1.1 Harmony Through Separation

‘Harmony through separation’. What a fantastic ideal! The very words kill each other, and the phrase presupposes a physical means to a spiritual goal (Paton, 1958, p. 16).

No other country has embarked on so thorough a reorganisation of its urban space for the purposes of segregation as South Africa. The apartheid policies pursued by South African governments between 1948 and 1994 were justified in terms of minimising friction between culturally distant population groups. Harmony, in so disparate a multi-racial society, was held to depend on separation. The groups in question comprised the majority African population (divided by apartheid planners into its major ethnic components, enabling the claim that South Africa had no majority group), the politically dominant white minority (comprising mainly Afrikaans and English-speaking elements, but regarded as a single group by the same planners), a coloured or mixed-race minority and a smaller Asian or Indian minority.

Such friction could, it was argued, be avoided only by separation, which apartheid sought to achieve at three spatial scales. Micro-scale segregation, commonly termed ‘petty apartheid’ in South Africa itself, involved separate amenities, from park benches to buses and post offices. Much of it fell within the powers of municipalities rather than central government, and some of it rested on the policy decisions of commercial service providers. Such segregation was less critical to the apartheid blueprint and it was the first to crumble, from the 1970s onwards.

More fundamental to apartheid planning, and the focus of this chapter, is meso-scale segregation of the four legally defined population groups in separate residential areas of towns and cities. If, as social geographers have suggested, ‘the greater the degree of difference between the spatial distributions of groups within an urban area, the greater their social distance from each other’ (Peach, 1975, p. 1), then segregation at this scale is an effective instrument for maintaining group divisions. It serves to prevent much of the voluntary association which might otherwise take place at the level of neighbourhood or social community, through sports or social clubs, churches and other organisations. Even where the desire to mix existed, as in some English-speaking churches or in liberal elements of white society, efforts to do so were hindered by transport problems, apartheid regulations and simply the self-conscious or contrived nature of events that try to overcome the ‘World of Strangers’ so powerfully portrayed by Nadine Gordimer (1962). Urban segregation was also vital to apartheid planning because it provided a territorial basis for segregated schools as well as health and other services. As apartheid developed, it also provided a territorial basis for local government, allowing some measure of autonomy for Africans, coloured people and Indians whilst allowing whites continuing control of not only their own residential areas but the central business districts and major industrial areas of cities. Finally, after the introduction of a tricameral parliament in the 1983 constitution, meso-scale segregation provided the fragmented territorial basis of the Indian and coloured chambers.

A basis for macro-scale segregation already existed in 1948, for blacks at least, in the form of reserves created by the Land Acts of 1913 and 1936. Under apartheid these were to be partially consolidated as the basis for potentially independent states successively known as ‘bantustans’, ‘homelands’ and ‘national states’. Between 1977 and 1981 four of these ten territories were granted full independence, but remained unrecognised outside South Africa. The intended statehood of all these territories was assisted by the building of new capitals in most of them and by industrial decentralisation policies that achieved limited success despite some of the highest subsidy levels anywhere in the world.

1.2 The Social Formation of Settler-Colonial and Segregation Cities

Notwithstanding the radical urban re-ordering and enforced resettlement undertaken after 1948, apartheid cities cannot be cut off from their roots, which Christopher (1983) even traces back to thirteenth-century English colonial practice. Until at least the 1950s, South African cities could be viewed in the context of African and colonial cities generally (Simon, 1984). Particular parallels have been noted with Nairobi (Fair & Davies, 1976; Western, 1984) and the techniques of French and Belgian planners in Africa (Winters, 1982). Abu-Lughod (1980) sub-titles her study of Rabat ‘Urban apartheid in Morocco’ and Western (1985) extends the comparison to one between Cape Town and the ex-colonial port cities of China.

South Africa’s apartheid cities shared certain key features common to the social formation of all colonial and settler-colonial cities. Forces of control, imposed to maintain relations of dominance, depended crucially on control of access to political power, but also included control of access to the means of production and levels of employment, and to the education, training and opportunity necessary for socio-economic mobility; control over land resources, their ownership, use and distribution, and over access to services and amenities; and control over spatial relations through segregation and urban containment. These controls were used, inter alia, to achieve formal urban development, to contain forces threatening to the economic and social interests (including health) of the dominant group, and to ensure that subordinate group housing was self-financing. Crises that arose as outcomes of incomplete control were typically met by progressive intensification of the means of control designed to maintain the social order, often coupled with minor concessions.

In what Davies (1981) recognises as the ‘settler-colonial period’ (1652–1923) early African urbanisation was largely uncontrolled, with only a gradual growth of segregationist pressures in most cities, although Indians, seen as a more direct threat to whites, attracted a more vigorous municipal response based on residential segregation, notably in Durban (Swanson, 1983) and many Transvaal towns. In the early twentieth century white fears of the spread of contagious disease (often exacerbated by the siting of ‘native locations’ close to sewage farms and refuse dumps) were a significant impetus to municipally ordained segregation (Davenport, 1971). In the mines, the compound system introduced in Kimberley in the 1880s was widely adopted (Mabin, 1986), but many Africans lived outside the compounds. Compounds were also adopted for Africans working for municipalities and in factories, warehouses and docks. Segregation in private housing areas was often achieved through the inclusion of racial exclusion clauses in suburban property deeds, but enforcement sometimes proved difficult.

Deteriorating economies in the African reserves and the growth of manufacturing employment led to rapid African urbanisation from the 1920s onwards. The number of urban Africans increased fourfold between 1921 and 1951, when it reached 2,329,000, representing 28% of all Africans (Maylam, 1990). Both white labour and commercial capital felt threatened by this African urban influx. The 1922 Stallard (Transvaal Local Government) Commission was concerned to restrict the number of urban Africans so as to minimise expenditure on their locations. The essence of ‘Stallardism’, which was to form the basis of official attitudes for several decades, is summed up in the oft-quoted dictum that ‘the native should only be allowed to enter the urban areas, which are essentially the White man’s creation, when he is willing to enter and minister to the needs of the White man, and should depart therefrom when he ceases so to minister’ (Transvaal, 1922, para. 42). This doctrine was embodied in the 1923 Natives (Urban Areas) Act of 1923 and had far-reaching implications for the provision of services, property ownership, participation in administration and the morphology of African townships (Smit & Booysen, 1977). This Act empowered, but did not compel, local authorities to set aside land for Africans in segregated locations, to house urban Africans or require their employers to do so, to implement a rudimentary system of influx control and to keep native revenue accounts so that income from fines, rents and beer halls could be spent on the welfare of locations. The Act also provided for an embryonic form of consultation. Some larger towns such as Johannesburg, Kimberley and Bloemfontein adopted the Act almost immediately. Smaller towns, fearing excessive financial responsibilities, took longer, but most locations were registered by 1937. The 1923 Act established a framework for subsequent legislation (Maylam, 1990). It embodied the central principles of segregation (and hence relocation) and influx control (the subject of further legislation in 1937 and 1945), the expectation that African urbanisation would be self-financing (leading to minimal house-building) and the co-optation of Africans, initially in an advisory capacity, to help the system operate.

During the Second World War wartime shipping shortages led to rapid industrial growth and accelerated African urbanisation, but manpower and materials were concentrated on wartime needs and housing lagged badly behind. The same disparity continued in the late 1940s as municipalities were reluctant to increase the rates to subsidise large-scale African housing and still doubtful about the permanency of the urban African population. Its dramatic growth alarmed whites and became a central issue in the 1948 election campaign, in which the National Party’s use of swart gevaar (black danger) arguments contributed to its unexpected victory. Had Smuts’ United Party retained power, both urbanisation and urban residential patterns would probably have been approached more pragmatically. Instead it was to take four decades for the realities of social and economic geography to challenge the artificiality of apartheid cities.

Davies’ (1981) model of the ‘segregation city’ (Fig. 1.1) represents South African cities before they were subjected to apartheid planning. It incorporates a central business district (CBD) which incorporates a small, peripheral Indian element to reflect Indian CBDs in Durban, Pietermaritzburg and Johannesburg. Coloured people never managed to establish such business districts and Africans were denied the opportunity to do so. Whites occupy most of the residential space, with a broadly sectoral differentiation of income groups, but with shortage of space forcing the most recent white housing developments to move out beyond a hitherto peripheral band of Indian and coloured housing. Low-income white housing and much of the African housing are close to the industrial sector, while barracks or compounds for African migrant workers are located within the industrial zone. The model includes some ‘islands’ of coloured and Indian housing within white areas and two extensive zones of mixing between whites and coloureds, reflecting central suburbs such as South End in Port Elizabeth, Woodstock and District Six in Cape Town.

Fig. 1.1
figure 1

The segregation city (Davies, 1981)

Full size image

These segregation cities were already highly segregated. In Durban, indices of segregation for 1951 include the following (where 1.00 indicates complete segregation): Indian/white 0.91; Indian/African 0.81; coloured/white 0.84; African/white 0.81 (Kuper et al., 1958). In Port Elizabeth W. J. Davies (1971, p. 148) calculates indices of 0.89 for white/others and 0.80 for coloured/African segregation, also in 1951. These figures are higher than those calculated by Duncan and Davis (1953) for American cities, despite the use of smaller enumeration areas in the latter. In South African cities African/white segregation would have been almost total if domestic servants had been excluded. This raises a critical question: if further segregation is sought, why not achieve it by small, piecemeal adjustment? The extensive re-zoning and wholesale movement implemented by apartheid planners clearly implies that something more than segregation per se was intended. The apartheid city was part of the National Party’s long-term blueprint for South Africa, and as such it was intended to provide for the future growth pattern of South African cities, providing a design that would allow for the growth of each population group in its own sector.

1.3 Re-Ordering The Cities: The Group Areas Acts

The Group Areas Acts of 1950 and 1966 produced distinctive apartheid cities which represented a major re-ordering of the segregation cities which preceded them. These Acts were a cornerstone of apartheid, and one to which P. W. Botha clung tenaciously to the end of his Presidency in 1989. By minimising inter-racial contact, they hindered transition from conflict pluralism to a more open pluralistic society (Lemon, 1980). Such a transition depends upon cross-cutting cleavages—individual allegiances and affiliations which cut across divides—which are most likely to be established within local communities. Race zoning served to keep people from knowing one another. Blacks, especially Africans, entered white homes and neighbourhoods largely as servants, while many whites never visited the African townships where most people lived. I recall an Afrikaans geographer saying ‘We who live in South Africa do not consider ourselves to live in apartheid cities’. Coming from a geographer this seems extraordinary, but it certainly confirmed the isolation in which it was possible for white people to live within their cities.

The 1950 Act imposed control of inter-racial property transactions and inter-racial changes in occupation of property, which were subject to permit. No less than ten kinds of area were defined, but the ultimate goal was the establishment of areas for the exclusive occupation of each group. In final form, group area proclamations applied to both ownership and occupation. ‘Border strips’ might be designated to act as barriers to ensure that no undesirable contiguity occurred, and ‘future border strips’ could be set aside with future needs in mind. ‘Frozen zones’ could be proclaimed if an area was considered suitable for proclamation but was not immediately required: control in such areas would include the freezing of property transactions. Control over occupation could be withdrawn at any time by special proclamation: the possibility of establishing such ‘open’ areas caused much controversy, especially when it was decided to retain a black (usually Indian) commercial district in what had been proclaimed a white group area (Lemon, 1987).

Initial implementation was the responsibility of the Group Areas Board. It needed the assistance of experienced surveyors, engineers and planners, which demanded local authority co-operation. Where this was refused, as in Cape Town, local authorities risked the imposition of an arbitrary zoning plan. When, after opportunity for objection and inquiry, the government finally approved the recommendations, their implementation was a matter for the Community Development Board, which dealt with housing, the development of group areas, the resettlement of displaced persons, slum clearance and urban renewal.

Group areas legislation radically extended control over private property, with the Group Areas Development Act of 1955 providing for expropriation, regulation of the sale price and compensation. White acceptance of such measures clearly rested on the assumption that others would be the victims—hardly surprising given that whites passed the legislation, whites implemented it and whites alone were represented on local councils. No less than 99.7% of whites already lived in what became white group areas (Christopher, 1989), with the inner city and suburbs generally proclaimed white and other groups consigned to the periphery. The vast majority of the 125,000 families moved in terms of group areas legislation were coloureds and Indians. Even greater numbers of Africans, almost certainly more than one million between 1950 and 1990, were moved in terms of amendments to pre-existing legislation, without even the semblance of consultation embodied in the Group Areas Acts. Ironically, the whole process led to the allocation of far greater financial resources to black housing than hitherto, but directed to fulfilling an ideological commitment rather than housing improvement per se, thus failing to address growing housing needs. In some cases people were physically better housed after removal, at least until their often poorly constructed homes deteriorated, but in other cases race zoning exacerbated overcrowding and squalor.

The human damage inflicted by race zoning was and remains immeasurable. It unquestionably accentuated and even initiated the racial antipathy it was (supposedly) intended to prevent (Pirie, 1984a). Irrespective of any physical improvement or deterioration of housing conditions, many people were emotionally impoverished by the destruction of their communities and remoteness from the areas where they had grown up, where they might have experienced the emotional plenty amid material shortage observed by Hart and Pirie (1984) in Sophiatown, Johannesburg. It is appropriately Cape Town, the city most affected by group area removals, which is the subject of the most penetrating study of their devastating human impact. Probing the images of the city held by its coloured people, Western (1981) demonstrates the humiliating damage which removal and destruction of homes had on self-esteem and security as people had to re-learn their place on the edge of a callous urban society. By engaging powerfully with people and place, Western gives flesh to his conceptual theme, which remains the theoretical message of apartheid cities: ‘human social relations may be both space forming and space contingent’ (p. 5).

The space formed has been modelled by Davies (1981, Fig. 1.2). An exclusively white CBD is surrounded by an extensive white residential core with space to expand in environmentally desirable sectors. Socio-economic patterns within white areas remained relatively undisturbed. Coloured and Indian group areas, and especially African townships (now including migrant worker hostels), were located peripherally within given sectors, also allowing for future expansion. Pushing the poor to the periphery invariably increased their journeys to work, especially where African townships were constructed across bantustan borders, as in Pretoria and the East Rand (Bophuthatswana and KwaNdebele), Durban (KwaZulu) and East London (Ciskei).

Fig. 1.2
figure 2

The apartheid city (Davies, 1981)

Full size image

Implementation of group areas proceeded slowly at first, and the real impact of the legislation was felt only in the 1960s and 1970s. This impact was not uniform, and the imprint of the colonial heritage survived eighty years of segregation and apartheid in some places (Christopher, 1990). In many small towns small communities of coloureds and Indians were deemed too insignificant to warrant a separate group area; sometimes such communities were resettled in a larger town (Christopher, 1991). On the eastern Witwatersrand in the 1960s attempts were made to concentrate coloureds at Boksburg and Indians at Benoni but the proclamation in the 1980s of some 57 new coloured and 15 new Indian group areas suggested that earlier policies of concentration had failed.

Overall, more than 1,300 group areas had been proclaimed by the end of 1987. Whites, who numbered 67% of the non-African urban population, were allocated 87% of the land in group areas. By 1985 10% of the total urban population still lived outside their designated areas, but only 5.7% in the Western Cape and 5.0% in Cape Town, where coloureds were removed from both African townships and once multi-racial suburbs now designated white (Christopher, 1989).

Segregation indices between whites and other groups were deceptively low in 1985, falling below 0.75 in 10% of towns and even below 0.5 in a few cases (Christopher, 1989). In some cases, mainly in Natal and the Transvaal, these figures are misleading because they exclude most of a town’s African population which lived in nearby bantustans. Otherwise low segregation indices are explained mainly by the housing of African live-in servants and also migrant workers in mine compounds in white group areas, whilst state institutions such as hospitals, prisons, military and police institutions also contributed to lower segregation indices.

The economic effects of race zoning fell disproportionately on Indians owing to the extent of their commercial involvement and reduction of this may have been a conscious purpose of the 1950 Act (Meer, 1971, p. 23). Dislocation was particularly great in the Transvaal and the Cape where the small Indian minorities were particularly dependent on commerce. By 1966 only 7.5% of Transvaal Indians remained unaffected by group areas proclamations. In Pretoria even Indian traders not directly affected lost business because of the removal of Africans and coloureds living close to the ‘Asiatic bazaar’. In Pageview, Johannesburg, where Indians had used every device including Supreme Court action to fight removal to distant Lenasia, the government eventually assisted the city authorities in building the ‘Oriental Plaza’, but the level of rents prevented many small traders from re-establishing their businesses.

Indian CBDs survived only in three cities. In Pietermaritzburg many Indian traders found themselves in the apex of a wedge of the city centre proclaimed for Indian and coloured residential use. After years of damaging uncertainty the country’s largest Indian trading area, in and around Grey Street in Durban, was proclaimed an Indian group area for trading and light industrial purposes in 1973, but its 13,000 residents were required to move; not all had done so when the decision was reversed a decade later. For Indian traders in many smaller Natal and Transvaal towns a forced move of two or three kilometres was enough to kill businesses which depended largely on white and African customers (Maasdorp & Pillay, 1977).

Indian traders recovered some ground when the 1966 Group Area Act allowed proclamation of group areas for a specific use rather than a specific group. The procedures were cumbersome but by 1983 there were 26 such areas, mostly small. Many Indians managed to continue to own properties and trade outside their group areas by nominally registering under white ownership, as in upper Church Street, Pietermaritzburg. Eventually the widespread use of this ‘nominee’ system encouraged legalisation of open trading areas in 1984. Procedures were again cumbersome, but 90 CBDs were open to trading by all races by the end of 1988, nearly half of them in the Cape, with Indians the main beneficiaries.

1.4 Group Areas Under Pressure

Legally enforced residential segregation came under increasing pressure in the 1980s. Low natural increase and the beginnings of a white exodus had produced a surplus of housing for whites, whereas other groups experienced substantial shortages (De Vos, 1986). Middle-class blacks increasingly bought homes in white areas using white nominees, and some through closed corporation deals. Thus ‘grey areas’—residentially mixed districts—were born, especially in Johannesburg. More liberal white cities such as East London, Durban and Pietermaritzburg began to seek varying degrees of de jure change, even threatening to defy the law (Pirie, 1987). Such authorities generally supported blacks using their right to apply for official consent to live in the ‘wrong’ area, and some 90% of the 2,716 applications made in 1985 were granted.

Judicial decisions also compromised race zoning, most notably the Govender ruling by the Appeal Court in 1982 which gave discretionary jurisdiction to the courts on eviction orders, ruling that such orders could only be imposed when a full inquiry regarding all the circumstances in each individual case had first been undertaken. When the Attorney General of the Transvaal announced that he had not of late been instituting any prosecutions in terms of the Group Areas Act it was clear that the legislation was no longer working. In 1986 the State President was also faced with strong demands from business firms, some of which had begun to buy homes for black executives in Johannesburg’s affluent northern suburbs, for repeal or drastic amendment of the Group Areas Act.

President Botha resisted such pressures, emphasising the critical importance of ‘own community life’ and the special need to protect low-income whites who were most vulnerable to change: they would face the greatest changes, but were among the least equipped members of society to do so, making them liable to social and political alienation. Yet the government had an interest, if it wished the growing black middle class to form a ‘bourgeois buffer’, to make it easier for middle-class and upwardly mobile blacks, especially Africans, to move out of segregated townships. In 1981 the Strydom Committee was appointed to advise the government on, inter alia, changes in group areas legislation. It recommended the replacement of the 1966 Act with a Land Affairs Act in terms of which ownership and occupation of land would be controlled by restrictions in title deeds. This was referred to the President’s Council, whose 1987 report endorsed the concept of local option: this would enable the government to maintain that it had responded to the claims of local democracy, enabling change in those areas that were challenging the law while conservative areas maintained the status quo.

The government’s legislative response was characteristically equivocal: it tabled three Bills, two of them enabling the introduction and management of ‘Free Settlement Areas’ (FSAs), but the third seeking to override the Govender ruling and strengthen the enforcement of segregation everywhere else. This perfectly illustrates Huntington’s (1981) model of reform in a multi-ethnic society: one step forward but two steps back as it attempts to reassure ‘standpatters’—its conservative electorate. The first two bills were passed, but the third was rejected by the Indian and coloured houses of Parliament and referred to the President’s Council, a body with a white, National Party majority empowered by the 1983 constitution to decide the fate of Bills supported by only one or two houses. Remarkably, the Council failed to pass the Bill, referring it back for reconsideration. In doing so it was effectively signing the death warrant of race zoning, recognising that the pressures for change were too great. By the late 1980s police investigations of complaints were leading to virtually no prosecutions (Urban Foundation, 1990) and by mid-1990 the government was reportedly issuing permits to blacks to live in the ‘wrong’ area by fax machine within 24 hours of application. In some instances officials even helped to protect blacks against hostile white neighbours and exploitation by white owners. The establishment of FSAs proceeded in 1990, but in an air of uncertainty once President de Klerk announced in April his hope of repealing the Group Areas Act. In September 1990 National and Democratic Party members of Johannesburg City Council united to ask for the whole city to be declared a FSA, while Cape Town City Council simply asked that the city be exempted from the Group Areas Act. The Act was duly repealed in June 1991, but the shadow of apartheid planning, like that of the colonial and segregation phases, will endure for decades to come, as subsequent chapters will reveal.

1.5 Urban Africans In The Apartheid City

That the majority population of South Africa’s cities should be discussed separately in what seems almost like a postscript reflects official policies that marginalised them during the apartheid years. They were regarded as a labour force whose workers (and their dependants) ‘belonged’ in the bantustans, and as a necessary evil whose numbers should be minimised. To achieve this, the central state assumed increasing power over their lives, notably in 1972 when municipalities surrendered control of urban Africans to 22 Administration Boards which became responsible for housing, influx control and the regulation of African labour (Bekker & Humphries, 1985). An Act of 1951 sought to restrict peri-urban squatting. Attempts to link urban Africans with their respective bantustans were given substance by attempts to segregate ethnic groups within the townships from 1954 (Pirie, 1984b). In practice this primarily affected the multilingual African populations in Witwatersrand townships, but even there it was applied mainly in new townships and was much less effective than group areas segregation.

Most fundamental was the Native Laws Amendment Act of 1952, which laid the basis for all state intervention to control the distribution of African labour between town and country, and between towns. Section 10 of the 1945 Natives (Urban Areas) Consolidation Act became mandatory, and was extended to include mineworkers. Africans could claim permanent urban residence only if they had resided there since birth, had lawfully resided there for fifteen years, or had worked there for the same employer for ten years. The Act also strengthened provision for the expulsion of Africans deemed surplus to labour requirements and sought to canalise labour through bureaux in rural areas. Hundreds of thousands of Africans were prosecuted annually for transgression of these ‘pass laws’. Even when these policies were at their peak they succeeded only in reducing rather than halting or reversing the growth of the urban African population.

During the 1970s it became clear that apartheid policies were consciously striving to create a dual labour force differentiating migrant ‘outsiders’ from ‘insiders’ with Section 10 rights, a stabilised population better able to meet the growing need for skilled labour in a more sophisticated economy. ‘Insiders’ would be used more effectively by allowing them much greater freedom in the job market. Such policies were recommended in the 1979 report of the Riekert Commission (South Africa, 1985a) and eventually reflected in the 1985 Laws on Cooperation and Development Amendment Act. To house the ‘insiders’ vast new townships were built, distant from white residential areas and often close to industrial areas. Accommodation was provided across the borders of bantustans where commuting was deemed practicable, as in Mabopane, Bophuthatswana and Mdantsane, Ciskei. In South Africa the rural–urban migration common throughout the developing world was thus dammed up behind artificial boundaries, adding to the disembodied nature of the apartheid city.

Meanwhile the legal definition of those who could potentially qualify for Section 10 rights was actually narrowed by the ‘independence’ of Transkei, Bophuthatswana, Venda and Ciskei, as children born after the ‘independence’ of the bantustan to which they were officially assigned could never acquire these rights (Budlender, 1985). As late as 1983 the government attempted to limit the effect of an unwelcome court judgement by amending Section 10 so that the family of someone who had worked for the same employer for ten years could only join him if he had a house of his own, rented or bought, or married quarters provided by his employer. By 1985 some 3.9 million Africans qualified in their own right and 1.7 million more as dependants. They were concentrated largely in Pretoria-Witwatersrand-Vereeniging, Port Elizabeth, Bloemfontein and Kimberley (Bekker & Humphries, 1985). But in Pretoria, the East Rand and Bloemfontein their numbers were steadily eroded as many were forced to accept offers of housing outside prescribed urban areas at Soshanguve, Ekangala and Botshabelo respectively. In Natal changes in Bantustan boundaries left few people formally qualified, but those who had been previously were treated as ‘administrative Section 10s’. In the Western Cape, few Africans in the Western Cape qualified owing to a policy of coloured labour preference.

After the onset of recession in 1982, stricter controls over the entry of ‘outsiders’ led to the breakdown of labour bureaux and mine recruitment as the labour market contracted and the labour surplus increased, with the result that vast areas and populations were virtually excluded from the urban labour market. But for sheer survival many rural households simply had to find access to urban economies. By 1990 there were an estimated 7 million informal settlers in or around the urban areas, including many living in backyard shacks in formal townships or living as lodgers: densities of fifteen or more people per four-roomed house became widespread. In the metropolitan fringes people settled on whatever land they could find, including farms, smallholdings, church and vacant land. Many faced repeated eviction and relocation, but for some at least their persistence was eventually rewarded as communities became established, as in the Winterveld of Bophuthatswana. Others were forced to settle even further from the cities, in the KwaNdebele bantustan, Botshabelo and even in bantustan locations far from the most broadly defined urban functional regions, as in the Nsikazi district of Kangwane, near Mmabatho in Bophuthatswana and in districts of Lebowa near Pietersburg.

New official thinking was reflected in the report of a President’s Council committee which viewed African urbanisation as not only inevitable but ‘an opportunity to use … its people in such a way that the end result will be an improvement in the quality of life’ (South Africa, 1985a, p. 26). It proposed a shift from discriminatory influx control to ‘orderly urbanisation’. This would involve directing urbanisation using indirect incentives and disincentives but also control measures, mainly through existing legislation. Influx control was duly abolished in 1986 with the repeal of the 1945 Act and partial or entire repeal of 33 other laws. Whilst this constituted a major reversal of key apartheid policies, the continuing allocation of the greater part of the cities to whites, coloureds and Indians left racial discrimination firmly in place, while the fact that Africans were the only group still seeking to urbanise made them the prime objects of any constraints. Land and housing were used as material constraints in distributing labour and population, in conjunction with decentralisation and deconcentration programmes developed under the regional development strategy which had been operative since 1982 (Tomlinson & Addleson, 1987).

In 1986 the Black Communities Development Amendment Act made possible full home ownership in the townships and enabled increased supply of land for African housing. The government indicated its intention to focus on supplying land and bulk infrastructure but withdrawing from direct housing supply, leaving this to the private sector. Its continuing determination to control urban growth was demonstrated by measures in 1988 to control slums and squatting, both criticised given the overall housing shortage.

African urbanisation also posed challenges for local government in apartheid cities. Prior to 1977 the only official representation of urban Africans was through Advisory Boards and Urban Bantu Councils, neither of which significantly influenced white municipalities or, after 1972, Administration Boards. Widespread unrest in 1976, the year of the Soweto riots, led to the establishment of Community Councils the following year, but these were compromised by their ambiguous relationship with the Administration Boards, especially with regard to housing allocation and influx control, and polls were low in the minority of wards contested. The Town Councils established in 1982 were regarded as independent Black Local Authorities (BLAs) but were subject to wider powers of ministerial intervention than their white counterparts. They were expected to be self-sufficient financially, despite pressure of population numbers on buildings and services which demanded large capital inputs. Most housing remained sub-economic rental property, while significant commercial and industrial development had been precluded by decades of apartheid legislation, thus revenue from these sources was wholly inadequate to provide for the upgrading of infrastructure. When BLAs resorted to rent increases rent boycotts became a key form of resistance, and a wave of violent unrest spread to much of the country until brought under control by police and army action in 1986. These financial and political pressures, and threats to councillors regarded as ‘collaborators’, caused the collapse of many BLAs.

Just as the failure of Community Councils was attributed to lack of autonomy, failure of BLAs was blamed solely on financial problems rather than their fundamentally unacceptable racially defined structures. The immediate government response was to use the National Security Management System to implement new policies and controls to stabilise the townships. During 1986 joint management centres (JMCs) were established in 34 townships identified as ‘high-risk’ security areas, with the object of regaining control through a combination of repression and upgrading of conditions (Boraine, 1989). They largely failed, facing growing resistance and a steady revival of civic and street structures, many of which proved able to negotiate directly with state and parastatal bodies.

The state responded to the financial unviability of BLAs by including them in proposed Regional Services Councils (RSCs), hitherto planned to include only whites, coloureds and Indians. RSCs were officially regarded as horizontal extensions of local authorities, charged with provision of bulk services and promotion of infrastructure ‘where the need is greatest’ (South Africa, 1985b). This meant largely African townships, although in practice areas across bantustan borders and informal areas were not included. RSCs were nominated by participating authorities from all race groups, but voting power was determined by the volume of services consumed, which still left whites with a controlling interest. For the most part, however, they operated consensually and did give redistribution the priority intended (Lemon, 1991). Nevertheless they failed to win greater legitimacy for the racially based local government structures on which they were based: the latter were undermined by community-based organisations and alternative civic structures which had emerged from popular action. In 1990 President de Klerk implicitly recognised this when he finally acknowledged the need for a non-racial system of local government.

1.6 Concluding Questions

By 1990, then, the foundations of apartheid cities were no more. Influx control laws had been repealed in 1986, the Group Areas Act in 1991 and the need for a non-racial system of local government officially acknowledged in 1990. After four decades of a system intended to provide a long-term framework to contain and separate South Africa’s major population groups at the urban scale, the speed of change was dramatic: most of the legislative changes had already taken place when Nelson Mandela became President in 1994. But this certainly did not guarantee rapid transformation of apartheid cities on the ground. Would their influence prove as enduring as that of the colonial and segregation cities before them? A generation after the political transition, to what extent are ‘homes still apart’? How far has post-apartheid urban government brought genuine redistribution and, in doing so, transformed South Africa’s cities? These are some of the questions which will be addressed in the case studies that follow.