The Royal African Society

The Rise and Decline of Urban in Author(s): Paul Maylam Source: African Affairs, Vol. 89, No. 354 (Jan., 1990), pp. 57-84 Published by: Oxford University Press on behalf of The Royal African Society Stable URL: http://www.jstor.org/stable/722496 Accessed: 26/03/2010 06:11

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=oup.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Oxford University Press and The Royal African Society are collaborating with JSTOR to digitize, preserve and extend access to African Affairs.

http://www.jstor.org THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA

PAULMAYLAM

IN DECEMBER1919 a deputation,representing the residentsof Ndabeni, an African township in , met the Minister of Native Affairs to discuss the future of the township. The residents of Ndabeni faced the prospect of relocationto a new township. A spokespersonfor the del- egation,expressing dismay at the impendingmove, told the ministerthat it appearedthat Africanswere not wantedin CapeTown:

that only their hands were needed at work, and that if some mysterious arrangementcould be devised wherebyonly their hands could be daily broughtto town for purposesof labourand their persons and faces not seen at all, that would perhapssuit their white mastersbetter. 1

This spokesperson,unnamed and forgotten, was touching upon both the essentialobjective and the fundamentalcontradiction of urbansegregation and apartheid. Much of state urban policy in South Africa has been directedtowards attaining the unattainable:the securingof labour-power withoutlabourers. Out of this fundamentalcontradiction has arisenmany furthercontradictions, conflicts and struggles. It is the aim of this articleto show how the policy and practiceof urban segregation,control and apartheidhave evolved (insofaras those policies and practiceshave affectedAfricans), and to examinehow the system has begun to show signs of collapsein more recentyears. The firstpart of the articletries to periodizethe evolution of the apartheidcity and to explain why and how it emergedin specificforms at certaintimes. In this section the focus will be on four main centres: Kimberley, Johannesburg,Cape Town and Durban. The secondpart examinesboth the primarymechan- isms andthe contradictionsof urbanapartheid. It considershow the urban Africanlabour force has been reproducedover time, and who has bornethe costs of reproduction. These issuesgave rise to considerabledebate among state policy makers,and to divisions and conflictswithin and between the centralstate, the local stateand capital. Indeed,much of stateurban policy over the yearshas had to be concentratedon managingthe contradictions, conflictsand strugglesthat have developedaround urban apartheid. This task of managementhas been achieved more successfullyat certaintimes The author lectures in history at the University of Durban, Natal. 1. B. Kinkead-Weekes, 'The development of popular resistance among local Africans 1918-1935' (unpublished paper, Workshop on the , 1985), pp. 8-9. 57 58 APRICAN APFAIRS than at other times. In the 1960s, for instance,the system held together. However, in the last decade or so the inherent strains and contradictions have shown up more starkly,and today the crisis of apartheidis in many ways most apparentin the urbanarena. Thefirstphase in theemergence of theapartheid city Urbanpolicies and practices in SouthAfrica can be saidto havedeveloped and changed over four main phases: (i) pre-1923; (ii) 1923-1950152;(iii) 1950/52-1979;and (iv) post-1979(this last phaseto be consideredat the end of the article). These phasesshould not be demarcatedtoo rigidly. There are threadsof continuitythat tie all four phasestogether. In some cases a new phasemerely marks a tighteningof thesethreads; in othercases there are more distinctbreaks with previouspatterns. In the first phase, before 1923, there was a relativelylow level of African urbanization. In 1904 the total urbanAfrican population of South Africa was officiallyestimated to be about337,000; by 1921this figurehad risento about587,000. During these firsttwo decadesof the twentiethcentury the percentageof urbanizedAfricans remained fairly constantat 12 to 13 per cent of the total Africanpopulation.2 There was also a high proportionof Africanmales to femalesin urbanareas. This demographicpattern reflected the fundamentalcharacter of the South Africanpolitical economyat the time. Although the Africanrural economyhad been experiencingstresses and strains since the latenineteenth century,it was still ableto providea relativelystable subsistence base. The manufacturingsector had not yet developedon any significantscale, and the urbaneconomy rested largely on mining and commerce. The low level of African urbanizationwas not a result of state enforcement. Indeed, the state apparatusfor controlling the African urban presence was largely undevelopedbefore 1923. A high degreeof regionalautonomy existed, as each province or municipality tended to devise its own regulations for controllingurban Africans. Although no centralized state control was exercised over African urbanizationbefore 1923, there were regionaltrends towardssegregating and controllingurban Africans. It thereforebecomes importantto con- siderthe dominantinterests and concerns that gave rise to thesetrends in our four main centres, and to examine the specific forms of segregationand control that emerged in these centres. In two cases the particularneeds of mining capital seem to have been decisive; in the other two, social considerationswere important. In the diamond-miningtown of Kimberley there emerged one of the earliestand most rigorousforms of urban labourcontrol in South African 2. See H. A. Shannon,'Urbanization, 190S1936,' South African3tournal of Economics,S ( 1937), pp. 16S90. THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 59 history. Closed compounds, accommodatingAfrican migrant workers, were introducedin Kimberley in the 1880s. GJaol-likeinstitutions, they were actuallymodelled on the De Beers convict station. The closed com- poundswere establishedat a time when diamondprices were fallingand the mineownerswere strugglingto reduceproduction costs. The compounds enabled them to exercise tighter discipline over their unskilled labour force, preventingdesertion and diamondtheft, and to ensurea morecertain supplyof labour. As Turrellhas argued,the compoundshelped to resolve a fundamentaldilemma for the mineowners:'On the one hand,they wanted experiencedlabourers in theirmines. On the otherhand, they did not want an organizedworking class in theirtown. It was this contradictionthat was bridgedby the closed compoundsystem.'3 This is not to say that all Africansin Kimberleylived in compounds. By 1892 about half of the town's African populationwere accommodatedin compounds, the other half living in the town or in locations.4 Thus Kimberleydid not representan earlymodel of rigidurban segregation. Its significancelies in the developmentof the compound as a model mech- anismof labourcontrol. Mabin has noted how the Kimberleypattern was followedelsewhere: 'The strictseparation of Africanand white workers both in the hierarchyof labourand in theirresidences became the modelfor mines and mining towns throughoutsouthern Africa.5 Johannesburgwas one of these towns to follow the Kimberley model. Johannesburg'smining compoundsdid not follow the exact model of the Kimberley 'closed' compound. Before the South African War Johannesburg'scompounds developed along rudimentarylines, but in the conditions of severe labour shortageafter the war they came to be more tightly controlled to curb absenteeismand desertion. The white com- poundmanager played a key role in tryingto enforcestrict discipline among the migrantworkers accommodated in the compound. He wasassisted by a group of African compound 'police'.6 Compounds also housed other Africans,apart from mineworkers,in Johannesburg. By the early 1920s about 5,000 municipalemployees were housed in compounds,and another 6,000 Africanworkers lived in privatecompounds attached to factoriesand warehouses.7

3. Rob Turrell, 'Kimberley's Model Compounds', 3tournalof AfricanHistory, 25 (1984), pp. 734; see also Alan Mabin, 'Labour, capital, class struggle and the origins of residential segregation in Kimberley, 188s1920',3rournal of HistoricalGeography, 12 (1986), pp. 7-13; and William Worger, 'Workers as criminals: the Rule of Law in early Kimberley, 187s1885', in Frederick Cooper (ed.), Strugglefor the City: migrantlabor, capital and the state in urban Africa(California UP, 1983), pp. 8s2. 4. Mabin, 'Kimberley, 188s1920', p. 18. 5. Mabin, 'Kimberley, 1880-1920',p. 22. 6. Sean Moroney, 'The development of the compound as a mechanism of worker control 190s1912', SouthAfrican Labour Bulletin, 4 (1978), pp. 3s1, 3s5. 7. Noreen Kagan, 'African Settlements in the Johannesburg Area, 1903-1923' (unpublished MA thesis, University of the Witwatersrand, 1978), p. 177. 60 AFRICAN AFFAIRS There were still, though, about another 60,000 Africans living in Johannesburgoutside the compoundsat this time. A small proportionof these, about 5,500, were housed in the municipaltownships, Kliptown and Western Native Township.8 Another 6,000 resided in Sophiatown, Martindale,Newclare or Alexandra,private freehold or leaseholdtownships that had been developed on privately owned land in the early twentieth century.9 Thousands more lived as tenants on white-owned properties, occupying shanties, out-rooms or disused warehouses or workshops. These were the slumyards,exploited by rackrentersand slum landlords.l? A fundamentalfeature of early Johannesburgwas that outside the com- pounds there was no strict patternof racialsegregation. Sophiatownand Martindale,for instance,in 1921provided homes for 1,457Africans and for 557 whites.ll The Vrededorp-MalayLocation and Doornfontein were also multi-racialcommunities. Early Johannesburgwas thus character- ized by tight controlexercised over mineworkersand some othersections of the labourforce, but also by the absenceof such control over a large pro- portion of its Africanpopulation. Johannesburgfollowed the pattern of earlyKimberley, but it did not presagethe laterera of tight segregationand control. Spatial patterns and control mechanisms in early Kimberley and Johannesburgwere largelyshaped by the originalmaterial base of the two towns, the mining economy. Cape Town and Durban, both older towns, developedfrom differentfoundations. In this earlyperiod merchant capi- tal predominatedin eachcentre, and theirlabour needs wererather different fromthat of the two miningtowns. In the late nineteenth century Cape Town still remainedprimarily a commercialcentre, an administrativecapital and a militaryheadquarters. Industrialproduction was minimal;factories were few in numberand small in size. Cape Town's labour market had a highly seasonal and casual character. The town's economycentred very much aroundthe docks;and the labourrequirements of the docks tended to fluctuatesignificantly from day to day. Similarly,the labourneeds of the fishingand building indus- tries variedfrom season to season.l2 At the turn of the centurythere was a growing demand for unskilled labour in Cape Town. Thousands of Africans came to the town, especially after 1898, so that by 1901 about

8. Kagan, African Settlements', pp .91-110. 9. Kagan, 'African Settlements', pp.25-43. 10. Kagan, 'African Settlements',p.43; Eddie Koch, ' "Withoutvisiblemeansofsubsistence": slumyard culture in Johannesburg,1918-1940', in Belinda Bozzoli (ed.), Town and Countryside in the Transtaal (Johannesburg, 1983), pp. 153-54. 11. Andre Proctor, 'Class struggle, segregation and the city: a history of Sophiatown 1905-40', in Belinda Bozzoli (ed.), Labour Tozvnshipsand Protest (Johannesburg, 1979), p.58. 12. Vivian Bickford-Smith, 'Cape Town on the eve of the (c.1985)', (unpublished paper, Cape Town, 1985), pp . 1- 13. APARTHEID IN SOUTH AFRICA THERISE AND DECLINE OF URBAN 61 bour- 10,000Africans lived there.l3 Although the predominantlywhite of racial geoisietended to live an exclusivesuburban existence, rigid patterns Town. residentialsegregation had not been historicallyentrenched in Cape CentralCape Town, for instance, had long been racially integrated. of DistrictSix, laterto become a 'coloured'area, housed a largeproportion whitesat the turn of the century. with The beginningsof residentialsegregation in Cape Town coincided Location thearrival of largenumbers of Africans. In 1890the Dock Native dock- wasopened, providingcompound-type accommodation for African the workers.14It was designed, in Bickford-Smith'sworks, 'to protect from the workerfrom other social evils of the town, to protect the town the first workers'.l5A stronger imperativetowards segregationcame in of capital yearsof the twentiethcentury, and it arose,not from the dictates butfrom the socialconcerns of middle-classresidents. was In 1901 Cape Town was hit by the bubonic plague. The outbreak the town. immediatelyassociated with the growing African presence in Mowbray, In l 899 delegations of white residents from Woodstock, of an ,Claremont and Maitland called for the establishment a location Africanlocation. The Cape Town Council soon decided that and sani- shouldbe built so that Africanscould be housedunder controlled Ndabeni tary conditions. The outcome was the rapid construction of Ndabeni, location,which was soon to accommodateabout 7,000 Africans. not the unlikethe Kimberleyand Witwatersrandmining compounds,was Ndabeni as creationof capital. Indeed, some employerswere critical of of town. they found it inconvenient to have their workers living out the Ndabeni was more the product of social pressure, exerted through see, urban rhetoricof disease and sanitation.l6 As we shall increasingly not always segregationwas often a source of contradiction,and it was functionalto capital. before There were some similaritiesbetween Cape Town and Durban there was 1923. Merchantcapital predominated in both towns, although WorldWar some industrialgrowth in Durbantowards the end of and after in each One. The economiesof both towns centredaround the docks, and difference case the labourforce was largelycasual and seasonal. The great a large was Durban's relative proximity to African reserves, making for proportionof male migrantsamong the Africanlabour force. growth of Cape Town, 188S 13. Vivian Bickford-Smith, 'The economic and demographic 1910, (unpublished paper, Cape Town, 1985), pp. 1S17. John Western, OutcastCape Town (London, 1981), p. 45. 14. at the beginning of the twentieth 15. Vivian Bickford-Smith, 'Black labour at the docks century', Studiesin theHistory of CapeTown, 2 (1980), p. 95. 16. Christopher Saunders, ' "ThecreationofNdabeni: urbansegregationandAfricanresist- Studiesin the History of Cape Town, 1 (1979), pp. 172-76; M. W. ance in Cape Town", policy in the Cape 'The sanitation syndrome: bubonic plague and urban native Swanson, 392-93, 409. Colony, 190s1909') 3tournalof AfricanHistory, 18 ( 1977), pp. 62 AFRICAN AFFAIRS If a segregationistimpulse was to be foundin early.Durban,it was directed more against Indians than Africans. As Swansonhas shown, in the late nineteenthcentury Indianscompeted with whites for 'space,place, trade'. The local state'sresponse was patternedon a threefoldapproach: 'residen- tial segregation,political exclusion, and commercialsuppression'.l7 The first failed to materializeas a scheme for an Indian location fell through. The third objective was more successfully attained, through the 1897 Licensing Act which gave municipalities arbitrarypowers to issue or withholdtrading licenses. The local state in Durban was more concernedabout controllingthan segregatingits Africanpopulation. By its very naturethe town's casual, or 'togt', labour force possessed a great deal of freedom and mobility. Regulationsrequiring 'togt' workers to register were first introduced in 1874. These weretightened in 1903,doubling the monthlyregistration fee to five shillingsand compelling'togt' workers to live in municipalor private compounds.18 By 1921 about 13,500 African workers(out of Durban's total African population of about 37,500) lived in such accommodation. Another 9,000 domestic workers, predominantlymale, lived on their employers'premises, while about 15,000Africans were without any formal accommodation.l9 In 1904 the Natal Parliamenthad passed a Native Locations Act, enabling municipalitiesto establish segregatedlocations. But Durbandid not followthe exampleof CapeTown in constructingsuch a location,except to build, ratherbelatedly in 1915-16, Baumannville,a small locationcomprising a mere 120 'cottages'for familyoccupation. More significantin Durban'scase was the developmentof a localbureau- cratic structure of control. The key to this structure was the revenue derived from the municipal monopoly of the manufactureand sale of sorghumbeer for Africanconsumption. Such were the profitsfrom this monopolythat the local state was able to financethe creationin 1916 of its own Native AdministrationDepartment, which would exercise increasing controlover Durban'sAfrican population. Of the four towns that we have examined,Kimberley may have had the tightest form of labourcontrol, in the shape of the closed compounds,but Durban had the most developed form of administrativecontrol over its Africanpopulation. Indeed, many of the practicesthat were evolving in Durban were to be borrowedby other municipalitiesor incorporatedin subsequent parliamentarylegislation. If one is seeking to discover the

17. Maynard W. Swanson, '"The Asiatic Menance": creating segregation in Durban, 1870-1900' International3tournalof African Historical Studies, 16 (1983), pp. 401-17. 18. D. Hemson, 'Class consciousness and migrant workers: dock workers of Durban' (unpublished PhD thesis, University of Warwick, 1979), pp. 2s25, 95-96. 19. Paul La Hausse, 'The struggle for the city: alcohol, the Ematsheni and popular culture in Durban, 1902-1936' (unpublished MA thesis, , 1984), pp. 119, 320. THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 63 origins of urbanapartheid practice, Durban probably provides the likeliest source. The particularityof the practicesand structuresthat evolved in Durban reflected the relatively high degree of municipal autonomy that existed duringthis earlyphase. Interventionby the colonialstate, or later by the central or provincial state, was minimal. Local administrativesystems and housing policies therefore tended to be diverse, and were to a large extent shaped by the nature of the local economy. In Kimberley and Johannesburghousing policy, with its emphasison the compoundsystem, was partlydetermined by mining capital. In the ports of Cape Town and Durbanthe pools of seasonaland casuallabour were less easilycontrollable, althoughin both townsattempts at controlwere made through the establish- ment of compound-typeaccommodation. But in all fourtowns, apartfrom the compoundshousing workers at the mines, docks,factories and commer- cial businesses, there was no developed policy or system for housing the urbanblack population. The segregatedtownships, which wereto become such a prominent feature of the South African landscape later in the twentieth century, were few in number during this early phase. One peculiarfeature of the few municipaltownships that had been built wastheir siting. As the 1914 TuberculosisCommission noted, sites were often ill- chosen, 'generally. . . not far from the town sanitarytip, the refuse dump, and slaughterpoles . . .X20Kliptown, Western Native Township andEastern Native Township in Johannesburgwere just three examplesof townships locatednext to sewagefarms.2l Generally,though, these few townshipsaccommodated only a smallpro- portionof Africanswho lived outside the compounds. Many lived in pri- vate leasehold or freehold townships;others rented backyardquarters in centralareas of the town. The degreeof controlexercised over their lives by the local statewas smallcompared with the tight restrictionsthat were to come later. The relative absence of control (outside the compounds)in this earlyphase partlyreflected the limited size of the urbanAfrican popu- lation at the time. However, from the 1920s the picture was to change significantly. Thesecond phase Our second phase covers three key decades:the 1920s, 1930s and 1940s. It is markedby certaincrucial developments and trends. First, therewas a significantshift in the natureof the SouthAfrican economy, with the growth of a manufacturingsector. Second, the deteriorationin the economies of the African reserves gained pace. Third, and closely connected with 20. T. R. H. Davenport, 'The beginnings of urban segregation in South Africa: the Natives (Urban Areas) Act of 1923 and its background' (Grahamstown, Institute of Social and Economic Research Occasional Paper, 1971), p. 7. 21. Kagan, 'African Settlement', pp. 178-79. 64 AFRICAN AFFAIRS these trends, there was a dramaticgrowth in the size of the urban African population. And fourth, the central state increasinglyintervened in the sphere of urban policy and practice. All of these developmentswere to have significantimplications for urbanAfricans. A varietyof factorsencouraged the growthof manufacturingindustry in these three decades: the stimulus of two world wars, state protectionist policies, the diversificationof mining capital, and an influx of foreign capital. Between 1921 and 1959-60 there was a doubling of secondary industry'srelative contribution to the grossnational product. Employment in the manufacturingsector as a whole grewat an averagerate of 5 3 per cent per arlnumbetween 1925 and 1939, and at a rateof 6 0 per cent for Africans. Accompanyingthis growthwas a shiftin the labourprocess . Mechanization increasedthe number of semi-skilledoperative jobs, many of which were filled by African workers. This trend not only altered the old division betweenskilled white and unskilledAfrican workers, but also affectedlocal andcentral state policies for administeringand controlling urban Africans.22 Contemporaryobservers gained a strongimpression that the economiesof the African reserves were deterioratingdrastically. The report of the 193(}32 Native EconomicCommission, for instance,drew attentionto the 'under-developmentof the Reserves'.23 A ratherless drastic picture has been painted by Simkins, who argues that between 1918 and 1954 the reserves went through a period of 'fragile productivity maintenance'. However,he goes on to say,this wasnot a caseofthe reserveeconomies being buoyant,but rathera caseof an unstableequilibrium being made possible by massiveoutmigration which limitedovercrowding in the reserves.24 There is evidenceof considerableoutmigration from the reservesto urban areas. Between 1936 and 1946 the total population of the reserves increasedannually by only 0 9 per cent, the small increasesuggesting sub- stantialoutmigration.25Between 1921 and 1951thetotalsizeoftheAfrican urban population rose from 587,200 to 2,329,000; and during the same period the percentageof urbanizedAfricans out of the total Africanpopu- lationalmost doubled from 14 per cent to 27 9 per cent.26 At the sametime the characterof the urban African population was changing. The pro- portion of urban Africans living under family circumstancesin 1936 is estimatedto have been about30 per cent, risingto 38 per cent by 1946.27 22. D. C. Hindson, 'The pass system and the formation of an urban African proletariat in South Africa: a critique of the cheap labour-power thesis' (unpublished PhD thesis, University of Sussex, 1983), pp.67-68, 125-28. 23. Report of the Native EconomicCommission, 193s1932, UG 22, (1932), part III, p.80. 24. C. Simkins, 'Agricultural production in the African Reserves of South Africa,1918-1969', 3rournalof Southern African Studies, 7 (1981), p.270. 25. Hindson, 'Pass system', pp. 129-30. 26. P. Maylam, 'Strategies of control and evasion: African urban history in South Africa c.190S1950' (unpublished paper, Durban, 1981), p.2. 27. Charles Simkins, Four Essays on the Past) Present and Possible Future of the Distribution of the Black Population of South Africa (Cape Town, 1983), p.22. THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 65 The rapid growth of African urbanizationin these three decades drew greater state intervention in the whole urbanizationprocess. However, throughoutthis secondphase there was considerableuncertainty and debate amongpolicy-makers, state ideologists and other opinion-formers as to what formthis interventionshotlld take. There weretwo poles of thinking. At one extreme was the view put forwardby the Stallard(Transvaal Local Government)Commission of 1922 that the right of Africansto be in urban areasrested solely on their willingness'to enterand to ministerto the needs of the white man'.28 At the otherpole was the liberalview that the stabiliz- ationof the urbanAfrican population should be encouragedin preferenceto the continuationof the migrantlabour system. The dominantview of the time amongstate ideologists and policy-makers was one that fell between these two poles. This compromise,reflected in the resorts of various governmentcommissions that sat in the 1930s and 194()s,recognized the impracticability,and to some extent the immorality, of Stallardism,but also stopped short of recommendingwholesale stabiliz- ation. This position was adopted in the 1932 Native Economic CommissionReport. The Commissionargued that it was 'undesirableto encouragethe urbanizationof the Native population')but went on to say:

It is perfectlyclear that a considerablenumber of Natives have become permanenttown dwellers. No good purpose is served by disregarding this fact, or by acting on the assumptionthat it is not a fact. In the interest of the efficiencyof urban industriesit is better to have a fixed urban Native population to the extent to which such population is necessarythan the presentcasual drifting population.29

This view was endorsedby the 1935 Young-BarrettCommittee, and by the 1948 Fagan (Native Laws) Commission. The latter concluded that 'the policy should be one of facilitatingand encouragingthe stabilizationof labour;but on the otherhand, migratory labour cannot be prohibitedby law or terminatedby administrativeaction'.30 Thus the dominantofficial view sraspragmatic: the realityof stabilization had to be recognized, but migrarlcywould necessarilyhave to continue. This view reflectednot only the realitiesof the urbanizationprocess, but also the particularlabour needs of differentcapitalist sectors. Secondary industry demandeda more skilled, and thereforemore stabilized, labour force, while the mining sectorcontinued its dependenceon migrantlabour. From 1923 the central state began to intervenemore in the business of regulatingand controllingthe urbanAfrican population. But it did so in 28. Rodney Davenport, 'African Townsmen? South African Natives (Urban Areas) Legislationthrough the Years',African Affairs, 68 (1969),p. 95. 29. Report of the Vative Economic Commission,193S1932, UG 22, (1932),part III, p. 72. 30. Report of the Nalive Laws Commission,1946-1948, UG 28, '1948), p. 50. 66 AFRICAN AFFAIRS the contextof ideologicaluncertainty and debate(just described),a change in the supply/demandpattern in the labourmarket, and a continuingvariety of local practices from one urban area to another. The state's growing interventionin the urban sphere in the three decades after 1923 laid the foundationfor the eraof 'highapartheid' and rigid labour control that was to follow from the 1950s. A legislative frameworkwas established;and, althoughlocal state autonomywas not overridden,the trend was towards that greatercentral state controlwhich was to become more entrenchedin the 1960sand 1970s. The 1923 Natives (Urban Areas) Act representedthe first major inter- vention by the centralstate in the business of managingthe urbanAfrican labourforce and ensuringits reproduction. The Act empoweredmunici- palitiesto establishsegregated locations for Africans,to implementa rudi- mentary system of influx control, and to set up advisory boards, bodies which would containAfrican elected representativesand which would dis- cuss local issues affecting Africans, but without any power to change policy. The Act also requiredmunicipalities to institute native revenue accounts,into which would be paid all incomederived from beer-hallsales, rents, fines and fees levied fromAfricans.31 The short-termsignificance of the 1923 Act was limited. Most of its provisionswere not obligatory,and few municipalitieswished to implement the Act. Segregation was gradually introduced as certain sections of municipalareas were proclaimedas 'white', compellingall non-exempted Africansand those not living on their employers'premises to move into a municipal location or hostel. Johannesburgand Kimberley began pro- claimingsegregated areas in 1924, Cape Town in 1926, and Durban in the early 1930s.32 Whole-salesegregation was impossiblebecause the procla- mation of segregatedareas could only be enforcedif alternativeaccommo- dation was availablefor those Africansbeing forced out of their existing quarters. The availabilityof such accommodationwas severelylimited as municipalitieswere unwilling to make substantialinvestments in the con- structionof townships. Other provisions of the Act also had a minimal short-termimpact. Many municipalitiesdid not establishadvisory boards while others did so belatedly. And few local authoritiesused the power grantedto them to instituteinflux controls;by 1937 only eleven towns had systematicallyimplemented such controls.33 The significanceof the 1923Act lies morein its broader,long-term impli- cations. It representedthe first majorintervention by the centralstate in 31. Davenport, 'African Townsmen?' p. 99. 32. Kagan, 'African Settlements', p.182; Mabin, 'Kimberley,1880-1920', p.21; Christopher Saunders, 'From Ndabeni to Langa', St?>diesin theHistory of CapeTown, 1 (1979), p.205; Paul Maylam, 'Shackled by the contradictions: the municipal response to African urbanization in Durban,1920-1950', AfricanUrban Studies, 14 (1982), p. l l. 33. Michael Savage, 'The imposition of on the African population in South Africa 1916-1984', AfricanAffairs, 85 (1986), p.194. THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 67 the process of African urbanization. It also provided a frameworkand foundation upon which subsequent legislation and policy were to be built. The key elements of later, more refined, urban apartheidpractice were to be found in the 1923 Act in embryonicform. These includedthe principleof segregationand the ensuingpractice of relocationthat aroseout of that principle;influx controlmechanisms; a self-financingsystem which shifted as far as possible the burden of reproductioncosts on to urban Africansthemselves; and an institutionfor cooptingpotential collaborators in the shape of the advisoryboard. During the following decades all of these mechanismswere to be tightened and refined, and the process of centralstate encroachmenton municipalautonomy was to gain ground. The tightening of control and the process of centralisationwere both takenfurther in the 1937 Native Laws AmendmentAct. The majorcon- cernof this measurewas influxcontrol. It providedfor the removalto rural areasof Africanssuplus to labourrequirements in anyparticular urban area; it made it more difficultfor Africanwomen to enter an urbanarea and for work-seekersto remain in an urban area. The Act increasedthe influx control and expulsion powers of local authorities,but it also increasedthe powers of the Minister of Native Affairs, who could now compel a local authorityto implementany sectionof the Natives (UrbanAreas) Act or have the section implementedby his own department.34 The 1944 Housing AmendmentAct set up the National Housing and PlanningCommission. This body was to havepowers to intervenein local housing policy, further weakening municipal autonomy.35 And the followingyear an act was passedto consolidatelegislation governing urban Africans. Influx control mechanismswere strengthened,and the central state'spowers of interventionwere furtherenhanced. So, by the time that the NationalParty came to power in 1948, a whole apparatusfor regulating and controllingthe movementand daily lives of urbanAfricans had already been constructed. However,while the machineryhad been created,it was not yet well-oiled or functioningas efficientlyas it was designedto. It was in our next phase that this optimal cfficiencywas to come closer to being achieved.

Thethird phase During the thirdphase, from the early1950s until the late 1970s,there was an intensificationof the patternsthat had evolved in the second phase. In particularmore and more powerto controlthe lives of urbanAfricans came

34. Hindson, 'Pass system', pp. 95-98; Davenport, 'African townsmen?', pp. 102-103. 35. Robin Bloch and Peter Wilkinson, 'Urban control and popular struggle: a survey of state urban policy 192s1970', Africa Perspective, 20 (1982), p. 21. 68 AFRICAN AFFAIRS to be vested in the centralstate and its agencies,further weakening munici- pal autonomy in th1ssphere. At the same time the actual controls selves that them- were imposed on urban Africans came to be tightened further. even The 1948 election victory of the NationalParty (lid not representsuch a major turning-point in South African history as it is often made out. Certainlythere were continuities in policy and practice predating 1948. However, it is also true that under the new Nationalistgovernment there were shifts of emphasisand changes of approachin the whole sphere of 'nativepolicy'. One importantshift, at the level of policy andideology, was awayfrom the Fagan Commission'sview that the growingpermanence of theurban African populationhad to be officiallyrecognized, and backto the old Stallardist notion that Africanscould only be temporarysojourners in urbanareas outside the . This was the view put forwardby the 1947report of the SauerCommittee, which set out NationalParty policy on the'colour question'.36 Accompanying this shiftof emphasiswas a greatercentralization of power. Fromthe 1950s the central state increasinglytook upon itself the task of regulatingthe presence of Africans in urban areas in accordancewith governmentpolicy. Local authoritiesfound their autonomyundermined asthe Department of Native Affairs intervenedmore and more. Many localauthorities were compliant,others were less so, leading to conflicts betweenthe central stateand local state. Such a disputearose in the 1950s betweenthe JohannesburgCity Council and the government over the latter'splan to uprootand relocatelong-established African communities in thewestern areasof the city (Sophiatown,Martindale and Newclare). United The Party-dominatedCity Councilopposed both the plannedremovals andthe manner of their implementation,objecting to the way in which Verwoerd,the Minister of Native Affairs, bulldozed aside municipal oppositionin carryingthrough the operation.37 Centralizationof statepower was further advanced with the establishment ofAdministration Boardsin 1972. The countrywas divided into twenty- tworegions, in each of which an AdministrationBoard was set up. The Boardswere to take on most of the functions previously exercised by municipalitiesin the administrationof urban Africans. The Boards becameresponsible for housing, influx control, and the regulation of Africanlabollr. They were to derive revenue from rents, levies on employers,and profits from liquor sales (the Boards took over municipal

36.Peter Wilkinson, 'Providing "adequate shelter": the South African state "resolution"of the Africanurban housing crisis, and the Patersin Southern 1948-1954',in D. C. Hindson(ed.), Working AfricanStudies, vol. 3 (Johannesburg,1983), pp. 69-70. 37.Gordon H. Pirie and DeborahM. Hart, western 'The transformationof Johannesburg'sblack areas',yournal of UrbanHistory, 11 (1985),pp. 397400. THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 69 liquor monopolies);and they were expected to financeall their operations out of this revenue.38 At the sametime as this process of centralizationwas occurring,controls over the presenceof Africansin urbanareas were steadilybeing tightenedin accordancewith the Sauer recommendations. Stricterinflux controlwas to be the chief mechanismfor achievingthis. A key measurewas the 1952 Native Laws Amendment Act. This restricted the right of permanent urbanresidence to Africanswho had eitherbeen born in the particularurban areaand had lived there continuouslysince birth, or had resided legally in the area for fifteen rears, or had worked for the same employer for ten years. The Act also made strongerprovision for the expulsionfrom urban areasof Africanswho were surplusto labourrequirements. Even tougher restrictionsupon entry irltourban areas and more severeexpulsion powers were introducedin 1964.39 Implementationof these measureswas uncompromising. For instance, prosecutionsunder the pass laws increasedfrom a total of about 280,200 in 1951 to about 631,300 in 1970. And the overall demographicimpact of tighter influx control was significant. From the early 1950s, although the size of the urban African population continued to grow, the rate of growthdeclined. At the sametime the proportionof Africansliving in the bantustans was increasing.40 Both of these demographic trends were largely the product of tighter influx control and the state's removalsand resettlementprogramme. Hindsonhas arguedthat influxcontrol after 1952 was not simplydesigned to enforcetemporary migration. It also aimed to stabilisea section of the urban Africanproletariat, thereby reinforcingthe differentiationbetween migrantsand those with more permanent(Section 10) urbanrights. This stabilisation went hand-in-hand with another dimension of the state's attemptto tighten controlover urbanAfricans, its housingpolicy. The short-termaims of state housing policy in urbanareas in the 1950s have been outlined by Morris: 'to remove black freehold rights in these areas, segregate the races, control movement and reduce the economic burdenof blackson the state and on local authorities'.4l The principleof residentialsegregation was entrenchedand rigidified in the 1950 Group Areas Act, which made compulsorywhat the 1923 Urban Areas Act had recommended. The principlewas put into practicewith the construction of vast Africantownships. Accordingto the apartheid'ideal' these town- ships were to be sited as far as possible from white residentialareas, but

38. Simon Bekker and Richard Humphries, FromControl to Confusion:the changingrole of AdministrationBoards in SouthAfrica 1971-1983 (Pietermaritzburg, 1985), pp. 7-20. 39. Savage, 'Pass Laws', pp. 19S95; Doug Hindson, Pass Controlsand the UrbanAfrican Proletariat(Johannesburg, 1987), pp. 61, 68-69. 40. Savage, 'Pass Laws', pp. 190-91. 41. Pauline Morris, A Historyof BlackHousing in SouthAfrica (Johannesburg, 1981), p. 69. 70 AFRICAN AFFAIRS reasonablyclose to industrialareas. Spatialseparation was to be reinforced by bufferzones and by naturalor otherbarriers. And townshipswere to be designedand sited in such a way that they could be cordonedoffin the event of riot or rebellion,and the resistancesuppressed in open streets.42 In the 1950s and 1960s constructionof such townships in many urban areasproceeded on a considerablescale. Fromthe late 1960s the provision of housing in urban areasoutside the bantustanswas slowed down as the state looked to ways of confiningas many Africansas possible within the bantustans,without upsetting the labour supply. Two main strategies were devised. In cases where industrialcentres were close to bantustans, townships would be relocated in the bantustansand the workers would commutedaily to theirworkplaces. The otherstrategy was to try to induce industries themselves to relocate to border areas close to bantustans.43 These strategieswere to representa last attemptto implementthe ultimate aim of Stallardand laterapartheid planners, namely to allow Africansinto urban areas outside the bantustansonly for the purpose of selling their labour. Overwhelmedby contradictionsand crisesthe statehas since been forced to abandonthis objective. The contradictionsinherent in the whole urban system of control graduallybecame more unmanageablefrom the early 1970s. And other developingcrises furtherundermined the apparatusof urban apartheid. The continuing collapse of the economies seriously weakened the base of the migrant labour system. Apartheid barriers like job reservationand influx control aggravatedthe growing shortageof skilledand semi-skilledlabour in urbanareas. And, most sig- nificant of all, black workersand urban black communitiesin the 1970s mobilizedand organizedthemselves with greatervigour and militancythan ever before. The 1973strikes and the 1976uprising were probably the key events leading to the eventual near collapse of the main pillars of urban apartheidin the 1980s. That collapsewill be examinedlater. Beforethat it is necessaryto analysemore closely those longstandingstrains and contra- dictionsthat preventedthe system,as idealizedby Stallardand Sauer,from ever becomingfully workable.

Contradictionsand crisesin urbanapartheid The centralcontradiction in the whole systemof segregatingand controll- ing urbanAfricans arose out of the unattainableobjective of tryingto secure the labour-powerof Africanswhile minimizingtheir presence as people. This objectivewas derivedfrom the interestsoftwo constituencies:capital's

42. Western,Outcast Cape Town, pp. 7s5. 43. Hindson,Pass Controls,p. 71. THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 71 need for labour,and the perceivedneed of most whitesfor protectionagainst the supposed dangersto their health and security arising from the urban Africanpresence. The effortto realizethis objectiveentailed certain basic policies. One was to channeland allocatelabour according to demandand thus limit the urbanAfrican presence according to labourneeds. Another was to ensure the reproductionof labour under controlled conditions at minimalcost to both capitaland the centraland local state. Implementationof these policieswas to be problematic. Both the quan- tity and quality of labour demandedby differentcapitalist sectors could changeover time, both in the shortterm and the long term, complicatingthe task of channellinglabour. Moreover the measuresand mechanismsof channellinglabour were often ignored or defied by the people over whom controlwas being sought. Nor couldthese controlmeasures do anythingto stall the process of African rural impoverishmentwhich was a significant factorin determiningthe movementof people to urbanareas. As the size of the permanentAfrican urban proletariatgrew, so did the cost of reproducingthat proletariatbecome more of an issue. The pro- vision of accommodationand transportfacilities for urbanAfricans became a financialburden. The question of who should be primarilyresponsible for bearingthis burden gave rise to strugglesand conflictsbetween capital and labour,between the stateand capital,between the localstate and central state,and betweenAfricans and the state. These divisionsinvariably arose out of effortsto shift the burdenof reproductioncosts fromone to the other. There is a great danger in slipping into a teleologicaland functionalist interpretationof the evolution of urban apartheid, tracing the gradual development of a monolithic, smoothly functioning system of control. (Indeed, the first half of this article,taken by itself, might seem to fall into that trap.) Although urban apartheidmight have taken on a monolithi appearancein the 1960s, it was always riddled with contradictions. As Cooper has noted, '[t]he ambivalenceand inconsistencythat run through the history of urban policy reflect the complexity of the issue of social reproduction'.44 It remains for us to examine these contradictions, inconsistenciesand complexities. They are to be found, for instance, in one of the key elements of urban apartheid,the pass system. As Hindson has shown, the main function of passcontrols has changedover time. Fromthe latenineteenth century they were used to maintainthe migrantlabour system;after the Second World Warthe controlswere gearedto the growthof an urbanAfrican proletariat; laterthe controlswere aimedat limitingthe size of the 'surplus'propulation in urbanareas.45 But at no time-didthe pass systemhave a single function;

44. Cooper, Strugglefor theCity, p. 8. 45. Hindson, Pass Controls,pp. 1>1 1. 72 AFRICAN AFFAIRS its purposeswere alwaysvaried. It has been designedprimarily to control the movement of Africansinto urban areas, and to channel workersinto those sectors where their labourhas been most needed. The system has also been used for curbingcrime, identifying political activists, and assisting tax collection.46 The pass system representeda majorinterverltion in the labourmarket. It has been arguedthat the chief purposeof influxcontrol was to perpetuate migrantlabour. But Hindson shows that this is a misleadingview: influx controlwas aimedat producinga systemof 'differentiatedlabour-power', in that it simultaneouslyenforced temporarymigration and promoted the stabilizationof a permanenturban African proletariat.47 Influx control was a key mechanismfor trying to cope with that funda- mentalcontradiction stated at the outset:the need to securea suitablesupply of labourwhile minimizing the presenceof Africansin urbanareas. It wasa contradictoryobjective, calling for both the inclusionand exclusion of urban Africans.48 And the functionality of the system was always limited. Although the apparatusof pass control was adaptedfrom time to time in response to shifting demographicpatterns or changinglabour demand, it was often unableto cope with short-termor seasonalfluctuations in labour demand. There wasr though, a much more fundamentalcontradiction. The whole systemof influxcontrol came to be inherentlyunstable, particularly in its laterphase. Fromthe 1950s a viciouscycle developed. The increasing displacementof the 'surplus'urbarl population to the reserves/bantustans aggravatedovercrowding and rural impoverishmentin those rural areas. This in turn put pressureon people to defy influx control by moving to urbanareas, where there was a better chanceof earninga subsistence. As Greenberg has put it, 'control has made necessary more controls; the successfuldamming up of labour[sic] in the Africanrural areas has created inducementsto burstthe dams'.49 Althoughinflux control has limitedthe growthof the urbanAfrican population, especially during the 1960s, it has never been able to achievean optimallimitatioru on growth. Africanscon- stantly violated influx control by moving where they wanted to go. The vastnumber of passlaw prosecutions is a measureof this. And the morethe controlswere violated,the more costly it becameto maintainthe machinery of control, in Greenberg'swords again, 'the greatly expanded coterie of clerks, managers, magistrates, inspectors and police'.50 The ultimate

46. Hindson, ' Pass Systems ', p. 5. 47. Hindson, Pass Controls,p. 10. 48. Savage, 'Pass Laws', p. 181. 49. Stanley, B. Greenberg, 'The state in contemporary South Africa: the contradictions of control' (unpublished paper, 1984), p. 25. See also Savage, 'Pass Laws', p. 205; Hindson, Pass Controls,pp. 47-48. 50. Greenberg, 'Contradictions', p. 8. THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 73 tendeney was for the whole system to breakdown, a tendeney that will be examinedin the finalseetion of this artiele. Attemptingto eontrolthe movementof Afrieanshas been one key element in the apartheid system. Another has been to exereise tight eontrol over the daily lives of Afrieansin urban areas,while at the same time ensuring the reproduetion of the labourforee. Both of these objeetiveswere to be servedat minimal eost to the stateand to eapital. But againefforts to realize these aims were to give rise to further eontradietionsand eonfliets. In partieularthe imperativetowards eontrol . was . . often . . . eontradietoryto the eost mlnlmlzatlon lmperatlve. At the eentreof this eontradietionwas the whole housingquestion. The maintenaneeof eontrol was elosely linked to the nature of the aeeommo- dation provided for urban Afrieans;and housing representedthe major component of reproduetion eosts. Moreover the whole debate about whether to opt for a migrant or a stabilized labour foree the Stallard/ Fagan/Sauer dabate had signifieantimplieations for Afrieanhousing. As wehave seen, migraneyand stabilizationwere not starkalternatives; a eom- binationof the two produeeda systemof differentiatedlabour-power. And differentiated labour-powerneeessitated a poliey of differentiatedhousing. Althoughit was neverfully realized,the aimwas to aceommodatemigrants in compoundsor hostels, and stabilizedfamiliesin formaltownships. Another eontradietionhinged aroundthe question of where aeeommo- dationfor urban Afrieansshould be sited. Many employerspreferred to aeeommodate their workers elose to the workplaee in eompounds. Workers werekept under tighter eontrol; they weremore immediately able;and avail- eliminatingthe need for transportto the workplaeeredueed repro- duetioneosts. In some eases, as we have seen, eompoundswere not just preferredby employers,but deemedto be an absoluteneeessity. This was theease at the Kimberleydiamond mines and Randgold mines. It was also toa lesser extent true for the Cape Town and Durban doeks where short- termfluetuations in labour demand necessitated having the work-foree livingelose at hand. However, eompoundsand hostels, whieh were often locatedin eentral areas of cities, seemed to pose problems of soeial and politiealcontrol. Enclavesof singleAfrican males were anathema to segre- gationistswho wantedas far as possiblethe Afrieanpresenee removed from centralbusiness and residentialareas outside of workinghours. During the course of the last fifty years or so the imperativetowards segregationand eontrolhas graduallyeome to weigh more heavily. ledto the It has constructionof more African townships. And, as the central residentialand businessdistriets have grown, so have these townshipsbeen removedfurther and furtherout to the urban peripheries. This periph- eralizationof African living spaee is well illustratedby the Cape Town example.As CapeTown expandedin the earlytwentieth eentury it began 74 AFRICAN AFFAIRS to impingeon Ndabeni, sited on what cameto be desirableindustrial Langa,the land. new townshipbuilt in the 1920s, was sited furtherout By the 1 of town. 950s white areaswere again impinging on Langa,so a new Nyanga,was township, establishedeven furtherout.5l In recentyears with the build- ing of Khayalitsha, the sitingof living spacehas been madeyet moreremote. This has beenthe generaltrend in manyof SouthAfrica's urban centres. It has been takento an extremein some areaswhere commuter townships been built, have often in bantustans,as an effortto ensurenot only the and physical social distanceof Africansbut also theirpolitical distance. The issue of siting living spacehas thus given riseto one centralcontradic- tion, between the imperativetowards segregation and control, on the one hand,and the necessityto minimizereproduction costs on the other. The peripheralsiting of townships raised the cost of transportingAfricans to theirworkplaces. And the building of a more formal style of housing to accommodateAfrican familiesinvolved an investmentof capital. This in turngave rise to divisionswithin the state,and betweenthe stateand capital asto who shouldbear these costs. Both the state and capitalwere concernedto minimizethe ducing cost of repro- the African labour force. The main componentof costswas reproduction accommodation,and the question constantly arose as to who shouldbear the main responsibilityfor housing costs. The centralstate, unwillingto burden taxpayers,tended to argue that the responsibilitylay withthe local state and with capital, both being the chief beneficiariesof cheaplabour in any particularurban area. The local state, unwilling to burdenratepayers, claimedthat capitalbenefited most from cheapAfrican labourand should thereforeeither contributedirectly to housing costs or giveworkers a high enough wage to enable them to pay economic rents. Capital,naturally unwillingto burdenitself with these costs if at all possible, andthe centralstate arguedtogether that all ratepayersbenefited from cheap Africanlabour and that they should therefore carry the main burden of housingcosts.52 Who was to bear the burdenin practice? Duringthe 1920s the responsi- bilitylay mainly with the local state. Both the 1920 Housing Act and the 1923Natives (Urban Areas)Act imposedon local authoritiesthe obligation toprovide housing. But generally municipalities failed to fulfil this obligation.In the case of Durban this was very much an evasion of responsibility,arising fromthe municipality'sunwillingness to drawon the generalborough fund for Africanhousing. In the case of Johannesburgit wasdue moreto the city's weakfinancial base. The city derivedlittle direct financialadvantage fromthe miningindustry as miningland was not subject

51.Monica Wilson and Archie Mafeje, 52. Langa(Cape Town, 1963), pp. 5-6. Maylam, 'Shackled by the Contradictions', pp. 6-7. THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 75 to rates. Taxes on the profits of mining and centralstate.53 manufacturingwent to the The upshot was that the African housing tension question became a source of between the central state and 1920s. the municipalities from the Municipalitieslike Johannesburgcalled controlmeasures in for more effective influx orderto reducethe size of the that requiredhousing. city's Africanpopulation While this was not immediately municipalitiesremained reluctant to forthcoming,the An official shoulderthe responsibilityfor governmentreport gave an housing. 1920s: indication of the situation in the

Demands for more adequate and effective Native government of the urban populationand for an improvement dence led to in their conditions of resi- the passing of the Natives Under this (Urban Areas) Act in 1923. Act the obligationwas fairly shoulderthe placed on the municipalitiesto financialburden of providing amenitiesfor the adequatehousing and urban Natives;but in the absenceof any ensure the proper effectivemachinery to carryingout of the intentions of municipalities,progress was Parliamentby the neitheruniform nor spectacular.54 Thiswas certainly the case in Durban where the underthe 1923Act, first township to be built Lamont,was only openedin 1934.55 While the central state was placing the burden of African municipalities,capital was also evading housing on sphere. the financialresponsibility in The low wages paid to African this couldnot workersmeant that chargeeconomic rentals for their municipalities and1930s housing. As it was, in the 1920s Africansin Johannesburgoften paid oftheir income in betweenone-third and a half rent.56 Municipalsub-economic ineffect subsidizing housing schemeswere capital. Municipalitiesresented role,as a Durban City having to play this Councilmemorandum of 1947 subsidyis in fact a wage shows: 'As a housing subsidy the Municipalityis Wages,and, by housing Native subsidizing Native employees on a sub-economic Councilis subsidizing in every case basis, the the employerof that labour'.57 Municipalsub-economic housing onlythe schemesfor a long time wages of Africanworkers, but subsidizednot building also, in a sense, the wages of workers. Before the 1950s, white accordingto the 'civilized 53.Eddie Koch, labour' 'The destructionof Marabi 1932-1938'(unpublished paper, culturc- urbansegregation in pp.187-89; Universityof the Johannesburg, A. W. Stadler, 'Birds in the WitwatersrandHistory Workshop,1984), 194s1947',ffournalof Southern cornfield:squatter movements in 54.Quoted African Studies, 6 (1979),pp. Johannesburg, in Koch, 'MarabiCulture', p. 11S17. 55.Lousie Torr, 'The 187. socialhistory of an urbanAfrican (unpublishedMA thesis, Universityof community:Lamont 56.Koch, 'Marabi Natal,Durban, 1985), p. 44. c.1930-1960' 57. Culture',p. 189. DurbanTown Clerk'sFiles, DurbanCity JudicialCommission on Native Council'sMemorandum, ch.8, p. 13. Affairsin Durban,vol. I, 76 AFRICAN AFFAIRS polieythe use of eheaperAfriean labour in buildingprojects was prohibited. The employmentof more expensivewhite buildersin township construc- tion raisedcosts by over 100per cent.58 Artifieiallyhigh constructioncosts and the low rentalsemanating from poor wagesput the squeezeon munici- palitieswhich becameincreasingly unable to meet the growingdemand for Afrieanhousing. By the late 1940s,as vast squattersettlements mushroomed around urban centres, the eentralstate began to reeognizethe seriousnessof the Afriean housing erisis and the predieamentof the munieipalities. Aecordinglyit beganto devise ways of reducinghousing costs. In 1951a Bill was passed permittingthe use of Afrieanartisans in buildingtownships. tlthiswas to eontributein the 1950sto a eonsiderablereduetion in the costs of construct- ing African housing.59 It was also in the early 1950s that the site-and- servieeapproaeh was offieiallyadopted. Controlledand servieedself-help housingsehemes for Afrieansredueed the finaneialburden while at the same time avoiding the dangers seemingly posed by the uneontrolledsquatter settlements.60 In the meantime,the NationalBuilding Researeh Institute, set up in 1946, had been trying to design new forms of standardized,low- eost, mass housing for low wage-earners,again with a view to minimizing eosts without endangeringeontrol.6l It was also in the early 1950sthat the stateattempted to shift more of this eost burdenon to capital. An earliereffort to do this hadfailed. The 1937 Native Laws Amendment Act had permitted municipalities to eompel employers to aeeommodatetheir blaek workers.62 However, the impli- eationof this would have been more eentrallysited eompoundsand hostels, eonflieting with the segregationistimperative. So the strategy subse- quentlydeveloped in the early1 950s wasnot to eompelemployers to provide aeeommodationfor black workers,but to make a direct finaneialeontri- bution towardsreproduetion eosts. This was the prineiple embodied in the 1952 Native Serviees Levy Aet. This measureobliged employersin cighteenmajor urban areas to pay to the loeal authority2s.6d. for every six days workedby an adult male Afriean;employers of domestieworkers and employerswho aecommodatedtheir own workerswere exempted. These eontributionswould be paid into loeal native revenueaeeounts and be used to financehousing and serviees.63 58. See Koch, 'Marabi Culture', pp. 19s91; Saunders, 'Ndabeni to Langa', p. 21 1; Torr, 'Lamont',. 114. 59. Bloch and Wilkinson, 'Urban Control', p. 25. 60. Wilkinson, 'Adequate Shelter', p. 78. 61. Bloch and Wilkinson, 'Urban Control', p. 26. 62. Bekker and Humphries, AdministrationBoards, pp. 124-25. 63. Wilkinson, 'Adequate Shelter', pp. 79-80. In the early 1970s the Native Services Levy Act was superseded. From 1974 employers were compelled to pay a monthly contribution to help finance the adminis,ration boards. Bekker and Humphries, AdministrationBoards, pp. 13S37. More recently the same principle of employer levies has come to be embodied in the emerging system of regional services councils. THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 77 In this examinationof the long-standing struggle between the central state, local state and capitalto shift the burdenof reproductioncosts on to each other, one importantpoint has been obscured. That is that urban Africansthemselves bore the majorpart of this cost burden. The crucial mechanismhere was the self-financingsystem of natise revenueaccounts. Durbanwas a pioneerin this system, institutinga nativerevenue account in 1908. The system came to be formalizedat a nationallevel by the 1923 Natives (UrbanAreas) Act. Native revenueaccounts were sub-accounts of generalborough accounts. Accruingto them were rents paid by African tenantsin municipalhousing, revenuefrom municipal beer sales, and other fees paid by Africans. This money was to be spent on administration,the provisionof services,and housing for Africans. Althoughnative revenue accountsoften deriveda substantialincome, it was never sufficientby itself to ensurethe requisiteprovision of housing for Africans;hence the massive housing shortagethat had developedby the late 1940s. Given this weak financial base, along with the unwillingness of the central state, local state or capital to bear the cost burden, thousandsof urban Africanswere forced to devise their own forms of shelter. By the 1940s vast shack settlementshad emergedaround many of South Africa's urban areas.64 Inherent in the growth of these settlementswere further contradictions. On the one hand they relievedthe centralstate, local state and capital of the considerablefinancial burden that would have been involved in providing more formal housing for shack-dwellers. On the other hand, the shack settlementswere zones that largely fell beyond the controlof the localstate. The settlements,lacking basic water and sanitary services,were consideredto be healthhazards. And they were deemedto be havens for criminalsand politicalactivists. Althoughmore researchis requiredinto the socialhistory of squattercommunities before we canobtain a clearerpicture, there is little doubtthat their existenceseriously weakened local state controlover urbanAfricans. The state itself was well awareof this; hence there followed the massive programmeof urban removalsand relocationthat beganin the 1950s. As we have seen, the exercise of control over urban Africans involved costs, which the variousparties wanting control were generallyunwilling to bear. This was one of the major contradictionsin the urban apartheid system. Moreover,control not only involved large financialcosts; it also generatedresistance, giving rise to furthercontradictions. Urban struggleshave many facets. As Cooperhas put it, Conflictover work discipline and housing, over what kinds of conduct would be legal or criminal, over what forms of social relations could

64. See, for instance, Stadler, 'Birds in the Cornfield', and Maylam, 'The "Black Belt": African squatters in Durban 1935-1950', Canadian3tournalof African Studies, 17 (1983). 78 AFRICAN AFFAIRS develop in workplace,urban residence, and ruralvillage, and over the valuesand culturesthat could develop inside urbanspace all shapedthe city and furtherpatterns of struggle.65

This has certainlybeen trueof twentiethcentury South African cities, where many such struggles have been played out. In some cases the city has simplybeen the arenain which strugglesthat werenot specificallyrelated to urbanissues, have been waged. But in many instancesthe struggleshave been in andof the city. They havebeen foughtover issuesconnected to the reproductivesphere or arisingout of the urbanapartheid system. The pass system has been one such issue giving rise to struggle. This strugglehas generallytaken two forms:first, defianceof the pass laws, and, second, campaignsfor the removalof the laws. As we have alreadyseen, defianceof the laws occurredon a massive scale. Between 1916 and 1984 nearlyeighteen million Africanswere arrestedor prosecutedunder the pass laws and influx control regulations.66 Millions more must have evaded arrest. Since 1910 there have been several organized campaignsagainst pass controls. In the 1910s there was a campaignin the Free State directed against the issuing of passes to women. In 1919 there was an anti-pass agitationon the Rand. In December1930 hundreds of Africansin Durban burnt their passes. There were attemptsto orgarlizeanti-pass campaigns in 1944 and 1946. The pass system was one of the targets of the 1952 defiancecampaign. In the mid-1950sthere was a massivenationwide pro- test against passes for women. And in 1960 both the ANC and PAC launchedlarge-scale anti-pass campaigns. Other forms of resistancehave centred around the housing question. Urban removalsand relocationin particularprovoked resistance. When Africanswere orderedto move from their existing homes in townshipsor shacksettlements to remote,bleak townships, they often resistedthe order. Hundredsof Africansin Cape Town, for instance, refused to move from Ndabeni to Langa in the late 1920s and 1930s until eventuallyforced to.67 Squatting in a shack settlement could itself be a form of resistance. Certainlyshack settlements arose partly out of a housingshortage, but there is evidenceto show thatshack-dwellers often preferredto live in areaswhich were largely free of strict state control. Stadlerhas shown how squatter movementsin Johannesburgin the 1940s repeatedlydefied or circumvented the local state'sattempts to removethem.68 Transport,another element in the reproductivesphere, was also a site of urbanstruggle. As Piriehas observed, '[o]n a materialplane, transport was 65. Cooper, Strugglefor thecity, p. 44. 66. Savage, 'Pass Laws', p. 181. 67. Saunders, 'Ndabeni to Langa', pp. 214-19. 68. Stadler, 'Birds in the Cornfield'. THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 79 an instrumentof marginalizationand dominationvia its underwritingresi- dentialsegregation, and via the time andfinancial sacrifices it imposed'. As an instrumentof dominationit also became a focus of resistance:'Covert resistanceto transportinadequacies featured fare evasionand disregardfor travel rules. Boycott of bus services was considerablymore spectacular and threatening'.69 Of the latter, perhapsthe best documentedexamples arethe bus boycottsat Alexandraand Evatonin the 1940sand 1950s.7? Stateurban policy and practicehas thus been riddledwith contradictions and inconsistencies. At the root of these was one fundamentalcontradic- tion: the attempt to attain the unattainable,the securingof labour-power with the minimalpresence of people. Out of this arosefurther contradic- tions, divisions and struggles. Urban apartheidbecame an issue around whichvarious contestants, the centralstate, local state, capital and labourj all sought to furtheror protecttheir particularinterests. In the 1960s, when the urbanapartheid system reached the peakof its functionality,many of the contradictionsand conflicts were submerged or contained. From the 1970s, however, the system became increasinglydysfunctional. In the 1980s it has shown signs of breakingdown.

Epilogue: the breakdozvnof urban apartheid During the courseof the twentiethcentury the evolving system of urban apartheidcame to rest on four main pillars:control over the movementof Africansinto urban areas;the regulationand regimentationof the lives of urban Africans through such mechanisms as segregation, curfews, and controlled housing; the developmentof a self-financingsystem whereby Africanscame to beara largeshare of their own reproductioncosts; and the cooptationof membersof the urbanAfrican petty bourgeoisiethrough the creationof local institutions,such as advisoryboards, community councils, andblack local authorities, institutions that provided a thin veneerof partici- pation and democracy. Although these pillars have never rested on firm foundations, there have been phases when they have looked very solid. However,in the last sevenor eight yearsmore and more cracks have begun to appearand there are signs that the pillars, and the whole urban apartheid edifice,are crumbling. The formaloperation of influxcontrol, as exercisedthrough the passlaws, wasabandoned by legislativeenactment in 1986. This did not meanthat all controlover the movementof Africansinto urbanareas was removed,but it

69. G. H. Pirie, 'Johannesburg transport 1905-1945: African capitulation and resistance', ffournalof HistoricalGeography, 12 (1986), p. 51. 70. See A. W. Stadler, 'A long way to walk: bus boycotts in Alexandra 194s1945', in P. Bonner (ed.), WorkingPapers in SouthernAfrican Studies, vol. 2 (Johannesburg, 1981), pp. 228-57; and Tom Lodge, BlackPolitics in SouthAfrica since 1945 (Johannesburg, 1983), pp. 153-82. 80 AFRICAN AFFAIRS did reflecta shift in state urbanizationpolicy. This shift had begun in the late 1970s with the report of the Riekert Commission. This report rep- resentedin parta returnto the thinkingof the FaganCommission of the late 1940s. The Riekertreport departed from rigid Verwoerdianideology by recognizingthe existenceof a permanenturban African proletariat. This permanentproletariat would constitute a classof 'urbaninsiders' who would be given preferentialaccess to employmentand housing. At the sametime stricter control would be exercised over the 'outsiders', those without permanenturban rights. The pass laws survivedthe Riekertreport. But duringthe early 1980s, therewas growingpressure even for the abolitionof the pass system. This pressurecame from sections of capitaland from organizations like the Urban Foundation. It was argued that the reality and inevitabilityof African urbanmigration had to be acceptedand that the marketmechanism should be left to determinethe movementof labour.7l For its partthe government probablybelieved that abolitionof the pass lawsmay have servedas a sop to internationalopinion. Furthermorethe machineryof pass control was becomingincreasingly costly to maintainduring a time of fiscal constraint. So in 1986 the notoriouspass was formallyabolished, to be replacedby the single uniform identificationdocument carriedby membersof other race groups. It would probablybe a mistaketo overemphasizethe significance of this change, as regulatiorlof the movementof Africansinto urbanareas would still be exercisedin more indirect ways, through control relatedto employmentand housing. However,the abolitionof passesis an indication that state policy has failed in its ultimateobjectives. It has been unableto construct that watertightdam to block the flow of 'surplus people?from impoverishedrural areas to the cities. With the damsconstantly bursting, state policy has now become concernedmore with channellingthe flow of people than haltingit. If influxcontrol was one of the more brittlepillars of the urbanapartheid edifice, the whole apparatusof segregationand controlhas been ratherless so. The basic infrastructureof urbanapartheid remains intact. Millions of Africansare still forcedto live in segregatedtownships on the outskirtsof towns and cities. And duringthe past few yearsof large-scaleurban unrest one of the key strategic functions of these townships has proven to be largely effective. One objective in the planning of townships was to site them in such a way that unrest and rebellion could be containedwithin them. Anotherwas to makethem easily accessibleto the police and troops. The fact that there has been a minimalspill-over of violenceoutside segregated blackurban areas during the past four years suggests that these objectives have largelybeen realized. 71. See, for example, RandDaily Mail, 30 April 1983, 31 October 1984, 6 November 1984, 6 December 1984. THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 81

There have, though, also been strong signs that the level of eontrolover urban Afrieans is weakening. Whereasonee eontrol eould be exereised administrativelythrough munieipal offieials and, later Administration Boardpersonnel, now eontroldepends heavily on a militarypresenee in the townships. And even that militaryeontrol has at times and in plaees been tenuous. At the height of the townshiprebellion in 1985and 1986the state lost eontrolof some blaekurban areas. Liberatedzones were ereated,and loealadministrative power was takenover by streeteommittees. Although state eontrolhas largelybeen regained,it has only been regainedby foree, and it seems unlikelythat it ean ever be regainedby administrativemeans. A furthersymptom of weakeningstate eontrol is the massive growth of shaeksettlements around the majorurban areas. As we have seen, housing representsa form of eontrol in itself. In the 1940s shaeksettlements had mushroomedin South Afriean eities. From the 1950s the state began to clearthese uneontrolledliving zones by moving their oeeupantsinto newly built townships. In the 1980s,there is a returnagain to the situationof the 1940s,although now it is on a mueh largerseale. It has been projeetedthat by the year 2000 there eould be almost three million blackpeople living in shacksettlements in and aroundgreater Durban alone.72 The statehas lost control of the blaek housing situation, and in losing that eontrol another pillarin the urbanapartheid edifiee is erumbling. The fiseal base of urban apartheidhas also been weakening. Until the 1980s this had generallybeen strong. Revenue, aeeruingfirst to munici- palities and then to AdministrationBoards, was derivedfrom a numberof sources:the profitsfrom beer and liquor monopolies,rents, state subsidies and employerlevies. The AdministrationBoards, when they were insti- tuted in 1972, took over municipalnative revenue aecounts whieh at the time mostly carried healthy surpluses. And for most of the 1970s the AdministrationBoards themselves eontinued to reeordsurpluses.73 From the early 1980s, however, revenue from these sourees has been declining. First, profitsfroln beer monopolies began to dropdrastieally, as more and more Africansshifted away from sorghumbeer consumptionand bought liquor at commercialoutlets. Then the government decided to privatizethe townshipliquor trade. In 1983 some AdministrationBoards began to sell their own liquor outlets to blackbusinessmen. Many Board officialsmuch resentedthe government'snew privatizationpolicy becauseit weakened the Boards' financialposition.74 The seeond most important source of AdministrationBoard revenue in the 1970s was the levy on employers,who, since 1974, had been requiredto pay a monthly eontri- butionto the Boardsto help financethe administrationof influxeontrol and

72. Natal lkIercury,26 September 1984. 73. Bekker and Humphries, AdministrationBoards, p. 132. 74. Bekker and Humphries, AdministrationBoards, pp. 37-38, 132-36. 82 AFRICANAFFAIRS surplusesfrom otherlabour allocation functions. During the early 1980s contributionalso declined.75 this on black Deficits from housing have also been rising. Expenditure of inflation,but housingincreased in the 1970s, partly under the pressure facilitiesafter the alsoas a result of a governmentcommitment to improve enough for these 197S77urban unrest. African wages were not high recently this com- increasedcosts to be passed on to residents.76 More in the shapeof ponentof the fiscalbase has comeunder even furtherpressure households widespreadrent boycotts. By August 1986 about 300,000 the stateabout countrywidewere involvedin the boycott,which was costing R30million a month.77 has cometo have The deterioratingfiscal base of Africanlocal government the struc- seriousimplications for anotherdimension of urban apartheid, elements of the turesof cooptation and collaboration. Like some other back to the 1923 urbanapartheid system, these structurescan be traced the establishment Natives(Urban Areas) Act. This act madeprovision for were not compelled ofNative AdvisoryBoards. Although municipalities would have its toimplement this provision,the idea was that eachtownship residents,elected ownAdvisory Board, comprising at least three township the Boards orappointed, as members. Even wherethey were established, held by a white werelargely powerless. The Board chair was always advisory bodies person;the Boards themselves were nothing more than but the latter (theycould make recommendationsto municipal councils9 they chose); more- could,and often did, ignore those recommendationsif the Africanpetty over,the Boardstended to be dominatedby membersof interests. In 1968 bourgeoisiewho used theirposition to furthertheir own Councils. The thegovernment provided for the creationof Urban Bantu Boardswas the onlyapparent difference between these and the Advisory municipalitieswere name. The Councils were equally powerless. And Councils had not compelledto establishthem; by 1977 only twenty-four created.78 been Community Urban Bantu Councils were superseded by the 1977 Councils were CouncilsAct. Unlike their predecessor,the Community and trading entrustedwith some powers;the allocationof accommodation mostimportant ones. sites,and the maintenanceof essentialservices were the circumscribed. But the position of the CommunityCouncils was severely minister. There Their powerswere subjectto the will and decisionsof the Councilsand was an uncertaindivision of power between the Community was weak, as AdministrationBoards. The fiscal base of the Councils the Boards. employerlevies and liquor profitsstill went to the Administration 136-39. Bekker and Humphries, AdministrationBoards, pp. 75. pp. 13947. 76. BekkerandHumphries, AdministrationBoards, Mail, 15-21 August 1986. 77. Weekly pp. 9S97. 78. Bekker and Humphries, AdministrationBoards, THE RISE AND DECLINE OF URBAN APARTHEID IN SOUTH AFRICA 83

And the Councils, which tended to be dominated by businessmen and traderspursuing their own interests, lacked any real popular legitimacy, especially in the largerurban areas where council election polls often fell below 10 per cent.79 The 1982 Black Local AuthoritiesAct, without abolishingCommunity Councils, providedfor four categoriesof local authorityto be established. Town councils in the larger centres, and village councils in the smaller centreswere the two most importantof these. Althoughthe councilswere entrustedwith slightly enlargedpowers under the 1982 Act, they were still subject to absolute ministerialauthority. And they still sufferedfrom a weak financialbase and a lack of popular legitimacy.80 These latter two problems tended to be mutually aggravating. The weak financial base forced the black local authoritiesto increaserents, which in turn fuelled popularhostility to the councils. In the past few yearsseveral black coun- cillorshave been murderedor have receiveddeath threats; and their homes have been attacked. In many areasthe black local authoritiessystem has completely collapsed in the face of popular resistance;in other areas the system only survives with the councillorsbeing given police and military protection. The blacklocal authorities faced popular rejection because they havebeen widely perceivedas an attemptby the stateto forceurban Africans to accept a form of sham democracyas a substitutefor full participationin a genuine nationaldemocratic system. In some respects the black local authorities are the bantustangovernments writ small. And they have increasingly come to perform some uncomfortabletasks on behalf of the state, such as removingillegal squattersfrom townships and evictingrent boycotters. Local governmentin Africanurban areas has in recentyears increasingly become a site of struggle. Communityorganizations have grown up all over the country to challenge a system based on puppet institutions and shamdemocracy. So stronghas been this challengethat yet anotherpillar in the urbanapartheid edifice has crackedand crumbled. The whole edi- fice had alwaysbeen brittle, liableto breakdown underthe strainof its own internalcontradictions. When faced with widespreadpopular opposition and resistance,it beganto buckleand collapse. The breakdownof urbanapartheid has in part reflectedthe wider failure of the whole apartheidsystem. The ultimate objective of the apartheid blue-print was to create politically and economicallyviable 'homelands', each characterizedby a particularethnic identity. The ofEcialhope was that, as the economies of the 'homelands' developed, more and more 79. Bekker and Humphries, AdministrationBoards, pp. 99-107; Heather Hughes and Jeremy Grest, 'The local state', in SouthAfrican Reviezv, 1 (Johannesburg, 1983), pp. 123-29. 80. Bekker and Humphries, AdministrationBoards, pp. 11s15; Jeremy Grest and Heather Hughes, 'State strategy and popular response at the local level', South African Review,2 (Johannesburg, 1984), pp. 52-55. 84 AFRICAN AFFAIRS Africanswould move from 'white' areasto the particular'homeland' that was ethnically appropriate,thereby reversing the direction of African migration. In their own 'homeland'Africans would be able to exercise politicalrights, and in so doing they would lose any claimto the franchisein greaterSouth Africa. It is well known that the grand apartheidmodel has never worked in practice. The economies of the 'homelands'have become steadily more impoverished,let alone economicallyviable. Far from Africansmoving voluntarilyinto the 'homelands',massive outmigration has occurredon an ever-increasingscale. This has put enormous pressure on the urban apartheidsystem and renderedmany of its controlmechanisms ineffective. Africanpeople have shown that they will move to those areaswhere they have a best chance of obtaining a subsistence, regardlessof the measures restricting their freedom of movement. At the level of daily existence Africanshave, on a huge scale, defiedthe apparatusset up to controltheir lives. At a wider political level they have organized themselves and engagedin concertedresistance to the apparatusof control. There can be little doubt that the townshipuprising of 1976 was a shatteringblow to an urban apartheid system that was already weakened by its own contra- dictions. The townshiprevolt that beganin 1984has furtherhastened the breakdownof urbanapartheid.