y
 s>
 UNIVERSITY OF THE WITWATERSRAND
 INSTITUTE FOR ADVANCED SOCIAL RESEARCH
 SEMINAR PAPER
 TO BE PRESENTED IN THE RICHARD WARD BUILDING
 SEVENTH FLOOR, SEMINAR ROOM 7003
 AT 4PM ON THE 4 AUGUST1997.
 TITLE: Bargained Liberalisation: The Labour Movement,
 Policy-Making and Transition in Zambia and
 South Africa
 BY G. ADLER and E. WEBSTER
 NO: 423
I. INTRODUCTION*
 African countries are faced with a double challenge: to democratise and to restructure their
 economies towards a new growth path. These are separate, but linked - and potentially contradictory
 - projects.
 There are two opposing views on the relationship between political and economic reform. An
 optimistic perspective reflected in much of the post-1989 policy perspectives of the World Bank and
 other donor organisations is to see these two projects as necessary and complementary. This is best
 captured in the work of Larry Diamond (1988; 1993; Diamond and Plattner, 1993), Richard Joseph
 (1990) and other advocates of "good governance," who raise political conditionality to an equal
 status with economic adjustment.' By contrast, a more sceptical reading of these two processes
 emphasises the contradictions between political and economic liberalisation. As Tom Callaghy
 argues, "if not handled properly, political conditionality might well impede rather than facilitate
 Africa's relinking to the world economy in more productive ways." In the absence of developmental
 democracies, which Callaghy doesn't expect to emerge any time soon, external donors should
 channel aid to "those actually undertaking difficult economic reforms and without being tied
 automatically to political change."(Callaghy, 1994:35) Thus, if there must be a choice between
 economic and political liberalisation, preference is given to the former.2
 A similar equivocation towards democracy can be seen in Tony Killick's (1995) recent work on
 national economic adjustment. On the one hand, he stresses the importance of democratic
 accountability, yet on the other he praises the beneficial effects in East Asia of the "insulation of
 policy-makers from pressures exerted by special interest groups." (Killick, 1995:28) His hesitation
 towards democracy is reflected in his attitude towards trade unions. There are seven references
 towards trade unions, and all refer to them as special interest groups that restrict labour market
 flexibility and create "rigidities" in the economy that slow down its capacity to adapt. (Killick,
 1995:22, 198, 252, 343, 349, 379, 381) This same view lies behind Callaghy's (and others') claim
 that the winners in adjustment are few, appear slowly, and are hard to organise, while the losers
 (including organised workers) are many, vociferous, and well organised - and may use their power
 under democracy to frustrate economic reform. In this sense insulation means protection from these
 constituencies who bear the brunt of adjustment, even if it means abandoning democracy.
 In contrast to this pessimistic view, a recent collaborative project on the policy options facing new
 democracies highlights the tension between political and economic reform as the key problem
 facing new democracies but reaches the opposite conclusion regarding the role of trade unions.
 (Przeworski et al., 1995) In this treatment, economic liberalisation engenders a transitional decline
 in consumption that undermines attempts to consolidate democracy while threatening a return to
 authoritarianism or - in the worst case - social disintegration, descending into "decentralized
 collective violence." (Przeworski et al., 1995:110) The dilemma facing new democracies, the
 authors argue, is "how to create incentives for political forces to process their interests within the
 democratic institutions, when material conditions continue to deteriorate." (Przeworski et al.,
 1995:113)
 The authors offer a social democratic alternative to neo-liberalism consisting of three
 recommendations. First, liberalisation must be accompanied by social policies that minimise social
 'This paper was commissioned by the Multinational Working Group on Labour Movements and Policy-
 Making in Africa of the Council for the Development of Economic and Social Research in Africa. The paper is
 based on documentary evidence, secondary literature, and interviews conducted in Zambia and South Africa among
 labour leaders, government officials and parliamentarians between 1995 and 1997. We wish to thank Jimi Adesina,
 Bjorn Beckman, Sakhela Buhlungu, Matthew Ginsburg, Karl Gostner, and Lloyd Sachikonye.
costs. Second, policies must be designed with a view towards growth, and finally, policies must be
 formulated and implemented through corporatist-style consultation and negotiation beyond the state
 and parliament to include unions, employers and other interest groups. (Przeworski et al., 1995:85)
 "Concertation" is central to their argument: it subjects the reform strategy to the competitive
 interplay of political forces, improves policy outcomes, builds support for the continuation of
 reforms, and helps consolidate democratic institutions. (Przeworski et al., 1995:82)
 To the extent that consolidation depends on concertation, trade unions are essential actors in the
 development of democracy, indeed, consolidation depends on strong union organizations that can
 participate in pacting: reaching and enforcing agreements with capital and the state. The authors
 assert, however, that concertation is impossible because unions "are too weak and too decentralized
 to serve as partners..."3 (Przeworski, 1995:82) The implications of this conclusion for the
 consolidation of democracy are, to be sure, distinctly troubling.
 What is striking about this particular conclusion, as well as those of Killick and Callaghy, is that
 they are not based on any sustained analysis of trade unions' actual involvement in consolidation,
 nor is the gap remedied elsewhere in the growing literature in the field. This shortcoming is not
 simply empirical - that unions have been overlooked as subjects of investigation - it is, more
 importantly, conceptual. In the literature on both the transition to democracy and democratic
 consolidation unions are grasped from one of two perspectives. Some view them as powerful
 special interest groups whose rent-seeking behaviour in defense of their privileges blocks necessary
 economic reforms. Others, from a functionalist perspective, view unions as collective organisations
 with the power to restrain their members' wage demands, a necessary condition for liberalisation.4
 In this approach transition and consolidation are seen as outcomes of agreements between elites:
 labour is assessed in terms of its capacity to deliver its constituency to the conservative pact. This
 is ultimately a narrow perspective that underestimates the broader strategic role that collective
 actors - in particular the labour movement - have been able to play in democratisation.
 By contrast, our view is that labour is a collective actor capable of shaping the character of
 democratisation through the disciplined and strategic use of power. A powerful and strategically
 sophisticated labour movement has the potential to reconfigure democratisation through
 participation in negotiated compromises. These compromises have the potential to discipline capital
 and the state and ensure that the social costs of adjustment are not borne by workers alone. Pacting
 should not be seen simply as functional to the consolidation of democracy, but as a conflictual
 process of class compromise, the results of which may produce workers' loyalty to democratisation.
 This approach thus emphasises the central role of the working class and working class organisation
 in understanding how democratic regimes are established.5
 In this paper we examine two contrasting cases of trade union involvement in liberalisation, Zambia
 and South Africa. We chose Zambia and South Africa as they are the most important cases in Africa
 where labour played a central role in a successful transition to democracy. However, during the
 period of democratic consolidation the Zambian labour movement has been marginalised and has
 played little role in economic policy making. Relatively unconstrained by labour, the Chiluba
 government has pursued a fairly orthodox neoliberal adjustment, with high social costs which may
 in time threaten democratisation.
 By contrast, the South African labour movement has achieved substantial - though uneven -
 outcomes from its engagement in policy making. Labour, through its collective power, has
 achieved an institutionalised voice in decision-making, which has forced government and business
 to negotiate certain fundamental economic policies, including privatisation, labour law and
 employment standards.
At the centre of both cases is the role of civil society in democratisation. In this paper we highlight
 the centrality of the labour movement, not as a substitute for civil society but as one of the most
 powerful formations within it. Under what conditions are civil society formations able to shape the
 outcome of democratisation? Under what conditions can they develop institutions that create new
 rules of the game to encourage a bargained liberalisation process?
 Bargained liberalisation would appear to be a contradiction. Liberalisation implies the opening up
 of an economy and its unilateral adjustment to the market as a global force. It assumes an
 omnipotent global market to which individual states and social movements must conform on the
 basis of international competition. Bargaining, on the other hand, implies reaching non-zero-sum
 agreement between independent actors based on mutually accepted rules. It also presupposes the
 creation of institutions that provide opportunities for these agreements to be reached and enforced.
 The extent to which these processes are seen as contradictory depends on one's prior view of
 globalisation. If globalisation is seen as omnipotent, than bargaining is by definition impossible. If,
 on the other hand, globalisation is seen as exaggerated, as a contradictory and uneven social
 process, in short as contestable, then bargaining opens up the opportunity for challenging a
 country's engagement with the international economy.6 Under bargained liberalisation, movements
 and institutional innovations become central to any understanding of responses to liberalisation.
 This paper focuses on labour as one of the key actors in bargained liberalisation.7 We stress two
 factors as central to labour's role in bargained liberalisation. The first focuses on labour's capacity
 to mobilise as an independent actor - from capital and the state/nationalist movement - a capacity
 founded on its own organisation of workers on the shop floor, its own finances, leadership, and
 organisational culture. The second concerns labour's capacity to use its power strategically: its
 ability to marshall its mobilisational capacity as an independent actor in pursuit of a broader
 transformative vision. This strategic capacity gives labour the ability to know when to compromise
 and when to advance in pursuit of its vision. Elsewhere we have called this approach radical
 reform, in which labour can use its power to influence policy-making in the direction of long-term
 social and economic transformation. (Adler and Webster, 1995)
 The independence and strategic capacity of the Zambian and South African labour movements will
 be assessed against the background of two key variables: the character of the state and the country's
 insertion into the international economy.
 The South African case represents an alternative to the two opposing views on the relationship
 between political and economic reform. It opens up the possibility - expressed but not investigated
 by Przeworski and others - of a bargained liberalisation process. At the centre of our argument is
 the significance of the emergence of an institutionalised voice for independent formations in civil
 society to shape the direction of political and economic liberalisation. Rather than viewing civil
 society uncritically as an adjunct of liberal democracy - as the "governance" theorists would have it
 - or as an impediment to the hard choices of economic adjustment, we argue for what Peter Evans
 has called "an effective state-society synergy." "The idea of 'synergy'," he argues, "implies that
 civic engagement strengthens state institutions, and effective state institutions create an environment
 in which civic engagement is more likely to thrive." (Evans, 1997, 3)
 The notion of a state-society synergy has important implications for policy making in Africa. First,
 without some form of bargained liberalisation one is likely to have neither political democracy nor
 economic reform. A prerequisite for bargained liberalisation is civil society formations - such as
 labour - which are both independent and capable of using their power strategically. It requires that
 such formations gain an institutionalised voice in policy formulation through which they can engage
 systematically with the state over economic and social policy. These conditions appear most
favourable in South Africa, though we argue that they could emerge in Zambia and are relevant
 elsewhere in Africa.
 II. ZAMBIA
 A. The Colonial Period: The Emergence of the Labour Movement
 The roots of labour organisation in Zambia extend back to the founding of the most important
 capitalist enterprises in the country: the huge mines of the copperbelt. Their establishment in the
 late 1920s was soon followed by collective action and organisation by black workers, who
 sidestepped the officially sanctioned dispute resolution system based on tribal elders.
 The first major strike occurred in 1935. Though brief and easily suppressed, it demonstrated that
 "African mineworkers were already conscious of a common interest and were capable of organising
 concerted resistance within the urban environment." (Roberts, 1979:202) The African workers'
 action prompted white miners in 1936 to launch their own organisation - the Northern Rhodesia
 Mineworkers Union [NRMU] - to protect their own interests against the threat of African
 encroachment.
 In 1940 the NRMU won wage increases after a short strike stimulated by wartime increases in the
 cost of living. But African workers - whose deprivations were far greater - launched a week-long
 strike across the copperbelt which was ruthlessly suppressed. A Commission of Enquiry later
 acknowledged the wage inequalities and granted modest increases for some workers, but refused to
 recognise African miners' status as workers or their right to form unions.
 Notwithstanding official indifference, African political and workplace organisation grew during the
 second world war. Welfare societies were established across the copperbelt and then in workers'
 home rural districts, bypassing government-sanctioned and tribally based Urban Advisory Councils.
 These societies provided the first colony-wide political network for Africans, and in 1948 their
 leaders organised them into the Northern Rhodesia African National Congress, the first nationalist
 party. (Roberts, 1979:204; Rotberg, 1965; Tordoff, 1974)
 Side-by-side with the development of community-based welfare societies, employees began to build
 representative organisations in the workplace. After the war these efforts were boosted by the new
 Labour Party government's support for "non-political" trade unionism in Britain's African colonies.
 (Davies, 1966; Cohen et al., 1978 ) Strong industrial unions for Africans emerged rapidly in
 Zambia. By 1948, there were unions at each of the copper mines, and in 1949 these merged to
 form the Northern Rhodesia African Mineworkers Union [AMWU]. Under pressure from the
 British colonial authorities the mining companies recognised the AMWU. It soon enrolled more
 than half the African workers on the mines, and became the cornerstone of the Zambian labour
 movement. (Roberts, 1979:205; Burawoy, 1972:13-16) Soon national unions were formed on the
 railways, and among government, building and timber, and garment workers.
 Notwithstanding the decentralised and apolitical emphasis in colonial labour policy, these
 organisations combined to form the Northern Rhodesia Trade Union Congress [NRTUC] in 1951.
 In the following year the NRTUC joined with the ANC to fight white settlers' efforts to establish
 the Central African Federation and gain direct rule over the colony. (Akwetey, 1994:44; Rakner,
 1992:72; Bates, 1971) Their resistance failed, and in 1953 Britain - now ruled by a Conservative
 Party dedicated to supporting white supremacy - granted self-rule to white settlers in the three
 colonies of Northern Rhodesia, Southern Rhodesia, and Nyasaland. (Minter, 1986; Roberts, 1979)
 There are two important points that emerge from this early history. The first is that the unions - the
AMWU in particular - developed from the bottom up, and were based on relatively strong factory
 structures in the key colonial industries. These gave the unions power well beyond their numerical
 size. Strong branch structures gave the miners the ability to stop production of the colony's crucial
 foreign exchange earner. But these same structures meant that national leadership could not simply
 impose its will on the rank and file. (Bates, 1971:202)8 The miners in particular demonstrated
 considerable militance. According to Bates, between 1950 and 1963 (the year before
 independence) there were 57 strikes and 143 other disputes at the Roan copper mine alone. (Bates,
 1971:95) The strikes were caused by a multiplicity of social and economic factors, but as the
 national struggle intensified, so racial conflict on the mines heightened. Union organisation gave
 rank and file workers the ability to challenge these conditions directly.
 The second is that prior to independence the unions were never subordinated to the nationalist
 parties. Indeed, support for the nationalist struggle did not imply absorption into the ANC, but was
 rather a constant point of debate in the labour movement, exemplified by the mineworkers* refusal
 to bring their members out in an anti-Federation strike in 1954. The NRTUC was divided between
 a radical leadership supportive of the ANC and the miners. The ANC itself split in 1958 (over the
 choice of strategies for opposing the Federation) with more militant members forming the Zambia
 African National Congress (which later changed its name to become the United National
 Independence Party - UNIP). (Chipungu, 1992)
 The realignment in nationalist politics precipitated a parallel split in the labour movement. The
 miners followed the ANC while all other unions affiliated to the NRTUC broke away and created the
 Reformed Trade Union Congress [RTUC]. In 1960 the RTUC's General Council resolved to support
 UNIP as "the most progressive party," though the two organisations remained formally separate. But
 the institutional divisions in the labour movement were soon healed. The pro-ANC leadership in the
 AMWU was removed in 1961, paving the way for the miners to rejoin the mainstream of the labour
 movement, which was reunified under the United Trade Union Congress (UTUC). While the UTUC
 re-endorsed the NRTUC's political support for UNIP, its two largest affiliates - the miners and the
 railway workers - remained aloof politically. (Rakner, 1992:73-75) But the legacy of these splits
 produced a caution in the labour movement around political questions: support should not
 compromise solidarity or autonomy.
 B. Zambia's First Independence: Labour's Struggle for Autonomy under Productionism
 After 1964 UNIP pursued labour policies generally similar to those of governments in other newly
 independent African states. Under the ideology of productionism trade unions were seen as vital to
 improving productivity, expanding output, and increasing foreign exchange for financing ambitious
 programs of social welfare and reconstruction. Under productionism unions were to be included in
 corporatist institutions and transformed from structures representing their members' interests into
 administrative arms implementing state policy. Above all, unions were expected to produce labour
 discipline.
 In the first years after independence in many countries this role demanded that unions quell rank
 and file opposition to the productionist regime itself - working harder for less money - and
 members' anger against expatriate managers and the untransformed colonial labour regime. In
 some cases - such as Tanzania and Kenya - unions sided with their members and were defeated and
 then reorganised into a compliant national union centre by their erstwhile nationalist allies.
 (Bienefeld, 1975, Bolton, 1978; Shivji, 1976, 1986; Sandbrook, 1975) In others, most notably
 Ghana, union leaders collaborated with the productionist regime and attempted to discipline
 members, with disastrous results for union organisation. (Crisp, 1984; Jeffries, 1975, 1978)
 The Zambian case falls between these extremes, and marks a rather different trajectory from unions
elsewhere in Africa. UNIP moved to consolidate pre-independence union formations into a single
 statutory trade union centre, the Zambian Congress of Trade Unions [ZCTU]. UNIP's aim was for
 the ZCTU to act as a labour arm of the party's national development strategy, communicating to
 workers the productionist ethic. But this policy was implemented in a voluntarist fashion with a
 minimum of coercion.
 Indeed, Bates' research demonstrates that rank and file workers were relatively uninterested in the
 new labour regime, and sought rather to gain the fruits of independence. As elsewhere on the
 continent, opposition to racist practices by expatriate managers was the spark for many strikes.
 (Burawoy, 1972) The number of recorded strikes increased dramatically after independence and
 remained high for the rest of the decade. (Rakner, 1992:76) Often these were strikes aimed as much
 against the union and the state as against management, as members were no more inclined to listen
 to the ZCTU's or UNIP's productionist arguments than they were to obey UNIP's colonial
 predecessors. Government policy was thus fatally flawed: if the union attempted to operate as a
 government agency and ignored this rank and file power, it would fail as a union and be unable to
 operate as a "labour arm" of government. (Bates, 1971:209) This "paradox of productionism"
 would prove to be an unresolvable dilemma for the Zambian government for the next three
 decades, and would serve as the foundation of union autonomy.9
 According to Rakner, this paradox was felt most strongly within the ZCTU itself. Its stronger
 affiliates - which predated the founding of the federation and were largely untouched by its creation
 - demonstrated a "distinctive unwillingness...to be subordinated to party control." The affiliates -
 rather than the federation itself - "fought against the attempts to incorporate the unions into the
 party and defended their autonomous position." Irrespective of the ZCTU's progovernment
 pronouncements, the affiliates sought solutions to their problems through the independent assertion
 of power in collective bargaining and through the use of the strike weapon. (Rakner, 1992:78-79)
 If the government's plans in establishing the ZCTU were ultimately unfulfilled, the reorganisation
 nonetheless had important unintended consequences for the labour movement itself. The traditional
 fragmentation of Zambian unions was overcome as unions competing in the same sector were
 merged into single affiliates. Furthermore, the ZCTU gave assistance to union organising,
 especially in the public service, which soon became a mainstay of the labour movement as the
 number of public service workers grew rapidly.
 In the early 1970s the government embarked on a second effort to bring unions into a productionist
 regime. Two factors account for the renewed attempt: moves towards a single party state,
 prompted by growing political opposition towards UNIP; and a new emphasis on import
 substitution and economic independence, towards which workers' continued "indiscipline" posed a
 serious challenge. The two combined to persuade the government to assert new control over
 political and civil society.
 UNIP's response to intensifying political competition in the late 1960s was the creation of a single
 party state in 1972, formally subordinating both the state administration and civil society formations
 to the party. The ZCTU and its affiliates were generally supportive of the single party state, except
 where it applied to them. The commission drafting the new constitution envisioned the labour
 movement as but one mass formation of the party, along with the women's and youth leagues. But
 the federation fought to maintain a separate and unique status in which it never affiliated to the
 party, but maintained seats on the party's highest decision making bodies. Through this formulation
 the unions hoped to maintain a critical alliance with the party. (Rakner, 1992:90) But the
 relationship between the party and the federation remained perpetually ambiguous.
The second factor prompting the new efforts to subordinate the unions to the dictates of
 productionism was UNIP's new commitment to economic independence. Zambia's political
 independence was perennially compromised by its economic and infrastructural linkages to the
 settler regimes in Rhodesia, Mozambique, Angola, Namibia, and South Africa. Economic
 independence was a necessity in a country dependent on its political enemies for skilled labour,
 manufacturing inputs, coal, petrol, and its main source of electricity. It was in response to this
 situation that Zambia embarked on an aggressive import substitution drive from the late 1960s.
 Nationalisation was a crucial component in the strategy, with the copper mines - nationalised in
 1969-1970 - representing the largest prize.
 Economic independence from the south and the diversification of the economy away from primary
 exports were explicit aims of the policy, and the unions were expected to play a crucial "patriotic"
 role in this regard. This updated version of productionism was distinctly less voluntarist than the
 first, and relied on important changes in the labour law to redefine the role of trade unions as actors
 within a state corporatist framework. The Industrial Relations Act (1971) was deeply contradictory,
 however.
 First, it undermined unions by introducing employee- rather than union-based works councils on the
 West German model in every workplace of 100 or more workers. Works councils enjoyed wide
 co-determining power which tended, however, to cut across the unions' traditional collective
 bargaining prerogatives and representational rights. For this reason they were viewed with
 skepticism by unions and employers alike, who tended to ignore them.
 At the same time, the new Act established political committees, an institution borrowed from the
 east block, and specifically Yugoslavian worker self-management. These committees ran parallel to
 works councils, but with a different aim: "to mobilise workers as party members, instill in them the
 spirit of patriotism, create awareness of national economic interests and guard against economic
 sabotage....[they were] intended to bring to workers a sense of discipline to improve productivity."
 (Rakner, 1992:84) As such they were immediately distrusted by workers.
 Finally, the Act sought to curtail workers' collective action through the extension of essential
 services determinations as a means of outlawing strikes. It also established a cumbersome
 arbitration procedure which made a legal strike practically impossible. Taken together these
 changes created more confusion than control. A works council, a political committee, and a shop
 stewards council might all be operating within the same factory, blurring lines of authority and
 responsibility.
 On the other hand, the Act brought about a dramatic strengthening of the labour movement. Three
 provisions - the principle of "one industry, one union", the mandatory affiliation of all unions to the
 ZCTU as the sole legitimate federation, and automatic and compulsory dues check off -
 significantly increased the power of the federation and its affiliates. Furthermore, the expansion of
 domestic industry and the growth in the state sector - both consequences of the development
 strategy - boosted employment levels and union membership, further increasing the resources
 flowing into the federation and its affiliates.
 UNIP, in attempting to create a labour federation strong enough to discipline workers in pursuit of
 its developmental goals, unintentionally assisted the development of a movement that could act
 independently of the state.10 Rather than increasing state control over labour, the creation of the
 single party state and the expansion of nationalised industry (with the state as employer) created
 new grounds for antagonism between UNIP and the working class, while changes in the labour law
 ironically created a more coherent and better resourced organisation that would in time give voice
8
 and force to workers' grievances. Soon a new leadership, with little direct experience of the union-
 party alliance during the independence struggle and not beholden to UNIP would take over the
 federation, symbolised by the election in 1974 of 30-year-old Frederick Chiluba as Chairman
 General.
 Workers' economic and political grievances grew in intensity by the end of the 1970s. The direct
 cause of discontent was the downturn in the Zambian economy caused by the unfortunate
 intersection of the world-wide fall in copper prices, the oil price shocks and rising interest rates.
 Zambia entered into a period of chronic balance of payments problems and - in the 1980s - a debt
 crisis, and into a series of structural adjustment programmes with the IMF and World Bank. But
 the proximate cause of the ZCTU's break from its ambiguous association with UNIP was not
 fundamentally economic. Rather it came about as a result of the labour movement's discontent with
 UNIP's extension of the single party state to local government.
 In 1980 the government announced a restructuring of local government - formerly composed of
 relatively autonomous local administrations - under the direct control of the party. The ZCTU
 strenuously opposed local government reform. When the reform bill was passed by Parliament the
 federation declared it would refuse to comply with the new structures, signalling the first act of
 disobedience towards UNIP.11
 The party responded rapidly to this breach of faith, suspending and later expelling 17 ZCTU and
 mineworkers' leaders from UNIP. Rather than accepting the federation's exercise of autonomy and
 criticism, the party leadership, according to Akwetey, "saw the ZCTU position as a challenge to its
 authority" and "began to perceive the ZCTU leaders as enemies of the party and issued threats to
 crush them. (Akwetey, 1994:59)
 These repressive responses provoked a wave of wild cat strikes on the copperbelt, where 50,000
 miners protested their union leaders' victimisation. The strikes soon spread across the country as
 postal and bank workers showed solidarity with the miners. Eventually the unions backed down:
 the ZCTU agreed to comply with the new local government act and their leaders were reinstated in
 the party. But a breach had opened in relations between the federation and the party, and before
 this wound could heal a new conflict emerged which would make the breach irreparable.
 Within months of the resolution of the local government crisis, a shortage of mielie meal and
 demands for equal pay with expatriate miners provoked a new round of wild cat strikes on the
 copperbelt. At one of these conflicts a minister's car was stoned, and UNIP arrested 7 MUZ and
 ZCTU leaders, including Chiluba and the federation's Secretary General, Newstead Zimba. They
 were detained for four months.
 The struggles over local government reform were the first serious conflict between the labour
 movement and the ruling party, though they were not the last. According to Rakner, the conflict
 demonstrated that the single party state could be challenged, but not from a position of alliance with
 the party: relations were never again the same between UNIP and the ZCTU. Moreover, the
 conflicts cemented the unity of the labour movement, and committed the MUZ to political struggle
 for the first time. But such political engagement would take place under the leadership of the
 ZCTU, now established as the policy and organisational centre of the labour movement. In less
 than a decade, the ZCTU was transformed from a crucial component in UNIP's state corporatist
 vision into the mainspring of opposition to the single party state.
 Opposition to UNIP intensified in the mid-1980s as a consequence of deepening economic crisis and
 growing official corruption. Both grew from the centralised command economy whose political
expression was the single party state. Efforts to diversify the economy ultimately yielded a highly
 indebted nationalised sector, which was nonetheless unable to break the legacy of dependence.
 Rather the system provided rapid enrichment for a narrow stratum of party officials while
 encouraging widespread cronyism and inefficiency. Both economy and polity were fused under the
 system of presidentialism, in which the state corporations served as a massive source of patronage
 in service of President Kaunda's political objectives. The combination of chronic economic crisis
 and growing political authoritarianism eroded whatever legitimacy UNIP continued to enjoy.
 (Baylies and Szeftel, 1992)
 As a result of the chronic debt crisis, Zambia was forced to implement seven different - and
 increasingly severe - stabilisation and adjustment programs between 1975 and 1986. Misplaced
 though they may have been as economic reform measures - by 1985 Zambia was the most indebted
 country in the world relative to its GDP - the IMF's efforts to roll back the state catalysed the
 growing opposition to the system, (van de Walle and Chiwele, 1994:9) With the support of the
 IMF and World Bank the government launched the Radical Reform Program, which envisaged a
 thorough liberalisation of the currency and interest rates, fiscal and monetary stabilisation, the
 reduction of consumer subsidies, and the privatisation of maize marketing, (van de Walle and
 Chiwele, 1994:10; Callaghy, 1990) The elimination of maize subsidies - an effort to reform the
 "urban bias" in development policies - prompted the most militant popular uprising since
 independence as thousands of people rioted on the Copperbelt in late 1986. (Riley and Parfitt, 1995;
 Rakner, 1992, Akwetey, 1994)
 The riots were repressed, but the government reinstated the maize subsidies. However the uprising
 was followed in early 1987 by a three-month rolling strike initiated by the ZCTU's Zambian
 National Unit of Teachers [ZNUT] who were soon followed by junior doctors and nurses and then
 by agricultural workers. By the end of April the country was at a stand-still as group after group of
 union members struck against austerity measures - including subsidy withdrawal and public sector
 wage freezes and retrenchments - which threatened their security as both workers and consumers.
 (Rakner, 1982:116) Kaunda chose the annual May Day celebrations to announce a break with the
 IMF and the government's intention to "go it alone" under the auspices of a New Economic
 Recovery Program [NERP].
 The ZCTU applauded the government's abandonment of the IMF-inspired liberalisation and urged a
 return to the statist system, notwithstanding that these policies had generated the crisis that
 prompted the external intervention in the first place. The NERP soon proved unworkable, and in
 1989 the government restored its links to the IMF, though on more stringent terms than before,
 with renewed demands to liberalise the economy. As part of liberalisation the government
 undertook the first major reform of labour legislation, aimed at removing the provisions of the 1971
 Industrial Relations Act that had strengthened the ZCTU. The reforms encouraged fragmentation
 and decentralisation by revoking the mandatory affiliation to the ZCTU and the "one-industry, one-
 union" provision, and requiring a two-thirds majority of the membership to apply funds in political
 campaigning. (Rakner, 1992:121).
 C. Zambia's Second Independence: Labour and Democratisation under Neoliberalism
 The shift in economic policy and the government's renewed efforts to rein in the unions precipitated
 the unions' final break with UNIP. The ZCTU's efforts to establish its autonomy and to gain
 influence over policy making had come to naught, and it now sought an alliance with other groups
 in civil society. The ZCTU's General Council had resolved in December 1989 to work towards the
 immediate reintroduction of multi-party politics, a demand that the federation repeated at a special
 UNIP national convention in March 1990 called to discuss democratisation. The UNIP leadership
 and Kaunda in particular opposed these demands. But mounting internal and external pressure -
10
 and a renewed outbreak of rioting in June caused by an increase in maize prices - forced the
 government to change tack. In July it lifted the ban on opposition groups' right to organise political
 parties, and within days the Movement for Multiparty Democracy [MMD] was launched, with
 Chiluba as Vice-Chairman. (Rakner, 1992:59-63)
 Over the next year MMD and other organisations - including foreign donors and human rights
 groups - fought with UNIP over the procedures for multiparty elections, which were eventually
 held in October 1991. The elections were a stunning rebuke for UNIP, as the MMD swept into
 office winning 125 of the 150 parliamentary seats. Chiluba, who had become MMD Chairman
 earlier in the year, won the presidency, receiving 76% of the vote. (Baylies and Szeftel, 1992:76;
 Bratton, 1992)
 With the relaunch of multiparty politics the ZCTU was the only organisation with a national
 structure and material resources to fight an election campaign. These considerable assets were put
 at the disposal of the MMD: both the federation and its affiliates mobilised their district committees
 in support of MMD candidates, and supplied a major share of the party's financial resources.
 (Akwetey, 1994:65) The commitment to the MMD arose out of the federation's original policy of
 support for the most progressive party, originally established by the then-RTUC in 1960 when it
 chose to support UNIP over the ANC. In 1990, however, this policy was invoked to justify
 opposition to the ruling party and support for the opposition. The ZCTU did not hope to become a
 labour party, nor to substitute for a political party, but chose to work for the reinstitution of
 democracy with whatever forces shared a commitment to multiparty politics.
 However, if the ZCTU provided the material basis for the MMD, its influence did not extend to the
 party's economic thinking. The MMD was not a labour party; rather it was a broad alliance of
 businessmen, professionals, and workers. These groupings shared little other than a commitment to
 multipartism and opposition to UNIP. Its leadership - and later its parliamentary caucus - was very
 largely composed of former cabinet ministers and parliamentarians who had fallen out with Kaunda
 and UNIP. These layers had come to oppose Kaunda's "socialism" and were strong advocates of
 economic liberalisation and engagement with the IMF. Where the ZCTU had also come to oppose
 the command economy, it however was staunchly opposed to economic liberalisation in general and
 structural adjustment in particular. But these positions were buried in the MMD, and the union
 federation reversed its long-held policy to adopt the MMD's stand. (Baylies and Szeftel, 1992) The
 ZCTU justified this about-turn in the hope that a democratically elected government would at least
 implement adjustment fairly.
 Within a year of taking power, for example, the MMD government had in fact enacted many of the
 liberalisation policies Kaunda had attempted and failed to implement after the 1989 rapprochement
 with the IMF: the MMD removed all maize subsidies - the issue that had perennially plunged the
 Kaunda regime into crisis - and moved equally rapidly to introduce market-based interest rates and
 a liberalised exchange rate regime. By 1994 it had moved to a "cash budget" to stabilize
 government expenditure, and began to undertake broader structural reforms, including the
 dismantling of marketing boards, privatisation, and the downsizing of the public service, (van de
 Walle and Chiwele, 1994:27-41)
 Despite these reforms, the new government soon began operating in a manner reminiscent of the
 party it had replaced. The presidentialist system had itself not been dismantled; indeed, Chiluba
 quickly reasserted his predecessor's habit of appointing cabinet members and senior civil servants
 according to the exigencies of ethnic and regional balancing rather than on merit. (Baylies and
 Szeftel, 1992:89-90) Nor did the multiparty elections lead to the hegemony of parliamentary
 politics: the presidency and the cabinet (particularly the core economic ministries) remained the
11
 centres of power, while the MMD as an organisation was reduced to an electoral shell with little
 influence on policy (van de Walle and Chiwele, 1994:20-21; 18).
 A number of observers have noted considerable continuity between the practices of the new
 government and UNIP, evident in the untransparent reshuffling of ministerial portfolios, the
 manipulation of the constitution and media for political advantage, and intimidation of political
 opponents. (Panter-Brick, 1994; Mphaisha, 1996) These practices came under the spotlight during
 the 1996 elections, in which the MMD was returned to office with an increased parliamentary
 majority. But some Zambian NGOs and most bilateral and multilateral aid organisations castigated
 the government's actions, and in the run-up to the elections foreign donors suspended program aid -
 including balance of payments support.12 (Human Rights Watch, 1996; Adams and Bevan, 1996)
 The ZCTU has criticised these undemocratic practices, but has remained relatively powerless to
 slow down, let alone reverse the government's economic liberalisation policies, nor to curb its more
 blatant political abuses. After the elections the federation found itself caught between its historic
 antagonism to liberalisation (indeed, its advocacy in the aftermath of the 1986 riots, of a return to
 commandist economic policies) and its support for the MMD's commitment to liberalisation. It had
 no feasible alternative to the new government's ambitious liberalisation program, save for criticism
 of the pace of implementation and concern for its social effects. Furthermore, the ZCTU had no
 clear policy for engaging government, beyond the general commitment to support the most
 progressive party. Nor did it possess a program for reconstructing the Zambian state, beyond its
 support for multiparty politics.
 The federation had some hope that it would at least meet on a regular basis with ex-labour leaders
 now in government, but such meetings were held infrequently, and when they did occur, the
 unionists' advice was ignored or the meeting became an opportunity to inform labour of
 government policies. (Interview, Chirwa, 1996)
 The federation was able to pressure for revision of the Industrial Relations Act by removing many
 of the anti-labour clauses enacted by UNIP in 1991, though it was unable to restore mandatory
 affiliation to the ZCTU or mandatory dues check off. One important addition to the: Act was the
 creation of a formal purchase on decision making through a new Tripartite Consultative Labour
 Council [TCLC] with power to "advise the government on all issues relating to labour matters,
 manpower development and utilisation..." (Akwetey, 1996:13)
 But the TCLC has not performed the role envisioned in the Act. It has no permanent secretariat,
 nor does it have any statutory right to review legislation. Its role is merely advisory, and its
 decisions are not binding. Furthermore, it meets at the convenience of the Minister of Labour, who
 for the first years of the first MMD government was Newstead Zimba - the former ZCTU Secretary
 General - who was identified by current ZCTU leaders as the most anti-labour member of Chiluba's
 cabinet. In 1996 the TCLC was convened only once. A top official of the Ministry of Finance
 referred to it as "that consultative body where once a year government informs labour of decisions
 that have already been taken." (Gostner, 1997:59-60; Confidential interview, Ministry of Finance)
 Privatisation has been a particularly difficult issue for the federation. In 1992 Parliament created
 the Zambian Privatisation Agency [ZPA], an independent parastatal, to conduct the sale of state
 assets according to market-based criteria. While it endorsed the principle of dismantling the
 command economy, the ZCTU been adversely affected by the policy. More than 150 firms have
 been privatised since 1992 leading to considerable job loss and heightened anxiety for those still in
 employment. (Ginsburg, 1997:61; Akwetey, 1996:18)13 The ZCTU, however, has advocated a
 more systematic approach to privatisation urging that profitable and strategic parastatals not be
12
 privatised, that the state retain a share holding in some privatised companies, and that the costs not
 be borne by the poor exclusively. But these positions have not been formulated into an alternative
 policy, nor has the ZCTU used what influence it possesses to challenge the government. Though it
 holds a seat on the ZPA board, it has not used its presence to contest the general policy or the terms
 of specific sales. (Confidential interview, ZPA, 1996)
 Notwithstanding this weakness at national level, some ZCTU affiliates have been able to use their
 shop-floor power to challenge these policies, or to fight for benefits once a sale has been concluded.
 This exercise of power has led to a number of militant confrontations between union members and
 managers. In May, 1996 workers at the newly privatised Dairy Produce Board protested against
 the new South African owners' decision to retrench without packages 200 of the company's 240
 workers. In March 1996, 200 members of the ZCTU's National Union of Plantation and
 Agricultural Workers nearly lynched three managers of Hybrid Poultry Farm for their bad attitude
 towards workers, while in October the Managing Director of the Medical Stores narrowly escaped
 a similar fate during a heated meeting on the future of the business. (Buhlungu and van der Walt,
 1997: 67-68) However, in the absence of a challenge to the national policy or a regular institutional
 presence in the privatisations themselves, these assertions of militance have been essentially reactive
 and unfocussed.
 The most important protest against liberalisation was waged by workers in the ZCTU's public
 service affiliates, the Civil Servants Union of Zambia [CSUZ] and the National Union of Public
 Service Workers [NUPSW]. In early 1996 these unions organised a week-long national strike to
 demand implementation of a 45% wage increase granted by the Industrial Court. (Buhlungu and
 van der Walt, 1997:67) The year before the same unions had taken steps to launch a legal national
 strike over the packages offered to workers facing privatisation or rationalisation. (Interviews with
 CSUZ and NUPSW executive members, 1997) Both actions helped delay government plans to
 downsize the public sector and to stabilise recurrent expenditure.14 Again, lacking both a coherent
 alternative policy and an institutional framework in which to bargain, labour can at best address
 some of the consequences of decisions, but only after they have been taken by government.
 As a result of its participation in the MMD the ZCTU helped bring about the first change in
 government in Zambia since 1964. But the federation did not gain either a formal or informal
 purchase on decision-making. As under UNIP, policy is still tightly held within the ambit of the
 presidency, the cabinet, and foreign donors. (Akwetey, 1996:19) This is especially true with
 respect to economic policy, which is perhaps even more concentrated in elite hands than was the
 case under the previous government, (van de Walle and Chiwele, 17-20)
 This result is an ironic outcome of democratisation in Zambia: labour's electoral ally has been able
 to implement programs its predecessor was powerless to pursue given labour's ability to block
 policy while in opposition. Van de Walle and Chiwele explain this irony in terms of Chiluba's
 overwhelming electoral mandate, which not only legitimises his rule, but has freed him from
 dependence on the ZCTU. This freedom was underscored by the MMD's continued electoral
 dominance in 1996, notwithstanding the ZCTU's neutrality in the election. Furthermore, the
 ZCTU's membership is concentrated in the major urban areas - the Copperbelt, Lusaka and
 Livingstone - such that most parliamentarians "have no reason to heed the power of organized labor
 in their district." Given the absence of any credible political opposition, the ZCTU cannot threaten
 defection from the government to gain leverage with the MMD. Moreover, though Chiluba himself
 may be more favourably inclined towards labour, this is not true of the MMD as a whole, nor of his
 cabinet ministers, who are drawn disproportionately from among liberals in the party. Given the
 fact that he knows the labour movement's strengths and weaknesses - and how far and in what ways
 he can push it - he has more space to pursue economic reform than Kaunda ever enjoyed. (1996:22)
13
 The parliamentary political terrain has thus become increasingly unfavourable for the ZCTU, a fact
 reflected in the confusion over the federation's political policy. The ZCTU claims no longer to be
 in an alliance with the MMD. Though a number of its members - especially on the copperbelt -
 worked actively for MMD candidates in 1996, the federation itself remained neutral, choosing to
 defend the democratic process itself rather than fight for its erstwhile ally. However in interviews
 in late 1996 individual leaders showed considerable confusion over the federation's political policy.
 Some expressed support for the policy of neutrality, while others claimed that the federation was
 still in an alliance. Others asserted that the alliance was focused only on the need to reinstate
 multiparty democracy, and - perhaps reflecting on past difficulties with UNIP - that no trade union
 federation could afford to be allied with a ruling party. From this perspective - which appeared to
 be dominant - labour should be involved in "industrial politics" in which the primary objective is to
 defend workers' economic interests. (Gostner, 1997:59)
 But aside from these political factors, the ZCTU's inability to influence policy must be understood
 against the background of the impact on the federation of economic liberalisation by taking into
 account the country's insertion into the international economy. In short, adjustment has not been
 kind to Zambian workers nor to their unions.
 In 1990 the ZCTU's membership stood at 353,445 in 19 affiliates.(See Table 1) However, between
 1991 and 1994 nearly 40,000 workers have been retrenched in Zambia, many as a result of
 privatisation.
 1 AM fr, 1. ZAMR1AN COM.KLHS Ol" \ H \ ? t t MO-NS,
 M'TH I M I S V N D M I MKt'KMIII1
 AFFILIATE AND SECTOR
 Airways and Allied Workers (AAWU)
 Civil Servants (CSUZ)
 Guards (GUZ)
 Hotel & Catering Workers (HCWU)
 Mirteworkers{MUZ) ':-"^:^^^^?.':r" '"^-""F^^^
 Building, Engineering ahd^Giiefai^orkers (N&B$3^:Y; : ; :
 Commerce & M M a 1 ; ^ ^
 Communication Workers (NUCW)
 Plantation & Agricultural Workers (NUPAW)
 Public Service Workers (NUPSW)
 Transport & Allied Workers (NUTAW)
 Railway Workers (RWUZ)
 Typographical Workers (ZTWU)
 University of Zambia & Allied Workers (UNZAAWU)
 Electricity and Allied Workers (ZEAWU)
 Health & Allied Workers (ZNUHAW)
 Financial 1 nstitutions; &M &l:pi|:W?r|c$r$. (tWi WiS^l-
 United Local Authorities Workers (ZULAWU)
 Total Membership
 Losses due to disaffiliations
 Losses due to retrenchment
 ZCTU Membership 1994
 1990
 8,558
 28,000
 13,000
 7,000
 6,319
 16,674
 65,000
 7,392
 8,858
 1,219
 2,794
 3,850
 1,219
 25,000
 353,445
 1994
 3,010
 29,000
 8,000
 9,693
 6,494
 18,332
 35,000
 7,896
 8,501
 2,500
 5,074
 3,600
 3,030
 !!pj!i;i;:s;:v: ?.
 22,000
 304,614
 142,484
 38,346
 162,130
 (Shading indicates breakaway unions). Source: Akwetey (1996:18)
14
 By 1994 the federation's membership had dropped to 304,614 indicating that a sizeable
 proportion of the loss of membership occurred because workers had simply "ceased to be
 members of the unions," a choice facilitated by the revocation of the mandatory
 membership clauses in the 1990 and 1993 Industrial Relations Act amendments. (Akwetey,
 1996:18)
 The most troubling impact of liberalisation has been the defection from the ZCTU of five of
 its largest affiliates, the first major split in the Zambian labour movement since 1959. The
 breakaway unions included the miners, teachers, and building and engineering workers
 (Chiluba's old affiliate). Together these defections meant the loss of nearly 50% of the
 ZCTU's remaining membership, and also deprived it of its leverage in crucial sectors of the
 economy.
 In part the split was caused by personality differences, as new leadership was elected to
 replace senior officials who left for government. (ZCTU, 1996) But at a deeper level, these
 federations were divided over their responses to government policy, particularly
 privatisation. While most civil service affiliates (such as the CSUZ and NUPSW) see these
 policies as inherently threatening, others, particularly those organised in the emerging
 private sector, see considerable opportunities deriving from privatisation. The MUZ, in
 particular, views privatisation as a boon, in that it promises increased foreign investment in
 existing operations, more prospecting and the development of new mines, and - in the
 medium term - more jobs.15 These affiliates' lack of concern for workers and affiliates in
 sectors more vulnerable to privatisation has led to accusations of sectoral opportunism.
 (Ginsburg, 1997:62)
 In 1995 the breakaway affiliates attempted to launch a rival federation to the ZCTU, known
 as the Confederation of Free Trade Unions. But this initiative has thus far foundered, as
 the Industrial Relations Act presently allows for recognition of only one national trade
 union centre. The unions then established the Federation of Free Trade Unions, but the
 FFTU is prevented from operating as a trade union. The ZCTU has sought to delay
 amendment of the labour law to allow for multiple union centres, though this is clearly a
 rear-guard action as the government has adopted the ILO conventions on freedom of
 association which reject monopoly unionism.
 The split has been extremely costly for the labour movement as a whole. The ZCTU has
 been deprived of the source of close to half its income from dues, making it more
 dependent than ever on foreign trade union donations. According to Akwetey, this has
 forced massive cutbacks in its educational and service functions to the affiliates. The loss
 of revenue thus threatens to place the unions in a vicious circle: retrenchments cannot be
 fought because the federation and its affiliates lack the resources to contest them or to
 develop alternative policy, while more retrenchments deplete resources further.
 Furthermore, the unions have been engaged in a protracted guerrilla war with the
 breakaway affiliates. Both sides have launched splinter unions to contest the other's hold
 over members, further draining resources while misdirecting attention away from their
 common adversaries. The government's removal of the requirement for mandatory
 affiliation to single industrial union - in other words, industrial pluralism - set up this
 rivalry, but the unions' inability to maintain solidarity and think strategically has been the
 direct cause of these troubles.
 The fight between unions underscores the most important aspect of liberalisation in Zambia:
 that pluralism and competition does not stop short of the main advocates of democracy.
 The unions were the earliest and most powerful supporters of multiparty democracy in the
15
 political sphere, but they could not resist the application of these same principles in the
 industrial relations sphere. UNIP was unable to make the transition from monopoly to
 competitive party politics, and now faces extinction as a political force. It remains to be
 seen whether the ZCTU will go the same way as the political system and the party to which
 it owes its birth and much of its strength.
 At various points since independence the ZCTU was able to combine mobilisation and
 interventions in policy to challenge the UNIP government. The unions used this
 combination of militance and institutional work to carve an autonomous space for
 themselves within the single party state. They were able to construct a similar strategy to
 contest local government reform and structural adjustment. The most important example of
 the ZCTU's strategic use of power was in their protracted efforts to remove the single party
 state and ensure the move towards multiparty politics. However, since 1991 the unions
 have not shown the same capability to contest the consolidation of democracy. They lacked
 clear alternative economic policies, nor did they have any serious strategy for engaging
 with or transforming the state once multi-party democracy had been reinstituted. In both
 the political and economic spheres the unions have been marginalised from policy making,
 which is increasingly insulated in a narrow sphere comprising the presidency, economic
 ministries of the cabinet, and multilateral organisations.
 Structural adjustment has in turn undermined the unions organisationally, contributing to
 the first splits in the movement since the colonial period, while it has generated
 considerable mass discontent and protest among the rank and file. However such militant
 action has been of a largely wild-cat and sporadic nature, and has been easily contained
 within the industrial relations system. Moreover, with the exception of strikes and
 threatened strikes by workers in the public service, these actions have been disarticulated
 from any sustained institutional interventions into policy making. In the absence of a
 broader strategy of challenging government policy and its articulation with militant action,
 labour is in danger of losing all control over the process of transformation which it led in
 the 1980s.
 III. SOUTH AFRICA
 A. Unions Under Apartheid: Building Union Autonomy under State-Led
 Industrialisation
 1. Organisational Independence
 While labour movements in much of the world have experienced declines in membership and
 influence during the 1980s and 1990s the South African labour movement has grown rapidly.
 The membership of trade unions grew dramatically from more than 700,000 in 1979 to
 nearly 3,000,000 in 1993. This involved a growth in union density from slightly more than
 15% to 58% over the same period. (Macun, in press)
 There are 213 registered trade unions, many of which are small craft unions or professional
 associations that are declining in membership.16 (Department of Labour, 1995) However the
 broad trend is towards nation-wide industrial or sectoral unions in the core of the economy,
 including mining, metal, textile and clothing, retail and commercial, chemical, food, and the
 public service. Thus, although there are many unions in the country, the majority of union
 members are concentrated in the large national industrial unions in these sectors.
 Many of the unions are affiliated to one of six federations, [see Table 2] On the face of it this
 suggests a degree of fragmentation in the labour movement, though the largest federation,
16
 COSATU has more members than all the other federations combined. However, the labour
 movement in South Africa is deeply divided on grounds of race and political orientation.
 Historically South Africa had a dualistic industrial relations system, as had other settler
 colonies in Africa. Trade unions with non-African membership were recognised under the
 Industrial and Conciliation Act of 1924, while Africans were excluded from the formal
 industrial relations system and subject to tight control through the pass system. The dualistic
 system ended in 1979 when the Act was amended to cover all employees except migrant
 workers, farm workers, domestic servants and employees in the public sector.17 The granting
 of formal trade unions rights to African employees legitimised the then-emerging black trade
 union movement, and membership figures skyrocketed.
 1 ABLE. 2: SOU 1 H Af- Klf ' tN UNION FKOF..HA1 !O>S.
 M : M B ? , U O I - A I T I I I V I 1 S , tNDMLMM.ftSmiM'W-l
 FEDERATION
 Congress of South African Trade Unions
 (COSATU)
 National Council of Trade Unions (NACTU)
 Federation of South African Labour (FEDSAL)
 Federation of Independent Trade Unions (FITU)
 South African Confederation of Labour (SACOL)
 United Workers Union of South Africa**
 (UWUSA)
 AFFILIATES
 15
 18
 16
 24
 4
 ?
 MEMBERSHIP*
 1,317,496(45%)
 334,733(12%)
 257,258(10%)
 236,000 (8%)
 54,290 (2%)
 ?
 The percentage in the membership column reflects the federation's membership as a proportion of
 total membership. Estimates of UWUSA membership fluctuate between 30,000 and 100,000, but
 reliable figures are not available. Source: Macun (forthcoming).
 COSATU and NACTU are essentially industrial unions of African semi-skilled workers,
 although COSATU is firmly committed to organising workers regardless of race, and has
 always had a significant number of coloured and Indian members. In recent years it has
 made some inroads among white workers as well. Indeed, its leadership has always included
 non-Africans in key positions.
 These demographic differences express themselves in four major political traditions within
 the democratic movement, which structure the differing perspectives on labour's relationship
 to politics: the Black Consciousness/Africanist tradition, white racist unionism, the national
 democratic tradition, AND the shopfloor tradition. (Fine and Webster,. 1989) NACTU's
 origins lie in the Black Consciousness political tradition, and the federation has more
 recently aligned with the Africanist tradition articulated by the Pan Africanist Congress,
 which broke from the African National Congress [ANC] in 1959 because of the latter's non-
 racialism.
 The other major federations, FEDSAL, FITU, and SACOL have their origins in the old
 white-dominated labour movement, and are themselves divided in terms of skill and
 ethnicity. FEDSAL represents white-collar workers, and recently has taken on larger number
 of black members; it recently merged with a number of independent public service unions to
 form the Federation of Unions of South Africa [FEDUSA]. FITU and SACOL, on the other
 hand, represent skilled and semi-skilled white workers respectively, while SACOL is
 essentially a movement of Afrikaans-speaking workers. FEDSAL and FITU are both
 politically unaffiliated, though they have been historically hostile to the democratic
17
 movement, while SACOL has been linked with the far-right wing Conservative Party.
 UWUSA was launched in 1986 with covert funding by the apartheid state. Its membership is
 almost entirely confined to Zulu-speaking semi-skilled and unskilled workers affiliated to the
 fnkatha Freedom Party led by Chief Mangosuthu Buthelezi. It has virtually no organised
 presence in the workplace, but it has often played a divisive role acting as scab labour
 undermining efforts by COSATU orNACTU. (Harvey, 1996)
 The national democratic tradition has the deepest roots in the labour movement, originating
 in the creation in 1955 of the South African Congress of Trade Unions [SACTU], which
 combined most of the then-existing unions organising among black workers. SACTU's
 alliance with the ANC involved a redefinition of its trade union role along the lines of
 political unionism. Lacking a strong power base on the shopfloor and faced with an
 increasingly hostile state and intransigent employers, SACTU mobilised the oppressed -
 across class lines - around the demands of the ANC's popular program, the Freedom Charter,
 adopted in 1955.
 SACTU's involvement in the Congress Alliance was premised on the assumption that South
 Africa could not be understood in class terms alone, but that political change necessitated a
 national democratic struggle to liberate South Africa from white rule that was seen as a form
 of "colonialism of a special type." To this end, the Congress Alliance was a multi-class
 movement of organisations representing the broad oppressed groups in South Africa under
 the leadership of the ANC and aiming to establish a national democracy. In form and
 content, the Alliance - and particularly the relations between unions and the nationalist party
 - parallelled other national liberation movements on the African continent. SACTU's
 political role brought it into direct conflict with the state and the organisation felt the full
 force of repression in the 1960s. By 1964 it had ceased public activities in South Africa and
 operated in exile until the unbanning of the ANC in 1990.
 When union activity reemerged before and after the Durban strikes of 1973, SACTU and the
 other formations of the Congress Alliance were in exile and their internal operatives were
 deep underground. The Durban unions ultimately merged in 1974 to form the Trade Union
 Advisory Coordinating Council [TUACC], the first progressive trade union centre since the
 repression of the early 1960s. In 1979 TUACC joined with other emerging unions from the
 Eastern and Western Cape to form the Federation of South African Trade Unions [FOSATU].
 (Friedman, 1987) In fact, the legal proscription of the nationalist movements meant that in
 their formative years these embryonic unions were able to develop leadership, organise their
 constituency, and define their strategies and tactics relatively independently from the
 ideological orientations and models of the ANC, the South African Communist Party
 [SACP], and SACTU. The space created by virtue of banning and exile meant that the new
 unions could develop innovative approaches to organising that differed from the populist
 strategies and tactics of the nationalist-linked unions of the 1950s.
 These emerging unions developed what has come to be known as the shopfloor tradition, best
 captured in a prominent 1982 speech by the FOSATU General Secretary, who proposed the
 development of a workers' political movement under worker control as an alternative to the
 national democratic tradition. (Foster, 1982) FOSATU's avoidance of political engagement
 was informed by an assessment of SACTU's errors, that its close identification with the
 Congress Alliance and its political campaigns caused its demise: political engagement, in this
 view, led to a neglect of workplace organising while inviting repression.
18
 These unions emphasised the development of factory-level leadership democratically elected
 by rank-and-file workers and directly accountable to them. (Pityana and Orkin, 1992, Baskin,
 1991) This particular strategic innovation came about in the early 1970s when the then-
 emerging black unions in Durban sought to avoid the growth of top-heavy office-bound
 leaderships. They were inspired by a programmatic commitment to worker control and to the
 need to survive state repression, which tended to be directed in the first instance at highly
 exposed national leaders. Furthermore, these unions were committed to building
 industrially-based structures, concentrating on shopfloor issues while remaining - for reasons
 of survival - unaffiliated to the exiled liberation movements.
 Through the early 1980s black workers were experiencing the effects of a deepening social
 and political crisis, which can now be seen as the fundamental contradictions of the apartheid
 order. The crisis spawned a range of new social movements, including student, youth, civics,
 and community-based unions which contributed to the growing opposition in the townships,
 especially from 1984. The community unions' increasing success forced FOSATU to
 reevaluate its strategy of political nonalignment and its efforts to isolate a class-based politics
 from the broader national democratic struggle, a position that threatened to marginalise
 FOSATU from its own base.
 The decisive break with political abstentionism was in November 1984 when FOSATU
 entered into joint action with student and civic organisations to participate in the first
 successful worker stay away since the Soweto uprising of 1976. (Labour Monitoring Group,
 1985) FOSATU's shift brought the shopfloor and national democratic traditions closer
 together, and facilitated efforts to promote unity among the emergent unions. This unity was
 finally achieved by the creation of COSATU, which combined the industrial union emphasis
 of the shopfloor tradition with the political orientation - alliance with the ANC and SACP -
 central to the national democratic tradition. In one of their first acts, the new office bearers
 of COSATU travelled to Lusaka and endorsed the ANC's leadership of the liberation
 struggle.
 COSATU's approach to the ANC marked the informal beginning of what is now known as
 the Triple Alliance between the ANC, SACP, and COSATU. It is significant that in this
 movement towards unity with the national democratic tradition COSATU entered the
 alliance not as a subordinate partner as was SACTU - lending its mobilisation power to the
 ANC's campaigns - but as an equal partner with an independent power base, strategy, and
 leadership. In this movement COSATU displaced SACTU.18
 The intensifying internal insurrection in 1985 and 1986 provoked massive repression in June
 1986. With the ANC and SACP still in exile, and internal political organisations under ban
 and leaders in detention or under trial, COSATU emerged as the de facto leader of the
 internal democratic movement. This position gave COSATU considerable influence over the
 course of internal politics. (Baskin, 1991)
 By 1991 COSATU unions alone included more than 25,000 shopstewards, giving an average
 ratio of one steward to 50 workers. The shopsteward system gave workers power in
 production, a capacity to intervene directly in struggles challenging managerial control over
 wages, working conditions, and the organisation of production. Furthermore, as shopstewards
 constituted the pool from which more senior union leaders were drawn, the system enabled
 rank-and-file members to exercise an unprecedented degree of control over their officials.
 Though the unions were initially dependent on foreign funding - largely from the European
19
 social democracies - by the 1990s they were self-financing. (Southall, 1995) With the growth
 in numbers and the shift from a flat-rate to a percentage-based subscription system the unions
 accumulated large sums of money which cushioned them from the withdrawal of foreign
 funders' support after the first democratic elections. More importantly, self-financing meant
 that the unions were not dependent on support from the state or the ANC for their
 livelihood.19 Finally, from 1990 COSATU experienced an increase in membership of nearly
 50% to stand at nearly 2 million in 1996. (Filita, 1997)
 A. 2. Strategic Sophistication: Radical Reform
 The independent unions were distinguished from their predecessors in South Africa not only
 by their size, organisational depth, and independence, but also by their strategic
 sophistication. They successfully combined a radical vision with a reformist approach which
 we have called radical reform. In pursuit of the long-term goals of ending apartheid and
 creating a socialist economy, the unions emphasized legal means of struggle.20 They sought
 inclusion of all workers within the industrial relations system, and decided to register their
 unions under the Labour Relations Act.2' Finally, they eschewed involvement in national
 political issues and refused to align themselves publicly to any political movement.
 Reforms were not regarded as ends in themselves, but rather as dynamic phases in a
 progressive struggle to achieve the long-term goals. With the strong backing of their
 members, shop stewards had the power to push for concessions from management, which not
 only created space for further advances, but also won concrete improvements in workers'
 conditions, thereby reassuring workers of the efficacy of direct action. There were two
 important components to the unions' approach to the strategic use of power: 1) democratic
 processes, to win voluntary consent from members for mobilization, and for restraint when
 necessary; and 2) tactical flexibility which included a capacity to distinguish principles from
 tactics, and to choose those tactics most likely to succeed, including negotiation and
 compromise.
 The negotiation of recognition agreements, which set out the rights and duties of shop
 stewards and trade unions in the workplace, was an important step in establishing "the rule of
 law" on the shopfloor. Trade unions did not win these rights without a fight. During the late
 1970s and early 1980s there were hundreds of strikes in support of demands for recognition.
 Throughout the early 1980s workers and trade unions used their increasing strength and
 rights on the shopfloor to fight for better wages and working conditions: leave, hours of
 work, safety, and pensions became matters for negotiation. But workers were not only
 becoming a formidable force on the shopfloor, they were also beginning to engage in
 centralized industrial bargaining in the Industrial Councils established under the Labour
 Relations Act. These joint employer-union bodies provided a forum where unions could
 engage in industry-wide negotiations, which could potentially be extended to include
 industrial restructuring issues.
 In its emphasis on political independence, tactical flexibility, and compromise with
 employers and the state the strategy stood in marked contrast to the political and military
 struggle then being waged by the ANC and its internal allies. Their strategy aimed at the
 state's overthrow through a national democratic revolution: a sudden shift in the balance of
 power in which the old ruling class was destroyed altogether. This strategy of revolutionary
 rupture stressed abstention from involvement in the apartheid state and all its institutions on
 the assumption that the leadership of any subject group would be coopted and the status quo
 would remain. By contrast, the labour movement developed a strategy of engaging rather
20
 than boycotting the state, from an independent and disciplined power base resting on strong
 factory structures, held together through practices of democratic accountability. The
 differences between these two strategic visions caused innumerable conflicts between the
 emerging unions on the one hand, and the exiled ANC and the developing internal political
 movement on the other. However, the strategy of radical reform led neither to revolutionary
 rupture nor to co-option, but to the legitimization of the union as an institution and the rapid
 growth of unions in the 1980s.
 The emergence of a powerful labour movement committed to radical reform opened up the
 possibility of a distinct role for the trade union movement in the process of transition to a
 new democratic order. Labour played a crucial role both in the initiation of the transition and
 in the transition itself. It was the best organised and the single most powerful constituency in
 the anti-apartheid resistance and through its protest action contributed to the chronic crisis
 that precipitated the transition. In the transition itself labour used its mass base to mobilise
 its constituents in support of its allies' positions in the negotiations and to bolster the ANC's
 campaign during the 1994 elections. But labour contributed more than collective muscle. Its
 economic and social policies supplied the core ideas for the ANC's electoral agenda - the
 Reconstruction and Development Programme [RDP] - and led to the development of
 innovative multipartite institutions that gave labour a direct voice in policy making.
 However, even as COSATU was consolidating its leadership position in the late 1980s within
 the democratic movement, negotiations were developing between the ANC and the apartheid
 state over what was to become a "negotiated revolution." (Sparks, 1994) The beginning of
 the transition to democracy, symbolised by the unbanning of the ANC and SACP in 1990 and
 their return to above-ground politics in South Africa initiated as well the ANC's return to
 dominance within the internal democratic movement.
 B. The Unions and the Transition to Democracy
 B.I. Autonomy in the Alliance
 Immediately following the unbannings, when the de facto alliance between the ANC, SACP
 and COSATU was formalised as the Triple Alliance, COSATU members expected that the
 federation would continue to play the same equal and independent role in alliance with the
 ANC that it had enjoyed over the previous five years. (Pityana and Orkin, 1992:58)
 In 1991 COSATU's Central Executive Committee resolved to attend the Conference for a
 Democratic South Africa [CODESA, the multiparty national negotiating forum] in its own
 right or not at all. (Pityana and Orkin, 1992:61) However, when COSATU applied for
 membership it was refused and it was to be represented indirectly, via its allies, the ANC and
 SACP. (Webster and Keet, 1992) CODESA thus became a parliament-in-formation
 responsible for drafting the interim constitution and deciding on transitional arrangements. It
 was comprised of political parties representing political society: civil society formations -
 including COSATU - did not obtain direct representation.
 This is the momentvwhen the ANC formally established its dominance within the Alliance,
 and COSATU (as well as other allied civil society formations) were reduced to a secondary
 role, influencing ANC policy through lobbying and pressure, rather than wielding a share of
 direct power over decisions. From this point political parties - not the civil society
 organisations which shouldered the 1980s internal resistance to apartheid - were to be at the
 centre of the transition.22
21
 B.2. Economic Policy: New Ideas and Institutions
 Notwithstanding the ANC's leading role in the alliance, COSATU was never subordinated.
 It retained separate leadership, finances, and policy orientations. Moreover, as it claimed a
 right to participate in the political process of transition, so it also sought to intervene in the
 process of restructuring the economy. COSATU's economic interventions had their origins
 in projects launched in the late 1980s to analyse the nature of the economic crisis facing
 South Africa, and at a deeper level, to begin exploring economic alternatives, fearing that a
 political transition would simply entrench a non-racial crisis-ridden capitalism. (Joffe et al.,
 1995:xi; Gelb, 1990) In 1990 this research effort shifted, to "develop an industrial policy
 that could address the poor performance of South African manufacturing." (Joffe et al.,
 !995:xi)
 The biggest challenge to economic liberalisation came not from the research effort, but from
 the struggles by workers to resist the unilateral restructuring of the economy. These
 struggles were exemplified by a three-day stay away in November 1991 to protest the
 apartheid government's IMF-inspired unilateral introduction of Value Added Tax [VAT].
 One of labour's core demands was for the creation of a National Economic Forum [NEF] in
 which the unions could extend the transition's negotiation process to include economic
 restructuring. This institution was composed of civil society formations: business
 associations and trade union federations. As a result, it allowed the labour movement to have
 a direct purchase on policy decisions, something it had lost in the political negotiations which
 took place between political parties. In this realm, COSATU was not subordinate to the
 ANC.
 In the perspective of Steven Friedman and Mark Shaw, who have conducted the fullest
 research on the unions' participation in the NEF, the unions were able to make gains through
 participation, however these were not as significant as the unions had originally hoped, and
 they came with costs. (Friedman and Shaw, forthcoming) The unions had proposed rather
 bold initiatives for industrial restructuring as well as a social plan to deal with the attendant
 social costs. The concrete gains made were less far-reaching: "agreement on a revised offer
 to GATT; overturning a government decision to raise the petrol price; [and] agreement on
 parameters for a public works programme and disbursement of money to organisations
 applying to run programmes." (Friedman and Shaw, forthcoming)
 Moreover, participation imposed costs on COSATU and revealed many weaknesses in the
 organisation. Many of its top negotiators and policy makers became tied down in a process
 that led them to neglect their basic union work. In addition, union negotiators were unable to
 report-back on a regular basis about developments in the NEF, and positions taken in the
 Forum could not effectively reflect a mandate from the organisation. This weakness was
 revealed in a 1994 survey of COSATU members in which 80% knew nothing about the NEF
 and had never been present when there was a report-back on the Forum. (Ginsburg and
 Webster, 1995:67-68)
 Researchers at the COSATU-I inked research unit NALEDI made an even stronger criticism
 about the lack of participation at the NEF.
 Elements of the movement's leadership agreed to positions that were not only
 unknown to the broader leadership, but were, and in many instances continue to be,
 opposed by them. Hence the continued demands by some members and officials not
 to 'sign' or 'join' the GATT. (Lloyd and Rix, 1995)
22
 Thus, not only were the members out of touch with developments in the NEF, but so were the
 federation's leaders, such that agreements could be reached at the NEF that did not reflect the
 positions of many in the federation. They point out that social democratic institutions such
 as the NEF should emerge only after substantial debate within the movement, "albeit often
 only at leadership level." This, argues Chris Lloyd and Steven Rix, has not been the case;
 instead, COSATU "has plunged into this institution without the guidance of such debate,"
 allowing a small leadership clique to make decisions "on the part of a broad movement
 whose thinking is in the opposite direction." (Lloyd and Rix, 1995:14)
 Finally, participation in the Forum revealed a glaring policy weakness in the federation, in
 particular that it did not possess detailed economic plans - the point made above with
 reference to the Industrial Strategy Project - but this point can also be extended to its lack of
 a social policy, and indeed the absence of a social plan. The absence of an economic
 framework was most obvious in "the NEF's macro-economic policy work group, which was
 meant to review current policy and, by implication, consider COSATU's proposed
 alternatives," however this group "failed dismally." (Friedman and Shaw, forthcoming) In
 the Forum the unions were able "to propose, but not to dispose," and as the 1994 elections
 came near the imperative became blocking and delaying policy until an ANC-led government
 could come to power.
 C. The Unions Under Democracy in an Era of Economic Liberalisation
 C.I. The Development of Multi-Partite Institutions
 One of the first Acts of the new democratically elected Parliament was legislation merging
 the NEF with an apartheid-era relic, the National Manpower Commission to create the
 National Economic Development and Labour Council [NEDLAC].23 NEDLAC is a statutory
 body composed of representatives of labour, capital, and the state, who sit in three chambers:
 Labour Market; Trade and Industry; Public Finance and Monetary Policy. A fourth -
 Development - chamber was created to include the broad range of organisations from civil
 society usually excluded from corporatist institutions: community and development
 organisations with a "mass base" and "definable national interests" which are independent of
 the state and not contesting parliamentary power.24 (Webster, 1995b:25-26) NEDLAC's
 central objective is to reach consensus and conclude agreements on social and economic
 policies before they go to parliament. Although parliament is sovereign and NEDLAC is an
 advisory body, a potential consensus between the social partners would be difficult for
 parliamentarians to disregard.
 The NEDLAC framework was first put to test in February 1995 when a new labour relations
 bill was tabled in the council, based on a negotiating document devised by a legal drafting
 team appointed by Minister of Labour Tito Mboweni. The bill proposed far-reaching
 changes to the industrial relations system, introducing for the first time enterprise-level co-
 determinist institutions called workplace forums, as well as extending union organising and
 representational rights while greatly strengthening the right to strike and legalising picketing.
 Indeed, the right to strike was subsequently entrenched in the new constitution following a
 COSATU-initiated general strike on 27 April 1996. At the core of the new system is a
 proposed statutory Commission for Conciliation and Mediation and Arbitration designed to
 shift work relations from adversarialism to a more participatory and co-operative style.
 The bill thus laid the foundations for a new labour regime based on co-decision making. It
 thereby extended unions' powers beyond wage-determination and distributional issues to
 production matters, making important inroads into managerial prerogatives. These rights go
 beyond the workplace, providing labour with access to sectoral and national economic
23
 policy-making through transformed sectoral negotiation chambers - called Bargaining
 Councils - and, ultimately, to NEDLAC. Taken together, these innovations provide labour
 with gains virtually unthinkable in any contemporary industrial democracy, and cut across
 global trends of deregulation of the labour market.
 Employers and unions deadlocked within the first month of discussions on the negotiating
 document, and labour embarked on a campaign of mass action over the next two-and-a-half
 months to support its demands, particularly for centralised bargaining and outlawing scab
 labour. In July the NEDLAC parties announced a breakthrough that enabled the social
 partners to reach consensus. The agreement was produced after 149 hours of formal
 meetings between employers, labour, and government, and in late-night "conversation
 groups."
 The bill was passed by parliament in September 1995, marking the first fundamental break
 with the industrial relations system established in the 1920s. The new system gives labour an
 institutional voice over economic and social policy and the potential to shift the balance of
 power in debates over the emerging labour regime. The opportunity was thus opened for
 labour to craft in the interests of working people policies over wages, employment standards,
 industrial policy, competition, privatisation, and trade.25
 Furthermore, the process through which this consensus was reached is remarkable in itself.
 Labour was able to exercise both mass action and negotiation in institutions to reach tripartite
 arrangements that drew the major stakeholders together into a compromise agreement in
 which all parties had an interest. Although tripartism is time-consuming and elusive and
 arrives at agreements not completely satisfying to all participants, they can improve the
 quality of decisions, build political bases of support for the proposed reforms, and help
 consolidate democratic institutions. (Webster, 1995a)
 If labour was able to develop a separate multipartite institution for influencing industrial
 relations and economic policies (see below), its interventions in the political transition
 occurred in a terrain where it did not hold an equal vote in determining outcomes. As noted
 above, the increasing centrality of political parties was reinforced once agreement was
 reached in CODESA on an election date, which was to be 27 April 1994. This immediately
 set in motion the development of an electoral programme and the construction of a political
 machine. In both of these areas COSATU would play an important role building and
 influencing the ANC - though increasingly as a junior partner to the ANC.
 C 2. Labour and The Elections
 The ANC's electoral campaign came to focus around one major initiative: the Reconstruction
 and Development Programme [RDP]. The RDP originated in an attempt by the National
 Union of Metalworkers of South Africa [NUMSA] to produce an accord that would tie
 COSATU's electoral support for the ANC to the latter's commitment to a working class
 program, (von Holdt, 1991; Gotz, forthcoming)
 The ANC adopted the reconstruction accord, now re-titled as the RDP, in 1993 for very
 particular reasons of its own. Its adoption, Gotz argues, was an electoral gambit, rather than
 an article of faith, adopted in haste without thorough consideration of COSATU's and the
 SACP's vision of radical change, and it laid the basis for the recasting of the document's
 meaning in a direction more attuned to the ANC's complicated electoral needs. In particular,
 its macroeconomic orientation was redefined away from the unions' commitment to "growth
 through redistribution" to a direction more attractive to domestic and international capital,
24
 with commitments to fiscal discipline and macroeconomic balance.
 The redrafted RDP was sharply criticised on both procedural and substantive grounds at a
 special COSATU congress held in August 1993. The redraft was presented for adoption
 without being debated within COSATU structures, and its content had been changed to
 include a statement that "coherent, strict, and effective monetary and fiscal policies will be a
 cornerstone of our RDP." (Etkind and Harvey, 1993) The re-formulation provoked
 considerable debate at the Congress over what was seen as the thin end of the neoliberal
 wedge. However, COSATU adopted the program without suggesting an alternative to this
 call for fiscal discipline at the heart of the RDP. The RDP soon became the paradigm within
 which all development policies were to be discussed, an extended wish list in which the
 homeless, the landless, workers, and international bankers could take equal comfort. In other
 words, it became all things to all people. (African National Congress, 1994)
 At the same special congress, COSATU constructed the other leg of its electoral strategy in
 support of the ANC. Twenty COSATU leaders were released to stand as parliamentary
 candidates on the ANC's list, including some of the organisation's most senior office bearers
 and strategists. A larger number of leaders were similarly released to stand on provincial
 lists (and later on local government electoral lists). The guiding idea was that such
 individuals would strengthen the capacity of the ANC and at the same time shape its
 direction towards labour's goals.
 Tt proved difficult to fulfill these aims. Many ex-COSATU leaders gained influence,
 especially those few became cabinet ministers and chairpersons of parliamentary
 committees, but COSATU as an organisation did not. Indeed, it suffered what has come to
 be known as a "brain drain" in which many of its leading figures, especially at regional and
 local levels, left the organisation. (Buhlungu, 1994a, 1994b) An estimated 80 key leaders left
 the organisation at the time of the national elections; a much larger number of lower-level
 leaders left in late 1995 after the local government elections. In the National Union of
 Mineworkers itself more than 100 shopstewards went into local government structures.26
 However, once in government these individuals found themselves detached from any
 accountability to their old federation while being isolated from the centre of power in their
 home, the ANC. The ex-unionists did not form a COSATU caucus in parliament, and as
 ANC members they are subject to party discipline; indeed, a clause in the constitution
 compels MPs to resign their seats if they leave or are expelled from their party, which gives
 party leaders extensive power over dissident members.
 The 1994 elections cemented the ANC's leadership within the Triple Alliance. By virtue of
 its overwhelming electoral victory, the ANC became the majority party in a legitimate
 Government of National Unity. As described above, the party leadership - principally the
 ministers and Deputy President Thabo Mbeki - now enjoyed considerable leverage over its
 MPs (including the ex-COSATU unionists in Parliament), not to mention the organisation's
 extra-parliamentary structures. Furthermore, it could now summon the vast resources of the
 state bureaucracy in pursuit of its policy goals. The transition from movement to
 Government had been initiated, if not completed.27 In practical terms, this meant that policy
 developments such as the RDP, that had once been the object of bargaining within the
 Alliance, now became government policies to be developed within state structures, including
 a bureaucracy whose composition was largely unchanged as a result of the CODESA "sunset
 clauses" that guaranteed the positions of the old white bureaucracy.
 From 1994 to 1996 the RDP became the ostensibly guiding document of the Government of
25
 National Unity, located in an RDP Office within the President's Office, under the immediate
 authority of Minister Without Portfolio Jay Naidoo - ex-General Secretary of COSATU.
 Within months the Alliance's election document was redrafted into an official government
 White Paper. (Parliament of the Republic of South Africa, 1994) However, the principles
 and programmatic objectives of the original RDP were reconceptualised as a long-term
 strategic vision to realign all government effort around clearly stated economic targets,
 leaving little room for labour-, or indeed civil society-driven change. (Gotz, forthcoming)
 Furthermore, the macroeconomic shift noted above became even more pronounced in the
 White Paper which significantly strengthened the emphasis on fiscal discipline that first
 appeared in the draft presented to the 1993 COSATU special congress. It developed a more
 robust commitment to an export-led growth strategy while loosening the RDP's original
 emphasis on a basic needs strategy stressing the need to reduce state expenditure,
 privatisation, and promoting private sector expansion. For example, the RDP placed ultimate
 responsibility on the government "for ensuring that housing is provided to all," and that
 housing be affordable "to even the poorest South Africans." (African National Congress,
 1994: 23). Instead the White Paper envisions a market-based approach to housing through
 "innovative financial institutions and instruments which promote domestic savings and
 extend financial services to those who do not have adequate access to these services."
 (Parliament, 1994: 25) But whereas the RDP articulated a vision of housing and community
 banks and mechanisms to make the financial sector responsive to popular needs (and
 maintained government as the provider of last resort), the White Paper drops any explicit
 suggestions on these lines, while a high interest rate regime puts housing finance beyond the
 means of most South Africans. (Parliament, 1994; Adelzadeh and Padayachee, 1994)
 C. 3. Contesting Economic Liberalisation
 The first contestation over liberalisation occurred in December 1995, when the government
 announced its aspirations to privatise important state assets. Labour responded by
 threatening a general strike and demanded the right to negotiate the terms of restructuring.
 After intensive discussions between the NEDLAC labour caucus28 and senior government
 ministers, a National Framework Agreement [NFA] was accepted in February 1996. In terms
 of the NFA all specific instances of restructuring must be negotiated between government
 and the relevant trade union parties. Most importantly, the government agreed that
 restructuring should not be done at the expense of employment and where there are negative
 consequences for workers a social plan must be negotiated. (National Framework
 Agreement, 1996; Rix and Jardine, 1996) This labour-sensitive agreement stands in marked
 contrast to the unilateral embracing of privatisation by the MMD government in Zambia.29
 The shifting meaning attached to the RDP provided a context in which government, business
 and labour began developing distinctive positions on economic policy. Business opened the
 public debate in January 1996 with a document advocating a standard neoliberal solution:
 economic deregulation; dramatic reductions in the deficit; trade liberalisation in the form of
 lower tariffs; rapid privatisation; lower corporate taxes; and increased labour market
 flexibility to be achieved through dual labour market policies. (South Africa Foundation,
 1996)
 Soon after business' document was published, COSATU released an alternative plan, "Social
 Equity and Job Creation," inspired by left-Keynesian thinking that has become the common
 position in the NEDLAC labour caucus. (Labour Caucus, 1996) The plan attacks
 neoliberalism and proposes instead a public and private investment policy geared towards job
 creation and growth. Central to the strategy is an active industrial policy to develop the
26
 manufacturing sector along with social adjustment measures to provide for the social costs of
 restructuring the economy. Importantly, the document proposes a redistributive fiscal policy
 based on a strongly progressive tax system that will redirect spending towards social services
 for the poor. The program also articulates a vision of regional reconstruction and
 development tied to support for a third world debt write-off and closer trade union solidarity.
 Finally, it incorporates demands for worker participation, both at the shop floor in the form
 of union-based Workplace Forums and at the sectoral and national levels.
 In the meantime, the Government was developing its own programme, the National Growth
 and Development Strategy [NGDS] which ambiguously straddled the neoliberal
 macroeconomic program of big business's approach and the redistributive and equity
 orientations of labour's document. (Inter-governmental Forum, 1996) However, this
 document was withdrawn because of opposition within the ANC and its allies, who criticised
 both its content and the process by which the document was written. As a result of this
 deadlock in the Alliance, the NGDS was never publicly released.
 However, South Africa's first major currency crisis - starting in February 1996 - quickly
 overtook this debate, as the value of the Rand dropped by more than 25%. (Lloyd and Rix,
 1996) In this context the government moved quickly to calm domestic capital and foreign
 currency markets by embracing a conservative macroeconomic framework developed by a
 team of policy makers including senior civil servants, representatives of the Development
 Bank of South Africa, and the World Bank. In June, after considerable internal disagreement
 within the Tripartite Alliance, Finance Minister Trevor Manuel finally released the new
 strategy, "Growth, Employment and Redistribution" [GEAR]. (Department of Finance,
 1996).
 GEAR puts forward ambitious targets. It aimed to achieve a "fast-growing economy which
 creates sufficient jobs for all workseekers." It also envisioned redistribution of income, the
 provision of sound services to all, and the achievement of 6% growth and the creation of
 400,000 jobs per annum by the year 2000. (Department of Finance, 1996: 1) The
 controversial aspect of the plan is the means identified to achieve these goals. Its integrated
 strategy highlights fiscal deficit reduction, gradual relaxation of exchange controls, reduction
 in tariffs, tax reductions to encourage private sector (and especially foreign direct)
 investment, and restructuring of state assets (privatisation).
 The ANC sought to win support for this shift in policy within the Alliance by a series of
 informal meetings with senior SACP and COSATU officials in the weeks before GEAR was
 released. ANC leaders stressed the precarious position they were in and the need to project
 confidence during the crisis, and that GEAR was the best means for doing this. The Alliance
 partners, while acknowledging the pressures on the government, queried the origins of
 GEAR and were critical of the lack of consultation. Indeed, the actual document was not
 tabled in the meetings - the COSATU and SACP leaders were shown only the section
 headings.
 However once GEAR was released, the Alliance partners were angered both by its content -
 which they now saw for the first time - and because the government asserted that it was
 "non-negotiable", a position that was reiterated by numerous ministers and spokespersons.
 These two factors provoked anger from COSATU as well as among many figures in the
 SACP. COSATU General Secretary Sam Shilowa publicly criticised GEAR - tellingly at a
 seminar marking the 75th anniversary of the SACP - describing the strategy as an
 "unworkable and unwinnable" plan that "poses serious difficulties for the working class and
27
 the country as a whole." Most significantly, Shilowa identified the distance the ANC has
 moved rightward from the original RDP by indicating that GEAR could never have emerged
 from the ANC before the 1994 elections. (Sunday Times, 1996; Sunday Independent, 1996)
 Despite these criticisms, GEAR was ultimately formally endorsed by the ANC's National
 Working Committee, on which COSATU and SACP representatives sit, though COSATU
 has never adopted the policy. Yet no formal agreement could be reached in either COSATU
 or the SACP on how to deal with GEAR: some felt that the policy should be given a chance
 and the government should not be attacked during a crisis; others wanted openly to oppose
 the policy. The issue threatened to be divisive both within and between the parties to the
 Alliance; as one participant put it, "the debate became so shrill that people walked away from
 it."
 However, silence over GEAR did not mean acquiescence, and during 1997 the terrain
 became more fluid. Firstly, line ministers experienced the consequences of GEAR'S fiscal
 restraint and the implications this policy held for delivering on the government's social
 programs. Secondly, GEAR has not been delivering on its ambitious promises of
 employment creation and redistribution. In this context, COSATU and the SACP seem to be
 avoiding a direct confrontation over GEAR, and instead are resisting it indirectly, a strategy
 best illustrated by their involvement in the revision of the Basic Conditions of Employment
 Bill. Business has advocated greater flexibility in the labour market while labour has sought
 a reduction in the working week, increased maternity benefits, and has resisted attempts to
 "level downwards" workers' conditions of employment. After more than a year of
 negotiations in NEDLAC COSATU declared a deadlock and embarked on a campaign of
 mass action, including a one-day general strike in June 1997, and promises to intensify action
 until their demands are met and a the draft bill is further revised.
 COSATU has pursued a sophisticated strategy, using its presence in NEDLAC to advance its
 argument, resorting to mass action when it has failed to win its demands there; but it has also
 indicated that it will use its alliance with the ANC in parliament to win revisions to the Bill.
 This is a clear example of how COSATU is able to act both inside and outside the state, using
 the combination of mobilisation and influence within the Alliance to force a bargained
 liberalisation.
 In the long term, it will not be easy to bridge business's and labour's positions since they
 come from different economic paradigms and represent different conceptions of labour
 regimes. A pact, it has been argued, would discipline all parties: if labour is given increased
 influence over investment policy this will provide the possibility that the benefits of growth
 will improve living standards for workers and the poor. On the other hand, business, too,
 will gain advantages from a pact through providing labour with a strong stake in economic
 growth. This would contribute to industrial relations and wage stability, and thereby
 encourage investor confidence by providing business with increased certainty over key cost
 variables. Finally, government would be forced to develop macro-economic policies with
 respect to the interests of both capital and labour; at a minimum this requires a commitment
 to redistributive economic and social policies that perforce imply some relaxation of the
 government's cautious fiscal approach. This could be the core of a class compromise that
 could bring about an economic shift similar to that produced in the political transition: a turn
 away from the current neoliberal policy towards a program much more oriented to the needs
 of labour and the poor.
 NEDLAC's institutionalised role in policy making provides some of the preconditions for a
30
 challenges any notion of ultimate subordination of trade unions and worker acquiescence.
 COSATU, too, has been reshaped by the transition process. It has lost layers of leadership to
 government and the corporate sector. It has had to learn new rules of the game, both in
 politics and in the market. It has for the first time opened a parliamentary office in Cape
 Town to lobby government and has begun to learn the diplomacy and power dynamics of
 dealing with an alliance partner in government. Arguably its biggest challenge is organising
 new segments of workers: both those now covered by the democratised Labour Relations Act
 (domestics, public service workers, farm workers), and the new sectors of the working class
 generated by globalisation (part time, casuals, home and contract workers). Notwithstanding
 its present numerical and strategic strengths, it might not master these new opportunities, and
 may in future shrink, facilitating its marginalisation.
 Transition is not a cab to be taken at will: it does not stop short of its advocates, who are
 themselves transformed by political and economic liberalisation. Nor however, is its
 destination predetermined - least of all by external agents. Civil society formations - we have
 here analysed trade unions - can hope to shape the character of this transformation through a
 combination of collective action and institutional engagement which affords them an
 opportunity to bargain the terms of liberalisation. Unions can use their power to create
 institutions that provide workers with a collective voice through which they can formulate
 rules that both constrain capital and the state and promote more egalitarian policies. Clearly
 these possibilities do not exist in all countries in Africa, but our study suggests that
 possibilities for forms of bargained liberalisation do not exist only in South Africa.
31
 ENDNOTES
 1. This approach was endorsed by the World Bank itself in its landmark 1989 assessment of structural adjustment,
 Sub-Saharan Africa: From Crisis to Sustainable Growth. (World Bank, 1989) For criticisms of its adoption of
 governance and reconsideration of the role of the state, see Beckman (1990), Ferguson (1995), Harsch (1993),
 Lemarchand (1992), Mkandawire (1994), and Schmitz (1995).
 2. Or as the political scientist Carol Lancaster put it - before joining the new Clinton administration as Deputy
 Administrator of USAID - "where a democratically elected government proved weak, inept, or corrupt, its
 replacement by a competent military government could, in the long run, lead to more rapid development and the
 prospect for a better standard of living for that country's citizens.... Where a new military government appears stable
 and competent, the United States should continue financing ongoing aid projects..." (Quoted in Harsch, 1993:27).
 3. In an earlier study, Przeworski gives three reasons why concertation is not feasible in less developed countries:
 unions represent only a small proportion of workers; they are not strong and centralised enough; and employers are
 hostile to them while government is ambivalent. (Przeworski, 1991:186).
 4. We have argued elsewhere that Przeworski's work fits into this category, (Adler and Webster, 1995) but the same
 orientation is also present in the work of J. Samuel Valenzuela (1989),
 5. This approach is broadly consistent with that taken by a recent comprehensive study of labour and
 democratisation in Latin America, the Caribbean and Europe by Rueschemeyer et al. (1992). See also a recent
 working paper by Collier and Mahoney (1995).
 6. See Mosley et al. (1991) who identify variables to explain the capacity of states to deviate from World Bank
 conditionalities. Working from the Bank's own evidence, they find that conditionalities are ignored nearly half the
 time, and that deviation occurs in a systematic fashion: some states are able to chart a relatively independent path
 while others are not. Working from a different conceptual framework, but advancing similar conclusions, Martin
 (1991) assesses IMF conditionality as an outcome of bargaining in which the IMF does not always get its way, and
 that deviations occur in a systematic fashion. For a summary of different approaches to globalisation, see Murray
 (1997).
 7. An analysis of labour's role in bargained liberalisation is not a substitute for consideration of the other key actors:
 business, other civil society formations, the state, and their interactions in the context of a particular country's
 insertion in die international economy. We hope that this more ambitious discussion will be the subject of further
 phases of this project.
 8. According to Bates, who surveyed mineworkers in the late 1960s, "the interests of the local branch officials
 remain those of the workers....The local leaders simply do not possess the resources necessary to induce the
 members of the mining community to act contrary to their conception of their interests. The union's structure is
 ill-suited for downward communications, especially when the viewpoints they wish to transmit are at variance with
 the interests of the workers." (Bates, 1971:202).
 9. Appreciation to Franco Barchiesi who coined the term and to the students in Dr. Adler's 1995 postgraduate
 seminar on "African Labour Movements: Struggles for Democracy," who helped develop the concept.
 10. There have been similar ironic developments elsewhere in Africa. For Nigeria, see Otobo (1992) and Bangura
 (1991); for Zimbabwe, see Sachikonye (1993) and Wood (1988).
 11. Indeed, the Mineworkers Union of Zambia declared that any leader of the MUZ (as the AMWU had been
 renamed after independence) standing in the local UNIP elections "should either withdraw or resign from leadership
 positions in the union." The union later dismissed sixteen shopstewards from their positions after they took part in
 party elections. (Rakner, 1992:101)
32
 12. Aid was restored at a recent meeting of donor nations at the World Bank's offices in Paris, but human rights
 organisations continue to lobby for donors to pressure Lusaka to cease harassment of opposition figures and parties,
 the media and NGOs. (Business Day, 1997)
 13. The World Bank has applauded the ZPA's record as a model for Africa. (World Bank, 1996)
 14. These strikes seriously delayed the government's liberalisation plans and contributed to the criticism by foreign
 donors - especially the IMF and World Bank - that it was not living up to its adjustment commitments. The
 government's "foot-dragging" - taken as a sign of bad faith by the international community - was rather an
 indication of labour's ability to frustrate the roll back of the public sector.
 15. Impartial assessments have favourably rated the medium-term growth prospects in mining, as reflected in the
 number of multinationals making serious bids for the state-owned Zambian Consolidated Copper Mines. (Bull,
 1996)
 16. However, the 1995 Labour Relations Act's provision preventing labour lawyers from representing workers at
 the new Commission for Conciliation, Mediation and Arbitration has led to the unintended consequence of a
 proliferation in small unions. Most of these have been launched by unscrupulous lawyers and labour brokers.
 17. Migrants were included under the Act through reforms during the early 1980s, while the latter groups were
 included only after the transition to democracy in the 1990s.
 18. SACTU ultimately disbanded and was absorbed into COSATU when the ANC was unbanned, by which time
 very few of its personnel remained. (Southall, 1995)
 19. By contrast, the SACP remains heavily dependent on financial support from the ANC. An unintended
 consequence of the new funding formula is that unions have amassed relatively large pools of money drawn from
 members' subscriptions. These have provided the basis for a controversial innovation: the creation of union
 investment arms in which excess funds are invested in private enterprises. (Naidoo, 1997)
 20. It is difficult to label the emerging unions' economic vision, in part because they never publicly advocated one,
 due to repressive legislation which made support for socialism a serious statutory offense. However, it is possible
 to identify their ideas as socialist, with two dominant strands of thought. The first was an "old left," largely
 associated with the SACP, and looking towards existing socialism for its economic and political models. The
 second - and more hegemonic in the unions - was broadly "new left" in orientation, associated with university
 lecturers and graduates active in the new unions, who were influenced by the 1960s shopstewards movement in
 Britain, and ideas of participatory democracy, including Yugoslavian workers' self-management. Despite
 differences, both approaches had a vision of a future South Africa in which there would be a fundamental
 redistribution of wealth and power; workers' would control industry and agriculture, and the economy would be
 run along planned lines. This vision is captured most clearly in Richard Turner's seminal book 1972, The Eye of
 the Needle.
 21. Unions sought recognition agreements from employers to gain access for their organizers to sign up members,
 and for shopstewards to meet during company hours. They aimed to establish forma! wage bargaining, as well as
 grievance and disciplinary procedures to protect their members' interests. Registration under the Labour Relations
 Act was viewed as a tactical necessity, offering further protection for the new unions and enabling them to
 participate in all the institutions of the industrial relations system. For an overview of the debate on registration and
 its key participants, see Maree (1987).
 22. An important symbol of this shift was the disbanding of the United Democratic Front [UDF], the organisation
 that had sustained hundreds of civil society organisations in the 1980s. Without the UDF these groups lost
 resources; more importantly they were deprived of the voice in national political affairs the UDF historically
 provided. Furthermore, the transition created unprecedented opportunities for upward mobility: many activists
 entered national and regional parliaments and local councils, the state bureaucracy, the corporate sector or withdrew
 from activism altogether. Though many such individuals continued to make contributions in these new spheres,
 their departures tended to deprive civil society of left intellectual leadership at the moment when it was most needed.
33
 This trend can also be discerned amongst non-political organisations in civil society, such as NGOs, that faced a
 fiscal crisis with the withdrawal in foreign funding that accompanied normalisation in South Africa. These
 dynamics were sharply felt in the alternative media, and led to the closure of alternative newspapers and magazines
 and publishers. (On the media, see Shand, 1996)
 23. The National Manpower Commission was established in the early 1980s as part of the labour reforms initiated
 by the Wiehahn Commission. It was a tripartite body appointed by the Minister of Labour, with the function of
 monitoring the new labour system and advising the government on manpower issues. It was immediately boycotted
 by the emerging unions until the transition to democracy, when they sought - unsuccessfully - to enter and transform
 it. (Schreiner, 1991)
 24. COSATU never conceived of NEDLAC as a narrow corporatist project based on the "Big 3" of labour, capital
 and the state. For arguments on the need for a multi-partite forum and the dangers of "narrow corporatism," see
 Bird and Schreiner (1992).
 25. This agreement did not win acceptance everywhere; in particular there was strong criticism from within
 COSATU, most prominently from the Witwatersrand Local, which condemned the consensus as a "miserable
 compromise." (von Holdt, 1995)
 26. The movements from COSATU to local and regional government and to the private sector were more disruptive
 than the deliberate secondment to the ANC's parliamentary list. According to private assessments within COSATU,
 these secondments - far from weakening the federation - in fact created space for new leaders to emerge: beyond
 a small number of capable individuals, the federation was able to divest itself of considerable "dead wood".
 27. In addition to any conscious diminution of the ANC as a movement, the transfer of cadre into all levels
 government and more recently a fiscal crisis necessitating wide-scale retrenchments have seriously weakened
 whatever grassroots capacity the organisation had been able to develop since its unbanning. As one confidential
 source suggested, "Shell House is an appropriate name, as the organisation has become a shell."
 28. Consisting of COSATU, NACTU, and FEDUSA.
 29. However, though labour won the battle over privatisation, it has had great difficulty taking advantage of the
 opening it created. (Ray, 1997)
34
 REFERENCES
 Books, articles and papers cited:
 Adam, Christopher S. and David L. Bevan. 1996. "Aid and the Credit Squeeze in Zambia: Causes and
 Consequences." Centre for the Study of African Economies Macroeconomic Report, October.
 Adelzadeh, Asghar and Vishnu Padayachee. 1994. "The RDP White Paper: Reconstruction of a
 Development Vision?" Transformation 25.
 Adler, Glenn and Eddie Webster. 1995. "Challenging Transition Theory: The Labour Movement,
 Radical Reform and Transition to Democracy in South Africa." Politics and Society 23:1.
 African National Congress. 1994. The Reconstruction and Development Programme: A Policy
 Framework. Johannesburg: Umanyano Publications.
 Akwetey, Emmanuel. 1996. "Hegemonic Unilateralism or Autonomous Multi-Actor Reciprocity:
 Dilemmas of Labour Regime Reform in Zambia and Ghana." Paper presented to the Workshop on
 Labour Regimes and Liberalisation," Institute for Development Studies, University of Zimbabwe,
 May.
 Akwetey, Emmanuel. 1994. Trade Unions and Democratisation: A Comparative Study of Zambia and
 Ghana. Stockholm Studies in Politics 50, Department of Political Science, University of Stockholm.
 Bangura, Yusuf. 1991. "Crisis and Adjustment: The Experience of Nigerian Workers." In The IMF,
 the World Bank, and the African Debt, Volume 2: The Social and Political Impact, edited by Bade
 Onimode. London: Zed Press.
 Baskin, Jeremy. 1991. Striking Back: A History of COSATU. Johannesburg: Ravan.
 Bates, Robert H. 1971. Unions. Parties, and Political Development: A Study of Mineworkers in
 Zambia. New Haven: Yale University Press.
 Baylies, Carolyn and Morris Szeftel. 1992. "The Fall and Rise of Multi-Party Politics in Zambia."
 Review of African Political Economy 54.
 Beckman, Bjorn. 1990. "Empowerment or Repression? The World Bank and the Politics of African
 Adjustment." Paper presented to a symposium on "The Social and Political Context of Structural
 Adjustment in Sub-Saharan Africa," Chr. Michelsen Institute, Bergen, Norway, 17-19 October.
 Bienefeld, Manfred. 1975. "Socialist Development and the Workers in Tanzania." In The Development
 of an African Working Class: Studies in Class Formation and Action, edited by Richard Sandbrook and
 Robin Cohen. London: Longman.
 Bird, Adrienneand Geoff Schreiner. 1992. "COSATU at the Crossroads: Towards Tripartite
 Corporatism or Democratic Socialism?" South African Labour Bulletin 16:6.
 Bolton, Dianne. 1978. "Unionization and Employer Strategy: The Tanganyikan Sisal Industry, 1958-
 1964." In African Labour History, edited by Peter C. W. Gutkind, Robin Cohen, and Jean Copans.
 Beverly Hills: Sage Publications.
 Bratton, Michael. 1992. "Zambia Starts Over." Journal of Democracy (April).
35
 Buhlungu, Sakhela and Lucien van der Walt. 1997. "The Voice From Below: Zambian Workers
 Speak Out." South African Labour Bulletin 21:3.
 Buhlungu, Sakhela. 1994a. "COSATU and the Elections." South African Labour Bulletin 18:2.
 Buhlungu, Sakhela. 1994b. "The Big Brain Drain: Union Officials in the 1990s." South African
 Labour Bulletin 18:3.
 Bull, Theo. 1996. "ZCCM - Gloomy Past, Hopeful Future?" Profit (November).
 Burawoy, Michael. 1972. "The Colour of Class on the Copper Mines: From African Advancement to
 Zambianization." University of Zambia, Institute for African Studies, Zambian Paper no. 7.
 Business Day. 1997. "Zambians Win Aid Pledges Worth More than $400m," (14 July).
 Callaghy, Thomas. 1994. "Africa: Falling off the Map? Current History (January).
 Callaghy, Thomas. 1990. "Lost Between State and Market: The Politics of Economic Adjustment in
 Ghana, Zambia, and Nigeria." In Economic Crisis and Policy Choice: The Politics of Adjustment in
 the Third World, edited by Joan M. Nelson. Princeton: Princeton University Press.
 Chipungu, Samuel N. 1992. "African Leadership Under Indirect Rule in Colonial Zambia." In
 Guardians In their Time: Experiences of Zambians Under Colonial Rule, 1890-1964, edited by Samuel
 N. Chipungu. London: Macmillan.
 Cohen, Robin, et al. 1978. "Introduction." In African Labour History, edited by Peter C. W. Gutkind,
 Robin Cohen, and Jean Copans. Beverly Hills: Sage Publications.
 Collier, Ruth and James Mahoney. 1995. "Labour and Democratization: Comparing the First and
 Third Waves in Europe and Latin America." University of California at Berkeley, Institute of
 Industrial Relations, Working Paper no. 62.
 Crisp, Jeff. 1984. The Story of an African Working Class: Ghanaian Miners' Struggles, 1870-1980.
 London: Zed Books.
 Davies, loan. 1966. African Trade Unions. Harmondsworth: Penguin.
 Department of Finance. 1996. Growth, Employment and Redistribution: A Macro-Economic
 Strategy. Pretoria: Government Printer.
 Department of Labour. 1996. Restructuring the South African Labour Market: Report of a
 Commission to Investigate the Development of a Comprehensive Labour Market Policy. Pretoria:
 Government Printer.
 Department of Labour. 1995. Annual Report. Pretoria: Government Printer.
 Diamond, Larry. 1993. Political Culture and Democracy in Developing Countries. Boulder: Lynne
 Reinner.
 Diamond, Larry. 1988. "Introduction: Roots of Failure, Seeds of Hope." In Democracy in Developing
 Countries, Vol. 2: Africa, edited by Larry Diamond, Juan J. Linz and Seymour Martin Lipset. Boulder:
 Lynne Reinner.
 Diamond, Larry and Mark Planner, eds. 1993. The Global Resurgence of Democracy. Boulder: Lynne
 Reinner.
36
 Etkind, Roger and Suzanna Harvey. 1993. "The Workers Cease Fire." South African Labour
 Bulletin 17:5.
 Evans, Peter, ed. 1997. State-Society Synergy: Government and Social Capital in Development.
 University of California at Berkeley, International and Area Studies Research Series, no. 94.
 Ferguson, James. 1995. "From African Socialism to Scientific Capitalism: Reflections on the
 Legitimation Crisis in IMF-Ruled Africa." In Debating Development Discourse: Institutional and
 Popular Perspectives, edited by David B. Moore and Gerald J. Schmitz. London: Macmillan.
 Filita, Templeton. 1996. Unpublished research findings on COSATU membership. National Labour
 and Economic Development Institute. Johannesburg.
 Fine, A and E. Webster. 1989. "Transcending Traditions: Trade Unions and Political Unity." In
 South African Review 5, edited by Glenn Moss and Ingrid Obery. Johannesburg: Ravan Press.
 Foster, Joe. 1982. "The Workers' Struggle - Where Does FOSATU Stand? Address by Joe Foster
 to the 1982 FOSATU Congress." Johannesburg: FOSATU.
 Freund, Bill. 1988. The African Worker. Cambridge: Cambridge University Press.
 Friedman, Steven. 1987. Building Tomorrow Today: African Workers in Trade Unions, 1970-1984.
 Johannesburg: Ravan Press.
 Friedman, Steven and Mark Shaw. Forthcoming. "Power in Partnership? Trade Unions, Forums,
 and the Transition." In Trade Unions and Transition in South Africa: Consolidating Democracy in a
 Liberalizing World, edited by Glenn Adler and Eddie Webster. London: Macmillan.
 Gelb, Stephen, ed. 1990. South Africa's Economic Crisis. Cape Town: David Philip.
 Ginsburg, David, Eddie Webster, et al. 1995. Taking Democracy Seriously: Worker Expectations
 and Parliamentary Democracy in South Africa. Durban: Indicator Press.
 Ginsburg, Matthew. 1997. "Zambia for Sale: Privatisation and the Labour Movement." South African
 Labour Bulletin 21:3.
 Gostner, Karl. 1997. "Playing Politics: Labour's Political Role in Zambia," South African Labour
 Bulletin 21:3.
 Gotz, Graeme. Forthcoming. "Shoot Anything That Flies: Claim Anything That Falls: Labour and
 die Changing Definition of the Reconstruction and Development Programme." In Trade Unions and
 Transition in South Africa: Consolidating Democracy in a Liberalizing World, edited by Glenn Adler
 and Eddie Webster. London: Macmillan.
 Harsch, Ernest. 1993. "Structural Adjustment and Democracy Movements." Africa Today 4th
 Quarter.
 Harvey, Suzanna. 1996. "Labour Market Killing fields." South African Labour Bulletin 20:2.
 Human Rights Watch. 1996. "Zambia: Elections and Human Rights in the Third Republic." Human
 Rights Watch/Africa Report 8:4(A).
 Inter-governmental Forum. 1996. "National Growth and Development Strategy." Pretoria:
 Government Printer.
 Jeffries, Richard. 1975. "Populist Tendencies in the Ghanaian Trade Union Movement." In The
37
 Development of an African Working Class: Studies in Class Formation and Action, edited by Richard
 Sandbrook and Robin Cohen. London: Longman.
 Jeffries, Richard. 1978. Class, Power and Ideology in Ghana: The Railway men of Sekondi.
 Cambridge: Cambridge University Press.
 Joffe, Avril, David Kaplan, Raphael Kaplinsky, and David Lewis. 1995. Improving Manufacturing
 Performance in South Africa: Report of the Industrial Strategy Project. Cape Town: University of
 Cape Town Press.
 Joseph, Richard. 1990. "Political Renewal in Sub-Saharan Africa: The Challenge of the 1990s." In
 African Governance in the 1990s: Objectives, Resources, and Constraints Working Papers from the
 Second Annual Seminar of the African Governance Program. Atlanta: The Carter Center of Emory
 University.
 Ki Hick, Tony, 1995. The Flexible Economy: Causes and Consequences of the Adaptability of
 National Economies. London: Routledge.
 Labour Caucus [COSATU, FEDSAL, NACTU]. 1996. "Social Equity and Job Creation: The Key to
 a Stable Future." Johannesburg.
 Labour Monitoring Group. 1985. "The November 1984 Stay away." South African Labour Bulletin
 10:6.
 Lancaster, Caroi. 1991. "Democracy in Africa." Foreign Affairs 85.
 Lemarchand, Rene. 1992. "Africa's Troubled Transitions." Journal of Democracy 3:4.
 Lloyd, Chris and Stephen Rix. 1996. "Towards an Economic and Employment Summit." Unpublished
 paper prepared for the General Secretary of the National Union of Metalworkers of South Africa.
 Lloyd, Chris and Stephen Rix. 1995. "Unions and Democratic Institutions." Unpublished discussion
 paper, National Labour and Economic Development Institute. Johannesburg.
 Macun, Ian. Forthcoming. "Growth, Structure, and Power in the South African Union Movement." In
 Trade Unions and Transition in South Africa: Democratisation in a Liberalising World, edited by
 Glenn Adler and Eddie Webster. London: Macmillan.
 Maree, Johann. 1987. The Independent Trade Unions, 1974-1984: Ten Years of the South African
 Labour Bulletin. Johannesburg: Ravan Press.
 Martin, Matthew. 1991. The Crumbling Facade of African Debt Negotiations: No Winners. London:
 Macmillan.
 Minter, Bill. 1986. King Solomon's Mines Revisited: Western Interests and the Burdened History of
 Southern Africa. New York: Basic Books.
 Mkandawire, Thandika. 1994. "Adjustment, Political Conditionality and Democratisation in Africa."
 In From Adjustment to Development in Africa: Conflict, Controversy, Convergence, Consensus? edited
 by Giovanni Andrea Cornia and Gerald K. Helleiner. New York: St. Martin's Press.
 Mosley, Paul, Jane Harrigan, and John Toye. 1991. Aid and Power: The World Bank and Policy-based
 Lending. London and New York: Routledge.
 Mphaisha, Chisepo J.J. 1996. "Retreat from Democracy in Post One-Party State Zambia," Journal of
 Commonwealth and Comparative Politics 34:2.
38
 Murray, Georgina. 1997. "Ideology and the Globalising Strategy of Big Business in Australasia in the
 1990s." Paper presented at the Annual Conference of the South African Sociological Association,
 Umtata, July.
 Naidoo, Ravi. 1997. "Union Investments: Preliminary Assessment and Framework Development."
 National Labour and Economic Development Institute, Johannesburg.
 National Framework Agreement. 1996. South African Labour Bulletin 20:2.
 Otobo, Dafe. 1992. "State and Labour: The 1988 Ban of the NLC." In Labour Relations in Nigeria,
 Volume One, edited by Dafe Otobo. Lagos: Malthouse Press.
 Panter-Brick, Keith. 1994. "Prospects for Democracy in Zambia." Government and Opposition 29:2.
 Parliament of the Republic of South Africa. 1994. "White Paper on Reconstruction and
 Development." Government Gazette, 23 November.
 Pityana, Sipho Mila, and Mark Orkin. 1992. Beyond the Factory Floor: A Survey ofCOSATU
 Shopstewards. Johannesburg: Ravan Press.
 Przeworski, Adam, et al. 1995. Sustainable Democracy. Cambridge: Cambridge University Press.
 Przeworski, Adam. 1991. Democracy and the Market: Political and Economic Reforms in Eastern
 Europe and Latin America. Cambridge: Cambridge University Press.
 Rakner, Lise. 1992. "Trade Unions in the Processes of Democratisation: A Study of Party Labour
 Relations in Zambia." Chr. Michelsen Institute, Bergen, Norway, Report no. 6.
 Ray, Malcom. 1997. "Privatisation: Pushing Ahead." South African Labour Bulletin 21:3.
 Riley, Stephen P. and Trevor W. Parfitt. 1995. "Economic Adjustment and Democratization in
 Africa." In Free Markets and Food Riots: The Politics of Global Adjustment, edited by John Walton
 and David Seddon. Oxford: Blackwell.
 Rix, Stephen and Conrad Jardine. 1996. "Privatisation: The Debate Over Restructuring of State
 Assets." In Against The Current: Labour and Economic Policy in South Africa, edited by Jeremy
 Baskin. Johannesburg: Ravan.
 Roberts, Andrew. 1979. A History of Zambia. New York: Africana.
 Rotberg, Robert. 1965. The Rise of Nationalism in Central Africa: The Making of Malawi and Zambia,
 1887-1964. Cambridge: Cambridge University Press.
 Rueschemeyer, D., Evelyn Huber Stephens, and John Stephens. 1992. Capitalist Development and
 Democracy. Chicago: University of Chicago Press.
 Sandbrook, Richard. 1975. Proletarians and African Capitalism: The Kenyan Case, 1960-1972.
 Cambridge: Cambridge University Press.
 Sachikonye, Lloyd M. 1993. "Structural Adjustment, State, and Organised Labour in Zimbabwe." In
 Social Change and Economic Reform in Africa, edited by Peter Gibbon. Uppsala: The Scandinavian
 Institute of African Studies.
 Schmitz, Gerald J. 1995. "Democratization and Demystification: Deconstructing 'Governance1 as
 Development Paradigm." In Debating Development Discourse: Institutional and Popular
 Perspectives, edited by David B. Moore and Gerald J. Schmitz. London: Macmillan.
39
 Schreiner, Geoff. 1991. "Fossils From the Past: Restructuring the National Manpower Commission."
 South African Labour Bulletin 16:1.
 Shand, Nina. 1996. "The Internal Dynamics o f Small' Media Non-Governmental Organisations in
 South Africa." MA Thesis, Department of Sociology, University of the Witwatersrand.
 Shivji, Issa. 1976. Class Struggles in Tanzania. New York: Monthly Review Press.
 Shivji, Issa. 1986. Law, State and the Working Class in Tanzania, c. 1920-1964. London: James
 Currey.
 South Africa Foundation. 1996. "Growth for AH: An Economic Strategy for South Africa."
 Johannesburg.
 Southali, Roger. 1995. Imperialism or Solidarity? Cape Town: University of Cape Town Press.
 Sparks, Allister. 1994. Tomorrow is Another Country: The Inside Story of South Africa's Negotiated
 Revolution. Sandton: Struik.
 Sunday Independent. 1996. "Shilowa Ends Weeks of Restraint and Damns Government's Economic
 Plan."
 Sunday Times. 1996. "Shilowa's Bombshell: COSATU Rejects Macro-Economic Framework."
 Tordoff, William. 1974. Government and Politics in Zambia. Manchester: Manchester University
 Press.
 Turner, Richard. 1972. The Eye of the Needle: Towards Participatory Democracy in South Africa.
 Johannesburg: Ravan Press.
 Valenzuela, J. Samuel. "Labour Movements in Transitions to Democracy: A Framework For
 Analysis." Comparative Politics 21:4.
 van de Walle, Nicholas and Dennis Chiwele. 1994. "Democratization and Economic Reform in
 Zambia." Zambia Democratic Governance Project, Monitoring and Evaluation Studies,
 USAID/Zambia.
 von Holdt, Karl. 1995. "The LRA Agreement:'Worker Victory1 or'Miserable Compromise'?" South
 African Labour Bulletin 19:4.
 von Holdt, Karl. 1991. "Towards Transforming SA Industry: A 'Reconstruction Accord' Between
 Unions and the ANC?" South African Labour Bulletin 15:6.
 Webster, E. 1995a. "The True Challenge to Tripartite Deals is Yet To Come." Business Day (20 July).
 Webster, Eddie. 1995b. "NEDLAC-Corporatism of a Special Type?" South African Labour Bulletin
 19:2.
 Webster, Eddie, and Dot Keet. 1992. "National Economic Forum: Parallel to CODESA?" South
 African Labour Bulletin 16:3.
 Wood, Brian. 1988. "Trade-Union Organisation and the Working Class." In Zimbabwe's Prospects:
 Issues of Race, Class, State and Capital in Southern Africa, edited by Colin Stoneman. London:
 Macmillan.
 World Bank, 1996. "Privatization in Africa: The Zambian Example." Findings, Economic
40
 Management and Social Policy 72 (October).
 World Bank. 1989. Sub-Saharan Africa: From Crisis to Sustainable Growth. Washington, D.C.: The
 World Bank.
 ZCTU. 1996. Unpublished statement on the 9th Quadrennial Congress of the Zambian Congress of
 Trade Unions. Kitwe.
 Interviews:
 Carolus, Cheryl. Acting Secretary General, ANC. Johannesburg, July 1996. (Webster and K. Gostner)
 Carrim, Yunis. ANC Member of Parliament. Cape Town, June 1995. (Webster)
 Chirwa, Alec. General Secretary, ZCTU. Kitwe, Zambia, December 1996. (Adler)
 Civil Servants Union of Zambia and National Union of Public Service Workers executive members.
 Lusaka, December, 1996. (Adler)
 Confidential interview. Ministry of Finance, Lusaka. December 1996. (M. Ginsburg)
 Confidential interview, South African Communist Party. Johannesburg, April 1997. (Adler and
 Webster)
 Confidential interview. Zambia Privatisation Agency. Lusaka, December 1996. (Adler)
 Erwin, Alec. Minister of Trade and Industry. Johannesburg, July 1997. (Webster)
 Godongwana, Enoch. General Secretary, National Union of Metalworkers of South Africa.
 Johannesburg, August 1995. (Gostner)
 Makgetla, Neva. Deputy Director General, Department for Public Service and Administration.
 Johannesburg, May 1997. (Adler and Webster)
 Nxesi, Thulas. General Secretary, South African Democratic Teachers' Union. Johannesburg, July
 1995. (Gostner)
 Olckers, Chippie. Director of Local Government Finance, Department of Constitutional Development.
 Johannesburg, June 1997. (Adler and Webster)
 Thobejane, Neil. General Secretary, National Education, Health and Allied Workers' Union.
 Johannesburg, August 1995. (Gostner)
 Vavi, Zwelinzima. Assistant General Secretary, COSATU. Johannesburg, September 1995 (Gostner)
UNIVERSITY OF THE WITWATERSRAND
 THE AFRICAN STUDIES SEMINAR SERIES
 Monday the 10th of April 2000, 7th floor Richard Ward Building, room 7003 @ 16:00
 The Neglected Role of Labour Relations in
 the South African Public Service
 By
 Glenn Adler
 Department of Sociology and Sociology of Work Unit
 University of the Witwatersrand
INTRODUCTION1
 As a consequence of South Africa's transition to democracy, public service workers now
 enjoy trade union and collective bargaining rights for the first lime in the country's history.
 These changes provide public servants with opportunities to bring their conditions of service
 into line with industrial relations 'best practice' in the private sector, and for black workers in
 particular, to redress decades of racism, employment insecurity, and low pay.
 Yet these reforms occur at a moment of straightened economic circumstances, which
 constrain possibilities for wage gains and employment growth. Moreover, the public service
 is undergoing massive restructuring, not only to overcome legacies of racial inequality and
 authoritarianism, but to transform the state into a more effective agent of development. The
 problems - under conditions of austerity - of addressing workers' wage demands and
 concerns for job security while redressing the legacies of apartheid will likely generate
 considerable conflict between workers and managers as well as between different groups of
 workers.
 These difficulties exist in many areas of the South African economy. But labour-management
 conflict in the public service is not a narrow industrial relations issue of interest to the parties
 alone. The public service is, with more than 1,1 million employees, the largest single
 employer in the country. Through its procurement policies it is the largest single consumer of
 goods and services. The wages and benefits it provides have a massive impact on effective
 demand, particularly in provinces where public servants make up a large proportion of total
 employment. Through its direct expenditure on social services and development, and
 indirectly through the economic effects described above it is the most important single agent
 of reconstruction. Finally, for vast numbers of citizens the public service is the provider of
 last resort: its actions literally determine whether people live or die.
 Thus labour relations conflicts in the public service will have crucial and long-lasting
 consequences for restructuring the state as an institution, and therefore for the government's
 ability to promote fiscal integrity, economic development, and service delivery. Such
 conflicts will thus inevitably spill beyond the boundaries of the public service to have an
 important impact on all citizens. But labour relations are nol simply a 'problem' to be
 avoided at all costs, for example by limiting public servants' right to organise or strike, or by
 exempting the service from labour relations legislation. If labour-management conflicts point
 to the existence of a problem, they also hold out the promise of a solution. Negotiated
 agreements between government and organised labour may provide means for finding
 constructive solutions to the tangle of difficulties currently confronting the public service.
 Public service labour relations is thus an important factor shaping the prospects for
 development in South Africa, and with it the fate of democratisation itself. But the possibility
 of developing modern labour relations practices in the public service, given its legacies of
 racially based paternalism and authoritarianism, and given fiscal constraints is a tall order.
 These practices do not emerge automatically as a consequence of the legislative inclusion of
 the public service in the Labour Relations Act and the extension of substantive and
 procedural rights to the sector. In the private sector these rights evolved over decades of
 1
  This is the introduction to a new collection. Public Service Labour Relations in a Democratic South Africa
 (Johannesburg: Witwatersrand University Press, in press), edited by the author. The book is the product of a
 collaborative research project sponsored by the National Labour and Economic Development Institute. Thanks
 arc due to Barbara Adair and Jeremy Baskin for detailed comments on an earlier version.
intense conflict, and remain imperfect, unevenly applied, and subject to reversals. It is
 difficult to imagine that the growth of labour relations in the public service will be any less
 fraught. However, if public service labour relations go wrong, the consequences could be
 profound for all South Africans. Thus the public service will be perhaps the single most
 important labour relations arena in the future.
 Yet the subject is almost entirely unresearched in this country. Much of the existing
 published material is descriptive and technical, oriented towards training officials in the rules
 and procedures of the public service itself. While such texts have their role, they are
 insufficient for preparing unionists, public servants, students and academics to confront the
 key issues in a public service undergoing transformation. The chapters in this collection are a
 first attempt to come to grips with this difficult but fascinating area of investigation.
 WHAT IS THE PUBLIC SERVICE?
 The public service is part of the broader public sector, and though the terms are often used
 interchangeably, they may refer to very different things. As this book focuses on the public
 .service, it is important at the outset to define the term.
 The public sector refers to all institutions public control where direct or indirect political
 control is exercised through appropriate legislation. In this light, the public sector includes a
 wide range of components: national and provincial departments, state-owned enterprises, and
 local authorities. These differ in terms of political control, relationship to the national budget,
 the method of wage determination, employment numbers and conditions, and the form of
 regulation and ownership. (See Figure 1.1)
 Figure 1.1: The South African Public Sector
 Government
 Adminis-
 trative
 agencies
 Service
 agencies
 Security
 Agencies
 Statutory
 agencies
 National Departments and Provinces
 Commer-
 cial Service
 State-Owned
 Enterprises
 Univer-
 sities &
 techni-
 kons
 Agricul-
 tural
 market-
 ing
 Boards
 Para-
 statal
 institu-
 tions
 ?
 Local
 Govern-
 ment
 Local
 para-
 statals
 Other
 local
 institu-
 tions
 Local Authorities
Public sectors vary greatly in scope and organisation between countries, depending on their
 own historical experience, politics, and needs. Moreover, in the last two decades the very
 nature of the public sector has undergone change under the influence of new market-based
 ideologies of governance and in the face of political shifts that have favoured right-of-centre
 political coalitions. As Waghorne points out in his chapter in this collection, there are no
 fixed definitions of either the public sector or - more narrowly - the public service that apply
 everywhere and for all time.
 In South Africa the public sector as a whole employed nearly 2 million workers in 1996.
 (CSS, 1997) It included:
 the public service, defined as national and provincial departments, which was the
 single largest component of the public sector, accounting for more than two-thirds of
 all employment;
 state-owned enterprises, such as Transnet and Telkom, which accounted for another
 15% of employment;
 local authorities, including city councils, municipalities, town councils, and the
 former regional services councils, which together employed approximately 13% of all
 public sector workers;
 and other components, including universities and tcchnikons, marketing boards, and
 associated institutions which accounted for less than 5% of total public sector
 employment. (See Figure 1.2)
 Figure 1.2: Public sector employment, December 1996
 U n ive r s i l i e s &
 T e c h n i k o n s - 3 .1
 S l s l e - 0 w n o d
 E n l e r p r i s e s ? 1 5 . 4 %
 P a ra s la to I
 I n s t i t u t i o n s - 1 . 3 %
 L o c a l A n I h o r i t i e s ?
 1 2 . 5 %
 A g r i cu l lu r o l
 M a r k e t i n g B o a r d s
 0 .1 %
 P u D l i c S e r v i c e
 6 7 . 6 %
 This book examines only the first and largest of these four components: national and
 provincial state departments. The emphasis means that some important dynamics in other
 areas of the public sector are ignored, such as privatisation of state-owned enterprises, local
 government restructuring, or the transformation of tertiary education. On the other hand, the
 narrower definition brings a number of advantages. By treating the public service as a single
 'industry' (albeit one divided into different sub-sectors) covered by common legislation, the
discussion is focused on relations between a single 'employer' and a limited range of unions.
 Both the employer and the unions bargain in the same institution: the Public Service
 Bargaining Council was established in 1993 and reorganised in 1997 as the Public Service
 Co-ordinating Bargaining Council (PSCBC). The approach not only simplifies discussion of
 a complex subject, but allows for clear comparisons and for somewhat stronger
 generalisations to be drawn.2
 Prior to 1994 the South African public service was complicated by the division of the country
 into 15 distinct administrations serving 11 different 'governments' which included the 10
 apartheid-era 'independent states' and 'self-governing territories.'3 The 1993 Interim
 Constitution of the Republic of South Africa, which came into effect following the first
 democratic elections in April 1994, eliminated these apartheid-era creations. It established a
 single public service for the country as a whole composed of a central administration and the
 administrations of the nine newly created provincial governments.
 As can be seen from Figure 1.2, the public service is the single largest component of the
 public sector. It includes national departments, provinces, and statutory agencies. It also
 covers educators employed in terms of the National Policy for General Education Affairs
 Act, 1984; personnel employed under the Correctional Services Act, 1959; and police
 personnel employed under the South African Police Service Act, 1995. The public service
 also includes the South African National Defence Force, the National Intelligence Agency,
 and the South African Secret Service.
 In 1996 total employment in the public service was nearly 1,3 million, though this had fallen
 to slightly more than 1,1 million by the end of 1998. (Fora detailed breakdown of public
 service employment see Baskin, Chapter 7) The vast majority of these workers were
 employed in the service delivery and security agencies.
 Since April 1994, greater clarity has emerged on the division of powers and functions
 between national government and provincial government and this has affected the location of
 public service workers. As both the interim and final constitutions defined many service
 delivery functions - such as education and health - as provincial competencies, the majority
 of public service workers are in fact employed in the nine provinces.
 LABOUR RELATIONS IN THE SOUTH AFRICAN PUBLIC SERVICE
 In focusing on labour relations this collection examines the range of interactions between the
 employer and organised workers in the public service. In the main these interactions focus on
 2
  Our choice should not suggest any privileging of the role of the public service within the broad public sector; it
 is rather a practical choice defined by a series of research imperatives. Given the general shortage of reliable
 analysis of the public sector as a whole, we felt it wise to cover somewhat less ground, but to do so in greater
 depth. We hope that the completion of systematic studies of state-owned enterprises and local government will
 someday allow for a comprehensive assessment of the public sector in its entirety.
 3
  These included the central administration of the Republic of South Africa and the four provincial
 administrations of the Transvaal, Orange Free State, Natal, and the Cape Province; the administrations of the
 'self-governing territories' of QwaQwa, KwaZulu, Gazankulu, Lebowa, Kwandebele, and KaNgwane; and the
 administrations of the 'independent states' of Transkei, Bophuthatwana, Venda, and Ciskei, commonly known
 as the TBVC states.
remuneration (wage and non-wage) and conditions of service (including job grading and
 training, discipline, dispute resolution, affirmative action). However in both practice and law
 this relationship extends to a number of issues well beyond the employment relationship,
 including involvement in policy making in areas such as public service transformation.
 For most of the twentieth century the public service was one of the most inhospitable sectors
 for worker organisation. The public service as a whole was excluded from the ambit of the
 Labour Relations Act (LRA), a condition that undermined collective action by all public
 servants, white or black. The very notion of'labour relations' scarcely applied: no public
 servants, white or black, enjoyed the right to join a registered trade union and bargain
 collectively with the employer. They did not enjoy the right to strike, nor were they covered
 by mutually accepted procedures for discipline, retrenchment or grievances. Instead white
 public servants benefited from the range of discriminatory practices that served all white
 workers: citizenship rights, job-reservation, and social welfare. Moreover they were
 enmeshed in a deeply paternalistic system founded on patronage within Afrikaner
 nationalism. By contrast, black public servants faced the barrage of restrictions confronted
 by all black workers: the denial of basic citizenship rights, exclusion from the industrial
 relations legislation, repression by the security forces, employer hostility. In addition until the
 1970s they were largely excluded from the racialised patronage networks within the state,
 except in the black homelands. While the wave of unionisation that washed over South
 Africa in the 1970s and 1980s swept away much of the employer unilateralism that
 characterised the workplace in the private sector, the public service remained an island where
 the employer's power reigned supreme.
 Since 1912 the Public Service Commission administered all aspects of the employer-
 employee relationship, including the determination of wages and conditions of service,
 discipline, and grievance handling. (See Macun and Psoulis, Chapter 4, in this collection) It
 unilaterally determined the nature of work through the determination of a Public
 Administration Standard (PAS) for each occupation, and developed a welter of regulations
 and legislation governing employment. No issue was too large or too small to be determined
 by the centralised PSC.
 White permanent civil servants (almost exclusively male) gained a measure of collective
 representation with the establishment of the Public Servants Association (PSA). But in line
 with the apolitical and professional public service ethos instilled by the British colonial
 model of administration, this was a restrained body that played a limited consultative role in
 the Public Service Joint Advisory Council. On the rare occasions when the PSA pursued an
 issue of importance to its members it could be ignored by the PSC.
 Yet the PSA's constituents did not lack reasons for complaint. As Posel argues in her chapter
 in this collection, white public servants were continually frustrated by the PSC's below-
 inflation rate wage increases, sanctioning of unfair dismissals, and deaf ear to employee's
 grievances. Curiously the PSC's arrogance never provoked the PSA to take collective action,
 and for Posel, the explanation is bound up with the system of patronage created by the NP..
 Labour relations under apartheid
 The NP politicised the PSC in efforts to remake the public service into an instrument for
 realising its apartheid policy, neatly reversing the long-standing demands for public service
 neutrality it expressed while in opposition before 1948. Instead the NP treated the public
service as a vast pool of patronage for party loyalists, even if this meant appointing
 incompetent whites unable to find employment elsewhere in the expanding post-war
 economy.
 In restructuring the state to pursue apartheid policies the public service was itself remade in
 apartheid's image. Not only did this entail the notorious creation of separate administrations
 for the TBVC states and self-governing territories, but it also led to a flurry of regulations to
 entrench the position of whites within the public service. These provisions contributed to
 duplication of work, the proliferation of occupational classes, a bewildering array of bonuses
 and allowances, and a massive wage gap marked by high levels of inequality on the basis of
 race and gender.
 The legacies of this political approach continue to structure employment relations in the
 public service as these very regulations were preserved during the transition. The legislative
 drafters of the LRA - largely ignorant of conditions in the public service - relied on the
 advice of senior bureaucrats who secured the conversion of all existing public service
 regulations into collective agreements. As a result, the same officials who had written these
 regulations to promote their own interests - such as favourable voluntary severance packages,
 extensive non-wage benefits, and generous leave provisions - ensured that these would be
 virtually enshrined in the new order.
 If NP hegemony created infinite ways to deliver benefits to Afrikaners, it also bound public
 servants in a subordinate position to the party and state elite. Most civil servants could not
 easily find work in the private sector, and were equally vulnerable should the impediments to
 recruitment of black public servants posed by job reservation be relaxed. As Posel argues,
 they knew 'that they depended on the party for their jobs, which could be revoked just as
 easily as they were supplied. The pressure for conformism and compliance were thus
 enormous.' These tendencies are clearly revealed Garson's and Marks' chapters in this
 collection, which describe the authoritarian culture common in schools and the police. In both
 cases promotion - if not survival - depended on fulfilling the commands of one's superiors.
 Public service labour relations, according to Poscl, were characterised by an increasing gap
 between the formal rationality of the bureaucratic PSC and the NP's party political project.
 Nor was the NP keen to resolve this tension. As PW Botha's 'total strategy' gathered
 momentum from the late 1970s the NP could use patronage to maintain loyalty while
 simultaneously deploying a public service discourse to discipline avowedly right-wing civil
 servants bent on undermining reforms.
 In this context the NP had little interest in pursuing the Wiehahn Commission's
 recommendation to extend labour rights to public servants. In the face of the historic upsurge
 in union organisation in the private sector during the 1980s the PSC ignored the
 Commission's advice and chose instead to extend the PSA's 'diffident and polite' 'non-
 adversarial' (in Posel's words) model to other public servants on a racial basis. Thus the
 1980s witnessed the launch of a number of relatively toothless staff associations defined by
 race, occupation and geographic location.
 The apartheid public service under pressure
 The decline of paternalistic unitarism in the public service occurred in tandem with the
 transition to democracy. In the second half of the 1980s the new union movement began at
last to make inroads into the public service. Union advances led by affiliates of the Congress
 of South African Trade Unions came first in those state departments most implicated in the
 state-society conflicts beginning in 1984. As state institutions became targets of township
 protest, increasing popular pressure came to bear on the black public servants responsible for
 services in black areas: teachers, police and soldiers, health workers. These workers had their
 own complaints accumulated as a consequence of years of discriminatory policies produced
 by the PSC. Now many were being asked by mass movement organisations to take sides in
 the growing civil conflict, for example by observing stay-away calls or embarking directly on
 industrial action, as in the 'chalk down' protests by teachers or the rolling mutinies by
 soldiers in 1989. (See Garson, Chapter 9 and Marks, Chapter 10)
 In these actions, black public servants' material grievances - apartheid on the job - merged
 with their own and their communities' wider discontent with NP rule. As elsewhere in the
 economy this yielded an explosive combination that produced militant social movement
 unionism. The budding unions in the public service were from the outset highly politicised, in
 the first instance because they were the only formations in the labour movement that faced
 the apartheid state both as oppressor and as employer. Where in the past this combination put
 black public servants in a highly vulnerable position and retarded union development, under
 conditions of mass mobilisation and widespread conflict with the state it transformed the
 public service into perhaps the most volatile labour sector in the country. With the growth of
 the mass democratic movement during the emergency and in the defiance campaign after
 1988 it became difficult for the state to pursue any strategy of unilateral reform from above:
 labour protest had transformed the state itself into an unreliable instrument of NP rule.
 Such volatility was heightened - not diminished - by the unbanning of political organisations
 in 1990 and the onset of negotiations over South Africa's constitutional future. Ordinary
 labour repression was made virtually impossible once the NP had committed itself to a
 negotiated transition and free political activity was restored. The signing of the Laboria
 Minute in 1990 between the government, COSATU, the National Council of Trade Unions
 and the South African Consultative Committee on Labour Affairs (a representative business
 body on labour relations issues) committed all parties to fundamental revision of labour law
 which included the long-delayed recognition of public servants' labour rights within a single
 Labour Relations Act.
 COSATU's efforts to establish a union beachhead in the public service date from 1987, with
 the founding of the National Education Health and Allied Workers' Union (NEHAWU).
 However these organising efforts bore little fruit until 1990. In education the federation
 played a central role in drawing together the disparate staff associations and more radical
 teachers' organisations that formed in opposition to the official bodies. The South African
 Democratic Teachers' Union (SADTU) was launched in 1990, uniting teachers across the
 disparate provincial and racial education authorities. In 1989 and 1990 police and
 correctional services officers staged a number of dramatic strikes, workplace occupations,
 and protest marches, and launched the Police and Prisons Civil Rights Union (POPCRU),
 which though it began as an independent union joined COSATU in 1994. (See Garson,
 Chapter 9 and Marks, Chapter 10)
 These three unions were largely responsible for the biggest strike wave in the history of the
 public service between 1989 and 1993. As Macun and Psoulis write in their chapter in this
 collection, many of those who participated were not union members at the time, but joined
 after taking part in the strikes. The issues motivating the strikers were in nearly all cases the
same: wages, union recognition, solidarity with dismissed or harassed workers, and
 discriminatory treatment. There were few strikes in the civil service, still the preserve of the
 PSA and the other old-guard staff associations, and weakly organised by the COSATU
 affiliates. But civil servants did take action in the homelands, whose own repressive labour
 laws became a major target.
 These strikes finally yielded a new labour relations dispensation, which saw core labour
 rights extended for the first time to public servants as a whole, with separate dispensations for
 educators and police personnel.4 The legislation allowed for recognition of public service
 unions and collective bargaining, created dispute resolution procedures, and established a
 bargaining council for the public service with separate councils for teachers and police. (See
 Patel, chapter 6 and the chapters by Garson and Marks)
 The new public service unions demonstrated considerable militance stemming from their
 double protest against apartheid and against the apartheid state as employer. During the
 transition his led to a second dual role: using their industrial muscle to block unilateral state
 restructuring (which would be deleterious to their members and would hamstring a future
 African National Congress government) and to push for transformation of the public service.
 In the context of the transition these interlinked imperatives made the new public service
 unions into the affiliates closest to the ANC.
 Under these circumstances the new public service unions grew extremely rapidly after 1990.
 However after 1994 these conditions would also produce a strange inversion. Under
 successive ANC-led governments the politicised unions would find themselves under
 pressure to conform to the policy orientations of their political ally. At the same time, the old
 staff associations would transform themselves from the 'diffident and polite' organisations of
 the apartheid era into more effective representatives of their members' interests.
 LABOUR RELATIONS IN A NEW DEMOCRACY
 Power changed hands in April 1994 to the ANC-dominated Government of National Unity.
 However, daunting obstacles stood in the way of developing a public service, in the words of
 the Reconstruction and Development Programme (RDP), 'capable of and committed to the
 implementation of the policies of the government and the delivery of basic goods and
 services to the people of the country.' (ANC, 1994:126)
 These obstacles posed problems for the government and the unions alike, notwithstanding
 that in the few years since public service workers won labour rights they registered their
 greatest gains ever. However these advances occurred in an increasingly contradictory
 context, in which the fruits of democratisation for workers (improvements in wages and
 conditions of service, job security, bargaining power, involvement in decision-making) would
 increasingly run up against competing imperatives of boosting expenditure on services,
 transforming the state itself, and maintaining fiscal discipline. These tensions were
 J
  Public Service Labour Relations Acl, 1993; the Education Labour Relations Act, 1993; and the South African
 Police Service Labour Relations Regulations, 1993.
 3
  The Government of National Unity came into existence following the elections and was comprised of the
 ANC, the NP and Inkatha Freedom Party (1FP).
temporarily contained with the adoption of a landmark three-year collective agreement in
 early 1996, but intensified with government's adoption of the Growth, Employment and
 Redistribution program (GEAR) in 1996. With the expiry of the agreement in 1999 the
 contradictions erupted into the first major conflict in the public service since the early 1990s.
 Transformation
 Given that its reason for existence was the promotion of apartheid, in both its culture and in
 its very institutions the public service was ill-suited to serving the development needs of the
 majority of the population. The NP's political projects shaped who was hired and promoted,
 where they were deployed and the character of their jobs. Most notably, the public service
 inherited by the ANC was strongly structured by a racial and gender division of labour
 resulting from decades of preferential hiring of white male Afrikaners. The priority given to
 party-political appointments meant that many public servants at all levels were hired for their
 loyalty rather than their ability to perform their jobs.
 Moreover, the apartheid project shaped the very institutional integument of the public
 service. It was characterised by an uneven racial/geographic distribution of its services, the
 racial duplication of entire - incompatible and financially wasteful - administrative systems,
 and the preponderant power of the coercive institutions of the national security state.
 Finally, the PSC's centralised rule-making power not only meant that modern labour relations
 practices were non-existent, but also made the public service highly inflexible. Ordinary
 workers were of course subject to its authority in the determination of wages and conditions
 of employment, but so, too, were senior public servants. Under PSC-rule senior officials were
 not managers in the conventional sense of the term: decision-makers with responsibility for
 the outcomes of their choices. Instead they were reduced to the status of administrators
 carrying out functions defined by the PSC within fixed regulations over which they had little
 influence. At both the top and the bottom of the public service hierarchy initiative,
 innovation, responsibility (and, therefore, accountability) were largely absent.
 Given these legacies, transformation was as much about destruction as creation: not simply
 grafting new institutions and practices onto the old, but an ambitious program to change the
 public service's purpose, structure, and personnel composition. As a result, it was
 quintessentially a labour relations matter as the remaking of the public service simultaneously
 meant the reconstruction of workers' jobs, their relations with other employees, with their
 superiors and with the public they were expected to serve. From this point of view
 transformation entailed overcoming racial and gender imbalances; a commitment to
 education and training to develop public servants' skills, eliminating discrimination in
 salaries and benefits, and changing the public service's authoritarian culture and outmoded
 work practices.
 Under the terms of the new LRA and in the common understanding of both government and
 labour, these changes were to be brought about through new processes of labour relations.
 On the one side these involved allowing workers' to organise and bargain freely with
 managers; on the other it required a change in the nature of authority, breaking the centralised
 power of the PSC and transforming administrators into managers endowed with decision-
 making power and responsibility. In short it entailed replacing the paternalistic model that
 had grown under apartheid with a system of labour-management bargaining that had
 developed in the private sector since the 1970s.
Given the enormity of transformation it is remarkable, looking back on the mid-1990s, that
 little explicit attention has been paid to its deeply contradictory character.6 It was occurring in
 a politically sensitive milieu defined by 'sunset clauses' negotiated in the transition talks
 which made it difficult to replace the senior old-regime personnel who were transformation's
 very targets.7 It was expensive: the enormity of redirecting public services to those who were
 once systematically denied such benefits entailed infrastructural and personnel expenditure of
 monumental proportions. Even assuming the goodwill of all parties and sufficient finances,
 these were - and remain - enormously complicated technical tasks: misguided policy and
 outright mistakes would be expected, and would inevitably cause delays. Transformation was
 rendered all the more difficult as service provision was to be maintained even as the
 instrument of delivery was undergoing change.
 Finally, transformation generates its own conflicts, not only between employer and
 employees, or between old and new or white and black public servants, but within and
 between these categories as well. Transformation creates new winners and losers, and spawns
 new collectivities and alliances as well as new adversarial relations. These will not always
 coincide with the cleavages inherited from the apartheid public service, and will likely cut
 across established patron-client relationships. Indeed, in some cases they may create new
 ones. Given these difficulties associated with transformation, it is remarkable that it has not
 provoked either widespread breakdown or all-out civil war within the public service.
 A strong case could be made that peace was maintained as a result of labour relations, which
 yielded improvements in wages and conditions of service, job security, and a relatively
 strike-free environment for most of the life of the Government of National Unity. However
 the means used to avoid war will not be readily available to the parties in the future.
 Dcmocratisation: Worker rights and collective bargaining
 The Public Service Labour Relations Act (PSLRA) came into operation in August 1993
 followed by enabling legislation in the education and police services sectors. These separate
 pieces of legislation were promulgated despite the agreement in the Laboria Minute to
 include the public service within the general labour relations dispensation. In part the
 legislation emerged from an immediate industrial crisis in the public service, which could not
 be systematically addressed outside of an appropriate legislative framework Yet the LRA
 itself was scheduled for fundamental renegotiation immediately following the installation of a
 new. government. The PSLRA was thus an interim measure to create, virtually from scratch,
 appropriate bargaining structures and powers, as well as agreements about which issues
 would be bargained where, and appropriate representation from the employer and workers.
 The Public Service Bargaining Council (PSBC) was established under the PSLRA and
 covered workers employed under the Public Service Act, excluding educators. According to
 (
 ' Though the Presidential Review Commission on the public service acknowledges the existence of'trade-offs'
 between transformation priorities, it does not begin to suggest what these should be, nor how they should be
 determined. (PRC, 1998: 2.1.4) Yet this acknowledgement is an important step forward over the earlier White
 Paper on transformation (DPSA, 1995), which did not even identify the problem.
 7
  The ANC's concession on 'sunset clauses' granted a measure of job security to public servants from the old
 order.
 10
Palel, both employer and employee representatives were weak and divided. The PSC - which
 had previously administered the public service unilaterally - was unable to adjust to the
 exigencies of bargaining, while its negotiators lacked basic labour relations skills. Employee
 representatives were similarly inexperienced and divided into two blocks, one composed of
 old-guard staff associations, the other of progressive unions led by NEHAWU. These
 problems entrenched adversarial ism in the PSBC. Agreement on a settlement for 1993/94
 was interrupted by the F.W. de Klerk government's unilateral implementation of an across-
 the-board (and below inflation) 5% increase. Nonetheless, a number of agreements were
 struck, including a wage settlement for 1994/1995.
 Union membership increased rapidly, and in June 1999 stood at 981 816 out of 1 016 440
 employees, giving an extraordinary union density figure of more than 96%.8 (PSCBC, 1999)
 SADTU provides a good example of the rapid growth: its membership increased from 40 000
 in 1993 to more than 210 000 in 1999, an increase of better than 420%. (See Figure 1.3
 below; for complete union membership figures see Macun and Psoulis, Chapter 4)
 Figure 1.3: SADTU membership, 1993-1999
 ?iOUOUU
 200000 ?
 lolXJUU
 100000 ??
 50000 ;
 0
 1993
 Source
 L ?
 1994
 : NALEDI
 r
 f.
 1995 1996
 (1999), p. 9
 ... _ _._
 - - ? - -
 --
 1997
 . . . .
 1998 1999
 Though union density and growth rates are impressive, these do not provide an accurate
 index of union strength. Given the much more permissive labour law environment after 1994
 and the relatively union-friendly orientation of government it would have been surprising if
 unions did not flourish. Their achievement is not, in this respect, exactly comparable to the
 organising gains registered by many private sector unions in the labour-repressive context of
 the 1980s.
 8
  The PSCBC excludes from its coverage members of the South African National Defence Force, the National
 Intelligence Agency and the Secret Service. However, a recent Constitutional Court decision granted members
 of the SANDF the right to join a trade union, though the court was silent on the extent of rights soldiers would
 enjoy. (Constitutional Court, 1999)
 11
Nor does density indicate the unions' degree of organisation. By contrast, in the 1970s and
 1980s many private sector unions grew comparatively slowly, but on a factory-by-factory
 basis where they laid down strong shop floor organisation. Their emphasis on democratically
 elected shop-steward structures not only gave these unions a deep pool of leadership, but
 provided unions with the capacity to respond to members' needs and the ability to embark on
 disciplined action in pursuit of their interests. (Friedman, 1987; Adlcr and Webster, 1995)
 For the most part unions in the public service have not grown in this fashion, and workplace
 and intermediate-level organisation is generally weak. This is certainly the case among the
 old staff associations, which never professed to operate as representative trade unions, but
 also remains true of COSATU affiliates, despite their federation's long-standing commitment
 to shop steward organisation and leadership.
 As Marks reveals in Chapter 10 on labour relations in the police service, neither the South
 African Police Union (SAPU) nor POPCRU have effectively translated into practice their in-
 principlc commitment to developing shop steward structures. Effective organisation at shop
 floor and intermediate levels remains elusive and is not encouraged by the still highly
 centralised character of both public service bargaining and administration. The unions'
 capacity problems at these levels may compromise their ability to bargain in newly created
 provincial bargaining councils, and their ability to interact with management at all levels.
 Finally, membership does not imply organisational loyalty. There is a marked tendency
 towards multiple union membership - where workers join one union to obtain benefits and
 another for political reasons - informally estimated at 80,000 or close to 10% of thetotal
 membership in the public service. This phenomenon has been relatively unknown in the
 private sector, where for black workers in particular union membership historically entailed a
 risky act of identifying with the unions' goals. 'Agency shop' provisions further dilute the
 meaning of union membership. Unions now receive 'agency' fees from public servants who
 are not members, but who nonetheless benefit from many of the collective agreements
 negotiated in the PSCBC. High agency shop fees induce workers to join a trade union,
 whether or not they identify with its goals.
 Taken together, these factors suggest that high union density in the public service does not
 mean that the organisations possess effective power in their interactions with the employer.
 Their priority will need to shift to converting their massive numbers into organisational
 strength, especially as there is no room for further growth except through mergers or by
 poaching from other organisations. Those who fail to convert quantity of membership into
 quality of organisation may well find themselves on the defensive against management and
 the victims of their more successful rivals' poaching efforts.
 The three-year agreement
 In the 1994 bargaining round in the PSBC unions collectively put forward a set of demands
 that combined the old-guard staff associations' interest in a high overall increase and the
 progressive unions' interest for uplifting those at the bottom. The position called for a Rl 500
 minimum monthly salary and an across-the-board 15% increase. The unions expected that
 these demands would be met, given their membership growth, a new labour relations
 dispensation and the perception that the new government would be amenable to improving
 public servants' conditions of service.
 12
Government rejected the proposal, and offered slightly more than a Rl 000 per month
 minimum wage, with a freeze on salaries at the high end. The offer was rejected by all the
 unions, many of whom threatened to walk out of the talks. Government then offered a
 variable increase, starting at 20% for employees in the lowest grade, falling to 5% at the top,
 but the PSBC nonetheless deadlocked. The impasse was broken through high-level political
 intervention by then-deputy presidents Thabo Mbeki and F.W. de Klerk, and a wage
 agreement was reached only after commitments were made to create task teams to conduct a
 comprehensive investigation of the public service salary structure. The investigation was not
 completed in time for the 1995 round of negotiations, but the process was accelerated by a
 wave of wage strikes in the private sector and a particularly fierce wildcat strike among
 unorganised public service nurses.
 The disparate positions of unions and government were eventually consolidated in a three-
 year agreement which not only went beyond the unions' original 1994 demand, but awarded
 the largest wage increase in the history of the public service, if not of the country as a whole.
 According to Baskin, (Chapter 7) the three-year agreement made a significant improvement
 in the minimum wage rate, which increased from R13 200 to R17 100 per year. The largest
 increase went to the lowest paid: Grade 1 workers received a healthy 29,5% increase, Grades
 2-6 all obtained increases higher than 35%. (Jacklin and Machin, 1998: 14) No employee
 received an increase of less than 7,5%.
 These increases - as large as they appeared - were made even more substantial by changes to
 the 'rank and leg' system of promotion. Where these promotions normally occur on an
 individual basis every two or three years, in 1996 all employees on Grade 1 were
 automatically advanced to Grade 2, whether or not they qualified on the basis of a
 performance review.
 Plans for further increases in 1997 and 1998 were not realised, given problems in financing
 the agreement (see below). Still, in both of those years public servants received average
 increases at or above the rate of inflation, though improvements for those in the bottom
 grades were considerably higher, while senior managers' salaries were frozen.
 As a consequence the agreement made major inroads into the public service wage gap. This
 legacy of the apartheid public service was reduced from 64:1 in 1989 to 25:1 in 1995 (after
 the first negotiations between the unions and the new government), then to 17:1 after the first
 year of the three-year agreement, and finally to 15:1 in 1998. (See Bhorat, 1997 and Patel's
 chapter on collective bargaining in this collection) Closing the wage gap through an
 improvement in conditions of the lowest-paid was one of the COSATU unions' core
 demands, and reversed the long-standing emphasis in the PSC to reward those at the top at
 the expense of the mass of black unskilled workers.9 The three-year agreement also sought to
 simplify and extend the range of non-wage benefits available to public service workers. For
 example, women were ineligible for the public service housing subsidy, while many General
 Assistants (and other categories, including some teachers) did not receive pensions and I3 lh
 checks as they were considered casuals. (See Baskin, Chapter 7)
 The hallmark of the three-year agreement was the link between wage improvements and
 9
  For example, De Klerk's above-mentioned unilateral 1993 5% wage implementation included a 20% increase
 for senior managers.
 13
broader restructuring and transformation of the public service. (See Patel, Chapter 6 and
 Baskin, Chapter 7) Improved conditions for those at the bottom were a key element of this
 vision, as were a reduction in the unwieldy occupational structure to 16 grades, reducing
 wage differences between grades, introduction of a new skills-based grading system, and a
 measure of job security through a moratorium on employer-initiated retrenchments. The new
 grading system aimed at overcoming the existing complex and arbitrary occupational
 hierarchy, which had been shaped by imperatives of maintaining the racial and gender
 hierarchy of the public service, rather than by an objective evaluation of job content or
 workers' skills. It was meant not only to break the apartheid division of labour, but to provide
 workers with career paths that could deliver advancement and further real improvements
 while developing the skills base of the public service.
 The agreement's increases were to be financed through 'right-sizing' the public service.
 According to Baskin, the parties agreed in principle to 'right-sizing': that 'the size of the
 public service be related to "available financial resources",' though the meaning of right-
 sizing remains contested. For government this meant reducing the public service wage bill as
 a proportion of public consumption expenditure, while the unions stressed that employment
 levels must be determined by service delivery requirements. In the context of a moratorium
 on retrenchments 'right-sizing' was to be achieved through a management plan and a skills
 audit to provide an accurate account of existing employment and skills levels and a link
 between these and service requirements. In addition, the agreement emphasised freezing or
 abolishing vacancies, a program of voluntary severance packages (VSPs), attrition, and in the
 last instance, retrenchment of'supernumerary' officials, though only after all possibilities for
 'redeployment' had been exhausted. Finally, the parties agreed to restructure pension
 benefits, both to make early retirement more attractive to long-serving public employees, and
 to achieve savings through a reduction in the state's contribution from 18% to 15%.
 The high-water mark of worker participation
 Not only did the post-1994 public service rapidly develop the familiar institutions of pluralist
 bargaining, but both unions and government alike shared a commitment to developing deeper
 forms of worker involvement in decision-making. (See Patel, Chapter 8) The workplace
 forum provisions in Chapter 5 of the new LRA stimulated important debates in South Africa
 on the merits of co-determination: the development of forms of consultation and joint
 decision-making between management and employees. However, Chapter 5 was not the
 stimulus for the extension of worker participation in decision-making; indeed, the public
 service was in fact exempted from the relevant sections of the new act.
 To a significant degree the interest in public service co-determination was impelled by
 temporary institutions and practices developed during the transition. Progressive unions made
 a concerted effort to check the unilateral prerogatives of the apartheid state by promoting the
 creation of broad transformation forums in public institutions in which a measure of power
 would be shared out among stakeholders. Many of these innovations survived the transition
 and provided models for new forms of public governance which were prominently featured in
 the White Paper on the Transformation of the Public Service. (DPSA, 1995) The White Paper
 envisioned worker representatives becoming directly involved in fundamental policy
 formulation as well as in implementation and monitoring at the enterprise level.
 At many levels prospects for worker involvement in decision-making were far greater in the
 public service than elsewhere. Central bargaining in the PSCBC not only gives unions a
 14
Voluntary Severance Packages
 VSPs were perhaps the most glaring failure. They were meant to cut public service
 employment levels while freeing up funds to pay for the wage and non-wage increases. They
 did nothing of the kind. According to the Presidential Review Commission, very few posts
 vacated by VSPs were in fact abolished; most were filled by new appointees. In some cases
 the very same public servant who received the VSP returned through the 'back door' as a
 consultant, though this practice violated the severance terms. As a result, 'although the long-
 term savings ... are ... likely to be much smaller than predicted, the short-term costs have been
 much higher ...' (PRC, 1998, 3.2.4.2)
 The VSPs shed neither 'dead wood' nor undesirable 'old guard' employees, but rather
 provided a lucrative escape for public servants with marketable skills. These departures were
 felt particularly sharply in health and education. Crucially, the VSPs did not yield anywhere
 close to the kinds of savings that were expected. As such savings were to be a source for
 funding the general wage increases granted in 1996 and projected for 1997 and 1998, the
 failure of VSPs put intense pressure on the fiscal sustainability of the entire agreement.
 The retrenchment moratorium
 The VSPs are not the only examples of unintended consequences issuing from the three-year
 agreement. A similar fate befell the moratorium on retrenchments, seemingly a major union
 victory. According to Baskin (Chapter 7), the moratorium did not prevent the loss of nearly
 170 000 jobs or more than 13% of the workforce as public service employment fell from 1,25
 million in 1996 to less than 1,1 million at the end of 1998. Ironically the cuts fell most
 heavily on the constituencies organised by the COSATU unions that fought most vigorously
 for the moratorium. They were heaviest at the bottom grades, NEHAWU's membership,
 though there was also an 8,5% decline in the police. These changes - along with the rehiring
 necessary to fill vacancies opened by VSPs - had a dramatic effect on the skills profile of the
 public service. The total number of jobs in the lowest grades (1 to 5) declined by over 19%
 while those in the middle category (6 to 9) increased by over 7%.
 As the decline in police numbers shows, the cuts did not follow delivery or development
 priorities. Instead they appear to have been motivated by managers' desires to defend their
 budgets by cutting back on the unskilled beneficiaries of the wage increases and extension
 and equalisation of benefits. However, employment growth in the relatively expensive
 middle-range categories put upward pressure on the wage bill and offset any savings from the
 job losses at the bottom. Declining employment delivered neither cost savings nor a closer fit
 between policy priorities and personnel. It was the worst of both worlds for both unions and
 government: neither good 'right-sizing' nor cost-saving 'downsizing'.
 Financing the agreement: Lack of alignment between the budget and bargaining
 Given these developments, it is not surprising that the fiscal basis for the 1996 increases
 quickly eroded. Total personnel costs as a percentage of the budget (excluding interest
 expenditure) increased from 46,6% in fiscal year 1995/96 to an estimated 50,6% in fiscal
 year 1999/00. The situation was worse in the provinces, where personnel spending rose from
 53% of provincial budgets in fiscal year 1995/96 to more than 59% in fiscal year 1998/99. It
 is this rapid increase in recurrent expenditure that provoked government demands for a
 16
powerful platform, but the scope of issues bargained extends beyond conventional
 distributive issues. According to Patel (Chapter 8) many unionists believe they have the
 ability to raise a wide range of issues in central bargaining and through transformation
 structures at the institutional level, and hence have little need for the comparatively modest
 powers available in Chapter 5 of the LRA. Moreover sectoral bargaining councils in
 education and the police allow for consideration of broader policy questions in those areas.
 Given the improvements in wages and conditions of service, involvement in policy-making
 and implementation, and the moratorium on retrenchments, it is not surprising that the public
 service was one of the least conflictual sectors in the country from 1994 to mid-1999.
 (Andrew Levy and Associates, various) Where the public service had been perhaps the most
 strike-prone sector in the years 1989 - 1993, with health workers, police, and teachers
 embarking on lengthy strikes, from 1994 this changed dramatically. Except for a prominent
 wildcat strike by nurses, largely in Gauteng and the Eastern Cape, there were no major strikes
 in the public service during this period, and comparatively few person-days were lost due to
 strikes.10
 The nurses' strike ironically demonstrated the efficacy of public service labour relations.
 Their actions were prompted by pent-up frustrations over wages and working conditions,
 especially the increased workload caused by the extension of free health services in 1994.
 Moreover, public service nurses were unable to express these demands due to the absence of
 an effective nurses' union. Far from a repudiation of the emerging labour relations
 dispensation in the public service, the strikes were an unfortunately clumsy effort by one
 occupational group to find a way of processing their demands through the system. In the
 aftermath of the strike most nurses did join unions, and participated in subsequent bargaining
 rounds in the PSCBC. There has been no repeat of the 1995 wildcat strike.12
 Assessing the three-year agreement: Quick fixes, deferred problems
 If the three-year agreement contributed to labour peace, it was unable to produce the
 conditions for sustainable transformation. Each of its key elements came undone.
 111
  This does not imply that no time was lost due to protest actions, demonstrations and stoppages, as opposed to
 formal strikes. For example, SADTU used these quite effectively in their winter 1998 resistance to the
 Department of Education's plans to terminate the contracts of temporary teachers. Such disruptions have been
 prevalent at the provincial and at enterprise levels - particularly in education - and have a definite, though
 impossible-to-calculatc impact on person days lost. Nonetheless, in the first five years of ANC rule there were
 few - if any - national or even sectoral strikes in the public service - a marked improvement over the early years
 of the transition to democracy.
 " These frustrations intensified after the South African Municipal Workers' Union (SAMWU) concluded a
 collective agreement which provided impressive increases to nurses working in municipal clinics, who often
 work in close proximity to those employed by provinciai,governments. In comparing SAMWU members'
 improvements with those provided in the PSBC - where nurses' interests were poorly represented - provincial
 nurses quickly discerned the gains possible through collective action, though not the difficulties involved in
 developing collective organisation. (See Adler, 1995)
 12
  This aspect of the nurses' strike has been widely misunderstood. For example Adam et al. (1997) claim the
 nurses struck 'against their own union', though at the time the vast majority weren't organised. Moreover, they
 take the event as evidence of'the inability of emerging corporatist arrangements to cope with a situation
 bordering on anarchy' rather than as an example of the inevitably messy stages in the evolution of these very
 arrangements. (See Adam et al, 1997: pp. 150-151)
 15
review of pay policy and for ending the retrenchment moratorium, proposals that would
 guide their position in the 1999 wage dispute. Increasing personnel costs were said to be
 'crowding out' other spending.
 For both sides this problem highlighted a larger issue: the lack of alignment between the
 government's budget process and the annual wage negotiations in the PSCBC. This problem
 would intensify with the adoption of the Medium Term Expenditure Framework (MTEF)
 from 1997. Under the MTEF budget planning would be projected forward in three-year
 cycles, largely fixing expenditure on the public service wage bill well in advance of PSCBC
 negotiations. As a result, collective bargaining would become less about the size of the wage
 bill and more about its division. From the unions' perspective this fundamentally undermined
 collective bargaining.
 A skills based grading system?
 If the sustainability of the wage increases came under scrutiny, so too did the supposed
 transformation in the basis for determining remuneration. The three-year agreement was
 announced as a fundamental shift towards a skills-based grading system, borrowed directly
 from developments in the private sector, particularly in the metal industry. It was on this
 foundation that other union-initiated remuneration reforms were premised: flattening the
 occupational hierarchy by reducing the number of job grades, narrowing differentials within
 and between grades, and closing the overall wage gap. The new system promised to provide
 workers' with opportunities for training and career - and pay - advancement.
 In practice the agreement ushered in no such thing. In a detailed study of public service
 wages and grading, Jacklin and Machin (1998) argue that the three-year agreement failed to
 include other initiatives on which a skills-based grading system depends: training, changes to
 work organisation arrangements and a reformed performance appraisal system. The
 inadequate opportunities for training meant that workers - especially at the bottom - would
 have limited or no opportunities for advancing within the system, and would likely be
 consigned indefinitely to their existing broad band.
 The consequences of failing to move to a real skills-based grading system were made worse
 by changes to the 'rank and leg1 promotion system. 'Rank' promotions refer to advances
 between employment grades, while 'leg' promotions refer to promotions within a grade.
 Public servants are normally eligible for promotion every three years, or at a minimum every
 two years if they are deemed 'preferentially promotable.' The benefits of promotion are
 substantial: rank promotions can result in as much as a 20% improvement in income. In the
 main, advancement within a broad occupational band (composed of a number of ranks)
 depends neither on a vacancy nor on funding for the post, nor in practice does it require
 meaningful assessment. Though in principle workers must be deemed 'promotable', in
 practice almost no one is held back. The 'rank and leg' system amounts to a regular bonus for
 employees unrelated to either individual performance or a rational assessment of the public
 service's skills requirements. Indeed, if a higher grade position is unavailable, workers will
 be paid at the higher rate even while continuing in their old job. The implications for the
 fiscus and for 'right-sizing' are thus considerable. The 1996 agreement in fact made the
 problems worse through the mass promotion of Grade 1.
 Recent figures from the South African Police Service illustrate the perverse effects generated
 by the promotion system (see Table 1.1). The fact that there are nearly five times as many
 17
sergeants and twice as many inspectors as constables is a reflection more of police officers'
 desire for and success in achieving promotion to senior positions than of the country's
 policing objectives and of the labour requirements necessary for achieving these. Nor do the
 figures reflect the police officers' skills nor the real content of their jobs: in cases where
 officers have been promoted to the rank of sergeant, they may in fact be working as
 constables: only their title - and salary - have changed.
 Table 1.1: Selected ranks and salaries in the South African Police Service, as at
 January 2000
 Rank
 Constable
 Sergeant
 Inspector
 Number of employees
 11,217
 55,312
 25,620
 Number of employees
 34,557
 40,881
 63,042
 Source: Unpublished statistics, Department of Public Service and Administration.
 In addition to rank and leg promotions, workers may also apply for promotion between the
 broad occupational bands - a 'post' promotion - but success depends on the availability of a
 vacant funded post and workers must possess appropriate qualifications. Grade 1 workers -
 including general labourers - may have been promoted en masse to Grade 2, but in the
 absence of proper training programs, they have few opportunities to gain the qualifications
 necessary for a post promotion into Grade 3 jobs, such as clerical positions. In general, the
 same applies to more skilled workers applying for promotion from Grades 7 to 8 - from
 clerical to administrative officer positions - though such advances are relatively easier than
 for labourers attempting to move up.
 In short the changes to the grading system, in the context of'rank and leg' promotions,
 produced destructive results. Not only is promotion expensive, but it remains delinked from
 either worker performance or the public service's employment needs while - in the absence of
 proper training programs - workers are effectively consigned to a life within their broad
 occupational bands. As Table 1.1 indicates, changes to the promotion system would not only
 allow a better fit between employment, remuneration and needs, but could in fact finance an
 expansion in employment at lower levels.
 The three-year agreement was signed two months before the launch of the government's new
 Growth, Employment and Redistribution (GEAR) macroeconomic strategy. And while it is
 probably true that the deal could never have been struck after GEAR, the strategy itself was
 not responsible for the three-year agreement's contradictions. Even before GEAR the
 government had been committed to a reduction in expenditure as a proportion of GDP, and to
 a reduction in the public service wage bill as a proportion of total government spending to
 make resources available for infrastructure development and other services. This put
 government in a very tight box for achieving its aims of transforming the public service. The
 equalisation of conditions of service could come about either through maintaining
 employment levels while reducing wages and benefits or improving remuneration while
 cutting numbers. Both the unions and government accepted the latter formulation, though
 with a priority on attrition, freezing posts, eliminating ghost workers, and reprioritisation of
 expenditure. Expansion of the wage bill would put pressure on other forms of government
 expenditure, in the absence of an increase in borrowing or increases in revenue through
higher taxation or other means, such as privatisation. But these steps are not only politically
 difficult, but fiscally problematic in the context of an untransformed state unable to spend
 effectively those resources it already commands.
 Given this context the three-year agreement was poorly designed: it was inadequately costed
 (Baskin, Chapter 7), while additional elements crucial to its success (like adequate training
 programs or a means to effect sustainable personnel reductions) did not materialise. Many of
 its possible consequences, such as the likely effect of VSPs - which were obvious to
 observers at the time - were not addressed. Its authors were generally unaware of the
 complexity of many public service systems, and entered agreements with little understanding
 of their likely effects. ANC officials were in fact dependent on the advice of'old guard'
 senior civil servants, who often ended up drafting new regulations, but with more concern for
 protecting their own prerogatives than for promoting transformation.
 1 f the three-year agreement is viewed as a way of transforming government under tight
 constraints, it must be the seen as a failure. However, it may be judged differently if viewed
 as a crisis management exercise adopted with imperfect information under extreme pressure
 following two years of relatively unsuccessful bargaining in the PSBC. Judged against these
 criteria it was a qualified success - certainly in terms of ensuring labour peace - though at the
 cost of avoiding tough choices and deferring the core problems that originally motivated the
 parties to find a solution.
 From agreement to conflict: 1999 and beyond
 If the negotiators in the PSBC were able to defer long-term problems, the participants in its
 successor body, the Public Service Coordinating Bargaining Council, had no such luxury.
 Not only were the defects of the 1996 bargain well known, but negotiations for a new
 agreement began in late 1998 in the midst of South Africa's worst economic crisis of the
 decade.
 The unions' initial common position was for a 15% across-the-board increase, while
 government responded with an offer of an average 5,7% increase. This massive gap remained
 through the announcement of the budget in March. At a PSCBC meeting at the end of April
 the government proposed a task team to review salary progressions, but would determine for
 itself whether to increase its offer based on the task team's recommendations. At this point all
 the unions in the PSCBC declared deadlock.
 In July the dispute was referred to arbitration, but within days the government improved its
 average offer from 5,7% to 6%, with employees in the lowest grades receiving 6,2%, and
 those at the top 3,5%. At the end of July, following a one-day strike led primarily by the
 COSATU unions, government's offer was improved to an average of 6,3%, with a 6,8%
 increase for teachers. By this time the unions had reduced their demand to an average of
 7.5%, with a somewhat higher increase for teachers. (PSA, 1999; Fraser-Moleketi, 1999)
 In early August a number of unions again embarked on a legal strike - supplemented by
 public demonstrations organised by unions that had stayed out of the earlier protest. This was
 an unprecedented development in public service labour relations: the first time that
 progressive and old-guard worker organisations pursued a common program of industrial
 action. In fact, the strike marked the transformation of many of the old-guard unions into a
 more robust form of unionism. To an extent they have begun to shed their diffidence and
 19
politeness, and have made use of the new possibilities in labour law for exercising power.
 At the same time, the unions linked to COSATU were caught between their commitments to
 their base and their participation in the alliance with the ANC. The tripartite alliance yielded
 many benefits in the past when the COSATU unions were able to break deadlocks through
 appeals to the party, and especially to President Mandela. In the run-up to the elections
 alliance discipline effectively diverted COSATU unions from pressing their case in the
 PSCBC or in the streets. And after the 1999 elections, the alliance route was effectively
 closed; indeed, leading figures in the ANC began attacking public servants' profligacy,
 labelling the unions as special - and selfish - interests. Indeed, informal efforts to broker a
 compromise were fruitless, and on 7 September government took the extreme action of
 unilaterally implementing its final offer - a step taken previously in the short history of public
 service labour relations only by F.W. de Klerk in 1993. (Fraser-Moleketi, 1999a)
 In issuing its unilateral increase government offered to 'work' with the unions on the 2000/01
 round of salary negotiations, and to develop a comprehensive remuneration policy. However,
 the invitation to return to a forum in which the unions had expended considerable energy in
 return for a minor improvement on the employer's offer - which was then unilaterally
 imposed - must have seemed eminently resistible. Indeed, at the time of writing the unions
 continue to boycott the PSCBC, on the grounds that the wage deadlock continues - and some
 of the unions have pursued the matter through arbitration.
 A 'score card' approach to assessing industrial conflict is seldom useful, but is even less
 appropriate in this case. Notwithstanding that the employer imposed its solution on labour,
 neither party's positions' addressed the fundamental questions facing the public service on
 which their own agendas depend: effectively linking budget, staffing levels, remuneration,
 and delivery priorities. This failure will have important consequences in the future.
 Given the enormity and difficulty of these questions it is unfortunate, but not surprising that
 the bargaining turned on the issue of wages rather than the unresolved transformation
 matters. Informants close to the PSCBC negotiations indicate that an agreement was in fact
 forthcoming on a new remuneration policy, including fundamental changes to the rank and
 leg system. However, given the obvious attractions the old policies hold for public servants,
 unions rightly calculated that they could not sell the agreement in the absence of a real wage
 increase.
 Explaining government's refusal to compromise is more difficult. In part its position was
 informed by the dynamics of the June 1999 elections. It effectively delayed proceedings in
 the PSCBC until after the poll. Then, as a result of the ANC's overwhelming mandate for a
 further five year term - with a near two-thirds parliamentary majority - government was in a
 politically unassailable position from which it could substantially increase its autonomy. The
 ANC likely concluded that the public service unions would be unable to mount a strike that
 could impose sufficient costs to force government to reconsider its position. Given the co-
 ordination between new and old guard unions COSATU's strategic and tactical choices were
 limited by the need to maintain a common front with partners untested in mass action. In
 addition, its choices were restricted by the exigencies of maintaining public support, which
 would not be easily forthcoming if the unions were seen to be disrupting 'delivery' in their
 fight against a popular government. Both of these conditions - government's unprecedented
 freedom of manoeuvre and COSATU's equally unprecedented constraints - no doubt
 reinforced the position of those in and outside of government committed to a strict
 20
interpretation of GEAR and the maintenance of fiscal discipline. The PSCBC negotiations
 enabled these interests to steal the line of march on the trade unions and 'hold the line' on
 wages.
 However, government's resolute stance on wages creates a number of substantial problems
 on its side. First, its position defers consideration of restructuring for yet another year.
 Meanwhile, the animosity generated by the dispute creates its own dynamics that make it
 more difficult to come to terms with these issues, when in the future - as is virtually
 inevitable - the parties must seek each other out to implement meaningful transformation of
 public service labour relations. The problems arc qualitative, not merely quantitative, and
 depend on agreement between the parties; they cannot be resolved through victory in a labour
 relations 'war.'
 KEY ISSUES IN PUBLIC SERVICE LABOUR RELATIONS
 It is difficult, in the current climate of mutual recrimination, to discern a means that would
 bring the parties back to good faith bargaining in the PSCBC. Yet it is equally difficult to
 perceive a route for restructuring the public service that does not pass through the PSCBC.
 On every dimension of transformation the parties arc faced with balancing competing
 imperatives that have been mutually acknowledged. There are few 'win-win' choices, and
 even fewer instances where the options can be rationally or neutrally determined. What are
 the key issues?
 The budget
 Since the beginning of formal labour relations in the public service a fundamental - and still
 unresolved issue - has been aligning negotiations over wages and non-wage benefits with the
 budget. Government draws up the budget beginning in the year prior to presenting it in
 parliament, in consultations between the Department of Finance, national departments, and
 provincial governments. Now that the Medium Term Expenditure Framework is in place,
 these discussions occur within three-year rolling budgets that have to a major extent fix
 allocations. The budget is then presented to parliament where its allocations may be
 interrogated.
 At its heart the issue concerns two competing and difficult-to-reconcile principles.
 Government draws up budgets, and parliament is supreme: it alone has constitutional power
 to approve money bills. On the other hand wage agreements in the public service depend
 upon good-faith bargaining between parties in the PSCBC not only over the allocation of
 funds available for salaries and benefits, but over their total amount. As personnel
 expenditure is by far the largest single item in the budget, even small changes can have a
 massive effect on government's ability to meet its other obligations.
 A bargaining process that generates agreements with little regard to these wider obligations
 would either upset other public priorities or would effectively render the agreement a dead
 letter. But the reverse is equally problematic, where the integrity of the budget trumps public
 servants' right to engage in collective bargaining, not only about the allocation of
 expenditure, but about its overall size. At present this is the case. In 1999 the budget
 effectively capped the amount available on personnel expenditure; the unions were unable to
 persuade government to change the allocation significantly nor could they marshal effective
 21
power to force it to do so.
 This is a common problem in public services in every democratic society. However,
 particular circumstances in South Africa add to the difficulty. GEAR targets limit room for
 manoeuvre, but these are not the origin of the difficulty. The overwhelming imperatives of
 service delivery and infrastructure development place burdens on government that for the
 foreseeable future will put pressure on the size of the public service wage bill. Attempting to
 overcome the legacies of the apartheid division of labour while at the same time retaining
 skilled public servants makes it difficult to trade off the needs of one employee constituency
 against another. Moreover, given the single-party dominant political system and the
 leadership's tight discipline over the parliamentary caucus, public service unions can exert
 little pressure on government to change the budget, either through lobbying or through the
 electoral process.
 There is no easy or logical answer, but the present arrangement is unsatisfactory: it has
 produced repeated conflicts since 1994 that are wasteful of time and resources and generate
 considerable ill-will. Furthermore it generates perverse consequences for government's own
 plans for public service restructuring. As Baskin notes in Chapter 7, limiting bargaining to
 dividing up a pre-set amount will tend towards majority rule at the PSCBC:
 and this is likely to take the form of'across-the-board' increases, or larger increases at
 the lower end of the salary scale. In the medium-term this will lead to erosion of the
 bargaining unit as lower-end employees are replaced by contracting out arrangements,
 and as higher-end . . . employees attempt to leave the bargaining uni t . . .
 A way forward is offered by the Medium Term Expenditure Framework (MTEF), though this
 will severely test both parties. The MTEF enables all citizens to learn government's long-
 term budget priorities. Such transparency for the first time allows for accountability, as
 government can now be held answerable for its performance. A major problem arises,
 however, if the MTEF's three-year commitments are perceived to be imposed from above or
 are unresponsive to needs.
 The MTEF includes - at least in rhetoric - mechanisms for popular participation in its
 formulation. Bui in practice such participation is de-emphasised by government, whose
 priority has been to establish executive authority over the budget process and to insulate it
 from external participation. (Gelb and Bethlehem, 1998) Such insulation effectively extends
 to parliament, notwithstanding the invocation of the 'supremacy of parliament.' Arguments
 that prioritise parliament's supremacy as against public service unions' involvement in the
 budget obscure the fact that parliament's budget-making powers are severely limited. Though
 the constitution gives parliament the power to amend money bills, such powers must be
 activated by a separate act - which has not been passed. This state of affairs 'renders the
 [parliamentary] budget process largely symbolic as the budget is determined wholly by the
 executive with Parliament having little or no say in its contents.' (Adair and Albertyn, 1999)
 The test for government will be whether it will tolerate even minimal forms of involvement
 in developing MTEF priorities - including the allocation for personnel expenditure - from
 civil society groups, including the public service unions. This would be in line with the
 values expressed in the White Paper on the Transformation of the Public Service and the
 Presidential Review Commission report, both of which stress civil society partnerships and
 empowering citizens to share the responsibilities of governance. (PRC, 1998: 3.1.2, 3.2.3)
 22
At the same time such participation would present a major challenge for the public service
 unions. It would require formulation of initial negotiating positions a year in advance,
 development of sufficient economic research capacity, and the development of cadres to
 engage with the MTEF process at the many provincial and departmental sites where and
 when it occurs. It also requires sufficient sophistication to engage with state technocrats and
 other stakeholders and the will to press the matter if their approaches are rebuffed, through
 public campaigns if necessary. Short of such an effort, their wage claims will carry little
 sympathy if presented as a fait accompli after the budget has been tabled in parliament.
 Bargaining structures
 Public service collective bargaining structures grew in equal parts from improvisation and
 imitation. Educators and police were able to gain a measure of recognition as a result of the
 crisis precipitated in these sectors by the struggle against apartheid. Whether or not there was
 a legal basis to do so the authorities in these sectors were forced to address workers'
 collective demands simply to ensure the continuation of state services in these areas. (See
 Garson, Chapter 9 and Marks, Chapter 10) The separate existence of bargaining councils is
 thus a by-product of the conflict that attended democratisalion, reinforced by the fact that
 historically ministers in these areas have legal authority to set wages and conditions of
 service.
 But the structures also evolved in imitation of the private sector and the development of
 centralised bargaining elsewhere in the economy. The PSBC was created as a vehicle for
 bargaining at the highest level of the public service as a whole. Its creation was facilitated by
 the existence of the PSC; unlike other sectors of the South African economy the
 determination of labour issues in the public service was already very highly centralised. In an
 odd way the establishment of centralised bargaining could occur much faster in the public
 service than elsewhere, as it meant transforming the existing PSC into a truly bipartite
 structure.
 However, unlike almost any other sector, the public service was composed of but one
 employer and a multitude of distinctly different 'industries' that were further divided into 9
 different provinces and more than 30 national departments. The federalism implicit in the
 constitution further complicates matters: some departments of state - education, welfare,
 health - are joint competencies of the national and provincial governments. Most of the
 managers and workers in these areas are in fact employed by the provinces. To further
 complicate things, different provinces organise their affairs in different ways, combining in
 one department functions that are housed in separate national departments. It has been a
 daunting task developing bargaining structures that suit the contours of the public service,
 without compromising the coherence of the sector as a whole and centralised bargaining. The
 same can be said about the kinds of issues bargained and the levels at which they are
 determined.
 The model of industry-wide bargaining councils provided under the new LRA was crudely
 extended to the public service, creating an unwieldy mix of more than 40 provincial and
 departmental bargaining institutions, alongside the PSBC, the Education Labour Relations
 Council (ELRC), and the National Negotiating Forum (NNF). (Adair and Albertyn, 1999)
 The framework was not wanted by anyone, however a replacement was not immediately
 perceived.
 23
Notwithstanding the conflict in the PSCBC over wages, considerable convergence occurred
 over the shape of bargaining. (See Baskin, 1999) Unfortunately not all the issues were
 resolved, nor formalised in agreements, and progress was thwarted by the dispute. In 1999
 the old NNF was replaced by a new Safety and Security Bargaining Council (SSBC), and its
 membership was extended from police to all employees of the South African Police Service,
 including civil servants. In 1999 two other structures were established, the Health Sector
 Bargaining Council (HSBC) and the General Administrative Bargaining Council (GABC).
 The former includes all employees of.the Departments of Health and Welfare (national and
 provincial) as well as public health and health and welfare professionals in other departments.
 Given that health is a provincial competence, the HSBC intends to establish provincial
 chambers for negotiating more detailed agreements. The GABC includes all employees not
 covered by the other bargaining structures. The ELRC continues to exist, and has developed
 provincial chambers in each province.
 I f considerable progress was made on the structures of bargaining there was less agreement
 about where issues are to be bargained. The underlying issue is finding a balance between
 common standards across the public service and accommodating the specific needs of the
 different provinces and departments. If wages and conditions of service are negotiated solely
 at the PSCBC, the sectoral and provincial institutions will be largely irrelevant. On the other
 hand the complete devolution of bargaining to the sectors and/or provinces could lead to
 intolerable conditions: either growing inequality between provinces or competition in which
 sectors attempt to outbid each other in pursuit of skilled staff. Solutions include converting
 bargaining in the PSCBC into agreements over minimum and maximum wages and
 progression ranges, rather than actual wages, or allowing it to ratify agreements made in
 other bargaining councils.
 Restructuring management
 A strong system of labour relations in the public service depends not only on strong unions,
 but also on capable management. However, the existence of public service management - in
 the true sense of the word - is a post-1994 phenomenon, and even then more rhetorical than
 real. One of the main aspects of transformation has been overcoming the centralised structure
 of administration symbolised by the old PSC. As Adair and Atbertyn write in Chapter 5, this
 logically entailed the devolution of authority to lower levels of the public service and
 converting administrators into managers endowed with both power and responsibility.
 As mentioned in the previous section, the Ministers of Education and Safety and Security
 already possessed powers to determine terms and conditions of employment for educators
 and police. Amendments to the Public Service Act in 1996 extended these powers to national
 ministers and, at the provincial level, to Members of the Executive Councils (MECs), who
 were given original powers to organise their departments, including the hiring and dismissal
 of employees. This parallels other legislation and regulations that make 'executing
 authorities' responsible for financial matters in their departments. The PSC's rigid provisions
 for disciplining employees were replaced by a collective agreement reached in the PSCBC in
 January 1999, which now allows for discipline to be managed according to standard labour-
 management procedures in a flexible manner and at the appropriate level.
 However managers have not yet been given power to determine pay and conditions of service
 for their employees. These powers are still vested in the Ministers for Education and Safety
 and Security for educators and police and the Minister for Public Service and Administration
 24
for all other public servants. In practice these are determined through bargaining in the
 PSCBC. Even the new powers granted under the 1996 amendments to the Public Service Act
 arc limited. According to Adair and Albertyn 'executing authorities1 may determine the
 organisational structure of their departments:
 however they may not retrench surplus employees and they must organise their
 department within the prescribed nationally determined grading system, which is
 inflexible and does not take account of sectoral needs and differences.
 Though extending responsibility with one hand, the amendments take power away with the
 other, as executing authorities' scope for autonomous action is highly constrained by
 decisions at the highest level. Moreover, 'even within the departments and institutions, all
 decisions, no matter what they are, have to be made by the highest authority1 as executing
 authorities who do not readily delegate powers to lower levels of management in their
 departments.
 According to Adair and Albertyn, 'decision-making thus remains highly hierarchical with
 senior and middle management having very little decision-making power.' The state of
 affairs does not foster accountability and responsibility, nor does it allow for meaningful
 interventions by workers or their trade union representatives, except at the very highest
 levels. Notwithstanding massive efforts to transform the old public service, and the
 increasing importance of collective bargaining in place of the unilateral determinations by the
 old PSC, 'managers still do not "manage," but are responsible for ensuring that the
 employees abide by a complex set of rules and regulations, rather than concentrating on
 output and service delivery.'
 The extension of financial responsibility to senior managers under these constraints produces
 perverse effects. Managers must fit their decisions about personnel deployment into
 calculations for meeting their financial obligations within fixed standards on wages and
 conditions of service. Personnel decisions are thus oriented towards budget concerns instead
 of being rationally linked to service delivery needs. Management tasks are made even more
 untenable as retrenchment has been blocked by the three-year agreement's moratorium, while
 redeployment is difficult, and dismissal - even for cause - is so time consuming as to be
 virtually impossible. The present conditions discourage creativity and initiative, and break the
 link between spending, development priorities, and the deployment of personnel.
 But even if one could find an appropriate balance between national norms and standards and
 decentralised decision-making, this is not the most profound management problem. Far more
 difficult is overcoming the legacy of apartheid-era hiring practices that allowed vast numbers
 of unqualified people to become managers. A system that is able to develop centrally-
 determined norms and standards while encouraging managerial flexibility depends ultimately
 upon skilled managers. Indeed it demands highly sophisticated men and women who can
 understand and effectively maintain this delicate balance. These arc not - in the main - the
 kinds of managers that the apartheid public service bequeathed to the new order. Their lack
 of proper managerial training is most glaring in the areas of financial management and labour
 relations, which seriously weaken their effectiveness in linking budgets and personnel with
 policy objectives. (DPSA, 1997)
 These problems are of course uneven: some provinces are worse than others (especially those
 that absorbed former TBVC and SGT administrations), and the same is true of national
 25
departments. But a number of high-profile public scandals have brought the problem of
 public service management into sharp focus. Recent reports from the Auditor General have
 shown that many departments are forced to roll-over their allocated budgets because they
 cannot disburse money. The Departments of Welfare and Health, for example, were unable
 to spend substantial portions of their national funds for poverty relief and AIDS prevention.
 Allegations of wide-spread corruption in the Department of Correctional Services have
 revealed that practices of patronage common in the old regime persist into the new, albeit
 deployed along ethnic - rather than merely racial - lines. These three cases are extremely
 worrying as they occurred at the national level of these departments, where it is assumed
 managerial competence is greatest.
 Adair and Albertyn argue that new legislation will need to be passed that completes the
 transformation of management in the public service. Nonetheless the task of developing and
 training management to perform the responsibilities thrust upon it remains an overwhelming -
 and exceedingly long-term - priority of transformation. Good management is a necessary
 condition for good labour relations in the public service.
 Right-sizing
 As mentioned above, one unfortunate consequence of the 1999 wage dispute was to defer yet
 again progress on 'right-sizing' the public service. The unintended consequences of the three-
 year agreement's retrenchment moratorium and VSPs have still not been addressed, which
 has stymied efforts to align the size and deployment of staff with development priorities and
 the budget. Right-sizing still means different things to management and labour. For the
 former the priority has been reducing the public service wage bill as a proportion of public
 consumption expenditure, while for the unions employment levels are to be determined by
 service delivery requirements.
 However both parties - at least in principle - recognise the other's priorities. Cutting the wage
 bill to reduce public expenditure without regard to the government's development goals
 makes a mockery of the budget, transforming it into an end in itself rather than a means to
 achieve political goals. In principle this position is rejected by government in all its policy
 documents, not only those on the transformation of the public service. The Presidential
 Review Commission observed that right-sizing:
 should be located within the macro-economic constraints established by GEAR, but
 should nevertheless be driven primarily by the need to improve the efficiency and
 effectiveness of service delivery rather than by narrow budget considerations. (PRC,
 1998:3.2.7)
 The DPSA's recent report on its skills and service audit suggests that restructuring is a
 qualitative matter that does not easily produce savings, and that savings may come at the cost
 of affecting basic service delivery. (DPSA, 1999) The audit revealed that the public service
 as a whole is not overstaffed; rather many departments have inappropriate staff complements
 for their needs. For example provinces that incorporated old homeland bureaucracies have
 too many unskilled labourers, but a serious undersupply of higher-level managers,
 professionals, and support personnel. In another example, the audit showed that many
 township and ex-homeland schools suffer from a severe undersupply of support staff,
 essential to effective education. The report advocated an active redeployment strategy, rather
 than simple retrenchment (itself an expensive step). Moreover, the audit confirmed Baskin's
 26
findings in Chapter 7 that while unskilled employment had declined, the number of skilled
 (and higher paid) employees - especially teachers - had expanded.
 On the other hand, determining staffing by reference to service delivery alone - without
 reference to budget constraints - is even more problematic. Yel this position is rejected - at
 least implicitly - in a number of union policy documents. For example, the report of
 COSATLPs September Commission on the Future of the Unions advocates an ambitious
 program for 'building a public sector that can deliver efficiently and effectively, and improve
 the working life of the members of the public sector unions.' (COSATU, 1997: 3.2) For
 COSATU this entails 'taking some responsibility for co-managing transformation,1 by
 exposing corruption and ghost workers. At the same time, the approach acknowledges:
 there are limits on government resources, and unlimited needs. COSATU should
 reject the limits defined by GEAR, and continue to contest them. But even without
 GEAR, there would be limits. This implies trade-offs between different priorities.
 COSATU still believes that a 'bigger, more labour-intensive' public service is necessary, but
 accepts that 'Participation with government. . . would mean that whatever choices are made,
 even hard ones, the unions would be willing to take joint responsibility for their
 implementation.' A similar position is at least implicit in a recent assessment of the public
 service strike written by two leading NEHAWU unionists. Though they disparage the notion
 of'crowding out,1 they do not reject the idea of trade-offs between components of the budget.
 (Makhura and Phadu, 1999)
 There is thus - in principle - considerable overlap between positions. It remains to be seen
 whether the parties are able to reach agreements that resolve this key problem of
 transformation. And, even if agreements are reached at the PSCBC, it is questionable whether
 either management or labour have the capacity to implement and monitor these at sectoral,
 provincial, and institutional levels. Even if both sides find an accommodation that enables
 them to escape from the current impasse, neither has the capacity as yet to ensure compliance
 from their constituencies. Faced with an unreliable management echelon itself in need of
 change, management's victory in the 1999 bargaining round may prove short-lived: it will
 find it difficult to fight a two-front war with both labour and its own managers.
 This state of affairs does not augur well for public service transformation. It also suggests a
 third possibility: that many managers and workers will find common cause in defending their
 inherited privileges against the restructuring agendas advanced by both government and
 organised labour. In this scenario, poor management structures may permit both old- and
 new-guard civil servants to direct public resources towards maintaining their own positions.
 There are many disturbing signs - as in the Department of Correctional Services - that
 apartheid-era patronage not only persists, but has opportunities to flourish in a situation
 precariously poised between the paternalism of the old order and the rational labour relations
 of the new. There is a considerable danger that the contemporary public service will
 recapitulate the situation under apartheid where the formal rationality of bureaucratic rules
 become increasingly distant from the real paternalistic relationships between management
 and labour. The public service may have been brought under the LRA, but it has not yet been
 remade in its image.
 Faced with these possibilities, recent suggestions by ministers and senior officials that the
 public service be exempted from many provisions of the LRA and the Basic Conditions of
 27
Employment Act may in fact backfire. Rather than promoting management hegemony or
 even fiscal discipline, such steps might in fact encourage practices of patronage quite at odds
 with formal models of managerial control. Though many may reject the notion, an employer
 truly interested in transformation of the public service and service delivery may have much to
 gain from bolstering effective unions and developing sound labour relations practices.
 CONCLUSION
 In the few years since South Africa's first democratic elections, incredible progress has been
 made in the development of a modern system of labour relations in the public service. Not
 only have unions and bargaining structures been established, but these have reached
 successful agreements on a range of distributive and non-distributive matters. The system
 has, for the most part, produced labour peace in what was until recently a storm-centre of
 industrial protest. These are remarkable achievements, not only considering the distance the
 parties have had to travel from the racist and authoritarian past, but also considering the fiscal
 and skills constraints facing the public service.
 At the same time, the labour relations system has been more adept at deferring major
 problems than resolving them. In part this stems from the inherent difficulties of the issues. In
 part it is a result of mistakes, and the unintended consequences of agreements reached in
 haste and under political pressure. In 1999 the system broke down in a dramatic wage
 dispute, which saw government acting in a unilateral manner not seen since F.W. de Klerk's
 presidency. Yet the attention given to the dispute obscured the agreements made over the
 same period on the structure of bargaining, management, and other issues. It also masked the
 fact that the parties' principled positions are not significantly different and that they have a
 mutual interest in reaching an accommodation.
 This does not imply that the new labour relations system can deliver agreements on the
 outstanding issues. The private sector has had nearly three decades in which to develop
 labour relations institutions, and effective managers and trade unionists to make them work.
 The public service has to catch up in less time, while facing more daunting issues, higher
 expectations from its intended beneficiaries, and the full glare of publicity when things go
 wrong. The stakes are high: labour relations will either unlock or slam shut the door to public
 service transformation, and will help determine whether the state can fulfil its obligations for
 development. To a great extent the fate of democratisation in South Africa rides on the
 outcome.
 28
REFERENCES
 Adair, B. and Albertyn, S. 1999. 'Collective bargaining in the South African public service:
 The creation of sector-based bargaining.' Unpublished research report, NALED1 long-term
 research project on public service labour relations. Johannesburg: National Labour and
 Economic Development Institute.
 Adam, H., Slabbert, F., and Moodley, K. 1997. Comrades in Business; Post-Liberation
 Politics in South Africa. Cape Town: Tafelberg.
 Adler, G. 1995. 'Govt Response to Labour Unrest is Inappropriate.' Business Day, 4
 October.
 Adler, G. and Webster, E. 1995. 'Challenging Transition Theory: The Labor Movement,
 Radical Reform, and Transition to Democracy in South Africa,' Politics and Society 23, no. 1
 (1995): 75-106.
 ANC. 1994. The Reconstruction and Development Programme: A Policy Framework.
 Johannesburg: Umanyano Publications.
 Andrew Levy and Associates. (1992-1999). 'Annual Report on Labour Relations in South
 Africa.' Rivonia: Andrew Levy and Associates.
 Baskin, J. 1999. 'The structure of negotiations on salaries in the public service: An ?
 assessment.' Unpublished research report commissioned by the Department of Public
 Service and Administration.
 Bhorat, H. 1997. "Public sector raises core labour issues.' Budget Watch, v. 3, issue 3.
 Central Statistical Services (CSS). 1997. 'Labour statistics - Employment and salaries and
 wages: public sector (December 1996).' Statistical release P0251, 24 April. Pretoria:
 Government Printer.
 Constitutional Court of South Africa. 1999. 'South African National Defence Union versus
 Minister of Defence and others.' Case Cct 27/98. 26 May.
 COSATU. 1997. 'The report of the September Commission on the Future of the Unions to
 the Congress of South African Trade Unions.' Johannesburg: Congress of South African
 Trade Unions. -^
 DPSA. 1999. 'Report on service and skills audit.' Pretoria: Government Printer.
 DPSA. 1997. 'The provincial review report.* Pretoria: Government Printer.
 DPSA. 1995. 'White Paper on the Transformation of the Public Service.' Pretoria:
 Government Printer.
 Fraser-Moleketi, G. 1999. 'Press statement by Ms. Geraldine Fraser-Moleketi: Government
 tables new salary offer to avert strike.' Ministry for Public Service and Administration. 27
 July.
 29
Kraser-Moleketi, G. 1999a. 'Press statement: Public service salary negotiations come to an
 end.' Ministry for Public Service and Administration. 7 September.
 Friedman. S. 1987. Building Tomorrow Today: African Workers in Trade Unions, 1970-
 19X4 (Johannesburg: Ravan Press).
 Cielb, S. and Bethlehem, L. 1998. 'Macro-economics for the masses?1 Siyoya! no 1,
 Autumn.
 Jacklin, N. and Machin, M. 1998. 'The emperor's new clothes: The government's approach
 to grading and wage policy in the public service." Unpublished research report, NALEDI
 long-term research project on public service labour relations. Johannesburg: National Labour
 and Economic Development Institute.
 Makhura. D and Phadu, T. 1999. 'Lessons from Ihc public service wage dispute.' African
 ('otntnunisf No, 152, third quarter.
 NALEDI. 1999. Unions in Transition: Cosatu into the New Millennium. Johannesburg:
 National Labour and Economic Development Institute.
 Patch I. 1998. 'Understanding the public service labour market: Prerequisite for its
 transformation.' Unpublished research report. NALEDI long-term research project on public
 service labour relations. Johannesburg: National Labour and Economic Development
 Institute.
 PSA (Public Servants Association). 1999. 'In the arbitration between the Public Servants
 Association of South Africa (and others) and the Government of the Republic of South
 Africa/ Sandton, 18 October.
 PSCBC. 1999. 'Membership Statistics 1999.' Pretoria: Public Service Co-Ordinating
 Bargaining Council.
 PRC (Presidential Review Commission). 1998. 'Developing A Culture of Good Governance:
 Report of the Presidential Review Commission on the Reform and Transformation of the
 Public Service in South Africa." Pretoria: Government Printer.
 30
UNIVERSITY OF THE WITWATERSRAND
 INSTITUTE FOR ADVANCED SOCIAL RESEARCH
 Seminar Paper
 to be presented in RW 7003 SEMINAR ROOM
 AT 4.00pm 19th SEPTEMBER 1994
 Title: Of Shop Floors and Rugby Fields: The Social Basis of
 Auto Worker Solidarity.
 by: Glenn Adler
 No 366
Uitenhage, Residential and Industrial Areas, c. 1970
 BLACK RESIDENTIAL AREAS
 ^ COLOURED RESIDENTIAL AREAS
 .1 WHITE RESIDENTIAL AREAS
 INDUSTRY
I. INTRODUCTION*
 Athol Fugard, John Kani, and Winston Ntshona's 1972 play, Sizwe Bansi is Dead, opens with the main
 character, Styles, sitting in his photographic studio in Port Elizabeth's New Brighton township,
 reminiscing about working at Ford some years before. He recalls the time when the General Foreman,
 "Baas" Bradley informed the workers about special duties for the day: they were to conduct an
 unprecedented "General Cleaning" of the factory. For hours the men struggled to remove accumulated
 dirt and oil from the machines and the factory floor. They were then handed brushes and were ordered
 to paint lines on the floor indicating danger areas, a concern for safety not previously shown by the
 company. Finally, they were ordered to take showers, and were provided with hitherto unknown luxuries
 such as hot water, soap and real towels, and were then given new overalls, tools, aprons, and gloves.
 Finally "Baas" Bradley gathered the workers together to explain their unusual work regimen,
 using Styles as translator:
 "'Tell the boys in your language, that this is a very big day in their lives.'
 'Gentlemen, this old fool says this is a hell of a big day in our lives.'
 The men laughed.
 'They are very happy to hear that, sir.'
 'Tell the boys that Mr Henry Ford the Second, the owner of this place is going to visit
 us. Tell them Mr Ford is the big Baas. He owns the plant and everything in it.'
 'Gentlemen, old Bradley says this Ford is a big bastard. He owns everything in this
 building, which means you as well'....
 'Styles, tell the boys that when Mr Henry Ford comes into the plant I want them all to
 look happy. We will slow down the speed of the line so that they can sing and smile
 while they are working.'
 'Gentlemen, he says that when the door opens and his grandmother walks in you must
 see to it that you are wearing a mask of smiles. Hide your true feelings, brothers. You
 must sing. The joyous songs of the days of old before we had fools like this one next to
 me to worry about.' [To Bradley] 'Yes sir!'"1
 The great man's visit is an anti-climax: Ford peers in from the factory door, abruptly returns to his
 waiting car and drives off. But if the visit was over, the working day was not: management immediately
 ordered the line speed doubled, pushing the workers to make up production lost during the general
 cleaning. Styles, angry at the games he must play to retain such a job, resentful of being "treated like a
 monkey," decides to leave wage labor behind him, and opens up his own business.
 The fictional Styles opts out of factory life, and takes his chances in the lower petite-bourgeoisie
 of independent township entrepreneurs. The character is in part modelled on John Kani's actual
 experiences as a worker at Ford; Kani, of course, quit Ford to become one of the most accomplished
 actors in modern South African theater. But Styles' decision illustrates the limited choices available to
 black workers as a whole in the auto plants in the late 1960s. Blacks in general, and Africans in
 particular, received poor wages for performing the worst jobs, and faced virtually unbridled supervisorial
 despotism. For coloureds, the promise of union protection was still some years away, while Africans
 were to wait another decade for rights of union representation. It was only in the years after 1968 that
 African activists at Volkswagen even dared speak privately about unions. Opting out was not an act of
 opportunism, but of desperation; the only alternatives were either the false face of the smiling, minstrel-
 like puppet Styles was forced to become, or waiting patiently until conditions changed and some kind of
 union activity became possible again.
 'This paper is drawn from two chapters of my doctoral thesis, The Factory Belongs to All Who
 Work In It: Race, Class, and Collective Action in the South African Motor Industry, 1967-1986,"
 Columbia University, 1994.
In June 1980, barely a decade after the events portrayed in the first scene of Sizwe Bansi is
 Dead, Volkswagen workers embarked on a strike which could only have appeared as an unattainable
 fantasy to workers in the 1960s. In constructing their union, the workers provided themselves with the
 possibility of another option beyond the narrow range offered to Styles and his brothers. To opting out,
 selling out, or minstrelsy was added the choice of achieving individual aims by joining a collective force in
 pursuit of collective goals.
 The 1980 Volkswagen strike was a watershed in the development of the autoworkers' union.
 Moreover, the Volkswagen workers' mass action was all the more unusual because workers were
 successful: they maintained discipline and were able to marshal! their collective force with a negotiation
 strategy which could yield positive results. But most unusual of all was that their solidarity crossed the
 dividing lines defined by the government-imposed racial classification system: Africans joined with their
 coloured co-workers in industrial action and mass protest. The 1980 strike demonstrated a rare cross-
 race solidarity in contrast to unionizing efforts elsewhere in the country, including the one developing in
 the automobile plants down the road in Port Elizabeth, where workers fractured on lines of race, leading
 to years of bitter recriminations and internecine battles between unionists. Nowhere else in the motor
 industry had the union achieved such clear unity across racial lines. Despite its importance, such cross-
 race unity has been taken for granted in all accounts of the union and the labor movement itself. The
 basis of this solidarity is the subject of this paper.
 In doing so, it will - for reasons of space - deemphasize shop-floor and broader social and
 political factors giving rise to militance which were, roughly speaking, similar for both coloured and
 African workers. Instead, it will emphasize social processes which help to explain why such militance was
 expressed in a non-racial manner uniting coloureds and Africans. A number of important social
 relationships, shaped in large part by the legacy of the uneven application of apartheid in Uitenhage,
 bridged racially defined community divides and brought coloured and African workers into close personal
 contact at work and off-work. Through these relationships both ordinary workers as well as labor
 activists came to identify each other as familiars rather than strangers, enabling them to devise a politics
 of solidarity that defied the statutorially defined racial categorizations. Non-racial unionism did not arise
 simply because workers shared similar grievances that enabled them to act collectively, but in large part
 because particular social movement activists were able to help workers define their problems in a non-
 racial way, and to provide non-racial class-collective solutions for these problems.
 II. SOCIAL MOVEMENT EXPLANATIONS OF COLLECTIVE ACTION
 One of the great difficulties in explaining collective action in South Africa - as elsewhere - is the
 tendency to separate macro-level or structural explanations from more agency-based and voluntaristic
 accounts of political and social events. In terms of theories of social movements, most explanations of
 collective action fall within the sphere, respectively of theories of New Social Movements (NSM), and
 Resource Mobilization Theory (RMT).2
 NSM stresses larger structural processes which give rise to social movements and which shape
 the movements' ideological character. In large part, NSM sees social movements as a product of the
 transition from industrial capitalism and modernity to post-modernity and post-materialism. Not only do
 these movements take up qualitatively new demands, but they do so in a qualitatively different manner.
 Eschewing participation in established bureaucratic structures and institutions (usually, though not
 exclusively including party systems), these movements prefer forms of mass mobilization, direct action,
 and participatory models of democracy: a new form of politics supposedly characteristic of post-modern
 societies.3
 RMT, on the other hand, focuses more on the internal life of social movements. Writers in this
 tradition spend less time analyzing the effects of social and political structures on movements; based on
 the assumption that grievances exist in all societies at all times, they prioritize the way organizations are
 created to mobilize such potential.4 RMT thus examines the creation of social movement organizations
(SMOs), the activists/entrepreneurs who create and staff such organizations, and the resources they
 create and use in pursuit of their goals.
 Neither approach, however, does a satisfactory job of addressing the problem of structure and
 agency. According to Klandermans and Tarrow, NSM "too easily assumes that mobilization potentials
 form spontaneously through societal developments," ignoring the fact that social movements themselves
 play a role in shaping peoples' ideas which impel them towards mobilization. On the other hand, RMT
 focuses too narrowly on how social movements mobilize, but largely ignores how changes in social
 structure create potentials for mobilization.5 Stated another way, NSM too often tends to view collective
 action as a mysterious reflex "exploding," "emanating," or otherwise arising unexplained from structural
 conditions, while RMT may tend to emphasize activists creating their politics little constrained by
 external conditions.
 Klandermans and Tarrow argue that an integration of the two approaches is necessary to
 understand "how the potential for social movements which emerges from the social and political
 structure...is translated into social and political action." They hope to illuminate processes which
 intervene between structural change and the activities of social movements.6 Klandermans has
 developed the concept of "consensus mobilization" to account for social movements' efforts to help
 potential members interpret the grievances arising from structural change. Consensus mobilization will
 involve at the outset the movement helping people to identify the source of their grievance in systemic
 rather than personal terms, and to accept that systemic solutions are required. Furthermore, movements
 modify people's views of their grievances and, most importantly, what they can do about them;
 movements must facilitate what McAdam refers to as "cognitive liberation," a crucial step in motivating
 individuals to act.7
 Perhaps the most important way in which consensus mobilization occurs is through what
 McAdam has labelled the "micromobilization context," which he defines as "that small group setting in
 which processes of collective attribution are combined with rudimentary forms of organization to produce
 mobilization for collective action."8 Recruits to social movement activism are seldom drawn out of the
 blue, so to speak, but are instead already integrated into preexisting social networks conducive to
 mobilization.9 Such networks provide "rudiments of organization" for a nascent movement, such as
 leadership, communication infrastructure, and other means which are necessary conditions for action.
 "As preexisting groups," McAdam argues, "these contexts provide the established roles and lines of
 interaction necessary for action to unfold." More subtly, and more importantly, these micromobilization
 contexts "supply the established structures of solidary incentives on which most social behavior
 depends....the myriad interpersonal rewards that attend ongoing participation in any established group."10
 Social movement activists may "appropriate" the established associational networks and emotional ties to
 such organizations and transfer them to the new movement.
 McAdam's insight is to show how such micromobilization contexts link macro conditions
 favorable to mobilization to individuals' decisions to participate. Through the strong interpersonal ties
 already existing in micromobilization contexts, social movement activists are in a better position to
 produce "cognitive liberation." In such contexts potential recruits to the movement are not meeting
 strangers, but activists whose credentials are previously established in the group. Not only is such trust
 important, but recruits may also feel loyalty to the group and to leaders which facilitates their acceptance
 of ideas, and this may provide a subtle pressure against deviation. Thus, while previously developed
 attitudes favorable towards the movement may "push" a recruit towards activism, "a prior history of
 activism and integration into supportive networks acts as the structural 'pull' encouraging the individual
 to make good on his or her strongly held beliefs."11
 The organizations and networks McAdam identifies do not have to be political in nature, nor do
 recruits have to be members of organizations. Micromobilization contexts may be formal organizations,
 formal or informal communication networks, and interpersonal networks. They may occur in all spheres
 of life: religious, school, youth group, or sporting bodies, as well as in ethnic, gender, and community
 groups. What is important is that through such prior engagement recruits will have already been
 introduced as part of apparently "normal" (i.e. natural) and "acceptable" activity to movement-like
organizational practices and rudimentary forms of collective action. In this way the existing, relatively
 non-controversial collective fabric of society provides the foundation on which activists build social
 movements.
 o The concept of micromobilization context helps to explain the way social movements translate
 structural possibilities into concrete action. For present purposes, it provides a crucial conceptual bridge
 which allows an explanation of the problem defined at the outset: how African and coloured auto
 workers came to act together, defining themselves as a "we/ as a collectivity with certain shared
 meanings, identities, and interests* which could be translated into collective action.
 III. FACTORY FLOOR ORGANIZATION IN THE AUTOMOBILE INDUSTRY
 A. BLACK WORKERS' ENCOUNTERS WITH RACIAL DESPOTISM
 When Elijah "Scoma" Antonie first entered the Volkswagen factory in February 1960, just one month
 before the Sharpeville emergency, Africans made up a small percentage of the VW workforce, and were
 confined to the most menial tasks in the factory. White supervision ruled the factory with an iron fist.
 Scoma was appalled by the blatant favoritism and discrimination:
 When I was introduced in that department it was said to me, "You see these whites who
 are standing here, they are all your foremen. Even if a white man says you must go and
 pick up shit, you must go and pick it up."
 His first reaction was aggressive: "I will never pick up shit. I have come to work here. If a white man
 has done shit, it is for him to pick up that shit." Scoma was not alone. Though workers were quiescent
 during the 1960s, they were certainly not content with their lot. Fury lay under the surface calm.
 Vuyo Kwinana entered the factory only in 1972. But when he arrived, he was told by his
 foreman that blacks cannot go to the toilet any time they pleased. "I have got to register my name
 according to the alphabet. A goes first to the toilet, then B second, then C third in order." On his first
 day Kwinana made a decision:
 I felt on that particular day, "I am not fit to work here." Half-past four will be too far
 for me to stretch, but I made it an option to stay in this department, clean out the
 waste, and throw it out the window.
 Both Scoma and Kwinana entered the factory with ideas which predisposed them to reject such
 practices. When he joined Volkswagen, Kwinana had been a member of the ANC underground since
 1957, and had served a two-and-a-half year prison term in the 1960s for being a member of the banned
 organization. Kwinana's father had been a leading - though conservative - figure in the ANC in East
 London, before the family moved to Uitenhage in 1951.12 Vuyisile Tote was one of the activists who
 had recruited him into the underground; Tole had been a worker at Uitenhage's Fine Wool Products
 [Veldspun] and had been dismissed during a locally celebrated 1955 strike.13 He had learned his trade
 unionism - and was himself recruited into the ANC and MK - via Raymond Mhlaba of the ANC and
 SACP and Themba Paulos of the ANC and TWIU.
 Scoma had not been a member of the ANC, but he had gained an understanding of trade unions
 through lectures on the ANC and SACTU given by Zola Nqini, a local ANC and SACTU activist. He
 had also heard stories about the exploits of TWIU unionists and their strike against Fine Wool Products
 in December 1955." Both Scoma and Kwinana developed second thoughts about confronting the
 Volkswagen foremen directly. Scoma was counselled by an older, more experienced worker:
 There was an ex-teacher who used to work in that department, and he called me to one
 side and said, "Hey, you've got to be very careful. You should go slow, because most of
the chaps here are very weak and they can sell you openly to these whites." And I was
 wearing a school blazer then. Then as time goes on, I checked my way, how to make
 these people aware that a trade union is very useful.
 Scorn a adopted a wise policy, given the era. From the beginning of the industry in the 1920s through to
 the 1970s, foremen possessed vast discretionary powers, including the right to grant incremental wage
 increases for workers, and most significantly, the power to hire and fire. It made tittle difference that for
 most of its history white workers formed the bulk of the labor force in the motor industry: supervisorial
 despotism was a hallmark of the industry irrespective of the racial composition of the workforce.
 B. THE FORMATION OF NUMARWOSA AND THE UAW
 The conditions Scoma described in the 1960s did not begin to change until after an Industrial Council
 was established for the Eastern Cape motor industry in 1969. Oddly enough, the first breach in the wall
 of management authority was made by the white Yster en Stal Uniet which managed to use access to the
 National Party and the Minister of Labour and threats of imposing Job Reservation to win union
 recognition in 1967 and the subsequent establishment of an Industrial Council. The existence of an IC
 for the industry meant that the legal space existed to bring coloured workers into the same framework,
 and in 1967 the Trade Union Council of South Africa [TUCSA] sponsored an effort to organize coloured
 automobile workers in Port Elizabeth.13 Within months factory-level organizing committees had been
 established at the three Port Elizabeth motor plants, Ford, General Motors, and Rover, and a new union
 was launched: the National Union of Motor Assembly and Rubber Workers of South Africa
 [NUMARWOSA].
 Discriminatory practices on wage rates, promotions, and working conditions were foremost in
 workers' minds. Each of the organizing committees set about building on these grievances to create a
 wider base within their respective plants. At first they attempted to create a network of contacts in
 different factory departments. Most contact people, while not resistant to the idea of a union, were
 reluctant to meet, even secretly. They viewed a unionizing drive with considerable suspicion,
 remembering the banning*;, treason trials, and other repression meted out to activists. Among those who
 were interested, few individuals were willing to take a leadership role. The factory organizers quickly
 identified a strategy to overcome workers' reluctance; they identified and engaged senior long-service
 workers who had some standing in the factory and in the coloured community, such as ex-teachers and
 rugby stars, and who were able - within the limits of the time - to stand up to foremen. The factory
 committee activists recruited such elder workers on an individual basis, and by bringing them into the
 project, eased entry for other workers who identified with them.
 As the organizing base grew, every department would have one or two volunteers helping the
 organizing committee. These individuals would emphasize the workers' concrete grievances, especially
 regarding foreman and pay, rather than broader political issues. Their efforts were in part facilitated by
 the legality of their initiative: though workers were certainly opening themselves up to victimization, their
 recruitment to NUMARWOSA did not contravene any existing laws. Finally, the activists would not
 expose any names to management until they had enrolled a substantial majority of the workforce.
 Starting in mid-1967 it took less than a year at Ford to sign up more than 50% of the coloured
 workforce. The union began holding regular general meetings in the community for the new
 membership. In April 1968, after ten months of organizing, the union elected an interim branch
 committee to administer the union's affairs, a leadership that could come out publicly as advocates of the
 union. After it became a formally registered union and gained entry to the IC in 1971, NUMARWOSA
 leaders began planning a more ambitious strategy of national industrial unionism. As it was slowly
 elaborated over the years 1971-1973, the strategy entailed an extension of its organizing efforts to Durban
 and the Transvaal, mergers with other unions organizing automobile workers, and the commitment to
 organizing on a non-racial basis.16
 NUMARWOSA's Executive Council took its first policy decision to approach African workers in
 late 1972." Its initial attempts were unsuccessful, partly because its organizing efforts were dissipated
across Port Elizabeth, East London and Uitenhage, and partly because, before the 1973 Durban strikes,
 there was no national impetus supporting African worker organization.
 Moreover, there were significant legal obstacles: as a registered union, NUMARWOSA could
 not have African members. The union initially sent stop order forms to the companies they thought
 would enroll African workers, hoping to create a fait accompli. Management rejected this approach, and
 NUMARWOSA's leaders had to find ways to bypass the legal restrictions on African representation in
 registered unions. In April 1973, the union adopted a familiar strategy through which coloured and white
 unions had long dominated African workers: it established a parallel union, the United Automobile,
 Rubber, and Allied Workers Union of South Africa [UAW].
 The aim behind the new approach was to organize Africans until the UAWs strength was
 comparable to NUMARWOSA's and then to merge the two organizations once the legal situation
 allowed such a move. In contrast to the typical TUCSA parallel union, whose leaders were appointed by
 the executive of the all-white union, the UAW would have leaders elected by the members, and a
 separate executive which could take decisions on behalf of the membership. The two unions held joint
 meetings, and formed a combined Advisory Council on which sat representatives from the executives of
 both organizations.18 The UAW was never seen by NUMARWOSA as a means of control over
 Africans, nor as a permanent structure, but as nothing more than a tactical device for organizing Africans
 into one union of auto workers, once the legal climate permitted. Though UAW could not yet have
 recognized shop stewards on the factory floor, its leaders functioned as de facto stewards, who would be
 brought into NUMARWOSA's existing structures after the two unions united. The merger would finally
 occur in October 1980: NUMARWOSA, UAW, and the WPMWU joined to form the National
 Automobile and Allied Workers Union [NAAWU], the first truly national industrial union within the
 emerging union movement.19
 C FACTORY FLOOR ORGANIZATION AT VOLKSWAGEN
 Volkswagen began hiring coloureds only in 1968, and NUMARWOSA immediately began
 organizing them.20 In May 1969 more than 50% of the coloureds at Volkswagen were organized, and
 NUMARWOSA was soon formally recognized by management. At that point, Volkswagen management
 took the extremely rare step of introducing a Liaison Committee as a forum to address minor grievances
 of African workers. Though Volkswagen workers would soon challenge it's structure, the Liaison
 Committee at first operated as in other companies in South Africa: worker members and the secretary,
 Elias Jonas, were appointed by management, rather than elected by the workers.21
 By contrast to NUMARWOSA's rapid gains among coloured workers, the process of organizing
 African workers was slow and painstaking, given their exclusion from the Labour Relations Act. Albert
 Gomomo was a worker in the Printing Department, which was in the area adjacent to Scoma, who
 worked in Dealer Aids. Gomomo had been a local activist in the PAC before it was banned, and had
 worked at the Gubb and Inggs woolwashery in the 1950s before coming to Volkswagen. Gubb and Inggs
 was near the Fine Wools plant, and Gomomo had been exposed to the TWIU activists who once
 organized in that plant. He was thus knowledgeable about trade unions and had a clear political
 perspective before coming to Volkswagen. Scoma did not know Gomomo, but Gomomo knew him, as
 Scoma was a player on the Swallows rugby club, which Gomomo served as an official.
 Gomomo approached Scoma, and their first discussions focused on rugby. Only later did they
 focus on ANC and PAC politics. But Gomomo gradually encouraged Scoma to come forward and take a
 more active role in building organization at Volkswagen. The workers' mounting complaints about the
 Liaison Committee spurred them to discuss trade unionism.
 Their first step was to talk with other workers in their respective departments whom they knew
 and trusted. Sometimes they would talk during lunch, or outside the factory after work in shebeens.
 They targeted other elder workers who had some experience of the ANC and PAC in the 1950s, and who
 had some experience in the trade union movement at that time. In their forays around the plant they
 would often encounter such workers, as many Volkswagen employees had worked at Fine Wools as their
first job. Mongazneli Davids and a Mr. Ndingane were among the first to enter the network. They were,
 on their own initiative, working in a similar fashion in their own departments, finding three or four like-
 minded workers to engage in regular discussions. They worked secretly, linking back to Gomomo, in a
 cell system somewhat analogous to that developed in political struggle in the late 1950s.
 In developing their organization, the network members discussed their general approach. They
 wanted to build a union, but realized that such a task was impossible for African workers at that time.
 They also developed their own critique of the failings of SACTU, in particular of its close links to the
 ANC, which saw committee members of the one organization serving on the committee of the other. As
 Scoma put it,
 That is a mistake. If the ANC is banned, it stands to reason SACTU is also going to be
 banned. Now in order to avoid that, we did not want to align ourself with the ANC
 directly. We wanted to be an independent body.
 Independence paid another dividend: the nascent organization was non-partisan, and attracted activists
 from the camps of both exile organizations, as well as apolitical workers. Furthermore, the political
 independence of the initiative helped reduce ordinary workers' fears of joining an organization whose
 membership could be construed as "political." The early 1970s were so repressive that even a hint of
 connection to the banned liberation movements could result in intervention by the security police; the
 union's political strategy helped remove their efforts from the political field, and altered ordinary
 workers' calculus of participation in favor of engagement.
 After the Liaison Committee was introduced, and workers became disenchanted with what they
 perceived as the stooge-like behavior of the management-appointed workers, the secret network came
 upon the idea that they could take over the Liaison Committee as a first step towards developing a
 union. Gomomo secretly introduced Scoma to contacts in other areas of the factory, canvassing them
 about putting him onto the Committee to ensure that such a step would not be seen as selling out to
 management. Gomomo recruited his younger brother, John Gomomo, as another candidate for the
 Committee.22 Both Scoma and John Gomomo had excellent work records, so neither could be easily
 victimized, and they were not members of the banned political organizations. They could come forward
 as acceptable public fronts for the growing network.13
 Scoma and Gomomo approached management to complain about the behavior of the
 management appointees to the Liaison Committee, and won the right to elect African workers to the
 structure. In 1971 both became members of the Committee, and once elected, they were able to win the
 appointment of a new secretary, Elcott Rhoxa, a staunch member of their circle. From the Liaison
 Committee they began to take up workers' grievances, operating as quasi-shop stewards. Much of their
 time was occupied fighting unfair dismissals and job upgrading cases, but they also pushed the company
 to finance community creches, a scholarship scheme for workers' children, and a sports stadium for
 Kwanobuhle township.
 Targeting the Liaison Committee was strategically wise. Volkswagen had itself established the
 committee, and it was less threatening to management to mobilize workers around its inadequacies than
 it was to organize a union. Ordinary workers would feel more comfortable enlisting in the project, as
 their fear of victimization by management was reduced. In this way, the activists' strategy helped reduce
 an important obstacle for ordinary workers to join the organization.
 The takeover of the Liaison Committee did not occur in a vacuum. From Scoma's first days on
 the Committee he had begun discussions with Chris "Papa" Williams, then NUMARWOSA President,
 and one of the strongest advocates in the union for the new policy of organizing Africans. Williams had
 been involved in non-racial rugby as a coach and manager, and, fortuitously, frequently found himself
 working in Albert Gomomo's department at Volkswagen. Through Gomomo the NUMARWOSA
 President was in close contact with the growing informal network among African workers. Scoma also
 had a number of close friends among coloured workers, against whom he played in local friendly rugby
 matches, and whom he knew from rugby administration. Through them he met the other
8
 NUMARWOSA shop stewards and began to cement friendships. The NUMARWOSA shop stewards
 and the insurgent Liaison Committee members began to meet regularly prior to Committee meetings, to
 discuss the agenda, and to ensure that the subject matter of the Liaison Committee did not undermine
 what the union was simultaneously negotiating with management. In this way the coloured union
 ensured they would not be undercut by African workers, and also that workers would present a united
 front to management. Their joint work developed to the point where the NUMARWOSA shop steward
 and the liaison Committee members met routinely to discuss all matters brought up before
 management.
 These informal joint meetings were in place by 1973 when NUMARWOSA decided to form the
 UAW. Indeed, as Williams was one of the stronger advocates of the NUMARWOSA policy, it is highly
 likely that his experience with Gomomo and the other African workers at Volkswagen influenced his
 thinking on the broader union policy, as he had very direct and practical experience that such bonds
 could in fact be built.
 With the creation of the UAW - at least on paper - the NUMARWOSA activists began wooing
 African workers. At Volkswagen the policy proceeded slowly, as the shop stewards invited the Liaison
 Committee insurgents to the NUMARWOSA offices in Port Elizabeth, and to the coloured union's
 offices in Uitenhage to discuss their joint problems. Slowly the idea developed that the Liaison
 Committee could serve as a base from which to organize African workers into the UAW, and seek some
 form of recognition from management.
 The social networks, which were the foundation on which the union grew, were not limited- to
 those between coloured activists in NUMARWOSA and Africans on the Liaison Committee. As the
 network Gomomo constructed was non-partisan, it intersected the ANC's clandestine cell system.
 Working underground, Congress members in Volkswagen strove to build the UAW by encouraging those
 within their influence to join.24
 As Vuyo Kwinana reflected,
 We knew NUMARWOSA was very powerful as a trade union. We had no doubts
 about it, but our main purpose was to bring the black workforce over to
 NUMARWOSA's side....We fully influenced our people to join UAW and join
 NUMARWOSA.
 Kwinana had no problems building a union which was politically "independent" from his own
 organization, nor was he under instructions to penetrate Volkswagen. But after experiencing
 discrimination in the company, he decided to help organize the workforce:
 I had to contact my organization afterwards, and told them about the plight of the
 workers. Then they said, "Do influence the workers as much as possible. We will try
 and assist you if we can." So my work was to destroy the Liaison Committee and
 establish a trade union at VW. That was my prime duties as an ANC worker, as an
 ANC member.
 Through workers like Kwinana the cells influenced the Liaison Committee members to "motivate them
 what direction to take, what strategy to apply."
 A cell member would secretly approach workers in his department whom he trusted and, after
 talking with them at length, would encourage them to move into other areas of the factory, from where
 they would remain in touch with him on a regular basis. Kwinana described the typical manner in which
 the cells - without announcing their identity - would influence workers:
 Scoma is in my department. He reports to me practically every day what their plans are
 in the Liaison Committee. We in turn, investigate from NUMARWOSA, because a
 friend of mine is...a coloured shop steward at Body Shop. He draws information from
this coloured chap, and bring them to us. And we sit down to analyze the information
 brought by this coloured and to the liaison Committee we said, This is wrong. These
 people must not move this way, they must move this direction." And then we start to
 influence now and encourage the Liaison Committee to move in this line.
 These individuals almost never became shop stewards or office-bearers of the union, but instead kept a
 low profile as ordinary workers on the floor, from where they could quietly exert their influence by
 discussing issues with other workers.
 Themba Dyassi was a central figure in the clandestine system. He was a member of Kwinana's
 cell, and a welder in the Body Shop. He first became involved in politics at school in the late 1940s, and
 after being expelled found work in Uitenhage with a building contractor, but lost his job in the
 victimization which followed the November 1952 Port Elizabeth stay-at-home.25 After finding work at
 Goodyear in Uitenhage, he entered the ANC cell structures, from where he was eventually recruited into
 MK. He was arrested in April 1964 while preparing to leave the country for military training. Sentenced
 to 2 1/2 years for membership in the ANC, he served most of his time on Robben Island, where he
 decided that upon his release he would return to factory work and attempt to build trade unions. In
 1967 he got a job at GKV in Uitenhage, and after a brief spell back at Goodyear took a job at
 Volkswagen in late 1974. Dyassi deliberately avoided taking a leadership role in the union, and especially
 avoided becoming a shop steward. He preferred to remain on the line:
 Being a shop steward is a very difficult job because most of the time you are at the
 offices. But being a worker, we arc amongst workers. You can actually see what is
 happening. Being a shop steward you are told what was happening there at the work
 place.
 From his position on the floor Dyassi could directly influence workers in his area, bring them into the
 union, and mediate between ordinary workers and the stewards. He recruited promising individuals into
 the union and into the ANC underground structure, and provided valuable political education and
 discipline, especially to younger members. He was part of the bedrock structure of both the union and
 the ANC: both organizations' direct link to ordinary workers whom the movements wanted to mobilize.
 The ANC cell network operated inside the factory, during lunch times and breaks, and outside
 the factory in the townships.
 We met informally outside the factory in the cell structure....we would ask a chap from
 VW, especially the coloureds, to come and drink with us over the Saturday, then we
 start questioning him, "What do you think about this and that?" We found that, more or
 less, they are very positive, they want to work closer, and sympathized greatly with the
 African. So we saw very fertile ground for us to influence them.
 The coloured worker would not have been aware that he was meeting with the underground; as far as he
 understood, he was meeting ordinary Volkswagen workers in a social gathering at which they "talked
 shop." Through these networks the cell members gained a broad purchase on developments inside
 Volkswagen, and given the members' commitments to building the union, solidified its presence in the
 factory.
 In Volkswagen NUMARWOSA developed something it lacked in virtually all its other
 workplaces: extremely close links between its leaders and prominent African workers who had legitimacy
 in the eyes of the African workforce. These connections made it far easier to present the idea of
 developing the UAW, which was, after all, a parallel union that could easily be construed as part of an
 oppressive strategy by coloured workers to dominate Africans. To have workers like Albert Gomomo
 and Vuyo Kwinana as advocates of the UAW helped clear away any misgivings about tjie organization's
 bona fides.
 For reasons which remain unclear, the organizing drive only got started in 1975, perhaps because
10
 of NUMARWOSA's preoccupation with withdrawing from TUCSA. But the Liaison insurgents affiliated
 to the UAW launched another drive between 1976 and early 1977. Volkswagen management was well
 aware that the liaison insurgents were organizing a union, but they had no idea how successful the drive
 had been. They were soon to find out. In mid-1977 an unlikely pair of visitors appeared at the plant one
 day to meet the Managing Director, Peter Searle and Ollie Rademeyer, the Industrial Relations Director.
 The delegation consisted of Werner Thonnessen, Assistant General Secretary of the International
 Metalworkers Federation, visiting South Africa for meetings of the South African Coordinating Council
 of the IMF, and Johnny Mke, President of the UAW. It was no mere courtesy call: Thonnessen and
 Mke brought with them stop order forms signed by more than 60% of Volkswagen's African employees.
 Management quickly decided that UAW was supported by the African workers, and recognized it. Thus,
 well before the Wiehahn labor law reforms were announced, Volkswagen recognized the UAW, gave
 them stop-order facilities, and scrapped the Liaison Committee. To comply with existing law the
 members' stop orders were technically defined as contributions to a burial fund.
 The four insurgent members of the Liaison committee immediately stood for election as UAW
 shop stewards, entering the structure established by the coloured union in the early 1970s. Scoma and
 John Gomomo were elected to the corps of 4 UAW and 4 NUMARWOSA stewards,26 and were taken ?
 on shop steward training courses conducted by NUMARWOSA.27 The two unions shared offices, held
 joint meetings, and made decisions together in a Branch Executive Committee which included elected
 representatives from both. UAW observers sat with NUMARWOSA in the Industrial Council, and
 together they plotted their strategy there. By the following year they were discussing strategies for
 merging the two unions. Thus, on the shop floor, the Liaison Committee insurgents - now recast in the
 role of UAW shop stewards - and the NUMARWOSA activists were realizing the strategy developed
 within NUMARWOSA in 1972-73 to build one non-racial industrial union of automobile workers.
 What made the Volkswagen experience unusual is that at no other plant in the entire country
 did NUMARWOSA-UAW succeed in unifying coloureds and Africans into a single unit. At Ford a
 racial division of labor meant that coloureds were congregated at its Neave plant, while its Struandale
 Engine and Cortina Plants had almost entirely African workforces. UAW organized workers at Engine
 Plant, but they never successfully united with NUMARWOSA at Neave, the way the two unions had
 done at Volkswagen. Nor did UAW succeed at the much larger Cortina assembly plant. Instead, after a
 series of wildcat strikes in late 1979 and early 1980 this plant was to be organized by a breakaway African
 grouping which would soon form the Motor Assembly and Components Workers Union [MACWUSA].
 At General Motors, too, the NUMARWOSA branch was long hostile to organizing Africans, and UAW
 never took off. After 1980 the overwhelming majority of African workers at GM would also be
 organized by MACWUSA.
 IV: FRIENDSHIP AND FAMILIARITY: SOCIAL NETWORKS OF SOLIDARITY
 An analysis of the workplace and labor process would certainly be important in accounting for why it was
 that unions formed as collective responses to shared conditions.28 Workers' cross-race solidarity,
 however, cannot be explained simply by reference to their insertion into a labor process which
 engendered a sense of cooperation amongst workers and grievances against management. Indeed, very
 similar working conditions were present in the plants when whites were the bulk of semi-skilled
 operatives, and did nof; encourage the development of solidarity among them. It certainly did not foster
 solidarity between white and black workers when the latter came into the factories in the 1960s. Indeed,
 similar conditions of work were present in the Port Elizabeth auto plants, and workers there did not
 develop a collective body Unking coloureds and Africans. An explanation of cross-race solidarity must
 therefore take into account social relations outside the factory.
11
 A. APARTHEID'S CHICKENS COMING HOME TO ROOST: THE CONSEQUENCES OF KABAH'S
 FAILED REMOVAL
 1. The Uneven Application of Racial Separation
 Since the 1840s, Uitenhage's African population has resided in the Kabah Location. Kabah's population
 was always racially diverse: between the 1930s and 1950s coloureds made up approximately 25% of those
 living in the "African" location. From the 1950s, however, many coloureds left Kabah as new
 Municipally-built sub-economic and economic letting schemes were started in Gerald Smith and Thomas
 Gamble townships, and a freehold private housing development opened known as Jubilee Park.
 Nonetheless, many coloureds remained behind in Kabah even into the 1970s, and these coloured housing
 schemes were in fact built in extremely close proximity to the Location. Social contact between Africans
 and coloureds was the norm rather than the exception, and indeed, before the passage of the Population
 Registration Act in 1950 and the application of the Pass Laws for Africans, the line separating coloureds
 from Africans was rather fluid, more notional than statutory.
 One of the main aims of apartheid legislation after 1948 was not simply the spatial segregation
 of residential areas, but the thorough separation of social, political, and economic activities. Group
 Areas, for example, were drawn in such a manner that a member of a certain group would not need to
 traverse another race's Group Area on the way to work or to shop. Virtually any form of contact was
 made extremely difficult. Other pieces of legislation enshrined this racist principle: new laws banned
 cross-racial sex or marriage, separate educational systems kept school children apart, while the Labour
 Relations Act of 1956 included clauses prohibiting different races from joining the same trade union.
 Of course if the "blueprint's" aim was to achieve full segregation, it proved extremely difficult in
 practice to sort out the diverse, racially heterogeneous areas of certain cities, especially in the Cape
 Province where Africans, coloureds, and Indians all had long traditions of property ownership in areas
 zoned for white occupation under the Group Areas Act. The extent of heterogeneity differed from place
 to place, while the struggle to impose apartheid legislation proceeded at an uneven pace and was often
 resisted by local authorities.9
 In Uitenhage, the central state's desire forcibly to remove Kabah consistently came undone,
 through a combination of resistance by the local Council, shortage of funds at the national level, and
 general bureaucratic inefficiency. Removals from Kabah to the new African township at Kwanobuhle
 began only in 1968, and even then, at such a slow pace that Kabah's African population remained
 relatively intact through the 1970s. With a huge urban influx due to the disastrous drought of the early
 1980s, Kabah's population in 1985 was as large as ever. Its persistence was a continued affront to
 apartheid planners' dreams for a racially ordered community.30 In Port Elizabeth, on the other hand,
 the extensive historical practices of local segregation, coupled with the removal of Africans from Korsten
 after 1955 meant that Africans and coloureds were almost entirely separated from one another by 1957-
 58.31
 2. Contact Across the Divide
 In contrast to the racially-engineered geography of Port Elizabeth, Kabah's delayed removal created
 structural possibilities well into the 1980s for contact between Africans and coloureds, who were either
 living in the same neighborhood or in close proximity. Most of these contacts grow out of the long
 history of contact in the community before it began to be chopped apart on the Procrustean bed of
 apartheid planning.
 Workers who grew up in Kabah have rich memories of contact with people from other races.
 Most coloureds lived in the south-eastern section of Kabah, the Baines Road and Middle Street area.
 One African worker remembered the area clearly:
 Middle Street, the street that was actually the dividing street between the coloureds and
12
 the Africans....So we used to grow up and play in the same street with the
 coloureds...that's where we got our Afrikaans. I didn't get my Afrikaans in school....We
 sort of got friendly with the coloureds. We regarded ourselves as one people.
 The same worker recalls going to church services at the mostly coloured Apostolic Faith church and at
 the Dale Street Congregational Church, the latter under the leadership of Rev. C.W. Hendrickse and
 later, his son, Rev. Alan Hendrickse.
 Other workers recall joining with coloured children to look after livestock on the north-eastern
 edge of Kabah. The bush provided respite for children in many ways, as one coloured worker recalled:
 Sometimes on hot days, there was a big dam...there at the other side of the hill, the
 other side of the bosch. We'd go there, learning each other how to fish....Well, we had
 no fishing rods. If you see a fish you have got a big stick, you just hit that fish with that
 stick. Sometimes you are fortunate, sometimes you are not fortunate.
 A coloured worker, who grew up with Africans in Blikkiesdorp, across Maduna Road from the main area
 of Kabah, recalled what he and his friends did for entertainment;
 It was near to the bush, we as guys got a big place alone, where we always played rugby,
 go and play in the bush, cowboys and crooks. It was a lot of fun to live there.
 The statutorily-imposed definitions of race and residential segregation upset what were in fact
 sometimes extremely close and comfortable relations between different "races." But the official
 categories did not mesh with people's direct experience in more ways than one. Color, alone, was hardly
 a clear demarcation of official racial identity, as one worker explained:
 You sometimes cannot easily identify a coloured from a black man. You can only
 identify him by listening to his speech, and you say, "Hey, this is a coloured man. I
 didn't know he was a coloured." Most people think that I'm a coloured...because of the
 color of my skin. It's what happens in most cases.
 For the people themselves, social practices, such as language or adherence to African customary practices
 like circumcision and lobola were more crucial to determining group inclusion than were biological
 considerations, such as skin color or facial features. Social practices are of course far more malleable
 than biological characteristics, and people can adopt or reject certain practices for a variety of reasons.
 Adoption came about through intermarriage in at least two cases among those interviewed. The
 mother of the worker who above described the Middle Street area, was the daughter of a Sotho woman
 and a coloured man, and was officially classified as coloured; the workers' father was a Xhosa. As a
 child the worker learned to speak Xhosa, Afrikaans, and seSotho. But his mother's brothers and sisters
 have coloured spouses, live in Uitenhage's coloured areas, and are legally considered to be coloured
 people. While his brothers and sisters are legally African, his cousins are legally coloured. This didn't
 present a problem for the worker concerned:
 I see them as my brothers and sisters. We are in fact one in the same people who see
 no difference at all. As far as they are concerned, I usually go there and sleep overnight
 and their friends in turn take us as their brothers and sisters. And also when they do
 come over to this end our brothers and sisters and our friends take them as the same.
 Nor did the worker experience strife with his cousins over participation in his customs:
 When we do make our customs, some of them attend, because when there is meat, we
 braai [barbecue] the meat, cook the meat, they take it, they have no objections at all.
 And I've seen lately that the coloureds also have adopted some of our customs, because
 some of the coloureds are actually African people. If your mother is coloured and your
13
 father is black, and you choose to be coloured, you go and live in a coloured area and
 you grow up as a coloured, but since your father was an African, you still have to abide
 by your father's customs. And unfortunately for you, you are staying in a coloured area
 where these customs are not being practiced. But, I've seen changes lately that most of
 the coloureds [who have African fathers] are being circumcised.
 These cases raise complicated questions about the processes through which people develop racial
 identities, a subject beyond the theoretical scope of this work. However, at a more prosaic level the
 social practices described above led to important interactions across legally-constituted communities,
 which reveal a great deal about workers' actual lived experience of race.
 In the case of this particular worker, he regularly (though not frequently) had occasion to visit
 relatives in the coloured areas and for them to visit him in Kwanobuhle. Nor were the visits fleeting
 encounters for trivial reasons, but were in fact sustained over weekends, and were organized around
 important, meaningful family rituals. For many other workers, visits across the apartheid divides were
 extremely common, and covered a range of activities from spur-of-the-moment socializing, to political
 and union gatherings, to sports.
 Almost two-thirds of the Volkswagen workers interviewed visited across the community/race
 divide on a regular basis. Well more than half the African workers visited into coloured communities,
 while almost all the coloured workers visited into African communities. On average, African workers
 paid more than 3 visits a month to coloured communities, while coloured workers paid just under 10
 visits a month to African communities. Over half the workers interviewed visited across the racial divide
 at least once a month. A small number visited less frequently, while fewer than 4 in 10 never paid such
 visits.
 By contrast, at General Motors African and coloured workers socialize across the racial divide
 rather less frequently than Volkswagen workers. African workers visited coloured areas an average of
 1.5 times per month, while coloured workers visited African areas on average 3 times per month. One
 quarter of the coloureds and a third of the African workers interviewed at GM paid one or more visits
 across the racial divide. However, well more than two-thirds of all workers paid less than one visit per
 month, or never visited at all.32
 The figures for coloured workers at GM roughly parallel the findings from Marianne Rome's
 1973 survey of coloured automobile workers in Port Elizabeth. She found that only 22% of the coloured
 workers interviewed paid one or more visits to African work friends in the month prior to being
 surveyed. These figures are not due to a general lack of sociability among coloured workers in Port
 Elizabeth: 84% of the workers surveyed had paid one or more visits to other coloured work friends
 during the same period.33 The image which emerges is of very frequent contact between Africans and
 coloureds in Uitenhage, though rather greater frequency of contact by coloureds to African areas than
 the reverse, but of very little similar cross-race contact between Port Elizabeth Africans and coloureds.
 In contrast to the apartheid ideal of total separation, the statutory divisions between Africans
 and coloureds broke down in practice as a result of social interaction in a variety of spheres: people
 joined together in pubs, weekend parties, sport, churches, even in inter-marriage. In this wide range of
 activities African and coloured workers daily created social bonds of friendship and affection, contacts in
 which they learned first-hand of each other's lives, and came to share experiences and understandings.
 In large measure these exchanges formed the social basis of solidarity which supports a wide range of
 oppositional and cross-race social movement activity. Of course the imposition of a more thorough-going
 apartheid in Port Elizabeth did not prevent the emergence of strong social movements there, but it did
 profoundly shape the character of those movements, rendering it far more difficult for joint involvement
 of Africans and coloureds in the same project.
 Of all the areas of social interaction, one in particular proved most important as the meeting
 ground between African and coloured auto workers. For both African and coloured men sport was one
 of the most important areas of community activity, and no sport had a greater purchase on their loyalty
14
 than does rugby football. It was the dominant group sport among young African and coloured workers
 growing up in die Eastern Cape from the 1940s to the 1960s, and the way it was played in the two towns
 captures the differences in social interaction between Africans and coloureds in Uitenhage and Port
 Elizabeth.
 B. KEEP YOUR EYE ON THE BALL: RUGBY, INTERACTION, AND SOLIDARITY
 1. The Game of Choice
 Rugby football is commonly associated with South Africa's white population, especially Afrikaners, where
 it assumes the dimensions of a civil religion.34 In most of South Africa rugby is a white game, while
 blacks largely follow and participate in other sports such as cricket, road running, boxing, and of most
 importance, soccer. In the Eastern Cape, however, rugby has long been the game of choice for both
 African and coloured people, and is followed with religious fervor.35 It is occasionally played between
 white and black teams, but more commonly, it is a game played by and between coloureds and Africans.
 Rugby is thus an important arena in which the two groups are brought together socially on a routine
 basis.
 Rugby was imported to the Eastern Cape by British missionaries, and formed an important part
 of the curriculum at mission schools. In the missionaries' perspective, proper training in cricket and
 rugby were considered crucial to weaning Africans from tribal pursuits and fostering an educated,
 assimilated elite on an English model.36 The first sides were based at these educational institutions.
 The first adult club was formed in Port Elizabeth in 1887, and the game quickly spread to towns
 throughout the region.37
 According to Peires, since the 19th century the game has been closely associated with the central
 concerns of black communities, serving not only as entertainment, but "as a vehicle of popular culture
 and as an outlet for personal achievement and ambition" which always attracted the leading civic and
 intellectual figures. In communities with few avenues for personal or political expression, "energies which
 might otherwise have been devoted to local or national service were displaced into rugby." Indeed, in the
 last twenty years, Peires convincingly argues, rugby was transformed from a substitute for politics to "an
 arena of national struggle, attacking a particularly vulnerable heel of the apartheid state."
 Such a transformation was aided by the National Party's own politidzation of sport and its
 application of rigid apartheid to such pursuits. Sport per se could not be banned, though incredible
 obstacles could be thrown in the way of sports administrators and players.36 Administrative pressures
 could be placed on undesirable sides or breakaway unions, by denying clubs access to facilities. In fact, it
 was the state's refusal in 1973 to recognize the Kwazakhele Rugby Union [KWARU], which had broken
 away from the official and racially-exclusive South African African Rugby Board [SAARB], which pushed
 KWARU into the arms of the South African Rugby Union [SARU], an affiliate of the anti-apartheid
 South African Council on Sport [SACOS]. KWARU was excluded from the well-serviced stadium in
 Kwazakhele, and state radio refused to broadcast the Union's matches. The heavy-handed action by the
 state rallied popular sentiment to KWARU, and virtually all of the better players and teams in Port
 Elizabeth moved over to the new non-racial union. KWARU became a nesting ground for non-racial
 political activists, including both Black Consciousness and ANC-aligned figures, who endorsed the goals
 of non-racial sport and opposition to state policy.
 The government's encouragement of the game, if only on a racially divided basis, created space
 for people to operate and (within limits) to pursue their own agendas. Rugby matches were among the
 only places where blacks could legally gather into large crowds, and the events could be quickly
 transformed into political events at half-time or during a break in play: an impromptu speech might be
 made, a freedom song sung, a hat passed around for a political cause, or a moment of silence observed.
 At a less overt level, the political symbolism would not be lost on the audience if, at the occasional match
 between black and white sides, the black team demolished its opponent.39
15
 2. Non-Racial Rugby and Social Interaction in Uitenhage
 At an organized level, Uitenhage teams had long played together in "friendly" matches pitting sides from
 Kabah or Kwanobuhle against sides from the coloured areas or from the district. Such contact grew
 much more frequent when both coloured and African sides broke away in the 1970s from their respective
 racially-exclusive official bodies to join the SARU-affiliated Uitenhage and District Rugby Union where
 they would play fixtures together on a regular basis, and would compete for local cups and the Union
 and SARU championships. By contrast, after KWARU broke away from the SAARB, it formed a
 separate union within SACOS, while coloured sides in Port Elizabeth continued to play under SARU's
 Eastern Province Rugby Union. Thus sides from Port Elizabeth's African and coloured townships would
 not play together within the same league, but only on the occasional friendly basis, or at the level of
 inter-provincial play for the SARU championship.
 Rugby brought African and coloured people in Uitenhage into extremely close contact. With the
 rising power of the SACOS-affiliated SARU in Uitenhage, teams began to integrate: in small numbers
 Africans joined once all-coloured teams, like Excelsiors, and some coloureds joined African sides, such as
 Swallows. These "cross-overs" were welcomed into the full range of club activity, which brought them
 into regular contact with teammates from across the race divide. One coloured worker who joined
 Swallows journeyed regularly to Kwanobuhle for practice and for matches. As a result, he established
 close friendships with his teammates, and would visit frequently:
 [Before the emergency] That was every weekend. I stay overnight sometimes in
 'Nobuhle and in Langa location. Oh, I have got a lot of friends there. I've got a lot of
 African friends.
 This particular form of integration was relatively unusual, as most teams remained strongly based in
 neighborhoods, and with the consolidation of Group Areas from the late 1960s, the racial separation of
 communities largely defined the character of the teams.
 Notwithstanding the internal composition of the sides, during the season clubs affiliated to the
 Uitenhage and District Rugby Union would play each other every weekend, which meant that for some
 period of time coloureds would visit African townships and Africans would visit coloured areas. Teams
 would at least travel through different communities and players would be exposed to conditions faced by
 others. As one African worker put it:
 We happen to be staying together. Even in my team, we've got some coloured chaps,
 and even in my [rugby] Union we've got a coloured team from Rosedale and Gamble.
 So on weekends, sometimes you play there during the weekends, or they come here, or
 we play at Langa. Always together.
 The keen dedication and single-mindedness sportsmen show on the field extended to then-
 entertainments after the match: the game would often be but a prelude to a full evening or even
 weekend of socializing:
  r
 Sometimes we play at their places. Some of the chaps they slept there. [Then] they
 take their coloured chaps, their friends which they have there, to Kwanobuhle.
 The parties are often raucous affairs of singing and dancing, well supplied with barbecued meat and
 alcohol, but with a strong camaraderie among teammates, as well as between members of opposing sides.
 In this way people from different officially-designated races spent considerable time together in the
 intimate, personal space of a family home.
 3. A Tale of Two Unions: NAAWU and the Uitenhage and District Rugby Union
 Relationships established through rugby for purely social purposes played an important role in cementing
16
 ties between workers which were drawn upon in the union's struggle for recognition and consolidation,
 Rugby was not simply a venue for social interaction between coloureds and Africans, but a site of contact
 between coloured and African workers. Workers in Uitenhage were keen sportsmen, both as participants
 and spectators. Half of those interviewed were themselves either active sportsmen or former athletes
 who had become administrators of sports clubs or leagues. Many more were simply spectators or club
 supporters. Though some of the shop stewards were retired athletes, none were simultaneously involved
 in sport and union activities, as sport consumed too much time.
 Some of the key activists in the formation of the union forged their initial acquaintances and
 friendships on Che rugby field. These relationships could be drawn upon in the workplace by the union
 struggling for recognition. Elijah Antonie, for example, was a player on Swallows football club when
 Albert Gomomo was an administrator, where Gomomo could take the measure of the man.
 Finally, sport provided a forum for conscientization, for the development and transmission of
 ideas which could serve simultaneously to promote athletic and workplace unity between coloured and
 Africans. SARU had been launched in 1962, but made no inroads among African rugby players in the
 Eastern Cape until the early 1970s. KWARU's break with the SAARB was driven by administrative
 conflicts rather than principle, and initially it sought refuge as a separate union under SAARB, rather
 than taking up sides with SARU. It was only in 1973, after the SAARB - backed up by the state -
 rejected KWARU's application that the new union joined the ranks of non-racial rugby. The same
 situation obtained in Uitenhage, where the apartheid rugby bodies remained strong into the early 1970s.
 It was only after KWARU's break from the racial structures that Uitenhage sportsmen began discussing
 the possibilities of non-racial sport.
 KWARU's existence provided an exit option for Uitenhage clubs dissatisfied - for whatever
 reason - with the Uitenhage African Rugby Board (affiliated to the SAARB). After the 1974 season the
 first split in the Uitenhage body occurred when one African team, Star of Hope, jumped to KWARU.
 The breakaway encouraged others to do the same, but sports activists on the Swallows Rugby Football
 Club had a more sophisticated strategy in mind.
 Swallows was a powerful club in the Uitenhage African Rugby Board, fielding impressive teams
 in all three divisions, and it was one of the biggest clubs in the Board. But it also contained a large
 number of consdentized political and union activists who carried ulterior agendas with them into the club
 rooms.
 Elcott Rhoxa was one such activist. He was once a player and later served as an official for
 Swallows. After obtaining his BA degree from the University of Fort Hare, he went to work at
 Volkswagen, where he became a member of the inner circle of union activists around Albert Gomomo -
 who was himself an official of Swallows - building representative shop floor structures for Africans and
 seeking unity with NUMARWOSA. When the union insurgents took over Volkswagen's Liaison
 Committee in the early 1970s, Rhoxa was elected as the Committee's secretary.
 Rhoxa began his career as a sports activist while at Fort Hare playing in the SARU-affiliated
 South Eastern District Rugby Union, where he was recruited to the cause of non-racial sport by the
 Union's vice-president, a staunch member of the Pan Africanist Congress. Rhoxa's adherence to non-
 racial sport grew out of a principle:
 We should integrate so that...a player can be judged according to his merit. We don't
 want to see people playing according to their racial grouping....So we felt we should have
 one board where everybody's talent or performance can be measured, because we
 cannot know if I am a star in an African Board and this one is a star in a coloured
 Board, and you are a star in a white Board, then we will not know actually who is a star
 in South Africa. What we wanted was that South Africa be represented by all people
 who live in it.
 His concluding line not only echoes the Freedom Charter, but also expresses an elemental concept of
17
 justice and equality that individuals should be allowed to succeed or fail on their own merit.
 But the cause of non-racial sport was promoted for both principled and instrumental reasons.
 According to Rhoxa, the attack on racial sporting structures simultaneously promoted the pursuit of
 broader goals extending well beyond the touch-lines on the rugby field:
 I felt that it is important now to politicize our members and the community through
 rugby, and then rugby now will be used as a platform to express our political aspiration.
 Rugby should not be played for mere pleasure. We should now use rugby as a weapon
 to break the racial attitude of the South African government. We felt that we should
 use it as a platform to consdentize our members. Because after we have joined the
 non-racial sport we started talking about the non-racial policy, how important, how
 advantageous it will be to us. And then the principles and policy of the African
 National Congress now also came into meetings, where it was also incorporated in the
 whole thing.
 Where open political organization was impossible in the vicious repression of the 1960s, sport presented
 an opportunity for continued activity.
 If we establish a political organization, it will be in jeopardy, because it's easy for the
 government to ban it. But it will be very difficult for the government to ban sport, and
 say people should not play sport. But now you must use sport as a channel through
 which we can express our aspiration. What was important was that people should be
 rallied in that [non-racial] direction, they should be reorganized and repoliticized, so that
 they should know how we should live in this country.
 Rhoxa worked with a number of other activists. Ronnie Boyce was also a student at Fort Hare, and
 along with other graduates from Lovedale and Healdtown, joined the efforts to bring Uitcnhage's African
 clubs over to the cause of non-racial sport. Boyce was one of the UAWs original complement of four
 shop stewards at Volkswagen. There were also a number of graduates of Limekhaya High School in
 Uitenhage and Newell High School in Port Elizabeth who joined their efforts, the latter spurred on by
 their involvement in the South African Students' Organisation [SASO] and the Black Consciousness
 Movement, both of which were extremely strong in Port Elizabeth. A younger generation of activists,
 raised in the student movement of the 1970s thus found common cause in sport with older workers and
 activists, such as Albert Gomomo and Scoma, who had direct experience of the 1950s and the repression
 of the next decade.
 Rhoxa tells the story of Newell alumni on Swallows who were pressured into non-racial sport as
 a result of growing embarrassment at having to represent an all-African team in the official structures
 while playing friendly matches against their former school-mates, now firmly in the KWARU fold. The
 sports activists manipulated these sentiments to encourage especially the younger players - who usually
 had direct or indirect experience of the student movements burgeoning around the country in the wake
 of the Soweto uprising - to leave the SAARB. In addition to these more political motivations, the sports
 activists generously folded into the mixture more pragmatic arguments calculated to appeal to the
 player's ambitions: that by remaining in the racial structures they would never have the opportunity to
 compete at an international level and that the splits which had occurred - and which would increase -
 would rob them of effective competition as the best teams left for the non-racial fold.40 His story
 provides an excellent example of how micromobilization contexts exert sometimes subtle moral suasion
 on potential recruits to a movement: activists may use the recruit's desire to be well thought of by his
 peers as a way of coaxing him into participation.
 Swallows did break with the Uitenhage African Rugby Board, but instead of following the Star
 of Hope Club into KWARU, the non-racial activists steered the club into the SARU-affiliated Uitenhage
 and District Rugby Union, making Swallows the first African club to join coloured teams in the cause of
 non-racial sport. Rhoxa was elected as the Swallows delegate to approach SARU for applications:
18
 So we decided to join the Uitenhage and District Rugby Union, which was then all
 coloured, now to show the people that we didn't join the SARU fold just because we
 wanted to accommodate ourselves, but what made us join them is that we felt we have
 adopted the principle of non-racialism.
 The principle was not uniformly accepted, however. Some Swallows players and officials stayed behind in
 the Uitenhage African Rugby Board, but even some of those who went over to Uitenhage and District
 maintained their old consciousness.
 The non-racial sport activists used various educational means to change these African
 sportsmen's attitudes. The activists' would often sit down with younger players and "preach" non-
 racialism:
 Now, fresh chaps who have joined us, we are trying to show them that rugby is not
 playing only for mere enjoyment. We have another cause, it is a matter of principle
 now, and that they should not look on rugby just as a sport. Some of them still have
 their racial attitude. When we play there [in the coloured areas] you see them shouting,
 "Coloureds," and all that, they use the word, "hottentots," which is that word for
 coloured. We try and tell them that, no, they should not speak in that manner now, and
 that they should see coloureds irrespective of color as a brother, since we play under
 non-racial sport.
 These discussions would take place informally and casually, among the players and officials:
 When we come to the meetings we are dealing only with the business, but outside when
 we meet, we speak rugby. We also talk about the subject of non-racialism, how
 important and how advantageous to us....We leave rugby practices, as we go home on
 the bus, we are watching sport, and during off-season, when we meet we have parties,
 we also talk about rugby.
 In this way they hoped to consolidate the political project of building non-racial rugby in the town.
 These discussions would include mostly Africans and some coloureds, both non-workers and
 workers, and among the latter would be a considerable number of Volkswagen employees. In addition to
 the activists, there were seven or eight Swallows players who were Volkswagen employees, as were the
 club's President, Secretary, and Vice-Secretary. The activists' preaching thus fell on a broader network
 of Volkswagen workers, who in turn could influence their co-workers.
 But Rhoxa and the Swallows were not alone. Other activists in the UAW, such as Kenny
 Tshaka, were also active in non-racial rugby politics; he later became a member of the UAW and then
 NAAWU Executive Committees at Volkswagen, but in rugby circles he was one of the non-racial
 activists from Star of Hope who chose not to move to KWARU. Instead, he launched a new club, Lions,
 which affiliated to Uitenhage and District. He played a formidable role conscientizing the mostly
 younger players who joined his team in the politics of non-racialism.41
 The efforts by African sports leaders to bring the most important clubs over to the non-racial
 sports movement parallels and reinforces the efforts by workplace leaders to build a united body under
 the auspices of NUMARWOSA. In both cases, coloured activists had built organizations in the 1960s -
 Uitenhage and Districts in the case of rugby and NUMARWOSA at the workplace - and had broken
 free of racially restrictive national bodies - the South African Rugby Federation and the Trade Union
 Council of South Africa - but had not succeeded in bridging the racial divide against which they were
 founded.
 In both cases, on a very local level African activists were themselves developing an inclusive
 policy of non-racialism which encouraged them to look for opportunities for broader unity. Indeed, the
 search for sporting and workplace unity often involved the very same individuals. Albert Gomorao, a
 crucial figure in the organization of workers at Volkswagen was simultaneously a major figure in the
19
 development of non-racial rugby. Gomomo happened to be working close to Chris "Papa" Williams, who
 was simultaneously a key administrator in SARU rugby and the President of NUMARWOSA. Through
 such personal and sometimes rather serendipitous processes activists from both sides in these two
 spheres found each other. Thus the familiarity which preexisted in rugby smoothed the process through
 which coloured and African activists organized on a non-racial basis in the workplace.
 VI: WAR OF MANEUVER: THE GREAT STRIKE OF 1980 AND THE
 CONSOLIDATION OF THE UNION AT VOLKSWAGEN
 Jurrie Harris and the rest of the NUMARWOSA-UAW negotiating team had only a vague idea what the
 next few weeks would hold for them when they convened a general meeting of Volkswagen workers on
 Sunday morning 15 June 1980 in Uitenhage's Jubilee Hall. The meeting was a report-back on the wage
 talks then taking place in the IC in Port Elizabeth. The talks deadlocked on the first day after
 NUMARWOSA-UAW demanded a "living wage" of R2.00 per hour, up from the prevailing minimum of
 R1.15, while the companies offered only R1.30. When the local magistrate banned the report-back as
 well as a hastily planned meeting later that day at a church in Kabah, the workers were incensed, and
 decided then and there to stop work the next day to continue the report back, this time on company
 premises.42
 On 16 June workers showed up for work at the normal time, but gathered on the lawns in front
 of Managing Director Peter Searle's office, to receive the aborted report-back. They directed their shop
 stewards, Harris, Johnson, and Mpushe to negotiate directly with top management, fearing that if they
 left the factory premises they would be locked out, lose their leverage over the company, and with it any
 chance to bargain.
 Instead of a meeting with management, the workers were addressed by the Labour Relations
 manager, Ivan Broderick, who told them management would respond to their demands, but only if they
 went back to work. His intervention was not, to say the least, well received. Scoma relates the events
 which followed:
 He promised Peter Searle that he knows Africans, so he will be able to speak to them.
 Then A&icans showed him that he did not know them. They let him come into the
 group, then they surrounded him, they started to beat him.
 Kwinana was also present at the meeting, and adds the following details:
 He wanted to address the workers himself, that's where he got the beating. Workers
 told him, "We are sick and tired of your tactics. You are not our leader. Our leaders
 are the shop stewards, and...we give our undertaking to the shop stewards to meet Peter
 Searle and top management, not you." We didn't recognize Broderick and the Labor
 Relations Department en bloc. We didn't want our complaints channeled through them.
 We want channels through our shop stewards and to the top management, because we
 wanted a direct reply.
 The workers forcibly rejected management's preferred mode of communication, and replaced it with
 their own chosen form of negotiation. The company had recognized NUMARWOSA and later the
 UAW in the hope of having a negotiating partner who could speak for the entire workforce. Now that
 "partner" was speaking with a unified voice backed up by thousands of workers gathered together on the
 company grounds. These spokesmen were not chosen at the spur of the moment, but had spent ten
 years establishing structures in the factory and building their own credibility as the workers'
 democratically elected leaders.
 The shop stewards remained locked in discussion with management until the early afternoon,
 and returned to the workers without a settlement, whereupon the workforce voted to strike. They
 walked out the gates, and triumphantly marched through "white" Uitenhage following a number of
 different paths back to their respective townships. The largest contingent, made up almost exclusively of
20
 African workers, headed southwest back to Kwanobuhle, crossing a large cement bridge over the
 Swartkops River, and passing a huge, foul-smelling sewerage plant before arriving home in the African
 township. The other group - made up of both coloureds and Africans - headed northwest along Durban
 Street and into the central business district, where they turned north towards Kabah and towards the
 coloured townships stretching off towards the foot of the Groot Winterhoek mountains. At the fork
 where JCamesh and Maduna Roads intersect, the workers separated, each to their own racially segregated
 residential areas: Africans and some coloureds towards Kabah, the mass of the coloureds towards the
 townships of Gerald Smith, Thomas Gamble, and Rosedale. The geography of the workers* march out
 of the factory itself highlights the racial and residential divisions between workers which elsewhere have
 undercut concerted worker action, but which Volkswagen workers had been able to overcome.
 Gcinuhlanga Gxowa participated in the walkout and march through town. T o me it was a new
 experience," he said, "because I was used to the school strikes, and it was my first year at work." Student
 strikes in 1977 had been violently suppressed by the SAP, but Gxowa was struck by the "system's"
 different reaction to the Volkswagen workers.
 Actually the police were just standing in the center of the town, when we were passing
 and moving as that huge group of people from work, moving through the town. They
 were standing. They couldn't do anything, because we were the strongest force that they
 had ever seen in Uitenhage, and it was just orderly, so nobody was fearing anything.
 Kwinana marched with some measure of pain and foreboding.
 I know what happens, I know the aftermath of a breakdown, what follows, because I
 know. I am the first one to be sacked, because they know they don't want me at the
 factory, because I was convicted for ANC activity before.
 Nonetheless he was jubilant.
 We are commanding. We are commanding. The workers are commanding now. No
 more the police or the Labour Relations Department.
 A new era of industrial relations had begun.
 The strike continued for another three weeks, during which time the workers met daily in the
 car park in front of the plant. There they would gather to receive report backs from the negotiators, and
 make collective decisions. In such a way workers could recreate the "collective worker" identity normally
 imposed by the factory and its division of labor, but weakened once they walked out the gates. The
 Uitenhage Black Civic Organisation, a civic body dominated by NUMARWOSA-UAW shop stewards,
 provided an organizational focus in the townships, holding meetings, distributing leaflets, and conducting
 house visits. Furthermore, the informal networks among shop stewards and workers, as well as the ANC
 cell system developed during the 1970s provided communication and support links between workers
 which bolstered the strike. The strike was eventually settled through the intervention of an IMF
 negotiator who held talks with Rademeyer to broker an agreement acceptable to both sides.43
 The workers' return to Volkswagen was as significant as their departure, because it was as
 uneventful as the walkout had been dramatic. Once the union had agreed to the settlement, all the
 strikers returned to work en masse, leaving no doubt that they were the true force representing the
 workers. During the strike the company had tried many times to crack the workers' solidarity: the gates
 remained open, and it offered to grant the wage increase to any worker who returned; VW even hired a
 helicopter to drop leaflets in the townships promoting their offer. No more than fifty or sixty workers
 responded. But once the union made its agreement, everyone returned.
 The conclusion of the strike reaffirmed and validated the workers' initial motivations to act: they
 had risked a great deal by striking, and had not only won important concessions, but emerged intact with
 no victimization by the company nor by the police. The lesson regarding the advantages of disciplined
 collective action was evident for all.
21
 In the years following the strike, the newly-merged union, NAAWU, would use the its
 resources - the strength and legitimacy of the shop steward system, informal networks, international
 intervention, and most importantly, the solidarity between African and coloured workers - again and
 again in a protracted "war of position" with management. In the immediate aftermath of the strike it was
 the first workplace where the new unions won the right to full-time shop stewards, workers paid by the
 company, who spent their entire working day involved in union affairs.44 By the mid-1980s NAAWU
 had 5 full-time and 26 part-time shop stewards, giving the union an extremely deep base on the shop
 floor. In building off of its initial success in 1980, it won further concessions regarding disciplinary and
 retrenchment procedures and protection from supervisorial abuses which provided substantial benefits to
 its members and expanded workers' control.
 VI: CONCLUSION: SHOPFLOORS, RUGBY FIELDS, AND WORKER
 SOLIDARITY
 The structure of residential life in Uitenhage and Port Elizabeth produced in Uitenhage many
 opportunities for cross-race contact between Africans and coloureds, while producing a far more
 thorough segregation in Port Elizabeth. The informal and formal networks described by the concept
 micromobilization context were heavily shaped by these social patterns, and tended to reproduce them.
 In Uitenhage these social networks tended to include Africans and coloureds, while in Port Elizabeth
 they did not, and tended to remain bottled up in workers' racially defined communities. In short, the
 micromobilization contexts - and the patterns of familiarity, trust, and affinity - crucial for building the
 union in Uitenhage were defined as much by class as by race, while in Port Elizabeth, race was a far
 more prominent factor.
 Relationships between African and coloured union activists in the workplace were thus facilitated
 by a range of prior contacts growing out of Uitenhage's racially mixed Kabah Location. These contacts
 were most pronounced in the micromobilization context of rugby. Rugby brought ordinary workers
 together in meaningful relationships which provided Africans and coloureds sustained involvement in
 each others' social lives away from work. At the level of union leadership, coloured rugby players served
 as intermediaries introducing African union activists to coloured shop stewards, softening their mutual
 distrust. In these ways, coloureds and Africans at all levels of the workforce could face each other as
 familiars rather than strangers; as acquaintances rather than with steely disregard for the alien "other;" as
 people with a store of common experiences rather than merely antagonisms; people to be included in a
 common organization around shared interests. The social experience of sport, as well as the broader
 social linkages sketched above constituted a set of bridging experiences and networks between coloureds
 and Africans, providing both the foundation and the means upon which a non-racial union could be built.
22
 NOTES
 1. The passage is quoted from Athol Fugard, John Kani, and Winston Ntshona, Sizwe Bansi is Dead, in
 Athol Fugard, Statements (New York: Theatre Communications Group, 1986), p. 7.
 2. Sidney Tarrow, "Struggle, Politics, and Reform: Collective Action, Social Movements, and Cycles of
 Protest," Western Societies Program Occasional Paper No. 21 (2nd edition), Center for International
 Studies, Cornell University, 1991; and Bert Klandermans and Sidney Tarrow, "Mobilization into Social
 Movements: Synthesizing European and American Approaches," in From Structure to Action: Comparing
 Social Movement Research Across Cultures, edited by Bert Klandermans, Hanspeter Kriesi, and Sidney
 Tarrow (Greenwich, CT: JAI Press, 1988).
 3. For an argument on the distinctiveness of new social movements, see Alain Touraine, The Voice and
 The Eye (Cambridge: Cambridge University Press, 1982); Ronald Inglehart, "Values, Ideology, and
 Cognitive Mobilization in New Social Movements," in Russell J. Dalton and Manfred Kuechler,
 Challenging the Political Order New Social and Political Movements in Western Democracies (Cambridge:
 Polity Press, 1990), pp. 43-66 and Dieter Rucht, The Strategies and Action Repertoires in New
 Movements," in Dalton and Kuechler, Challenging the Political Order, pp. 156-175. The strongest
 criticism against the newness of NSMs has been levelled by Gunnar Olofsson, "After the Working-class
 Movement? An Essay on What's 'New* and What's 'Social' in the New Social Movements," Ada
 Sociologica 31:1 (1988).
 4. Dalton, Kuechler, and Burklin, The Challenge of New Movements," p. 9.
 5. Klandermans and Tarrow, "Mobilization into Social Movements," pp. 13-14.
 6. Klandermans and Tarrow, "Mobilization into Social Movements," p. 10.
 7. McAdam explains cognitive liberation in terms of three interrelated cognitions: that people who
 ordinarily accept the legitimacy of authority come to believe that their rulers are unjust; that ordinarily
 fatalistic people, who accept the inevitability of existing arrangements, begin asserting rights that imply
 demands for change; and that people who ordinarily think of themselves as helpless come to believe they
 have the capacity to change their lot. In the absence of any of these three cognitive developments,
 people will lose either the motive or will for collective action. Doug McAdam, Political Process and the
 Development of Black Insurgency (Chicago: University of Chicago Press, 1982).
 8. Doug McAdam, "Micromobilization Contexts," in From Structure to Action edited by Klandermans et
 al., p. 134-135.
 9. Doug McAdam, "Recruitment to High-Risk Activism: The Case of Freedom Summer" American
 Journal of Sociology 92:1 (July 1986).
 10- McAdam, "Micromobilization Contexts," p. 136.
 11. McAdam, "Micromobilization Contexts," p. 151.
 12. V.M. Kwinana was a high school teacher in East London, and was the dominant figure in the town's
 more conservative "Congress B" faction; he also led the opposition to the activist boycott strategy
 endorsed by the 1949 Cape ANC Conference. Interview with V. Kwinana; Lodge, Black Politics, p. 57.
 13. In December 1955 131 members of the SACTU-affiliated Textile Workers' Industrial Union at the
 Fine Wools struck work in protest of forced overtime, after the company refused their demand for an
 overtime rate of 2 shillings per hour. The workers were convicted of conducting an illegal strike, but the
 union chose to fight their convictions in the Grahamstown Supreme Court, which upheld their appeal.
 Not only did the workers win a wage increase, but through their legal action they established an
 important precedent that workers could not be compelled to work overtime unless their contract
23
 specified such an obligation. See Ken Luckhardt and Brenda Wall, Organize...or Starve! The History of
 the South African Congress of Trade Unions (New York: International Publishers, 1980), 230; "Judges
 uphold appeal by Native strikers," Eastern Province Herald 23 May 1956. For years afterward, the
 TWIU's victory would resonate in Uitenhage workers' historical memories.
 14. The lectures most likely occurred in the alternative schools set up by the ANC during the boycott of
 Bantu education after 1955.
 15. The International Metalworkers' Federation provided a small grant to support the work of a part-
 time organizer. The man TUCSA entrusted with this task was Edgar Deane, secretary of the TUCSA-
 affiliated furniture union. Despite his later prominent role in the politically moderate and muddled
 TUCSA, where he was a member of the NEC, Deane had a solid left background in coloured politics,
 having played leadership roles in the Franchise Action Council, in the Congress-allied South African
 Coloured People's Organization, and in the Labour Party. See R.E. van der Ross, The Rise and Decline
 of Apartheid: A Study of Political Movements among the Coloured People of South Africa, 1880-1985
 (Cape Town: Tafelberg, 1986), 255-278, 307; Gavin Lewis, Between the Wire and the Wall: A history of
 South African 'Coloured'politics (Cape Town: David Philip, 1987), 267-274; Thomas Karis and Gail
 Gerhart, eds, From Protest to Challenge, v. 3, (Stanford: Hoover Institution Press, 1977), 12.
 16. Most prominently, the Western Province Motor Workers' Union [WPMWU]. The search for allies
 brought NUMARWOSA into contact for the first time with the Durban grouping of fledgling unions
 organized by the Trade Union Advisory Coordinating Committee [TUACC], an alliance which would
 later provide the backbone of the Federation of South African Trade Unions and of the hegemony of
 shop-floor based industrial unionism in the new South African labor movement.
 17. National Union of Motor Assembly and Rubber Workers of South Africa, Minutes of the Executive
 Council Meeting, 26 November 1972, "Organising of African Labour."
 18. National Union of Motor Assembly and Rubber Workers of South Africa, "Minutes of the Executive
 Council Meeting," 24-25 March and 18-19 August 1973. Fred Sauls, NUMARWOSA's General
 Secretary, along with the Union's secretariat drew up the draft constitution of the UAW, which was then
 presented to the first general meeting of the new union for amendment and approval.
 19. The links between NUMARWOSA-UAW and WPMWU grew closer through their joint experience
 in the South African Coordinating Council of the International Metalworkers Federation. In this context,
 both groupings lined up on the same side in opposition to the racist Confederation of Metal and Building
 Unions [CMBU], and slowly began to realize their common interests. They would be thrown together
 again in the talks which preceded the creation of the Federation of South African Trade Unions. For a
 discussion of the SACC, see Eddie Webster, Cast in A Racial Mould: Labour Process and Trade
 Unionism in the Foundries (Johannesburg: Ravan Press, 1985), Chapter 8: "The International Factor: The
 IMF in South Africa, 1974-1980."
 20. The company employed coloured workers as a result of pressure from the government's threatened
 implementation of a Coloured Labour Preference Policy for the Eastern Cape.
 21. For these reasons paison Committees would be the type of "reform" most preferred by management
 after the 1973 Durban strikes, when the old system of excluding Africans from the industrial relations
 system became palpably unworkable. Volkswagen was one of 118 companies who introduced "non-
 statutory" Liaison Committees well in advance of the strikes.
 22. See the interview with Gomomo, "Union Organisation in Volkswagen," South African Labour Bulletin
 11:1 (September 1985), and "Profile: John Gomomo," South African Labour Bulletin 15:3 (September
 1990).
 23. Indeed, John Gomomo went on to become Uitenhage organizer for the Ciskei National
 Independence Party - in the eyes of both management and the state, hardly a dangerous political
 association. He went on to become a full-time shop steward at VW, Vice-President and then President
24
 of the autoworkers' union, and is today President of the Congress of South African Trade Unions. In
 1990 he, along with other prominent ex-FOSATU worker leaders publicly announced their supposedly
 long-standing affiliation to the South African Communist Party.
 24. This would include actual ANC cell members and the broader network of workers with whom they
 had contact, but who were unaware of their actual identity.
 25. The action was called to protest the declaration of a state of emergency in the Port Elizabeth and
 Uitenhage magisterial districts in the aftermath of the New Brighton riots. The riots were in turn
 prompted by intensifying state repression against the Defiance Campaign and efforts by the Native
 Affairs Department to impose the provisions of the Urban Areas Act on Port Elizabeth.
 26. The other shop stewards were Boyce, Mpushe, Harris, Reddiffe, Spiers, and Isaacs.
 27. The company soon organized a European tour for the senior shop stewards from the unions,
 including Yster en Stal, to observe industrial relations systems in Germany, Belgium and Great Britain.
 Gomomo made extremely effective use of the trip, cementing bonds to the IG Metall union in Germany
 and the International Metalworkers Federation which would serve the union for years to come.
 28. For further discussion of the labor process and the racial division of labor in the automobile plants,
 see Glenn Adler, "Skills, control, and 'careers at work': Possibilities for worker control in the South
 African motor industry" South African Sociological Review 5:2 (1993).
 29. For a complete discussion of the application of the Urban Areas and Group Areas Acts, see Chapter
 4 of my doctoral thesis, "Racial Segregation and its Impact on Working Class Communities in Port
 Elizabeth and Uitenhage."
 30. The state's most recent effort to remove Kabah was facilitated by the nation-wide crackdown under
 the 1986 State of Emergency. Despite moving approximately 35,000 people from Kabah to a site and
 service camp on the outskirts of Kwanobuhle, not all Africans were removed from Kabah - in particular
 the residents of some 700 formal houses in the McNaughton section of the old Location remained. As
 the emergency waned, as migrations from the rural areas continued, and as the powers formerly granted
 under the Urban Areas and Group Areas Acts fell away, the state was helpless to prevent the massive
 re-population of Kabah, the situation which obtains today.
 31. For further discussion of racial segregation in Port Elizabeth, see AJ. Christopher, "Formal
 Segregation and population distribution in Port Elizabeth" Contree 24 (1988); "Apartheid Planning in
 South Africa: The Case of Port Elizabeth" The Geographical Journal 153:2 (July 1987); and "Race and
 Residence in Colonial Port Elizabeth" South African Geographical Journal 69:1 (1987). See also Jennifer
 Robinson, The Power of Apartheid: Territoriality and State Power in South African Cities-Port
 Elizabeth, 1923-1972," Ph.D. diss., Cambridge University, 1990; and Bev Taylor, "Removing Port
 Elizabeth's Notorious Korsten Slums in the 1950s," and "South End and Fairview Removals and the
 Development of Bethelsdorp," unpublished papers (1990).
 32. The figures from General Motors must be treated with great caution, as the research was interrupted
 by a major strike over the firm's proposed disinvestment, when fewer than half the projected interviews
 were completed. For a discussion of the strike, see my article, "What's Good for General Motors?: Black
 Workers' Response to Disinvestment, October-November 1986" Journal of Southern African Studies 15:3
 (April 1989). For a discussion of the effects of the strike on research possibilities see my article, The
 Politics of Labor Research During a Liberation Struggle: Interviewing Black Workers in South Africa," in
 International Annual of Oral History, 1990 edited by Ronald J. Grele (Westport, CT: Greenwood
 Publishing, 1993).
 33. Marianne Roux, "Perspectives and Values of Industrial Workers: The Emergence of Socio-Economic
 Consciousness Amongst Coloured Workers in the Motor Manufacturing Industry in the Eastern Cape,"
 Ph.D. diss., Stellenbosch University, 1977, p. 163, 159.
25
 34. The centrality of sport, and rugby in particular, in white culture and politics in South Africa remains
 entirely undeveloped as a subject of academic inquiry. The forthcoming work by Dr. John Nauright, of
 the University of Otago in New Zealand should rectify this serious gap in our understanding of South
 African society. For a brief discussion of the role of sport in Afrikaner politics in the 1960s and 1970s,
 see Wilkins and Strydom, The Super-Afrikaners, Chapter 14: Sports Policy. The debate among
 Broederbonders and Nationalist politicians around the inclusion of Maori players on the New Zealand
 rugby team constituted one arena of conflict which precipitated the 1969 split in the National Party,
 leading to the formation of the Herstigte Nationale Party.
 35. For social histories of sport in South Africa, see Tim Couzens, "An Introduction to the History of
 Football in South Africa," in Town and Countryside in the Transvaal: Capitalist Penetration and Popular
 Response, edited by Belinda Bozzoli (Johannesburg: Ravan Press, 1983); Andre Odendaal, "South
 Africa's Black Victorians: Sport and Society in South Africa in the Nineteenth Century, in British Culture
 and Sport at Home and Abroad, 1700-1914, edited by ?; and Jeff Peires, "Rugby in the Eastern Cape: a
 history," Work in Process 17 (April 1981).
 36. See Odendaal, "South Africa's Black Victorians," 195.
 37. Neither rugby nor cricket caught on with Africans in Natal, the Orange Free State, or the Transvaal.
 Odendaal accounts for the lack of interest in Natal by pointing to the very prominent role of American
 rather than British missionaries in African education in the Province, while the lack of interest in the
 Afrikaner republics to the north may stem from a similar lack of penetration by English missionaries and
 the far stronger restrictions on African education and an African middle class there. In Peires' words,
 rugby "was never adopted by the northerners who saw something ridiculous in the spectacle of a grown
 man running around with a ball in his hands." Peires, "Rugby in the Eastern Cape," p. 1.
 38. Peires, "Rugby in the Eastern Cape," 3, 5.
 39. Years after the event, black workers still spoke fondly of the game in the 1970s between an African
 side composed of players from the official SAARB and a touring British side, where the visitors won
 convincingly, but in which a Volkswagen player, (and later full-time shop steward) Aaron Balintulo
 scored a try. By contrast, the workers would laughingly emphasize, the British team completely
 overwhelmed the all-white Springbok rugby team, which was shut out.
 40. The Springboks held the claim to represent South Africa as a whole and the racial bodies were not
 recognized overseas.
 41. Tshaka played a surprisingly similar role in the factory. He was never elected as a shop steward, but
 was extremely important behind the scenes as an activist and "conscientizer" among the rank-and-file.
 42. For a detailed chronology and a superb analysis of the factors leading to the strike, see Martheanne
 Finnemore, The mobilisation of Black workers in the auto industry in the Eastern Cape, 1970-1983," MA
 diss., University of Port Elizabeth, 1984.
 43. The IMF negotiator was Karl Casserini. The real impasse occurred when the company increased its
 offer to R1.40 per hour, which was rejected by the workers, who remained out on strike. After that
 Volkswagen threatened to fire workers who remained out, but delayed action while the talks with
 Casserini gathered mpmentum. Eventually the company increased its offer to R1.42, and agreed to meet
 the R2.00 living wage target within 18 months; though falling well short of their original demand, the
 workers accepted the offer and returned to work.
 44. Volkswagen was the first company to agree to the demand, and was preparing to implement the
 agreement, but Ford in Port Elizabeth jumped first and announced their agreement, and have been
 wrongly credited with the innovation.
UNIVERSITY OF THE WITWATERSRAND
 INSTITUTE FOR ADVANCED SOCIAL RESEARCH
 SEMINAR PAPER
 TO BE PRESENTED IN RICHARD WARD BUILDING
 SEVENTH FLOOR, SEMINAR ROOM 7003
 AT 4 PM ON THE 18 SEPTEMBER 1995
 TITLE: 'Render unto Caesar1: The central state,
 local government and struggles over
 segregation in Port Elizabeth and Uitenhage,
 1948-1962
 BY: GLENN ADLER
 NO: 386
xzz
 ? ? WHITE MUNICIPAL HOUSING SCHEMES
 ^ ^ COLOURED MUNICIPAL HOUSING SCHEMES
 M 5 I AREAS OF COLOURED AND AFRICAN RESIDENCE
 L * ? I AREAS OF WHITE RESIDENCE
 f ." I AFRICAN LOCATIONS
 SCHAUDERV1UE
 'DOWERVULE
 -McLEANVILU
 -KENSNGTONVILLE
 North End
 n\
 P O R T E L I Z A B E T H
 IENO
 Map 1. Pdrt Elizabeth Residential Areas
 and Municipal Housing Schemes, 1943.
 . UITENHAGE .
 Map 2. Uitenhage: Town and Locations, 1923
SLACK RESIDENTIAL AREAS
 COLOURED RESIDENTIAL AREAS
 |V?$ f l ASIAN RESIDENTIAL AREAS
 |_ ' . | WHITE RESIDENTIAL AREAS
 \ / / \ INDUSTRY
 Mount Road
 P O R T E L I Z A B E T H .
 Mill Park
 KWA NOBUHLE | | BLACK RESIDEWT1AL AREAS
 COLOURED RESIDEMTIAL AREAS
 | . ' .] WHITE fiESIDEMTIAL AREAS
 Y//A
 Map 3. Port Elizabeth Group Areas,
 in the period of high segregation, c. 1975
 Map 4. Uitenhage Group Areas, c. 1970
I. INTRODUCTION*
 In early December 1968 the Town Council of Uitenhage began the first forced removals of
 African people from the 124 year-old inner-city location of Kabah to the new township of Kwanobuhle
 on the southern Municipal boundary. The removals were the beginning of a comprehensive plan to
 make UitenHage conform to apartheid urban policy principles by removing Africans from Kabah, where
 they had historically lived side-by-side with coloureds, and reproclaiming the old location as part of the
 growing coloured Group Area of the town. However, in the two decades that followed these removals,
 Kabah was never fully conquered by the state. Both central and local governments perennially failed to
 realise the goal of comprehensive social engineering due to shortages of funds, bureaucratic inefficiency,
 continued migrations from Cape country districts, and the stubborn efforts of ordinary people to stake a
 claim to urban space. By contrast, in 1968 Port Elizabeth was well along the path of near-total
 segregation. A decade earlier Africans had been virtually entirely removed to proclaimed townships, and
 in the late 1960s the Municipality was embarking on a program of removing coloureds from "white"
 areas. The result of these programs was that Port Elizabeth soon became among the most segregated
 cities in South Africa.
 The Kabah removal was the culmination of more than twenty years of policy struggles within the
 South African state, and explicit rejection by the Uitenhage Town Council of both the Urban Areas and
 Group Areas Acts. It is quite remarkable that at the moment of greatest National Party power in the
 1960s it proved quite unable to remake the world according to its will. Not only did it face profound
 resource problems and resistance from below, rendering the "purist vision" of apartheid unattainable in
 the short term, but the institutional means available to the central state to enforce its policies required it
 to rely on relatively undependable local authorities as essential sites of delivery of services and exercise of
 control. In July 1961, as the Uitenhage Town Council was about to conceded to central state pressure,
 the local pro-Government newspaper Die Oostcrtig reminded the Councillors, "'Render unto Caesar that
 which is Caesar's' is a good idea for them to bear in mind."1
 Local authorities were nominally responsible for maintaining immediate conditions of
 accumulation for industry and commerce, and for the reproduction of the white as well as black working
 classes. These tasks were conducted, if not more commonly avoided, under severe structural constraints
 of fiscal dependence on white ratepayers. This dependence would lessen after 1973 when the central
 state ultimately usurped authority over the townships, but until then the NP government ruled through
 local institutions to which it could not simply dictate. Local constituencies heavily shaped Municipal
 responses to central policies, and local officials varied widely in their ability or willingness to comply,
 especially within UP strongholds. A politics of accommodation was necessary in which policy outcomes
 were the result of a rough and tumble political process of coercion and concession, though increasingly
 on terms defined by the central state. Nonetheless, it was a process of accommodation, which introduced
 a considerable measure of variability in the application of central policies.
 If the central state could, in the end, demand that local authorities "Render unto Caesar," this
 seemingly confident assertion of power reveals in fact its opposite. In St. Matthew's account, Jesus' reply
 to the Pharisees marks the difference between a merely earthly power with its essential limitations, and
 that of a transcendent, omnipotent God. Similarly, once confronted with a challenge to its authority the
 central state had little to fall back on other than a dictatorial exercise of power, even vis-a-vis its own
 institutions, revealing the shallow and fragile nature of its own sovereignty. In this respect the invocation
 of ultimate authority marks its very limitation, as ultimately the central state was forced to rely on these
 very same institutions to exercise its power. This paper looks backward from the Kabah removals,
 tracing the genesis of the segregation policies themselves to demonstrate the essentially contested nature
 of state power in South Africa.
 "A previous version of this paper was presented in 1988 at the seminars of the Southern Africa
 Research Program at Yale University and at the Postgraduate Seminar of the Institute of Commonwealth
 Studies, University of London. The paper is drawn from a chapter of my doctoral dissertation, "'The
 Factory Belongs to All Who Work in It': Race, Class, and Collective Action in the South African Motor
 Industry, 1967-1986," (Columbia University, 1994).
As such, the research is influenced in important ways by Deborah Posel's insights into the nature
 of the South African state, particularly in her assessment of the state as an actor in its own right
 "engaged in ceaseless processes of struggle and accommodation, both internally and with subordinate as
 well as dominant classes."2 It is curious that while her arguments about struggles over the meaning of
 apartheid have gained wide currency, there has been far less recognition of her invitation to investigate
 more thoroughly the coherence of the apartheid state itself.
 This paper thus undertakes an institutional analysis of the development of policies of segregation
 in Port Elizabeth and Uitenhage. In addition to giving a sustained account of the evolution of such
 policies over the first decades of NP rule, it also self-consciously locates the discussion far away from the
 mainstream of debate on apartheid. Though both cities experienced rapid population increases
 throughout the period under review, the migrant labour system played a negligible role in the regional
 economy. In this respect, influx control and the contradictory relations between administrators,
 employers, and migrant and "urbanised" workers are relatively less important than in any other
 metropolitan area in South Africa. Instead, territorial apartheid and residential segregation was a far
 more central issue facing the central state and local authorities, as both attempted to deal with the racial
 ordering of space. Though this is an important issue elsewhere, arguably it is of greater salience in the
 Cape Province, given the enduring legacy of African, coloured, and Indian property (and franchise)
 rights. An immediate problem confronting the NP government after 1948 was reversing and eliminating
 patterns of residential integration, between whites and blacks and between those statutorily defined as
 Africans, coloureds, and Indians.
 In this respect, the paper makes a departure of sorts by examining the virtually simultaneous
 enactment of the web of apartheid legislation in the two municipalities. Other analysts have tended to
 focus exclusively on either the pass laws, or the Urban Areas Act, and few have linked their
 implementation to the enforcement of the Group Areas Act. As a result scholars have missed the
 combined impact of these policies on localities.
 n. URBAN SEGREGATION BEFORE APARTHEID
 A. SEGREGATION IN COLONIAL UITENHAGE AND PORT ELIZABETH
 Segregation in Port Elizabeth and Uitenhage dates from their founding. Cape colonial towns
 were the realm of white settlers, but many different groups were present at their start.3 The black
 population included many Malay artisans and traders, as well as Khoi labourers, and initial residential
 divisions did not exclusively follow colour lines.4 AJ. Christopher argues rather that informal patterns of
 class and social differentiation characterized residential growth in Port Elizabeth: mixed-race people were
 free to own land and some people of colour clustered in new industrial areas where many white workers
 also lived.5
 However, white colonists did not consider the majority of indigenous people - "Africans" - to be
 citizens of the colony capable of owning property, and more importantly, viewed them as foreigners,
 temporary residents present only to labour. Africans were relegated to segregated, physically isolated
 "locations" on land deemed unsuitable for white occupation. In both Port Elizabeth and Uitenhage the
 first separate residential areas - initially for Khoi and later for Mfengu - were established on London
 Missionary Society ground close to the town centres.6 After being created as Municipalities in the 1840s
 both towns established formal locations close to these original LMS locations.7 Over time, these
 locations expanded and many entrepreneurs opened private locations to accommodate the increased
 African population.
 By the time of Union in 1910 the pattern of segregation in Port Elizabeth diverged from that of
 its neighbour. The inner-city locations were the target of a sustained removal effort, perpetually delayed
 by the absence of funds to compensate site holders or to provide alternative accommodation in new
 areas. The removals finally occurred under the pretext of plague clearance, and most Africans in Port
 Elizabeth were moved to New Brighton, 8 kilometres from the centre of town.8 But a loophole in the
 new law - Africans could be compelled to leave town but not to enter the location - enabled many to
 avoid the heavily controlled and expensive barracks housing at New Brighton and to move instead to
Korsten, a land development beyond the Municipal boundaries.9
 Despite population increases, Uitenhage's original locations - including the oldest, Kabah - were
 never forcibly removed.10 These were built on the periphery of the town, and did not impede the
 further expansion of either commercial or white residential areas. Furthermore, there was little upward
 pressure on land prices - an important stimulus for the white speculators behind the removals campaign
 in Port Elizabeth - as Uitenhage experienced a prolonged recession due to the chronic decline in the
 wool trade on which the town's economy was based.
 Thus at the end of the colonial era in 1910, Uitenhage retained its first town locations on their
 original sites within the Municipal boundaries, the most important being the Kabah Location. By
 contrast, Port Elizabeth had passed through a period of prolonged struggle over the inner city locations,
 eventually leading to their destruction and the removal of the African population beyond the town border
 to the formal Location at New Brighton or to the growing squatter settlement at Korsten.
 B. INDUSTRIAL DEVELOPMENT AND URBAN SEGREGATION BEFORE 1948
 These differences in segregation patterns would increase as the two towns took steps to
 accommodate the rapid increase in the black population that accompanied industrialisation before and
 after World War II. Both Municipalities designed new housing schemes for African and coloured
 people, but these tended to reinforce the trends mentioned above. The Port Elizabeth Municipality built
 new housing for Africans, but within New Brighton; furthermore, its first foray into providing
 subeconomic housing for other workers yielded a series of segregated coloureds-only and whites-only
 housing schemes scattered throughout the city. The only exception to this pattern was Korsten. In
 Uitenhage, after consolidating four locations into the "old location" at Kabah in 1936, Municipal housing
 schemes for Africans and coloureds were located either in or adjacent to Kabah. As a result, Port
 Elizabeth's policies promoted a more thorough-going segregation between coloureds, Africans and
 whites, while in Uitenhage the divide was drawn between blacks and whites.
 1. Port Elizabeth
 As Jennifer Robinson's research shows, local industrial development sparked by World War I
 was not an automatic or natural reflex to market conditions, but bore the imprint of an entrepreneurial
 local authority seeking to promote Port Elizabeth industry and seize whatever opportunities were open to
 them. The expansion of the footwear industry and the first investments by the automobile multinationals
 were heavily encouraged by a Council willing to assume large debts to fund the expensive infrastructure
 essential to manufacturing and to offer fully serviced industrial townships ready for development. The
 resulting growth in the manufacturing workforce also prompted the Council for the first time to provide
 housing for Port Elizabeth's working class, and the Council's policies for public housing and for
 administering New Brighton Location earned it a reputation (which its promoters eagerly sought to
 embellish) as a liberal, progressive city.
 While relying on the physically separated racially exclusive Location, the "liberal" Council
 eschewed the familiar mechanisms of pass controls for Africans formalized in the 1923 Urban Areas Act
 in favour of a paternalistic and pacifying administrative system, codified in extensive regulations executed
 through a personalistic Superintendent backed by a network of African "headmen."
 If the physical segregation of Africans within urban areas was accomplished through an inherited
 framework of colonial restrictions, culminating in the Urban Areas Act of 1923, separating whites from
 all other people of colour was a more complicated process. Some Africans, coloureds, Indians, and
 Malay possessed property and voting rights in the Cape. But following World War I, the informal
 processes of segregation which date from the earliest colonial times intensified as a result of two separate
 processes. First, new private housing developments, especially those aimed at white buyers, began
 featuring restrictive covenants in their title deeds. The new language prevented purchase by anyone
 other than the defmed group, which in the vast majority of cases limited ownership to whites.11
 The other factor heightening the extent of separation between white and coloureds was the
 character of state provision of working class housing. The 1920 Housing Act enabled municipalities to
construct "economic" and "sub-economic" housing, but it mandated that such schemes be constructed on
 racially separate lines. The Port Elizabeth Municipality developed a number of such housing projects
 between 1923 and World War II, during which time more than 1,400 houses were built for Europeans,
 2,000 for coloureds, and nearly 4,000 for Africans.
 Those built for Africans were constructed as further extensions to New Brighton Location,
 known as White Location (310 houses begun in 192S) and McNamee Village (3,506 houses, built between
 1938 and 1947). But the new Council houses designated for coloureds and whites were scattered around
 the city on available land, without reference to any overall framework of spatial separation of the
 races.12 [See Map 1] These single-race schemes were often cited adjacent to one another, divided by a
 road or minor physical buffer strip, in contrast to the principles of broad racial zoning characteristic of
 later apartheid policies which saw whole sections of cities reserved for particular groups. Indeed, some
 of the Council housing schemes for coloureds were located in the southern and largely white residential
 areas of the city. These projects formed literal islands of racial occupation, with little intermixing
 between coloureds from the housing schemes and whites from the surrounding private developments. In
 1936, while there were several areas in the city where substantial numbers of coloureds resided near
 whites, such as South End, North End, and Sidwell, (though rigid street-level separation predominated),
 there was only one area within the city boundaries where large numbers of Africans lived: Korsten.
 Korsten's fast population growth and generally unregulated character contributed to the creation
 of a densely settled, unserviced area with high rates of disease and infant mortality. Officially Korsten
 was described as a slum, and the Municipal health department repeatedly lobbied the Council to have it
 removed on such grounds. Slum or no, Korsten was still a relatively attractive area for people who
 couldn't afford or obtain, or didn't want housing within New Brighton. It still lacked the more overt
 forms of Council control present in the location, and it was extremely close to industry. The very
 conditions which favoured settlement by Africans made the police anxious: they were worried by the
 absence of clear administrative jurisdiction over Korsten, by the irregular patchwork of wood-and-iron
 shack settlement making direct control difficult, and by the "intermixing" of Africans and coloureds which
 meant repeated violation of laws restricting liquor consumption by Africans. Many members of the Port
 Elizabeth City Council shared these concerns, and an additional interest: inspired by the expanding
 manufacturing sector, the Council eyed Korsten's central location as a prime spot for development of
 new industrial townships, converting "slums" into revenue-producing factory land.
 Once Schauderville (for coloureds) and McNamee Village (for Africans) were launched in 1935
 and 1938, the Municipality was able to begin condemning Korsten houses under the Slums Act and
 moving their inhabitants to the new segregated townships. But the removals never succeeded in
 emptying Korsten, nor in remedying the "slum" conditions which had nominally motivated the Council.
 In 1946 the population of Korsten had declined, but there were still 10,225 Africans and 3,614 coloureds
 resident in the area. Not only had the population of the town vastly expanded during the war, but
 according to Robinson many of the people who moved to McNamee Village and Schauderville were
 unable to afford the rents charged and ended up moving back to Korsten or to new squatter settlements
 even farther from town at Veeplaas, Klemskool, Missonvale, and Bethelsdorp. In 1951 the African and
 coloured population in Korsten in fact returned to 1936 levels.
 Thus, before World War n, the local government of "the Progressive city" had already embraced
 full racial segregation in its housing policies, albeit with important differences from the apartheid policies
 which would be enacted by the National Party government after 1948. The principle of segregation was
 applied to Municipal housing schemes, and tolerated, if not encouraged in the provision of restrictive
 covenants, but outside of the Location there was not yet a comprehensive vision of full-scale separation
 of residential areas. Certainly the scattered islands of racially exclusive housing estates created tricky
 planning problems in later years when central and local planners sought to create large, unified single
 race residential zones. But in its early embrace of separation between Africans, coloureds and whites,
 Port Elizabeth had a well-established tradition of segregation which would ease the Council's
 accommodation with post-1948 policies.
 2. Uitenhage
 In contrast with Port Elizabeth, the bulk of Uitenhage's industrial growth and population
increase occurred after World War II. Prior to that time the Council initiated no Municipal housing
 schemes, nor did it establish a formal African location. Furthermore, coloureds and Africans lived in the
 same Municipal Locations. In part Uitenhage's anomalous development stemmed from its small size and
 its delayed manufacturing growth, both of which limited its fiscal base. Within the Locations a clear
 policy of laissez-faire characterized official policy regarding black residence. After isolating Africans and
 coloureds in areas separated from whites, the Municipality paid scant attention to even the minimal levels
 of investment and administrative attention required by the "Location Strategy" used in Port Elizabeth.
 In the 1920s blacks resided in four proclaimed Locations in Uitenhage: Kabah (the largest),
 Doornhoek, Gubbs, and Oatlands, situated respectively to the north, west, south, and east of town. [See
 Map 2] Between 1921 and 1936 the African population more than doubled from 3,187 to 6,588, while
 the coloured population grew by 42% from 3,052 to 4,334. For the first time in the town's history the
 9,437 whites were not a majority of the population.13 The new black population, however, was not
 composed of migrants. There was no marked imbalance in gender ratios; in fact throughout the period
 there were slightly more women than men in the town.14 The indicators strongly suggest that the black
 population was not only becoming larger, but also more stabilized and permanently "urban."13
 Kabah had electric street lamps, central water hydrants, and daily rubbish removal from central
 points. But conditions were generally deteriorating through the 1930s, and were even worse in the three
 other Locations which shared none of Kabah's amenities.16 Depression and drought brought economic
 contraction, and combined with a steady inflow of people from the countryside contributed to a large
 increase in unemployment.17 The Locations expanded to accommodate the increased population, as
 residents built their own houses, with the more well-off securing Council loans for purchasing building
 supplies.18
 Against a background of growing population, unemployment, and uncertain finances, the
 Municipality sought means to cope with the continued movement to Uitenhage. It asked the government
 to proclaim influx controls under the Urban Areas Act, though it never applied its more draconian
 provisions.19 In 1936 the Council moved to reorganize the locations by demolishing the three smaller
 Locations and moving the residents to Kabah. Rather than incurring the expense of building a separate
 coloured residential area, the Municipality intended to separate the groups within the Location.20
 Furthermore, the Council refused to subsidize new housing; instead, the new residents of Kabah would
 be encouraged to build their own homes with materials purchased from the Municipality on instalment.
 Both decisions would prove fateful for the future development of Kabah and segregation in Uitenhage.
 The Location would be out-of-step with evolving national policy regarding the separation of coloureds
 and Africans, and self-built housing would soon reproduce the slum conditions of the old locations.21
 The Councillors saw themselves providing a valuable service for African and coloured residents
 of the three locations. "In fact," wrote the Town Clerk, "the hutholder from Oatlands Location is moving
 from what is an unhealthy and badly laid out slum to what the Council hopes shortly to make into a
 model location."22 It was not to be. High death rates for Africans, especially a staggering rate of infant
 mortality testified to the grim conditions of life in the "model location."23
 The problems in Kabah prompted the Municipality's technical staff directly concerned with
 control over Africans to develop new policies aimed at bringing Uitenhage into line with Location
 administration elsewhere. In 1943 the Location Superintendent composed a lengthy memorandum
 examining Kabah's problems, highlighting the continued presence of coloureds (a violation of the Urban
 Areas Act) whom white officials believed contributed to the illicit liquor trade, and the "highly
 unsatisfactory" housing, which could only be overcome through provision of separate sub-economic
 housing schemes for coloureds and Africans.24 Embracing the principle of segregation, the
 Superintendent did not mention removal of the Kabah Location itself.
 The Superintendent's policy considerations offer some insight into the views of the Municipality's
 Native Affairs bureaucrats. While he favoured segregation, he reminded the Council that it had resolved
 in October 1943 to make representations to the Minister for Native Affairs to amend the Urban Areas
 Act to allow Africans to acquire urban freehold property. As property holders, he reasoned, Africans
 would develop pride of ownership, and would be encouraged to erect suitable houses for themselves,
 freeing the Council from housing expenditures. The development of such a class of people would require
higher wages and right of freehold (within segregated areas), and recognition "that the detribalised urban
 Native is part and parcel of the urban areas...a member of a community separate from, but no less
 important than, the European urban population."2*
 The same themes were echoed two years later in a joint report by the Medical Officer of Health,
 Acting Town Engineer, and the Location Superintendent. This time, however, their solution entailed the
 removal of the entire African population: Kabah lacked room for expansion, and it was too close to the
 white areas, diminishing housing values and blocking further expansion of the [white] town. The report
 introduced a whole new policy proposal: the Kabah Location itself would have to be demolished and
 moved to Sandfontein, a municipally-owned farm east of town.25
 The planning schemes of the Municipal technical staff remained on the drawing board for the
 duration of the Second World War. But their plans were extremely important in building an intellectual
 and policy framework for segregation in the post-war period, embracing the related ideas of extensive
 provision of freehold title, sub-economic [subsidized] public housing, separation of coloureds and
 Africans, and possible removal and resiling of the Location itself.
 Towards the end of the war, the Uitenhage Council took two steps to secure the town's position
 in the post-war period, hiring a consultant to devise a general town plan and making an aggressive effort
 to attract industrial investment. The Council hired T.B. Floyd as its planning consultant, based on his
 reputation as one of the foremost participants in the town planning movement of the 1940s.37 In
 Uitenhage, in addition to laying down recommendations for industrial, commercial, and residential zoning
 regulations, Floyd's plan dealt at considerable length with the long-running black housing problem.
 Floyd's planning vision was more comprehensive than that of the local officials, and surprisingly
 conformed very closely to the principles of the future Group Areas Act. "The object of town planning,"
 Floyd told a Uitenhage audience, "was, where possible, to zone a town in such a way that the European,
 coloured and Native sections of the population were separated by clearly defined natural
 barriers...Another principle was the elimination of situations which required the Non-European element
 to traverse the areas set aside for Europeans."28
 The Kabah location failed the test of both "principles": coloureds and Africans were mixed
 together, the location was separated from white areas by a single street, and it was located on the
 northern side of town, requiring workers to walk across Uitenhage to reach the heavy industries situated
 to the south. Floyd's recommendations incorporated many of the themes developed in the wartime
 reports by the town Council bureaucrats, most importantly, moving the location east to Sandfontein.
 Africans would be provided with sub-economic housing away from white residential areas and close to
 the factories where Africans provided unskilled labour. The coloured section of the location would
 remain in place, and sub-economic housing would be provided for them towards the west, away from the
 white and present African areas.29
 But his argument for removing Kabah was immediately attacked by various Councillors, and
 ultimately rejected. Many Councillors recalled the earlier removal of residents from Oatlands, Gubbs
 and Doornhoek Locations, and the Council's unfulfilled promises to provide a "model location." By an
 overwhelming vote the Council rejected Floyd's proposals for removing Africans from Kabah, and agreed
 to remodel Kabah on its present or an adjacent site, with emphasis given to developing a comprehensive
 sub-economic housing scheme.30
 With the advent of World War II housing construction came to a stop throughout the country,
 even as thousands of people flooded into the towns for jobs in industries stimulated by the war. Housing
 problems were'a major cause of the urban protest movements of the 1940s, and to cope with the housing
 crisis the Smuts Government created the National Housing and Planning Commission.31 Previously
 Municipalities were unwilling to build housing, as most Africans could not afford rent levels necessary for
 recouping the debt incurred in construction. Loans from the NHPC, however, came at a reduced rate of
 interest, and included a formula whereby the central state would assume a large percentage of the losses
 on a sub-economic scheme.32
 Following NHPC policy the Uitenhage Council solicited funds for creation of sub-economic
housing schemes for Africans, coloureds, and whites, as envisaged in the amended town plan.33 A plan
 for an African sub-economic scheme was approved by the Council in 1947, and early the next year the
 central Government approved a loan of ?99,390 for construction of 320 pairs of semi-detached units in an
 extension to Kabah. The area would be named McNaughton Village after the Location Superintendent
 serving during the consolidation of the four locations in the 1930s, and officials looked forward to the
 eventual expansion of the scheme to include 2,000 semi-detached units (4,000 dwellings), half built by the
 Municipality and half by the NHPC. In addition, the Council hoped to have a large portion of the town
 commonage surveyed and set out into residential plots which Africans would be permitted to purchase
 outright for construction of their own homes.3*
 Finally, funds would be sought for a sub-economic scheme of 100 units for coloureds in a newly
 cleared area west of Kabah, to be called Gerald Smith Township, and for a 60 unit sub-economic scheme
 for whites close to a new industrial area at the south end of town. The Council soon applied for a
 further sub-economic loan to expand McNaughton Village by 320 more semi-detached pairs. Thus,
 under the terms of the new NHPC sub-economic schemes, the Council agreed for the first time in its
 history to provide formal housing for its African population, and to provide racially segregated
 sub-economic and economic housing for coloureds and whites. It also developed a housing scheme for
 coloureds at Jubilee Park where plots would be available for purchase. Any removals would be voluntary
 with the promise of a better dwelling than that left behind in Kabah or the town.
 Between the end of the war and the accession to power of the Nationalist government in 1948,
 the Uitenhage Town Council had approved a vision of development which rejected forced removals and
 aimed instead at stabilization of the African population in Kabah and a slow, voluntary segregation
 between Africans and coloureds. With the provision of central government funds through the NHPC the
 Municipality hoped to achieve what the earlier location reorganization had failed to accomplish: the
 establishment of serviced, hygienic, and above all low-cost housing for Location residents in compliance
 with the segregation provisions of the Urban Areas Act. Aside from receiving approval for their project
 and technical advice from the NHPC and the provincial townships board, the Council planned the future
 of the Location virtually without participation by representatives of the central state.
 Rather than developing a long-term plan, the Council decided to look into the immediate
 problems of Kabah, hoping that "by the time we have dealt with the more urgent problems of Kabah, the
 land ownership question should be much clarified."33 Indeed, after the election of 1948 the land
 ownership question would be solved, but in a manner not much to the Council's liking. But even as the
 Council dedicated itself to a plan to carry the town into the 1950s, its own industrial program was
 generating conditions which would foil all its efforts at control.
 Between 1945 and 1946 the Council's industrial drive yielded dramatic results, transforming the
 political economy of the town forever.36 The former key industries, woolwashing and the railway
 workshops, were now overshadowed by two representatives of the newly expanding automobile industry.
 The town's fortunes, linked to the Cape Midlands hinterland and to external markets since the
 widespread development of wool production in the 19th century, were now tied firmly to the premier
 twentieth century international growth industry.
 The immediate effect of the boom was to provide employment. For the Port
 Elizabeth-Uitenhage metropolitan area total black industrial employment increased from 36% of the
 industrial workforce in 1935/36 to 56% in 1949/50. In absolute terms, the black industrial workforce
 increased from 4,288 in 1935/36 to 17,215 in 1949/50, an increase of over 300%.37 Though no figures
 are available for Uitenhage alone, an estimate of the town's black industrial workforce in 1949/50 would
 be approximately 2,500. The industrial workforce thus grew at a much faster rate than the African
 population as a whole, which increased from 6,588 in 1936 to 16,942 in 1951, or by more than 150%, with
 the fastest growth coming in the five years after the war when the new factories came on line.38 While
 the population of Kabah swelled, the number of Africans outside the Location doubled from 1,177 in
 1947 to 2,127 in 1951, as families crowded into lodgings in the western end of town.39
 Despite the introduction of limited influx controls in 1933, the African population had become
 the single largest group in the town. But the African population growth was part of a general migration
 to the town, which included a sharp increase in the coloured and white population as well. As was the
8
 case in the 1930s, the increase in population was not a function of formal migrancy; indeed, the evidence
 suggests that whole families were moving to Uitenhage, or that families would soon join a husband who
 had found work in town.40
 The pressing task for the Municipality was the straightforward issue of building enough houses
 and overcoming the squalid conditions in the Location. But as early as 1947, as the contours of
 Fagan-inspired policy were becoming clear, the Departmental Executive Committee cast serious doubts
 on the likelihood of the policy ever succeeding. The number of applications from employers to the
 Location Superintendent asking for plots to be allocated to Native employees grows daily more
 formidable."41 Unemployment had virtually been eliminated, and employers complained about the
 limited availability of labour in Kabah and pressed the Municipality to allow them to bring in more
 Africans from outside. Indeed, a Council committee recommended the repeal of the limited influx
 control measures enacted in the 1930s.
 Industrialization brought about what the Municipality had long sought to avoid: a sharp and
 uncontrolled increase in the black population. By 1948 the Council had to act quickly to implement the
 now modest looking project agreed upon two years before, when it optimistically hoped to demolish
 sub-standard dwellings and relocate African and coloured occupants to separate Municipal schemes. In
 the months before the National Party electoral victory, the Council lacked a viable long-term plan for
 accommodating its burgeoning population, resources to build houses, and the power to regulate the
 town's population growth. A Departmental Committee of Enquiry into the administration of Kabah
 argued that such processes were natural, "and without turning a location into a concentration camp, it
 seems impossible to prevent it."42 It was the intention of the National Party - in spirit at least - to do
 just that.
 ID. 1948 AND BEYOND: THE CENTRAL STATE, LOCAL COUNCILS, AND POPULAR
 RESISTANCE TO APARTHEID
 The United Party Councillors in Uitenhage and Port Elizabeth entered the 1948 elections
 hopeful that the national housing question would soon be solved along lines amenable to their local
 policies: that the Fagan recommendations would be given statutory force upon the re-election of the UP
 Government. But the 1948 election changed the political calculus almost completely, as the National
 Party's narrow electoral victory cleared the way for enactment of apartheid legislation. Locally, HNP
 members were elected to Parliament and to the Councils, which remained, however, in UP hands. The
 shift was a small, but nonetheless significant watershed. For the first time since Union in 1910, the
 Uitenhage district did not elect a UP/South African Party member to Parliament.
 A. THE GROUP AREAS ACT
 The NP campaigned on an electoral platform advocating a fundamental departure in urban
 policy, especially regarding the reproduction of African labour power. According to Hindson, the short
 term policies expressed in the UP's Fagan Report and the NP's Sauer Report had much in common, but
 their "long term trajectories were radically different." Where Fagan accepted the inevitability of African
 urbanization, seeking to channel influx through "restrained" transitional segregationist measures, Sauer
 "wished to reverse African urbanisation and to build up reserve economies so that they could support the
 entire African population."43 In response to the urban crisis of the late 1940s, the Sauer Report
 advocated a compulsory national system of labour regulation and labour control.44
 The new legal framework which the NP rushed to create in its first years in office gave statutory
 force to many of the recommendations of the Sauer report.45 These laws, jointly linked, aimed at
 regulating and limiting the movement of Africans to the cities, encouraging long-term shrinkage of
 African presence in the urban areas, and strengthening social and political segregation between all race
 groups. Amendments to existing legislation tightened influx control by linking it to controls over
 employment. African labour contracts would be registered, employers would be forced to pay a Native
 Services Levy into the local Native Revenue Account, and a reorganized labour bureau would be
 established as the institutional means for limiting movement from rural areas and distributing labour
 between the various sectors of the economy.46
Urban policies after 1948 were not merely a function of the political struggles over migratory
 labour, however. Municipalities and the central state developed initiatives to reorder, redistribute, and
 control urban populations as a whole, whether migratory or settled, African or "non-African1*. Rather
 than viewing the pass laws and the Urban Areas Act in isolation as a response to the problems of
 migrant labour, it is important to see how they intersect with other central state policies, especially with
 the Group Areas Act [GAA].4T
 In repealing a contradictory patchwork of provincial legislation, the Group Areas Act sought a
 comprehensive system of segregation which would affect all race "groups," including whites, by imposing
 control throughout South Africa over inter-racial property transactions and inter-racial changes of
 occupation.48 The law depended on the prior identification of every citizen in terms of the Population
 Registration Act, and the placement of all individuals within designated racial groups. These groups
 would be assigned to a specific, racially homogeneous physical space: the group area. It specified a
 series of stages through which control would be progressively tightened leading to the ultimate
 declaration of full group areas, racially pure in terms of both ownership and occupation. The Land
 Tenure Advisory Board [LTAB], a legislative holdover from an earlier effort to control Indians, later
 reconstituted as the Group Areas Board, was given extremely broad powers under amendments to the
 Act to inquire into proposed group areas and to recommend proposals for proclamation by the
 Minister.49
 The impact of the Act was profound: it not only gave the state dramatic new powers to remake
 the full physical field of urban life, but it universalised the systematic mechanisms of separation which
 were formerly wielded only against Africans.50 The "pure" segregated residential areas were to be
 separated by great distances and wide "buffer strips." Finally, the Act posed a major threat to the
 stability of the Indian and coloured petite-bourgeoisie, by limiting their access to trading stands or factory
 sites in central cities.31
 For local Councillors convinced of the efficacy of gradual and voluntary segregation and an
 integrated commercial sphere, the implications of the Act were profound. But the full power of the Act
 cannot be grasped when it is viewed in isolation: its awesome powers operated (admittedly unevenly and
 sometimes contradictorily) in the evolving thicket of apartheid legislation. In combination the GAA and
 the Urban Areas Act created enormous new state powers. If the latter Act carried the authority for
 segregating and controlling the movement of Africans, the former defined how the segregated areas were
 to be cited vis-a-vis all other racially defined sections.
 B. THE POLITICS OF NON-COMPLIANCE IN VITENHAGE
 Kabah - where it was planned at all - developed according to a vision quite at odds with the
 newly-evolving apartheid framework. In 1951 the 4,200 coloured residents of Kabah made up nearly a
 quarter of the Location population, and accounted for more than 60% of the coloured population of
 Uitenhage.52 The remainder of the coloured residents lived in the town itself, interspersed with poor
 whites and Africans in the older western sections. On its eastern side the Location was separated from
 the white sections of town by a single street, while another road on its south-western side divided Kabah
 from the proposed coloured sub-economic scheme which the Council had been shepherding through the
 national bureaucracy since 1947.
 Not only were coloureds and Africans living in close proximity in Kabah, not only were
 coloureds and Africans living in the town itself, but the Location was too close to the white sections of
 town and too close to the proposed coloured sub-economic scheme. Coloureds would have to be
 removed from Kabah to satisfy the requirements of the Urban Areas Act, and they would have to be
 moved a suitable distance away from Kabah and from the soon-to-be-declared "white" town to satisfy the
 buffer zone requirements of the GAA; furthermore, the Kabah Location itself would have to be moved.
 The wide gap between the legal ideal of racially pure areas and the incompatible conditions on the
 ground defined the terrain of the local political struggle over national urban policies.
 How would the systematic policies be brought to bear on a seemingly intractable situation on the
 ground? The policy changes after 1948 were not meant to be instantly implemented. Hendrik Verwoerd,
 Minister of Native Affairs, accepted that no one would consider "stopping the development of South
10
 Africa-not even for a period."33 As Posel makes clear, in distinguishing between "purist" and "practical"
 interpretations of policy, there were limits on the purist approach, the extent of which would be
 established through real political struggle.34 Since local governments were the primary apparatus of the
 state for intervening in the labour market, such bodies became a major site of conflict.
 As a first step toward achieving its national goals, the Government enacted subtle, but extremely
 important changes in its legal relationship with local authorities. Under 1945 amendments to the Urban
 Areas Act the Minister was empowered to proclaim controls on African entry into an urban area "if
 requested to do so by a resolution...of any urban local authority...."33 Some municipalities, such as Port
 Elizabeth, had opted out of the system, and remained "open cities" where Africans were not subject to
 the pass laws. Uitenhage's limited influx control measure from 1933 was never seriously enforced, and
 the Council looked forward to its repeal.
 The Native Laws Amendment Act of 1952 reversed the scope of application of influx control:
 rather than being applied only where local authorities had requested it, influx control was automatically
 instituted throughout the Union. Under the new law the local authority was limited to asking for an
 exemption from influx control, and even then, the Government could refuse. The same Act provided for
 a compulsory national labour bureau system for all urban and rural areas, repealing a 1949 provision for
 voluntary establishment of labour bureaux.56 Local authorities were stripped of an extremely important
 statutory power, while influx control was legally, though not yet in practice, extended from a piece-meal
 program to a national system.
 As a second step, the central government shuffled the bureaucracies responsible for specific
 urban policies. The impact of the shifts quickly became apparent to local officials. In late 1951 S.B.
 Featherstone, the Uitenhage Town Clerk, received notification from the Secretary for Native Affairs in
 Pretoria that funding for the Municipality's proposed coloured sub-economic housing scheme would be
 delayed due to its "close proximity to the proclaimed Native location."57 The coloured housing scheme,
 laid out under the housing policies of the previous Government, was cited outside the African area, in
 compliance with the Urban Areas Act (1945), but was just across the road from the Location, in violation
 of the new GAA.
 Featherstone was incredulous. The Department and the Municipality both sought separation of
 Africans and coloureds in Kabah, "And yet you tell us summarily that the proposed site for a Coloured
 Housing Scheme is no good because it is too near the Native Location. Under existing conditions
 coloureds and Natives live together in the Location. Surely it is better to develop a separate area for
 coloureds alongside the Location even if it is near the Location/58
 The Town Clerk's anger was heightened because the Council had expected rapid approval of its
 proposal. It had already purchased materials and hired labour, and delays increased expenditure for
 storing supplies and retaining workers. More importantly, the Municipality had received a formal
 circular a mere four days before receipt of the Secretary's letter announcing that the Native Affairs
 Department now had authority to review Housing Commission plans and possessed veto power over the
 choice of sites for housing schemes.59 Previously such power was vested only in the National Housing
 and Planning Commission. The day before the Secretary's letter was written the Town Clerk had
 received an undertaking from the Technical Adviser to the National Housing Commission that he would
 personally discuss the matter of the coloured housing scheme with the Native Affairs Department.60
 The adviser was Adolph Schauder, architect of Port Elizabeth's "liberal" "Location Strategy" and a
 prominent figure in the UP; yet he was not informed of the content of the Native Affairs Secretary's
 letter. The Municipality confronted a new and bewildering rearrangement of bureaucratic forces, linked
 into power networks to which the UP Councillors enjoyed little or no access. In January 1952 the
 Council reached agreement with the Commissioner for Coloured Affairs and with the Native Affairs
 Department on the provision of a buffer strip between the proposed coloured scheme and the Location.
 Over 2,000 people would have to be removed and 300 houses demolished to make way for the buffer
 strip, which the Town Clerk derided as "the task which the Government Departments concerned have
 imposed upon the Council in return for the right to build a Coloured Housing Scheme...."61
 If the fight over buffer strips was the preliminary bout, application of the GAA was the main
 event. The Council's role in the application of the Act was essential and the Land Tenure Advisory
11
 Board began wooing it in 1951. The LTAB lacked detailed information about the town, and the
 Council's participation would legitimate the group areas project in the eyes of many local constituencies
 who were suspicious of the policy.62
 As a first step the Council accepted the task of conducting a survey of the town to establish the
 precise racial character of ownership and occupation. This is...the only thing which we as the local
 authority should do at this stage," Featherstone, the Town Clerk reported to the Finance and General
 Purposes Committee. He hoped that conducting the survey would satisfy the central state that the
 Municipality was cooperating while buying time for the Council. "I do not think we should allow
 ourselves to be rushed into action," he wrote, "My attitude is not an apathetic one but I am not keen that
 we should be the first in the Cape to invoke the machinery of the Act and make what might prove costly
 blunders due to ignorance of all the implications."63
 On 17 October 1951 the entire Cape Province was declared a "specified area" in terms of the
 Group Areas Act, essentially freezing existing conditions.64 The proclamation was a central state
 initiative further depriving local authorities of discretion in responding to the GAA. For the next year the
 Municipality refused to take the initiative in a series of minor dealings with the LTAB. Then in August
 1953, the Secretary of the LTAB in Cape Town asked the Council whether it had reconsidered its
 position on the next stage of implementation, declaration of "defined" areas.65 In defined areas the level
 of control over "interpenetration" would be increased: unoccupied property could not be built upon nor
 existing buildings extended without ministerial approval.
 The step was a particularly dangerous one for the Council to take. "Specified" areas, the freeze
 on inter-racial changes of ownership and occupation, had been established from above by the central
 government in a province-wide proclamation in which the Council had no part. But because Municipal
 officials possessed the most accurate understanding of conditions on the ground, the Government
 expected them to play a major role in "defining" individual properties owned or occupied by individuals of
 one race in an area to be reserved for another. By participating in efforts to define areas, the Council
 itself would directly harm vested interests in the town, especially black businesses in white areas. At the
 end of August the Council's Finance and General Purposes Committee decided to accept the LTAB's
 request and directed the Town Clerk to report to the Committee any cases of "possible penetration."
 But a curious lobbying effort began at the next full Council meeting where the members were to
 vote on the Finance and General Purposes Committee recommendation. During the meeting, the
 Council received an urgent letter from the Uitenhage Indian Congress, asking the Council to refrain
 from making any decision on the question to allow them time to study the Committee's proposals.
 Application of the Act, the letter warned, would "have grievous detrimental and widely devastating effects
 on...properties and the very existence and destiny of the future of the Indian community."66 Three of
 the four authors were shopowners, and the Vice Chairman, D.R. Lalla, owned four properties. Their
 request was accepted, and the matter was referred back to Committee.67
 At its next meeting the Committee received the deputation from the Uitenhage Indian Congress,
 including the four shopowners. The group stressed "the cordial relations which have always existed
 between the Indian community and other sections of the community in Uitenhage," and said that the
 Indians living or trading in town had been present there for more than fifty years, a presence which they
 did not regard as "penetration." The group asked that since the Act placed sole authority for racial
 zoning in the hands of the LTAB, which would eventually make its own representations to the Minister,
 "the Council should wait until this happens and not take the initiative...."68 After considering the
 deputation's arguments, the Committee decided that the Council should defer the matter indefinitely, "to
 remove any suspicion on the part of the non-European sections of the community that the Council is
 acting against their interests...." The next week the Committee's new recommendation was carried by the
 full Council.69
 The Council learned the Government's response within two weeks of the decision. The Finance
 and General Purposes Committee received notification from the National Housing and Planning
 Commission that the NHPC and LTAB had decided not to consider any applications for housing funds
 whatsoever until the LTAB had granted approval for the planning of group areas. Whether or not the
 LTAB meant what it said, the local officials who had just refused to cooperate with group areas
12
 legislation were on notice that in the future they would have to seek approval for all housing plans from
 the very authorities they had just snubbed, and Uitenhage had numerous housing requests pending before
 the NHPC.
 The LTAB approached the Council again in June, 1954, but in two years of debate Councillors
 could not decide whether "to take the initiative to see that people were treated in the most humane way
 possible" or to refuse to participate in applying a law likely to cause the "economic strangulation of the
 Chinese and Indian traders...."70 In support of the latter position, Councillor Gerald Smith "objected to
 the very principle of the Act and...did not think the Council should take the lead...[A]s responsibility for
 the Act and its administration rested solely with the Government it should be left to them to apply
 it...."71 When neither the full Council nor a special subcommittee could break the deadlock members
 essentially threw up their hands. In March 1956 the Council decided to "leave the problem to be handled
 by the LTAB."72 The position became the official policy of the Council for the next six years.
 At the meeting in October 1953, when the Council originally rejected cooperation following the
 protest by the Uitenhage Indian Congress, the leading NP member, Councillor Frans Conradie, reminded
 the members that Indians were but 1% of the town's population. Speaking in Afrikaans, he "could not
 understand why the Committee had so quickly turned around as a result of formal objection against the
 move which had been put forward by the Indian community." Indeed, why did the Council reverse itself
 after meeting a single Indian organization?
 The Act would surely cause hardships, and the Councillors knew that they would fall most
 heavily on a small but important group: Indian, Chinese, and coloured businessmen and property owners.
 The question of removing coloureds from Kabah was no longer a policy issue: from the 1940s the body
 had accepted that goal. Councillors rationalized residential segregation on the grounds that hardships
 inflicted on blacks by removal were minimized because families received better houses in the new
 schemes than those they left behind in Kabah.
 But the Group Areas Act's limits on ownership and occupancy threatened to deal a mortal blow
 to the economic well-being of those owning businesses in town. A 1959 Municipal survey identified 34
 Chinese and 29 Indians as property owners in the town, holding at least 100 properties with a taxable
 value of more than ?200,000. A 1955 survey identified more than 340 black-owned dwellings and nearly
 80 shops.73 Furthermore, as a result of the historical patterns of development on the western end of
 town whites owned at least 107 dwellings and 18 shops in likely coloured, Indian, Chinese, or Malay
 Group Areas, and would suffer hardships under the Act. Thus the holdings of a number of individuals -
 black and white - were rendered insecure. To expropriate all or part of these properties the Council
 would lose part of its tax base, be forced to pay compensation, and face the prospect of lengthy and
 costly litigation.
 Indian, Chinese, and coloured businessmen were not only property owners, but voters, as well.
 Though in 1956 the NP succeeded in its protracted constitutional battle to remove coloureds from the
 parliamentary voters' roll, they did not fully disenfranchise them: coloureds and Indians in the Cape
 remained eligible to vote in local elections into the 1970s.74 A 1959 survey of "non-white" voters listed
 950 "coloureds," amounting to 19% of the registered electorate.73 In a town where the Afrikaner
 electorate was increasing, and voted almost exclusively for the NP, a UP Councillor needed all the non-
 Afrikaner votes he could muster, and would be ill-advised to be seen threatening Indian, coloured, or
 Chinese interests.76 If the Uitenhage Indian Congress and other bodies had special access to the
 Council, it was indeed, as Councillor Conradie emphasized, not because of their numbers, but because of
 their economic influence and political weight.
 C. APARTHEID IN PE: THE CONSEQUENCES OF COMPLIANCE
 1. The Urban Areas Act: Defiance, Riots, and the End of the Liberal "Experiment"
 Port Elizabeth, like Uitenhage, emerged from World War II with a severe housing crisis. The
 town's population had grown almost as rapidly as Uitenhage's and for the same reasons: economic
 expansion during and immediately after the war, and from drought in the hinterland. The population
 expansion, and the near absence of Municipal construction had contributed to a housing problem far
13
 worse than that which prompted the program of removals and segregated housing schemes of the late
 1930s. Indeed, the efforts to remove Africans from Korsten in 1937 had temporarily reduced the number
 of people living there, but only slightly, and Africans continued to constitute the overwhelming majority
 of the population. By 1951 there were actually more Africans resident in Korsten than before the
 removals. The City Council reaffirmed its pre-war Location strategy, and refused to comply with the new
 national influx control legislation. Yet, as Robinson shows, these policies were increasingly out of step
 with post-war reality.77
 Most importantly, the City Council faced a new threat from below: the increasingly organized
 African working class in Port Elizabeth, led by a resurgent African National Congress and Communist
 Party. Economic downturn in the late 1940s exacerbated the problems of social reproduction, especially
 as the motor industry was hit by restrictions on CKD material imposed to redress the national balance of
 payments crisis. The deteriorating social conditions, and especially the apartheid proposals of the
 National Party Government generated a resurgence of black political protest, which was as strong in Port
 Elizabeth as in any other area of the country.
 A renewed ANC with local leadership drawn from the ranks of working class residents protested
 proposed rent increases in 1945 and 1947, and in 1949 staged a militant four-month boycott of a steep
 rise in bus fares by South African Railways and Harbours.78 The Eastern Cape, and especially Port
 Elizabeth and Uitenhage, were also the centres registering the strongest response to the ANC's 1952
 Defiance Campaign.79 The growing political mobilization around the campaign, and the intensification
 of both the central and local state's police response led in October to a violent riot. At the New
 Brighton train station a Railway policeman shot a man whom he suspected of stealing a tin of paint. The
 angry crowd which gathered at the scene killed a passing truck driver, and moved on to destroy
 white-owned properties in the township, killing 3 more whites in the process, while the police killed seven
 Africans.80
 The National Party disingenuously laid responsibility for the riots at the door of the Port
 Elizabeth City Council and its "liberal" policies. The importance the government attached to its attack on
 the Council is shown by the prominence of its chief spokesman: none other than Hendrik Verwoerd,
 apartheid's architect as Minister of Native Affairs, who would later serve as Prime Minister during the
 most brutal period of NP rule. The riots provided Verwoerd with a convenient opportunity to discredit
 resistance to his Native Affairs policies in the run-up to the 1953 elections, in which the NP was in no
 sense assured of success.81 Four days after the riot he travelled to Port Elizabeth for talks with NAD
 and SAP officials, pointedly snubbing the Mayor and City Council by sending a departmental official to
 convey his views. That night, in a public speech at an NP rally, Verwoerd pronounced the liberal
 "experiment" a failure, and threatened to impose influx control himself if the Council failed to act.82
 In the aftermath of the riot Port Elizabeth came to resemble a military zone, with Permanent
 Force armoured convoys carrying out "routine training exercises" on the city streets.83 Armed police
 were placed on every bus travelling to New Brighton, more armed policemen were on duty at the bus
 terminus, and the vehicles stayed out of the township, letting off passengers half a mile away. The ANC
 called for a transport boycott until the police were removed, and the following day the boycott was
 virtually total. Four days later the SAR&H System Manager met ANC leaders, the police were removed,
 and the action called off. But that very same day the City Council made decisions which would give the
 ANC far greater cause for concern.84
 On 28 October, ten days after the riot, the City Council voted by an overwhelming majority to
 request the Minister of Justice to enforce the Riotous Assemblies Act banning all open-air meetings by
 Africans in the city and to request the Minister of Native Affairs to impose a curfew.45 Ten days later
 the powers were granted by the Minister of Justice, who also banned 52 African leaders in the Eastern
 Cape under the Suppression of Communism Act, and invoked War Emergency Regulations giving the
 police power of summary arrest. The emergency restrictions applied not only in Port Elizabeth, but in
 Uitenhage, Peddie, King William's Town, and East London, all of which were important centres during
 the Defiance Campaign.
 As Robinson points out, in the face of pressure from the central government and its local
 supporters - including a growing proportion of PE voters - liberal Councillors could retain belief in the
14
 benevolence of their policies by laying responsibility for the rising tide of protest on agitators rather than
 on the material conditions Council policies had themselves spawned. At the October 28th Council
 meeting Urban Areas Commissioner Brand's presentation of the pass laws strongly emphasised their
 utility in weeding out undesirables, an act of salesmanship keenly attuned to the Councillors' own
 self-justifying perspective.86
 By its actions the City Council may have hoped to preempt the central state from imposing an
 even more draconian response, but the moves only heightened the crisis. In response to the ban and
 curfew, the ANC Eastern Cape Regional Committee called for a stay-at-home on November 10, "to
 continue until God Almighty has changed the hearts of the City Councillors."87 In the days before the
 stay-at-home police reinforcements rushed to Port Elizabeth, while the Council, local employers and
 senior representatives of the SAP and Ministries of Native Affairs and Labour hammered out common
 strategies. The City Council and Government departments threatened to dismiss any employees taking
 part in the action, though local industries adopted a more lenient approach of "no work, no pay," perhaps
 as a result of experience gained from months of strikes sparked by dismissals of workers arrested in the
 Defiance Campaign. Before the stay-at-home, the local Executive Committee of the ANC negotiated
 with the Mayor, and agreed to limit the strike to one day in return for the Mayor's promise to reduce the
 curfew to three months and the ban on meetings to one month.
 The protest, like the bus boycotts and the Defiance Campaign before it, was a total success in
 both Port Elizabeth and Uitenhage as nearly all African workers stayed away from work. Retribution
 was swift, however, as approximately 5,000 Africans were fired, with the harshest responses coming from
 Government departments.88
 The central government blamed the stayaway on the Council's "liberal" policies, and in its
 aftermath made renewed calls to apply influx control. As noted above, the amended influx control
 legislation removed local discretion in its application, and through a Parliamentary proclamation made
 the laws automatically applicable in all towns with more than 20,000 African residents, including Port
 Elizabeth. Though the Council had requested an exemption from the legislation, Verwoerd rejected their
 appeal. Instead he accused the Council of "dereliction of duty" in failing to carry out obligations under
 the law, such as inaugurating a labour bureau and registering African service contracts.89
 Central ministries in Pretoria were closely monitoring events in Port Elizabeth, and applying
 pressure to force compliance with their aims. When Eiselen, Secretary for Native Affairs heard press
 reports about the Mayor's promise to the ANC to rescind the curfew and ban, he rebuked the Town
 Clerk for the Council's ignorance of the emergency laws they had themselves invoked, which could only
 be altered by the national Government. After informing the Town Clerk that Verwoerd would not
 recommend withdrawing the restrictions, "even when the position has returned to normal," [emphasis
 added] Eiselen warned the Council "not [to] compromise itself or the Honourable the Minister by
 making any promises which may not accord with the desires of the Minister."90 Once mounted on the
 tiger of political repression, the Council found it very difficult to avoid ending up inside.
 On 15 January 1953 the Council agreed to implement the influx control provisions of Section 10
 of the Urban Areas Act and authorized establishment of a local labour bureau.91 Clinging to its legacy
 of liberalism while caught between a militant black resistance movement and an increasingly
 interventionist central state, the Council justified its capitulation on grounds that it would administer the
 restrictions more fairly. It asserted that the laws would allow greater control over the "tsotsi [gangster]
 element," and increased protection for "bona-fide" Africans against those from outside flooding the job
 market. The draconian provisions of the Urban Areas Act would therefore be administered in the best
 interests of Port Elizabeth's Africans!
 2. Group Areas
 In contrast to the intense political fight over the Urban Areas Act, the Council had far fewer
 qualms about compliance with the GAA. Despite the Council's preoccupation with the local emergency
 and influx control, city agencies had been busy for over a year conducting the initial research for planning
 group areas. Rather than rebelling against the GAA, as had the Uitenhage Town Council, a majority of
 the Port Elizabeth Council saw the legislation as a logical extension of a long-standing Municipal
15
 segregation policy.
 In mid-1952 a Joint Town Planning Committee was established, consisting of representatives of
 the Port Elizabeth and its adjacent Walmer Municipality, the Port Elizabeth Divisional Council
 (governing the unincorporated area of the Port Elizabeth magisterial district), and representatives of
 central state ministries. The Joint Committee developed the first proposals for Group Areas with a view
 to coordinated planning for the entire metropolitan area.92
 In December 1952 the Committee forwarded its proposals to the Land Tenure Advisory Board.
 As a result of years of local segregation - especially in the proliferation of racially restrictive covenants in
 new housing developments - the Committee felt that many areas in the city could be declared group
 areas immediately. There were a number of "problem" areas, however. Though coloureds had been
 segregated in Municipal housing schemes, these were scattered like islands throughout mostly white
 residential areas, so there was no single zone of coloured residence in the city.
 To accommodate these areas, the Council deviated from the principles embodied in the Act and
 pursued by the LTAB, which sought to create wide zones of single-race residence with sufficient land for
 expansion, and separated by broad buffer strips.93 By contrast, the Council's initial proposals minimized
 the disruption and tremendous expense which would ensue from moving people from the many scattered
 areas and rehousing them in a single zone.
 First, its definition of "coloured" included Malay, Chinese, and Indian people, though the
 Population Registration Act defined these as separate groups, and the GAA stipulated that each "group"
 be accommodated in a separate area. Furthermore, it declared Group Areas wherever large
 concentrations of coloureds were already living, whereas the Act demanded that they be relocated to a
 single, large group area. The plan established "free areas" in zones where large numbers of coloured and
 Indian businessmen already owned shops. Finally, in an effort to soften the blows on present citizens,
 the proposals envisioned segregation as an extremely long-term process, and granted families 60 years to
 move if they lived in an area specified for a different group.
 Ultimately the LTAB (renamed the Group Areas Board in 1955) overrode all these deviations,
 and in 1960 submitted its recommendations to the Minister of the Interior. The scattered islands of
 coloured residence, including South End, Fairview, Lea Place, Stuart Township, and the pockets in North
 End and Korsten were declared part of the white group area, though evictions were to be deferred for
 the time being. Proposals for free areas for coloured and Indian traders were rejected, and all were to
 be moved to the border of the coloured group area. With minor exceptions the recommendations of the
 Group Areas Board served as the basis for the final group areas proclamation in Port Elizabeth on 30
 May 1961, eight years after the first proposals were submitted to the LTAB. Notwithstanding the local
 officials' rationalizations that cooperation with the government would permit a fairer application of the
 Act, their proposals were uniformly rejected in favour of a pure vision of segregation dictated from
 above. [See Map 3]
 The rather abstract process of planning did little to change the immediate conditions. Despite
 differences between the Council and central government over the disposition of areas like South End,
 both bodies agreed that Africans resident in the town must be removed to New Brighton, and that
 coloureds be removed from Korsten to the existing coloured schemes in Schauderville. Actual removal
 and - more important - funds for building alternative accommodation in the officially "preferred"
 residential area were far more intractable problems.
 3. Forced Removals
 The Urban Areas Act and Group Areas Act gave the Municipality interlocking legal authority
 and powerful new state institutions to remake Port Elizabeth on the Procrustean bed of apartheid social
 engineering. To conform with the new vision of racial planning, the local and central state would have to
 destroy the residential checkerboard which had resulted from generations of piecemeal official and
 private segregation. People living or owning property in areas defined for other "groups" were to be
 moved to the "proper" area, where alternative accommodation would have to be provided to prevent
 further squatting. Korsten and nearby Dassiekraal were the state's first targets in its effort to transform
16
 Port Elizabeth into an apartheid city.
 Unlike the removals and housing schemes of the late 1930s, these policies were meant to bring
 wholesale changes in residential and commercial ownership and occupancy, permanently altering the
 urban map. The earlier removals from Korsten had been hampered by the slow provision of housing in
 McNamee Village, and by the limited authority to evict Africans under the Slums Act. By contrast, the
 second action against Korsten was bolstered by the provisions of the amended UAA, making it an
 offense for Africans to live outside of a formal location.9* Not only could Africans be moved en masse,
 but the enhanced repressive arsenal provided by new pass laws and labour bureaux could prevent them
 from returning, and passes could be used to differentiate between "illegal" Africans and legitimate
 coloured residents. Furthermore, the Government's acceptance of "Site and Service" housing in the
 Kwazakhele extension to New Brighton enabled it to move people at a rate unencumbered by the slow
 pace of house construction. Finally, Africans owning property in Korsten would ultimately be
 expropriated, not under the UAA, but under the provisions of the GAA, which prevented ownership by
 persons outside of their prescribed group area.
 The Korsten removals began in April 1956, and by the end of 1957 more than 28,817 people had
 been moved to Kwazakhele. Little resistance was offered, despite the strength of ANC organization in
 Port Elizabeth.95 By 1960 there were slightly less than 2,200 Africans living in Korsten.
 As Taylor points out, the Korsten removals signified "the conclusive containment and control of
 the majority of the identified working class.?" In 1951 31% of Africans in Port Elizabeth lived in the
 "wrong" area, outside designated Locations. By 1960 the figure was reduced to 16%, and by 1985 to
 4%.96 The greatest increase in segregation of Africans in Port Elizabeth during the twentieth century
 occurred over the period of the second Korsten removals, so that by 1960 segregation between coloureds
 and Africans was almost as thorough as between whites and Africans.77 Despite the massive population
 growth for both coloureds and Africans over the period, the removals policy succeeded in "canalising"
 such growth into racially segregated areas defined and controlled by the state.
 The state's preoccupation with removing Africans to Kwazakhele had postponed any attempts to
 address the backlog of housing for coloureds. The natural increase of the population was enough to
 cause considerable overcrowding in Schauderville, Dower, and other coloured housing schemes. These
 housing projects were inadequate to start with, but the huge migration of coloureds to Port Elizabeth
 during the 1950s came at a time when very few houses - and no sub-economic schemes - were built by
 the Municipality for coloureds.98 The proclamation of group areas in 1961 added immensely to the
 problem, as coloureds living in South End, North End and other parts of the city would have to be
 rehoused in the new coloured group area stretching north from Schauderville.
 These removals occurred throughout the 1960s, and though the bulk of the coloured population
 had been relocated by 1970$ some smaller areas were cleared only in 1984. The pace of removals was
 slowed by a number of financial and political problems." The central government's unwillingness to
 fund sub-economic housing schemes created immense stumbling blocks to the removals mandated by the
 GAA, while the Municipality was unwilling to subsidize the losses which would inevitably result from
 placing poor coloureds in economic housing where they couldn't afford the rent.
 The removals were startling simply in terms of the magnitude of people moved: in 1985 290
 people lived in Fairview, compared to some 10,400 in 1970. Of 6,556 coloured residents of South End in
 1960, only nine remained twenty years later. By 1980 only 586 coloureds lived in North End where 1,695
 had lived in 1960, and virtually of these people were gone within three years.100
 In the space of two decades the NP's apartheid policies reversed community patterns which had
 developed over 150 years in North End and South End. In the case of South End the term "erased" is
 literally appropriate: save for two mosques, the original buildings were razed; even the rectangular street
 grid was destroyed and a new curvilinear pattern superimposed on the old. The demolitions resulted in
 the devastation of vast tracts of the city to make way for racially acceptable "redevelopment," which
 arrived decades later. In the interim, areas subjected to group areas clearances stood empty and forlorn,
 mute monuments to apartheid. These systematic acts of destruction were aimed not simply at uprooting
 groups of people, but at destroying all vestiges of entire communities. These outlying coloured areas
17
 were far from the centre of the city, often far from work, and were not only cut off from the African
 townships and white areas, but were differentiated by class and were also isolated from each other.101
 In remaking the residential map of Port Elizabeth, the removals orchestrated under the amended
 Urban Areas and Group Areas Acts also transformed social relationships within and between legally
 defined "groups." Coloureds and Africans were effectively physically separated in the Port Elizabeth
 municipal area by the late 1950s, while coloureds were gradually removed from white areas by the early
 1970s. In spite of the tremendous surge in population after 1951, when the coloured and African
 population more than doubled, by 1970 these increases were almost entirely accommodated within the
 exclusively defmed boundaries of segregated group areas and African townships.
 D. V1TENHAGE: THE CONSEQUENCES OF NONCOMPLIANCE
 1. Passes and Protest: Policies Towards Africans
 At the same time that Port Elizabeth Councillors were collaborating closely with the LTAB in
 the further segregation of the city, the Uitenhage Town Council dug in its heels at every stage of the
 process. Of course Port Elizabeth was a more prominent target than the smaller city, and came under
 intense pressure from the central state to conform. The NP government was especially aggressive after
 the 1952 riots launched Port Elizabeth's local debate into the mainstream of national political struggles
 over Verwoerd's policies towards Africans. Such concerted pressure was not applied in Uitenhage until
 the late 1950s and early 1960s, a full decade after the Port Elizabeth Council agreed to cooperate. The
 delay is a crucial factor in accounting for differences in segregation between the towns. Where Africans
 were generally removed to distant New Brighton by 1957, the old Kabah Location remained intact well
 into the 1970s. Even more important, where segregation between Africans and coloureds in Port
 Elizabeth - well advanced before 1948 as a result of local processes - was virtually completed by the late
 1950s, in Uitenhage such processes were just beginning by the late-1950s, and reached their peak only in
 the mid-1970s.
 The UP majority on the Uitenhage Town Council was extremely sympathetic to the plight of
 Indian and coloured property owners under the GAA, and sought to exert local discretion where
 possible, but their attitude towards African residents of the town was considerably different. Much of the
 discretion in policy towards Africans had been removed from local authorities and transferred to the
 central state bureaucracy by the NP's amendments to the UAA. The Uitenhage body followed it's
 neighbour's traditional approach to African affairs, believing the influx of Africans could be successfully
 controlled by means other than the strict pass control regulations of Section 10 of the Urban Areas Act.
 In January 1951 the Native Affairs Committee, chaired by the liberal Councillor E.S. Rens, had resolved
 not to apply those elements of Government's labour control where local authorities were still allowed a
 measure of discretion.101
 It was still necessary, for example, in terms of the Native Laws Amendment Act for the Minister
 to consult with the local authority regarding establishment of labour bureaux. At first the Council agreed
 to study the question. But it was not until mid-1953, after Port Elizabeth established a bureau, that the
 Council agreed in principle to the creation of one in Uitenhage. But before the Council could act
 definitively, and without consultation, Government Notice no. 2004 of 1953 declared Uitenhage a
 "prescribed area" under the Urban Areas Act, unilaterally establishing a Native Labour Bureau.103
 Nor was it mandatory for Uitenhage to apply the Native Services Levy Act, as the original law
 [Act no. 64 of 1952, section 2] applied automatically only to towns with more than 20,000 African
 inhabitants.10* The Council did not desire proclamation of the Act in the town, as local industry was
 extremely hostile to its provisions.105 At the end of 1957 the Municipality was again discussing
 application of the Act to Uitenhage, and while the Mayor expected the law would be applied "sooner or
 later," the Council applied to be exempted from the levy.106 Opposition to these measures was never
 couched in terms of the infringement of rights of Africans. Rather the primary considerations of the
 Council were the costs associated with the projects, burdens imposed on business, and doubts about the
 effectiveness of the controls themselves.
 Soon the Native Affairs Department would take a hard line on the Native Services Levy, but
18
 only as a power play within a much more important fight: the removal of Kabah. The question of the
 Location's future had been a non-issue since the Council's 1945-46 commitment to upgrading Kabah. In
 a September 1955 meeting with the LTAB to discuss group areas planning, Council representatives very
 briefly referred to the Location, but all agreed that "at this stage it would be an uneconomic and
 impractical proposition to try and move the Location."107
 But in July 1957, Brownlee, Chief Native Commissioner for the Eastern Cape, bluntly informed
 the Council that under new Native Affairs Department policy no housing loans would be granted to a
 local authority whose Location did not conform to buffer strip requirements. A 500 yard buffer strip
 would be required between Kabah and the white suburbs and between the Location and the new
 coloured sub-economic scheme.108 The Municipality had not cleared the buffer zones in 1952 when it
 was ordered to move 1,800 persons and 300 dwellings. In the interim many squatters had moved to
 those areas and the NAD was demarcating a far larger buffer zone, which would necessitate moving the
 most densely settled sections of the Location: over 11,100 persons and 1,242 dwellings.109
 But Brownlee had something else in mind. Rather than go to the expense of creating a buffer
 strip, which would involve extensive compensation to those moved, while leaving Kabah intact, the
 Council, he suggested, should consider moving the entire Location. Brownlee recommended that the
 new Location be sited to the south, in the direction of Port Elizabeth, which the NAD was promoting as
 a point of concentration of black residential areas in Port Elizabeth and Uitenhage. The Chief Native
 Commissioner, perhaps unconsciously, fully resuscitated T.B. Floyd's 1945 town planning
 recommendations. To induce conformity, Brownlee offered to delay for one year the extension of the
 Native Services Levy Act to Uitenhage. Even so, in late October 1957 the Council rejected the
 unsolicited advice from NAD, on the grounds that the resiling of Kabah Location would be "quite
 impracticable."110
 It did not take long for the other shoe to drop. In reply to the Council's action and without
 warning NAD announced in late October that the Native Services Levy Act would apply to Uitenhage
 retroactively as from 1 October. The Town Clerk applied to the Department to extend the period, as
 local officials had no means to begin collecting the levy, but Brownlee refused, and demanded that arrear
 service contract fees be collected from business or the Municipality would be held liable for the
 uncollected monies on its Native Revenue Account.111
 The Council rapidly set to work investigating the potential costs for upgrading or resiting Kabah.
 In August 1958 the Council resolved in principle to acquire a farm southeast of the town for a new
 African location. The Town Clerk stated the Council's motives most succinctly: "the Council basically has
 no particular concern as to whether the Location remains on its present site, or is resited...the issue now
 has come down to one purely of finance...."112
 After a two-year dispute about which farm to purchase, the renamed Bantu Affairs Department
 [BAD] stepped in and persuaded the Council to purchase the farms "Boschoogte" and "Naroes," to the
 south-west of town across the Swartkops River. The Councillors were worried about the cost of
 developing the farms, and some reconsidered upgrading Kabah. In a meeting with the Council the
 Under-Secretary of BAD, Louis Smuts reiterated the Department's demand for the Council to relocate
 Africans and reserve Kabah for coloured housing, but dangled the carrot of financial support for the
 move. "If the Council and the Department worked together," he said, "the problem which appeared to be
 a major one would not be found to be so difficult."113 Still the Council hesitated. Some members
 favoured buying the two farms suggested by BAD, others championed another site, while some opposed
 relocation entirely. All were worried about embarking on a project riskier than anything previously
 attempted by the Council.
 When Smuts returned for another session of the Council-in-Committee in November 1960, he
 laid down the law. "The Department," he informed the Councillors, "would under no circumstances allow
 the Kabah Location to be extended" and another site had to be found. The UAA, he asserted, provided
 that it was the duty of the local authority to provide accommodation on a site approved by the Minister
 of Bantu Administration and Development, and he would impose a decision if the Council did not
 act.114 Only one Councillor advocated retaining Kabah at its present site. In January 1961 the Council
 reversed its 1945 decision and subsequent determinations regarding Kabah, and agreed to remove the
19
 Location.
 The Council's decision cannot be separated from the climate of growing political opposition by
 blacks, but especially by Africans which developed in Uitenhage, as elsewhere in South Africa during the
 late 1950s. In contrast to the earlier deliberations on the application of group areas to Uitenhage, the
 decisions regarding resiting of Kabah were made behind closed doors, in secret Council sessions.
 Whereas prominent members of the coloured, Indian, and Chinese committees had the opportunity to
 address the Councillors, the Council welcomed no representatives of the African people. Some
 Councillors did indeed provide a channel of communication to the ANC, and were willing to at least
 tolerate its presence in the township. But these attitudes became unpopular as protest movements
 developed in the town through the 1950s. As their suffering grew under the repressive central state
 policies increasingly being enforced through the local authority, African residents mounted more direct
 forms of protest against apartheid and its institutional embodiments, both nationally and locally. Politics
 polarized increasingly between the ANC and the Government, with the Council drawn inexorably towards
 the latter.
 Uitenhage was a stronghold of the ANC throughout the decade, and was one of the
 organizations' most active branches in the Cape. Uitenhage was one of the best-mobilized centres in the
 Eastern Cape during the 1952 Defiance Campaign.115 Unlike Port Elizabeth, where many protests
 occurred at the New Brighton train station, civil disobedience in Uitenhage generally occurred in the
 centre of the "white" town at the market square, railway station, and law courts. Yet these were
 generally peaceful protests and many of the criticisms of unjust laws were shared (at least privately - and
 unbeknown to the protestors) by UP Councillors. Whatever liberal sympathy might have existed for
 Africans, however, did not extend to subsequent Congress campaigns, which met a far less tolerant
 response.
 The 1955 school boycott was extensive in the town, even after ANC provincial and national
 leaders distanced themselves from the campaign.116 The most serious confrontations occurred over the
 application of Pass Laws to women in 1957 when police violently dispersed anti-pass protesters, and
 made mass arrests.117 The education boycotts and pass protests were followed by the prominent 1958
 General Election stay-at-home, and by a number of more localized strikes and protests.
 In part, the change arose from the altered party composition of the Council. The newly-elected
 NP members of the Council made important political and ideological interventions on the body:
 bi-lingualism in all Council meetings and documents was first introduced in 1954 for example, and NP
 members led efforts to segregate town parks and the morning market.
 Perhaps the most significant NP representative was Frans Conradie, who served between 1953
 and 1957, when he was elected to the Provincial Council. His rise in the NP was meteoric: married to a
 prominent Afrikaans writer, he was a local and Eastern Cape Representative on the South African
 Bureau of Racial Affairs, Member of the Provincial Council, then member of the Executive Council of
 the Cape, and later MP. But these positions were facilitated through Conradie's real connections,
 growing out of his membership in the Broederbond, where he served on the Executive, and from the
 early 1960s, as Convenor of its secret watchdog committee on Coloured Affairs, and as a member of the
 committee on Relations with English.118 These connections would prove invaluable for the NP and the
 central state in the upcoming efforts to extract compliance from the local Council. Conradie's law
 partner, G J. Pieterse, was also a member of the Broederbond, and served intermittently on the Town
 Council from 1958 to the early 1970s. Pieterse was secretary of the first branch of the Afrikaanse
 Sakekamer to open in a town once dominated by English-speaking businessmen.119 It was these
 individuals who led the local attack against the Faganite tradition on the Council, and especially against
 the ANC. Their first major initiative was a 1954 attempt to thwart the planned annual meeting of the
 Cape ANC at Kabah even though a permit had been granted and the ANC had already paid the fees for
 the Location hall.120
 As in Port Elizabeth after the 1952 riot, the internal and external NP actors were able to
 intensify pressure on the UP Council by using the local "unrest" as a pretext to justify the increased
 restrictions provided through the UAA and GAA. Councillors who had previously opposed stronger
 controls now sought methods, including registration of service contracts, to eliminate such "undesirable
20
 elements" from the town.121
 Where Indian organizations, which represented a mere 1% of the population, had been able to
 influence Council policy directly, African organizations, most especially the ANC, could mobilize large
 constituencies throughout the 1950s, but could not gain the ear of the Council. In contrast to both the
 way it handled the group areas legislation, and the manner in which it discussed T.B. Floyd's 1945
 recommendations for removing Kabah, the Council agreed to the main lines the NAD policy without
 concern for the interests of those most affected, especially once the Government eased the Council's
 fiscal burden.
 2. Group Areas, Again
 When the Secretary of the East Cape Committee of the Group Areas Board (the transformed
 LTAB) saw press reports in mid-1961 that the Council had agreed to resite Kabah, he wrote to the Town
 Clerk to renew the campaign for group areas.122 Since Kabah would eventually be cleared, plenty of
 prime land would be opened up for possible coloured settlement. The thorny group areas problem could
 be simplified.
 On 17 July 1961 the Council met to consider his request and a UP Councillor managed to push
 through a resolution declaring "that the Council does not wish to discuss the question...or to submit
 proposals..." But where in years past large majorities stood on principle in rejecting any participation
 with the GAA, by 1961 the Councillors lamely claimed to be too preoccupied with plans to resite the
 African location to take up such a vexing issue. And even then the resolution passed only narrowly.
 "Council snubs Group Areas Board," was the headline in the Johannesburg-based Sunday Times,
 while the Port Elizabeth Evening Post editorialized that "Most Councillors have stood firm against the
 great hardships which compulsory racial grouping would bring." But the pro-Government Die Oosterlig
 interpreted events differently. "The constituency in which the town lies has sent representatives of
 apartheid to the Assembly and Provincial Council. The majority in the City Council is therefore a mere
 island in the political sea." The editorial concluded, "'Render unto Caesar that which is Caesar's' is a
 good idea for them to bear in mind."123
 The Afrikaans daily had a far clearer reading of the prevailing balance of political power. In the
 1958 General Elections the NP greatly increased both its number of seats and its share of the popular
 vote. In both the 1960 referendum and the 1961 general election it one a majority of the (white) popular
 vote, which translated into a massive parliamentary majority. The local UP Councillors could hardly
 claim to be representing the majority of white opinion when their national party was being consistently
 beaten in election after election. In June 1962 the Council accepted without opposition a resolution to
 submit group areas proposals to the Department of Community Development.
 However, agreement was possible only once a compromise had been forged to reduce the
 hardships for Indian and Chinese traders. The proposal came from none other than G.J. Pieterse, of the
 local law firm Conradie, Pieterse & Campher. Through his past participation on the Council, Pieterse
 was well aware that the body had no objections to residential segregation, but that it balked at any
 restrictions on business, which "should be left undisturbed."124 The Council's position clearly
 contradicted the principles of the GAA, and would lead to permanent deadlock. But Pieterse identified
 a way around the impasse: he recommended that the Council petition the government to declare a free
 trade zone, allowing "non-white" businesses to continue in something like their current form.
 Pieterse developed the strategy while preparing an application for a local Indian businessman
 who sought a group areas permit. His client was none other than A.S. Kooverjee, a landowner,
 businessman, and member of the Uitenhage Indian Congress. Kooverjee was also a central participant in
 the delegation which had successfully pressured the Council to reverse itself in 1953 and refuse to
 cooperate with the LTAB. Kooverjee had made a shrewd choice of lawyers. Pieterse's partner was now
 an insider in Broederbond and Cape provincial government policy-making circles; the lawyers' knowledge
 of the Council and their connections to the responsible central ministries made them extremely
 well-positioned to broker a group areas deal between the Council and the government.125
21
 With hopes of obtaining such an exemption, the Council's major objection to the GAA fell away,
 and four months later it voted unanimously to cooperate with the central Government in the planning of
 group areas and agreed to submit formal proposals for application of the Act to the town. From that
 date until the proclamation of group areas in Uitenhage on 20 October 1967, the Council cooperated
 with the Government. [See Map 4] A familiar argument - well rehearsed in Port Elizabeth in 1952 -
 served as justification for its action: "Now...the Council has decided to prepare a plan, arguing that if it
 does not do so, the Government will intervene...and impose something far more drastic for the victims of
 the Act."126 By "trying not to be cruel," as the Evening Post put it, the Council reversed its
 long-standing position of noncompliance with a centrepiece of apartheid legislation. An era of Municipal
 contestation of apartheid policy had come to a close.
 Local Afrikaner notables had spun a new web of influence to the central state and to the NP,
 shouldering aside the old UP political network whose connections to the centre had been severed since
 1948. Where Councillors and officials, especially the Town Clerk, had faced a decade of governmental
 coercion and brazen interventions, the 1960s promised a different mode of politics. The space for
 opposition had all but disappeared, and the new power brokers no longer had English surnames. The
 character of politics in the town was transformed more thoroughly than at any time since the
 reestablishment of British rule at the Cape in 1806.
 Uitenhage Town Councillors, unlike their counterparts in Port Elizabeth, had consistently framed
 their choices as either noncooperation with an unjust policy, or working within the legal framework to
 apply the harsh regulations as fairly as possible: "trying not to be cruel." At the end of the day, local
 defensive battles could delay but not reverse central policy. Faced by a central state willing to wield new
 resources in an effort to gain control, local officials accepted the second option, to work within the rules
 of the game as established by the NP, fearing that policies would be forced upon them, as occurred in
 Johannesburg and Cape Town. These rules were as much ideological as legal or bureaucratic, and they
 became the basic framework within which the Council would pursue policy towards Africans and
 coloureds until the 1980s.
 In playing by the rules, however, the Council was aided by a compromise which carefully
 accommodated the one black constituency towards which Councillors could least afford to be cruel.
 Coloureds, Indians, and Chinese businessmen and property owners had access to the Council by virtue of
 property and the franchise, and they used their assets at key points to influence the Council. For
 Africans, no such avenues existed to the Council or to the state. Possessing neither property nor the
 vote, they had little to offer to the Council. Indeed, the issues motivating their growing protest
 movement in the 1950s could not be satisfied except through a transformation of a political system in
 which the Council itself played a crucial role, and which even the most liberal Councillor had no interest
 in altering. The few whites who were willing at least to entertain African grievances became increasingly
 isolated as the decade passed and the tide was running against their form of paternalism. Between 1957
 and 1960 the fate of the Uitenhage's century-old African community was sealed behind closed doors
 where economic efficiency and the tax burden to be borne by the largely white ratepayers were the only
 factors taken into account. Council hostility to the removals dissipated once the central state agreed to
 assist in the financial burden of relocation.
 Not only did the Council agree to virtually the same policy they had rejected when it was
 suggested in 1945 by their own expert town planning consultant, but they revived another legacy of the
 town's past. By purchasing the farms "Boschoogte" and "Naroes" for the new township, they unknowingly
 sited the new township, Kwanobuhle, "Place of Beauty," on land used as a concentration camp during the
 Anglo-Boer War.
 TV. CONCLUSION
 A central theme of this paper is the growing central state control over local affairs, particularly
 over the reproduction of the urban black population. Before 1948 the Council was relatively unhindered
 by the central state in its efforts or (more accurately) non-efforts to house blacks. In practice the relative
 freedom meant inaction, as neither the Council nor the Government would supply sufficient funds for
 upgrading black communities. But through the 1950s, and especially after 1957 the space for local
22
 opposition narrowed considerably. Three key factors account for this process.
 First, the central state formally narrowed the scope of Municipal discretion in implementing
 central state policies; the respective Ministers came to wield virtually dictatorial powers to pursue policies
 over the heads of recalcitrant local authorities. The process culminated in the early 1970s when
 Municipalities were formally stripped of authority over their coloured, Indian, and African residents.
 Second, the central state began to exert greater bureaucratic pressure on the Councils by
 reorganizing lines of authority. Central ministries which had not previously been involved in particular
 issues now held veto power over local decisions.
 Third, the new bureaucratic machinery at the national level and the political will to use it gave
 the central state much greater power to manipulate local authorities' fiscal dependence, thus enabling it
 to force agreement with national policy. On the one hand, departments could threaten to withhold funds
 for completely unrelated projects to force compliance with directives of NAD or the LTAB. On the
 other hand, extraordinary grants of funds could be made to defray the cost to local authorities of
 implementing national policy, thus enabling the central state to buy compliance. These deals were
 brokered through a new stratum of Afrikaner political operators, replacing the old UP businessmen and
 politicians who dominated many Councils before 1948, and linking the local authority to the central state
 and National Party.
 Over the previous decade the central state had reorganized important parts of the state
 apparatus, increased its penetration of the state as a whole, and had developed its legal/ideological and
 administrative capacities to intervene more directly in civil society in pursuit of its policy goals. By
 1961-62 as the costs of opposition increased, as the benefits to be gained by cooperation increased, and
 as the likelihood that the NP would never be replaced by a UP government grew, Councillors wavered in
 opposition and quietly went ahead with implementing apartheid policy. In the end, however, the central
 state was able to force local authorities to comply with apartheid policy - "to Render unto Caesar" - but it
 could not remake the world in its own image. The Council's stalling efforts unintentionally succeeded in
 delaying proclamation of Group Areas in Uitenhage until 1967 (six years after Port Elizabeth, even
 though Uitenhage presented far fewer planning problems), and the removal of the Kabah Location did
 not start until December 1968. The removals to Kwanobuhle were continually stalled by funding
 problems, which delayed the construction of more houses, thereby further slowing group areas removals,
 since Africans had to be removed from the land in Kabah before housing for coloureds could be built.
 Both projects were thus consistently slowed so that through the 1970s there were still parts of town
 where coloureds and Africans lived either in the same area or very close together.
 When the state was finally able to embark on the segregation project, it faced the new political
 economy of the 1960s boom: the expanding motor industry, and a burgeoning industrial workforce.
 These demands on local authorities were not accommodated in the central state planners' blueprints. As
 the UP Council had been caught in the late 1940s with a dated solution unable to cope with
 contemporary problems, so too the Government's agenda for reorganizing urban residential and
 commercial space could not cope with changes brought about by the economic boom of the 1960s.
 At the very moment when Africans and coloureds were moving into semi-skilled and skilled job
 positions in the expanding automobile industry, the state was running into extreme difficulties in its task
 to segregate the black working class in Uitenhage. In Port Elizabeth, the segregation between coloureds
 and Africans was completed in the 1950s, before the industrial boom, and subsequent growth in the
 working class could then be channelled into racially separate communities. In Uitenhage, the delays in
 the application of group areas and urban areas legislation meant that the 1960s expansion occurred in
 communities not yet reconstructed by apartheid planning principles. Working class formation in Port
 Elizabeth was thus fractured on lines of race between coloured and African workers, bottled up in
 spatially separate residential areas deliberately planned by apartheid bureaucrats to impede social
 interaction across communities. In this respect, the application of apartheid planning principles to Port
 Elizabeth intensified a pattern of racial segregation in place from the colonial era, and led to the almost
 total separation of races. Class became relatively less salient as a principle of residential ordering, and
 communities became differentiated on racial grounds.
23
 By contrast, the pattern of segregation in Uitenhage was not nearly as well-ordered before 1948.
 While blacks and whites were generally separated, the Kabah location was a mixed race community,
 where most coloureds lived either with or in close proximity to African people. As a result of struggles
 between the central state and the City Council, and later due to financial difficulties in building suitable
 accommodation, the NP's racial policies did little to alter these patterns, at least not before the mid-
 1970s. As a result, working class formation in Uitenhage occurred on a terrain not fully defined by
 apartheid principles, not fully differentiated on racial grounds, which allowed possibilities for contact and
 coordination across racial lines in common communities. These social links created rich bonds of
 familiarity between coloured and African workers, and networks of social interaction in Uitenhage, while
 the more thorough application of apartheid in Port Elizabeth impeded the development of such bonds -
 indeed, was targeted at destroying them. Such bonds between African and coloured workers would soon
 give a powerful boost to the growth of non-racial trade unionism in Uitenhage from the early 1970s,
 when industrial organization revived.
24
 NOTES
 1. "Koppige Stadsraad," Die Ooslerlig, 21 July 1961.
 2. Deborah Posel, The Making of Apartheid, 1948-1961: Conflict and Compromise (Oxford: Oxford University Press,
 1991), vii.
 3. A,J. Christopher, "Formal Segregation and population distribution in Port Elizabeth," Contree 24 (1988): 5.
 4. A.J. Christopher, "Race and Residence in Colonial Port Elizabeth," South African Geographical Journal 69, no. 1
 (1987): 6.
 5. Christopher, "Race and Residence," 5.
 6. J.B- Peires, The House ofPhalo: A History of the Xhosa People in the Days of Their Independence (Berkeley:
 University of California Press, 1982), 74-76. Gary Baines, "The emergence of black labour unionism in Port Elizabeth
 and the disturbances of October 1920," (MA diss., Rhodes University, 1989.
 7. W.S.J. Sellick, Uitenhage Past and Present: Souvenir of the Centenary, 1804-1904 (Uitenhage: Uitenhage Times,
 1904), 70. Baines, "The emergence of black labour unionism."
 8. Maynard Swanson, "The Sanitation Syndrome: Bubonic Plague and Urban Native Policy in the Cape Colony,
 1900-1909," Journal of African History 18 (1977): 387-410. See also Joyce F. Kirk, "The African Middle Class, Cape
 Liberalism and Resistance to Residential Segregation at Port Elizabeth, South Africa, 1880-1910," (Ph.D. diss.,
 University of Wisconsin, 1987), 193. Christopher, "Race and Residence," 10; Gary Baines, "The colonial origins of
 segregation: The case of Port Elizabeth's Native Strangers' Location," paper presented to the Eleventh Biennial
 Conference of the South African Historical Society, University of Stellenbosch, January 1987.
 9. Baines, "Black Labour Unionism;" and Kirk, "The African Middle Class." See also Joyce F. Kirk, "A 'Native'
 Free State at Korsten: Challenge to Segregation in Port Elizabeth, South Africa, 1901-1905," Journal of Southern
 African Studies, 17, no. 2 (June 1991): 309-334.
 10. Cape of Good Hope, Reports on the Public Health for the Year 1896 [G.42-'97], 145; Census of the Cape of Good
 Hope 1904, Preliminary Report of the Director of the Census [G.80-1904], 58.
 11. Christopher, "Formal Segregation," 7.
 12. Port Elizabeth Municipality [henceforth PEM] 220/71. Housing Department, "Report on Housing," 20th September
 1963.
 13. Republic of South Africa, Department of Statistics, Population of South Africa, 1904-1970 (Pretoria: The
 Government Printer, 1976). John A. Banach, "The Cape Midlands: Its Demography (1911-1960) and Regional Income
 (1954/55-1959/1960)," Occasional Paper Number 14, Institute for Social and Economic Research, Rhodes University,
 1969, pp. 53-58.
 14. By contrast, the national ratio of African men to women in towns was 2.2:1 in 1936. Doug Hindson, Pass Controls
 and the Urban African Proletariat (Johannesburg: Ravan Press, 1987), 53. Cape Archives (henceforth CA) 3/Uit 42
 File 261o, Uitenhage Municipality Location Census, 23 June 1930; CA 3/Uit 37 File 261 v. 9, July 1934-May 1935,
 Uitenhage Municipal Locations, n.d.
 15. These demographic characteristics distinguish both Port Elizabeth and Uitenhage from almost all other urban areas
 in South Africa, where for the first half of this century the African population was overwhelmingly comprised of
 migrant workers.
 16. CA 3/Uit 38 File 261 v. 12 August 1936-January 1937, letter from Town Clerk to the Secretary for Native Affairs,
 19 August 1936, "Proposed Abolition of Oatlands Location".
 17. Among adults, the unemployment rate in 1930 was less than 10%; in 1935, it had risen to 36%. CA 3/Uit 42 File
 261o, Uitenhage Municipality Location Census, 23 June 1930 and CA 3/Uit 37 File 261 v. 9, July 1934-May 1935,
 Uitenhage Municipal Locations, n.d.
25
 18. CA 3/Uit 38 File 261 v. 11, November 1935-July 1936, letter from Town Clerk, Uitenhage to Town Clerk,
 Graaff-Reinet, 9 April 1936, "Housing Schemes for Cape Coloureds."
 19. Union Gazette 2143, 13 October 1933, Proclamation No. 222, 1933. CA 3/Uit FUe 261 v. 9, July 1934-May 1937
 Report of Locations Sub-Committee, 17 August 1934, p. 3.
 20. CA 3/Uit 38 File 261 v. 11, November 1935-July 1936, letter from Town Clerk to Town Clerk, Graaff-Reinet, 9
 April 1936, "Housing Schemes for Cape Coloureds."
 21. CA 3/Uit V. 37 File 261 V. 9 Location Matters General July 1934 to May 1935, "Overcrowding and dilapidated
 Houses at Locations: Recommendations referred by Health Committee." CA 3/Uit 37 File 261 v. 9 July 1934-May
 1935, Minutes of Native Affairs Committee.
 22. CA 3/Uit 38 File 261 v. 12 August 1936-January 1937, letter from Town Clerk to the Secretary for Native Affairs,
 19 August 1936, "Proposed Abolition of Oatlands Location."
 23. CA 3/Uit 4/1/191 File 26 v. 1, Census and Statistics January 1937-October 1943, Extract from report of Medical
 Officer of Health, 7 August 1941, "Vital Statistics for Year Ending 30 June 1941."
 24. CA 3/Uit 4/1/225 Housing General 74 v. 1, March 1940-April 1945, extract from report of Location
 Superintendent, 7 December 1943.
 25. Historical and Literary Papers Library [henceforth HLP], University of the Witwatersrand, Margaret Ballinger
 Papers, A410/B2.14.13 Correspondence, Native Affairs, Land 1941-1951, letter from A. Park, Mayor of Uitenhage to
 Mrs. M. Ballinger, 3 May 1944, "Right of Natives to Acquire Freehold Property within Urban Areas."
 26. CA 3/Uit 4/1/225 Housing General 74 v. 1, March 1940-April 1945, Joint Report of Medical Officer of Health,
 Acting Town Engineer and Location Superintendent to Housing Committee, "Sites for Proposed Sub-Economic Housing
 Schemes," 11 January, 1945.
 27. The Tunes [UitenhageJ, 4 December 1946; see also Floyd's own books: T.B. Floyd, Town Planning in South Africa
 (Pietermaritzburg: Shuter and Shooter, 1960), 48-49, and T.B. Floyd, More About Town Planning in South Africa
 (Pietermaritzburg: Shuter and Shooter, 1966).
 28. Uitenhage Municipality [henceforth UM] 13/1 "Interim report on general lines of development for Uitenhage," by
 T.B. Floyd, May 1946.
 29. UM 13/1, Report of Proceedings of Town Council, 27 June 1945 "Town Planning Scheme," and UM 13/1 Extract
 from Council Minutes, 17 September 1945; The Times, 25 July 1945, "Town Planning at Uitenhage."
 30. UM 13/1, Extract from Council Minutes, 17 September 1945, and UM 13/1, Minutes of Council, 19 November
 1945, extract of Report of Native Affairs Committee, 12 November 1945.
 31. See A.W. Stadler, "Birds in the Cornfield: Squatter Movements in Johannesburg, 1944-1947," in Labour,
 Townships and Protest, edited by Belinda Bozzoli {Johannesburg: Ravan Press, 1979); and Amanda Younge, "Housing
 policy and housing shortage in Cape Town: 1942-1980," Africa Perspective 21 (1982): 12-13.
 32. Floyd, Town Planning in South Africa, 84. See also Peter Wilkinson, "Providing 'Adequate Shelter': The South
 African State and the Resolution of the African Urban Housing Crisis, 1948-1954," in Working Papers in Southern
 African Studies, Volume 3, edited by D.C. Hindson (Johannesburg: Ravan Press, 1983), 66-67, 76.
 33. CA 3/Uit 4/1/184 File MJ20 v. 1, Location Enquiry, June 1947-October 1952, letter from Town Clerk to Chief
 Native Commissioner, King William's Town, 11 September 1947, "Uitenhage Municipality: Proposed Native Village."
 34. UM, letter from Town Clerk to Mr. S.I. Misrock, 21 March 1950.
 35. CA 3/Uit 4/1/184 MJ 20 v. 1, Location Enquiry, June 1947-October 1952, "Report of Departmental Executive
 Committee on Enquiry into Administration of Kabah Location," 21 July 1948.
16
 36. The first coup came with the decision by US tire manufacturer Goodyear to Bite its only South African factory in
 Uitenhage. In the same period the state-funded Industrial Development Corporation financed a new textile company,
 Fine Wool Products [Veldspun], employing more than 500 persons, and taking advantage of the town's proximity to the
 sheep farms of the Cape Midlands. The South African Railways began a phased expansion of the existing railway
 workshops, boosting employment well beyond 1,200, and a tanning company and another textile firm opened factories.
 Finally, in 1946, Studcbaker formed South African Motor Assemblers and Distributors [SAMAD] to manufacture cars
 under license from the American parent, the first roots of the firm which would eventually become Volkswagen of
 South Africa.
 37. B.D. Phillips and C. de Coning, "Secondary Industry in the Port Elizabeth/Uitenhage Region: a Structural
 Analysis," Research Report no. 2, Institute for Planning Research, University of Port Elizabeth, 1969, p. 57.
 38. Between 1936 and 1946 the African population increased 5.2% per annum, but between 1946 and 1951 rate jumped
 to 13.8% per year; the rate of increase between 1936 and 1946 exactly parallels the national average, while the rate for
 the following five years far surpasses the national figure of 6.6%. Hindson, Pass Controls, 67.
 39. CA 3/Uit 4/1/184 MJ20 v. 1, Location Enquiry, June 1947-October 1952, Record of Proceedings of Departmental
 Executive Committee Enquiry into Administration of Location Department, 21 November 1947.
 40. CA 3/Uit 4/1/191 File 26 v. 2, Census and Statistics, March 1944 to April 1950, "Population. Uitenhage Town and
 Location," 10 April 1947; UM File 26 v. 3, Census and Statistics, May 1950 to August 1956, Letter, Uitenhage
 Magistrate to Town Clerk, "Population Census, 1951," 16 July 1951.
 41. CA 3/Uit 4/1/184 MJ20 v. 1, Location Enquiry, June 1947-October 1952, p. 7.
 42. CA 3/Uit 4/1/184 MJ20 v. 1, Location Enquiry, June 1947-October 1952.
 43. Hindson, Pass Controls, 59.
 44. Stanley Greenberg, Legitimating the Illegitimate: State, Markets, and Resistance in South Africa (Berkeley:
 University of California Press, 1987), 38.
 45. As Posel stresses, the report was less a bold blueprint than an unwieldy compromise between different
 interpretations of apartheid within Afrikaner nationalism. Not only did the report itself fudge the differences between
 "purist" and practical versions of apartheid, but the legislation derived from the Sauer Report was subjected to further
 concessions to the UP opposition and (if only indirectly) to the popular resistance mounted by the ANC, SACP and
 other forces in civil society. See Posel, The Making of Apartheid, Chapters 2-4.
 46. Hindson, Pass Controls, 59; and Deborah Posel, "The Meaning of Apartheid before 1948: Conflicts of Interests and
 Powers within the Afrikaner Nationalist Alliance," Journal of Southern African Studies 14, no. 1 (1987): 136-139.
 47. Act no. 41 of 1950, Statutes of the Union of South Africa, 1950 (Pretoria: The Government Printer, 1950). The
 Act required extensive amendment throughout the 1950s, to overcome initial inconsistencies and oversights, and to
 address new problems unforeseen by the Government. In addition, legal tests, and as will be clear below, popular
 challenges as well as resistance within the state continually sent the Government's lawyers back to the legislative
 drafting table. See Alan Mabin, Comprehensive Segregation: the Origins of the Group Areas Act and its Planning
 Apparatuses," Journal of Southern African Studies 18, no. 2 (June 1992): 405-429.
 48. Muriel Horrell, Laws Affecting Race Relations in South Africa (Johannesburg: South African Institute of Race
 Relations, 1978), 71; see also L.R. Dison and I. Mohamed, Group Areas and Their Development (Durban:
 Butterworths, 1960).
 49. F.P. Rousseau, Handbook on the Group Areas Act (Cape Town: Juta, 1960), 44.
 50. The Act was the official justification for the forced removal of more than 830,000 people between 1960 and 1980.
 Surplus People Project, Forced Removals in South Africa, Volume 1: General Overview (Cape Town: Surplus People
 Project, 1983), 6.
 51. Horrell, Laws, 262-3.
27
 52. CA 3/Uit 4/1/191 File 26 v. 2, Census and Statistics, March 1944-April 1950, "Population: Uitenhage Town &
 Location."
 53. Quoted in Greenberg, Legitimating the Illegitimate, 39.
 54. Deborah Posel, "Interests, Conflict and Power: the Relationship Between the State and Business in South Africa
 During the 1950s," paper presented to the Annual Conference of the Association for Sociology in Southern Africa,
 University of Cape Town, 1985, p. 15.
 55. Simon Bekker and Richard Humphries, From Control to Confusion: The changing role of Administration Boards in
 South Africa, 1971-1983 (Grahamstown: Institute of Social and Economic Research, 1985), 205-210. The authors
 identify the important amendment to the Urban Areas Act, but do not dwell upon its consequences for local authorities.
 56. Hindson, Pass Controls, 63.
 57. UM 74L/1, letter from Secretary for Native Affairs to Town Clerk, "UM Proposed Coloured Township adjoining
 Native Location."
 58. UM 74L/1, letter from Town Clerk to Secretary for Native Affairs, 7 November 1951, "Uitenhage Municipality:
 Proposed Assisted 3/4% Housing Scheme for Coloureds."
 59. See Wilkinson, "Providing Adequate Shelter," 77.
 60. UM 74/5, 30 October 1951, "Report of the Town Clerk to Housing Committee."
 61. UM, Town Clerk to Native Affairs Committee, 24 January 1953; Memorandum from Superintendent of Kabah to
 Town Clerk, 9 December 1952, "Proposed Coloured Housing Scheme & Buffer Strip."
 62. William J. Davies, "Patterns of Non-White Population Distribution in Port Elizabeth with Special Reference to the
 Application of the Group Areas Act," Series B, Special Publication no. 1, Institute for Planning Research, University of
 Port Elizabeth, 1971, pp. 25-6.
 63. UM Town Clerk's Report to Finance and General Purposes Committee 8 September 1951, "Group Areas Act."
 64. By Proclamation no. 220 of 1951. The proclamation introduced "occupational control" and "control of acquisition"
 such that no one could acquire ownership of a property from an owner of a different racial group, nor could anyone
 occupy property which had been occupied by a person of a different group prior to the proclamation.
 65. UM, Report of the Town Clerk to Finance and General Purposes Committee, 8 August 1953.
 66. UM, 207/2, Letter to Mayor from Uitenhage Indian Congress, 21 September 1953, "Defined Area of the Group
 Area Act."
 67. UM, Town Council Minutes, 21 September 1953.
 68. UM 207/2, Report of the Finance and General Purposes Committee, 12 October 1953.
 69. UM 207/2, Report of the Finance and General Purposes Committee, 12 October 1953.
 70. UM, Report of Town Clerk to Finance and General Purposes Committee, 12 June 1954; UM, Record of meeting
 between the Government's Reference and Planning Committee and Representatives of the Uitenhage Municipality, 9
 June 1954.
 71. UM, Report of the Finance and General Purposes Committee, 5 December 1955.
 72. UM, Minutes and Proceedings of Town Council-in-Committee, 27 March 1956.
 73. UM, Town Treasurer to Town Clerk, "Proposed Establishment of Group Areas in Uitenhage," 4 October 1961;
 UM 207/2, "Meeting of Subcommittee appointed to draft Proposals for Group Areas," 4 March 1955.
28
 74. In 1958 the NP failed in its efforts in the Cape Provincial Council to remove coloureds from the common Municipal
 roll, and place them on a separate roll to elect local boards in their own Group Areas. Coloureds remained eligible as
 municipal voters in the Cape until 1972- Horrell, Laws, 57-9. See Gwendolen M. Carter, The Politics of Inequality:
 South Africa Since 1948 (New York: Praeger, 1958), especially Chapter 4, for an excellent account of the crisis caused
 by the National Party's schemes to eradicate coloured franchise rights entrenched in the 1910 constitution; and Leonard
 Thompson, The Cape Coloured Franchise (Johannesburg: Institute of Race Relations, 1949).
 75. UM, Town Clerk "Concentration of Coloured Group in Uitenhage," 30 October 1959.
 76. A.S. Du Toit, "Housing and Population Study of the European Population of Uitenhage," Research Report no. 5,
 Institute for Planning Research, University of Port Elizabeth, 1970, p. 9.
 77. Jennifer Robinson, "The Power of Apartheid," Chapter 7, "Forging a new mode of control: Riots, Removals and
 Regulations."
 78. Tom Lodge, Black Politics in South Africa Since 1945 (Johannesburg: Ravan Press, 1983), 52-54.
 79. Port Elizabeth and Uitenhage accounted for more than 30% of the national total of those arrested during the civil
 disobedience campaign. Lodge, Black Politics, 46; see also David Carter, "The Defiance Campaign: A Comparative
 Analysis of the Organization, Leadership, and Participation in the Eastern Cape and the Transvaal," Collected Seminar
 Papers on the Societies of Southern Africa in the 19th and 20th Centuries, vol. 2 (London: Institute of Commonwealth
 Studies, University of London, 1971).
 80. In September 10 ANC Eastern Cape leaders were arrested in coordinated swoops, charged under the Suppression of
 Communism Act. The state action sparked a new round of protest as 350 people were arrested the next day defying
 segregation at rail stations, the biggest single protest since the Defiance Campaign began. Almost as a prelude to the
 October riot, a crowd gathering at New Brighton station grew angry when a Railway policeman accosted a woman
 defier, prompting the crowd to surge around the officer. Unlike the policeman who shot the suspected paint thief on
 October 18, this officer wisely "released the woman...and the crowd settled down." "350 Defiers arrested in PE 24
 hour protest campaign," Evening Post, 12 September 1952. The event was but one of a growing number of such
 clashes between Port Elizabeth blacks and police reinforcements and army detachments that had been rushed to the
 area.
 81. Even more violent riots had occurred in East London and on the Witwatersrand, where influx control was firmly in
 place, but Verwoerd used the New Brighton event as a pretext for a very public campaign against his opponents.
 82. "Verwoerd issues ultimatum to P.E.," Evening Post, 23 October 1952; "Bus Boycott in New Brighton," Evening
 Post, 21 October 1952.
 83. "Rumours follow column," Evening Post, 22 October 1952.
 84. "Bus Boycott in New Brighton," Evening Post, 24 October 1952; "Bus boycott called off. Armed guards are
 removed," Evening Post, 28 October 1952.
 85. At the meeting, G.A. Brand, Urban Areas Native Commissioner for the Cape Eastern Area, reiterated Verwoerd's
 views on the failure of the "experiment" and his threat to intervene if the Council failed to apply influx control. PEM,
 Minutes of Proceedings of Council-in-Committee, 28th October 1952; "Meetings banned in P.E. from today," Evening
 Post, 7 November 1952; "Curfew and Riotous Assemblies Act now in force," Eastern Province Herald, 8 November
 1952.
 86. The position was reiterated in subsequent approaches to the Council by other state officials. The Native
 Commissioner for Port Elizabeth wrote to the Town Clerk that under the Native Laws Amendment Act only workers
 who are required for industries will be allowed to enter the town, while those who are removed will be barred from
 reentering. "What effect, if any," he wrote, "do you consider will the removal of a large number of unemployed and
 'won't work' natives have on the defiance campaign and crime in general?" PEM 169/7, Native Affairs in City,
 February 1952-November 1952, letter from Native Commissioner PE to Town Clerk, 5th November 1952.
 87. PEM 169/7, Native Affairs in City, February 1952-November 1952, letter from G.X. Tshume, Secretary ANC
 Eastern Cape Regional Committee to Town Clerk, 3 November 1952; see also Lodge, Black Politics, 54.
29
 88. PEM 169/7, Native Affairs in City, February 1952-November 1952. Schedule of returns submitted in reply to
 Departmental Circular no. 79/1952. PEM 169/8, Native Affairs in City, November 1952-December 1952, Minutes of
 Proceedings of Native Affairs Committee, 14 November 1952; "Hundreds without their jobs today," Evening Post, 11
 November 1952.
 89. Robinson, "The Power of Apartheid," Chapter 7. "Levy Act is law from January 1, P.E. must comply," and "P.E.
 told to control influx, Minister sends reproof," Evening Post, 14 November, 15 December 1952.
 90. PEM 169/8, Native Affairs in City, November 1952-December 1952, letter from W.W.M. Eisclen, Secretary for
 Native Affairs to Town Clerk, 1 December 1952.
 91. PEM 169/9, Native Affairs in City, January 1953-May 1953, Minutes of Proceedings of Special Meeting of the
 City Council, 15 January 1953; Minutes of Proceedings of Council-in-Committee, 15 January 1953.
 92. Davies, "Patterns of non-white population distribution," 110-112.
 93. "Outline of P.E. planning scheme," Evening Post, 5 December 1952; "P.E. policy has made race zoning easier,"
 Evening Post, 6 December 1952; Davies, "Patterns of nonwhite population distribution," 115-117.
 94. Robinson, "The Power of Apartheid," Chapter 7.
 95. For a convincing explanation, see Janet Cherry, "The Making of an African Working Class - Port Elizabeth 1925-
 1963 (MA diss.. University of Cape Town, 1992).
 96. Bev Taylor, "Removing Port Elizabeth's Notorious Korsten Slums in the 1950s," unpublished paper, 1990, p. 24;
 and A.J. Christopher, "Apartheid Planning in South Africa: The Case of Port Elizabeth," The Geographical Journal
 153, no. 2 (1987): 200.
 97. Christopher, "Formal segregation," 11-12.
 98. Bev Taylor, "South End and Fairvtcw Removals and the Development of Bethelsdorp," unpublished paper, 1990, p.
 5.
 99. Taylor, "South End and Fairview Removals."
 100. Department of Geography, University of Port Elizabeth, and PEM 196/'', Union Government Census, 1960.
 101. Christopher, "Apartheid planning in South Africa," 202.
 102. UM, Minutes of Natives Affairs Committee, 29 January 1951.
 103. UM, Town Clerk's report to the Native Affairs Committee, 26 September 1953.
 104. Horrell, Laws, 77; Union of South Africa, Statutes of the Union of South Africa, 1952 (Pretoria: Government
 Printer, 1953), 979.
 105. UM, Report of the Town Clerk to Finance and General Purposes Committee, 6 October 1956, "Round Table
 Conference with Industry."
 106. UM, Town Council of Uitenhage, Mayoral Review 1956/1957, 6 September 1957, p. 21.
 107. UM, 207/3, Record of Meeting of Sub-Committee of Finance Committet; with Eastern Cape Committee of Land
 Tenure Advisory Board, 30 September 1955.
 108. UM, Report of the Town Clerk to Native Affairs Committee, 27 July 1957, "Visit of Mr. Brownlee."
 109. UM, Committee Reports and Minutes, January-June 1958, Special Report of Town Clerk to Native Affairs
 Committee and to Finance and General Purposes Committee, "Further Development in Native and Coloured Areas and
 Associated Matters," 14*15; UM 74L/2, Memorandum from Superintendent of Kabah to Town Clerk, "Proposed
 Coloured Housing Scheme & Buffer Strip," 9 December 1952.
30
 HO. UM, Minutes of Proceedings of Town Council, 21 October 1957, Report of the Native Affairs Committee, 30
 September 1957.
 111. UM, Minutes of Proceedings of Native Affairs Committee, 27 January 1958, "Registration of Native Service
 Contracts: Date of Operation in Uitenhage."
 112. UM, Town Clerk's Report to Special Committee Appointed to Investigate Further Developments in Native and
 Coloured Areas, 3 February 1960.
 113. UM, Minutes and Proceedings of Council-in-Committcc, 7 March 1960, "Proposed Resiting of Location."
 114. UM, Minutes of Proceedings of Council-in-Committee 3 November 1960, "Proposed Resiting of the Location."
 115. UM, Committee Reports and Minutes, Minutes of Finance and General Purposes Committee, 10 November 1952;
 and "Strike Hits Town's Textile Factories," Eastern Province Herald, 11 November 1952; "More than 600 strikers lose
 their jobs," Eastern Province Herald, 12 November 1952.
 116. See Lodge, Black Politics, 126-128. UM, Committee Minutes and Reports, July-September 1955, Report of
 Superintendent of Kabah to Native Affairs Committee, 23 July 1955, "School Boycott ANC."
 117. See Julia Wells, "The history of black women's struggle against pass laws in South Africa," (Ph.D. diss.,
 Columbia University, 1982); Cherryl Walker, Women and Resistance in South Africa (London: Onyx Press, 1982);
 Richard de Villiers, "Resistance to the extension of passes to African women, 1954-1960," African Studies Institute
 seminar paper, University of the Witwatersrand, 1979; and Tom Lodge, Black Politics, Chapter 6: "Women's protest
 movements in the 1950s," especially pp. 144-146. UM, Committee Reports and Minutes, Monthly Report of
 Superintendent of Kabah to Native Affairs Committee, 28 September 1957; and "Defiant women set fire to books,"
 Eastern Province Herald, 30 July 1957.
 118. Ivor Wilkins and Hans Strydom, The SuperAJrikaners; Ifiside the Afrikaner Broederbond (Johannesburg: Jonathon
 Ball, 1980), 398-405; also "September 17 is nomination day," Eastern Province Herald, 11 August 1956.
 119. "Sakekamer branch to meet today," Eastern Province Herald, 27 July 1957.
 120. Though their main proposal was turned back by the UP majority, the NP councillors pushed the Council to grant
 the permit only on condition that the SAP be allowed to attend each session, and after the affair, the NP members
 successfully urged the Council to adopt strong regulations modelled on those in New Brighton to control meetings.
 Councillor Ernest Reds "pointed out that the African National Congress was in any event holding weekly meetings in
 the open in the Location and that as this was not an illegal organization their meetings could not be stopped." During
 debate on the issue, Councillor Conradie said, "We cannot close our eyes to the fact that we are giving hospitality to an
 organization which, although not declared illegal by the Government, is no friend of the White man." UM, Town
 Clerk's Report to Native Affairs Committee, 27 March 1954; Minutes of Native Affairs Committee, 29 March 1954;
 Minutes of Town Council, 20 April 1954. "Move to stop A.N.C. talks fails," Eastern Province Herald 22 June 1954.
 121. UM, Report of the Superintendent of Kabah to Native Affairs Committee, 23 July 1955; UM, Committee Minutes,
 January-March 1956, Minutes of Proceedings of Native Affairs Committee, 14 February 1956.
 122. UM, Report of Town Clerk to Finance and General Purposes Committee, 11 July 1961.
 123. Sunday Times [Johannesburg], 23 July 1961; "Justice stand in Uitenhage," Evening Post, 1 August 1961; "Koppige
 Stadsraad," Die Oosterlig, 21 July 1961.
 124. UM, Town Council Minutes, 19 February 1962.
 125. UM, Town Council Minutes, 19 February 1962.
 126. "Trying not to be cruel," Evening Post, 14 June 1962.
UNIVERSITY OF THE WITWATERSRAND
 A F R I C A N S T U D I E S I N S T I T U T E
 African Studies Seminar Paper
 to be presented in RW 7003 (7th floor) at
 4.00p.m. MONDAY 5TH OCTOBER 1992
 Title: Skills, Control, and 'Careers at Work1: Possibilities for
 Participatory Management in the South African Motor Industry.
 by: Glen Adler
 No 326
INTRODUCTION
 Much of the writing on the South African motor industry assumes that the coming of
 the multinational automobile company brought with it the transplantation of the
 repertoire of production techniques developed under the rubric of Fordism. Whether
 inspired by Taylonst assumptions about scientific management or Braverman-inspired
 criticisms of the degradation of work, there is general agreement in the literature on
 the generally low levels of skill, high levels of monotony, and general dehumanisation
 associated with work in the auto plants. While most analysts pay close attention to
 the low-volume character of the local industry when discussing the crisis of high-cost
 production in a saturated market, no one has investigated the obvious point that low-
 volumes should have a dramatic impact on the nature of work itself.
 This paper departs from the conventional view of the labour process in the motor
 industry, and argues that analysis until now has failed to come to grips with the
 complexity of work in the industry. Indeed, rather than viewing the industry as one
 dominated by deskilling and dehumanisation driven by rapid mechanisation, the
 perspective developed here is that jobs require far greater thought and direct physical
 control over execution than is allowed for in the Fordist model. Workers not only
 perform more complicated tasks than similar workers in the United States and
 Europe, but are invariably able to perform a very wide range of jobs as well.
 Workers are not reduced to mere appendages of machines controlled ultimately by
 supervision or production engineers in time and motion departments. Rather, an
 elaborate system has developed in South African automobile plants where workers
 develop a wide range of skills, which may be described as "allrounder" status. Instead
 of seeing workers as possessing few or no skills, this paper makes use of the concept
 "tacit skills" to argue that automobile workers possess a wide array of carefully
 developed abilities.
 Workers' problems in the industry grow less out of the conventional grievances
 identified in the Fordist approach - monotony and degradation - than out of a deeply
 felt injustice at the heart of the system of job grading, reward, and promotion. The
 formal job grading system bears little similarity to the actual tasks workers perform,
 failing in the main to recognise their tacit skills and allrounder status. Nor is there
 any comfortable fit between jobs and level of education. Thus, while workers'
 knowledge on the job - their skills - are crucial ingredients in the company's success,
 workers themselves remain largely unrecognised and generally unrewarded for their
 contribution. As a result, workers develop a sense of deep frustration in their
 working lives, leading to profound demotivation, with little sense of a future. These
 attitudes are important products of the labour process, and have extremely serious
 implications for productivity in the industry.
 The technical division of labour is not the source of these problems; rather they grow
 out of an informal personalistic system of control dominated by white foremen, which
 has two distinct tap roots. Supervisorial domination was the mode of control in the
 factories from their inception in the 1920s through to the 1960s when trade unions
 and industrial relations departments usurped many of tha functions of foremen.
 Supervisors' continuing control over promotion is a legacy of this system, though in a
covert form adapted to the modern system of labour relations. The second root lies
 in the historical practices of job reservation in which whites protected their own
 positions while controlling the pace and form of entry of blacks into the factory
 through manipulation of definitions of skill and job categories.
 Supervisors' informal practices around promotion make use of the dominant modes of
 these two historical traditions: first, the importance of personalistic ties of patronage
 to particular foreman, and second, a deep-seated racism which privileges whites.
 Thus, despite the transfer to the industrial relations system (labour relations, trade
 unions, and the Industrial Council) of powers of hiring and firing, wage increases, and
 discipline, in actual practice, foremen maintain great, though hidden powers which
 allow them to subvert the formal promotion system seemingly monopolized by labour
 relations.
 By admitting the existence of tacit skills amongst automobile workers, this paper
 hopes to lay bare an extremely important area of conflict in the factories. At the
 same time, however, it also aims to expose an extremely valuable asset workers
 possess. Their depth of skills has afforded management with a very rare flexibility in
 deploying manpower, a priceless ability to rotate workers as the volume of work or
 absenteeism demands. Such capacity - virtually unremarked in the literature - is one
 which has been very carefully cultivated by automobile companies in Japan, one
 universally viewed as central to their success in cutting production costs.1 Japanese
 producers have similarly viewed workers' job knowledge as a resource to be cultivated
 in the quest for improved productivity, indeed, managements bestow upon such
 workers higher status, better wages, and greater job security. The view that South
 African automobile workers are essentially cheap, relatively unskilled labour denies
 them their dignity and obliterates abilities they have struggled to develop. But, more
 importantly, the view also ignores a crucial factor which makes production possible at
 all. If developed and rewarded, workers tacit skills could play an extremely important
 role in efforts to improve productivity in the industry.
 Thus far, however, these grievances associated with skills and promotion have not
 resulted in worker actioa In most cases the informal system is taken as an inevitable
 fact of life, impervious to change. Some companies - perhaps with eyes open to the
 true material basis of production in their plants - are already attempting to build
 systems of workplace participation, especially schemes involving workers in fault-
 finding and redesign of jobs.
 But such objective conditions may also serve to promote an intervention by the labour
 movement into the organization of production. In other countries trade unions have
 won effective control over precisely such processes, greatly expanding worker control
 in production. Perhaps the conditions described in this paper could serve as the basis
 for unions to colonise the internal labour markets of firms, and gain control over
 crucial elements of the labour process. In so doing, they might place themselves in a
 powerful position to influence the development of the industry, while at the same
 time providing their members with substantial rewards.
PART I. SKILLS, CONTROL, AND CAREERS AT WORK
 Issues of skill and control are closely linked concerns in any labour process. Workers'
 possession of skill, their knowledge of the job, confers on them some measure of
 control over production. This control obviously becomes a problem under systems of
 private ownership in which the workers are not themselves the owners of the means
 of production. How can a private employer ensure that his goals will be met in the
 time, and at the quality, and cost desired?
 i
 Frederick Winslow Taylor understood that workers' possession of skills stood behind
 the process known as soldiering, workers' conscious efforts to restrict output to protect
 their earnings and conditions of work.2 It was in workers' self-interest for owners to
 remain ignorant of how fast work could actually be done, to preserve their own
 estimate of the value of their work. Taylor's famous management principles aimed to
 overcome soldiering by systematically gathering the knowledge of workers, separating
 conception from execution of work by transferring "brain work" from workers to
 management, and by formulating precise instructions for each worker which define
 the tasks to be done, the procedures to be followed, and the time allowed. These
 principles were meant not simply to routinise work, but, more importantly, to
 transform power relations fundamentally by transferring skills - and therefore the
 control over production - from workers to management
 Taylor's innovations coincided with the experiments by American automobile
 producers at the turn of the century which wpuld have even more dramatic affects on
 workers' skills and control over production. These experiments transformed an in-
 dustry based on specialised production by artisans of small batches of expensive units
 into the quintessential mass production industry. Henry Ford's manufacturing
 breakthrough was the culmination of advances from a range of mass-production
 industries. He combined innovations in fabrication techniques and materials from the
 metal working industry with a simple, well-designed car adaptable to a multitude of
 driving conditions and markets.3 But Ford's major breakthrough was to integrate
 those advances with experiments in two separate areas of production to form the
 continuous production system.
 The first was to master advances in new machinery and transport mechanisms to
 facilitate the continuous flow of materials in assembly. Ford's engineers developed a
 myriad of single-use machines fed by unskilled workers to perform simple, precise,
 and rapid operations. These machines were meant to eliminate all hand work from
 the machining of parts, thereby removing any imprecision in manufacture, and were
 were arranged in the exact sequence of manufacturing operations from the first sub-
 assemblies through to the final product. Experiments in transport mechanisms linked
 these machines together; gravity slides and conveyors moved the job between each
 worker while providing a ready supply of components when and where they were
 needed.
 Ford combined these innovations in production organisation with Taylor's "scientific
 management" techniques for dividing skills and routinizing work.4 The modern as-
 sembly line led to rapid deskilling and fragmentation of work, which permitted greatly
increasing efficiency and accuracy of production while vastly reducing the dependence
 on skilled labour and workers' ability to resist the intensified pace of work.3
 The image of assembly work as fundamentally low skilled activity derives from these
 interlinked legacies of Taylor and Ford. Oddly enough, their descriptions parallel the
 Marxist tradition which emphasises the increasing desldlling and growing alienation of
 workers as an inevitable consequence of the growth of capitalism.6 Managers are
 portrayed as omnipotent, and their best-case plans are taken as reality, while workers'
 own roles are ignored, particularly their resistance to managerial initiatives.
 Jean Leger's recent thesis on South African mining contains a lengthy examination of
 another approach to workers' skills. In elaborating the concept of tacit skills, Leger
 argues that the labour process is always a dual process of conflict and cooperation, in
 which management never gains absolute control over production.7 In Leger*s
 discussion skill is not seen narrowly as the product of some educational or training
 programme. Rather, all human activity entails the exercise of some skill. He
 identifies two important aspects of "tacit skills." First, what are often assumed to be
 routine tasks may involve conscious and unconscious processes which are learned
 experientially; they cannot be articulated in formal statements, nor learned through
 detailed instructions. Such an approach may help to define the skills present in a
 familiar task like riding a bicycle, or the more specialised tasks associated with "the
 tricks of the trade." Second, social skills must also be taken seriously as essential
 components of any labour process: cooperation, trustworthiness, congeniality, and
 obedience.
 These skills are described as "tacit" for the simple reason that they are not formally
 recognised nor are they rewarded. Powerful groups may have interests in denying
 their existence, while those in possession of such skills may be powerless to articulate
 their interests. Rather than seeing skills as some objective datum, Leger prefers to
 view them as social constructs. The social construction of skill results from social
 processes which involve power relations between individuals and groups in which
 certain occupations are defined as skilled or unskilled.8
 This approach does not hold that all skills are equal, rather that the common
 tripartite distinction between skilled, semi-skilled, and unskilled "grossly underrate^]
 the working knowledge required and the skills exercised in these occupations....there
 is no such thing as unskilled work and...the term is humiliating to the workers so
 labelled."9 Craftsmen learn their skills through apprenticeship programs, and through
 on-the-job experience where they master the art of applying abstract principles to
 specific problems. Their skills are general, and may be applied in diverse situations.
 Their skills are formally recognised and accredited. They carry their skills with them.
 Tacit skills, however, usually take the form of "plant specific" skills, in which workers
 acquire a store of very detailed knowledge of the particular processes in their factory
 or industry. These skills are seldom developed in formal training programs, but can
 only be learned on the job through watching other workers or through informal
 education conducted by factory elders. The fact that workers are taught "by doing,"
 rather than in formal training is not an index of the lack of skills associated with the
work, or of the value of the skills to overall production, but of the particular kinds of
 skills.
 Such skills may be especially well-developed in older factories where machines
 "develop quirks that only their operators can anticipate and control."10 They may
 not be recognised in any formal sense, but they are essential to the smooth operation
 of the plant; they therefore also possess power to disrupt production. At the same
 time, workers with plant-specific skills are extremely vulnerable to plant closure or
 relocation: they cannot transfer their accumulated experience into a certificate which
 allows them to compete for another job. They may come to share with management
 a commitment to maintaining the long-term viability of their enterprise. Leger*s
 approach indicates the importance of seeing the labour process as an affair in which
 workers are more than pawns in management's schemes. If workers are active
 participants in production, their positions are in part the result of their own actions
 and designs.
 The American sociologist Charles Sabel has produced a concept which helps to
 capture workers' own involvement in production. The "career" refers to the series of
 tasks a person chooses "that successively challenge and require the development of
 whatever powers one takes as the measure of human worth." Different work groups
 have different ideas of success; "they differ about which powers define dignity, which
 jobs count as disgraces and which as accomplishments... Sabel asserts that the
 worker's world can be understood as a whole, in which a range of experiences and
 ambitions "combine according to stylistic canons that the worker recognizes as his
 own." Insofar as these canons shape a worker's responses to events, his behavior
 cannot be understood without taking them into account
 Taken together, the two concepts of tacit skills and careers at work will help to
 provide a different understanding of the labour process in assembly line work in
 South Africa, and simultaneously, to grasp the deep frustration workers encounter
 when their skills go unrecognized and unrewarded and their career aspirations
 unfulfilled.
 PART II. THE FORDIST FALLACY AND THE SOUTH AFRICAN AUTOMOBILE
 INDUSTRY
 The South African industry dates from 1924, when Ford opened an assembly plant in
 Port Elizabeth. General Motors followed its American competitor to Port Elizabeth
 in 1926, and the two companies dominated the industry until the post-war period
 when European competitors entered the South African market12 In the 1960s
 Japanese companies opened plants, and in a decade began to dominate the market.
 But the decisive transformation in the industry occurred in the early 1960s, when the
 government passed import-substitution legislation forcing multinationals to source an
 increasing percentage of their components within South Africa. Previously the
 industry had been organised on an assembly basis only, relying on parts imported
 from overseas in a disassembled form.13
The program quickly led to the expansion of the industry, as a number of new foreign
 producers rushed to establish assembly plants. Assemblers could easily oatct A t first
 stage of local content through inclusion of components already produced in South
 Africa. Subsequent stages, however, would be impossible to meet without further
 investment in manufacturing operations by the assembly firms and components
 suppliers.14
 The 1960s expansion had a dramatic impact on employment Overall employment
 grew by 437%, increasing from 8,918 in 1961 to 39,011 in 1975, an annual rate of just
 over 29%. As significant as the overall increase, however, was the change in the
 racial composition. Where whites accounted for just over half the workforce in 1961
 (52.1%), coloureds 27.6%, and Africans 203%, the figures virtually reversed fifteen
 years later, with Africans amounting to 44%, whites 34%, and coloureds 22%.w
 Not only had the automobile industry developed a majority black workforce, but
 blacks were increasingly clustered in the main semi-skilled occupations at the heart of
 the assembly and manufacturing operations. The changing size of the workforce,
 racial composition, and the advance of blacks to higher skill levels should not obscure
 an important qualitative change in the character of employment These workers were
 no longer employed solely in assembly work, but in a wide range of occupations
 within the new manufacturing facilities in each plant Whole new workforces were
 needed at those firms opening up engine plants and press shops, with skills never
 before required in the assembly industry: skilled workers such as tool and die makers,
 machinists, and electricians; semi-skilled die setters, sheet metal workers, motor
 mechanics; a whole range of semi-skilled operators of lathes, drill presses, grinding
 machines, and presses; and new categories of quality control technicians to monitor
 the manufacturing output
 The government's reluctance to intervene to limit the market meant that growth was
 characterized by extreme fragmentation. Virtually every MNC automobile producer
 entered, and soon South Africa reproduced in miniature the competitive dynamics of
 the international market16 Car companies competed through the demand-side
 practices of product differentiation used in the European and American markets,
 resulting in the proliferation of models and variants, fragmenting still further the
 already competitive market The latest models were provided, promising frequent
 turnover and high retooling costs, and the motor firms were routinely the most
 prolific advertisers among South African corporations.
 What possibilities as may have existed for internal economies of scale were lost, and
 the companies resorted to battles for the relatively well-off, upper segment of the
 market, which meant white consumers, corporate fleet purchasers, and the
 government These conditions had a dramatic impact on the labor process, placing a
 premium on appearance - which rendered the system sensitive to sabotage and other
 forms of resistance - and the multiplication of models created high job content for
 workers required to assemble a number of different models and variants on the same
 line.
The transition to manufacturing had a dramatic on the nature of the labour process in
 the industry. Yet, the available accounts have often misunderstood the character of
 work in the automobile plants, as the studies have been based on a logical fallacy,
 that firms could be treated as a black box in which the internal labour processes
 could simply be deduced from the conditions of mass production in the home
 country.17 This is particularly true in the structuralist Marxist tradition of analysis of
 South African industrialisation. In this approach, the motor plants have been singled
 out as pioneers in the transition from manufacture to machinofacture in South Africa,
 the latter being a profound change in relations of production in which handicraft
 processes are replaced by machines tended by semi-skilled operatives, marking the
 "real subsumption" of labor to capital.
 Rob Davies has applied such a perspective in explaining the origins of the "civilised
 labour policy" of the Pact Government General Motors, according to Davies, was
 "perhaps the most mechanised single plant in South Africa during the Pact period,"
 and it was in the company's interest to employ relatively well-paid whites, because
 they had internalised appropriate capitalist virtues which suited them to such
 mechanised work."18 Leaving aside theoretical criticisms of the concepts
 machinofacture and subsumption, the depiction of the labor process and workforce at
 both GM and Ford is entirely inaccurate for the years 1924 to World War Two. The
 first automated assembly lines were installed only in 1948, and it is impossible to
 accept Davies' claims about machinofacture prior to that date. To compound the
 problem, the period when more recognizably "Fordist" techniques were adopted -
 from the 1950s onward - coincides with the erosion, rather than the strengthening of
 the white monopoly on employment
 Yet, even more sober and empirically attentive analysts fall victim to the deductive
 fallacy. Roger SouthalTs otherwise excellent assessment of the rise of unions in the
 automobile industry nonetheless attempts to read off the South African labour
 process from Robert Blauner's analysis of alienation and the assembly line in the
 United States.19 Similarly, the work of researchers at the University of Port
 Elizabeth, under the direction of the industrial psychologist Roux van der Merwe,
 evaluated the skill levels of jobs in nine manufacturing sectors in Port Elizabeth,
 including the motor industry. They too, appear to take the division of labour at face
 value in their assessment of workers' jobs, and conclude that median job complexity
 for auto workers was below the average for all the sectors analysed in the survey.20
 The same study was uncritically drawn upon in an otherwise superb study by Jan
 Roux who concludes that auto workers have low skill levels, and have been "stripped
 of any form of control, both over the ordering of tasks, and over the speed at which
 they had to be executed."21
 The shortcoming of each of these presentations is that in decontextualising their
 studies of the labour process, they ignore a crucial feature of the South African
 industry: it's low volume character. It is rather surprising that such an obvious feature
 of the industry is not taken into account. Most of the same authors readily admit the
 effect low volumes have on profitability: producers are forced to spread their
 investments over a limited output; economies of scale are impossible, unit costs are
high, and producers are thereby forced into raising prices in a highly competitive
 market High prices further restrict demand, and so the vicious cycle begins again.
 However it is posited here that these same conditions have a decisive impact on the
 labour process, which fundamentally differentiates South African production from the
 processes in metropolitan countries.22 First, the level of output from any South
 African plant is but a fraction of the output from a plant in the industrialised
 countries. In addition, South African producers, oriented as they are to the white
 market and with no hope of economies of scale, do not compete by cutting prices, but
 by differentiating their product through a range of makes and models, offering a wide
 array of options.
 Workers on the line therefore operate in a very different environment to workers in
 the United States, Europe, or Japan. An assembly line in such plants may be
 producing as many vehicles in an hour as a South African line produces in a day;
 workers on the line nave far less time to perform their tasks, and are of necessity
 given fewer tasks to perform. According to a retired Ford (South Africa) Production
 Director,
 In a North American line you'd probably get 60 [vehicles] an hour, he's
 probably loaded to about .9 of a minute, he only gets time to put on the
 front door handle and the rear door handle. Where here a chap is in a
 plant producing a vehicle every two minutes. So his work load changes
 quite significantly.
 The South African worker may be responsible for a number of operations which in
 America would be performed by a number of workers.
 The proliferation of makes and models complicates the picture. An American plant
 might be responsible for producing a single make of car, and workers on line would
 face the same model - with the same operations required - over and over again
 throughout the day. On a South African line, a worker may encounter a number of
 different makes, each with an assortment of models (different engine sizes, two door,
 four door, etc.). The condition means that the South African worker must be able to
 identify the part appropriate to a particular car. According to the Ford Director,
 See, basically what determines work content is the number of parts you
 handle. If you take a line in the US plants...you probably find that the
 basic vehicle has got something in the neighborhood of about 5,000 line
 items. There might be a couple of thousand or so of options. So the
 labour is handling possibly between 6,000 and 7,000 line items. A line
 item count in a plant like Volkswagen or Ford Motor Company here or
 Toyota will be 17,000 to 18,000....So you've got fewer number of
 operators assembling those to the product, it stands to reason they are
 going to have to know more about the product than people at a higher
 rate of production.
 8
In sharp contrast to the repetitive monotony of overseas assembly lines, in South
 Africa the greater job content means workers are performing a variety of tasks in a
 constantly changing order
 We were building 150 units a day with a mixture....The repetitiveness of
 work was alarmingly low. As a matter of fact, in many instances
 operators were only repeating the exact same cycle eight times a day,
 although they were operating on 150 units a day. Whereas in the states
 it could be 60 times an hour.
 Workers must master the operations necessary to fit the appropriate part in the
 proper manner within the time allotted, or else he will find himself "drifting out of
 station," or moving down and interfering with the worker below him on the line. For
 this to be done successfully, the worker must learn a variety of tacit skills, usually
 from more experienced colleagues on the line:
 Q: Did you get any help from the other workers?
 A: Yes.
 Q: What did they do?
 A: Sometimes they helped me. Sometimes they showed me the "how to,"
 the art of doing i t
 Q: "The art" What do you mean?
 A: You see, say you fit a tail light, and you are working in a line that is
 moving. Someone will tell you, "This thing has got four screws. Put the
 light in and fit the washers, the bolts."
 Later the same worker expanded on how he learned the procedures:
 Or maybe, if you do something like as I have been fitting in there, this
 window winder, you may fiddle in putting it in, not knowing the exact
 way....Say that the thing is laid this way, you are checking it another
 way, and sort of holding it on the side, so someone will tell you, "No,
 put it this way and in." I mean, the art
 But fitting parts to a body on a moving line presents its own difficulties, which the
 worker must learn to master. Here a worker describes the process of "working up the
 line," whereby he places the part in a number of vehicles at once, and then fits each
 one, working back to his starting point:
 Someone will tell you, "Before you do that, take six lights [walk up the
 line] and put in the different cars and come back with them." You see,
 the art, the way of doing it is easier for you. More than taking it from
 [the bin] and putting it in the car. The car is here at that time. Now
 you left it there...you have to come back [to the bin] again and take a
 light, and the car is going forward, see. Now by that time you are
 drifting out of station.
Furthermore, a South African plant will integrate a for wider range of production
 activities than first world plants. Organizationally, an assembly plant in the United
 States or Europe will only assembly cars, and will be fed from a separate engine plant
 and a separate body plant, which may even be located in different regions and in
 separate divisions of the corporation. In South Africa low volumes mean there are
 not sufficient economies of scale to warrant the establishment of separate plants.
 Most companies conduct all or most of these operations under one roof, so that the
 workforce will contain an extremely wide range of occupations.
 Most workers in assembly and basic metalworldng operations learned the work from
 other workers. Rising up the graded job hierarchy, workers in intermediate
 occupations requiring formal certification, such as metal finishing acquired their skills
 in the Volkswagen training centre, or in private study which enabled them to pass
 national trade tests.23 Still higher up the hierarchy, Volkswagen trains a wide array
 of artisans through formal apprenticeships, and a die-setter, two electricians, and a
 toolmaker were among those interviewed who had acquired their skills through such
 in-house programs.
 Workers lower down in the hierarchy were often indeed performing "simple" tasks,
 such as fitting components on the assembly line or spot welding in the Body Shop.
 Nonetheless, many had acquired multiple skills to the point where they claimed the
 ability to perform all the jobs on their line or in their department Many had
 developed such skills through their own initiative, by watching other workers. Some
 had done so by being temporarily reassigned by the foreman to another job when a
 worker was absent. One worker explained the source of his "training" as follows:
 I learned these other jobs only by helping other people. Monday, then
 the man who is doing the job is not here, and then they take somebody
 from our line and then I must learn how to do his job....The other jobs
 there, nobody ever taught me how to do them, no, I have just learned
 them by seeing.
 And on the same lines, another worker said:
 I'm only responsible for fitment Door fitment, bonnet fitment, fender
 fitment, and tailgate fitment That's what I'm doing here. I'm fitting
 the bonnets. If for instance, the guy who is fitting the fenders, he's not
 at work, I have to do that, as well.
 One operator in the Press Shop (at the bottom of the pay grades), reports that he can
 do the work of the die-setter (at the top of the grades):
 A: If the setter is not there, I handle it,I can control the department,
 because the foreman is coming to ask us, "Do you know this part?" I
 say, "Yes." "How do you do this part?" I say, "All right, I show you how
 to set this machine for that part. How to make it right."
 Q: How did you learn that?
 10
A: Since I have a long time doing that job, no one has ever learned me
 how to do it, but I only see, on this job now, they did like this, they did
 like this.
 Q: Did that foreman that you described, did he show you those things?
 A: No, he didn't show me. He only worked in front of me. I am just
 keeping an eye when they are working.
 These allrounder skills possessed by line operators provide management with an
 extremely valuable asset: workers can be rotated to different posts as the volume of
 work or absenteeism demands. There is no expensive retraining required, and no
 interference from cumbersome work-rules, as stipulated in union contracts in the
 United States or Britain. Such manpower flexibility certainly enables management to
 cut costs, by relieving it of the need to retain temporary workers or a larger-than-
 necessary staff.
 From the workers' point of view however, such aUrounder status does not prove to be
 such an asset. While it is possible to achieve a formal allrounder status, known as
 utility man, and to jump a pay grade as a result, many workers claimed to possess
 these skills without being recognized. Not only don't they receive a full grade
 increase as a result of their skills, but in fact they are frustrated that they often
 perform the work of higher graded workers, without receiving even a temporary
 increase. One worker commented as follows:
 Well, I'm easy to go there, but some other, they have a question. "Why
 must I go and do that job? That is Grade 5 and I'm Grade 2, and I'm
 not supposed to be there. I'm supposed to help the same grade as I."
 Another worker stressed the same theme:
 They promise such things. Once they say, "If there are more than about
 five jobs in the same line, they will give you something. But they ask
 how many jobs you know, you know seven jobs, eight jobs, then you get
 nothing. Now the only thing they ask, as far as I am concerned, is when
 there is anybody who is absent, they just check on that note that
 somebody can do that job,and they take you and put you there. You do
 different type of jobs during the week, although you get nothing for
 that.
 Despite developing multiple skills, and in some cases, certificated skills, workers were
 divided on the marketability of their abilities. Those who had obtained nationally
 recognised qualifications felt secure in their ability to find work elsewhere, and felt
 they had learned skills at Volkswagen which would be of use to them in the future.
 At the other extremes, some workers who had not obtained formal or informal
 allrounder status felt they had attained very little. A Press Shop operator
 complained, "No, no, no, no, you just take your small child and come and operate
 this." Another, who installed the electrical harnesses in the vehicles, complained that
 he did not know other operations in the car, and described his job as "an empty
 bucket, it's not enough information." Yet others, especially allrounders in the
 11
assembly lines, and especially metal finishers and painters felt that their skills could
 provide them with a livelihood in the township, repairing cars. Whether or not
 management fonnally recognises such workers as possessing skills, a majority of those
 interviewed felt they had acquired abilities which would be of use to them in the
 future.
 PART III: THE RISE OF INDUSTRIAL RELATIONS
 Prior to the 1970s, foremen possessed vast discretionary powers, including the power
 to grant incremental wage increases for workers.24 But the foreman's disciplinary
 arsenal did not end with wage setting. He also enjoyed the ultimate discretionary
 authority: the power to hire and fire. The contractual basis for hourly-rated
 production workers was the so-called "hourly basis," which meant simply that they
 could be fired on an hour's notice. There were no formal procedures for appeal.
 One retired General Foreman from Ford described the dismissal process in the
 following way:
 A: A fellow will come to you in the morning and I'll say, "Well, George,
 where were you yesterday? Where were you the day before?"....Now
 just about two or three days after that, he's out of work again. Well he
 says, "My wife is not well. She's just had a baby"...There's a weak link.
 That fellow, you've got to get rid of. Get rid of him straight away,
 because he's upsetting these fellows around this area. You see, when
 you're working with a team of people, and you've got one bad egg,
 there's encouragement somewhere....So you've got to tie that down.
 And the only way to tie that down is to get rid of him.
 Q: And the decision making would be entirely up to the foreman?
 A: Up to the foreman, yes.
 Q: And if he had a dispute, would there be anywhere for him to go?
 A: No, no, no, you just give it to him. You know, you just get your
 General Foreman, you superintendent, and you tell him, "Okay, I fired
 this fellow and this is the reason why." And then I'll just take in
 another man.
 Ollie Rademeyer, former Industrial Relations Director for Volkswagen reinforced the
 story by relating an anecdote from General Motors:
 You've probably heard it. The biggest optimist in the work force is the
 guy who brought his sandwiches to work, because he had no assurance
 that he'd still be there at lunch time.
 A Volkswagen worker relates the same view from the receiving end:
 Ag! We sort of were ill-treated by the white people here. Just like if,
 say, you are a foreman in my department, if you don't like me, you have
 to just chase me very, very easily away, drive me away, out of the
 12
factory....The guy who was responsible for that section used to say,
 "Well, I don't like this man. He must go." And then he go, whether he
 likes it or not
 These most blatant supervisorial abuses began to be curtailed in the 1960s, when a
 union began organising white workers. At first the firms replied as they always had,
 with an iron fist They had successfully resisted "interference" from unionising drives
 for over 30 years "with typical American methods," according to W.G. Ballinger, one
 of many who had attempted organise workers. In effect, he wrote, "this means that
 attempts to organise the workers breaks on the rocks of victimisation."25
 In the 1960s, however, white workers used political channels and made a successful
 appeal to the government for protection. The Minister of Labour launched an
 Industrial Tribunal investigation of the industry for "unfair labour competition"
 between whites and blacks, which resulted in Job Reservation Determination No. 16
 of 1964. In essence, all jobs then occupied by whites were permanently reserved for
 whites; in addition, a number of restrictions were placed on the sorts of jobs blacks
 could occupy. For example, no black person could be employed welding jobs outside
 of a jig. The Determination would have been disastrous for the companies - if it
 were enforced - but the government offered exemptions from the regulation. As
 Rademeyer relates:
 So we then entered the period of the glorious exemptions. I've never in
 my life seen such a farce as that thing....Many jobs that could have been
 performed...by whites because they were out of jigs, we put into jigs. I
 mean, a jig is a jig. You can take a job that really can be easily
 performed outside of a jig and just build an angle-iron frame around it
 and say, "that is a jig. It can now be performed by a non-white"....if
 there was a difficult job where they couldn't get whites and they had
 non-whites, they'd just put a jig around it...Then [the government
 inspectors] would say, Yes, but it wasn't done in a jig before." We'd
 say, "Ja, but you see this model, the 1962, is more technical."
 But the determination achieved its aim: to restrict the companies' ability to
 introduction black workers into the factory. In essence, no black could be employed
 if there were whites available to perform the work. Rademeyer is quite explicit about
 the bind the companies were in:
 The companies played the game to a certain extent in that any genuine
 white worker who presented himself at your gate and he hadn't been to
 prison for six or seven months in the last two years, you said, "Okay,
 give him a try. He's white." Any bloke they referred to you, "Give him
 a try." So that you didn't actively kick whites out.
 Job Reservation would have a long-term impact on the company. More than ten
 years after implementation of the policy, an audit of the application of the "Sullivan
 Principles" at Ford discovered that 45% of white workers interviewed did not have
 Standard 8, though all were employed in the highest wage grades; none had entered
 13
the company below the middle grades. For black workers, on the other hand, a
 Standard 8 certificate was a necessary qualification for advancement to the higher
 grades, and very few were employed there.26 Indeed, more than twenty years after
 implementation of the determination, black workers were still complaining about the
 incompetence and lack of qualifications of white workers and supervisors who were
 nonetheless senior to them.
 In essence the companies exchanged recognition of the white union for exemptions on
 job reservation, and in 1969 an Industrial Council was established for the industry. A
 union for coloured workers began organising in 1967, and was admitted to the IC in
 1971, and it soon began organising among African workers in the Eastern Cape
 plants.
 The development of unions is paralleled by the growth in the industrial relations
 bureaucracy at each of the companies. Formerly personnel and human resources
 departments had been extremely minor divisions of each firm, handling pensions,
 benefits, and the security guards, but from the late 1960s they became rapidly
 developing power centres. The staff expanded along with the general growth in
 employment, but also to keep pace with new functions, especially those related to
 regulating labour relations on the shop floor. Functions which were once the
 preserve of foremen were transferred to the new bureaucracy, and subjected to formal
 rules and procedures. The former Industrial Relations Manager at Ford commented:
 We've taken it away from the [foreman] for the sake of good labour
 relationships. I mean for no other damn reason To ensure that if you
 get two people breaking the same rule,that the punishment is going to
 be the same, given the same circumstances.
 The former Production Director at Ford echoed the same view:
 I'm fully of the opinion that a lot of the need for a union was as a result
 of foreman malpractice....It was just a case of saying to the foreman:
 "Look, you follow this set of rules so that anything that you do with your
 labour is consistent with what other foremen are doing."
 Hiring and firing was now formally performed by Labour Relations, while wage rates
 were the subject of negotiations between employers and unions in the Industrial
 Council. Discipline, though meted out by line supervision, was increasingly subjected
 to review and appeal through company-union grievance procedures, conducted
 between shop stewards, labour relations officers, and foremen. In short, the formal
 discretionary power of foremen was substantially reduced by the development of the
 internal and Industrial Council industrial relations system.
 One foreman in the Volkswagen Body Shop, a particularly tough, big, barrel-chested
 man, recalled the time when supervisors wielded such power:
 Those were nice days! When we had control of the [pay] increments,
 we had control over the workers. You can kill his wife or his children,
 14
but don't take his money away. Now there's no incentive for workers to
 work harder.
 In the late '60s, early 70s, if you don't like a guy's face you could fire
 him....Now you've got to get your facts straight
 The former Industrial Relations Manager at Ford substantiated the view:
 [The foreman] could no longer give an employee a kicker. He's got to
 check with somebody else before he could kicker an employee. Do you
 know what a lacker is? A kick in the assl
 Indeed, the institution of a regularised, rule-bound industrial relations system is one
 of the factors most often quoted by workers at Volkswagen in accounting for what
 they most like about the company. They are protected in a variety of ways by the
 union, particularly from unfair dismissal and victimisation by supervision. One
 Volkswagen worker eloquently summed up the change:
 We used to say, "If the master say or do that, you used to go and run to
 do that without questioning why." But now we've got why now. They
 don't like that why....I mean, because we used to go and do whatsoever.
 If he said, "Go there and do that and whatsoever," now we haven't got
 the right, before '80 to ask him, "Why?" He must explain why he send
 me there, because Fm working here.
 Mailer's recent research on Volkswagen uncovered the same response. When asked
 why they joined the union, over 80% of black workers cited the need to defend
 workers' rights, including protections against arbitrary discipline or dismissal.27
 PART IV. INFORMAL AUTHORITY: THE FOREMAN'S REVENGE?
 Despite the development of formal systems for regulating work noted above, foremen
 nonetheless continue to possess extremely important power over the most basic
 factory processes, especially promotion. It is asserted by virtually all the interviewees
 that foremen continue to posses power over wage increments made in the time
 between across-the-board increases mandated by the Industrial Council. Workers,
 especially those in higher grades, claim these increments account for a none-too-
 frequent difference of fifty cents or one rand in hourly pay rates between white and
 black workers performing the same tasks. In addition, foremen seem to retain
 considerable power over the distribution of overtime work, which interviewees claim
 is parcelled out on a racial basis. During periods of short time such unequal
 treatment can have a very immediate effect on workers' pay packets.
 Within the official system, job grading and promotion are subject to the strict controls
 of the labour relations system. According to the formal system, job vacancies are to
 be filled by the Personnel Department after notification of any vacancies by
 supervision. The Personnel Department advertises posts on in-company bulletin
 15
boards, conducts examinations of applicants, and selects appropriate candidates based
 on the results of the examination.
 In practice, however, the system almost never worked in the formally described
 manner. Instead, virtually all workers interviewed described an informal system in
 which foremen still wield ultimate control over who gets promoted. When asked to
 describe the procedures relating to promotion, only one worker said that the actual
 system in fact corresponds to the formal rules. Those who responded to the question
 (18 workers) claimed the real promotion procedure operated through favoritism on
 the part of foremen, racism, or identified a complete mismatch between education
 and promotion. Many expressed indignation at the promotion of less qualified whites
 ahead of higher qualified blacks.
 Well, say, for instance, there's a job open for an Assistant Foreman.
 What they will do, if there is a white guy they know that would also be
 in a position for that job, and myself also make application for that job,
 they prefer that white guy to go into that job before they consider me.
 And, in a lot of things, you will find that they will mostly push the white
 guy to go into certain jobs before they push any black guys into that job.
 Another worker commented:
 How do you like...being taken to be an operator in the lines here, while
 the very supervisor, or the superintendent, or even your manager is less
 educated than yourself? How would you like that? That's something
 wrong, man. That's something wrong, really.
 And another:
 We did sit last time with a foreman that can't even make out a time
 sheet He asked somebody from the people working under him to do
 his time sheets. I mean that's unfair, in my opinion.
 Racism was one frequently stated complaint, but others complained equally bitterly
 about favoritism, in which foremen promote workers - African or coloured - who are
 tied to them. One worker commented that promotion is not a case of what you
 know, but of "Who do you know?" The informal procedure is to approach one's
 foreman to ask for promotion. The system seems to be wired in such a way that
 foremen can obtain advancement for select individuals. According to the procedures
 (and comments by foremen themselves), they should not have such power, but actual
 practice is different:
 I found out when inside of Volkswagen, when you apply for a job here
 you are wasting your time. That is number one...You just couldn't apply
 for a job inside. You just have to go to the foreman or supervisor
 concerned to tell him that you want that job....You just don't have to
 apply for the job inside. If you apply you are wasting your time. You
 can apply, and you are being called for an interview. They interview
 16
you and they say, "Well call you back." You will never be called back
 again.
 Q: How do you find out about a job?
 A: You nave to move around to the people concerned. Tell them you
 want the job...you want to work under his belt, you want to work under
 him.
 Another worker repeated a similar story, but gave more details of the actual hidden
 mechanisms of supervisorial control over promotion:
 I heard about the job. It wasn't on the board. I heard about it,
 somebody told me about the job. So I went up straight to the foremen,
 asked him where it was, like that, if there was a vacancy and he told
 me, That you must first bring in your matric, the only requirement is
 that you must have some experience of repairman job...
 Another worker extended a discussion of these mechanisms, and provided this most
 telling story of the way he received a promotion from the lowest pay grade to die-
 setting, one of the highest graded and apprenticed jobs:
 I applied for die setting. Actually, my general foreman...he told me
 about the vacancies but the notices wasn't on the board yet. So I had to
 see this guy ???*?.
 The foreman had informed the worker of an extremely important opening, even
 before it was relayed to Personnel, and referred him to the foreman in the
 department concerned:
 I went straight to his office and I told him about this vacancies that will
 occur in the Press Shop. And he asked me, who told me about this, and
 I said, no, **** [his general foreman]. He said, It's confidential. It's
 only between me and him." And he said, "Okay, fill in a form, and
 bring your certificate tomorrow."
 The worker retrieved his Standard 8 certificate and returned to the foreman:
 "Okay, you are in for the thing." So it's about three days after that I
 started training die setter, and I've qualified last week.
 Through personalistic ties to foreman, this particular worker made a leap over 4 wage
 grades into the artisanal "elite." He now receives more than twice his original wage,
 and possesses a certificated skill he can take with him anywhere in the country.
 Finally, one worker captured the informal - and hidden - character of such processes
 with a thoroughly apposite piece of Afrikaans vernacular:
 I did make application for a foreman job, and through the time that I
 was waiting for the application, somebody told me long before it was
 out, "[Another] man is gonna get the job." So what we are calling in
 17
Afrikaans is smokkel. It's not a man that has got the capabilities and
 the man that has got the qualifications that do get a job, it's a man that
 has a link with a supervisor or with a manager that do get a job and
 that's unfair.
 What is the basis of such personal ties?
 Many workers claim that race is the fundamental divide, with whites being favoured
 above all. But the worker described above is coloured, and African workers have also
 benefitted from such relationships on the basis of possessing (or being seen to
 possess) appropriate attitudes:
 You know, the situation here is politicised, racialised. The relations
 between workers and the bosses are also regulated by racial
 considerations, political considerations, you know, attitudes that can
 affect your progress, if you are deemed or are seen as somebody who
 have your own mind. Then you suffer in this set-up. You suffer in this
 set-up. You must just be moderate. And then you get the right chances
 for moving up....If you have definite political inclinations, chances of
 advancing are minimal. It is like that.
 One Grade 7 worker with a matric, explained with contempt the approach which
 would likely bring success:
 So if we go as far as assistant foreman, and if you are lucky enough, and
 then you have got to be a friend of the boss, buy him pies and a cool
 drink, and let him shout at you like anything, let him say disgusting
 things about you, and you don't say anything against him, you just smile
 and scratch your head all the time, and then you'll probably be a
 foreman.
 The general feeling amongst workers was that they could do little to change the
 system. Some saw the practice as inherent in the racial system of South Africa, and
 nothing short of total political transformation would ever alter it. One man summed
 up the situation by saying it made him utterly demotivated. Another protested, "I
 don't even bother to apply any more. I just want to quit."
 Thus, despite the formal - indeed the quite substantial - deracialisation of the
 workplace and the institution of a representative system of labour relations,
 fundamental processes on the shop floor remain part of an informal system of
 personal control by white supervisors. While workers may no longer be dismissed or
 denied wage increases by despotic supervision, their futures with the company, their
 expectations of a career at work, remain subject to their foreman's discretion.
 These informal practices seem to violate fairly deeply held values of justice, especially
 as related to equality of opportunity. One might have assumed that black workers
 would come to expect such treatment as a norm, and many were extremely clear in
 explaining the informal promotion system as a normal part of South African life: as in
 18
'that's just the way things are.' But such recognition doesn't seem to have blunted the
 sense of morality. I f they can upgrade me," one worker said when asked about his
 plans for the future, "if they can send me to the training school, I can go there. Then
 if I failed, I won't blame them, FU blame myself." He seemed to suggest that he
 could adapt to individual failure, but could not accept conditions in which he did not
 even get the opportunity to test himself. The informal promotion system therefore
 corrupts the relationship between, on the one hand, skill, education, and desire, and,
 on the other hand, the chances for promotion. The system plays havoc with workers'
 expectations and calculations about the future, rendering their "career at work" into
 little more than the plaything of their foreman's caprice.
 We have to meet that particular foreman of that particular job that you
 want to do. Then you speak to him. If he likes you, then he will accept
 you. If he doesn't, then it is bad luck.
 The absence of clearly defined job ladders, which would at least allow workers some
 expectation of a career at work, could possibly have contributed to the view of most
 of the workers, who could not envision themselves working at VW in five or ten
 years' time.28 Only a small number identified working at Volkswagen as a
 possibility. Others expected to be in their own business, unemployed, seeking further
 education, or dead! A large number didn't know what they would be doing. One
 worker summed up the pessimistic mood: 1 don't think I'll be here in 10 years, if god
 spares me!"
 PART VI. THE CONSEQUENCES OF INFORMAL AUTHORITY
 The pessimism voiced by workers is an unfortunate consequence of the organisation
 of control in the labour process in the industry, rather than resignation with factory or
 assembly line work per se. As described above, contrary to popular assumptions,
 workers in the industry possess an array of skills which are certainly useful to the
 company, and to a number of the workers themselves. Given the ability to translate
 their tacit skills into a career at work, such skills could also provide workers with a
 basis for personal development and self-fulfillment. By not recognizing the
 importance of tacit skills, both the automobile companies and the union are
 tarnishing an extremely valuable asset through tolerance of a system of informal
 despotism.
 Labour relations systems in other countries, most notably the German system of co-
 determination have institutionalised workers1 control over the very processes which in
 Volkswagen are controlled by foreman. Manpower training and promotion, in short
 the processes central to establishing careers at work, are part and parcel of the legal
 system of co-determination, and unions have entrenched rights to participation in
 manpower policy-making. As a result, trade unions are able to achieve employment
 stability and extensive training for their members. Wolfgang Streeck makes the
 argument that such deepening of skills has provided German producers with a highly
 qualified work force, and argues that such capacity is an important part of the
 foundation for the German producers' extremely successful strategy of building and
 19
marketing extremely reliable high-quality vehicles.* In short, Streeck notes a very
 close connection between union participation in training and promotion, and
 economic success.
 Some companies in South Africa, notably Nissan, have seemingly come to a similar
 conclusion. They have developed policies adapted from Japan which tap into workers
 tacit skills, and while providing higher status for participating, though not yet financial
 rewards. Regular communication groups, the so-called "Green Areas" where workers
 meet to discuss each day's work, are attempts to marshall workers' tacit skills to
 improve production. Volkswagen, too, is experimenting with such policies, and has
 promoted "ikayalam" ["my house" in Xhosa] where groups of ten "workers gather in
 the foreman's area at the start of each working day and 'specify the work content for
 the day and decide who does what and when.1"30 According to Mailer, the initiative
 aims "to coordinate common interests on the shop floor by drawing workers into co-
 operative, non-union structures in production."
 It is not clear, either at Nissan or Volkswagen, whether such structures will offer
 workers any substantial rewards, most especially a "career at work." No automobile
 company has yet come forward offering guarantees of job security or lifetime
 employment characteristic of the German and Swedish co-determination systems or
 Japan. It is clear, however, that South African unions have not responded to these
 management initiatives, except to reject them. But perhaps these responses are short-
 sighted. While many of these programmes - including the one at Volkswagen -
 appear to be schemes for avoiding unions and replacing them with company-
 controlled structures, the material basis which permits the experiment may also be
 claimed by unions. Tacit skills are at the core of these participation schemes, and
 unions may be able to develop their own schemes to develop such skills. If they can
 colonise skill development - perhaps in a variation of the German model - they will
 not only be making an effective inroad into central managerial prerogatives, but they
 will also be able to provide substantial material rewards to their members.
 20
ENDNOTES
 Methodological note: the interview material reproduced in this paper formed part of
 the research for my doctoral dissertation. 40 Volkswagen workers were organised in
 a cluster sample covering all sections of the factory, stratified according to skill, age,
 and wage grade. The interviews were conducted using standardised, open-ended
 questions, to allow scope for comparison of results while allowing respondents room
 to expand on their ideas.
 1. In Sweden and Germany the systems of co-determination which spawned such
 flexibility were forced on employers by powerful labour movements (in Germany) and
 by a coalition of labour and Social Democratic governments (in Sweden).
 2. For an excellent review of Taylorism, see Jean-Patrick Leger, "Talking Rocks' - An
 Investigation of the Pit Sense of Rockfall Accidents Amongst Underground Gold
 Miners/ Ph.D., University of the Witwatersrand, 1992, pp. 7-11.
 3. Alan Altshuler, et. al, The Future of the Automobile, p. 14. Hounshell painstakingly
 reconstructs the development of mass production at Ford and places the advances in
 the context of the historical evolution of industry in America; David A. Hounshell,
 From the American System to Mass Production, 1800-1932 (Baltimore: Johns Hopkins
 University Press, 1984), especially Ch. 6, The Ford Motor Company & the Rise of
 Mass Production in America."
 4. In HounshelTs subtle analysis of the advent of mass production at Ford, he makes
 a fine distinction between Taylorist scientific management and Ford's practice.
 Taylor took production hardware as a given and sought revisions in labor processes
 and the organization of work; Ford engineers mechanized work processes and found
 workers to feed and tend their machines~?the machine ultimately set the pace of
 work at Ford, not a piece rate or an established standard for a 'fair day's work.9"
 Hounshell, From The American System, pp. 252-253.
 5. The historic developments at Ford's Highland Park plant allowed a massive
 expansion of output. From 1,700 vehicles produced in 1903, Ford's first year,
 production grew to 10,000 in 1908 with the introduction of the Model T, to 300,000 in
 1914 when the assembly line was fully installed, to 1.9 million in 1923 when Model T
 sales peaked. At that point Ford was producing 44% of the world's output (excluding
 the company's factories in Canada and England). Ford cut production costs by more
 than half within five years of introducing the assembly line, and fifteen years after
 introduction, the price of a Model T dropped to one-third its original leveL Rhys
 Jenkins, Transnational Corporations and the Latin American Automobile Industry
 (Pittsburgh: University of Pittsburgh Press, 1987), p. 14. See also Stephen Meyer III,
 The Five Dollar Day: Labor Management and Social Control in the Ford Motor
 Company, 1908-1921 (Albany, NY: State University of New York Press, 1981).
 According to Meyer, Ford's "Five Dollar Day" was not intended primarily as a
 21
motivational device (as, after all, Ford relied oa the pace of the Use rather than
 normative means to motivate workers) but as a strategy to obtain the best labour
 available by outbidding his Detroit competitors.
 6. Harry Braverman, Labor and Monopoly Capital' The Depodotion of Work in the
 Twentieth Century (New York: Monthly Review Press, 1974); see Peter Armstrong's
 review of criticisms of Braverman and a defense of his d*Afflfng argument, in
 "Labour and Monopoly Capital," in Richard Hyman and Wolfgang Streeck, eds*, New
 Technology and Industrial Relations (Oxford: BlackweU, 1988).
 7. Leger's work follows Cressey and Maclnness, who argue that the labour process
 must be understood in terms of the dual nature of the commodity. Peter Cressey and
 John Maclnness, "Voting for Ford: Industrial Democracy and the Control of Labour,"
 Capital and Class, No. 11 (Summer 1980). These authors reject the Marxist concept
 of the real subordination of labour to capital. Capital may purchase means of
 production and labour as exchange values, but the actual use value may only be
 realised by workers themselves. At some point, any labour process must enable
 workers to produce use value; hence, workers will always control the detail of work,
 and at the heart of the labour process will be a cooperative relationship between
 capital and labour. See the extremely valuable and dear examination of these
 concepts in Judy Mailer's recent book, Conflict and Co-operation: Case studies in
 worker participation (Johannesburg: Ravan Press, 1992).
 8. Leger explains that In the case of women workers, feminist writers have shown
 how "women's work" is systematically devalued and defined as unskilled, even in
 cases where women are now doing jobs previously performed by men. Similarly,
 Leger examines the processes where the definition of skills on South African mines
 depended on the race of the worker performing the Job. Leger, p. 63.
 9. Leger, p. 13.
 10. Charles Sabel, Work and Politics: The division of labor in industry (Cambridge:
 Cambridge University Press, 1982), p. 62. Sabel quotes from William Kornblum's
 description of a Chicago mill: "[W]hile the mill hand may have no generalizable
 skills, he is intimately familiar with the idiosyncracies of a particular mill. This is a
 familiarity which may take years to acquire~.The machinery seems to have a
 personality of its own and the men who coax steel through it know they cannot be
 easily replaced with new men. This contributes to an egalitarian spirit among mill
 men of different seniority; all are pitted together against the whims of a
 cantankerous old steel mill/
 11. Charles Sabel, Work and Politics: The division of labor in industry (Cambridge:
 Cambridge University Press, 1982), p. 80.
 12. For a survey of the development of the industry, see my paper "From the
 'Liverpool of the Cape' to the 'Detroit of South Africa': The Automobile Industry and
 Industrial Development In the Port Elizabeth-Uitenhage Region," Paper presented to
 the Conference on Port Elizabeth's Place in the History and Historiography of South
 Africa, Port Elizabeth, September 1992.
 22
13. The exceptions included components which could be competitively produced in
 South Africa, such as soft trim, paint, glass, and tires. I? the late 1950s, however,
 such locally produced items accounted for a mere 18% of the value of the car.
 14. In 1971 investment in the industry exceeded R200 million, three-quarters of which
 had been made in the previous ten years. In component manufacturing, investment
 rose from approximately R15 million In 1961, the year before the Local Content
 program took effect, to an estimated RlOO million ten years later. Through 1967 the
 motor firms had invested about R155 million in their assembly and manufacturing
 operations, Increasing to R200 million in 1971. Hie overall sales of passenger cars
 increased rapidly from 75,938 in 1961 to 229,031 in 1975, but the growth of
 components industry was faster still. The index of the physical volume of production
 for motor vehicles increased from 85.1 in 1963 to 199 in 1975, compared to an
 increase in parts and accessories production from 843 in 1963 to 4793 in 1975. The
 actual value of locally manufactured components bought by the motor firms
 increased from just over R13 million in 1961 to R180 million in 1973. See Ieuan L.
 Griffiths, The South African motor industry," The Standard Bank Review, No. 631
 (October 1977), p. 14,19; PJ. Nieuwenhuizen, The South African Component
 Manufacturing Industry (Johannesburg: Federation of Component Manufacturers
 Associations, 1977), p. 13; and NJ. Swart, The South African Motor Industry-in an
 International Context (Pretoria: die Afrikaanse Handelsinstituut, 1974), p. 168.
 15. Unpublished employment figures, January 1960-January 1987, National As-
 sociation of Automobile Manufacturers of South Africa, Pretoria; and Industrial
 Tribunal, "Report," 1964, p. 8. In the combined automobile manufacturing and com-
 ponents industry, employment more than doubled from 30,400 in 1964 to 64,700 In
 1975, while the ratio of black to white workers increased from 1.64:1 to 222:1.
 Nieuwenhuizen, The South African Component Manufacturing Industry, p. 67.
 16. In 1972-73 the following MNCs were represented in the South African passenger
 car market, either through wholly-owned subsidiaries or on a contractual basis with
 South African-based assemblers:
 Subsidiaries
 Ford
 General Motors
 Volkswagen
 Chrysler
 Leyland
 BMW
 Contractual Agreements
 American
 Datsun
 Toyota
 Mazda
 Peugeot
 Mercedes
 Motors Citroen
 Renault
 Fiat
 Alfa Romeo
 Volvo
 There were 21 manufacturers represented in South Africa (including commercials),
 producing 44 models of passenger cars (with 229 variants), for a market of 342,000
 vehicles. The Volkswagen Beetle was the model with the maximum production,
 amounting to 21,000 units. By contrast, there were 10 manufacturers in Brazil,
 producing 38 passenger car models for a market of 609,000, and the model with the
 23
maximum production amounted to 149,000 units. Swart, The South African Motor
 Industry, p. 189-190,196.
 17. Perhaps this fallacy results from the fact that most research on the industry
 relies on published statistical sources, closed-question surveys, and interviews with
 production management, rather than on participant observation and intensive
 Interviewing of shop floor workers.
 18. Rob Davies, Capital, State and White Labour m South Africa, 1900-1960 (Brighton:
 The Harvester Press, 1979), pp. 203-208.
 19. Souths]] asserts that the South African labour process facilitates effective
 unionisation, "not least because the bulk of the work on the assembly line is
 excruciatingly monotonous." He then gives a precis of Burner's examination of
 highly rationalised assembly line work, in which "job fragmentation and deskilling
 were Inherent In the process^Jeaving workers with minimal autonomy, with the
 unskilled being subject to the greatest pressure." He then asserts, "Notwithstanding
 South African manufacturers' occasional tendency to ascribe to blacks an animalistic
 sense of rhythm, there Is no reason to suppose that the latter find assembly line
 production any less alienating than do workers elsewhere." Roger Southall,
 "Monopoly Capital and Labour's Response: Emergent Unionism In the South African
 Motor Industry," mimeo, University of Ottowa, 1984, p. 17.
 20. See the contribution by Joy Welch, "The Technology Dimension," in Sylvia Miller
 and Roux van der Merwe, eds^ Labour Turnover and Organisational Effectiveness,
 Industrial Relations Unit, University of Port Elizabeth, 1980.
 21. Jan Roux, "Labour Control in the Eastern Cape AutomobUe Industry, 1970-1974,"
 Africa Seminar, Centre for African Studies, University of Cape Town, September
 1984.
 22. Of course the same conditions would apply in other low-volume countries
 producing for their Internal markets, such as Brazil, Mexico, Argentina, the
 Philippines, Australia, and generally most third world automobile assembling
 countries.
 23. The Industrial Council agreement establishes the wage grades and subsumes each
 job in the factory within the categories. At the time of research, the agreement
 consisted of 8 grades, from Grade 1 (the lowest) to Grade 8. Minimum pay levels
 were also established in the agreement, as well as periodic across the board increases
 for the life of the agreement.
 24. In fact, the only automatic increase occurred when workers moved from beginner
 to permanent status after three months on the job. In the absence of any systematic
 wage setting mechanism, all increases were based on merit, as decided by the
 production manager in consultation with the foreman. Unless he remained on good
 terms with supervision, a worker would have no other means short of a Wage Board
 investigation to increase his wages.
 24
25. Report by Representative W.G. Ballinger to the South African IVades and Labour
 Council," 20 December 1939, Ballinger Papers, University of Cape Town A3.II.11.1
 [from a microfilm copy held at CPSA, Reel 1], p. 2.
 26. Jan Roux, "Capitalist Control and the Labour Process in the Eastern Cape
 Automobile Industry," BA (Hons), African Studies, University of Cape Town, 1984, p.
 111.
 27. Mailer, Conflict and Cooperation, p. 112.
 28. Workers may also have been expressing realistic assessments of the likelihood of
 VW surviving through economic crisis, political upheaval, and international
 sanctions.
 29. Wolfgang Streeck, "Industrial Relations and Industrial Change: the Restructuring
 of the World Automobile Industry in the 1970s and 1980s," Economic and Industrial
 Democracy, V. 8 (1987),
 30. Mailer, Conflict and Co-operation, p. 137.
 25
UNIVERSITY OF THE WITWATERSRAND
 A F R I C A N S T U D I E S I N S T I T U T E
 Afr ican S t u d i e s Seminar Paper
 t o be pre s en t ed in RW 7003 SEMINAR ROOM
 AT 4.00pm 4 th October 1993
 Title: Political Instability and African Entrepreneurship
 by: Fred Ahwireng-Obeng
 No 349
POLITICAL INSTABILITY AND AFRICAN ENTREPRENEURSHIP
 Fred Ahwireng-Obeng
 Department of Economics,
 University of the Witwatersrand,
 South Africa.
 NOT TO BE QUOTED UNDER ANY CIRCUMSTANCES.
 SUMMARY
 Several close observers of African economic development believe that political
 instability (PI) has stunted economic progress in Africa by decreasing private
 entrepreneurial activity. However, whether African entrepreneurs themselves
 regard PT as an obstacle or an opportunity, how they respond to it and how it
 affects their behaviour as entrepreneurs are questions yet unanswered. This
 inquiry is the central focus of the paper. The evidence from the African
 business and political literature, periodicals and daily newspapers suggests
 no simple, direct relationships between PI, the volume of entrepreneurial
 activity and national welfare. On the other hand, different kinds of PJ tend
 to create different sets of opportunity conditions for profitable
 entrepreneurial initiatives. African entrepreneurs have responded rationally
 to potentially high private pay-offs resulting from PI and resorted to
 survival strategies to offset perceived and real obstacles. A considerable
 part of these initiatives are socially unproductive, suggesting the need for
 African governments to forge mutually beneficial coalitions with entrepreneurs
 in order to increase national welfare. The real problem is whether reliable
 coalitions are feasible.
1. INTRODUCTION
 The African historical record lends no support to the familiar idea that
 political instability (PI) decreases entrepreneurial activity. However, the
 evidence does confirm that different kinds of PI encourage or deter different
 forms of entrepreneurship. As a result, three decades of unstable political
 climate in sub-Saharan Africa have changed the character of its indigenous
 business sector. Indeed, the identity of entrepreneurs, the formality or
 informality of their enterprises and their degree of creative adaptation,
 expansion or innovation have all been determined in part by political
 instability.
 Those who see political instability as a curb on enterprise argue that
 it induces frequent and unforeseeable changes - and hence uncertainty - in the
 rules of business activity. This interferes with politically feasible sets
 of coalitions assembled by the entrepreneur and so prevents him or her from
 operating profitably in the legal or formal economy. Under conditions of
 instability, state controls intensify excluding the entrepreneur from the
 mainstream economy and leaving him to carry on his activities underground.
 The resulting reduction in formal economic activity reduces the state's
 limited tax base. Moreover, continued instability often leads to withdrawal
 of foreign financial assistance. As a result, economic development, is stunted
 (Mbaku 1989, pp.58-59). Another version of this thesis asserts that
 uncertainty arising from PI diminishes the entrepreneur's rewards and
 contribution to the economy by reducing the availability and marginal
 productivities of factors of production and disrupting the timing and
 organization of the production process. Production failure is usually
 accompanied by increased risk of capital loss and emigration of
 entrepreneurial and managerial expertise to escape financial hardships and
 personal political harassment. In this way, the scope of entrepreneurial
 activity is curtailed (Fosu 1992, pp.830-831).
 A major conceptual problem underlying many investigative studies on
 Africa is that they are based on simplified assumptions about the nature of
 political instability (McGowan and Johnson 1984; Mbaku 1987; Fosu 1992).
 Fosu's study, for example, applies an index of political instability at the
 elite level, taken to mean successful, attempted or planned coups d*etat as
 proxy for political instability in multiple regressions based on a production
 function framework to show that political instability has a deleterious impact
 on economic growth (and entrepreneurship) in sub-Saharan Africa. It ignores
 other important kinds of PI which are the outcome of contradictions in the
 role of the modern African government.
 The present paper defines political instability as "the condition in
 political systems in which the institutionalized patterns of authority break
 down and the expected compliance to political authorities is replaced by-
 violence intended to change the personnel, policies or sovereignty of the
 authorities through injury to persons or property. It is the result of
 conflict between different groups over the values according to which rewards
 are allocated in the political system" (Morrison 1989, pp.122-123). Five
 types of conflict - elite, factional, communal, mttss and popular - are
 examined. The definition of entrepreneurs used here is taken from Marsden a s
 those who "innovate and assume risks. They hire and manage labour- forces.
 They open up markets. They find new combinations of materials process and
 products. They initiate change and facilitate adjustment in dynamic
 economies" (Marsden 1990, p.l). The aim of the paper is to examine from the
 available literature how seriously African entrepreneurs regard the problem
 of political instability, how they respond to it, and how it affects their
 behaviour as entrepreneurs. In the next section the origin of PI is traced
to the fragility of modern African governments. The opportunity conditions
 so created are outlined in a descriptive model in section three. Section four
 examines entrepreneurs' responses to these conditions and section five
 concludes the paper.
 2. POLITICAL INSTABILITY AND THE MODERN AFRICAN GOVERNMENT
 Claphara (1991, p.91) suggests that the modern African government
 typically contradicts itself in exercising its functions (the purpose for
 which it exists), its power (what it is able to do) and its accountability
 (the relationship between itself and the people it governs). First, the state
 provides basic public needs and the opportunity for individuals to pursue
 their own goals. However, it frequently grants favours and opportunities for
 individuals and groups at the expense of their fellows. Second, the country's
 economic power base and the government's legitimacy, or the moral basis for
 government, are usually weak, which means it either exercises too much power
 to survive or too little for fear of rebuttal. Third, its own values and
 those of civil society which it governs are often irreconcilable. These
 contradictions usually leave the government fragile and at odds with civil
 society. The government naturally resorts to the politics of control rather
 than practise political democracy in order to safeguard its stability and
 power. Private enterprise, the means by which individuals acquire wealth and
 power outside government, becomes a threat which must either be frustrated or
 lured to subjugate to government control. The conflict that ensues is not
 only between government and private business but also between government and
 other social groups - students, professionals, academics, chiefs and religious
 leaders - who are affected by government action. Another dimension of the
 conflict takes place among a broad range of social groups competing to
 influence political decision (Chazan et al. 1988).
 This kind of conflict is classified as elite conflict and takes the form
 of lobbying, petitions, demonstrations and strikes in a vicious interaction
 of bad politics and poor economic management. When intense it escalates into
 coup attempts and successful coups d'etat. And when injudiciously handled the
 elites themselves organise their constituencies on the basis of class and
 ethnic identity to make factional demands. For example, divisions frequently
 arose between ideologues and technocrats in Nyerere's Tanzania and in Mobutu's
 Zaire, in which party functionaries, senior civil servants and military
 personnel contested among themselves to curry favour with the president. The
 continuous factional conflicts in Nigeria, Cameroon and Sudan are caused by
 ethno-regional leaders competing for political control. In Kenya, Zambia and
 Nigeria government has responded to factional conflicts by appointing
 factional leaders to poli tical decision-making positions. Some miljtary
 regimes in Ghana {between 1969 and 1979) and Nigeria (in 1979) avoided
 factional conflicts by succumbing to demands to transfer power to civilian
 governments. However, the Binfran succession from Nigeria (1967-70) resulted
 from failure of the Federal Government to accede to factional demands.
 To the extent, that the Biafran succession developed into a full-scale
 war, it may be classified together with the Katanga succession in Congo (1960-
 63), the Sudanese civil wars (1955 and 1962-72) and the civil wars in Chad as
 communal conflicts. By definition, they are "violent actions by members of
 one or more conimmin I groups to challenge the state's territorial integrity and
 protest existing distribution of authority and reward with the aim of
 redistributing them among communal groups or between communal groups and the
 national government" (Chazan et al. 1988, p.391; Morrison, 1991, p.123).
 Another typo of PI is mass conflict and involves revolts or revolutions
 in which insurgents of a political movement attack agents and supporters of
established political authority (Morrison 1991, p.124). Mass conflict is rare
 in Africa with exceptions such as the Ethiopian revolution which transformed
 the entire social order following the overthrow of the monarchy. However,
 popular conflict, in the sense of popular acts of protest and disengagement
 from formal authority by communities and groups who are socio-economically
 dislocated, is widespread in Africa in the form of "informal" activities.
 This form of response reflects a quest for protection from uncertainty, an
 attempt to control accessible resources in as many ways as possible and a
 willingness on the part of the deprived to reduce susceptibility to misery
 (Chazan et al. 1988, p.199). In Ghana, Nigeria, Somalia, Zaire, Kenya, Angola
 and elsewhere, the informal activities are organised through various
 institutions such as trading alliances, craft guilds and women's associations.
 But they are not always non-violent. Anti-social activities ranging from
 armed robbery and gangsterism to black market activities have thrived in the
 past three decades (Chazan et al. 1988, pp.199-200).
 No African country is free from any one of these five kinds of political
 conflict. Different types of conflict appear in different combinations in
 various countries at stages of the process of political maturing. Political
 instability does not necessarily have a negative impact on the entrepreneurial
 role. Out of the various types of conflict, new kinds of political coalitions
 appear and new entrepreneurial opportunities emerge. The next section
 explores this thesis.
 3. OPPORTUNITY CONDITIONS FOR AFRICAN ENTBEPRENEURSHIP; A DESCRIPTIVE MODEL
 The following model is based on suggestive evidence from the
 entrepreneurship literature and defines the theoretical basis of the paper
 (Broehl 1982; Wilken 1987; Baumol 1988):
 (a) Entrepreneurs will generally undertake activities most likely to yield
 personal wealth and position regardless of whether the activities
 contribute to productivity increase, economic growth or increase in per
 capita income (national welfare).
 (b) As entrepreneurship is not always productive, it is likely to influence
 national welfare not only by its size but also by the direction it takes.
 (c) The size and direction of entrepreneurship are dependent on the
 entrepreneurs' private profits or the rewards and incentives offered by
 current institutional arrangements.
 (d) In less developed areas, such as Africa, entrepreneurs seek and realize
 business advantage more often through political coalitions than through
 market interactions. Private profits can be substantial under conditions
 of political instability.
 (e) Profits are a function of the type of political conflict which in turn
 determine the degree of threat to the security of government. The degree
 of insecurity of government determines the extent of government control,
 the nature of government-business relationships) and so the security of
 person and property of the entrepreneur, uncertainty and risk of capital,
 loss.
 (f) The government-business variable appears in a general model of
 entrepreneurial, supply {the size of eutrepreneurship) as S, where
 productive entrepreneurship (+E) is equated with national welfare (Y) and
 is a dependent variable on the function of economic conditions (X) and
 non-economic conditions (Z) and S such that:
 +E = Y = f(X, I) + S
 The S variable is assumed to be promotional, protective or corrective of
 business or antagonistic to it. Moreover, it is exogenous and so
 influences all other variables namely Z X and E.
(g) Determinants of the X variable are the availability and quality of
 production factors and size and composition of the market. They influence
 E and Y to the same extent.
 (h) The non-economic factors in Z which promote enlrepreneurship include:
 social integration and mobility, non-marginality, high entrepreneurial
 legitimacy and pro-entrepreneurship ideology and high institutional
 (social and political) security.
 The relationship between insecurity arising from political conflict arid
 entrepreneurship has received diverse interpretations in the literature. Many
 authors believe that insecurity of all forms increases governmental
 interference and control, creates uncertainties and threatens the entrepreneur
 with unnecessary risk as asserted by the primacy of the PI thesis (Easterbrook
 1949; Brozen 1954; Brandenburg 1964; and Katzin 1964). Peterson and Berger
 (1971) as well as Kirzner (1973) hold the opposite view that entrepreneurship
 is more likely to be enhanced under turbulent conditions than umler stable
 conditions. However, others believe that a minimal security or moderate
 certainty is necessary (Cole 1959; McClelland 1961 and Kuznets .1966). These
 viewpoints contrast the Schumpeterian idea that the essence of
 entrepreneurship is to create insecurity in stable conditions (Schumpeter
 1961). But, Redlich (1958) suggests that insecurity does not necessarily
 decrease entrepreneurship, although it may result in business loss or fear of
 intimidation by some entrepreneurs who wi11 be milikely to increase their
 insecurity by assuming a more entrepreneurial role. We add thai, as different
 kinds of political conflict generate a different mix of political coalitions
 and opportunity conditions, different; levels of private pay-offs become
 attainable which in turn determine the direction and size of entrepreneurship.
 We now examine specific entrepreneurial responses to the mix of
 opportunity conditions generated by political instability in Africa.
 4. ENTREPRENEURIAL RESPONSES TO POLITICAL INSTABILITY
 The pre-independence historical record of African entrepreneurship re-
 iterates four themes: First, African entrepreneurship was positively elastic
 to political action. That is, the role of political authority was critical
 to its success or failure (I]iffe 1981; Hopkins 1988). Second, the
 continent's entrepreneurs were drawn from diverse social, economic and ethnic
 groups. Third, the majority of entrepreneurs showed considerable originality
 in adapting to their external working environments. Fourth, the
 entrepreneurial spirit stimulated a wide range of activities in trading,
 artisanal work, farming, transport and small scale manufacturing (Hopkins
 1983). These entrepreneurial activities have shifted strategically since thn
 time of independence to take advantage of political instability or to avoid
 its potentially ill effects. Also, the political role of African
 entrepreneurs has changed from passive to active agents for reform. The
 classic idea of Milton Friedman (1962) that "the business of business is
 business" no longer applies here.
 African entrepreneurs have performed this new-found role with
 considerable self-interest. After nearly two decades of PI and revolutionary
 climate in South Africa, characterised by boycotts, worker stay aways,
 international sanctions and administrative breakdown in blank residential
 areas, private sector organisations have formed alliances to promote the cause
 of political reform. Even so, these organisations, including black business,
 local white business and foreign business, pursue different strategy sets
 reflecting their different power bases and constituencies. Black business
 supports and promotes protest strategies but white business rejects them and
 rather restricts its involvement to economic issues. Foreign business,
however, deals with a more eclectic range of issues and a wider variety of
 strategies. This diversity restrains successful involvement (Lee et al. 1991,
 pp.116-124).
 Elsewhere in Africa the private sector lias largely failed to bring
 pressure to bear on governments to relax their domination of public policy for
 a more consultative approach. In 1984, the Rawlings Provisional National
 Defence Council (PNDC) administration in Ghana decided that the share of the
 private sector must exceed that of the public sector by 1991 (Tangri, 1991),
 Despite this commitment, the security and revolutionary arms of government,
 continued to harass indigenous entrepreneurs for unproven offences. The
 government failed to protect them from the impact of its market-based reform
 policies including devaluations of the local currency, positive and high
 interest rates, and unfair competition from cheap imported goods. Further,
 government intervention strengthened through stringent rules and regulations
 in employment, taxation, investment and import-export procedures. The private
 sector responded with anger and frustration and by 1992 more than 500 local
 industries had closed down (West Africa 1992, p.1572).
 According to a recent World Bank report reflecting a vision of Africa's
 economic future, the central role of its entrepreneurs is to stimulate demand
 for low-cost, products and services that they produce by creating jobs (World
 Bank 1989). There is, however, considerable difference of views on which
 entrepreneurs, formal or informal, are better disposed to perform this task.
 Formal or modern entrepreneurs are those who hire labour and are part of the
 formal economy; and informal entrepreneurs are the self-employed or micro-
 entrepreneurs who employ unpaid family members and who usually operate outside
 official regulations (Marsden 1990, p.2).
 The report projects that informal sector employment will rise by 6
 percent annually from 1990 to 2020. The sector's contribution to GDP will
 rise by 7.5 percent annually and its labour productivity wil.l increase by 1.5
 percent. However, labour productivity in the formal sector is expected to
 grow marginally at two percent annually. This means that the formal sector's
 contribution to total GDP will decline from 47% to 32%.
 This scenario - the expansion of informal entrepreneurship and
 contract ion of the formal - can also be inferred from the evidence of the
 response of Africans in production, commerce and the services to the various
 types of political instability. Generally, high intensity political
 instability constrains large-scale production and creates the envi ronment
 conducive to micro and informal production. Low intensity political
 instability such as popular conflict induces informal entrepreneurship in all
 sectors. Consequently, the dualistic structure of entrepreneurship tends to
 shift towards the informal.
 Elite Conflict
 In several African countries, elite conflicts explode into military coups
 d'etat to usher in new governments with new macro-economic policies including:
 price management/liberalisation (such as exchange rate adjustment, reduction
 or increase in subsidies); budgetary reform and institutional restructuring
 (such as entrepreneurship policy). Invariably, African governments differ in
 their attitude toward indigenous entrepreneurs. Some, J ike the Row!ings
 regime, the governments of Guinea, Mali in the 1960s under Keita, Benin, Congo
 Rrazaville, Angola, Mozambique, Somalia and Tanzania discouraged indigenous
 African entrepreneurs from emerging. The obvious response was the growth of
 informal, parallel or black markets. Some used political power to acquire
 private property and business interests. This is parasitic capitalism typical
of Zaire and Zambia under Kuanda, the last years of Tubraan and the early years
 of Tolbert of Liberia. However, the political leaderships in Kenya and
 Nigeria have always allowed local entrepreneurs to play an important economic
 role (Iliffe 1982, pp.77-87).
 Frequent changes in economic policies usually affect public servants,
 wage earners and the unemployed the most. Their response is to drift, to the
 informal sector, thus creating competition and reducing their earning
 capacity. To adjust their tastes to match their declined purchasing power,
 they embark on innovative local repairs. When goods become expensive or
 scarce, or commodity prices offered to producers are low, informal market
 activities are the popular response. Producers and speculators smuggle to and
 from neighbouring countries (Chazan et al. 1988, pp.254-256).
 An alternative response, quite popular in Ghana and Nigeria, is for
 displaced urban dwellers to "go back to the land", that is return to their
 villages of origin to undertake subsistence agricultural production. A few
 cases of urban professionals turning to plantation agriculture have, however,
 been reported (Ewusi 1987; Chazan et al. 1988, pp.254-256).
 Large-scale migration, another adjustment strategy, is dominant in West
 Africa. During the early years of Rawlings' rule, many Ghanaians migrated to
 Nigeria and Cote ij'lvoire. This meant, in effect, a flight of entrepreneurial
 talents to the host countries. However, most immigrants are usually
 restrained from undertaking large-scale, formal business and are obliged to
 mount informal enterprises.
 Factional Conflict
 The political conflict in South Africa may be considered as a special
 case of factional conflict,taking the form of a racial conflict between blacks
 and whites. At the corporate level, it is a conflict between the black owners
 of labour and white owners of capital backed by a racist government. As a
 result, the private sector functions in an unstable, hostile environment,
 characterized internally by the sanctions of political uncertainty at the
 workplace and outside it and external.] y from international economic isolation.
 Decades of living under siege have made the industrial sector structurally
 uncompeti I.ive and entrepreneurs are concerned more with protection and
 survival than with expansion (Financial Week, April 2-8 1992, pp.24-25).
 Given (lie importance of political stability as a consideration for
 investors, South Africa ranks 59th just behind China, Poland, Hungary, Israel,
 Saudi Arabia and Hong Kong (41st.) in the 1991 International Country Risk Guide
 of 129 countries of which the US places ninth. In view of South Africa's
 relatively unimpressive ranking", its private sector suffers severely from
 flight of capital and lack of foreign investment.
 Communal Conflict
 The evidence on entrepreneurial responses to communal confict is rather
 inconclusive. Large-scale agriculture has been difficult under war conditions
 although small-scale farming has had mixed fortunes. During the Moaambican
 war Lonrho, a multi-national, committed about $20 mill ion in agribusiness
 joint venture and was earning about one million dollars profit annually by
 1989. However, it spent 15% to 20% of its production cost on hiring private
 militias to protect its estates from Renamo attacks. High security cost meant
 that only large-scale investment in agriculture could be profitable (African
 Business, June 199.1). The war destroyed traditional coping strategies among
 farming communities and undermined the existing exchange entitlement network
(Mozambique Information Office 1992, p.2). The Sudanese experience was
 however, different. In the midst of the civil war, the besieged people of
 Juba turned faeces-ridden waste ground in the city into food pjots to produce
 about, a third of their basic food requirement (Daily Times, 6 January, 1992,
 p.10).
 War conditions could also stimulate profitable entrepreneurial
 initiatives in eco-tourism and the ivory trade. The rebel war in Rwanda
 boosted neighbouring Uganda's tourist industry when bush fighters sent
 Rwanda's mountain gorillas scurrying for safety across the border into Uganda.
 The Ugandan government responded by inaugurating a profitable eco-tourism
 programme (Africa Analysis, 18 Sept. 1992, p.14). Similarly the war in
 Mozambique launched a buoyant ivory trade, for guns and war material but
 drastically reduced the country's elephant population from about 55 000 in
 1974 to 13 350 in 1992 (New Ground, Spring 1992, pp.18-19).
 In mining and general trading, the civil unrest in Zaire and the war in
 Angola gave mixed business prospects for South African entrepreneurs. By
 November 1991 the turmoil in Zai re had sent cobalt prices from $15.50 in
 September to $25.27 (the average price during the 1970 unrest was $70).
 Stocks fell fast, and as Zaire dominates world production the situation
 benefitted South African producers - Rustenburg Platinum and Impala Platinum -
 who extract it as a by-product of platinum. South African exporters found
 lucrative business in Zaire, from fresh produce, mining and manufacturing
 equipment to processed foodstuff, taking advantage of the superior quality and
 cheaper prices of their products. The South Africans appeared also in newly
 accessible Mozambique, rebuilding hotels, breeding cattle and seeking
 contracts for the supply of commodities ranging from plastics to shipbuilding
 items (The Star, 9 October 1991; The Citizen 23 April 1992; Business Times,
 10 November 1991). However, the $125 million trade between Zaire and South
 Africa frequently came to a temporary standstill during periods of loot-ing,
 arson and riots in Zaire.
 The Angolan war created many other unforseen business opportunities for
 the unemployed, Zairean immigrants and local merchants. The brief return to
 peace and stability in that country in 1992 meant that patrolling soldiers
 in key diamond areas be withdrawn to ensure free mobiLi ty. Wlien the
 government introduced a law to legalise the right of Angolans to own and sell
 diamonds with the aim of mopping up diamond stocks among the population, the
 unemployed expressed their entrepreneurial proclivity and did a $500 rnillion-
 a-year business in diamond sales (Southern African Economist, October 1992,
 p.40). At the onset of the war in 1975, most of the Portuguese immigrants who
 had dominated the private sector left the country. The openings left in a.
 wide range of skilled and unskilled vocations were filled by Zairean and other
 immigrants. Parallel markets opened up everywhere to supplement the centrally
 planned economy which could not afford to provide the people's needs. When
 the traditional caravan network of long-distance trade collapsed from the
 destruction by the war of the country's communication infra-structure, the
 merchants in the highland areas grasped the opportunities of ai r? transport to
 ferry goods to other provincial areas. This private enterprise expanded to
 reach supply sources in Lisbon, Rio dc Janeiro and Paris (Birmingham, 1988).
 Other parallel market operations were illegal and inimical to economic
 growth. Refugees from the Mozambican civil war f1ocked to the country's
 informal sector. Rut it became ruled by Mafia-style entrepreneurs, taking
 advantage of the poor and war-displaced people, by hiring them as hawkers to
 evade tax. Similarly, the paraI lei market in politically unstable Sierra
 Leone became so pervasive that it impinged on the functioning of the formal
 economy. Most foreign exchange transactions including exports, imports and
8
 migrant remittances were channelled through this market due to lack of
 confidence in the country's political and economic management (African
 Economic Digest, 13 July 1992).
 Popular Conflict
 Socio-economic conditions in most African countries reveal sustained
 deprivation and neglect. These circumstances induce collective informal
 entrepreneurship within community groups. The groups conduct self-help
 schemes, operate occupational collectives and undertake profitable commercial
 ventures. Others, however, run smuggling and black market syndicates which
 could be socially undesirable. The frustrations of popular conflict manifest
 themselves also in petty thefts, industrial sabotage, embezzlement, armed
 robbery, thuggery, tax evasion and other forms of criminal entrepreneurship
 (Chazan et al. 1988, p.199).
 The policy of apartheid illustrates an extreme case of this type of
 conflict. Apartheid literally means that the black communities in South
 Africa were relatively deprived of basic human needs. Their response was to
 evolve a distinctive, informal sub-culture, an informal means of human
 expression and survival. For example, they introduced shebeens as underground
 focal points for recreation similar to the Bri tish public house and the
 American speakeasy, when the white regime prohibited the consumption of
 "European" alcohol (Nkosi, 1965, p.14; Sampson, 1956, p.703; Nakasa, 1975,
 p.57; Hart, 1991, p.77). Similarly stokvels or rotating savings and credit
 associations became an immensely popular means by which blacks acquired
 venture capital in a society where they were deprived access to formal sources
 of finance. The informal culture is so deep-seated that transforming
 informal, black entrepreneurship into formal types, is one of the major
 problems in South Africa's transition to a non-racial, economic democracy.
 5. CONCLUSION
 A curvilinear relationship between political instability and private
 entrepreneurship can be implied from the African experience. Entrepreneurs
 operating' large manufacturing enterprises have been adversely affected
 particularly by high-intensity political instability such as wars. Prolonged
 interruptions to production, insecurity and uncertainty have squeezed many out
 of business. At the same time the trading business has become lucrative.
 Also, bo tli high and low intensity Pi have provided favourable sets of
 opportunity conditions for micro and informal entrepreneurship. The evidence
 suggests further that socially unproductive black-market activities are
 widespread, which means that private profits are considerable. In effect,
 African entrepreneurs have reacted rationally to a wide range of politically
 unstable s i t, ua t. i ons.
 Political instability has modified features of African entrepreneurship
 in various notable ways. Industrial entrepreneurs have reacted to the
 disastrous impact of instabi 1 ity on their sector by consulting more frequently
 with their governments than ever before to influence public policy. However,
 the interactions are not always cordial. Besides, two changes have
 accompanied the preponderance of informal business. First, a new breed of
 operators has joined the sector. This includes well-educated or high.ly-
 skilled public-sector workers who have lost their jobs, those seeking to
 supplement their incomes, and families displaced by conflict particularly the
 more vulnerable women and children. Moreover, the new breed are more
 sophisticated in their innovations such as substituting for imported inputs
 and creating new markets under chaotic political conditions.
9
 The expansion of the informal sector indicates the capacity of the
 entrepreneurs for adjustment and survival. But the sector alone provides
 neither the panacea nor the basis for sustained economic development, African
 entrepreneurs have shown also that given growth-directed incentives from the
 mixed fortunes of political instability, they have the potential and dynamism
 to undertake productive economic activities on a wider scale. A successful
 policy to promote productive entrepreneurship wi 11 depend on government-
 business coalitions accomplished to protect the interests of both parties
 while seeking to maximize national welfare. Nonetheless, this scenario is
 improbable for practical reasons.
 We have depicted the modern African government as a fragile, insecure
 institution constantly drawn into various types of institutional conflicts
 with its constituencies. Consequently, it is difficult to imagine that
 governments will engage their business sectors in mutually trustworthy
 coalitions. Coalitions are least likely in Tanzania, Ghana, Benin, Angola and
 Mozambique where evidence of African socialism or ant i-entrepreneur-ship policy
 is common. In Zaire, Ivory Coast and Zambia where parasitic capitalism was
 practised by the political elites any coalitions are likely to be shaky. The
 most, probable scenario could be found in Nigeria and Kenya where post-
 independence governments have consistently nurtured indigenous
 entrepreneurship (Hi f fe 1983, pp. 64-87). Ultimately, African governments
 will have to show transparency, openness, predictability and the rule of law
 to win the confidence of their entrepreneur class.
10
 BIBLIOGRAPHY
 ADEDEJI, A. "Stable Foundations" Africa Forum, Vol. 1 No. 3, 1991, pp.52-32.
 AFRICA ANALYSIS, "Uganda to Profit from Hairy Refugees", 18 September 1992,
 p.34.
 AFRICA BUSINESS, IC Publication, London.
 AFRICA ECONOMIC DIGEST, London and Lagos.
 AKINIRADE, 0. "Threats to Security and Stability in Nigeria; Perception and
 Reality". Geneve Afrique, Vol. 26, No. 2, 1988, pp.47-82.
 ANGOLA ECONOMIC NOTES, "Diamond Smuggling Reaches Crisis Levels", Vol. 2, No.
 4, April 1992, pp.1-3.
 ANYANG NYONG'O, P. "Political Instability and the Prospects for Democracy in
 Africa". Africa Development, Vol. 13, No. 1988, pp.72-77.
 BAUMOL, W.J. "Is Entrepreneurship Always Productive?" Journal of Development
 Planning (Special Issue on Entrepreneurship and Economic Development;) No.
 18, 1988, pp.85-94.
 BAYNHAM, S. "Security Issues in Africa". Africa Insight, Vol. 21, No. 3,
 1991, pp.181-189.
 BIRMINGHAM, D. "Angola Revisited". Journal of Southern African Studies.
 Vol. 15, No. 1, October 1988.
 BRANDENBURG, F. "Contributious to the Theory of Entrepreneurship and Economic
 Development: The Case of Mexico". Inter-American Economic Affairs, 16,
 pp.3-23, 1962.
 BROEHL Jr. , N.G. "Entrepreneurship in the Less Developed World". In
 Encyclopaedia of Entrepreneurship, eds Kent, C.A. Sexton, D.L. and Vesper,
 K.H. Prentice Hall Tnc. , Englewood Cliff, NJ, 1982.
 BROZEN, Y. Determinants of Entrepreneurial Ability, Social Research, 21,
 pp.339-364, 1954.
 BUSINESS DAY, Johannesburg. South Africa.
 BUSINESS TIMES, Johannesburg, South Africa.
 CHAZAN, N. et al. Politics and Society in Contemporary Africa, Lynne Reinner
 Publishers, Boulder, Colorado, 1988. Chapter 7.
 CLAPHAM, C. "The Africa State". In Rimmer D. et al. Africa: 30 Years On.
 Hieneman, London, 1991. Chapter 7.
 COLE, A. Business Enterprise in its Social Setting. Oxford University Press,
 Cambridge, 1959.
 DAILY TIMES. Blantyre, Malawi.
 EASTERBROOK, W.T. "The CJimate of Enterprise". American Economic Review.
 39, 1949, pp.322-325.
 ELKAN, W. "Entrepreneurs am) Entrepreneurship in Africa" World Bank Research
 Observer. Vol. 3, No. 2, July 1988, p.171-188.
 EWUSI, K. Structural Adjustment and Stabj.1 isat ion Policies in Developing
 Countries: A Case Study of Ghana's Experience in 1983-86. Ghana
 Publishing Corporation, Tema, 1987.
 FINANCIAL WEEK, Johannesburg.
 FOSU, A.K. "Political Instability and Economic Growth: Evidence from sub-
 Saharan Africa" Economic Development and Cultural Change. Vol. 40, No.
 4, 1991.
 FR.EIDMAN, M. Capitalism and Freedom. University of Chicago Press, Chicago,
 1962.
 HART, D. "The Informal Sector in South African Literature". In Preston-Whyte
 and Rogerson, C. (Eds) South African Informal Sector. Oxford University
 Press, Cape Town, 1991.
 HOPKINS, A.G. "African Entrepreneurs!! ip: An Essay on the Relevance of
 History to Economic Development". Geneve Afrique, Vol. 26, No. 2, 1988,
 pp.7-28.
 ILIFFE, J. The Emergence of African Capitalism. Macmillan, London, 1983,
 Chapter 4.
u
 KATZIN, M.F. "The Role of the Small Entrepreneur". In M.J. Herskowits and
 M. Horwitz (Eds) Economic Transit Jon in Africa. Evanston, 111. , North
 Western University, 1964.
 KIRZNER, I .M. Competition and Entrepreneurship, Chicago, University of
 Chicago, 1973.
 KUZNETS, S. Modern Economic Growth: Rate, Structure and Spread. New Haven,
 Conn., Yale University Press, 1966, p.451.
 LEE, R et al. "Speaking or Listening? Observers or Agents of Change?
 Business and Public Policy: 1989\90". In Transition to Democracy, Lee,
 R. and SchJemmer L. (eds), Cape Town, Oxford University Press, 1991,
 Chapter 6.
 MARSDEN, K. African Entrepreneurs: Pioneers of Development. International
 Financial Corporation Discussion Paper No. 9, World Bank, Washington, 1990.
 MBAKU, .T.M. "The State and Economic Growth and Development in Sub-Saharan
 Africa" Journal of Third World Studies, Vol. 14, No. 1, Spring 1987,
 pp.49-68.
 McCLELLAND, D.C. The Achieving Society, New York, Free Press, 1961.
 McGOWAN, P. J. and JOHNSON, J.H. "African Military Coup d'Etat and
 Unde[-development: A quantitative Historical Analysis". Journal of Modern
 Political Studies, 22, No. 4, 1984, 633-666.
 MORRISON, D.G. B.lack Africa: A Comparative Handbook. Macmillan, London,
 1989. Chapter 5.
 MOZAMBIQUE INFORMATION OFFICE. Mozambique Special Report No. 10. London,
 May, 1992.
 NAKASA, N. The World of Nat Nakasa. Ravan Press, Johannesburg, 1975.
 NEW GROUND, Newtown, Johannesburg, South Africa.
 NKOST, L. Home and Exile. Longman, London, 1965.
 PETERSON, R.A. and BERGER, D.G. "Entrepreneurship in Organisations: Evidence
 from the Popular Music Industry" Administrative Science Quarterly. 1971,
 16, pp.97-106.
 REDLICH, F. "Business Leadership: Diverse Origins and Variant Forms".
 Economic Development and Cultural Change. 1958, 6, 177-190.
 SAMPSON, A. Drum: A Venture into the New Africa. Collins, London, 1956.
 SCHLEMMER, L. "Business and Politics in South Africa". South Africa.
 International. Vol. 21, No. 1, July 1990.
 SCHUMPETER, J.A. The Theory of Economic Development. New York, Oxford, 1961.
 SOUTH AFRICA FOUNDATION. "South Africa in Transition: Risks and Prospects".
 Johannesburg, 1992.
 SOUTHERN AFRICAN ECONOMIST. Harare.
 STEEL, W.E. AND WEBSTER, L.M. "How Small Enterprises in Ghana Have Responded
 to Adjustment". The World Bank Economic Review. Vol. 6, No. 3, September
 1992, pp.423-438.
 TANGRI, R. "The Politics of Government-Business Relations in Ghana". Journal
 of Modern African Studies. Vol. 30, No. 1, 1991, pp.97-111.
 THE CITIZEN. Johannesburg, South Africa.
 THE STAR, Johannesburg, South Africa.
 THE WEEKLY MAIL. Johannesburg, South Africa.
 WEST AFRICA. London, UIL.
 WILKEN, P.H. Entrepreneurship: A Comparison and Historical Study. Norwood,
 N.J., Ablex Publishing Corp., 1979.
 WORLD BANK. Sub-Saharan Africa: From Crisis to Sustained Growth. Washington,
 1989.
UNIVERSITY OF THE WITWATERSRAND
 INSTITUTE FOR ADVANCED SOCIAL RESEARCH
 SEMINAR PAPER
 TO BE PRESENTED RN THE RICHARD WARD BUILDING
 SEVENTH FLOOR, SEMINAR ROOM 7003
 AT 4PM ON THE 14 OCTOBER 1996.
 TITLE: The UN's Experience of Demobilization,
 Disarmament, and Reintegration in Southern
 Africa
 BY CHRIS ALDEN
 NO: 411
THE ISSUE OF THE MILITARY:
 THE UN'S EXPERIENCE OF DEMOBILIZATION, DISARMAMENT AND
 REINTEGRATION IN SOUTHERN AFRICA
 By
 Chris Alden
 Department of International Relations
 University of the Witwatersrand
The Issue of the Military: The UN's Experience of Demobilization,
 Disarmament and Reintegration in Southern Africa
 "What about the people who were at the battlefront during
 the liberation of our country, and who are not educated?
 What are we going to do? Many of us are illiterate people
 who only know how to fight. Now we are being threatened by
 intellectuals who were lucky to be sent abroad by SWAPO to
 study for the benefit of the Namibian nation."1
 In early 1995, Boutros Boutros Ghali released an amended version
 of his now famous Agenda for Peace. According to the Secretary
 General, the composition elements of the United Nations' peace
 support programme consisted of preventive diplomacy, peace
 making, peace enforcement, peacekeeping and peace building.2
 More than any other single element in Boutros Ghali's typology,
 the military stands at the crossroads of all points on this peace
 continuum: it is the trigger of conflict, it is primary
 instrument of violence, it is the enforcement arm of
 international law and it is the ultimate symbol of
 reconciliation. From technical issues of disarmament and
 demining to the human dimension of reconciliation and
 reintegration into society, the question of the military ranks
 as the most complex and volatile components of any peace support
 operation. Devising an approach which takes into account both
 the multi-faceted nature of demilitarization and its varied time
 frame is one of the pressing challenges for the international
 community in the wake of the onset of peace.
 This article will analyse the United Nations' experience of
 demilitarization in Southern Africa with reference to the issues
 surrounding the transition from an emergency situation to a
 developmental context. The three UN peace support operations in
 Southern Africa, namely Namibia, Mozambique and Angola, will be
 examined to assess the international organisation's role in that
 process. In so doing, the article highlights a number of themes,
 including the importance of developing a regional approach to
 demilitarization; the imperative of cooperation both within the
 UN itself and with the international donor and NGO community as
 a whole; and the need to develop a greater understanding of the
 efficacy of "targeting" demilitarization programmes towards ex-
 combatants in light of the broader goals of demilitarization.
 I. Between Emergency and Development: Theoretical Parameters of
 Demilitarisation and Peace Building
 Demilitarisation of combatants, a process which encompasses the
 demobilisation and disarmament of troops and their reintegration
 into society,3 holds as its premise that combatants are
 particularly dangerous elements to interject into the fluid post-
 conflict situation. Both the perpetrators and, as often, the
 objects of brutalizing violence, combatants have the capacity to
 disrupt the fragile peace settlement, either by returning to open
 hostilities with their opponents or resorting to armed banditry
in the aftermath of formal peace. Accordingly, it has become a
 kind of cardinal principle that these self-same volatile elements
 and circumstances can be offset through a targeted programme
 which builds on the structured demobilization and disarmament of
 combatants in conjunction with some form of monetary and
 educational incentives designed to facilitate their peaceful
 reintroduction to civilian life. This two phased approach to the
 question of demilitarization of combatants is composed of both
 the short term objectives of emergency assistance and the long
 term objectives of development. As such, it is well-ensconced
 within the emergency/development continuium as it is currently
 manifested in peacekeeping operations and subject to all of the
 complexities inherent in that process.
 The demilitarization programmes instituted by the UN have, in the
 main, recognized the necessity of extending the scope of
 demilitarization beyond the short term objectives of
 demobilization and disarmament. In the lexicon of the UN, peace
 building represents the transitional point between the
 demobilisation of combatants and their full reintegration into
 society.
 Peacemaking and peace-keeping operations, to be
 truly successful, must come to include
 comprehensive efforts to identify and support
 structures which will tend to consolidate
 peace... (T)he concept of peace-building as the
 construction of a new environment should be
 viewed as a counterpart of preventive
 diplomacy...4
 Nevertheless, further conceptual clarity on the matter of
 emergency concerns and development imperatives, and their
 relationship to the objectives of demilitarisation, is needed if
 UN peace support operations are to act effectively in this area.
 In this same vein, it is therefore noted that the controversies
 which often attend project design and implementation in
 demilitarization are themselves a product of this disjuncture
 between the emergency and development ethos.5 Indeed, it is the
 contention of one SRSG that this disjuncture is one of the major
 failings of UN Peace Support Operations.6 More than a mere
 intellectual oversight, to be recognized and corrected, the gap
 between emergency and development ethos is rooted in and finds
 institutional expression in the differing UN agencies which make
 up peace support operations. Another element which further
 complicates the situation is the role of the international NGO
 community which itself is divided along the emergency and
 development fault line and, accordingly, acts with varying
 degrees of autonomy from the UN peace support operation. This
 has the Janus effect of, at times, contradicting the aims of the
 mission and, in other instances, ameliorating the oversights or
 failings of the UN.
 II. The UN, Demilitarization and the Southern African Context
In the aftermath of colonialism and apartheid, intensified by the
 struggle to dislodge these elements through force of arms,
 Southern Africa is a landscape awash with the brutal legacies
 conflict. Super power intervention further contributed to the
 immolation of the physical, social and indeed, spiritual terrain
 of the region. And while the conflicts in Namibia, Angola and
 Mozambique were often disaggregated into discrete confrontations,
 the common denominator of apartheid South Africa, a regional
 hegemon anxious to reshape Southern Africa in its own image, and
 an extensive ideologically-based external intervention, belies
 cognitive approaches (and the attendant policy implications)
 which are predicated on the inviolability of the state system.
 Demilitarisation, as it has become abundantly clear in the wake
 of UN involvement in Southern Africa, is not a national problem
 but rather a regional one.
 The three cases from Southern Africa -- Namibia, Angola and
 Mozambique -- provide concrete examples of UN involvement in
 demilitarization, highlighting the strengths and weaknesses of
 its changing approach in the region. Most notably, the UN
 refined its demilitarisation programme to include both short and
 long term components as well as moving towards an integrated and
 comprehensive approach to the issue of the military. In each
 case study, the following areas will be examined: the
 demilitarization programme as conceived in the peace settlement;
 the demilitarization programme as organised by the UN peace
 support operation; and the record of implementation.
 Namibia, and UNTAG
 Established in the wake of the signing of the New York Accords
 in December 1988, Namibia was the first post-Cold War mission of
 the UN, the largest peacekeeping operation since the ill-fated
 mission to the Congo and the most complex ever undertaken by the
 organisation. As such it posed a series of challenges to the
 UN's capacity to conduct a multi-dimensional operation that
 incorporated elements of traditional peacekeeping as well as
 novel components in such areas as police and elections
 monitoring. Its successful resolution, despite some serious
 incidents, gave considerable encouragement to the international
 community as to the efficacy of extending the UN role in
 peacekeeping world-wide.
 A. The Peace Agreement and Demilitarization
 The agreement formally ending the conflict in Namibian had in
 fact been drawn up in 1978 under the auspices of the UN's Western
 Contact Group. Security Council Resolution 43 5 (1978), which
 called for the withdrawal of South African military and
 administrative control over the territory and democratic
 elections, formed the basis for the cessation of hostilities and
 the transition to independence. Following the assent of all
 parties to the New York Accords, the Security Council passed
 resolution 629 (1989) which officially established the mission
 to Namibia. Accordingly, the UN was charged with the supervision
of the cease fire, monitoring of the conduct of the South West
 African Police (SWAPOL) police and observing the election
 campaign. The time table for the operation was to cover twelve
 months and contained the following steps:
 * a cease fire was to be formally established on
 1 April;
 * the disbanding of the SWATF and SWAPO and the
 reduction of SADF, from a high of 32,000 to nil
 by 8 November 1989;
 * elections were to be held 16 November 1989.
 All of this was to take place in tandem with the Cuban withdrawal
 from Angola, overseen by UNAVEM I (see below).
 Overseeing the implementation of the entire process would be a
 Special Representative of the Secretary General and a South
 African appointed Administrator-General. To fulfil the stated
 aims of SC Resolution 435, UNTAG would be structured to include
 a civilian component, a police component and a military
 component. The civilian component consisted of the Special
 Representative's Office, which was supported by 4 2 smaller
 political offices established across Namibia's ten regions, and
 would provide overall direction and coordination to the mission,
 as well as prepare the foundation for UNDP involvement after the
 mission; an Independent Jurist who would arbitrate on matters
 relevant to the peace process; the UNHCR, which was responsible
 for the repatriation of refugees in advance of the elections; an
 electoral division, which would oversee voter registration and
 monitor the elections in November 1989; and a logistics division.
 The police and military components are discussed below.
 B. UNTAG and Demilitarization
 In the Namibian case, the UN mission's involvement in
 demilitarization was conceived wholly in terms of short term
 objectives, that is to say, the demobilization and disarmament
 of combatants. No provisions for reintegration of former
 soldiers were introduced into the programme, nor were they the
 recipients of any targeted projects, assistance or funding beyond
 that provided in the course of their stay in the reception areas.
 The first aspect of the United Nations Transitional Authority
 Government's (UNTAG) role in demilitarization centred around the
 supervision of the confinement and withdrawal of South African
 troops from the territory and the concomitant repatriation of
 South West African Peoples Organisation (SWAPO) guerrillas from
 their bases in southern Angola and Zambia. Linked to this was
 the monitoring of the cease fire between the former foes. To
 fulfil these objectives, UNTAG was to bring in a team of 2 00
 military observers and 7,500 peacekeepers. Their task was to
 monitor the following:
 With
* the restriction of the South African Defence
 Forces (SADF) to bases within Namibia by 1 April
 1989 and their subsequent withdrawal from the
 territory, with the first reduction to 12,000
 troops by 14 May 1989, 1,500 troops by 1 July
 1989 and total withdrawal by 8 November 1989;
 * monitor the dismantling of the SWATF, commando
 and ethnic units and their disarmament by 1 April
 1989;
 * monitor the disarmament and repatriation of
 SWAPO forces, including the confinement of
 selected forces to bases in Angola and Zambia7;
 * policing of official entry points into the
 country and reception areas for returnees.
 The second aspect of UNTAG's role in demilitarization was the
 monitoring of the Namibian police and those elements of the SADF
 which had taken up civilian functions during the transition
 period. Complicating the situation further was the incorporation
 of the notorious counter-insurgency unit, Koevoet (Crowbar) into
 SWAPOL. With SWAPOL given the principal role in maintaining law
 and order during the transitional period, it would be the UN
 mission's 360 CIVPOL members job to ensure that it did not engage
 in acts of intimidation against the population during the run up
 to the election. The use of CIVPOL in fulfilling this task was
 one of UNTAG's innovations: no other mission had undertaken such
 a direct and extensive policing role.
 C. Implementation
 The implementation of the demilitarization programme was subject
 to the problems and constraints facing UNTAG. Despite having
 literally years to prepare for the mission, when it came time to
 implementing the objectives of SC Resolution 435 (1978), it was
 clear that little work had been done to realize the actual
 mechanics of the operation. In the first instance, delays in the
 passage of enabling legislation by both the Security Council,
 which only authorized the particulars of the operation on 16
 February, and the General Assembly, which gave its approval two
 weeks later, cut an already narrow margin for deployment of UNTAG
 to the bare minimum. At the heart of these delays was a dispute
 over the mission's budget, which pitted SWAPO and African states
 anxious to see UNTAG maintain its full force strength in the
 event of a return to hostilities and the Security Council
 members, who believed that conditions had changed substantially
 in the territory to warrant a reduction in size and cost of the
 mission. Ultimately, the Security Council prevailed and UNTAG
 peacekeeping troops were decreased to 4,650 (while officially
 remaining at the original figure), reducing overall costs from
 US$700mn to US$446mn.
 With regard to the actual mission, some problems emerged due to
 uncertainties as to UNTAG's chain of command. For instance,
while the Special Representative of the Secretary-General, Martti
 Ahtisaari, was nominally head of the mission, given the
 specialized agencies institutional and financial independence
 from New York, initially there were difficulties in coordination.
 This was to impact upon the already problematic communication
 between New York and the field headquarters, as well as adversely
 effecting communication between the different components of the
 mission in Namibia.8 Unwieldy procurement procedures, which
 obliged UNTAG to forgo purchasing material from nearby South
 Africa so as to maintain adherence to international sanctions,
 and insufficient logistical arrangements in the field, was a
 further obstacle.9 Nevertheless, the overall picture of the
 mission was one of cooperation and successful coordination
 between the different elements of the mission.
 The first and most dangerous crisis came in the early days of the
 UN mission with the unexpected infiltration of hundreds of SWAPO
 guerrillas from their bases over the border. As the SADF had
 been confined to bases, SWAPOL undertook to repel the incursion
 alone. SWAPO contended that the intention of the incursion was
 merely to move its forces into Namibia as part of the larger
 demobilization exercise while the South African officials
 declared it to be an outright violation of terms of the peace
 agreement. Though UNTAG had already been alerted to possibility
 of SWAPO incursions at least a month before it actually took
 place, it was clear that neither New York nor the field office
 in Windhoek had made provisions for this contingency.10 Faced
 with a choice of unilateral action on the part of South Africa
 and the implications of such action, UNTAG agreed to allow the
 SADF to suspend its confinement and respond to the incursion by
 releasing six battalions. Over 200 SWAPO members were killed
 and, after a meeting of the Tripartite Commission11, Sam Nujoma
 announced that SWAPO would return to its bases outside Lubango,
 in Angola, under UNTAG escort. By 13 May, the cease fire was
 back in place.
 Another issue which clouded the demilitarization process was the
 incorporation of Koevoet into the police force which had,
 inexplicably, won UN approval in advance of the UNTAG mission.
 With a deserved reputation for human rights abuses, there were
 fears that intimidation of the electorate would be effectively
 sanctioned in the guise of policing. Complaints by CIVPOL, whose
 task it was to pair up with SWAPOL, as to both the conduct of
 Koevoet and that of the police in general provoked a crisis.12
 After much negotiation, Ahtisaari and Pienaar, the Administrator-
 General, worked out a compromise which saw Koevoet members leave
 the police and CIVPOL numbers increase substantially to 1,500.
 The last former Koevoet members left the police at the end of
 October 1989.13
 Much of the formal disarmament and demobilization process was
 effectively run outside of the UNTAG framework with the UN
 playing the role of observer or monitor. In the case of SWAPO,
 disarmament took place in Angola and Zambia followed by
 registration of repatriated refugees, inclusive of former
 guerrillas, at selected Assembly Areas over a seven day period.
As noted above, the SWATF disbanded in advance of the UNTAG
 mission and, though its arms were stored on South African
 military bases, the UN had difficulty gaining access to these
 facilities to verify their contents.14
 Despite continued reports of intimidation and covert South
 African assistance to an anti-SWAPO coalition, the elections went
 ahead on schedule. With nearly all of eligible Namibians voting,
 SWAPO was elected to power with 57% support to rival DTA's 28.5%,
 while the smaller parties picked up the rest of the votes.15 The
 UNTAG mission, the UN's first significant post-Cold War peace
 support operation, was viewed as a success.
 Angola and UNAVEM II
 In the case of Angola, the UN peace support operation was
 reconstituted three times, in the first instance because its
 mission was extended and later due to the failure of the previous
 effort. The United Nations Angola Verification Mission (UNAVEM)
 I, running from 1989 to 1991, was the complement to the UNTAG
 mission in Namibia, monitoring the withdrawal of Cuban troops
 from Angola. UNAVEM II, created in the wake of the successful
 implementation of both UNAVEM I and UNTAG, was a more ambitious
 mission that undertook both a monitoring and, in some instances,
 facilitating role in what was to have been Angola's transition
 from war to peace. UNAVEM II's failure, exacerbated by a host
 of mitigating circumstances, cast a shadow over the UN's ability
 to conduct multi-dimensional peace support operations.
 A. The Peace Agreement and Demilitarization
 The signing of the New York Accords in December 1988, which
 included provisions for the withdrawal of Cuban troops from
 Angola and South African forces from Namibia, was to see the
 first direct UN involvement in Angola. Responding to a request
 by the Angolan and Cuban governments, on 20 December 1988 the
 Security Council passed SC Resolution 626 (1988) establishing a
 United Nations mission to verify the withdrawal process in
 Angola. In this capacity, the United Nations Angola Verification
 Mission (UNAVEM I) oversaw the departure of 50, 000 Cuban
 soldiers, a process which was scheduled to take 3 0'months. In
 order to complete its assignment, UNAVEM I was assigned a team
 totalling 70 military observers under the command of Brigadier
 General Pericles Ferreira Gomes, the Chief Military Observer.
 The first contingent of military observers arrived on 10 January
 1991 and they began taking up positions across the country.
 Working from January 1989 to May 1991, the UN mission was able
 to complete the pullout of Cuban troops ahead of the established
 time table of July 1991.
 Concurrently, peace talks were held under the auspices of the
 Portuguese government with the introduction of representation
 from the United States and the Soviet Union in the later stages
 of discussion. With the failure of the Angolan government's 1991
 offensive against the Unita stronghold in the southeastern corner
of the country underscored the inability of either party to
 secure victory through force of arms. At the sixth round of the
 talks in early, the Angolan parties finally agreed to a
 comprehensive peace settlement which included a cease fire,
 demobilization and the country's first democratic elections.
 The Acordos de Paz para Angola, or Bicesse Accords, were signed
 on 1 May 1991 in Portugal. They consisted of:
 * a cease fire between the Government and Unita set to
 take effect on 31 May 1991 and to be administered by
 a Joint Commission staffed by both parties and
 monitored by the UN;
 * the demobilization of Government and Unita troops
 and their integration into a unified national
 military, to be completed in advance of the elections;
 * elections to be scheduled between September and
 November 1992 and monitored by the international
 community.
 In a further agreement known as the "triple zero option", neither
 side was permitted to purchase new armaments and the
 international community (primarily in the form of the United
 States and the Soviet Union) was charged with ensuring that no
 transactions occurred.
 In order to implement the peace agreement, the Bicesse Accords
 established a series of commissions to administer the
 introduction of its provisions. A Joint Political-Military
 Commission (CCPM), staffed by senior representatives from the
 Government and Unita, was created to manage the overall
 implementation of peace agreement. The representatives from
 Portugal, the United States, and Russia were given observer
 status on the commission while the UN was confined to the
 position of "invited guest". Linked to it was the Joint
 Commission for the Verification and Control of the Cease-fire
 (CMVF) , set up to oversee the cease fire process. The Joint
 Commission on the Formation of the Angolan Armed Forces (CCFA)
 was the third institution created by the Bicesse Accord. In
 addition, a number of working groups charged with overseeing
 specific tasks, such as demobilization, de-mining, the police,
 and humanitarian assistance, were established under the auspices
 of the commissions .16
 The success of the withdrawal process encouraged the two main
 Angolan adversaries, the Government and Unita to request further
 UN assistance in the fulfilment of their own bi-lateral peace
 process. Though the UN had acted as an observer to the
 negotiations at Bicesse, its role in the structuring of the peace
 process had been severely circumscribed by one of the parties to
 the final agreement. Unita had wanted a strong UN presence in
 the country with substantial powers while the Angolan government,
 anxious to uphold its sovereignty, sought to limit the UN role.17
 The end result was that the position of the UN in the process
 8
would be, according to the plan, limited to monitoring and
 verifying a range of aspects of demilitarization of the conflict.
 These were:
 * the end of hostile actions between the Government
 and Unita, including propaganda by 15 May 1991;
 * the cantonment of all military forces to designated
 Assembly Areas by 1 August 1991;
 * the demobilization of all military forces, including
 the collection and disposal of weapons;
 * the creation of a new national army;
 * and, the neutrality of the national police force,
 especially with regard to its fulfilment of provisions
 for human rights.
 B. UNAVEM and Demilitarization
 The demilitarization programme in Angola was essentially short
 term in emphasis, though there was some provisions made for long
 term components. The short term programme focused on the
 demobilisation and disarmament of combatants through a jointed
 supervised process which included a limited UN role as monitors.
 This heavily circumscribed role for the UN was one of the factors
 which contributed to both the collapse of the demilitarization
 programme and the popular perception of overall UNAVEM failure.
 As for the long term components of demilitarisation, provisional
 plans were made for targeted programmes that undertook training
 of ex-combatants, but these were neither integrated into a broad-
 based approach to reintegration nor {with the outbreak of
 fighting) were they allowed to operate.
 UNAVEM's involvement in Angola's programme for demilitarisation
 centred on monitoring the cease fire, assuring the neutrality of
 the national police force, observing the process of
 demobilisation, and supporting the formation of a new national
 army. Marshalling its limited human and financial resources to
 this task, UNAVEM created teams of observers who were attached
 to counterpart teams of Government and Unita representatives
 ^established by the various commissions. Their role was to check
 if the Angolans were carrying out their work in a fair and
 unbiased fashion; there were no enforcement provisions as such
 provided for in the Bicesse Accords and certainly none directly
 available to UNAVEM. UNAVEM monitors were obliged to fan out
 across the six main regions of the country -- Luanda, Huambo,
 Lubango, Saurimo, Luena and Mavinga -- and take up positions in
 one of the eighty-two locations there.16 From this vantage, UN
 officials were to carry out their mission to monitor the
 fulfilment of the peace agreement.
 The verification of the cease fire was the responsibility of the
 (CMVF). To fulfil this mandate, the CMVF sent teams to each of
the fifty Assembly Areas and the thirty-two "critical points".
 While the Assembly Areas had, where possible, UNAVEM military
 observers attached to them, the designated "critical points",
 ports and other areas which could serve as conduits for lethal
 material, were permanently staffed by UNAVEM observers. In
 addition, UNAVEM developed a mobile reaction team to investigate
 violations in either those areas which it had no permanent
 presence or places outside of its established monitors.
 Monitoring the conduct of the National Police was another
 responsibility of the CMVF. The incapacity of UNAVEM to carry
 out its mission was perhaps most vividly underscored in this
 area. With only 89 monitors on hand, it was immediately obvious
 that UNAVEM would not be in a position to fulfil its stated
 duties. In a belated and inadequate response to these
 difficulties, the strength of the UN police observer contingent
 was increased in May 1992 to 126 personnel.19
 Demobilization was the responsibility of a working group formed
 by the CCPM. The programme for demobilization called for the
 cantonment of the Government's FAPLA and Unita's FALA forces at
 50 designated Assembly Areas by 1 August 1992. Once in place,
 the majority of the estimated 120,0 00 FAPLA troops and 65,000
 FALA troops would be disarmed and brought into civilian life.
 At the same time, an equal number of soldiers from both sides,
 numbering 50,000, would be assigned to the new army, the Forcas
 Armada, Angolanas (FAA). Training for those selected for service
 with the FAA would be the responsibility of Britain, France and
 Portugal, under the overall supervision of the CCPM. UNAVEM
 officials, operating in teams of two, were to be posted at each
 of the Assembly Areas where FAPLA and FALA were to gather.
 Notably, while some financial support targeting the demobilized
 soldier's integration into civilian life had been mooted amongst
 some elements of the international donor community, no concrete
 provisions for its dispersement had been developed at this late
 stage.
 C. Implementation
 The process established by UNAVEM for the investigation of
 allegedly cease fire violations proved to be inadequate.
 Compounding problems was the slow deployment of UNAVEM observers,
 caused both by delays in New York and difficulties created by the,
 two parties, which meant that the verification of demobilisation
 and disarmament was not fully operational until 3 0 September
 1991. All in all, there were sixteen incidents which, according
 the UNAVEM officials, in themselves could have resulted in a
 renewal of significant conflict.20 Nonetheless, open hostilities
 were by and large suspended during the build up to the elections
 in late September 1992.
 The neutrality of the national police force was a source of major
 dissent between the government and Unita. The creation of the
 Rapid Intervention Police, popularly known as the "Ninjas",
 invoked severe criticism from Unita officials as they were
 outside of the established demilitarisation programme. Trained
 10
by the notorious Spanish Guarda Civil, the Rapid Intervention
 Police were composed of several military units drawn from the
 army and the security forces and numbered approximately 4,000.
 The actual demobilization of Angolan soldiers was a dismal
 failure. Conditions in the Assembly Areas were appalling,
 lacking all the pre-requisites for such an operation such as
 proper shelter, food and water. Rioting Government soldiers,
 protesting the absence of basic necessities in the Assembly Areas
 and the overall slowness of the demobilization process,
 threatened to demobilize "spontaneously". The unwillingness of
 the Government to provide the requisite transportation and
 foodstuffs for the assembled troops proved to be a major
 stumbling block in the demobilization process. Efforts by UNAVEM
 military observers to develop an accurate count of the number of
 demobilized soldiers were hampered to the extreme by the shortage
 of UN personnel. Insufficient controls for monitoring the
 movement of troops, causing UNAVEM military observers to resort
 to weekly estimates of encamped troops, rendered UN monitoring
 irrelevant.
 In contrast to FAPLA, Unita forces remained largely disciplined
 and under central control. However, the leadership proved to be
 generally unwilling to engage its troops in the demobilization
 process. Unita cited the lack of preparation on the part of the
 international community for the integration of its troops into
 civilian life as a reason for withholding from the process.21
 The difficulties of gaining access to Unita-controlled territory
 further impeded UNAVEM in its efforts to keep abreast of the
 situation in the field.
 Despite these shortcomings, by May 1992 70 per cent of the
 160,000 soldiers were in the Assembly Areas but of these, only
 6,000 had actually been demobilized.22 By June the figure had
 increased to 20,000, with 85% of Unita troops in place and 37%
 of Government forces in the Assembly Areas. By 7 October (a week
 after the election) UNAVEM officials claimed that 96,620 FAPLA
 troops had been demobilised.23
 Accusations and counter accusations of violation of the
 demobilization process increased as the deadline for disarming
 Angolan troops neared. As noted above, Unita claimed that the
 Government converted several military units into paramilitary
 units and placed them under the control of the national police.
 At the same time, the Government reported that Unita was
 illegally holding 20,000 troops in reserve in Cuango Cubango
 province. A UNAVEM investigation team subsequently verified
 that several hundred unaccounted FALA troops were in fact in that
 area but, due to the lack of UN personnel, were unable to bring
 them into the demobilization process.24
 Linked to the problems of demobilization were obstacles hampering
 the establishment of the new national military. These included
 lack of facilities (adequate or otherwise) for soldiers,
 apparently a result of both the illegal sale of government
 material by the military and the re-allocation of resources to
 11
training and equipping the Rapid Intervention Police.25 Despite
 the failure to move forward on establishing a new national army,
 in order to maintain the facade of adherence to the Bicesse
 Accords, the Government and Unita held an official ceremony on
 2 7 September inaugurating the FAA.26 As late as October, only
 two of the sixteen were rehabilitated for use and 8,800 soldiers
 had been integrated into the new national army.
 Against this increasingly somber background, over 90% of the
 eligible Angolans voted in the election, with dos Santos winning
 49.6% to Savimbi's 40% in the presidential race, while the MPLA
 secured 54% to Unita's 34% for seats in the legislature. Short
 of the requisite 50% for the presidency, the two leaders were
 obliged to conduct a run off election; this was never to
 happened. Savimbi, claiming fraud, remobilized his troops and
 began a military campaign to consolidate UNITA positions in early
 October. Within a few weeks, Luanda itself was gripped by
 fighting as the government launched a counter strike against
 Unita. Persistent efforts by UN officials to broker a cease fire
 between the two warring Angolan factions came to naught and, by
 the end of January 1993, the Secretary General observed that a
 state of civil war existed in Angola. The continued violations
 of truces forced the downsizing of UNAVEM II and with resignation
 of the Special Representative in May 1993.
 Mozambique and ONUMOZ
 The demilitarisation programme in Mozambique, in contrast to the
 Namibian and Angolan examples, adopted an approach which
 deliberately sought to link short term objectives with long term
 goals. Utilizing the international NGO community to a greater
 extent than in previous operations, and cognizant of the recent
 debacle in Angola, the UN mission to Mozambique {ONUMOZ) put
 together a coordinated effort which rivalled UNTAG in size but
 far exceeded it in breadth.
 A. The Peace Agreement and Demilitarisation
 The General Peace Agreement (GPA) signed by the Government of
 Mozambique and Renamo in October 1992 called specifically for UN
 participation in the areas of the monitoring of the cease fire,
 providing humanitarian assistance and monitoring of the
 elections. The SRSG, Aldo Ajello, in cooperation with the
 Western powers which had been party to the negotiations in Rome,
 convinced the Government and Renamo to formally establish the
 Supervising and Monitoring Commission (CSC) as the central
 authority overseeing the implementation of the GPA. The CSC's
 mandate included the settling of disputes between the parties,
 any question of interpretation of the GPA and a coordinating role
 for the subsidiary commissions to be established. The Cease Fire
 Commission, the Commission for the Reintegration of Demobilizing
 Military Personnel and the Joint Commission for the Formation of
 the Mozambican Defence Force were created to managed specific
 aspects of demilitarization.
 12
The time table set by the parties to the GPA was as follows:
 * the cease fire to commence 15 October 1993;
 * demobilisation to be completed by 1 April 1993
 and the subsequent formation of a new national
 army;
 * and elections to be held in October 1993.21
 B. ONUMOZ and Demilitarization
 The demilitarization programme employed by ONUMOZ sought to
 integrate the international community's efforts in supporting
 short and long term demilitarization. Short term components
 included the monitoring of the cease fire and the supervision of
 the demobilization of combatants. Long term components focused
 on projects which targeted ex-combatants, including provisions
 for a two year subsidy, job referal and training programmes.
 Underlying the UN's demilitarization plan was the desire to
 separate the demobilization process from that of the electoral
 process.28 And, in contrast to UNAVEM II, sufficient financial
 and human resources were made available to fulfil these
 ambitions.
 Monitoring the withdrawal of foreign troops from Mozambique, a
 condition Renamo had insisted upon at Rome, was the first task
 of the UN demilitarization programme in Mozambique. It was
 decided to bring 7,500 UN peacekeeping troops into the country
 to monitor their departure and take up positions along the
 transport corridors. The Cease Fire Commission (CCF), composed
 of the Mozambican parties, representatives of the international
 community and members of ONUMOZ, worked to ensure that peace was
 maintained in the rest of the territory. The terms of the
 Commission were specific: it was to investigate allegations
 brought to it of violations of the cease fire agreement signed
 in Rome.
 The demobilization of combatants formed the second part of
 ONUMOZ's involvement in demilitarization. A specialized unit
 attached to the mission planned and implemented demobilization.
 A team of four UN personnel were assigned to each of the 49
 Assembly Areas where they were to oversee the process, including
 the registration of combatants, disarmament, the selection of
 soldiers for the new army and formal demobilization. Education,
 entertainment and general logistical arrangements were also their
 responsibility.
 The last short term component of demilitarization was the
 creation of a new national army. It was initially envisaged that
 the new army would consist of 30,000 soldiers, equally divided
 between former Government and former Renamo troops and trained
 by the British, French and Portuguese.29 The Joint' Commission
 for the Formation of the Mozambican Defence Force was to be the
 point of intersection between the Mozambican parties and key
 members of the international donor community.
 13
The long term component of demilitarization, namely the
 Information Referral Service (IRS) and the Reintegration Support
 Scheme (RSS), fell under the auspices of the United Nations
 Office for Humanitarian Assistance Coordination (UNOHAC). These
 latter measures were to be integrated into the actions taken by
 the Commission for Reintegration (CORE), the institution
 established by the Rome Agreement to direct the process of re-
 integrating soldiers and refugees into civilian life. The IRS
 was conceived as a mechanism for providing demobilized soldiers
 with access to information on the job market as well as basic
 information on aspects of the reintegration programme.
 Originally falling under the auspices of CORE (which proved to
 be ineffective), its eleven provincial offices were actually run
 by IOM.30 Using a Trust Fund administered by the UNDP, the RSS
 was to provide demobilized soldiers with eighteen months of
 subsidies in the form of cash disbursements given at local
 branches of the Banco Popular de Desenvolvimento.31 By providing
 a reasonable assurance of financial support for an extended
 period of time, it was hoped that the former combatants would
 find employment in their districts and, concurrently, integrate
 into the local community. To assist in this process, vocational
 kits consisting of agricultural tools, seeds, and food rations
 for up to three months were given to de-mobilized soldiers upon
 departure from the Assembly Areas. UNOHAC used a host of
 development and refugee support agencies to implement these
 programmes. For example, in the area of de-militarization,
 UNOHAC contracted out to the International Organization for
 Migration (IOM) to organize transportation of demobilized
 soldiers while for long term assistance, UNOHAC turned to the
 United Nations Development Programme (UNDP) who, in partnership
 with the Ministry of Finance and IOM, put into play the RSS.
 Finally, the monitoring of conduct of the election campaign was
 the responsibility of the UN's Civpol. Numbering 128 {later
 expanded to 1,114), Civpol was attached to Mozambican police
 units to assure that the electoral process was conducted in an
 atmosphere free of intimidation and, concurrently, that human
 rights of citizens were respected.
 C. Implementation32
 With a six month delay in bringing the mission up to full
 strength in Mozambique, a new time table had to be set which
 called for the concentration of troops in the Assembly Areas to
 begin in September 1993 with full de-mobilization completed in
 May 1994; the new army to be operational by September 1994; and
 the election to take place in October 1994.
 The initial step in implementing the demilitarization programme
 was to bring in the UN peace keeping forces to monitor the
 withdrawal of Zimbabwean and Malawian troops. Despite delays
 that held up their introduction until August 1993, a total of
 6,000 peacekeepers were finally put into place. Concurrent with
 the introduction of UN troops was the monitoring of the cease
 fire between the Government and Renamo by the CCF. As the
 majority of violations involved unauthorized troop movements
 14
rather than shooting incidents, the task of monitoring was
 confined to assessing claims of position and territory.33 In the
 interim, the GPA mandated separation of forces was abandoned as
 adherence to it would have delayed the peace process further.
 The next phase of demilitarization, the demobilization of
 Government and Renamo troops, proved to be exceedingly
 problematic. In an especially ambiguous passage, the GPA had
 allowed for "dual administration" of territory controlled by the
 Government and Renamo (which was to be reconciled before the
 elections through a joint commission on territorial
 administration), thus giving way to conflicts over everything
 from internal movement to taxation. In addition, the clash over
 siting of Assembly Areas, which fell victim to strategic
 manoeuvring the Government and Renamo, was only brought to a
 close through a compromise brokered by the SRSG.34 As a result
 it was not until November 1993- that twenty of the 49 Assembly
 Areas were officially opened to receive troops while the rest did
 not become operational until February 1994.3S The slow pace of
 demobilization, a product of lengthy indecision and confrontation
 by the Mozambican parties, fuelled dissent amongst combatants
 housed within the Assembly Areas. Months of confinement in the
 monotony of the camps gave way to demonstrations and riots,
 targeting both local military and UN officials.36
 With pressure to commence the demobilization from both the
 international donor community and the soldiers themselves,
 spurred in part by the approach of newly established deadlines
 for the mission, ONUMOZ decided that the first demobilizations
 should take place mid March 1994. The assembly phase was
 completed on 15 August, in spite of uneven cooperation from the
 Government and Renamo, with the final total of registered
 soldiers being over 86,000.37 Nevertheless, approximately 5,000
 Government troops and 2,000 Renamo troops remained outside of the
 official demobilization process.38
 The establishment of the Forcas Armadas de Defensa de Mocambique
 (FADM) was another contentious aspect of the demilitarization
 programme. Training of trainers was slow to start, while joint
 command of the new army was only agreed upon in January 1994 and
 80 officers were appointed in June to command the newly created
 infantry battalions. Delays in the supply of new equipment and
 the renovation of inadequate training facilities, coupled with
 the prolonged process of identifying new soldiers forced the
 compression of training into six weeks. Unhappiness over the
 prospect of being forced to continue in the military brought
 about strikes and desertions.39 By election time in October
 1994, less than 10,000 soldiers had completed their training and
 Mozambican officials were lowering the target size for the FADM
 to 15,000 at arms.
 The long term component of demilitarization, introducing measures
 for the long term maintenance of the demobilized troops, was
 taken up by UNOHAC's IRS and its RSS. With the able assistance
 of IOM (which played the principal role in reintegration as CORE
 failed to materialize), the transportation of demobilized
 15
soldiers and their dependents to their chosen destinations
 occurred with minimum problems. The subsidy scheme commenced
 without significant hitches while the IRS offices received
 numerous enquiries and assisted in job placement.
 The elections of October 1994 were marred by one last crisis.
 The Renamo leader, Afonso Dhlakama, pulled his party out hours
 before the polling was to begin; however, concerted pressure on
 the part of the Western powers and the SADC leadership caused him
 to reverse his decision. The result was a turnout of over 85%
 and the election of the Government's candidate, Joaquim Chissano,
 to the presidency by a margin of 53% while Dhlakama received 33%.
 In the legislative elections, the Government party, Frelimo, won
 129 seats to Renamo's 112, with the rest of the seats going to
 smaller parties. Declaring the mission a resounding success,
 ONUMOZ withdrew from Mozambique by late December.
 III. The Future of Demilitarization and UN Peace Building
 If the Southern African cases cited here are to serve as any
 guide, it is principally to suggest that the issue of the
 military and the UN's role in the demilitarization process
 deserves renewed consideration on a number of levels.
 First, demilitarization knows no boundaries. To cite but a few
 examples, some of the demobilized soldiers from Koevoet were to
 re-appear as Unita soldiers after the collapse of the Angola
 peace agreement. In Mozambique, official disarmament proved to
 be hopelessly inadequate, with a further 22,000 arms unearthed
 from amongst 146 undeclared weapons caches, and many arms finding
 their way into neighbouring South Africa and Zimbabwe.40 The
 successful transformation of the regional heritage of conflict,
 especially those issues of a military nature, requires an
 atmosphere of active and institutionalized cooperation between
 the states of Southern Africa. An approach which is predicated
 upon an understanding of economic assymetry and porous borders
 in the region would in all likelihood stand a better chance of
 minimizing post-conflict problems than one which treats each
 conflict as a discrete, state-based situation. To cite two
 examples, programmes which seek to coordinate the disarmament
 phase of demilitarization in one area with an increase in cross-
 border surveillance would, hopefully, act to stem the traffic in
 small arms; equally, it is reasonable to assume that economies
 of scale could be achieve in developing a regional approach to
 the costly task of demining. The preliminary structures emerging
 out of the SADC framework are a significant step towards
 providing a forum for devising regional strategies to the address
 post-conflict management issues.41 In this regard, it is
 possible in future settings for the international community and
 in particular the UN system, through the initiation of a short
 and long term demilitarization programme integrated across state
 boundaries, to make a stronger commitment to the promotion of an
 enabling environment for tackling the post-conflict problems of
 demilitarization on a regional basis.
 16
Second, implementation of demilitarization programmes requires
 not only conceptual clarity as to the emergency/development
 continuum but ongoing coordination between the UN, the
 international donor and NGO community. Clarity between emergency
 and development is, with a few exceptions, well understood in
 terms of actual conceptualization of demilitarisation programmes;
 this can be readily seen in the sophisticated approach to
 demilitarisation applied during the ONUMOZ mission. However,
 translating that into a coordinated approach that effectively
 pools the resources of the relevant UN and non-UN agencies as
 well as the host government and donors, is still proving to be
 problematic. In fact, while' the three case studies cited here
 seem to present a picture of deepening cooperation and
 coordination both within the UN system and the NGO community,
 events in Angola indicate a disturbing trend away from that
 approach. UCAH, the local coordinating body established by the
 Department of Humanitarian Affairs in Luanda, maintains its
 distance from UNAVEM III while the NGOs are at pains to ensure
 their independence of action from UCAH.42 Enduring institutional
 biases which range from considerations of mandate to more prosaic
 matters keep the international community from acting effectively
 in collaborating together in demilitarisation. Furthermore, the
 perchance for UN and NGOs to introduce elements into
 demilitarisation programmes which can only be construed as self-
 serving {e.g., promoting the inclusion of a component into the
 demilitarization programme whose sole virtue is that it matches
 the institution's selected expertise), endangers what is already
 a volatile situation.
 And third, further analysis of the long term component of
 demilitarisation needs to be undertaken. Studies by the World
 Bank and private consultancy firms do not as yet provide
 sufficient evidence to state unequivocally that the utilization
 of a targeted approach to reintegration is the most efficatious
 policy.43 These preliminary findings, especially as they rely
 heavily on case studies whose reintegration programmes are very
 much in progress, should be treated with caution. The selection
 of target groups, which spill over into the post-conflict
 development process, can act to re-affirm the ex-combatants
 identity and give him a sense of entitlement and expectations
 that is beyond the means of the post-conflict state. At its most
 egregious, the Namibian government's belated formulation of
 Development Brigades, with their severe cost implications and
 attendant social problems, is a marked failure of targeting.44
 At the same time, it is readily understood that for many
 soldiers, resort to arms -- whether in the form of military
 service or banditry -- is the only real means of "employment"
 available to them. A programme which actively responds to the
 immediate financial and employment deficencies of former
 combatants while recognizing the longer term imperative of social
 integration would be ideal. An earlier version of Mozambique's
 Reintegration Support Scheme, which provided incremental
 financial assistance over a set time period and linked such
 support, effectively, to a fixed domicile, was one effort to
 couple targeting concerns with the promotion of integration into
 the local community.4S Still, given the centrality of
 17
demilitarisation to the peace process and its sustainability, it
 behooves the international community to develop a better
 understanding of this process.
 With the future of multi-dimensional peace support operations of
 the kind cited in this paper in question, perhaps it is important
 to re-affirm a central point about the UN and demilitarisation.
 Succintly put, should UN peace support operations be involved in
 the long term component of demilitarization? The answer is a
 qualified yes. The overlap between short and long term
 components of demilitarisation, whether it is in the area of the
 linkage of the demobilisation process to reintegration measures
 or developing local capacity to manage long term reintegration
 projects, necessitates integrated planning at the earliest
 possible moment in the demilitarisation process. And, it is
 logical to assume that a coordinating authority such as the UN
 would serve as an integral part of that process. However, with
 its uneven record in Southern Africa and elsewhere, the logic of
 the centrality of UN involvement is less appealing. For
 demilitarisation to succeed against all odds, and the post-
 conflict state is rife with crippling problems, it needs a better
 UN than the one we have today.
 18
ENDNOTES
 1. Letter to the Editor, The Namibian 23 July 1990.
 2. Boutros Boutros Ghali, Agenda for Peace 1995, New York: United
 Nations, 1995, pp.12-29.
 3. See the World Bank, Demobilization and Reintegration of
 Military Personnel in Africa: The Evidence From Seven Country
 Case Studies, Africa Regional Studies Report No. IDP-130,
 Washington, DC: World Bank, 1993; Chris Alden, "Swords into
 Plowshares? The United Nations and Demilitarization in
 Mozambique," International Peacekeeping, Summer 1995, pp.175-176.
 4. Boutros Ghali, op.cit., p. 61.
 5. Undoubtedly the clearest expression of this overlap between
 the immediate concerns of peacekeeping and those of development
 can be found in the area of demining.
 6. Interview with Aldo Ajello, 16 September 1994.
 7. It should be noted that disarmament and confinement to bases
 was organized through UNAVEM I while repatriation was linked to
 the UNHCR's programme.
 8. UNTAG: Description and Analysis of the Mission's Operational
 Arrangements 9 September 1991 (draft), pp.265-266.
 9. Ibid., pp.263-268.
 10. Private communication.
 11. The Tripartite Commission was established by the Brazzaville
 Protocol of 1988 and included South Africa, Angola and Cuba, with
 the United States and the Soviet Union serving as observers.
 12. UNTAG, op.cit., pp.208-209.
 13. V. Fortna, "United Nations Transition Assistance Group in
 Namibia", in W. Durch, ed. , The Evolution of UN Peacekeeping:
 Case Studies and Comparative Analysis, New York: St. Martins
 Press 1993, p.359.
 14. Interview with Omar Halim, Deputy Chief of Staff, UNTAG, 16
 March 1996.
 15. AWEPA, Consolidation of Democracy in Namibia, AWEPA Electoral
 Observer Mission, December 1994, Amsterdam: African-European
 Institute, 1995, pp.7-9.
 16. A. Vines, "Angola and Mozambique: The Aftermath of Conflict",
 Conflict Studies 280, RISCT, London, May/June 1995, p. 12.
 17. M. Anstee, "Angola: the Forgotten Tragedy --a Test Case for
 U.N. Peacekeeping," International Relations vol XI, No. 6,
 19
December 1993, p.497.
 18. Fortna, op.cit., p.3 95.
 19. "United Nations Angola Verification Mission II," United
 Nations Peace-Keeping Operations: Information Notes 1993 Update
 No. 2, New York: United Nations November, 1993, p.26.
 20. T. Ohlson and S. Stedman, The New is Not Yet Born: Conflict
 Resolution in Southern Africa, Washington, D.C.: Brookings, 1994,
 p.lll.
 21. Vines, op.cit., p.9.
 22. Ibid., p.110.
 23. "United Nations Angola Verification Mission II," United
 Nations Peace-Keeping Operations: Information Notes 1993 Update
 No. 2, New York: United Nations, November 1993, p.21.
 24. World Bank, op.cit., p.41.
 25. Vines, op.cit., p.10.
 26. Human Rights Watch/Africa, Angola: Arms Trade and Violations
 of the Laws of War Since the 1992 Elections, New York: Human
 Rights Watch, 1994, p.13.
 27. AWEPA, General Peace Agreement 1992, Amsterdam: African-
 European Institute, 1992, pp.34-36, 42-44, 48-50, 56-64.
 28. Interview with Ton Pardoel, Chief Technical Unit Officer, 26
 May 1994.
 29. AWEPA, General Peace Agreement 1992 Amsterdam: AWEPA/African-
 European Institute, 1992, p.30.
 30. IOM press release, Maputo, 24 May 1994.
 31. Each Government and Renamo soldier received six months of
 their regular salary plus bonuses (with a minimum of 75,000
 meticais) from the Government, with half of that given to them
 at the point of official demobilisation and half given to them
 in the district of resettlement. Upon completion of the
 Government subsidy programme, the a donor established provided
 a further eighteen months of support.
 32. For a full account, see C. Alden, "Swords into Plowshares?
 The United Nations and Demilitarization in Mozambique"
 International Peacekeeping 2:2, Summer 1995.
 33. AWEPA, Mozambigue Peace Process Bulletin, August 1993, No.
 5, p. 5.
 34. J.P. Borges Coelho. and A. Vines, "Pilot Study on
 Demobilization and Re-integration of Ex-Combatants in
 20
Mozambique," Refugee Studies Programme, Oxford University 1995,
 pp.3-6.
 35. An additional difficulty was the issue of the Government's
 para-military forces, something not adequately addressed in the
 GPA. Estimated to number 155,000, Government militias were
 scattered across the rural areas, often only nominally under the
 authority of district or provincial officials. It was only in
 January 1994 did they actually begin to disarm with two-thirds
 of their number demobilized by July. Interview with Colonel Pier
 Segala, Cease Fire Commission, 14 September 1994.
 36. CCF, "Problems/Incidents in Assembly Areas and Other Areas,"
 ONUMOZ, Maputo, September 1994.
 37. Africa Confidential 23 September 1994, vol 35:19, pp.3-4.
 38. Ibid., p.4; interviews with senior ONUMOZ officials.
 39. AWEPA, Mozambique Peace Process Bulletin, July 1994, No. 10,
 p.5.
 40. Africa Recovery, December 1994, p.14.
 41. The Inter-State Defence and Security Committee, which grew
 out of a Front Line States initiative, is one example; the
 controversial Association of Southern African States (ASAS) is
 another. J. Cilliers, "Towards Collaborative and Co-operative
 Security in Southern Africa: The OAU and SADC," in J. Cilliers
 and M. Reichart, eds., About Turn: The Transformation of the
 South African Military and Intelligence, Halfway House, SA:
 Institute for Defence Policy, 1996, pp.208-216.
 42. Interview with senior UCAH officials in Luanda, 12 August
 1995; interview with NGO officials in Luanda, 12-14 August 1995.
 43. See for example World Bank, op.cit.; R. Muscat, "Conflict and
 ..Reconstruction: Roles for the World Bank", policy paper,
 Washington, D.C., 21 November 1995; Creative Associates, "Other
 Countries Experiences in Demobilisation and Reintegration of Ex-
 Combatants" , workshop proceedings and case study findings,
 Washington, D.C., March 1995.
 44. 0. Angula, "Development Brigades: New Deal to Train the
 Forgotten", Namibia: Development Briefing 2:10/11, April/May
 1993, p. 3; S. Shikangalah, "Development Brigades: The Namibian
 Experience", in J. Cilliers, ed., Dismissed: Demobilisation, and
 Reintearation of Former Combatants in Africa, Halfway House:
 Institute for Defence Policy, 1996, pp.70-71.
 45. This clause was changed in the final version, allowing ex-
 combatants to change the bank which administered their subsidy
 to accomodate those individuals who wished to move residence.
 21