decline in marriage
    • Womens increased movement: from migrating men to moving women?
  • Informal settlements, the informal economy, and the sexual economy
  • Conclusions
  • Acknowledgments
  • References

---