The geographies of community-oriented unionism: scales, targets, sites and domains of union renewal in South Africa and beyond

Abstract

This article takes the increased interest in union renewal and community politics within the geography discipline as its starting point, and argues that the various cases described in the literature should be understood as socio-spatial responses to a new geography of work. Encouraged by the literature on contentious politics, I provide a conceptualisation of how workers reconfigure political spaces through community orientation. A concrete case of municipal unionism in Cape Town, South Africa is used to explore these geographies along four dimensions: the scalar strategies of trade unions, their targets for direct action, their sites of recruitment and the domain of mobilisation these campaigns and alliances stake out for themselves. It is argued that community-oriented unionism tends to entail a reorientation of each of these dimensions. Moreover, union strategies correspond to particular challenges faced in times when local labour markets undergo neoliberal restructuring, such as fragmentation of work and rescaling of labour relations.

Introduction

Work by labour geographers over the last 20 years has explored the relationship between collective action and space, although such work tends to be restricted to a certain kind of collective actor, i.e. organised labour. Early entries in the literature (e.g. Herod 1997; Martin et al. 1993), while ground-breaking in their own right, to some extent portrayed a traditional notion of the organised worker as male, white and employed in industrial work. Since then, labour geographers have addressed an array of worker subjects, including low-paid women, workers of colour and migrant workers (e.g. McDowell 2008; Pastor 2001). By doing so, they demonstrated labour geography’s commendable progress in transcending the traditional worker stereotype. Another step away from labour-centrism can be seen in studies of union engagement with other civil society organisations (e.g. Sziarto 2008; Wills 2001) and those exploring the links between workers and consumers in global production chains (Hale and Wills 2007; Hartwick 1998; Johns and Vural 2000). Much of this literature looks at trade unions, which in different ways try to renew their strategies, often under unfavourable conditions.

This article focuses on union renewal through community orientation. Community-oriented unionism is defined here as the range of ways in which unions mobilise non-members and communities on the basis of social disempowerment and overlapping political interests. Use of such a broad definition is intentional. Staeheli (2008) was right when she argued that it is problematic to make rigid academic claims about what community (or the community) is or why various actors choose to engage politically in community struggles. In labour geography and critical sociology, certain ideal types are often upheld: ‘social movement unionism’, which refers mostly to the national or international scale of mobilisation (and often in emerging economies; Fairbrother 2008; Moody 1997; Waterman 1999), and ‘community unionism’, which describes mobilisations based in the local labour market (Tattersall 2008; Tufts 1998; Wills 2001). Here, I argue that both these ideal types correspond to a set of tendencies discernible in the strategies and rhetoric of trade unions around the world. Few unions fully live up to the ideal. However, these tendencies can be observed in different contexts: always under construction, internally and externally contested and often evolving in a non-linear manner.

In this article, I argue that such community orientation tends to be a response to particular socio-spatial challenges facing workers and organised labour. I suggest a model that captures key tendencies found in the case study-based literature. These categories are then applied to the South African Municipal Workers’ Union (SAMWU) in Cape Town, as it tries to renew its organising strategies in the post-apartheid era. The text emanates from three additional motivations: first, I explore union–community interactions and experiences beyond the neoliberal core countries (UK, US and Australia) which hitherto have dominated these debates; second, I examine community orientation as tendentious in worker struggles around the world, with markedly different points of departure than those described in the extant literature; third, and related, instead of presenting a success story of ‘community unionism’, this case exemplifies how a re-orientation towards a more community-based approach is contested and unfolds a response to changes in the political landscape and in the labour market. In short, my argument is that community-oriented unionism has a geography to it, one related to forms of fragmentation and rescaling – in the workplace and in the labour market. Before I delve into the main argument of the article, I will use as a starting point the recent debate on the spatialities of contentious politics.

Organised labour and the spatialities of contentious politics

Thought-provoking interventions by Leitner et al. (2008) and Nicholls (2009) have explored how a geographical understanding can help us explain why political mobilisation follows certain trajectories and occupies certain spaces (and not others). Political actors are themselves spatial in the form of networks with certain mobilities and ways of organising their politics. Social movement organisations are the main subject in these accounts, and their spatiality is seen in relation to the organisation of the state. This has paved the way for a fruitful exchange between descriptions of horizontally organised networks, or rhizomes if you will (cf. Woods 2003), and the way in which these ‘contentious politics take[s] on state institutions, whose spatiality has traditionally been dominated by nested scales, ranging from the national to the local’ (Leitner et al. 2008). To achieve their political goals, social movements deploy particular scale frames that correspond with, or even challenge, the scalar organisation of the state. The spatialities of different actors intersect with each other to create ‘social movement spaces’ where activists forge ties in certain places. Some networks also demonstrate a mobility that allows them to bind different places together though social movement mobilisation, creating new political geographies (Nicholls 2009).

Leitner et al. (2008) stressed that the motivation for exploring the spatialities of contentious politics should not be to identify which spatial concepts – scale, space, mobilities, networks or assemblages – are appropriate for describing social movements and which are not. Rather, they seek to uncover how the complex spatialities of these politics co-implicate each other and how they interact with the spatialities of other political phenomena, in particular the political practices of the state at different scales. Here lies the point of entry of this article – looking at a case where union politics and community politics interact in a context of social change and state restructuring.

Studies of trade unions have tended to look at the world from a different perspective than that of social movement theory. In other words, the position we have ascribed to the key subject – organised labour – in the socio-political landscape differs markedly. In the Marxian critical geography tradition, labour as an abstract category is placed in the midst of the capital accumulation dynamic (Castree et al. 2004). Wherever we direct our analysis, we find labour is at the heart of capitalism. As Herod (2001) demonstrated, this insight can be translated into political practice; a relatively small group of striking workers can have a perplexing impact on global economic systems. However, when we start exploring the expanding and uneven nature of economic processes, it becomes evident that labour as a political phenomenon is much more diverse; different groups of workers are embedded in variegated and uneven ways in a globalising world economy (Coe and Jordhus-Lier 2010; Webster et al. 2008). While a Marxian approach to labour tends to locate it in relation to the capitalist production process, the politics of labour must be understood in a much wider sense (cf. Cumbers et al. 2008). The agency of organised labour is also differentiated through its ability to interact with other actors and other spheres of social life. Limiting our understanding to the capital–labour binary can constrain labour in an analytical sense but also in a practical political sense.

While labour geographers have been predominantly using scale as their spatial language, social movement scholars have been attentive to how contentious politics can be understood through multiple spatial concepts, paying attention to how activists themselves conceptualise their activities spatially and to the performative potential of this spatial language (Marston et al. 2005; McFarlane 2009; Nicholls 2009). That said, the two bodies of literature have much in common and are in my opinion complementary in nature. Both offer analyses firmly embedded in time and place. As Nicholls (2009) rightly pointed out, geographical proximity can allow social capital and strong ties to develop, even in political landscapes characterised by heterogeneity and translocal flows. This is also a prerequisite for union–community alliances to develop over time (Park 2007). Case studies in labour geography have explored local politics in metropolitan areas, thus providing a counterpoint to the literature on globalisation and transnational networks. While labour geographers have mainly been concerned with the link between recent neoliberalism and union renewal, Wills and Simms (2004) traced the history of community unionism in the UK through several stages: from initial community-based forms of unionism, led via a successful period of formal political representation during Fordism, to the reciprocal community unionism emerging in times of neoliberal restructuring.

Moreover, social movements and trade unions are both approached as relational phenomena. The spatial imaginary in the literature on contentious politics not only describes relations within networks and movements, but also relations with the state and the natural environment (Leitner et al. 2008; McFarlane 2009). This literature has, however, not specifically addressed trade unions. Labour geographers, on the other hand, have been increasingly concerned with the relationship between trade unions and other actors in civil society (Tufts 1998; Wills and Simms 2004). By drawing on this body of case studies, we can begin to ask whether instances of community unionism have dynamics in common and, taking a cue from the contentious politics literature, to explain these commonalities through different spatial terms. This is not an exercise to be done merely for its own sake, but because the community orientation visible in trade union movements across the world is inherently spatial. Redrawn union geographies have fundamentally altered local political landscapes and can be interpreted as responses to the changing economic and political geographies in which they unfold.

Socio-spatial challenges for organised labour: fragmentation and rescaling

In the literature, community-oriented union renewal is often situated in contexts characterised by neoliberal restructuring of public services (Lier and Stokke 2006; Moon and Brown 2001) or in the low-paid service sector of deindustrialising societies (Ryan and Herod 2006; Wills 2005). In Western societies, a shift in emphasis from the public to the private sector, and from industrial work to post-industrial service work, represents a move away from the strongholds of organised labour and into terrain in which trade unions are less experienced and labour regulation is less developed. Membership declines and weakened union power has followed as a result.

Neoliberalism has reconfigured the relations among the state, the market and civil society (Leitner et al. 2007) into what could be described as a new geography of work. This is observable at the global level, where labour geographers have argued that (actual or perceived) capital mobility has a disciplining effect on the political agency of workers in export-oriented industries (Peck 1996; Walker 1999). Other geographers have focused on the microgeographies of restructuring as they play out in the local labour market. One key concept in this regard is the fragmentation of workplaces and workforces (Reimer 1999). Fragmentation refers to all barriers that keep workers from constructing solidarity and shared identity. In some instances, fragmentation occurs across space, i.e. through the atomisation of workplaces (Rhee and Zabin 2009), or through outsourcing parts of a workforce. Simultaneously, fragmentation occurs in time, through shift work, irregular working hours and fixed-term contracts. Moreover, workers in the same workplace are fragmented in a contractual sense and exist across a continuum from formal, full-time employment to acute precariousness. In sum, the social spaces in which workers interact, forge ties of solidarity and recruit new union members are circumscribed and dispersed.

Fragmentation can also be used to describe changes in the organisation of capital: elusive networks of labour market intermediaries and direct employers make it challenging for workers to identify which party holds the employer responsibility (Benner amd Dean 2000). Even though the employer is accessible, power to influence workers’ conditions might lie with other actors in local or global production and distribution networks. Wills argued that as ‘subcontracted capitalism’ is becoming paradigmatic, ‘workers have to identify the “real employer” at the top of contracting chains’ (2009, 456). This prompts innovative ways of creating solidarity between producers and consumers and puts pressure on points in the value chain other than direct employers (Hartwick 2000; Johns and Vural 2000). The effects of fragmentation are complicated by processes of labour migration and cultural diversification, creating a more socio-culturally heterogeneous workforce (Wills 2008). This can further erode worker solidarity and social cohesion. Trade unions have historically played ambiguous roles in this respect: some have been guilty of cementing divisions between ethnic groups or full-time/part-time workers; others have managed to forge solidarity between different identities.

It might be worth asking why ‘fragmentation’ is the best term to describe these processes, and not some other term such as labour market segmentation (Kalleberg 2003) or labour process restructuring (Gough 2003). Admittedly, the multiple processes that constitute fragmentation suggest that it is a chaotic concept. Still, what is appealing about this terminology is that it resonates with the experiences of a particular subject position – that is, workers trying to organise. Collective action is, by definition, based on constructing notions of solidarity between people who might otherwise differ from each other. Certain workplaces and local communities have traditionally provided relatively stable social and material bases for constructing strong ties over time (Darlington 1993; Wial 1993). Fragmentation of work represents a socio-spatial challenge to this solidarity.

The scalar strategies of particular labour struggles have been explored through several well-known case studies (Castree 2000; Herod 1998; Sadler 2000; Tufts 2007). The concept of scale has remained useful for labour geographers, despite receiving criticism in recent debates (e.g. Marston et al. 2005). Organised labour is traditionally structured in well-defined hierarchies and has achieved significant political victories through explicitly embracing scalar strategies. Neoliberalism represents not only fragmentation, however, but also a rescaling or denationalisation of state power (Jessop 2002), including the de-/re-regulation of labour markets and labour control (Peck 1996). When labour regulation and systems of collective bargaining are being weakened at particular scales, and when new scalar constructions such as entrepreneurial cities come to the fore, trade unions are forced to rethink their scalar strategies to remain relevant and influential. They can no longer take for granted that strong workplace and national organisations are sufficient, or even the most appropriate way, to build solidarity and engage politically – suggesting alternative bases of political leverage.

Similar challenges, different contexts

When discussing union renewal, it is easy to inadvertently draw a straw man, a caricature, of the ‘traditional trade union’: an interest organisation for male factory workers, an organisation that loses members and fails to establish a presence in the new economy, a rigid bureaucracy unable to break away from a hierarchical culture and a movement failing to reflect a heterogeneous and multicultural workforce. Of course, unions have always broken with this norm. Still, I would argue that it is hard to approach Western trade unionism without using the shift from Fordist to post-Fordist employment relations as a backdrop. This shift also signals the relative decline of unions that shared many of the straw man’s characteristics, unions that in fact were incredibly successful in representing workers in different places and at different times in history, e.g. across the Rust Belt during Fordism and in Nordic welfare states. Unions’ main site of recruitment is the workplace, as factory floors prove fertile soil for collectivism. While labour parties were used as vehicles to pursue a wider social agenda, trade unions based their mobilisation in the sphere of production – wages and employment conditions being typical issues in this regard. In addition to the emphasis on national organisation mentioned above, this era was also characterised by a sophisticated relationship developing between labour and capital through social partnership and collective bargaining. The boss was the typical target for action, with the strike being labour’s main weapon. However, this approach has also represented constraints, and many trade unions have actively distanced themselves from this stereotype by pursuing a community orientation.

While having a certain general applicability, this picture is not representative of trade unions in a development context. The majority of wageworkers live and work in the global South. Post-colonial trade unions have their own historical experiences, which have also been shaped by societal changes, albeit from very different starting points. Whereas the relatively coherent Keynesian welfare states gained legitimacy through wage redistribution to a broad, unionised working class that enjoyed collective bargaining rights, worker organisations in developing countries mobilise in economies in which the informal and agricultural sectors are dominant and unionised workers represent a small minority of the population. Still, it can be argued that some of the challenges identified in the mainly Anglo-American case study literature point to problems also facing unions in the global South. It is not the task of this article to provide a detailed overview of the union challenges across the developing world (see Lambert and Webster 2001; Lindell 2010; Munck 2002). Rather, by using a case study, I will show how South African trade union experiences can be brought into a fruitful exchange with an analytical framework based on the Anglo-American labour geography literature. Before doing so, some characteristics of post-apartheid unionism in South Africa must be mentioned.

South African unions have a different point of entry than trade unions in Western economies as they try to carve out a political agenda in the post-apartheid era. With a track record of social movement unionism during their struggle against the apartheid state, they are well aware of the potential strength of alliances between trade unions and political movements (von Holdt 2003). Entering the neoliberal paradigm of the 1990s, they were suddenly expected to act as responsible partners to a democratic government. When the African National Congress (ANC) became South Africa’s first democratically elected government in 1994, it was united with its previous allies in the resistance movement, notably the Congress of South African Trade Unions (COSATU) and the South African Communist Party (SACP). It was hoped that this tripartite alliance would translate the social movement unionism of the resistance into considerable union influence in national politics. Since the early 1990s, however, two related political developments have fundamentally circumscribed organised labour’s influence.

First, civil society went through a post-apartheid transformation. A growing gap emerged between those aligned with the ANC on the one hand and diverse oppositional groups on the other (Habib and Kotzé 2002; Oldfield and Stokke 2007). By the end of the 1990s, numerous community groups had started openly resisting what they perceived as attacks on their livelihoods (Ballard et al. 2006; Miraftab and Wills 2005). They were concerned with the cost and quality of basic services and rights to adequate shelter and land. Forced water and electricity cut-offs and house evictions initiated by local governments and banks encountered fierce resistance, as did the lack of land distribution. Often referred to as ‘social movements’, these groups were not using elections or strikes as their means of influence, but rather mass mobilisation and direct action (Desai 2003). The demands of these movements were in line with the progressive political agenda the union movement claimed to be promoting. However, their unpredictable methods and confrontational political approaches threatened the image of the trade unions as responsible social partners. Government officials and the media vilified some of these social movements, and state repression and frequent clashes with the police weakened their organisational ability. To COSATU-affiliated unions, this made the prospect of collaborating with them less appealing (Lier and Stokke 2006).

Second, the tripartite alliance became an arena in which power struggles relating to class, ethnicity and personality played out. It soon became apparent that the multi-class character of the ANC organisation and its representatives in government was susceptible to pressure from associations, think tanks and managers of domestic capital and from international financial and development agencies (Bond 2000). The alliance soon proved insufficient for organised labour to secure working class interests (Webster and Buhlungu 2004). COSATU launched nation-wide anti-privatisation strikes in 2000, 2001 and 2002; continued pressure from left-wing forces within (and outside) the tripartite alliance arguably led the government to reconsider many of its more radical neoliberal targets and to increase certain forms of social spending (Harrison 2006). The transition to a (neo)liberal democracy was also marked by the reluctance of state authorities to recognise trade unions as legitimate actors outside formal systems of industrial relations or corporatist forums such as the National Economic Development and Labour Council (NEDLAC). Relations between organised labour and government authorities have continued to be tense throughout the 2000s.

While COSATU unions bring together workers from all previous racial groups, they face challenges of representation in the post-apartheid economy. On the whole, the South African working population is dominated by people living in poverty and unemployment who make ends meet through social networks based in the urban informal economy and in rural agriculture. This has led some observers to claim that trade unions represent neither the poor nor the majority of South African workers (Friedman 2002; Seekings 2004). The erosion of standard employment relationships as a result of liberalisation, neoliberal restructuring and deregulation has led to a further erosion of the traditional base for trade unions. Webster (2006) estimated that of a 12 million strong workforce in 2006, just over half were in standard employment relationships; in only two of COSATU’s 21 affiliates was the percentage of members outside standard employment relationships more than 10 per cent. Casualisation and informalisation of employment is leading to what Webster and Buhlungu (2004) labelled a crisis of representation. Such a crisis occurs when organised labour, in part through its own success, ends up representing a well-protected, but declining, core of workers while being non-representative of the working class as a whole. Through a very different historical trajectory, then, South African labour is confronting many of the same challenges experienced by trade unions in neoliberal countries such as the UK and Australia.

Community-oriented unionism as a socio-spatial response

Encouraging stories of union renewal have demonstrated the ability of trade unions to reclaim representativeness where these new geographies of work are taking shape. Their differences notwithstanding, one can argue that some commonalities warrant closer scrutiny. Figure 1 suggests that community-oriented union politics can be understood as related, yet distinct, socio-spatial responses. When trade unions, even those not typically classified as ‘community unions’, choose to engage more strategically with communities, they can do so according to at least four different spatialities. While scale is one of these, I have consciously moved beyond the language of scale. This is to enable the model to better distinguish how trade union organisations are articulated spatially from how they relate to authorities (e.g. employers, the state), potential new members and political alliances.

Figure 1.  Union renewal through community-oriented unionism Source: The author

I further differentiate these four distinct, yet interrelated, spatialities of union renewal by using Cox’s (1998) notion of spaces of engagement/dependence as sensitising tools: the lower two correspond with unions’‘spaces of dependence’ (the relatively localised relations forming the social basis of the organisation), whereas the upper two movements refer to the ‘spaces of engagement’ where unions seek influence beyond their shop-floor mandate. While workers are attached to certain places through their livelihoods, workplaces and social networks, their unions are indirectly attached to the same places through their paying members, who also constitute their political constituency. While workers might experience outsourcing in the form of wage reduction or greater social risk, the union organisation suffers in the form of membership loss and fragmented recruiting grounds. By referring to studies of union renewal from geography and sociology, I hope to provide a fuller meaning to how these four tendencies appear in practice.

An expanded domain of mobilisation refers to the ability and willingness of trade unions to ‘move beyond their traditional workplace boundaries’ (Lambert and Webster 2001, 350) to find ‘common cause between unions and those groups cemented around affiliations of religion, race, gender, disability and sexuality’ (Wills 2001, 468–9). Many of the unions in question have actively sought to become involved in issues not directly related to employment contracts, conditions of service or wage levels. Such a shift does not necessarily increase the trade union’s membership directly, but it can serve to widen its political base through alliance-building and coalitions. This has brought unions into contact with religious organisations, women’s movements, environmental justice groups and community-based organisations (Sadler 2003; Sziarto 2008). In fact, an expanded domain of mobilisation often entails a new gender dynamic in the trade union organisation (Jepson 2005; Munro 1999). Whether the motivation is to seek political support for the union’s own agenda or to boost the political struggles of others, the active construction of a ‘common cause’ is seen as a key success factor in such alliances (Nissen 2004). In some cases, this is a question of bringing together already mobilised constituencies; in other cases, unions actively create political consciousness in relatively quiescent communities.

Related to this, several unions have shown an ability to identify alternative sites of recruitment when the workplace is fragmented or inaccessible. Reaching beyond the workplace can allow unions to get in touch with otherwise unreachable groups of workers. The long-running Justice for Janitors (J4J) campaign of the Service Employees’ International Union (SEIU) in Los Angeles is a good example in this regard (Savage 2006). In a context of deunionisation, organisers made use of churches, residential blocks and community centres to recruit workers otherwise spread out and isolated in high-security worksites downtown. Here, colleagues found they were closer to each other geographically and had more social interaction in their time off work than during work hours. A recent example from Johannesburg resonates with this experience: when the trade union General Industrial Workers Union of South Africa (GIWUSA) began organising unemployed members and labour broker employees in addition to its permanently employed members, it decided to set up territorially defined township locals complementing its existing workplace locals, thus reaching its expanded membership base more effectively (Appolis 2009).

Broad-based recruitment and alliance strategies enable unions to engage with other spaces of power in new and innovative ways. Campaigns, discourses and networks can create new scales of organising that transcend the workplace and administrative regions of the trade union. Alliances between groups of workers and other social constituencies can be based in local neighbourhoods or framed as city-wide campaigns. Living-wage campaigns epitomise this strategy, and Walsh’s (2000) analysis of the Baltimore experience depicted how the scale of organising was changed from the workplace scale to the metropolitan scale. Only then was it possible for workers from different workplaces to force their wages ‘out of competition’ by compelling city authorities to pass a living-wage ordinance (see also the ‘geographical unionism’ of Wial 1993). In his insightful analysis of union renewal, Tufts (2007) emphasised the relationship between various scales of organising as a mutually reinforcing dynamic and observed a ‘spatial circuit of union renewal’ emerging in the hotel sector in Toronto. Here, support for organising efforts at the workplace level was bolstered by city-level community campaigns, national coordination and international support. A similar argument is made by Castree (2000) on labour internationalism during a British dock dispute. Community-oriented unionism should be understood as multiscalar, and renewal along each of the four movements depicted in Figure 1 can in principle occur in a variety of political scales. Processes of decentralisation and new forms of urban governance in the developing world also provide an impetus for trade unions to create more dynamic urban campaigns (Lindell 2010). In South Africa, where a strong national trade union movement has been in place since the mid-1980s and where metropolitan governance is seen as a driver of development and democratisation, recent struggles around production and consumption have indeed centred on the urban scale.

Finally, community-oriented unions have been able to identify alternative targets for action in a socio-economic landscape in which the direct employer has limited ability to change the terms of employment: the Coalition of Immokalee Workers managed to successfully redirect its focus from the tomato growers to its fast-food chain customers (Drainville 2008); the J4J campaign enabled workers to exert pressure on building owners instead of their subcontracted employers (Savage 2006) and, in the case of garment homeworkers in Toronto, a solidarity campaign was directed at the retailers instead of the producers (Tufts 1998). Action at the point of sale can serve to bring the political struggle into spaces of consumption. Buyer-driven commodity networks often have their locus of power situated with retailers. Therefore, unions have sought to engage with consumer groups to influence alternative targets for action in response to the alienation and spatial separation between consumption spaces and production spaces in commodity chains (Hartwick 1998; Johns 1998). Labour market intermediaries such as temporary staffing agencies also represent a blurring of employer responsibility that requires unions to think unconventionally (Wills 2009). This is a pressing issue in South Africa, where so-called ‘labour brokers’ have mushroomed in the last decade, bridging the informal and formal economy in ways that have encouraged casualisation and outsourcing of both private- and public-sector work.

While union–community interaction by some has been celebrated as progressive organisational innovation (Waterman 1999), others have portrayed it merely as a reflection of contradictions in the political economy (Park 2007). With this model, I suggest that the best analysis of community-oriented unionism is one that appreciates the strategic agency of workers’ organisations while recognising that their politics are embedded in and shaped by structures in the labour market and in the wider political economy. When labour is marginalised, fragmented and progressively more exploited in the capital–labour relationship, the basis of workers’ social reproduction is pushed further onto social networks (Herod and Aguiar 2006). However, so is their politics: shared identities and political alliances are forged with family members, communities, neighbourhoods and associational life. This expanded notion of organised labour should be interpreted not only as a defensive tactic in times of crisis, but also as having transformative potential. By juxtaposing socio-spatial challenges for organised labour with the different socio-spatial dimensions of union responses in the literature, I have tried to sketch out an analytical framework with a specific geographical sensitivity. These dimensions are interesting, as they seem to correspond to some wider trends receiving attention in a critical geography tradition. In the remainder of this article, I will focus on these issues in a context I have become familiar with through two separate fieldwork periods in 2003 and 2007, namely the politics of municipal governance in the city of Cape Town.

Sustaining and reinventing community-oriented unionism in Cape Town

The South African Municipal Workers’ Union was established in 1987 through a merger of five local government unions, the biggest of these being the historically coloured Cape Town Municipal Workers’ Association. In contrast to most other South African unions, SAMWU still has its national headquarters in Cape Town. An affiliate of COSATU, it is seen as one of the more politically progressive unions within this federation. The Cape Metropolitan branch of SAMWU (the Metro branch) has long been the majority union in Cape Town, with approximately 18 000 members in the mid-1990s. It experienced a dip in membership during the early 2000s, during which a rival union, the Independent Municipal and Allied Trade Union (IMATU), surpassed it in size. Historically, SAMWU has its stronghold among blue-collar workers in service delivery and has now recruited many black, Xhosa-speaking workers. IMATU has also changed from being a white union for skilled workers to recruiting workers of all colours across the municipal organisational hierarchy. SAMWU has now regained its position as the biggest union in Cape Town, with a current membership of 12 370 (SAMWU representative 6 April 2011).

Over the last 20 years, the Metro branch has opposed cost-recovery reforms, rationalisation and outsourcing in the municipality. These trends can be traced back to the late 1990s, when metropolitan administrations started to fear that they were heading towards fiscal crisis. Increased spending resulting from social redistribution and development, combined with limited subsidies from national government, placed huge pressure on municipal budgets. This was also the case with Cape Town, and consultants were hired to come up with solutions. It was argued that if several fragmented metropolitan administrations were to merge into one ‘unicity’ with unprecedented developmental responsibilities, local government would have to undertake a cost-cutting and rationalising exercise (Unicity Commission 2000).

Tufts (2007, 2384) argued that as economic globalisation has assigned renewed ascendancy to the urban scale, ‘workers who are needed to build “competitive cities”’ may be able to extract material gains as a consequence. For workers whose demands directly affect a city’s expenditure, however, the situation can be quite the opposite – and this became evident as the picturesque tourist destination of Cape Town set its sights on becoming a self-professed ‘world class city’. The drive to reduce salary costs led city authorities to reduce permanent staff from 33 000 to 19 000 within less than 10 years, mainly through natural attrition and a moratorium on new appointments. With fewer staff, outsourcing of service delivery and support functions became widespread. By early 2007, Cape Town had also hired between 4000 and 5000 contract staff via labour broker agencies to fill vacant positions.

Neoliberal reforms also meant that municipal service delivery systems in Cape Town were refashioned as utilities and, hence, run more like businesses. Cost-recovery principles in basic services have had serious implications for user fees and the collection of payments, including that inability to pay led to cut-offs, pre-paid water meters were installed and widespread problems with debt and arrears existed (McDonald and Smith 2004). Since 1999–2000, community organisations have started to emerge, mobilising around these issues in Cape Town (Miraftab and Wills 2005; Oldfield and Stokke 2004). Numerous township-based ‘anti-eviction committees’ organised and protested in the streets, resisting disconnections and evictions. Service delivery has since remained an intensely contested political issue in Cape Town.

These New Public Management-inspired reforms have had significant implications for municipal workers and service users in Cape Town. The response from organised labour, however, has been ambiguous. While the conservative IMATU union’s relationship to Cape Town throughout the restructuring process was based on collaboration, consultation and consent, SAMWU, in contrast, can look back at a decade of confrontation, negotiation and (at times) non-collaboration with Cape Town.

You see, we don’t play a social or political role at all. (IMATU representative 8 June 2007)

We believe that privatisation is profit-driven, and we believe that profit should not be made on basic services, considering that 60 per cent of the people of this country is poor […] So, I don’t know if you wanna call that ideological? (SAMWU representative 26 June 2007)

SAMWU tried to marry the broad-based union tradition of the apartheid days with a parallel strategy of social partnership in local government and mirrored Wills and Simms’ (2004) observation that decline is often met by a combination of organising and partnership. The union has sought to preserve its community orientation during times when the premise for such engagement has changed. This is not to say that SAMWU is a ‘community union’. In many respects, it has developed into a political union in the mould of Northern European unions that negotiate institutional forms of influence with a high level of mobilisation. Complicating this picture is the fact that, partly as a result of the labour movement’s failure to politically oppose privatisation and outsourcing, the union has seen its recruitment base increasingly pushed into the private sector. By using the four analytical categories outlined above, I will examine SAMWU’s strategies in Cape Town in the 2000s, when the restructuring of Cape Town gained momentum.

1 Defending the domain of mobilisation through social movement unionism

SAMWU officials and members sympathised with many of the service delivery and eviction protests in Cape Town. After all, municipal workers experienced neoliberal restructuring in both the workplace and as service users. Still, many officials were reluctant to engage with organisations outside the ANC fold, although attempts were made in Johannesburg (Barchiesi 2007; Buhlungu 2006). SAMWU’s Metro branch initiated the Cape Town Anti-Privatisation Forum to bring together local organisations mobilising around service delivery issues. However, this strategy ran into a series of contradictions, as expressed by the national chairperson at the time:

How do we find the balance between still seeing the ANC […] as the only viable vehicle which creates space for workers to basically express themselves […] and the realities on the ground whereby communities are becoming more and more disillusioned? (SAMWU representative 1 December 2003)

Interestingly, unions and social movements have found working together to be problematic in part because of their different spatialities: hierarchical versus rhizomatic, nationally coordinated versus spontaneous and local, formalised engagement with the state versus informal and confrontational (Lier and Stokke 2006). Consequently, broad-based coalitions have largely been left in favour of more targeted, single-issue campaigns.

One of the more successful campaigns has been the so-called Water for All Campaign (WFAC). WFAC emerged in the mid-2000s and was driven by a committee consisting of activists from SAMWU and various other organisations opposing pre-paid water meters, which were seen as an instrument of neoliberal cost recovery. According to senior Cape Town management, strict cost-recovery mechanisms were necessary to combat a ‘culture of non-payment’ and to secure the financial viability of these services. In SAMWU’s opinion, cost recovery was a way of forcing poor and unemployed people to accept the market logic in basic services. Pre-paid water meters and similar devices were also seen as a health risk: adequate hygiene levels and fire safety depend on access to water. SAMWU officials upheld that it was paradoxical that employers and government authorities had been trying to restrict workers’ right to strike, referring to their role in delivering essential services:

Workers are of the view that how can you call a service essential when you shut them off for the poor with electricity and basic services like water? (SAMWU representative 26 June 2007)

Through marches, petitions and several public meetings, WFAC made sure that this aspect of Cape Town’s service delivery regime was put on the agenda. The initiative also brought organisations aligned with the ANC in contact with groups in vocal opposition to the government party. Bridging this divide was only possible through the personal networks that have evolved over time in left-wing circles in the city. Through such alliances, the union was able to strategise beyond its traditional constituency of permanent local government employees. These organisational alliances are networks that can prove crucial for ‘developing common political identities and alternative imaginaries’ (Leitner et al. 2008, 162).

SAMWU engaged in such alliances while at the same time insisting on a broad social agenda in its industrial relations with Cape Town employers. Bringing non-workplace issues such as access and quality of service delivery to the table in these forums was met with resistance from local government authorities. This tension was visible in the South African Local Government Bargaining Council: union representatives complained that the employer side was depoliticising the forum by marginalising elected politicians in favour of technocrats; union representatives felt they were discouraged from raising issues concerning service delivery, occupational health and safety and workplace racism.

2 Negotiating scales of organising

Defining what is ‘appropriate’ trade union politics relates closely to securing access to political spaces. Even though SAMWU built a strong, coordinated union organisation at a city level, as well as at a national level, it has increasingly struggled in using this space to achieve significant political victories. The metropolitan scale has been consolidated into a distinct level of local government with an amalgamated financial structure and democratic elections. Throughout this process, SAMWU tried to strengthen its Metro branch while maintaining constructive relations with employer representatives. However, as the 2000s progressed, the relationship between Cape Town and SAMWU became increasingly unstable. Negotiations around restructuring issues ended in stalemate. In Cape Town’s view, the delayed restructuring process called for expediency and depoliticised industrial relations. SAMWU’s strategy of establishing a social partnership with Cape Town management during this period of intense restructuring of the municipal organisation was reduced to a defensive fight for the union’s space of engagement (Cox 1998). A statement from a senior Cape Town manager is telling in this regard:

SAMWU says: ‘We are a partner in service delivery because our members do the work, so if you want to restructure you partner with us – we decide together how we restructure’. Now, when an employer says: ‘Sorry, that’s not how it works, I run the business, you only have members’, then obviously there’s gonna be a fall-out. (Cape Town senior manager, interview 22 June 2007)

Further complicating the union’s grip on the metropolitan scale were new government policies influencing the management of service delivery. A national stalemate within the ruling alliance around privatisation was displaced to the municipal level through prescribed local consultation processes among residents, local states and trade unions, a much more time-consuming exercise for the union than challenging privatisation at a national scale.

Centralised city-level bargaining also faced an impasse, and Cape Town proposed a new decentralised industrial relations regime that included the introduction of so-called ‘local labour forums’ in the mid-2000s. Union representatives were initially reluctant to embrace this decentralisation; they saw a potential fragmentation of responsibility and feared this would weaken its political strength. These dilemmas were manifested in several failed attempts to negotiate with the employer, followed by protests and strike action. However, recently the union has managed to use these decentralised spaces for its own benefit, increase its influence and enhance its own capacity on particular sectoral issues (SAMWU representative 6 April 2011).

3 Exploring a new frontier for recruitment through externalised labour

As the policy of using labour brokers and private contractors gained terrain, SAMWU became less representative of the municipal workforce. One of the most significant strategic shifts in SAMWU in recent years has been the efforts to organise externalised labour, i.e. private-sector employees working for the municipality. Targeting these workers as potential members has been imminent in the COSATU federation since the late 1990s. This shift was motivated not only by solidarity and working-class ideology, but also by pragmatic concerns; as long as this growing layer of workers remained unorganised, working unconditionally for low wages, COSATU’s members were in danger of becoming uncompetitive (cf. the notion of ‘accommodative solidarity’ in Johns 1998). In the mid-2000s, SAMWU adopted policies and launched legal disputes in relation to the growing level of externalised labour. The union managed to successfully unionise some private contractors, but there were persistent problems in effectively representing these workers because of issues of institutional coverage.

The greatest recruitment potential was in the private waste companies, large companies enjoying relatively stable contractual arrangements with Cape Town due to their capital investments. SAMWU’s Metro branch was able to organise workers at some of these employers; almost 90 per cent of the workforce at Wasteman was signed up, which represented the union’s biggest success in the sector. Other companies posed greater difficulties: labour broker companies, for instance, have mushroomed across the city. Responding to staff shortages and the outsourcing policy of the city of Cape Town, these companies supply cheap, temporary labour. Many of them benefit from preferential treatment through so-called Black Economic Empowerment policies. In the labour broker sector there is little stability and workers constantly change workplaces. Moreover, SAMWU is unlikely to gain majority representation in these companies, as labour brokers have their main market in the private sector.

Community-based subcontractors are a third type of company that has been difficult to unionise; these are small companies that either directly or indirectly supply Cape Town with labour and services. Their employees perform municipal tasks, but do not enjoy anything close to the employment standards and wage levels of municipal employees. The accountability and monitoring of these workers are effectively out of the hands of the city of Cape Town, and because of the workers’ fragmented, unstable employment nature, few attempts to organise this workforce have been made. In fact, after union involvement helped a small group of sewage workers in the Khayelitsha township to secure decent work wear and some contractual security, the workers asked SAMWU to ‘pull back’ (SAMWU representative 26 June 2007); they had achieved what they set out to do and were afraid of employer sanctions. This perhaps suggests a need for SAMWU to rethink their sites of recruitment, in line with the lessons learnt from the J4J campaign in Los Angeles (Savage 2006). These ‘community companies’ are often coordinated by ANC-aligned community organisations in the Xhosa-speaking townships. Efforts to bring these jobs in-house face hostility from some of these organisations and their leaders, who see a threat to their authority and to residents’ casual income. In this way, casualisation has served to circumscribe the potential for union–community mobilisation by pitting different groups of workers against each other in competition for employment opportunities.

Notwithstanding its efforts, SAMWU’s presence in the private waste and labour broker sectors still remained relatively insignificant in Cape Town: from zero members in 2000 to approximately 400 in 2007, a number that had increased somewhat by 2011 (SAMWU representative 6 April 2011). Unionisation rates are generally low in externalised municipal services, and SAMWU’s overall private firm membership accounted for less than five per cent of national membership in 2007. In other words, if the goal was to undermine the rationale for outsourcing by effectively raising the conditions of these workers, the union still has a long way to go. One of the main obstacles is the sector-based structure of the South African industrial relations regime. While SAMWU has gained recognition in the local government bargaining council, it has not done so in the bargaining council where externalised workers belong. The most tangible of SAMWU’s attempts to engage political authorities on issues of externalised labour was its legal challenge to the jurisdiction of organised private firm workers.

4 Points of production and regulation as targets for action

The mobilisation against pre-paid water meters through WFAC also provides a telling example of how community orientation helped identify new targets for political action. Among those involved in the WFAC were representatives from the heavyweight National Union of Metalworkers of South Africa (NUMSA), whose involvement signalled the commitment by another big union with strong activist credentials. Moreover, NUMSA happened to organise workers in manufacturing plants that were planning to manufacture and assemble such water meters. NUMSA’s involvement in WFAC prompted its shop stewards to meet with managers at a manufacturing plant outside Cape Town, where the managers received an unwritten guarantee ensuring that this kind of water meter would not go into production. NUMSA had expressed reluctance to use its industrial muscle on this issue, however, and going on strike to block the production of pre-paid meters was unlikely. Complicating matters further were the political differences and conflicts of interest between NUMSA and SAMWU; the latter also wanted to ban the production of pre-paid electricity meters, a much more important source of employment for NUMSA members.

Ultimately, it was the Cape Town politicians who decided to use these controversial devices and senior municipal managers were in charge of implementing them. Given its interaction with city management in industrial relations, SAMWU could have exerted some pressure in this regard. As noted above, SAMWU faced strong resistance to bringing service delivery-related issues into forums of negotiation where SAMWU representatives met city officials. Together with WFAC activists, SAMWU used its organisational capacity to host meetings in community halls. As mentioned, this strategy was in fact fairly successful in the Cape Town case: in 2005, through community meetings and advocacy, SAMWU received a guarantee from the mayor of Cape Town that the rollout of these devices would not continue. The WFAC alliance played its part in challenging the spread of pre-paid meters in poor Cape Town households. While mobilising through community meetings was the key strategy, union involvement also opened two alternative targets for actions: the point of production, which was the plants where NUMSA members worked, and the point of regulation, which was the Cape Town organisation where SAMWU members worked and with which SAMWU representatives tried to maintain a social partnership.

Conclusion

In this article I have argued that union renewal – embryonic but observable across the world – represents a complex mesh of negotiating partners, allies and political targets that change the spatialities of unionism. Moreover, there seems to be a link between certain patterns of restructuring and neoliberal change and unions’ scalar strategies, their political targets, their recruitment strategies and their role in civil society. Through differentiating between four spatial dimensions of community orientation, we can take another step away from binaries such as old/new social movements and business/community unionism. Rather than simply stating that unions (need to) ‘branch out’, we can read union renewal as the redrawing of how unions expand their membership, with whom they build alliances, who is targeted by their political tactics and how they build their organisations.

In many ways, the case of SAMWU adds to an already well-documented diagnosis in the literature of unions being put to the test, although some interesting differences also emerge when we choose cases from a very different historical-political context. Where processes of fiscal and political decentralisation accompany neoliberal forms of governance, trade unions are forced to rethink their existing scalar arrangements. Municipal workers in Cape Town have historically enjoyed powerful union representation at the metropolitan level. In the post-apartheid era, they were forced to recreate the urban scale by engaging with an ambitious local government restructuring agenda. In contrast to the urban entrepreneurialism experienced in the UK and North America, South African municipalities are also the cornerstones of the country’s development path. Entrepreneurial ambitions are coupled with huge expectations of service delivery rollout, creating an even stronger incentive for labour practices that deliver ‘more for less’. While this adds significance to the municipal scale, it also intensifies a battle over priorities wherein municipal workers are defending their jobs and affordable basic services.

However, scale is not an exhaustive term that captures every socio-spatial dimension of organised labour. We should also recognise how union renewal entails different targets for action to challenge (real) employers, state institutions and other spheres of power. Trade unions struggle to achieve results vis-à-vis formal employers in labour markets transformed by ‘subcontracted capitalism’. The notion of new targets is becoming particularly relevant in South Africa where labour market intermediaries have experienced rapid growth since the 1990s serving both the public and private sectors. SAMWU has sought alternative targets for action by singling out business actors and state institutions with influence over the state of service delivery, not limiting itself to the municipal organisations mandated to engage with union members as an employer. The union has also enjoyed some success in recruiting employees in private companies, thus making it more representative of the post-apartheid municipal workforce.

SAMWU combines strategies reflecting its political past and future challenges. While ‘community unionism’ represents something new for many trade unions (or at least a remnant of a very distant past), SAMWU’s Metro branch has attempted to sustain and reinvent the social movement unionism of the anti-apartheid struggle into the 2000s. When union renewal is sought through engaging with community elements, this opens new spaces of engagement and targets for action. This is not to say that community orientation guarantees success. While SAMWU’s attempts to mobilise around service delivery reveal great potential for community solidarity, the union’s efforts to reverse casualisation have often pitted it against community organisations that perceive union members as competitors for the same jobs. Moreover, echoing Castree’s (2000) observation that the success of labour internationalism relies on strong national and local structures, SAMWU’s community engagement would exert limited political leverage without a solid base built through years of signing up members in municipal depots across the city and of mobilising formal employees around workplace issues. Following from this, the application of the case to the general model (see Figure 2) juxtaposes the trade union’s original basis of mobilisation (the centre) with its reorientation through new strategies (the arrows).

Figure 2.  SAMWU’s strategic engagement with a neoliberalising municipal labour regime and the wider community Source: The author

As South African unions have been on the defensive since the 1990s, and with the local branch having limited resources at its disposal, engaging in broad-based mobilisation, actively seeking new sites of recruitment and identifying targets for their campaigns has been exhausting – even disorienting. However, even limited success in one area can lead to mutual reinforcement in membership growth, effective political alliances and successful campaigns and bargaining. In addition, as with many other trade unions, the alternative for SAMWU would be to engage in defensive unionism, which would eventually weaken the union to represent only a small, protected core of workers.

Acknowledgements

I am grateful to my PhD supervisors Neil M. Coe and Noel Castree for engaging in the development of this argument, and for the comments from the editor and three anonymous reviewers on previous versions of this manuscript. I remain responsible for any misinterpretations in this article.