CENTRE FOR

SOCIAL SCIENCE RESEARCH

Social Surveys Unit

RACE, DISCRIMINATION AND

DIVERSITY IN SOUTH AFRICA

Jeremy Seekings

CSSR Working Paper No. 194

May 2007
Jeremy Seekings is Professor of Political Studies and Sociology at the University
of Cape Town. He has held visiting appointments at Yale and Oxford universities.
His books include Heroes or Villains? Youth Politics in South Africa in the 1980s
(Johannesburg, 1993), The UDF: A History of the United Democratic Front in South
Africa, 1983-2001 (Cape Town, London and Athens, OH, 2000), and Class, Race and
Inequality in South Africa (co-authored with Nicoli Nattrass, New Haven and London,
2005, and Pietermaritzburg, 2006). From 2001 to 2006 he edited the journal Social
Dynamics, and is currently co-editor of the International Journal of Urban and Regional
Research. He is presently conducting research on the politics of welfare state-building
in developing countries.

This paper was discussed at a conference on “Legal and Political Remedies to

Discrimination”, at the University of California, Los Angeles, February 2007.
Race, Discrimination and Diversity in
South Africa

Abstract
The end of apartheid has brought a resurgence of research into racial identities,
attitudes and behaviour in South Africa. The legacy of systematic racial
ordering and discrimination under apartheid is that South Africa remains deeply
racialised, in cultural and social terms, as well as deeply unequal, in terms of
the distribution of income and opportunities. South Africans continue to see
themselves in the racial categories of the apartheid era, in part because these
categories have become the basis for post-apartheid ‘redress’, in part because
they retain cultural meaning in everyday life. South Africans continue to inhabit
social worlds that are largely defined by race, and many express negative views
of other racial groups. There has been little racial integration in residential
areas, although schools provide an important opportunity for inter-racial
interaction for middle-class children. Experimental and survey research
provide little evidence of racism, however. Few people complain about racial
discrimination, although many report everyday experiences that might be
understood as discriminatory. Racial discrimination per se seems to be of
minor importance in shaping opportunities in post-apartheid South Africa. Far
more important are the disadvantages of class, exacerbated by neighbourhood
effects: poor schooling, a lack of footholds in the labour market, a lack of
financial capital. The relationship between race and class is now very much
weaker than in the past. Overall, race remains very important in cultural and
social terms, but no longer structures economic advantage and disadvantage.
Post-apartheid South Africa is thus the precise opposite of Brazil.
Introduction
In a world in which racial labeling and discrimination are regrettably
commonplace, the South African system of apartheid stood out as an extreme
attempt to order a society explicitly and systematically according to racial
categories. Many aspects of apartheid were not unique to South Africa. In the
aftermath of slavery, colonial (and especially settler) societies in Africa
generally practiced racial segregation. In the USA and Brazil, most black
people were denied the vote through literacy and other qualifications (until the
1960s in the USA, and as late as the 1980s in Brazil). In much of Latin America
and the Caribbean, as well as the southern states of the USA, white people
owned the land whilst landless black people worked for them. Racism and
racial discrimination have been almost universal in the twentieth century. The
concept of ‘apartheid’ has even been applied to cities in the USA in the late
twentieth century. Indeed, apartheid – as implemented by the National Party
government in power from 1948 to 1994 – was built on the foundations of
racialised colonial and settler societies in which a minority of white settlers –
farmers and workers – lived amidst an indigenous or ‘native’ majority. What
made apartheid unique was its systematic depth and breadth, as the powers of a
modern state were deployed to order society along ‘racial’ lines, going far
beyond racism and racial discrimination to generalized social engineering
around state-sanctioned racial ideology and legislation.

It would be astonishing if post-apartheid South African society was not shaped

profoundly by the experience of apartheid, remaining distinctive in terms of the
social, political or economic roles played by ‘race’. Despite the rhetorical
commitment to non-racialism of the major ‘liberation movement’ (the African
National Congress), together with its allies inside the country, during the
struggle against apartheid, and despite the abolition of apartheid-era racial
legislation and the adoption of a widely lauded constitution, race does indeed
remain ever-present in contemporary South Africa. To a large extent this is due
to a deep-rooted and enduring consciousness of race in society. To some extent
it is due to factors that reflect choices made by post-apartheid political elites: the
use of the race card in public life, including in politics, and new policies of
racial discrimination involving, especially, affirmative action in employment,
with the stated objective of redressing the disadvantages experienced by non-
white South Africans (either collectively or individually) under apartheid.

Racial discrimination in economic life against black people has been largely
ended in South Africa. Some lingering discrimination by white employers
against black people no doubt persists, but it is probably more than offset by the
effects of affirmative action. Persistent racial inequalities reflect class
stratification rather than racial discrimination, as we have argued at length
elsewhere (Seekings and Nattrass, 2005). Income is distributed within the
African population almost as unequally as within the population as a whole, as
opportunities have expanded rapidly for many African people to move into
better-paid occupations at the same time as many others languish in poverty
because of poor schooling and chronic unemployment. Yet society remains
highly racialised. Inter-racial contact, yet alone marriage, remains very limited.

Telles (2005), in his recent study of Brazil, collates a wide range of data
showing how important race is, but in some rather than all respects. Contrary to
the ideology of ‘racial democracy’, racial discrimination seems significant in
economic life. Yet, in terms of identities and social interactions, Brazilians are
remarkably non-racial. Telles distinguishes between vertical relationships, in
which race is important, and horizontal ones, in which it is not. Post-apartheid
South Africa appears to be the opposite of this. The vertical dimension of
racism appears to have been largely eliminated (or perhaps even reversed), but
the horizontal dimension appears resilient (or perhaps has even increased, as
racial differences within the increasingly multi-racial middle class have grown
and become more visible). In this paper, I examine the evidence for this.

Apartheid and Democracy in South Africa

The foundation of apartheid was the system of racial categorization enshrined in
law by the 1950 Population Registration Act (and subsequent amendments).
The Act provided for all South Africans to be classified into one of three basic
racial categories:
A white person is one who in appearance is, or who is generally
accepted as, a white person, but does not include a person who,
although in appearance obviously a white person, is generally
accepted as a Coloured person. A native is a person who is in fact or
is generally accepted as a member of any aboriginal race or tribe of
Africa. A Coloured person is a person who is not a white person nor a
native.
Later, a fourth category – Indian – was added, for people of South Asian
descent, the label ‘native’ was replaced by the labels ‘Bantu’ and ‘Black’.1
Racial classification was recorded in official identity documentation. From
1970, the ‘black’ category was further sub-divided into ethnic or linguistic
groups (such as Zulu and Xhosa) (Christopher, 2002).

1
The four-category schema had been used in the censuses of 1921, 1936 and 1946.
This racial categorisation was largely ‘common-sensical’ and consensual, based
on agreed and broadly co-terminous factors (descent, language or culture, and
appearance). In difficult or contested cases, classification was not based on
either descent or purely biological markers. Instead, the cultural markers of
‘appearance’ and ‘general acceptance’ were most important. Whilst informal
‘rules’ about appearance – including about skin colour or hair – were used, they
were used inconsistently, and appearance was generally interpreted in terms of
social standing or class. Overall, judgements about social standing (friends,
work, name, dress, deportment, tastes) were most important in contested cases.
The 1951 national Population Census provided the first opportunity for mass
racial classification. Race was determined by census enumerators, who had no
specific expertise and received no special training. In ambiguous cases,
therefore, classification reflected the prevailing social prejudices of white
people. People could appeal to a Race Classification Appeal Board. Although
the Appeal Board tended to find in favour of the applicant, there were very few
appeals, reflecting the generally consensual basis of classification (Posel, 2001a,
2001b).

Ambiguous and contested cases generally involved the very small minority of
‘coloured’ people. ‘Coloured’ was a composite and diverse category including
the descendents of relationships between white and black people, the
descendents of ‘Malay’ slaves brought from South-east Asia (categorised
separately in 1951 but not thereafter), and (after 1970) descendents of the
indigenous Khoi and San who inhabited the Western Cape prior to the arrival of
either white or black people and did not speak Bantu languages. Whilst
segregation between white and black preceded apartheid, introducing
segregation between white and coloured was a primary objective of the
apartheid state.

Systematic racial classification was required because the apartheid project

entailed three broad objectives. The first was ideological: to maintain racial
purity by preventing the ‘mixing’ or ‘dilution’ of ‘white blood’. There should
be no inter-racial sex, and hence no inter-racial marriage. To prevent
temptation, there should be residential segregation by race. Pre-1948 legislation
prohibiting sex and marriage between white and black people was extended to
cover white and coloured people. The 1950 Group Areas Act led to the forced
removal of almost one million people, mostly coloured people removed from
mixed residential areas when they were declared ‘white’ areas. In hospitals,
patients were supposed to be handled by nurses of the appropriate racial group
(although almost all doctors were white). Segregation was extended to other
areas of social interaction: education (with separate schools and universities for
each racial group), transport (separate railway carriages), and most municipal
facilities such as parks and beaches. Where complete segregation was not
possible, partial segregation was implemented through providing separate
entrances and counters (at stations and post offices, for example).

The second objective was to ensure and then protect the privileged economic
position of the white minority. The apartheid state inherited policies that
reserved land for white ownership and better-paid occupations for white people
(through the ‘colour bar’). Under apartheid, these were enforced more
emphatically, but the emphasis shifted to racial discrimination in public
education. White children from poorer white families were provided with the
skills required to sustain a privileged position in the labour market. The massive
investment in the education of white children was so successful that the colour
bar became largely redundant. At the same time as protecting wages, the
apartheid state sought to protection the profits earned by white-owned capitalist
enterprises (including, especially, farmers). Whilst the cost of ‘white’ or skilled
labour was inflated under apartheid, the cost of ‘black’ or unskilled labour was
depressed. ‘Influx control’ policies and the pass laws restricted where black
people were allowed to live and hence work, ensuring not only that urbanization
among black people was curtailed (in stark contrast to most other parts of the
developing world at the same time), but also that white farmers were guaranteed
a supply of labour despite being unable to pay competitive wages.

The third objective of apartheid was to maintain the political dominance of the
white minority. In the 1950s, the apartheid state was preoccupied with
removing coloured voters from the existing common voters roll, but thereafter
the primary concern was the political threat posed by the already
disenfranchised ‘native’ or black majority, i.e. the ‘swart gevaar’ (or ‘black
danger’, in Afrikaans). The apartheid state sought to restrict the political rights
of black people to the native reserves, or ‘bantustans’ or ‘homelands’ as they
were later renamed. Powers were devolved first to appointed chiefs, and later to
semi-elected but largely compliant banstustan governments. In 1975, the
Transkei homeland was the first to become notionally ‘independent’, with its
citizens losing South African citizenship.

Apartheid not only ordered but also transformed South African society.
Discriminatory education and privileged family backgrounds provided white
children with the advantages of class, such that explicit racial discrimination in
the labour market became unnecessary. Economic and industrial policies
designed to reduce dependence on black workers and boost incomes for (white)
skilled workers and professionals resulted in capital-intensive growth at the
same time as rising unemployment (among unskilled black people), and thus
both inequality and poverty. Forced removals of the unemployed and their
dependents (i.e. the ‘surplus population’) from white-owned farms and towns to
the reserves resulted in massive over-crowding and the consequent destruction
of a smallholding agrarian society in those areas.

But apartheid was unable to transform the country’s demographics. The white
minority was too small to sustain economic growth: employers demanded semi-
skilled and then skilled black workers to produce goods and a larger pool of
consumers to buy them. From the 1970s, and especially in the 1980s, the
apartheid state began to dismantle racial discrimination in the labour market, to
invest more heavily in secondary schooling for black children, and to lift
restrictions on the growth of an urban black middle class and a stable, urban
working-class. ‘Petty apartheid’ – i.e. the segregation of parks and other public
facilities – was dismantled. Racial discrimination in public welfare was slowly
reduced. In 1984, the apartheid state reversed its earlier policies and sought to
co-opt the small coloured and Indian minorities into an anti-black coalition,
through racially segregated representation in a Tricameral Parliament (with
separate chambers for white, coloured and Indian Members of Parliament).
Finally, in the face of country-wide revolt from below and intensified
international pressures, the apartheid state began to negotiate an end to all
aspects of apartheid. In 1994, South Africans – including in the notionally
independent homelands – voted in the country’s first non-racial elections.

Much of the apartheid-era racial legislation was abolished during the transitional
period, i.e. prior to 1994, and the post-apartheid government led by the ANC
completed the process. Despite the abolition of the Population Registration Act,
however, a battery of new legislation has been introduced providing for racial
discrimination in favour of non-white, and especially African, people. The 1998
Employment Equity Act and 2003 Black Empowerment Act have the effect of
requiring private employers to transfer equity to new owners and to appoint new
employees on the basis of race, privileging the members of ‘designated groups’,
especially African people. The criteria for racial group classification are not
defined, and employers (as well as educational institutions that discriminate by
race in student admissions) seem to shift uneasily between allowing people to
self-classify and threatening that ‘false’ self-classification is a disciplinary
offence!2 The ideological, political and economic objectives of post-apartheid
policies of racial discrimination remain opaque: the effect is to accelerate very
rapidly the growth of an African elite and middle-class, but it is unclear what is
the relative importance of a concern to stabilize capitalist democracy, Africanist
ideology, social justice and the redress of previous disadvantage, and
straightforward greed and self-interest on the part of the new political elite.

2
This is an important topic for further research. Cases of ambiguous classification are likely to be more
common among the current and aspirant middle classes than among the population as a whole. It is said that
members of the Constitutional Court worry that racial classification would be contested and they would have to
formulate an explicit approach to the matter.
Race and Identity after Apartheid
Post-apartheid South Africa is characterized by the dual legacies of apartheid:
cultural diversity and economic inequality that both have racial characteristics.
Apartheid entrenched racialised identities and fostered racial division at the
same time as exacerbating inequality in the distribution of income. The post-
apartheid state faces the challenge of tackling these legacies of racialised social
engineering.

South Africans tend to see their society in racialised terms. Asked about racial
identities, only a tiny proportion of South Africans aver the apartheid-era
categories of ‘African’ (replacing, in official use, the former label ‘black’),
white, coloured and Indian. How people categorise themselves also accords
closely with how (they say) other people see them.3 In other words, there
remains a close correlation between official apartheid-era racial classifications,
post-apartheid self-classification and post-apartheid classification by other
people. This does not mean, however, that these are the only identities that
South Africans have. Asked who they are, South Africans will often say they
are South Africa, and that they are proud of this (see Grossberg et al., 2006).
Increasingly, they are likely to employ class identities (working class, middle
class, poor). And many also use non-racial cultural identities, including
religious ones (Christian, Muslim) or ethnic ones (Xhosa, Zulu, Afrikaans).

Self-description of skin colour by self-

classification of race

50

40
African
percentage

30 Coloured
20 White

10 Black

0
0 1 2 3 4 5 6 7 8 9 10
skin colour (0=very pale, 10=very dark)

Note: ‘African’ and ‘black’ were categories chosen by people who were or would have been classified
as black under apartheid; they are reported separately here because it is possible that self-assessed skin
colour correlates with the choice between ‘African’ and ‘black’. Source: 2005 Cape Area Study.

Figure 1: Self-assessed skin colour by self-reported race, Cape Town, 2005

3
This and the next paragraph are based on the findings of the 2005 Cape Area Study, conducted in Cape Town
(see Seekings et al., 2005).
Racial categories themselves mean different things to different people. Asked
for reasons for their racial self-classification, white South Africans typically
refer to their physical appearance or descent. African and coloured South
Africans do not refer to physical appearance or descent, but instead emphasise
‘culture’ (and, in the case of coloured people, the categorization of the apartheid
period). There is little difference in the distribution of self-assessed skin colour
among African and coloured people, but white people see their skin colour as
distinctly paler (see Figure 1). (Interviewers were also asked to assess the
respondents’ skin colour, and their assessments accorded closely with the
respondents’ self-assessments). Whilst almost all South Africans use racial
categorization in everyday life, it seems to be white South Africans who hold
onto biological conceptions of race.

This is in part because white South Africans – like white people in many other
contexts – take their culture for granted. Culturally, whiteness is invisible to
most white people (Steyn, 2001). African people are much more conscious of
their cultural distinctiveness. Speaking different languages at home, often
attending different churches, and perhaps above all retaining distinctive beliefs
about, for example, ancestors, witchcraft (see Ashforth, 2005) and family
(Russell, 2003). The end of apartheid has also been accompanied by a
resurgence of coloured identity. Under apartheid, coloured identity was defined
by the intermediate status of coloured people in the racial hierarchy: aspirations
to assimilation into white society and fears of relegation to the status of African
people combined with widespread feelings of shame as well as marginality.
After apartheid, a racialised conception of ‘colouredness’ has grown stronger,
with renewed affinities to whiteness and deepened racism towards African
people (Adhikari, 2006).

In recognition of cultural diversity, post-apartheid nation-building in South

Africa employed the discourse of the multi-cultural ‘rainbow nation’ rather than
building a common non-racial South African national identity. The national
anthem thus combines elements of both the hymn associated with the liberation
movements and the apartheid-era anthem, and is sung in four major languages.
Official multi-culturalism serves, however, to reproduce the culturally-based
racial identities of the past.

In 1998, the then deputy-president Thabo Mbeki emphasized racial inequality in

a controversial speech in which he described South Africa as comprising ‘two
nations, the one black and the other white’.
One of these nations is white, relatively prosperous, regardless of
gender or geographical dispersal. It has ready access to a developed
economic, physical, educational, communication and other
infrastructure. This enables it to argue that, except for the persistence
 of gender discrimination against women, all members of this national
have the possibility of exercising their right to equal opportunity, and
the development opportunities to which the Constitution of 1993
committed our country. The second and larger nation of South Africa
is black and poor, with the worst-affected being women in rural areas,
the black rural population in general and the disabled. This nation
lives under conditions of grossly underdeveloped economic, physical,
educational, communication and other infrastructure. It has virtually
no possibility of exercising what in reality amounts to a theoretical
right to equal opportunity, that right being equal within the black
nation only to the extent that it is equally incapable of realization.
(Hansard, 29 May 1998, col. 3378)
Mbeki here was drawing on a tradition of referring to ‘two nations’ that
originated in mid-nineteenth century, class-divided Britain (in the writings of
Benjamin Disaraeli) and was popularized in still race-divided USA in the late
twentieth century.

For Mbeki, the project of ‘nation-building’ in South Africa entailed bridging the
divides between the racially-defined ‘nations’ above. ‘Nation building is the
construction of the reality and the sense of common nationhood which would
result from the abolition of disparities in the quality of life among South
Africans based on the racial, gender and geographic inequalities we all inherited
from the past’, he said in 1998 (quoted in Roefs, 2006: 77). Note that there is no
mention here of class inequalities. The implication is that the ‘national
question’ – defined in racial terms – has precedence above the ‘social question’
of class-based inequalities. Programmes of race-based affirmative action are to
have precedence over pro-poor or inter-class redistribution.

On average, white South Africans remain privileged after apartheid, and most
African people remain poor. But data on the average person within racial
categories ignores the rapid increase in inequality within those categories,
especially within the African population. The rapid growth of the African elite
and middle class, at the same time as unemployment locks many other African
people into chronic poverty, has resulted in incomes (and opportunities) within
the African population being distributed nearly as unequally as in South Africa
as a whole, and this inequality is as extreme as anywhere in the world. It would
be more appropriate to view South Africa in terms of three ‘nations’: the almost
entirely African poor, the mostly African working classes, and the multi-racial
middle-classes and elites (see Seekings and Nattrass, 2005).4

4
The two nations description is especially inappropriate in the Western Cape and Cape Town, where the
demographics of race and class are distinct.
The growth of the African middle class has been the result primarily of the
deracialisation of education and of the labour market, and secondly of
discriminatory post-apartheid policies of affirmative action. The public service
implemented affirmative action rapidly after 1994. The 1998 Employment
Equity Act required mid-sized and large private employers to set targets for the
transformation of their workforce and to report on their progress in achieving
these targets. The growth of an African elite has been the result primarily of
discriminatory policies of ‘black economic empowerment’ in business. The
2003 Black Economic Empowerment Act set in motion a massive redistribution
of corporate ownership from the old white elite to the emerging black elite. The
African middle class and elite are the agents of a reinvention of African culture:
‘African’ names and dress are adopted, and supposedly traditional rituals are
practiced with new-found fervour.

‘Post-apartheid’ South Africa is thus characterized by a paradoxical combination

of features. Race is no longer coterminous with class, with opportunities for
upward mobility opening rapidly for some African people whilst opportunities
remain limited for many others. Class is increasingly important. Racism has
almost certainly declined. Yet race retains its central position in identities and
culture, and political parties can and do continue to play the race card. The
priority attached to the rhetoric and policies of affirmative action suggests that
the national question takes precedence over the social question, but at the same
time African elites rhetorically recommit themselves to non-racialism and a
concern for the poor.

Maré has wondered how far the ‘ordinariness’ of racial consciousness (or ‘race-
thinking’) in post-apartheid South Africa is the consequence of post-apartheid
policies of racial categorization.
To meet with the requirements of the Employment Equity Act, to gain
admission to universities, to claim travel allowances, to play in sports
teams, to provide information for tax purposes, to ask the National
Research Foundation for funding, to register births and so on, each
requires a statement of race belonging. … There is no opportunity in
these forms to avoid the issue. At every level there is an official, from
the government minister responsible to the company personnel officer
or employment equity manager, to monitor adherence or compliance
or progress. No provision is made for alternatives to the basic ‘four
races’ of apartheid South Africa, or to reject such classification.
Leaving the space blank, which remarkably few seem to do, means
that someone else is required to complete it to balance the books.
(Maré, 2001: 82)
Even if the state was to abandon any such administrative categorization, it is
hard to imagine that South Africans’ acute consciousness of race would vanish.
South Africans’ racialised identities and perceptions of others have strong roots
in civil society. ‘Race-thinking’ persists uneasily alongside a strong
commitment to transcending the racial divisions of apartheid. It is the deep-
rootedness of race-thinking (among African as much as if not more than white
people) that makes debates over, especially, affirmative action, so complex and
invidious.

Racism and ‘Race Relations’

South African ‘race relations’ was a prominent topic for research until the mid-
1960s (MacCrone, 1949; Kuper et al., 1958; Crijns, 1959; van der Berghe,
1967), but thereafter social scientists largely neglected race in favour of Marxian
studies of class formation and conflict. Pyschologists continued to use social
distance scales in studies, mostly conducted among samples of university
students (see surveys in Foster and Louw-Potgieter, 1991), whilst a trickle of
studies began to explore the effects of inter-racial contact on attitudes in the later
apartheid period (Preston-Whyte, 1976; Foster and Finchilescu, 1986).

Since the end of apartheid, research has diversified in a range of new directions,
despite the fact that most people continue to live in residential areas that are, in
practice, racially segregated, and most children continue to attend schools with
children of the same ‘race’. Indeed, one of the most striking findings of post-
apartheid survey research is how few South Africans enjoy much inter-racial
contact. In a survey conducted in 2000-01 by James Gibson (see Gibson, 2004),
white, coloured and Indian respondents were asked a series of questions about
African people, and African people the same questions about white people.
Only 16 percent of the respondents, weighted appropriately, reported having ‘a
great deal’ of contact at work with members of the designated group, whilst only
6 percent reported having ‘a great deal’ of such contact outside work. Another
13 percent reported having ‘some’ such contact at work and another 13 percent
reported having some such contact outside of work. Eight percent said they ate
meals ‘quite often’ with members of the designated group. A tiny 4 percent said
they had ‘quite a number’ of friends in the designated group, with another 20
percent saying they had ‘only a small number’ of such friends. Overall, one in
three South Africans reported any of the above; two out of three South Africans
said that they had little or no contact with members of the designated group.
 Figure 2: Cross-racial contact, Cape Town, 2005
100

percentage 80

60

40

20

0
african coloured white
all of my 5 closest friends are [sam e race as me]

most of my 5 closest friends are [same race as me]

all or most of the 5 people at w ork w ith w hom I w ork m ost closely are
[same race as me]
I have not socialised w ith people w ho are not [sam e race as me] in past 7
days

Subsequent research in Cape Town – which has distinctive demographics, with

African and white minorities alongside a coloured near-majority – found what
appear initially to be rather higher levels of cross-racial contact (see Figure 2).
Only 40 percent of African and coloured respondents said that all of their five
closest friends were from the same racial group as them, and only about 20
percent of white respondents said the same. For sure, large majorities of
respondents in all three racial groups said that all or most of their closest friends
were from the same racial group, but the data suggest that, in Cape Town at
least, only a minority of the adult population moves in entirely mono-racial
social circles. Even smaller proportions of working people say that they work in
mono-racial working environments. But only a minority said that they had
actually socialized with people from other racial groups in the past seven days
(see the fourth column in each set of columns in Figure 2).

The apparently higher level of inter-racial contact in Cape Town probably

reflects two factors. The first is methodological: the survey (the 2005 Cape
Area Study) did not ask white and coloured respondents about their interaction
with African people specifically, nor did it ask African people about their
interaction with white people specifically. It is likely that much ‘cross-racial’
contact is between with white and coloured people, or between African and
coloured people, neither of which would have shown up in Gibson’s earlier
study. Secondly, it is likely that there are higher levels of cross-racial contact in
urban than in rural areas. There are certainly many more opportunities for such
contact in urban areas.
 % agreeing with statement

South Africa would be a better place without

any of them

I often don't believe what they say

They are racist*

I feel uncomfortable around them

It is hard to imagine ever being friends with

one of them

They are untrustworthy

It is diffiult to understand their customs and

ways

0 10 20 30 40 50 60 70

Note: responses to the racism question are actually disagreements with the statement ‘most … are not
racist’. Source: 2000-01 Truth and Reconcilation data-set; my own calculations.

Figure 3. The limits to social deracialisation, 2000-01.

Perhaps unsurprisingly, given the general lack of contact, Gibson found

evidence of inter-racial suspicion and distrust on the part of non-African
respondents towards African people and among African respondents to white
people. The bars in Figure 3 show the proportion of the weighted sample who
agreed (or agreed strongly) with each of seven statements about the designated
group. Almost one in five South Africans agreed that South Africa would be a
better place without the designated racial group. Almost half agreed that they do
not believe what members of the designated racial group said, that they feel
uncomfortable around them, and that they find it hard to imagine ever being
friends with one of them. Almost two out of three South Africans agreed that it
is difficult to understand the customs and ways of the designated group.
Without exception, larger proportions of African respondents agreed with these
statements about white people than did white, coloured or Indian respondents
when asked about African people (see Gibson, 2004: 123-4). The 2003 South
African Social Attitudes Survey (SASAS) similarly found that large minorities
of every racial group agreed that most members of their own racial group were
racist, whilst a large majority of African people thought that most white people
were racist and large majorities of white and coloured people thought that most
African people were racist (Roefs, 2006: 89-90).

Gibson shows that there is an inverse correlation between most forms of inter-
racial contact and racial distrust: the more contact that respondents report with
members of the designated group, the less likely they are to agree with
statements indicating prejudice or wariness. The exception to this is contact at
work, which has no significant effect on inter-racial attitudes. Gibson shows
that contact is especially important to white, coloured and Indian respondents
(ibid: 139-42).

Limited social deracialisation does not mean that there has been no perceived
improvement in race relations. A series of surveys have found that South
Africans believe that race relations have improved since the end of apartheid.
According to Gibson’s 2001 survey, 16 percent said that race relations had
improved a great deal, and a further 45 percent said that race relations had
improved somewhat. In 2003, a survey conducted by the Kaiser Family
Foundation, the Washington Post and Harvard found that as many as 68 percent
of South Africans believed that race relations were better than they had been
under apartheid, and as many expected that race relations would continue to
improve over the next five years (Hamel et al., 2004). The 2003 SASAS also
found that most African, coloured and Indian people (but less than half of white
people) said that race relations had improved since 1994 (Roefs, 2006: 90-1).
But the same surveys found evidence that improved race relations did not mean
good race relations. Race relations remains a pressing problem for 49 percent of
Gibson’s respondents and a further 33 percent described them as important.

Figure 4: Discomfort and friendship across racial

lines, Cape Town, 2005

100
80
percentage

60
40
20
0
african coloured white
I feel uncom fortable around people w ho are not [sam e race as m e]

I do not feel uncom fortable …

I cannot im agine ever being friends w ith people w ho are not [sam e race as
m e]
I can im agine being friends …

The 2005 Cape Town survey also probed extensively the range of respondents’
contacts and attitudes across racial lines. Figure 4 summarises the proportions
of respondents agreeing and disagreeing with the statements “I feel
uncomfortable around people who are not [same race as respondent]” and “I
cannot imagine ever being friends with people who are not [same race as
respondent]”. Very small proportions of respondents who white and coloured
respondents and only a small proportion of African respondents agreed with
either statement. (Again, the Cape Town survey suggests lower levels of
discomfort than Gibson’s national survey, probably for the reasons discussed
above).

African respondents

Coloured respondents

White respondents

Figure 5: Attitude toward a family member marrying someone, according to race of

respondent and of the prospective spouse

The Cape Town survey asked more testing questions about attitudes towards
cross-racial marriage. All respondents were positive about marriages to
members of their own racial group and relatively hostile to inter-racial marriage,
but they did not discriminate significantly according to the precise inter-racial
combination (see Figure 5). Thus African respondents were more-or-less
indifferent between kin marrying white and kin marrying coloured people (C),
coloured respondents were more-or-less indifferent between white (W) and
African (A), and white respondents were more-or-less indifferent between
coloured, Indian (I) and African.5 (Further detail is provided in the Appendix).
5
It is curious that there is not more research on inter-racial relationships in post-apartheid South Africa. There
would be value in both accurate quantitative research and qualitative research – including into the cultural milieu
of inter-‘racial’ households.
Without longitudinal data, it is hard to assess just how much attitudes have
changed over time. Largely anecdotal evidence suggests that explicit and overt
racism has declined. But it seems likely that a consciousness of racial difference
has been resilient. If so, this is surely linked to the limited extent of residential
desegregation or of racial integration in other spheres of life.

Inter-Racial Integration and Interaction

Patterns of residential segregation have not broken down to any great extent
since the transition to democracy. Analysing data from successive population
censuses, Christopher (2001) shows that South African towns and cities began to
desegregate, racially, in the 1990s. White segregation levels (as measured by a
standard segregation index) peaked in Cape Town in the 1991 census (and in
South Africa’s other major cities in either the 1985 or 1991 censuses).
Segregation levels for all racial groups declined between the 1991 and 1996
censuses. But the pace of desegregation was very slow indeed. Christopher
(2005) updates his analysis using 2001 census data. Slow desegregation
continues, but ‘the vast majority of the urban population continues to live in
highly segregated suburbs’ (2005: 267).

Indeed, most new housing areas established after the end of apartheid are as
segregated as the older neighbourhoods established (or remade) under the
apartheid Group Areas Act. ‘Choices’ about where to live are, of course,
severely limited by economic inequalities. Prices of housing and of land make it
almost impossible for low-income African households to move into middle class
suburbs. In the absence of detailed studies about ‘residential choice’, however,
it remains unclear how important are social networks (i.e. in what other areas do
households that might move have connections or friends) or social preferences
(i.e. to what extent do people prefer to live in neighbourhoods with other people
sharing a similar culture, that is to some extent coterminous with former racial
classifications). The fact that people were forced to move under apartheid does
not mean that now, after apartheid, people might be making choices that reflect
the tragic cultural legacy of apartheid itself.

Although most South Africans continue to live in mono-racial neighbourhoods,

there is a fast-growing literature on the atypical neighbourhoods where there is a
degree of racial integration. Residential desegregation a variety of forms, which
can be divided into two broad categories: Existing neighbourhoods can become
more ‘integrated’ or integration can occur through the construction of new, more
integrated neighbourhoods. Each of these categories can be divided into sub-
categories (see Table 1).
Table 1: Forms of residential integration
Category of integration Examples Studies
Integration in 1. Institutional residences (e.g. Many None
existing university residences, police
neighbourhoods barracks
2. Apartment blocks Hillbrow Morris (1999)
3. Private housing, through Ruyterwacht, Teppo (2004),
movement of higher-income African Muizenberg Lemanski
people into formerly coloured/white (2006c)
areas, or of coloured people into
formerly white areas
4. Private housing through the None (probably
movement of downwardly-mobile because
white people into formerly negligible).
coloured/African areas, or of
coloured people into formerly
African areas
Integration in new 5. State-driven low-income housing Hout Bay, Saff (1998),
neighbourhoods projects in high-income areas Noordhoek, Dixon et al.
Muizenberg, (1994)
Milnerton
6. Planned ‘integration’ in new Delft South, Oldfield (2004),
state-financed and state-allocated Westlake Lemanski
housing projects (2006b)
7. Voluntary ‘integration’ in new Summer Greens Broadbridge
private housing areas (2001)

All of South Africa’s major cities have central areas of apartment blocks which
were formerly reserved for white occupation but into which there has been very
substantial in-migration. The classic study of such an area is Morris’ account of
Hillbrow, in Johannesburg, which began to desegregate as early as the late
1970s (Morris, 1999; see also Crankshaw and White, 1995). In Hillbrow, black
in-migration prompted white flight, so that racial integration was temporary
only. Such cases presumably harden rather than dissipate racial animosity or
ambivalence.

Variant 5 also fosters relatively little actual integration. Cases of low-income

housing developments in middle-class suburbs, where land prices are high,
attract considerable publicity and controversy. In Cape Town, there are just four
small areas of this sort (Imizamo Yethu in Hout Bay, Masiphumelele in
Noordhoek, Westlake near Muizenberg, and Marconi Beam in Milnerton). In
each case, the low-income settlements grow far faster than expected and there
are chronic problems in providing minimal services and infrastructure (yet alone
upgrading the area). Residents of neighbouring low-income and middle-class
areas rarely share transport (the former use public transport, the latter private) or
even shops (with different supermarkets catering for different groups of
consumers), and almost never interact socially. But it is impossible to assess
whether negative responses from the existing, white middle-class population are
due to their race or class (for different interpretations, see Saff, 1998; Dixon et
al., 1994).

This leaves variants 3, 6 and 7, i.e. desegregation in old or new neighbourhoods.

There are two case-studies of variant 3 from Cape Town, of a formerly poor
white area (Ruyterwacht) and of a mixed-income, formerly white area
(Muizenberg). In Ruyterwacht – which was the site of racialised and possibly
racist protests in the early 1990s (Jung and Seekings, 1997) – low housing prices
made it attractive for upwardly-mobile coloured households. The non-white
population rose from almost zero in the 1980s to 23 percent by 1996 and an
estimated 40 percent by 2000. Most of the new residents were young, coloured
families, with small children. They were better educated and had higher
incomes than most of the existing white residents, and their houses were
noticeable for their newly-built garages, second stories, swimming pools and
satellite dishes. Many of the new residents are Muslim, which might be
expected to add another element of discord in a neighbourhood that was hitherto
exclusively Christian. But a typical comment from their white neighbours was
that ‘here, our coloureds are good’ – expressing both an engagement with post-
apartheid realities and an inability to discard entirely the racial discourses of the
past. Faced with two incidents of rape, white and coloured neighbours joined in
a neighbourhood watch. In Teppo’s account, ‘hierarchist’ white residents and
their new coloured neighbours ‘mbrace one another across racial lines, perhaps
reluctantly at times, but knowing full well it is the only choice for both groups if
they wish to keep their suburb secure’ (Teppo, 2004: 231).

Muizenberg, on the False Bay coast, shows equally dramatic transformation

(Lemanski, 2006c). Between the 1996 and 2001 censuses, the white share of the
population fell from three-quarters to only just over one-third. Muizenberg had
become an extraordinarily mixed area, with very similar proportions of white,
coloured and African (including immigrant African) residents – although racial
desegregation has been concentrated in the less expensive sections. In
Muizenberg, racial desegregation did not lead to social interaction and
integration. Most white children attend non-local schools further down the
Peninsula, use their cars to shop outside the area, and tend to socialise
elsewhere. The long-standing, richer, white residents view local facilities as
deteriorating as a result of racial and class desegregation. Churches provide rare
sites for racial interaction, although cross-racial membership need not mean that
there is much inter-racial interaction.

Summer Greens is an example in Cape Town of variant 7, i.e. a new, middle-

class residential area into which have moved people from different racial groups.
Summer Greens is home to a lower middle class that combines both upwardly
mobile African and coloured families, and downwardly mobile white ones
(Broadbridge, 2001). By 2000, about one half of the suburb’s residents were
coloured, one-third white, and the rest African. The developers encouraged a
‘village atmosphere’ and sense of community: residents were not allowed to
build high walls (instead, the entire suburb is walled), and the streets were
designed to be public spaces rather than just transport routes. Residents,
especially those who moved here from coloured or African areas, were very
positive about security, but few had close contact with their neighbours. White
residents complained about ‘low class behaviour’ by people who ‘aren’t very
sophisticated, you know’, as well as of cultural differences. One white man did
not want his children to play with African children, because of the risk of AIDS.
There were few facilities in the suburb that could facilitate inter-racial
interaction. The most positive interpretation is that racial integration is a slow
progress, especially in middle-class suburbs where people emphasise their own
privacy and have limited interaction with any neighbours.

The picture in low-income areas is even less encouraging. Researchers have

studied two areas in Cape Town where the state explicitly sought to mix low-
income African and coloured people in public housing projects: Delft South
(Oldfield, 2004) and Westlake (Lemanski, 2006b). In Delft South, residents’
social lives, shopping patterns, and schooling divided along racial lines, as
African and coloured residents continued to utilise the racially-defined networks
and facilities that they had used prior to moving into a desegregated residential
area.6 Neighbours interacted more in Westlake, however, perhaps because the
neighbourhood’s location, surrounded by wealthy suburbs with expensive
facilities, meant that poor African and coloured residents were compelled to use
the same facilities because these were the only ones they could afford to use..

Overall, very few South Africans live in racially integrated neighbourhoods, and
few of those that do so live in neighbourhoods that can be described as
meaningfully integrated across racial lines. Even when the market or the state
throws people from different racial groups together in a neighbourhood, there is
little interaction, and racial othering and prejudice remain commonplace.

Whilst there has been very little residential desegregation, there has been some
desegregation of schools. This has been made possible by the absence of zoning
restrictions. Large numbers of young people do not attend the closest schools,
and some undergo lengthy commutes to attend schools that are distant from their
homes. Overall, school desegregation remains modest: Most African children
attend schools in townships or rural areas where all of the other children are also
African. It is only a small and fortunate minority that is able to get access to the
better schools found in formerly coloured and, especially, formerly white areas.
6
My own interviews with residents of Delft indicate a deep racial divide between African and coloured
neighbours.
But a minority of children – including many white, Indian and coloured
children, together with a small minority of African children – do undergo the
experience of attending a multi-racial school. The pioneering study of the racial
desegregation of a public school by Dolby (2001), based on research in a
Durban high school in 1996. Dolby’s subject school had been a white school in
a lower-income white residential area. By 1996, two-thirds of the students were
black. The one-third that were white were typically poor-performing students
from poorer homes who could not escape to better, more expensive schools.
Almost all of the teachers, however, were white, and they sought to preserve the
school’s ‘white’ identity through, for example, compelling boys to play rugby
and wear the school blazer. Students group along racial lines on the school
field. Racial epithets are common in the corridors. The combination of
threatened or resentful white teachers and students, and a growing majority of
black students, made for an explosive mix.

The environment at Dolby’s school was very highly racialised, but school pupils
did not simply reproduce an apartheid-style conception of race in terms of
biology, history or past culture. Rather, they renegotiated race around the
dynamics of taste (especially clothes, music and clubs). Black pupils define
blackness not in terms of Zulu tradition (‘this is the [19]90s’, one girl protests to
a white teacher who anachronistically imagines that Zulu-speaking girls attend
the ‘traditional’ Zulu reed dance) but in terms of global African-American
culture (and icons such as Michael Jackson, Michael Jordan, Whitney Houston,
and top rap-artists). But coloured students also draw on African-American
fashion in defining their cultural identity: Levi jeans, Dickies chino pants,
baseball warm-up jackets. White students draw on different global influences:
more sexy clothing, techno music, and so on. Black, Indian, coloured and white
students value different clothes, listen to different music, play different sports,
go to different clubs. Deviants – such as an African girl who wears clothing
perceived as ‘white’ – are humiliated or ostracized. By policing the adopted
cultural markers of race, the students themselves actively reproduce racial
difference and division.

Similar research was conducted in schools in Cape Town (Soudien, 1998, 2004)
and Johannesburg (Dawson, 2003). Most recently, Gooskens (2006) digs into
this same vein of research, drawing on fieldwork conducted among children
attending a formerly white school in southern Cape Town. Gooskens also finds
that perceptions of similarity and difference are based on gender, lifestyle, class,
religion, moral values and language, rather than race per se – although race
remains ever present in their thoughts and language. In Gooskens’ account,
adolescents blend a rejection of racial identification or categorization with racial
name-calling.
Malls are an important site of inter-racial interaction. Nkuna (2006) describes
how young people of all races try to construct a new multi-racial identity in The
Zone, a mall in Johannesburg. The identity is based on similar tastes in clothing
(specific brands), music, and other markers of fashion (hair, body piercing).
Because these entail spending money, this is a middle class world. Almost half
of the young people at The Zone are students in higher education, and almost all
attended private schools or formerly white schools in the suburbs. Very few
attended township schools. The culture is not only consumerist, but embraces
American styles whilst rejecting ‘parochial’ South African fashions (such as
kwaito music).

Survey data suggest that the workplace is an important site of inter-racial

interaction. The 2005 Cape Area Study found that very small minorities of each
race group said that they worked in mono-racial environments (see Figure 2
above). Unfortunately, there appears to be almost no published research on the
everyday reality of inter-racial interactions in the post-apartheid workplace.
Recent studies provide only snippets of information or analysis on race. Von
Holdt (2003, 2005) provides a compelling account of the importance of race in
understanding dynamics in a steelmill at the end of apartheid, but his research
stops in the mid-1990s. He emphasizes how the job colour bar and racial
segregation of facilities were accompanied by the routine use of violence to
maintain white baaskap (supremacy).7 Bezuidenhout (2005) describes four
engineering factories, where the basic racial order has barely changed with the
transition from apartheid to democracy, although there is no suggestion that it is
maintained now through violence. He also points to the racialised perception
among African workers that they are still discriminated against, even though this
discrimination is now based on seniority within the company rather than
(explicitly) race. Kenny’s (2005) study of supermarkets points to the
importance of white security personnel in perpetuating a racial order. The most
interesting window into race in the workplace is in research by a black scholar in
a car-manufacturing plant. ’Ordinary workers’, Masondo writes, ‘feel more
comfortable’ in mono-racial social groups, especially because of linguistic
barriers. But black and white salaried staff do interact: they ‘are always together
during lunch hours’ (Masondo, 2005: 165). Black workers complain that the
new black managers ‘treat us the same way as the white managers did. They
shout at us as if we are their children.’ (168).

Racial interactions in the workplace clearly warrant further research. To what

extent, or in what ways, or how often, do inter-racial interactions transcend
workplace hierarchies? What are the consequences of racialised hierarchy (or
its erosion) in the workplace for the reproduction (or erosion) of race-thinking?

7
Carrim (1986) provided a much more optimistic account of African-Indian interactions.
Discriminatory Attitudes and Experiences
Ethnographic research in neighbourhoods and schools suggests that racial
differences and divisions remain pronounced, but finds little evidence of the
kind of brutal racism associated with white South Africans in the early apartheid
period. Experimental research on behaviour and survey –based research on
attitudes supports this assessment. Experimental research on race has been
pioneered in South Africa by Justine Burns. In one of Burns’ experiments,
secondary school students in Cape Town played the ‘dictator game’, in which
players are given money and then choose how much to pass onto anonymous
‘partners’, whose photo they have seen but otherwise know nothing about.
Using a photo allowed Burns to test for the effect of the partners’ race, or at
least race in terms of physical appearance. Burns found that there was no direct
race effect, i.e. that players did not discriminate against partners who appeared
to be racially different. This behaviour appeared to be motivated by an aversion
to inherited inequality, and racial appearance was taken as a proxy for inherited
inequality (Burns, 2004).

The participants in Burns’ experiments know that they are in an experiment, and
this might affect their behaviour. The participants do not know, however, that
inter-racial interaction is the focus of the research. In surveys, respondents
might also select responses in the knowledge that they are being researched, but
the use of ‘vignettes’ can help to disguise the focus of the research.
Respondents are presented with one or more vignettes describing a situation,
followed by a question or series of questions related to the situation. Sniderman
and Piazza (1993) used vignettes to examine ‘modern’ forms of racism in the
USA. In their ‘laid-off worker’ vignette, respondents were presented with a
scenario in which a person (or subject) is retrenched, and are then invited to
suggest how much (if any) financial assistance that person should receive from
the government whilst looking for work. The scenario varies insofar as the
subject (or retrenched person) is given different characteristics: white or black,
male or female, younger or older, single or married, with or without children,
and dependable or not dependable. The 2003 Cape Area Study, conducted with
a small sample in Cape Town, employed a variant of the ‘laid-off worker’
vignette to probe the effects of race on perceptions of distributive justice By
including a range of characteristics for each subject, the respondent’s attention is
being diverted in part at least from the racial characteristic.

Questions about distributive justice are a telling test of one dimension of racial
attitudes because the official ideology of apartheid emphasised that each racial
group (and each ethnic group within the African population) should look after its
own: white South Africans were not responsible for the poverty of black South
Africans; rich South Africans were only responsible for poor South Africans if
they were members of the same racially-demarcated ‘community’. One might
expect that the over-riding racialisation of society under apartheid and the
continuing salience of race have resulted in a close correlation between race and
attitudes toward distribution or distributive justice. The government, African
National Congress and the media frequently accuse white South Africans of
being opposed to ‘transformation’, i.e. to redistributive social and economic
policies. If this was the case, then we would expect to find that South Africans
will assess the desert of other members of their own racial group (i.e. ‘insiders’)
more favourably or positively than that of members of other racial groups (i.e.
‘outsiders’).

The results of the ‘laid-off worker’ experiment in Cape Town in 2003 suggested
that the race of the respondent and the race of the subject were of little import in
whether a respondent considered a subject deserving. For example, white
respondents did not discriminate significantly against African or coloured
subjects. But there were clear (and counter-intuitive) race effects on the amount
that the respondent said that the subject should receive per month from the
government. White respondents were more generous, perhaps because they had
a more inflated view of what constituted a ‘minimum’ income; more curiously,
black and coloured respondents as well as white respondents suggested that
larger grants be made to white than to African or coloured subjects (Seekings,
2005).

The 2005 Cape Area Study extended this vignette (as well as asking several
other vignette-style questions), with a larger sample, but also confined to Cape
Town. Instead of asking about the scenario of a retrenched worker, respondents
in 2005 were presented with a wider range of circumstances in which a subject
might be considered deserving of financial assistance. Respondents were first
told that ‘The government provides grants to some people in need, for example
old-age pensions to elderly people. I am going to describe a situation, and then
ask you what the government should do to help the person involved.’ A specific
subject was then described. The subjects varied between interviews. Firstly, the
general circumstances of the subject varied. Some subjects were described as
retrenched workers, others as people who were sick; some were disabled and
others abandoned by husbands; and so on. A range of other social and
demographic characteristics – including race – were varied also. The 2005 data
showed most of the same patterns as the 2003 data: the race of the subject made
little or no difference, white respondents were a little less positive in their initial
assessment of desert, but a lot more generous in the sums they ‘awarded’.
Unlike 2003, there was no indication that respondents were more generous to
white subjects. In this dimension of social attitudes, race plays little effect, and
there is no little or no evidence of racism or racial discrimination. These
findings from survey vignettes are not dissimilar to Burns’ findings using data
from field experiments.

White South Africans have no qualms, however, in expressing opposition to

race-based policies, such as affirmative action and BEE. Several studies suggest
that there is wide and strong support for government interventions to help the
poor, but only among African people is there a majority in favour of race-based
affirmative action (in employment), black economic empowerment or
redistribution of land (ILO, 2004; Roberts, 2006).

Survey data suggest that most South Africans believe that ‘race relations’ have
improved since the end of apartheid, and neither surveys nor field experiments
provide evidence of significant racial discrimination in attitudes or experimental
behaviour. But discrimin-ation might persist in other domains, and it is even
more likely that discrimination is perceived as continuing.

The 2005 Cape Area Study asked about recent experiences of discrimination.
Respondents were asked whether, in the five years since 2000, they had ‘been
treated worse than other people or benefited’ because of their race. Very few
respondents said that they had experienced negative racial discrimination (see
Figure 6). Most African respondents said that they had benefited because they
were black, whilst most coloured and white respondents said that they had
neither benefited not been treated worse.

Figure 6: Experiences of discrimination in past 5 years, by

race, Cape Town, 2005

80

70

60

50

% 40

30

20

10

0
african coloured white total

treated worse both benefited neither

This was followed by a series of questions about experiences in specific settings

(see Figure 7). Almost all African respondents reported that they had been
watched or followed in shops, compared to a minority of coloured respondents
and very few white respondents. Larger proportions of African respondents than
coloured or white respondents reported experiencing each of the other four
situations (being treated with less respect, being treated worse in restaurants and
shops, being treated by people as if they were afraid of you, and being treated by
other people as if they were better than you). But in these other four situations
the proportions of African, coloured and white respondents who reported that
they had had the experience were not massively different. For example, just
over one half of African respondents reported being treated with less respect,
compared to over one-third of coloured and white respondents.

These results are broadly consistent with the findings of the countrywide 2003
SASAS. Most South Africans reported that they never feel that they are being
discriminated against. A larger minority of white and Indian people report
experiencing discrimination than among African or coloured people.
Discrimination is perceived as occurring primarily at work (especially by
African people), when applying for jobs (especially among white and coloured
people), and in shops (especially among white people, curiously) (Roefs, 2006:
88-9).

Figure 7: Experiences of discrimination, Cape Town, 2005:

Respondents reporting that they have often or sometimes been treated ...

80

70

60

50

% 40

30

20

10

0
african coloured white total

with less respect worse in restaurants and shops as if they were afraid of you
as if they were better than you watched or followed in shops

Unfortunately there is no experimental research in South Africa similar to the

work conduced by Pager in the USA, assessing the extent and patterns of racial
(or other) discrimination in the labour market. If such research was conducted,
however, it is likely that it would find that in occupations in which they are
applicants from all racial groups (i.e. excluding unskilled employment), racial
discrimination is practiced in favour of black applicants through affirmative
action and BEE policies.
Discrimination and Disadvantage
The effects of race in the labour market were much studied in the early 1990s.
From the 1920s to the 1970s, racial discrimination generally confined African
people to low-paid occupations. When African and white people were in the
same occupation – for examples, teaching and the police – white employees
were paid more than their African counterparts. But this picture began to
change dramatically from the 1970s. Crankshaw (1997) demonstrated the
steady, and at times rapid, rise of African workers into better-paid occupations
in the later apartheid period. Moll (2000) showed that the share of inequality in
the distribution of wages that was accounted for by inter-racial differences
declined from 65 percent in 1980 to 42 percent in 1993, whilst the share
accounted for by intra-racial differences rose from 35 percent to 58 percent. The
racial wage gap declined but still remained large, with median earnings for
African workers only about one-quarter of the median for white workers, in
1995 (Bhorat and Leibbrandt, 2001: 83; see also Burger and Woolard, 2005:
19). But a series of studies demonstrated that this persistent racial wage gap was
due primarily to differences in education, skill, location (urban/rural), and
economic sector, rather than by racial discrimination per se. Moll (2000) also
found that racial discrimination amounted to 20 percent of mean African wage
in 1980 but just 12 percent in 1993. Whilst several other studies used data from
the mid-1990s to re-examine racial wage discrimination, there has been a dearth
of studies using post-1995 data. Several recent studies of the labour market pay
no attention at all to racial discrimination, focusing instead on the
unambiguously pressing topic of unemployment and job creation (e.g. Burger
and Woolard, 2005; Oosthuizen, 2006).8

There appears to be just one study of racial discrimination using post-1995 data.
Burger and Jafta (2006) uses a set of decomposition techniques on data from
OHSs and LFSs between 1995 and 2004 to assess changes over time in the
‘unexplained’ part of the racial gap in both racial employment and (formal
sector) wage gaps. – with ‘unexplained’ meaning unexplained by other readily
measured variables such as years of schooling and location. They find that there
has been a narrowing of the racial wage gap since 1994 at the top end of the
wage distribution, but not overall. The unexplained element remains, i.e. being
white apparently continues to earn a premium in the labour market, essentially

8
Some studies persist in the tradition of Mbeki’s “two nations” analysis. Moleke (2006), for example, concludes
that ‘race still matters’ and ‘the labour market is far from being deracialised’ on the basis of evidence of racial
inequalities, i.e. differences in living conditions, skills and so on between the ‘average’ white person and average
African person. This kind of analysis implicitly defines racialisation in terms that render it impossible to
deracialise the labour market: given persisting inequalities within the African population, ‘deracialisation’ would
require reducing the skills and living conditions of a large portion of each generation of white South Africans to
the levels of their most disadvantaged African compatriots.
because white people earn higher returns on their education than do Indian,
coloured or, especially, African people.

As most of these econometric studies emphasise, returns to education surely

continue to vary by race because of the enormous but unmeasured differences in
the quality of education, combined with the similarly unmeasured benefits that
social capital bring to young people from middle-class backgrounds. Taking
such factors into account would surely reduce considerably the ‘unexplained’
component of the racial wage gap, and reduce further the importance of racial
discrimination relative to other factors such as inequalities in real skills and
useful contacts.

Burger and Jafta’s work points to the importance of distinguishing between

different sections of the labour market. Unlike in (say) Brazil or the USA, there
are few unskilled white workers competing with black workers for low-paid
employment (and of the small number of unskilled white workers, some might
have hidden class advantages, for example young people with part-time jobs as
waitresses). It is at the top end of the labour market that the effects of persistent
racial discrimination against African people or affirmative action in their benefit
would be concentrated. There are unfortunately few studies of the top end of the
labour market, especially among young entrants. But some data suggests that, in
some sectors, patterns of discrimination have changed markedly over very short
periods of time. In the late 1990s, the public sector was the primary venue for
affirmative action. The proportion of public sector managers who were African
rose from 30 percent in 1995 to 51 percent in 2001. The proportion of senior
managers who were African rose from 33 percent to 43 percent (Thompson and
Woolard, 2002). As many as 70 percent of all African graduates get their first
job in the public sector (cited in Altman, 2006: 69). In the early 2000s,
legislation has pushed larger private sector employers to similar shifts in
employment patterns. There has also been a dramatic shrinkage in the racial
wage gap among managers (cited in Altman, 2006).

A second, complicating factor in the analysis of racial discrimination in labour

market outcomes (i.e. employment and wages) using cross-sectional data is that
substantial numbers of younger white people have emigrated, or are at least
outside of the country for long periods of time. Whilst they might be outside the
South African labour market, their choice is probably not entirely exogenous to
conditions in the labour market. I am unaware of any studies that examine the
real effects of affirmative action legislation on the labour market for school-
leavers or, especially, university graduates, but there is no shortage of anecdotal
evidence that young white men and women believe that affirmative action
policies and practices are foreclosing opportunities for employment, and that
this perception influences decisions about emigration. If it was true that white
graduates are emigrating to avoid unemployment (perhaps because they would
‘choose’ unemployment over employment in occupations that are inferior to
those to which they aspired), then emigration would cause analyses of cross-
sectional data to under-estimate the effects of affirmative action.

Panel studies offer a promising way forward for the empirical analysis of
patterns and dynamics of advantage and disadvantage in post-apartheid South
Africa. One such study is the Cape Area Panel Study (CAPS), which
interviewed a representative sample of almost 5,000 young people (aged
between 14 and 22) in 2002, and has since re-interviewed this panel three times
(in 2003-04, 2005, and 2006) (see Lam et al., 2005). CAPS has collected
detailed data on schooling and entry into the labour market, as well as on sexual
and reproductive health and experiences within families and households. Most
data are collected from the young people themselves, but data are also collected
from parents and other older household members, and data on individuals and
households have been combined with community- and school-level data. The
problem with panel studies is that data-collection and -cleaning are so time-
consuming that there are inevitable delays before panel data are available for
analysis and longer delays before analyses are completed.9 A second problem is
that attrition results in incomplete data. Given the difficulties of collecting data
from or on emigrants, it is likely that CAPS will be able to offer only a partial
picture of how and why and with what consequences young people enter the
labour market. The South African Human Sciences Research Council is also
conducting panel studies among cohorts of matriculants (i.e. students writing the
grade 12 examination) and university graduates, which will provide a fuller
picture of what is happening at the top end of the labour market.

It is likely, however, that panel studies will confirm the following. Most
children from poor neighbourhoods – almost all of whom are African – grow up
in home environments that are unconducive to educational success, and attend
schools where the quality of education is very poor. Many remain in school
until their late teens, but are unable to acquire many skills. Their ability to find
employment is constrained by their lack of skills and experience, their location
far from most job opportunities, and their lack of contacts with jobs who could
help them find employment. Many move into the underclass of chronically
unemployed, with intermittent short spells of unskilled work. On the other
hand, children from middle-class neighbourhoods – who comprise rapidly rising
numbers of African as well as Indian and white children – attend better schools,
enjoy the benefits of middle-class home environments, and gain work
experience through part-time jobs (especially in school holidays). They move
into higher education and then into the labour market. White middle-class
9
Most data from the first three waves of CAPS became available publicly and for free, over the internet, in
December 2006.
children enjoy the relative benefits of wider and deeper social networks, but the
disadvantage of being white in an affirmative action environment.

Conclusion
The available evidence suggests that race remains of enormous social and
cultural importance despite a decline in economic importance. Earnings and
incomes reflect race far more than class. This raises questions, however, about
the meaning of class in the South African context. In its intellectual seedbed in
north-west Europe during the industrial revolution, there was generally a close
relationship between ‘objective’ class positions (in terms of relationships to the
means of production) and everyday cultures. As E.P.Thompson argued
famously, the working class was made culturally as well as through changes in
the form and shape of capitalism. In South Africa, ‘race’ – understood as a
social and cultural phenomenon, not biologically – has shaped profoundly
cultural change, interacting complexly with the growth of modern state and a
capitalist economy.

Insofar as this is the case, then South Africa would appear to be the opposite to
Brazil, where race is of limited cultural and even social importance but of
continuing economic signifance. In the terms used by Telles (2005), in Brazil
there is racism in terms of vertical relations but not of horizontal relations,
whilst in South Africa there is racism in terms of horizontal relations but not of
vertical relations. In Brazil, inter-racial marriage and racial discrimination in
employment are both common. In South Africa, after apartheid (and subject to
caution with respect to the extent and effect of affirmative action), neither is
common.

But the available evidence on post-apartheid South Africa is sadly limited.

Little progress has been made yet with respect to two key kinds of study. First,
data from panel studies is yet to be used to explore precisely how and why ‘race’
shapes progress through school and into the labour market. Secondly, there are
still too few studies of how race, class and culture are made and understood in
the lived experience of South Africans, at home, in neighbourhoods, in schools
and in workplaces.10 In addition, there is a dearth of empirical research on how
employers and others comply with official requirements to categorise people,11
and on the extent or effects of pro-African racial discrimination (i.e. affirmative

10
Such studies might usefully draw on Bourdieu.
11
Teppo (2004: 209-10) has a wonderful account of how school teachers do so: by getting the students
themselves, collectively, to count students by race.
action) in employment.12 In South Africa, as in Brazil, we are only beginning to
unravel the complicated interactions of race, class and culture in the
contemporary context.

12
The case of the University of Cape Town (UCT) might be instructive. The Employment Equity Act requires
employers such as UCT to determine targets for racial transformation, and to report on their progress in meeting
these targets. In the absence of will to retrench long-standing employees – i.e. staff appointed under apartheid –
employers have to achieve transformation through the slow pace of natural turnover. This means that most or all
new appointments have to be from designated groups. Whatever the rhetoric of formal policy declarations, the
reality of appointments is increasingly that young white men, and to a lesser extent young white women, cannot
be appointed to permanent positions. ‘Transformation’ is slow, but there is in effect a ‘colour bar’ on young
academic appointments, as there was under apartheid.
Appendix
Attitudes towards inter-racial marriage within family, by race of respondent
and race of prospective spouse
Apartheid-era racial Race of prospective spouse
classification of Black/African coloured indian white
respondent
mean 1.75 -0.15 -0.49 -0.16
African sd 0.75 1.49 1.44 1.47
n 408 408 408 408
mean 0.55 1.26 0.63 0.78
Coloured sd 1.21 0.96 1.17 1.13
n 478 478 478 478
mean -0.25 -0.02 -0.12 0.87
White sd 1.08 1.05 1.03 1.02
n 252 252 252 252
mean 0.68 0.64 -0.28 0.61
Other sd 1.52 1.33 1.52 1.35
n 35 35 35 35
Note: possible scores range from -2 to +2; negative scores indicate disapproval, positive scores
approval. sd standard deviation.
Source: Cape Area Study, 2005
References
Adhikari, Mohamed (2005), Not White Enough, Not Black Enough: Racial
Identity in the South African Coloured Community (Ohio UP, 2005; Double
Storey, 2006).

Altman, Miriam (2006), ‘Wage determination in South Africa: What do we

know?’ Transformation 60: 58-89.

Ashforth, Adam (2005), Witchcraft, Violence and Democracy in South Africa.

Chicago: University of Chicago Press.

Battersby, J. (2004), ‘Cape Town’s Model C Schools: Desegregated and

Desegregating Spaces?’, Urban Forum 15.3, 279-291.

Bezuidenhout, Andries (2005), ‘Post-colonial workplace regimes in the

engineering industry in South Africa’, in Edward Webster and Karl von Holdt
(eds), Beyond the Apartheid Workplace: Studies in Transition (Pietermaritzburg:
University of KwaZulu-Natal Press): 73-96.

Bhorat, Haroon, and Murray Leibbrandt (2001), ‘Correlates of Vulnerability in

the South African Labour Market’, in Haroon Bhorat, Murray Leibbrandt, Muzi
Maziya, Servaas van der Berg and Ingrid Woolard, Fighting Poverty: Labour
Markets and Inequality in South Africa (Cape Town: University of Cape Town
Press).

Broadbridge, Helen (2001), Negotiating Post-Apartheid Boundaries and

Identities: An Anthropological Study of the Creation of a Cape Town Suburb,
unpublished PhD thesis, Stellenbosch University.

Burger, Rolof, and Rachel Jafta (2006), ‘Returns to Race: Labour Market
Discrimination in Post-Apartheid South Africa’, University of Stellenbosch:
Stellenbosch Economic Working Paper no 04/06.

Burns, Justine (2004), ‘Inequality Aversion and Group Identity in a Segmented

Society’.

Carrim, Yunus (1986), ‘Trade Unionism in Natal: Shopfloor relations between

Indian and African workers’, South African Labour Bulletin 11,4: 44-64.

Christopher, A.J. (2001), ‘Urban Segregation in Post-apartheid South Africa’.

Urban Studies 38,3: 449-466.
_____ (2002), ‘”To define the indefinable”: population classification and the
census in South Africa’, Area vol.34, no.4;

_____ (2005), ‘Further Progress in the Desegregation of South African towns

and cities, 1996-2001’, Development Southern Africa 22,2 (June): 267-76.

Crankshaw, Owen (1997), Race, Class and the Changing Division of Labour
under Apartheid (London: Routledge).

_____ and Caroline White (1995), ‘Racial desegregation and the origins of
slums in Johannesburg’s inner city’, International Journal of Urban and
Regional Research 19,4: 622-38.

Crijns, A.G.J. (1959), Race Relations and Race Attitudes in South Africa
(Nijmegen: Drukkerij Gebr. Janssen N.V.).

Dawson, Marcelle (2003), Identity and Context: Friendship and Friction in a

South African High School (M.A. dissertation, Rand Afrikaans University).

Distiller, Natasha, and Melissa Steyn (eds, 2004), Under Construction: Race
and Identity in South Africa Today (Johannesburg: Heinemann).

Dixon, John, Don Foster, Kevin Durrheim and Lindy Wilbraham (1994),
‘Discourse and the politics of space in South Africa: the squatter crisis’,
Discourse and Society 5,3.

Dolby, Nadine (2001), Constructing Race: Youth, Identity and Popular Culture
in South Africa (Albany: State University of New York Press).

Foster, Don, and Joha Louw-Potgieter (eds, 1991), Social Psychology in South
Africa (Johannesburg: Lexicon).

_____ and Gillian Finchilescu (1986), ‘Contact in a non-contact society: The

case of South Africa’, in M. Hewstone and R. Brown (eds), Contact and
Conflict in Inter-Group Encounters (Oxford: Basil Blackwell).

Gibson, James (2004), Overcoming Apartheid: Can Truth Reconcile a Divided

Nation? (Pretoria: Human Sciences Research Council Press).

Gooskens, Imke (2006), ‘Boundaries and crossing points: Children, geography

and identity in the Fish Hoek Valley’. Social Dynamics 32,1 (June): …
Grossberg, Arlene, Jaré Struwig and Udesh Pillay (2006), ‘Multicultural
national identity and pride’, in Udesh Pillay, Ben Roberts and Stephen Rule
(eds), South African Social Attitudes: Changing Times, Diverse Voices
(Pretoria: Human Sciences Research Council): 54-76.

Hamel, Elizabeth, Mollyann Brodie and Richard Morin (2004), ‘A Survey of

South Africa Ten Years into the New Democracy’, paper presented at the 59th
annual conference of the American Association for Public Opinion Research,
Phoenix, Arizona (May).

ILO (2004), Economic Security for a Better World (Geneva: International

Labour Office).

Kenny, Bridget (2005), ‘The “market hegemonic” workplace order in food

retailing’, in Edward Webster and Karl von Holdt (eds), Beyond the Apartheid
Workplace: Studies in Transition (Pietermaritzburg: University of KwaZulu-
Natal Press): 217-41.

Klandermans, Bert, Marlene Roefs & Johan Olivier (eds) (2001), The State of
the People: Citizens, Civil Society and Governance in South Africa, 1994-2000
(Pretoria: Human Sciences Research Council).

Kuper, Leo, Hilstan Watts and Ronald Davies (1958), Durban: A Study of
Racial Ecology (London: Jonathan Cape).

Lam, David, Jeremy Seekings et al. 2005. ‘‘The Cape Area Panel Study
(CAPS): An Introduction and User’s Guide’ (Cape Town: Centre for Social
Science Research, University of Cape Town)

Lemanski, Charlotte (2006a), ‘The impact of residential desegregation on social

integration: Evidence from a South African neighbourhood’, Geoforum 37: 417-
435.

_____ (2006b), ‘Spaces of Exclusivity or Connection: Linkages between a

Gated Community and its Poorer Neighbour in a Cape Town Master Plan
Development’, Urban Studies 43,2 (February): 397-402.

_____ (2006c), ‘Desegregation and integration as linked or distinct? Evidence

from a previously “White” suburb in post-apartheid Cape Town’, International
Journal of Urban and Regional Research 30,3: 564-86.
MacCrone, I.D. (1949), ‘Race Attitudes: An Analysis and Interpretation’, in E.
Hellman (ed.), Handbook on Race Relations in South Africa (Oxford: Oxford
University Press).

Maré, Gerhard (2001), ‘Race counts in contemporary South Africa: “an illusion
of ordinariness”’, Transformation 47: 75-93.

Masondo, David (2005), ‘Trade liberalisation and work restructuring in post-

apartheid South Africa: A case-study of BMW’, in Edward Webster and Karl
von Holdt (eds), Beyond the Apartheid Workplace: Studies in Transition
(Pietermaritzburg: University of KwaZulu-Natal Press): 149-172.

Moleke, Percy (2006), ‘The state of labour market deracialisation’, in Sakhela

Buhlungu, John Daniel, Roger Southall and Jessica Lutchman (eds), State of the
Nation: South Africa, 2005-2006 (Pretoria: Human Sciences Research Council
Press): 202-22.

Moll, Peter (2000), ‘Discrimination is declining in South Africa but inequality is

not’, Journal for Studies in Economics and Econometrics 24,3: 91-108.

Morris, Alan (1999), Bleakness and Light: Inner-city transition in Hillbrow,

Johannesburg (Johannesburg: Witwatersrand University Press).

Norval, Aletta (1996), Deconstructing Apartheid Discourse (London: Verso).

Nkuna, Lucert (2006), ‘”Fitting-In” to a “Classy Place”: The Zone and Youth
Identity’, in Peter Alexander, Marcelle Dawson and Meera Ichharam (eds),
Globalisation and New Identities: A view from the middle (Johannesburg:
Jacana): 261-74.

Oldfield, Sophie (2000), ‘The centrality of Community Capacity in State Low-

Income Housing Provision in Cape Town, South Africa’, International Journal of
Urban and Regional Research 24,4: 858-72.

_____ (2004), ‘Urban networks, community organising and race: an analysis of

racial integration in a desegregated South African neighbourhood’, Geoforum
35.2: 189-201.

O’Meara, Dan (1996), Forty Lost Years: The apartheid state and the politics of
the National Party, 1948-1994 (Johannesburg: Ravan).
Oosthuizen, Morne (2006), ‘The post-apartheid labour market: 1995-2004’,
DPRU Working Paper no. 06/103 (Cape Town: Development Policy Research
Unit, University of Cape Town).

Posel, Deborah (1991), The Making of Apartheid, 1948-1961: Conflict and

Compromise (Oxford: Clarendon Press).

_____ (2001a), ‘What’s in a Name? Racial categorisations under apartheid and

their afterlife’, Transformation (Durban) no.47.

_____ (2001b), ‘Race and Common Sense: Racial classification in twentieth

century South Africa’, African Studies Review (October).

Preston-Whyte, Eleanor (1976), ‘Race attitudes and behaviouir: The case of

domestic employment in white South African homes’, African Studies 35,2.

Roberts, Ben (2004), ‘The Happy Transition? Attitudes to Poverty and

Inequality after a Decade of Democracy’, in Udesh Pillay, Ben Roberts and
Stephen Rule (eds), South African Social Attitudes: Changing Times, Diverse
Voices (Pretoria: Human Sciences Research Council): 101-30.

Roefs, Marlene (2006), ‘Identity and race relations’, in Udesh Pillay, Ben
Roberts and Stephen Rule (eds), South African Social Attitudes: Changing
Times, Diverse Voices (Pretoria: Human Sciences Research Council): 77-97.

Russell, Margo (2003), ‘Are urban black families nuclear? A comparative study
of black and white South African family norms’, Social Dynamics 29,2 (winter):
153-76.

Saff, Grant (1998) Changing Cape Town: urban dynamics, policy and planning
during the political transition in South Africa (Washington DC: University Press
of America).

Seekings, Jeremy (2005), ‘The Colour of Desert: Race, Class and Distributive
Justice in Post-Apartheid South Africa’, CSSR Working Paper no.126 (Cape
Town: Centre for Social Science Research, University of Cape Town).

_____ and Nicoli Nattrass (2005), Class, Race and Inequality in South Africa
(New Haven: Yale University Press).

_____, with Tracy Jooste, Mirah Langer and Brendan Maughan-Brown (2005),
‘Inequality and Diversity in Cape Town: An Introduction and User’s Guide to
the 2005 Cape Area Study’, CSSR Working Paper no.124 (Cape Town: Centre
for Social Science Research, University of Cape Town).

Sniderman, Paul and Thomas Piazza (1993), The Scar of Race (Cambridge MA:
Belknap Press).

Soudien, Crain (1998), ‘“We Know Why We’re Here”: The Experiences of
African Children in a “Coloured” School in Cape Town’, Race, Ethnicity and
Education 1,1: 7-29.

_____ (2004), ‘Constituting the Class: An Analysis of the Process of Integration

in South African Schools’, in Linda Chisholm (ed.), Changing Class: Education
and Social Change in Post-Apartheid South Africa (Pretoria: HSRC Press).

Steyn, Melissa (2001), Whiteness Just Isn’t What It Used To Be: White Identity
in a Changing South Africa (Albany: SUNY Press);

Telles, Edward (2005), Race in Another America: The Significance of Skin

Color in Brazil (Princeton: Princeton University Press).

Teppo, Annika (2004), The Making of a Good White: A Historical Ethnography

of the Rehabilitation of Poor Whites in a Suburb of Cape Town (PhD thesis,
University of Helsinki).

Van der Berghe, Pierre (1967), Race and Racism: A Comparative Perspective
(New York: John Wiley).

Von Holdt, Karl (2003), Transition from Below: Forging Trade Unionism and
Workplace Change in South Africa (Pietermaritzburg: University of KwaZulu-
Natal Press).

_____ (2005), ‘Political transition and the changing workplace order in a South
African steelworks’, in Edward Webster and Karl von Holdt (eds), Beyond the
Apartheid Workplace: Studies in Transition (Pietermaritzburg: University of
KwaZulu-Natal Press): 45-71.