Skip to main content

News and opinions

06 April 2026

Cosatu after 40 Part 1: The rise and fall of COSATU’s ‘one industry, one union’ aspiration

The new wave of trade unions that emerged in South Africa during the 1970s was marked by non-racialism, democracy, and worker control. The unions’ growth, slow at first but then accelerating, was due in part to their organisational effectiveness, the gains they were able to make for workers, and the anti-apartheid politics they adopted. By the early 1980s, the new independent unions had become a major force in the labour relations system and a key part of the fight for political democracy in South Africa.

Introduction

The new wave of trade unions that emerged in South Africa during the 1970s was marked by non-racialism, democracy, and worker control. The unions’ growth, slow at first but then accelerating, was due in part to their organisational effectiveness, the gains they were able to make for workers, and the anti-apartheid politics they adopted. By the early 1980s, the new independent unions had become a major force in the labour relations system and a key part of the fight for political democracy in South Africa.

An important moment in this phase of trade unionism in South Africa came with the formation of the Congress of South African Trade Unions (COSATU) in November 1985. COSATU brought many of the new trade unions together in a federation1 that was able to coordinate their strategies, build the strength of affiliated unions, and serve as a unified political voice. COSATU’s significance has been considered mainly in terms of its political impact, but an important and somewhat overlooked aspect of the federation was the model of trade unionism it was to pursue. This was articulated in COSATU’s founding principle of ‘one industry, one union’, which embodied the strategy to organise workers into broad-based industrial unions. This aim was not unique: COSATU’s intention to structure itself in this way followed the predominant global trend of industrial unionism, which in South Africa had previously been distorted by racial exclusion.

At present, COSATU remains committed to its founding principle of ‘one industry, one union’, although much has changed in the 40 years since it was adopted. While the goal to establish one union in every industry always faced challenges, in recent years what momentum there was has stalled as the boundaries of COSATU’s industrial unions have become blurred, competition between affiliates for members has risen, the practice of affiliates organising beyond their industry scope has led to divergence from the industrial union model, and the ability of the federation to manage relations between its member unions has declined. Although the federation has adopted resolutions that aim to ensure conformity to the ‘one industry, one union’ principle, affiliates have increasingly gravitated towards a multi-sectoral form of union organisation.

The challenges COSATU faces with respect to the industrial union model are not unique. The trend is well advanced in Europe. The particular confluence of sector and occupation, or bundle of occupations, that held together the ‘one sector, one union’ idea of the ‘industrial union’, be it in metal engineering, chemicals, textiles and clothing, construction or transport, was common in many European countries 30 or 40 years ago, but is today the exception. Industrial decline; the blurring of occupational profiles in many sectors and erosion of sectoral borders; union decline, union mergers and the rise of the multi-sector union; the loosening of sector agreements; and finally, the severing of the links that once held industrial unions and political parties together are all part of this transformation of modern-day trade unionism in Europe (Visser, 2012, pp. 130-131).

Some of the economic, technological, and political factors that have undermined industrial unionism in Europe can also be identified in South Africa. De-industrialisation has been a feature of the last three decades alongside the rapid growth of the services sector. The labour market has undergone changes marked by the externalisation of work (through outsourcing) as well as rising informalisation, while in the formal workplace the way work is organised continues to be in flux. These factors have accelerated the disintegration of industrial boundaries of some sectors, which has taken its toll on union organisation. These developments have corresponded with the ‘internal’ shifts within COSATU that have contributed to the goal of ‘one industry, one union’ receding as affiliates indulge in intensified competition for members and expand their boundaries across more and more industries.

COSATU is aware of these developments but is yet to develop a strategic response to the erosion of its industrial union model. In developed countries, unions have embarked on what is described in a growing literature as trade union revitalisation or transformation (see, inter alia, Frege & Kelly, 2003; Frege & Kelly, 2004; Ibsen & Tapia, 2017; Pillay, 2017; Visser, 2019; Peels & Mwamadzingo, 2022). Revitalisation has generally involved a strategic re-evaluation by unions of their structure and their organising models as they seek to reposition themselves and shore up declining membership. COSATU’s principled adherence to the industrial union model has arguably obscured how much has changed on the ground, to the extent that rather than responding to these shifts, the federation finds itself unable to pursue key mergers to completion or effectively manage demarcation disputes between affiliates.

These developments have implications beyond organising strategy. From its inception, and in parallel with the ‘one industry, one union’ model, COSATU has had a strong commitment to consolidating and strengthening centralised collective bargaining. Although organisation and collective bargaining have different dynamics, there is a synergy between the organisational structure of unions and the level at which bargaining takes place.

The Labour Relations Act, no 66 of 1995, provides for the establishment of bargaining councils that define their scope in relation to a geographical and a clearly defined industrial area. Concluding collective agreements, and extending those agreements to an industry and area, depends on the representativity of the employer and trade union organisations in the bargaining council. Changes to the organising scope of unions and to their membership within sectors usually impacts representativity, and therefore the ability of unions to improve conditions and terms of employment for their members (as well as non-members) through industry-wide collective bargaining.

A further factor that has effected a major change to the ‘shape’ of COSATU, and the continued relevance of industrial unionism, is the huge growth of its public sector2 affiliates.3 This growth has occurred in a relatively short space of time and has mirrored and partly offset the shrinking of private sector affiliates, especially those in the manufacturing sector. It has resulted in a very different federation to the one established in 1985.

 

Overview of industrial unionism in South Africa from the 1970s

South Africa’s trade union movement historically reflected the divisions among the country’s workers and also reinforced the divisions in South African society (Kraak, 1993, p. 174). The Industrial Conciliation Act, no 11 of 1924, was the legislative cornerstone that divided workers and the trade unions they joined. The Act excluded African workers from its scope, which meant that a union with African members could not register and could not participate in an industrial council.4 The Act shaped the labour movement until the reforms that flowed from the Wiehahn Commission removed the exclusion of unions representing African workers from the ‘official’ labour relations system.5

At the time of the reforms there were 167 registered unions, the majority of which were members of one or more of the 13 federations, councils or joint committees that existed, with the remaining 61 unions unaffiliated. Race was the main factor dividing unions, with many unions and federations restricted to White members, while others organised White, Coloured and Asian workers or just Coloured and Asian workers. The result was a fragmented labour movement that excluded African workers, who made up the majority of the labour force. These racial restrictions and the changing demographics of the labour force ensured that few registered unions achieved a significant size. For example, the largest federation, the Trade Union Council of South Africa (TUCSA), which was open to unions organising White, Coloured and Asian workers, had only two unions with more than 20,000 members, and 45 affiliates representing fewer than 5,000 members. The main Whites-only federation, the South African Confederation of Labour (SACOL), with 23 affiliates, had a similar profile: it had three unions with more than 20,000 members and 14 with less than 5,000 members. Amongst the unaffiliated unions it was much the same: two had more than 20,000 members but there were 58 with less than 5,000 members (Miller, 1982, pp. xxiv-xxv; 122).

The union movement therefore predominantly comprised small unions, with a scattering of medium-sized unions and only a handful of relatively large unions. A factor that overlay this pattern was the intersection of skills with the racial demographics of the labour force from the earliest days of industrialisation. This resulted in many craft or quasi-craft unions, most of which represented only White workers, persisting well into the 20th century (Miller, 1982; Hendrie & Kooy, 1979, pp. 1-26).

There were few industrial unions, and most of those that existed had not organised a significant proportion of an industry. The exclusion of African workers, furthermore, meant the inevitable decline of the registered union movement as Africans became an increasing proportion of the total labour force.

Although unions representing African workers had been excluded from the legislated labour relations system, they had never been prohibited. Unions organising African workers, however, always operated in a harsh environment, rejected by almost all employers, opposed by many registered unions, and repressed by the government. Unions for African workers had therefore seldom been able to secure a strong foothold in workplaces or to forge collective bargaining arrangements, and membership fluctuated in the years before 1979. Friedman describes African unionism over this period as a series of waves that gathered and “threatened to bowl over all in their path. But they all broke harmlessly on the shore and were swept out to sea”(1987, p. 11).

This was the environment in which independent unions that arose in the wake of the 1973 strikes in Durban found themselves. They adopted a cautious approach by building factory level organisation in individual plants and factories, but most grew rapidly. Many of the new unions had a strong industrial focus, such as the Metal and Allied Workers’ Union (MAWU) and the National Union of Textile Workers (NUTW), but the approach of building union membership one factory at a time initially saw growth confined to one or two major urban centres.

Other unions had different structures. The Western Province General Workers’ Union (WPGWU) had more of a regional orientation and organised across industries, although it tended to focus on four or five industries where much of its membership was concentrated. A union such as the Food and Canning Workers’ Union (FCWU) was different. It was a registered union that pre-dated the 1973 strikes, but it had established a separate unregistered union for African workers, the African Food and Canning Workers’ Union (AFCWU). The two unions, although separate to comply with the legislative restriction on racially mixed unions, functioned much like a single union, with most of its membership in the Western Cape province (Kraak, 1993, p. 189; Theron, 2016, p. 43).

Another of the new unions, the South African Allied Workers’ Union (SAAWU), had a different structure. It described itself as a federation comprising unions in a number of industries, mainly situated in the Eastern Cape region. This federal structure was however never formalised and in practice SAAWU operated as a general union. Its growth was impressive: by 1980 SAAWU had approximately 75,000 members. Other general unions of significance were the General and Allied Workers’ Union (GAWU) and the Municipal and General Workers’ Union of South Africa (MGWUSA) (Kraak, 1993, pp. 190-194).

Almost from the moment the new unions emerged there were efforts at coordination and attempts to build unity. One such effort saw the formation of the Trade Union Advisory Coordinating Council (TUACC), which aimed to support the activities of new unions through sharing resources and adopting common policies. This initiative led to the establishment of the Federation of South African Trade Unions (FOSATU) in 1979, which brought together three registered and nine unregistered unions with a total of approximately 45,000 members (including, see above, MAWU and the NUTW) (Friedman, 2011, pp. 24-25).

FOSATU was explicitly a federation of industrial unions which aimed to establish unions in all industries on a national basis. Its organising model recognised that to effectively wield power, unions comprising mainly African workers had to organise a majority of workers within industries. It also recognised that such an organising model could divide workers along industry lines, but FOSATU aimed to counter such tendencies by facilitating coordination and cooperation between affiliates to maximise unity and power (Friedman, 2011, p. 67). The industry focus was, furthermore, balanced by the aim of the unions to remain open and inclusive of all Black workers.6

FOSATU’s industrial union organising strategy was successful: by 1981 its membership had more than doubled to 94,614 (Miller, 1982, p. 11). FOSATU was, however, still a relatively small player: union membership for the entire labour movement stood at 1,225,454 in 1982 (Macun, 2002, p. 31). Although FOSATU was the leading proponent of industrial unionism, it was not the only home for industrial unions. The Council of Unions of South Africa (CUSA), formed in 1980, brought together a number of industry-based unions, including the National Union of Mineworkers (NUM), which went on to become the largest union in the country, the South African Chemical Workers’ Union (SACWU), the Transport and Allied Workers’ Union (TAWU), and the Building, Construction and Allied Workers’ Union (BCAWU). CUSA also followed a strategy of building unions on a foundation of plant-based, shop steward committees (Kraak, 1993, p. 188).

 

There were other organising models pursued by the new independent unions, while there were also differences regarding registration under the Labour Relations Act7 and whether to opt for collective bargaining within the industrial council system, not to mention political differences. The new unions were also competing with each other for members on the ground and there were splits over “‘poaching’ of members and demarcation of territory” (Kraak, 1993, p. 194). Achieving a coherent system of industrial unions was therefore not a foregone conclusion.

However, the differences and rivalry within the new union movement, along with continued state repression and a recession that caused widespread retrenchments, gave impetus to the search for greater unity. The process started with unity talks in Cape Town in August 1981, and the momentum was sustained by a series of meetings over the next few years. A feature of this phase was the mergers between many of the unions in anticipation of the establishment of a new federation. Ultimately, the process culminated in the formation of COSATU in November 1985. A number of founding principles were agreed to early in this process, one of which was the principle of ‘one industry, one union’ (Lewis, 1985, p. 33).

Following its establishment, the new federation’s implementation of the principle initially focused on a number of industries in which existing affiliates were to merge within a two-year deadline. The industries were: food and beverages; textiles, clothing, and leather; mining; paper, wood, and printing; metals, automobile components, and assembly; chemical and petroleum; commercial and catering; transport, cleaning, and security; local government; and domestic workers. It was also decided that new unions would be formed in the agriculture and construction sectors (Kraak, 1993, p. 194).

This phase of mergers proceeded slowly at first but then gathered momentum and by the end of 1988 only two mergers, in the retail and transport sectors, were outstanding. The merger of the Commercial, Catering, and Allied Workers’ Union of South Africa (CCAWUSA) and the Hotel and Restaurant Workers’ Union (HARWU) was delayed by divisions within CCAWUSA and it was only in 1990 that the South African Commercial, Catering, and Allied Workers’ Union (SACCAWU) was formed (Baskin, 1991, pp. 400-402). The other merger, between the Transport and General Workers’ Union (T&GWU) and the South African Railway and Harbour Workers’ Union (SARHWU), took much longer to finalise: the South African Transport and Allied Workers’ Union (SATAWU) was established only in 2000.

Despite these delays, the momentum behind the merger process following the establishment of COSATU resulted in remarkable progress. The original 33 affiliates were reduced to 12 industrial unions, which was in line with the industries COSATU had originally demarcated (Bennett, 1987, pp. 75-78). At the same time, membership increased significantly from 462,359 workers in 1985 (which was a third of all registered union members) to 1,258,853 workers in 1991 (46% of total registered union membership). This growth cemented COSATU’s position as the largest and most powerful union federation. The next biggest federation was the National Council of Trade Unions (NACTU) with 327,000 members in 23 affiliated unions (Macun & Frost, 1994, p. 77). NACTU had also committed to organising along industry lines.

At that time COSATU’s affiliated unions were concentrated mainly in the manufacturing sector, where it represented approximately 48% of the sector’s total workforce.8 NUM, which had joined COSATU after leaving CUSA, represented approximately 44% of those employed in the mining industry. COSATU was therefore structured along industrial lines but was also a federation that was predominantly organised in industry. By contrast, COSATU had a very limited presence in the public sector; in 1991 its three affiliates represented only 7% of the sector’s workforce (Macun & Frost, 1994, p. 78).9

While the unity talks that led to the formation of COSATU were taking place, the issue of whether to participate in the statutory industrial council system was being debated, mainly among FOSATU-affiliated unions. This led to MAWU and the NUTW deciding in 1983 to join industrial councils in their industries (subject to certain conditions) (Lewis, 1987, p. 174).10 After COSATU was formed more of its affiliates followed suit in the goods transport and the paper, wood, and chemical industries, which started to align the industrial scope of organisation in COSATU with collective bargaining arrangements. At this point there was no framework for centralised collective bargaining in the public service.

 

Challenges to the industrial union model

Managing mergers

The initial merger process that followed the launch of COSATU was not straightforward. Many of the merger talks made very slow progress and were marked by political differences between affiliates, the fears of individual unionists that they would lose their position and influence, and concerns amongst smaller unions that they would be swallowed up by the larger unions (Baskin, 1991, pp. 115-116). By its second national congress in July 1987, COSATU had not secured political and organisational unity and there were tensions within recently merged unions and in those that were still expected to merge (Baskin, 1991, p. 212).

One problem was that there was no definition of what constituted an ‘industry’, and although industries had been identified, their boundaries had not been clearly demarcated. According to Baskin, the solution had been to give “the Central Executive Committee (CEC) some flexibility by empowering it to ‘evaluate the viability of the sectors’” (1991, p. 115). There were soon a number of applications before the CEC, which decided to adopt a ‘make no exceptions’ approach, but seemed to apply this differently depending on whether the unions involved could reach agreement. So, the request for a separate cleaning and security sector was refused, as was an application to separate ‘commercial’ (i.e. wholesale and retail) from ‘catering’ (i.e. restaurants, etc.). In both these cases the unions directly involved did not agree on the requests to split the sectors (Baskin, 1991, p. 115).11 On the other hand, the CEC ratified an agreement between the relevant unions that local government workers should be organised separately from health and hospital workers.

The CEC used a different approach when there was a dispute over whether a very large enterprise, in this case Eskom, fell into the mining or metal engineering industry. The dispute was between NUM and MAWU,12 which both had members at Eskom. The CEC decided in favour of NUM on the basis that power stations were closely linked, both structurally and geographically, to the supply of coal by mines organised by NUM.13 While this dispute suggested one criterion for the definition of an industry, it was a dispute about the boundary of an industry rather than whether it could be split into two viable industries for the purpose of union organisation.

The notion of what an ‘industry’ was for the purposes of the industrial union model was therefore contested from the outset and competing logics emerged where the context or circumstances differed. This situation appears to have been perpetuated because of the CEC’s failure to develop a set of criteria or principles to inform such decisions or, alternatively, when it treated each case as unique and made its decision based on the specific circumstances and context. It is therefore not surprising that the demarcation of industries remains contested.

Two features of the early merger phase in COSATU were a clear end-goal and strict oversight of mergers to ensure that the goal would be achieved. The unity talks that preceded the formation of COSATU had agreed that the aim would be a ‘tightly-knit federation’ with large national industrial unions as the essential building blocks. Furthermore, the constitutional structures of COSATU, in particular the CEC, were premised on there being a small number of large industrial affiliates. Until the mergers took effect CEC meetings consisted of up to 100 representatives from a very diverse group of unions, which resulted in unwieldy, time-consuming, and frustrating meetings. Mergers would reduce the number of representatives in the CEC and make its work much more manageable.

The resolution on mergers was therefore of fundamental importance to the future of COSATU. These factors no doubt spurred the leadership to adopt what has been described as a ‘forced march’ approach to the merger process, albeit one that allowed for some flexibility (Baskin, 1991, p. 110). In Baskin’s assessment, this approach was a precondition for strong, national, industrial unions that would be able to “rationalise their operations” and provide a full range of services and support activities (e.g. research, education programmes, and a health and safety division). This would ensure the establishment of a “foundation for both a quantitative and qualitative growth in membership” (Baskin, 1991, p. 448).

The aim to focus on making COSATU into a ‘tightly-knit federation’ continued during the 1990s, but signs of a slowdown in the merger process became evident towards the end of the decade, and it subsequently went into reverse which was evidenced by an increase in the number of affiliates.

One reasons for this was the admission of new affiliates. COSATU originally channelled the admission of new unions through the established affiliate in the relevant industry (Baskin, 1991, p. 112), and the applicant union would directly enter merger negotiations with the established affiliate as the route to joining the federation. At some point, however, COSATU appears to have departed from this approach, for example, when it admitted the Public and Allied Workers’ Union of South Africa (PAWUSA) despite having existing affiliates in the public sector. This was a sign that the commitment to establishing a ‘tightly-knit federation’ was weakening, and the policy of ‘one industry, one union’ was the resulting casualty.

 

Similar backsliding can be seen early on in the public sector. When COSATU was launched it was decided to delay the formation of a single public sector union until organisation in the sector had strengthened (Baskin, 1991, p. 303). When, two years later, the National Education Health and Allied Workers’ Union (NEHAWU) was launched with a focus on only parts of the public sector (i.e. health and education), COSATU’s justification was that it was an “interim step” to the establishment of a single public sector union (Baskin, 1991, p. 198). Soon thereafter, however, the federation admitted the Post and Telecommunication Workers’ Association (POTWA) 14 and in 1990 the South African Democratic Teachers Union (SADTU) joined. This contradicted policies in two respects: first, it was a further step away from the aim of having one public sector union and second, with respect to SADTU, it breached the prohibition on occupation-specific unions.

The explanation by a former COSATU leader that “COSATU stressed industrial or clear sector unions, so NEHAWU was the broader public sector union and SADTU was the teachers’ union”, provides no clarity or confidence in the leadership.

Public sector unions grow rapidly

Public sector unions, including the COSATU affiliates, grew rapidly in the 1990s, especially after the passage of the new Labour Relations Act. The Act brought the public service within its scope and provided dedicated provisions for centralised collective bargaining structures in the public service. Between 1991 and 2012, COSATU’s public service unions increased their membership from 7% of total union membership to 39% and by 2013 their combined contribution to COSATU’s income stood at 38% (Bischoff & Maree, 2017, p. 170). But the growing significance of the public sector within COSATU, and the union movement as a whole, meant there was a need to accommodate a broad ‘sector’ with different internal contours and complexities, including dedicated statutes establishing employment conditions for different parts of the sector, not to mention some occupation-oriented unions. Furthermore, the increasing tendency of a number of these unions to organise in the private sector (e.g. private schools and hospitals) blurs the line between the private and public sectors and adds to the complexity regarding union boundaries.

The establishment by COSATU in 1988 of a public sector co-ordinating committee (PSCC) to strengthen unity efforts has not provided a solution (Baskin, 1991, p. 303); if anything, the problem has worsened. COSATU’s ability to enforce its policy to achieve the goal of a single public sector union has weakened as its public sector affiliates grew and their financial contribution to the federation increased. As one interviewee noted:

Because NEHAWU and SADTU [have] become strong now in the federation. I think they are the strongest unions in many ways, even financially. So, you can imagine for an institution like COSATU that relies heavily on subscription or affiliation fees from unions, it would not be an easy thing for them to manage.

As the goal of a single public sector union receded, the commitment to the ‘one industry, one union’ policy weakened in COSATU as a whole.

 

Private sector unions slide into multi-sectorialism

In the private sector an increasing number of COSATU affiliates have amended their constitutions to add new industries to their organising scopes. This often resulted in a tit-for-tat amendment by affected unions, for example, SACCAWU included the clothing and textile industry in its scope in response to the Southern African Clothing and Textile Workers’ Union (SACTWU) adding the retail sector to its scope. While these were not the only reasons for such developments, the result now is significant multi-sectoralism and overlaps in the organising scope of affiliates.15

An examination of the organising scopes of COSATU affiliates in 2022 found that most affiliates that organise primarily in the private sector have scopes that span at least four major sectors,16 while some organise in multiple industries but also have a catch-all clause in their constitutional scope which, on paper, makes them general unions (e.g. SATAWU has a provision that it is open to all workers, including atypical workers). There are, furthermore, a number of unions that have overlapping organising scopes (e.g. the Agricultural Food and Allied Democratic Workers’ Union [AFADWU] and SACTWU overlap in three sectors). Admittedly, this is the situation only ‘on paper’; the unions do not necessarily actively organise in all the sectors or might have only small numbers of members in ‘non-core’ sectors. However, these organising scopes are clearly not in line with COSATU’s objective of ‘one industry, one union’ and also open the door for expansion of organisation beyond an affiliate’s ‘core’ sector. It should be noted that, contrary to the international literature, which suggests that the shift from industrial unions to multi-sectoralism is driven by the need to recruit new members, in COSATU’s case this trend has been accompanied by declining membership.

The multi-sectoral trend is even more advanced in the rest of the union movement. An examination of the constitutions of a sample of 35 unions not affiliated to COSATU (out of 206 non-COSATU registered unions),17 found that 29 could be categorised as general unions (they either defined themselves as general unions, or have a catch-all proviso covering all workers, or their scopes include six or more major sectors). The sample included very big unions such as the Association of Mineworkers and Construction Union (AMCU), the National Union of Metalworkers of South Africa (NUMSA), the Food and Allied Workers’ Union (FAWU), and Solidarity; medium-sized unions such as the National Union of Public Service and Allied Workers (NUPSAW), the General Industries Workers’ Union of South Africa (GIWUSA) and the United Association of South Africa (UASA); and many very small unions with only a few hundred members.

The majority of these unions had not amended their organising scope to add sectors, that is, they had registered as general unions or as unions organising in six or more major sectors. Only ten unions had amended their original organising scopes to add more sectors, including AMCU, NUMSA,18 and SACWU, all of which were ostensibly industrial unions that have chosen to become multi-sectoral or general unions.

As important, in terms of the implications for the ‘one industry, one union’ policy, is that COSATU affiliates have got away with this expansion of organising scope. This is either because unions made the amendments without seeking endorsement from the CEC,19 in breach of a CEC standing resolution requiring unions to submit intended amendments to it, or the proposed amendment was submitted to the CEC and there were no objections to it despite it extending the scope of the union into other sectors in which there are existing affiliates.

Furthermore, some affiliates have attempted a more ad hoc expansion of their membership into workplaces outside their constitutional scope. Evidence of this practice is found in the rising number of organisational rights disputes referred to the Commission for Conciliation, Mediation and Arbitration (CCMA) as a result of employers having challenged union demands for organisational rights on the ground that the proposed members fell outside the constitutional scope of the union.

Where these disputes went on to the Labour Court, the judgments tended to go against the unions (Godfrey et al, 2021, pp. 687-690). Finally, in 2020 the Constitutional Court in NUMSA v Lufil Packaging confirmed that an employer was entitled to refuse organisational rights to a union in respect of a workplace that did not fall within the constitutional scope of the union. The judgment has likely slowed the ad hoc expansion of organisation to new sectors, but it has also had the effect of encouraging unions to amend their constitutions to add sectors to their organising scopes so as to pre-empt employers’ challenges.

 

The problem confronting unions is the apparent tension between the aim of inclusivity, that is, to provide a home to all workers, and the need to maintain the level of organisational integrity that the industrial union model requires. However, union leaders face a difficult choice if workers from another sector approach them seeking representation. Strict adherence to the COSATU policy means that the union leader must turn the workers away. In some instances, it might be possible to refer the workers to another COSATU affiliate in the relevant sector, but often workers do not want to join that union. In fact, their lack of confidence in that union is why they approached the other union leader, and they threaten that if he/she refuses to let them join the union they would prefer to go to a non-COSATU union. Union leaders generally believe that in such a situation the best option is to admit the workers to the union. After that they would deal with the difficulties of “how we service them and also … how we negotiate for them”. This is why, interviewees explained, “we have [members at] some companies that really do not fall in our scope”(Jacobs et al, 2023, p. 75). Union inefficiencies, or workers’ perceptions of inefficiencies, has therefore been a further factor contributing to the decline of the industrial union model in COSATU.

COSATU fragments

More dramatic consequences followed the decision by NUMSA in 2014 to broaden its constitutional organising scope to include, amongst others, the private security, construction, and industrial chemical industries. NUMSA explained that it intended to expand its scope so that it could pursue a new organising model based on value chains rather than industries. The view in COSATU, however, was that it would lead to increased poaching of members from other affiliates, therefore exacerbating a problem “that had festered for many years” (Craven, 2016, p. 121). There were a number of additional factors that exacerbated the highly public dispute, including political differences over whether COSATU should remain in an alliance with the ANC. The eventual outcome, the expulsion of NUMSA from COSATU, was a major setback for worker unity and the industrial union model (Bezuidenhout & Tshoaedi, 2017, p. xviii). COSATU lost its largest affiliate, which was followed by the departure of FAWU.

These splits in the federation gave additional impetus to COSATU’s shift to the public sector. One consequence was that NUM had to then extend its organising scope to the metal engineering industry to incorporate (and rescue) the Liberated Metalworkers’ Union of South Africa (LIMUSA), which was a union COSATU had established as a rival to NUMSA after its expulsion. NUM is now spread across the mining, construction, and manufacturing industries.

The related dismissal of COSATU’s General Secretary, Zwelinzima Vavi, followed a few months later. Vavi was preparing to spearhead the formation of the South Africa Federation of Trade Unions (SAFTU), which NUMSA and FAWU joined as founding members together with 22 other affiliates representing almost 700,000 workers in total (SAFTU, 2017). SAFTU’s affiliates are active in all the key industries in which COSATU unions operate and their continued rivalry with COSATU’s affiliates has significantly undermined any hope of the latter resurrecting its ‘one industry, one union’ objective.

The gradual unravelling of industrial unionism amongst the largest affiliates of COSATU was facilitated by changes in industry, including de-industrialisation, and the rapid growth of unionism in the public sector. It was given impetus by the gradual decline of what had been a strong centralisation of strategy and policy within COSATU’s early period and the intensification of political and organisational divisions among affiliates. Roughly three decades after its formation, the exit of key industrially based affiliates from COSATU has provided the grounds for greater competition between affiliates of different federations. These shifts require not only revitalisation within COSATU, especially in relation to its presence in the private sector, but also a rethink of its founding principles and strategic priorities.


REFERENCES

Baskin, J. (1991). Striking Back: A History of COSATU. Ravan Press.

Baskin, J. (1994). Discussion Paper: Centralised Bargaining and COSATU. National Labour and Economic Development Institute.

Bennett, M. (1987). COSATU’s two constituencies. Indicator South Africa, 5(1), pp. 75-78

Bezuidenhout, A. & Tshoaedi, M. (2017). Preface. In A. Bezuidenhout & M. Tshoaedi (Eds.), Labour Beyond COSATU – Mapping the rupture in South Africa’s labour landscape. Wits University Press, pp. xv-xxi.

Bischoff, C. & Maree, J. (2017). Public Sector Unions in COSATU. In A. Bezuidenhout & M. Tshoaedi (Eds.), Labour Beyond COSATU – Mapping the rupture in South Africa’s labour landscape. Wits University Press.

COSATU. (2018). Constitution, as amended by the 13th National Congress.

COSATU September Commission. (1997). https://mediadon.co.za/september-commission/

Craven, P. (2016). The Battle for COSATU: An insider’s view. Bookstorm.

Fergus, E. & Godfrey, S. (2016). Organising and bargaining across sectors in South Africa: Recent developments and potential problems. Industrial Law Journal, 37, pp. 2211-2236.

FOSATU Central Committee. (1982). Principles of Collective Bargaining. South African Labour Bulletin, 8(1), pp. 81-85.

Frege, C. M. & Kelly, J. (2003). Union Revitalization Strategies in Comparative Perspective. European Journal of Industrial Relations, 9(1), pp. 7-24.

Frege, C. M. & Kelly, J. (Eds.). (2004). Varieties of Unionism. Strategies for Union Revitalisation In a Globalizing Economy. Oxford University Press.

Friedman, M. (2011). ‘The Future is in the hands of Workers’: A History of FOSATU. Mutloatse Arts Heritage Trust.

Friedman, S. (1987). Building Tomorrow Today: African Workers in Trade Unions, 1970-1984. Ravan Press.

Godfrey, S., Jacobs, M. & Fergus, E. (2021). Multi-sectoral Workplaces and Trade Unions: The Unforeseen Consequences of Demarcation Principles and Organisational Rights. Industrial Law Journal, 42, pp. 669-693.

Hendrie, D. & Kooy, A. (1979). Trade unions in South Africa: Some statistics. University of Cape Town.

Ibsen, C. L. & Tapia, M. (2017). Trade union revitalisation: Where are we now? Where to next? Journal of Industrial Relations, 59(2), pp. 170-191.

Jacobs, M., Godfrey, S., Macun, I., Fergus, E. & Collier, D. (2021). Mapping and understanding the demarcation challenge.

Jacobs, M., Godfrey, S. & Macun, I. (2023). The changing sectoral boundaries of COSATU’s affiliates.

Kraak, G. (1993). Breaking the Chains. Pluto Press.

Lewis, J. (1985). Trade Union Unity. South African Labour Bulletin, 11(1), pp. 29-33.

Lewis, J. (1987). Overview: The Registration Debate and Industrial Councils.

Macun, I. (2002). The Dynamics of Union Growth in South Africa.

Macun, I. & Frost A. (1994). Living like there’s no tomorrow. Social Dynamics, 20, pp. 67-90.

Miller, S. (1982). Trade unions in South Africa 1970-1980.

National Union of Metalworkers of SA v Lufil Packaging (2020). Industrial Law Journal, 41, pp. 1846-1868.

Peels, R. & Mwamadzingo, M. (2022). Trade union revitalization.

Pillay, D. (2017). Trade union revitalisation in South Africa.

SAFTU. (2017). Inaugural Congress. https://saftu.org.za/launch-congress-2017

Theron, J. (2016). Solidarity Road. Jacana Media.

Visser, J. (2012). The rise and fall of industrial unionism. Transfer, 18(2), pp. 129-141.

Visser, J. (2019). Trade Unions in the Balance.


ENDNOTES

1 In this article we refer to COSATU as “the federation”.

2 We mainly use the term ‘public sector’ although the majority work in the public service.

3 This growth is not limited to COSATU.

4 An industrial council was a centralised collective bargaining structure.

5 The Wiehahn Commission was appointed in 1977.

6 That is, African, Coloured and Asian workers.

7 The new name for the Industrial Conciliation Act.

8 The unions were CWIU, FAWU, NUMSA, PPWAWU, and SACTWU.

9 The affiliates were NEHAWU, SADTU, and SAMWU.

10 See FOSATU Central Committee, 1982.

11 Baskin does not elaborate.

12 MAWU later became part of NUMSA.

13 NUMSA had not transferred its members at Eskom.

14 Now the Communication Workers Union (CWU).

15 Some constitutions allow extension beyond sectors.

16 Sectors defined at Standard Industrial Classification level.

17 Data from Registrar of Labour Relations.

18 NUMSA applied to amend scope.

19 COSATU leaders were unaware of some amendments.


BIOGRAPHIES

Shane Godfrey is retired but continues to actively research as an Honorary Research Associate at the University of Cape Town and Research Associate of the Centre for Transformative Regulation of Work (CENTROW) at the University of the Western Cape.

Mario Jacobs is a researcher and lecturer at CENTROW, Faculty of Law, University of the Western Cape.

Ian Macun is a former academic and public servant (in the Department of Labour), and is currently a part-time commissioner with the Commission for Conciliation, Mediation and Arbitration and an independent researcher.