From a ‘Decent Work’ Approach to a Marxist Analysis of Labour Control and Labour Agency in Global Production: Reviewing Research on Labour in GPNs

Abstract

This chapter reviews literature on labour in GVCs and GPNs. It argues that within the interdisciplinary literature on labour issues in GVCs/GPNs, two parallel research strands have emerged that are characterised by very different conceptual approaches: (1) a ‘Decent Work’ approach underpinned by the institutionalist perspective of the ILO Decent Work Agenda and (2) a ‘Marxist Political Economy’ approach, which is based on the assumption that the exploitation of labour is an inherent structural feature of capitalist production systems. Situating this study within the second research strand, this chapter then reviews the contributions and shortcomings of existing literature on labour control and labour agency in GVCs/GPNs. In doing so, the chapter highlights the limitations of existing scalar approaches for studying labour control and labour agency in GVCs/GPNs, which have not paid enough attention to how dynamics of labour control and labour agency at different levels influence each other. Against this background, this chapter argues that to gain a more nuanced understanding of the ‘architectures of labour control’ underpinning specific GPNs as well as of workers’ and unions’ networked agency strategies, a relational analytical approach can be beneficial.

This study aims to contribute to and establish itself within the broader debate of economic geography on labour in global production networks (GPNs). In particular, it aims to contribute to two strands of research on labour in GPNs that are concerned, firstly, with structures and mechanisms of labour control in GPNs; second, with the conditions and strategies for the agency of workers and unions in GPNs. In this regard, it is important to note that rather than exclusively working with the GPN framework, contributions to these debates have also worked with the Global Commodity Chain (GCC) and the Global Value Chain (GVC) framework. Therefore, in the following literature review, I include studies focussing on labour control and labour agency working with either of these approaches. When developing my own relational approach to labour control and labour agency in Chapter 3, I will, however, draw on the GPN framework since its network perspective fits best with the relational perspective adopted in this book.

In the remainder of this chapter, I will first demarcate the conceptual features of the GCC, GVC and GPN approaches and introduce central analytical concepts of the GPN approach (Sect. 2.1). After that, I outline two basic conceptual approaches to labour in GPNs—the ‘Decent Work’ approach and the Marxist Political Economy approach—and position this study within the latter approach (Sect. 2.2). Sections 2.3 and 2.4 then give an overview of the current state of research on labour control and labour agency in GPNs. Each chapter sketches three main strands of research on labour control and labour agency, respectively, summarising their conceptual and empirical contributions and highlighting their shortcomings.

1 From Linear Commodity Chains to Relational Production Networks: Opening Up Analytical Space for the Role of Labour

In this section, I introduce the three main analytical frameworks that have underpinned past studies of labour in global production systems: the Global Commodity Chain (GCC), the Global Value Chain (GVC) and the Global Production Network (GPN) framework. Whereas the GPN framework was the first framework to explicitly open up analytical space for labour, an increasing number of GCC/GVC studies has also tackled the enabling and constraining conditions for union and worker agency in global production (see e.g. Anner 2015b; Riisgaard and Hammer 2008, 2011; Selwyn 2012, 2013, 2015). This development is in line with the general trend towards closer integration of GCC, GVC and GPN studies over the last two decades, with the three approaches coming to form one wider interdisciplinary research community (Coe and Yeung 2019: 775). Therefore, when displaying the state of the broader research debate on ‘labour in GPN’ in the remaining sections of this literature review chapter, I will include conceptual approaches concerned with the control and agency of labour in global production systems working with either of the three approaches. Nevertheless, important conceptual differences remain between the GCC, the GVC and the GPN framework, particularly with regard to their potential for producing relational accounts of the constraining and enabling factors for the agency of workers and unions within global production systems. I argue that when it comes to understanding the agency of workers as embedded in and constituted through manifold networked relationships, the GPN framework provides the best conceptual tools (c.f. Cumbers 2015: 136). To illustrate this argument, in the following, I briefly introduce the central ontological assumptions and conceptual tools of the GCC, the GVC and the GPN approach.

The Global Commodity Chains (GCC) framework was introduced by the American sociologist Gary Gereffi (1994, 1999) in the 1990s and subsequently found wide application within economic geography. The declared aim of the GCC framework was to provide a framework for analysing different modes of organising international “production systems that give rise to particular patterns of coordinated trade” (Gereffi 1994: 96). As the name suggests, the GCC framework is underpinned by a linear chain ontology that conceptualises international production systems through the lens of their sequential ‘input-output structure’, defined as “the set of products and services linked together in a sequence of value-adding activities” (Gereffi 1994: 97). As two other central dimensions for analysis, Gereffi (1994) has introduced ‘territoriality’, referring to the specific geographical distribution of the activities involved in the production, distribution and sales of a specific commodity, and ‘governance structure’, referring to the power relations between firms in a specific production chain that determine “how financial, material and human resources are allocated and flow within a chain” (Gereffi 1994). The most important conceptual contribution of the GCC framework has been in the dimension of governance with the distinction between buyer-driven and producer-driven commodity chains (Gereffi 1999: 41f.). Whereas producer-driven commodity chains are controlled by large industrial multinational enterprises and found typically in capital and technology-intensive sectors, buyer-driven commodity chains are typically controlled by large retailers and branded merchandisers selling labour-intensive mass consumption commodities. Retailers and branded merchandisers—typically originating from the Global North—act as lead firms. They capture the biggest share of added value in the commodity chain by controlling the higher value-adding steps of design, marketing and retail, while outsourcing the typically labour-intensive production to independent suppliers in low-wage countries (Gereffi 1994: 97). At the same time, retailers and branded merchandisers exercise significant control over the parameters of the production process by providing detailed specifications regarding product design and quality as well as by setting prices and lead times (Gereffi 1999: 55). Considering that the global garment chain is a prototypical buyer-driven chain, the distinction between buyer-driven and producer-driven value chains has been a focal point for studies concerned with inter-firm power relations and its impact on labour in the global garment sector.

Notwithstanding the important to date influence of the distinction between producer-driven and buyer-driven chains provided by the GCC framework, it was also subjected to various critiques, particularly by economic geographers. Firstly, critics pointed out the GCC framework’s almost exclusive emphasis on ‘governance’ as an analytical dimension, which was perceived to make an overly crude distinction with its exclusive focus on buyer-driven and producer-driven commodity chains as two sole modes of governance (Dicken et al. 2001). Second, critics argued that due to its linear chain ontology and the resulting focus on inter-firm relations, the GCC framework possessed limited capacity for analysing the role of state, labour and civil society actors in shaping global production systems (Dicken et al. 2001: 100). Lastly, economic geographers criticised the GCC framework for paying insufficient attention to the ‘specific social and institutional contexts’ at the national, regional and local level into which firms are embedded and hence to account for how commodity chains shape local dynamics and vice versa (Henderson et al. 2002: 441).

Reacting to the first critique, Gereffi et al. (2005) presented the Global Value Chains (GVC) framework as a further development of the GCC framework, which provided a more nuanced typology of inter-firm power relationships in international production systems. The GVC framework hence distinguished between five types of value chains, each characterised by a specific type of inter-firm relationships: (1) markets, i.e. value chains that are characterised by arm’s length relationships between lead firms and suppliers; (2) modular value chains, in which ‘turn-key’ suppliers carry out technology-intensive production of specialised inputs or ‘modules’ for lead firms; (3) relational value chains characterised by complex interactions, high levels of trust and co-dependence between lead firms and suppliers; (4) captive value chains, in which a large number of rather small suppliers are ‘transactionally dependent’ on large buyers; and (5) hierarchy, a type of value chain that is characterised by vertical integration of all steps of production (Gereffi et al. 2005: 83f.). While already demonstrating a more fine-grained understanding of power relations within GVCs, the GVC framework, however, continued to focus on inter-firm relations around linear input-output structures, while largely ignoring other sets of relationships in international commodity production systems (Bair 2008; Coe et al. 2008: 275).

It is against this backdrop that a group of economic geographers of the so-called Manchester school set out in the early 2000s to develop the GPN framework as an alternative heuristic approach for analysing relationships in global production systems (Dicken et al. 2001; Coe et al. 2004, 2008; Henderson et al. 2002). As opposed to the linear ontology of the GCC and GVC framework, the GPN framework is underpinned by a relational ontology. This relational ontology conceptualises global production systems as constituted through “highly complex network structures in which there are intricate links – horizontal, diagonal, as well as vertical – forming multi-dimensional, multi-layered lattices of economic activity” (Henderson et al. 2002: 442, emphasis in original). Hence, the GPN framework recognises that each production network inevitably contains a vertical value chain dimension, i.e. a set of relationships linking actors throughout the linear sequence of stages manufacturing stages to distribution and consumption. At the same time, the GPN framework highlights that actors at each stage of the vertical dimension are also embedded into various sets of relationships at the horizontal dimension that constitute place-specific local, regional and national political economies (Coe et al. 2008: 274ff.) (Fig. 2.1).

Fig. 2.1

Source Author’s own elaboration based on Coe et al. (2008)

Horizontal and vertical dimension in the GPN framework.

In this light, Coe (2015: 185) has argued that the ontology of the GPN framework can best be characterised as ‘territorial cum relational’, since it integrates network relationships that link actors at different ‘nodes’ of the GPN on the one hand, and territorially embedded, institutionalised multi-scalar regulatory dynamics on the other. As a result of this particular ontology, the GPN framework can furthermore include a wide range of non-firm actors, such as labour, governments, civil society organisations and consumers, as “constituent parts of the overall production system” (Coe et al. 2008: 275). Through the relational lens of the GPN framework, it is from the interactions of these societal and state actors with firm actors at multiple levels that global production systems emerge in the form of networks. These networks are at the same time relational and structural: “Networks are structural, in that the composition and interrelation of various networks constitute structural power relations, and they are relational because they are constituted by the interactions of variously powerful social actors” (Dicken et al. 2001: 94). In this line, GPNs are understood as “contested organisational fields”, in which various actors with their own interests “struggle over the construction of economic relationships, governance structures, institutional rules and norms, and discursive frames” (Levy 2008: 944).

The GPN framework has introduced three central analytical dimensions: value, power and embeddedness (Henderson et al. 2002: 448ff.). The analytical dimensions of ‘value’ examines how different actors within the production network create, enhance and capture value. The GPN framework understands value in this context as encompassing “both Marxian notions of surplus value and more orthodox ones associated with economic rent” (Henderson et al. 2002: 448). ‘Power’ as an analytical dimension in turn raises questions about which actors exercise power in which ways to secure or increase their share of value within the production system (Henderson et al. 2002: 450). As opposed to the GCC/GVC framework, which conceptualises power very narrowly as lead firm power over suppliers, the GPN framework recognises various actors as potentially capable of exercising power within GPNs, including, for example, multinational and domestic firms, local and national state agencies, international organisations, trade unions and consumer organisations (Henderson et al. 2002: 450f.). Lastly, the analytical dimension ‘embeddedness’ introduces two different types of embeddedness that characterise GPNs: network embeddedness and territorial embeddedness (Henderson et al. 2002: 453f.). ‘Network embeddedness’ refers to the fact that GPNs link actors across territorial boundaries “regardless of their country of origin or local anchoring in particular places” (Henderson et al. 2002: 453). ‘Territorial embeddedness’ in turn refers to the fact that the actors and activities that GPNs’ links are at the same time ‘grounded’ in specific places for two reasons. First, most economic activities are spatially fixed in particular locations due to the immobility of the needed production infrastructure or labour force. Second, the actors in GPNs are embedded in place-specific social relationships, institutions and cultural practices that shape their interests and actions (Coe et al. 2008: 279). The distinction between network and territorial embeddedness reflects the distinction between relationships at the vertical value chain dimension of the GPN characterised by network embeddedness and relationships at the horizontal dimension of the GPN characterised by territorial embeddedness.

With its understanding of the labour process as a central moment of value creation and of labour as a potentially powerful actor in GPNs, the GPN framework also opened up conceptual space for analysing the role of labour in global production systems—both regarding the conditions of work and regarding the agency of workers and their organisations (c.f. Cumbers et al. 2008; Rainnie et al. 2011). As a result, over the past decade a vibrant research field tackling labour issues in GVCs/GPNs has emerged. Within this broader research field, two different approaches to studying labour in GVCs/GPNs can be distinguished: a ‘Decent Work’ approach and a ‘Marxist Political Economy’ approach. Debates on labour control and labour agency in GPN—which represent the analytical point of departure of this study—are generally underpinned by the ‘Marxist Political Economy’ approach. Therefore, to demarcate the research field that this study aims to contribute to, in the next Sect. 2.2 I briefly sketch the main theoretical-philosophical assumptions and concepts informing both approaches.

2 Contrasting Approaches to Analysing Labour in GVCs/GPNs: The ‘Decent Work’ Approach and the ‘Marxist Political Economy’ Approach

With labour issues coming into the focus of both GVC and GPN analysis over the past decade, two parallel research strands emerged that are characterised by very different conceptual approaches to analysing labour in global production systems. In this book, I refer to these two approaches as the ‘Decent Work’ approach and the ‘Marxist Political Economy’ approach. As the name suggests, the ‘Marxist Political Economy’ approach is underpinned by a Marxist political economy perspective. This perspective focuses on the relations between labour, capital, the state and consumers and is based on the assumption that the exploitation of labour is an inherent structural feature of capitalist production systems (c.f. Swyngedouw 2003; Rainnie et al. 2011). The ‘Decent Work’ approach to analysing labour in GVCs/GPNs in turn is underpinned by the institutionalist perspective of the ILO Decent Work agenda and focuses on the links between value chain governance and working conditions as a means for promoting social development (Barrientos et al. 2011b; Mayer and Pickles 2010). As a result of these different perspectives, the ‘Marxist Political Economy’ research approach and the ‘Decent Work’ research approach are characterised by rather different agendas and concepts of labour (Fig. 2.2).

Fig. 2.2

Source Author

Research approaches to labour in GVCs/GPNs—‘Marxist political economy’ approach and ‘Decent Work’ approach.

In the following two sections, I first outline the philosophical foundations, the conception of labour, central theoretical concepts, main research interests and criticisms of the ‘Decent Work’ approach (Sect. 2.2.1) and then of the ‘Marxist Political Economy’ approach (Sect. 2.2.2). It is important to note that the representation of these approaches in this chapter is a stylised one that aims to highlight their different conceptual and philosophical underpinnings. In this light, Table 2.1 gives an overview of the characteristics of each research approach to labour in GPN.

Table 2.1 Characteristics of the ‘Decent Work’ approach and the ‘Marxist political economy’ approach to labour in GPN

2.1 The ‘Decent Work’ Approach to Labour in GVCs/GPNs

The ‘Decent Work’ approach to labour in GVCs/GPNs was originally conceived by scholars from the GVC-school to provide policy-oriented conceptual tools for promoting the ILO Decent Work Agenda (Barrientos 2007; Mayer and Pickles 2010; Barrientos et al. 2011a). The ILO Decent Work Agenda was introduced in 1998 to tackle the increasing informalisation, deregulation and flexibilisation of labour markets under globalisation, especially in the growing export industries of developing economies in the Global South (Lerche 2012: 18). Whereas the ILO and GVC scholars saw the integration of developing economies into global value chains as an important motor for employment and economic development, they also recognised the need to improve the conditions of work and employment in newly industrialising countries (Barrientos 2007: 1; ILO 1999).

In this context, the ILO Decent Work Agenda was introduced to achieve a fair globalisation and poverty reduction through promoting rights at work, employment and income opportunities, social protection and social security as well as social dialogue (ILO 1999). ‘Decent Work’ was conceived in this context as a term that should converge the ILO’s four strategic objectives and thus designate work that takes place “under conditions of freedom, equity, security and dignity, in which rights are protected and adequate remuneration and social coverage is provided” (Barrientos 2007: 1). In the seminal report from 1999, the ILO Director-General pleaded that the concept of Decent Work “must guide [the ILO’s] policies and define its international role in the near future”. In this light, the ILO Decent Work Agenda must be interpreted as the main pillar of the broader ILO development agenda, which sought to promote the ILO’s principles and rights at work in international development policies and initiatives (ILO 1999; Lerche 2012: 18ff.). As the main strategy to achieve its goals, the ILO’s Decent Work Agenda proposed institution-building, particularly in the fields of worker participation and representation, social dialogue and social protection. Intellectually, the ILO Decent Work Agenda thus drew from debates in development theory and policies about promoting globalisation or development ‘with a human face’ as well as on institutional economics (ILO 1999).

GVC scholars started to engage with the ILO Decent Work Agenda in the 2000s to complement the perceived one-sided focus of the GVC framework on economic development with a social dimension (c.f. Gereffi and Korseniewicz 1994; Gereffi 1994, 1999, 2005; Gereffi et al. 2005). This engagement gave rise to a research strand within GVC and GPN studies concerned with identifying potentials, barriers and strategies for promoting decent work in GVCs/GPNs to promote economic and social development. Barrientos et al. (2011b: 320) formulate as the central question of the ‘Decent Work’ approach “how to improve the position of both firms and workers within GPNs”. On the one hand, the ‘Decent Work’ approach is hence interested in the impact of individual firms’ strategic choices on working conditions and labour rights. On the other hand, the ‘Decent Work’ approach to labour in GVC/GPN aims to conceptualise workers “beyond their role as factors of production, highlighting them as human beings with capabilities and entitlements” (Barrientos et al. 2011b: 322). In this view, the well-being of workers mainly depends on access to rights and resources that enhance their well-being. Access to these rights and resources is, however, mediated through institutional arrangements encompassing employers as well as government institutions and communities (ibid.).

In this line, the ‘Decent Work’ approach to labour in GVCs/GPNs proposed has put forth two main analytical concepts: the concept of regulatory governance and the concept of social upgrading. Mayer and Pickles (2010: 2) introduced the concept of regulatory governance as a type of governance that “constrains the behaviour of profit-seeking firms that might otherwise tend to exploit workers, leading to poor working conditions, lack of job security, constraints on worker organisation, and general downgrading of industrial relations systems and practices”. Regulatory governance can take on the form of public governance comprising governmental rules, regulations and policies and private governance comprising inter alia social norms, corporate codes of conduct, CSR initiatives, consumer campaigns, social movements and other non-governmental institutions (ibid.)

To promote Decent Work in GVCs/GPNs, Mayer and Pickles (2010) argued that it is central to overcome the present ‘governance deficit’ in the globalised economy resulting from “limited [governance] capacities in the emerging economies, weak international institutions, increasingly challenged institutions in advanced industrial countries and everywhere greater emphasis on facilitation than on regulation” (Mayer and Pickles 2010: 3). According to Mayer and Pickles, this ‘governance deficit’ has led to the deterioration of working conditions and to a shift away from formal, regular and secure employment towards more flexible, informal and insecure employment in global production. Pressure by consumer movements has then led to the rise of new private forms of regulatory governance, such as corporate Codes of Conduct (CoC), Multi-Stakeholder Initiatives (MSI) or Global Framework Agreements. There are, however, significant limits to the potential of private regulatory governance mechanisms alone improving workers’ rights and working conditions. Against this background, Mayer and Pickles argue that private governance mechanisms need to be complemented by initiatives that aim to strengthen public governance at the national and international level (Mayer and Pickles 2010: 14f.).

The concept of social upgrading, in turn, was introduced by Barrientos et al. (2011a) as a counterpart to the concept of economic upgrading, informing GVC studies until then. Whereas economic upgrading refers to the process by which economic actors move from lower value-added activities to higher value-added activities (Gereffi 2005: 171ff.), social upgrading designates the “process of improvement in the rights and entitlements of workers as social actors, which enhances the quality of their employment” (Barrientos et al. 2011b: 324). The central assumption of the concept of ‘social upgrading’ is that economic upgrading does not automatically lead to social upgrading (Barrientos et al. 2011b; Milberg and Winkler 2011; Rossi 2013). Whether social upgrading occurs in GVCs mainly depends on which market pressures prevail and on which competitive strategy firms adopt: Firms may compete via quality—a strategy that depends on a skilled workforce and thus is likely to promote social upgrading processes. Or firms may compete via price, which would most probably hinder processes of social upgrading or even effect social downgrading, e.g. deterioration of working conditions and workers’ rights (Barrientos et al. 2011b: 333).

Central research topics of studies adopting a ‘Decent Work’ approach to labour in GPN have been the quantity and quality of employment generated in GVCs/GPNs, with a particular focus on: export sectors in developing and emerging economies (Barrientos et al. 2011a), the relationship between economic and social upgrading (Milberg and Winkler 2011; Rossi 2013; Pyke and Lund-Thomsen 2016), as well as the effectiveness of private governance mechanisms such as codes of conduct (Locke et al. 2007; Egels-Zandén and Merk 2014).

Studies engaging with labour in GVCs/GPNs with a ‘Decent Work’ approach need to be given credit for problematising the initial implicit assumption of the GVC framework that processes of economic upgrading automatically bring about processes of social development. Moreover, they have produced relevant insights into the limitations of private governance mechanisms for improving workers’ conditions. However, research adopting a ‘Decent Work’ approach to labour in GVCs/GPNs has also encountered criticism, especially by scholars advocating a Marxist political economy approach for labour in GVCs/GPNs (Arnold and Hess 2017; Rainnie et al. 2011; Selwyn 2013; Werner 2012).

Marxist scholars have argued that the ‘Decent Work’ approach to labour in GVCs/GPNs and especially its conceptualisation of social upgrading present two crucial analytical and political weaknesses. According to Marxist scholars, the first and major analytical weakness of the ‘Decent Work’ approach is its inability to understand the systemic processes of exploitation characterising capitalist social relations as root cause for indecent work (Selwyn 2013: 75f.). Given its intellectual foundations in institutional theory, the ‘Decent Work’ approach in GVC/GPN is based on the assumption that “given the right institutional context, capital does not exploit labour” (Selwyn 2013: 82). According to Marxist scholars, this assumption is problematic because “the upgrading analytic centres industrial change on the relations of power and dynamics of competition among firms, rendering the social relations that mediate the production of exploitable workers and the conditions of their exploitation marginal to the analysis” (Werner 2012: 407). Politically, these scholars criticise that the emphasis on promoting social dialogue in the ‘Decent Work’ approach serves the interests of capital by de-legitimising adversarial bargaining and more militant forms of labour agency (Arnold and Hess 2017: 2191; Standing 2008).

As a second weakness, Marxist scholars hold that the ‘Decent Work’ approach diminishes the role of the agency of labour (Selwyn 2013: 83). In the institutional notion of governance promoted by the ‘Decent Work’ approach, labour is conceived as just one among many institutional actors that could possibly work together to strengthen regulatory governance and, thus, constrain “the behaviour of profit-seeking firms” (Mayer and Pickles 2010: 2). At the same time, the ‘Decent Work’ approach is based on the assumption that firms cooperate with other stakeholders and engage in regulatory governance, not necessarily because they are forced to do so by labour and/or other actors, but because “capital can be persuaded that workers are vital to its reproduction […] and by extension deserving of socio-economic rights” (Arnold and Hess 2017: 2191). Due to this assumption of possible shared interests between capital and labour, according to Marxist scholars, the ‘Decent Work’ approach ignores that in reality the institutional arrangements regulating the labour process are often the result of potential or real struggles between capital and labour (Selwyn 2013: 83). As a result, studies approaching labour through a ‘Decent Work’ approach have tended to produce top-down strategies for improving workers’ conditions, which focus on the collaboration between elite bodies (such as lead firms, governments and international organisations), while discounting the role of workers and trade unions (Selwyn 2013: 76).

In summary, in the perspective of the ‘Decent Work’ approach, improvements in workers’ conditions can be brought about through market pressures on firms to develop workers’ capacities on the one hand, and through strengthening regulatory governance institutions, particularly at the national and international level, on the other. Therefore, even though labour is conceived as a ‘social agent’ entitled to rights, workers and their organisations only play a minor role as active agents in the framework of the ‘Decent Work’ approach (Selwyn 2016: 792ff.).

Building on the criticisms of the ‘Decent Work’ approach, I argue in this book that to understand the nature of ‘indecent work’ and to conceive effective strategies for improving workers’ conditions, we need to give analytical priority to labour in two regards. First, when analysing the roots of ‘indecent work’, we need to prioritise to the conditions of the labour process “as a fundamental process of creating surplus value under capitalism that is at the heart of all systems of commodity production” (Cumbers et al. 2008: 371). Secondly, to develop effective strategies for improving workers’ rights, priority needs to be given to the agency of workers and their organisations, i.e. trade unions. To summarise, “the social relations of production, class conflict and resistance” (Cumbers et al. 2008: 372) and the “politically contested state-capital-labour relations” (Arnold and Hess 2017: 2184) should be at the core of analysing the conditions and role of labour in GPNs.

At the same time, these two arguments are the main ideological tenets of the “Marxist Political Economy” approach introduced in the next section.

2.2 The ‘Marxist Political Economy’ Approach to Labour in GVCs/GPNs

Studies on labour in GVCs/GPNs adopting a ‘Marxist Political Economy’ approach are underpinned by the assumption that social and economic phenomena are shaped significantly (yet not exclusively) by the nature of capital-labour relations (Swyngedouw 2003: 44). As a result, studies on labour in GVCs/GPNs taking on a ‘Marxist Political Economy’ approach are generally concerned with delivering a critical analysis of social relations of production, class conflict and resistance, and of the resulting material conditions under capitalism (Cumbers et al. 2008: 372). Production is defined in this context in its most general sense as “any human activity of formation and transformation of nature and includes physical, material, and social processes as well as the human ideas, views and desires through which this transformation takes place” (Swyngedouw 2003: 44). The term production, thus, represents all forms of economic activity, not only limited to the production of physical goods. The production process is furthermore conceptualised as an integral part of a set of wider social, political and environmental processes and relations, of which capital-labour relations are the most decisive ones in capitalist societies (ibid.). The ‘Marxist political economy’ approach, thus, adopts a fundamentally relationist view of the economy, which is thought of as an interlinked network of processes and relations of production, exchange and consumption (MacKinnon and Cumbers 2011: 29).

The intellectual foundations of the ‘Marxist Political Economy’ approach to labour in GVCs/GPNs lie—as the name suggests—in Marxian political economy, history and theory, as well as in several academic strands that build on the work of Marx, such as Marxist economic geography, labour geography and Labour Process Theory (LPT). While recognising that capitalist relations are historically and spatially contingent and intersect with culturally specific relations of gender, ethnicity, etc. (MacKinnon and Cumbers 2011), geographical studies taking on a ‘Marxist Political Economy’ approach still adopt six rather universalist theoretical assumptions—originating from Marxian theory—about the nature of social and economic relations under capitalism.

The first and most important assumption is that under capitalism, the individual and collective form of production is characterised by “a fundamental social division between those owning the means of production (capitalists), and those only owning their labour, which they need to sell to as labour force to capitalists in order to secure their own short- and medium-term survival” (Swyngedouw 2003: 44).

Second, the socially accepted goal and driving force of processes of production, exchange and consumption under capitalism is profit-making. Hence, the capitalist market economy is necessarily expansionary and growth oriented (Swyngedouw 2003: 47).

Third, profit in the form of surplus value is generated in the labour process through the transformation of labour power (or abstract labour) into actual work (or concrete labour) (Thompson 2010). Although surplus is generated by living labour, it is appropriated in the form of profit by the owners of capital. Therefore, the labour process under capitalism is inherently exploitative in nature since workers only receive a part of the value that they generate in the form of wage or salary (Swyngedouw 2003: 47).

From this follows the fourth assumption, that capital-labour relations under capitalism are inherently antagonistic and conflictive due to opposed interests: Whereas capitalists seek to maximise the generation of surplus to ensure profits and investments and thus the process of capital accumulation, workers seek to ensure their own reproduction, i.e. their short-term and mid-term survival for which they need means in the form of salaries or wages. This conflict of interests leads to continuous tension and potential labour unrest (Swyngedouw 2003: 47f.).

Fifth, in addition to the inter-class struggle between labour and capital, economic relations under capitalism are characterised by the intra-class struggle between individual capitalists competing over the conditions of surplus production, appropriation and transfer (Swyngedouw 2003: 48).

From this double nature of inter- and intra-class struggle then results the sixth assumption that to ensure the process of capital accumulation, capital needs to exercise control over labour (Cumbers et al. 2008: 370). To ensure continued generation of profits, capital needs to overcome the ‘indeterminacy’ of labour, which results from two characteristics of labour as a special production factor. On the one hand, as Thompson and Smith (2009: 924) point out, “hiring labour power does not guarantee an automatic outcome or product for the buyer, as the capacity to work remains within the person of the worker”. On the other hand, workers under capitalism have the burden and freedom to decide to which capitalist they want to sell their labour power. As a result, capital needs to strive to control labour time in the production process, on the one hand, and over the deployment of labour in labour markets on the other hand (ibid.).

Starting from these six basic assumptions, Marxist economic geographers such as David Harvey (1982), Doreen Massey (1984), Jamie Peck (1989, 1992) and Andrew Jonas (1996), and labour geographers such as Andrew Herod (1997, 2001b) and Noel Castree (2007) have further developed Marxian theory by theorising the role of space, place and scale in constituting and shaping capital-labour relations. These scholars have pointed at the variety of geographically and historically specific forms of capitalist production systems in which class relations intersect with locally specific ‘cultural’ relations, such as relations of ethnicity or gender (see e.g. Hudson 2004; Jonas 1996; Massey 1984). Moreover, Marxist economic geographers and labour geographers have advanced our understanding of how spatial asymmetries are crucial in constructing and reproducing asymmetrical capital-labour power relationships (see e.g. Castree et al. 2004; Herod 2001a): Under globalised capitalism, the spatial asymmetry between capital and labour, i.e. the relative mobility of capital in comparison to the relative immobility of labour, has been aggravated due to new developments in logistics and information and communication technologies, allowing capital to set up, manage and control geographically dispersed global production networks. However, as Harvey (1982) points out, even in the era of globalisation, capital is never completely mobile since it depends on locally fixed material infrastructure and institutional settings for production and the labour process to take place. Capital is thus caught in a permanent tension between the need for being fixed in one place for a sustained period, and the need for mobility to seek locations offering more cost-efficient conditions for production.

Drawing on these central assumptions about capital-labour relations under capitalism, the ‘Marxist Political Economy’ approach to labour in GPNs is based on a two-fold notion of labour. From a ‘Marxist Political Economy’ perspective, the notion of labour encompasses, on the one hand, the labour process as “fundamental process of creating surplus value under capitalism”. On the other hand, the notion of labour refers to workers and their organisations as sentient socio-economic actors, who actively shape the geographies of capitalism (Cumbers et al. 2008: 371f.). This two-fold notion of labour has in turn given rise to two distinct strands of work within the ‘Marxist Political Economy’ approach to labour in GVCs/GPNs: (1) a strand of work concerned primarily with the dynamics of the labour process and mechanisms of labour control in GVCs/GPNs; (2) a strand of work concerned with conditions and strategies for labour agency in GVCs/GPNs. Studies within the first strand of work concerned with labour control in GVCs/GPNs have focussed on institutionalised labour control dynamics at the local, regional and national level within GVCs/GPNs that ensure the reproduction of the labour process at specific nodes of a GPN (see e.g. Baglioni 2018; Pattenden 2016; Smith et al. 2018; Wickramasingha and Coe 2021). Labour control dynamics can be defined most broadly in this context as encompassing, on the one hand, dynamics of exploitation, i.e. dynamics that ensure the production of surplus value, and, on the other hand, dynamics of disciplining, i.e. dynamics that mitigate or prevent workers’ resistance (c.f. Baglioni 2018). Studies within the second strand concerned with the conditions and strategies for labour agency in GVCs/GPNs have in turn focussed on workers’ and unions’ strategies for improving their material conditions within GVCs/GPNs (see e.g. Alford et al. 2017; Cumbers et al. 2008; Hastings 2019; Pye 2017).

Whereas both strands are underpinned by the assumptions of the ‘Marxist Political Economy’ introduced above, the research strands on labour control in GVCs/GPNs and on labour agency in GVCs/GPNs, however, share intellectual properties with rather different schools within (neo-)Marxist research. Studies concerned with labour control in GVCs/GPNs draw predominantly on work from Marxist economic geography and Labour Process Theory (Jonas 1996; Kelly 2001, 2002; Thompson and Smith 2009; Thompson 2010. Studies concerned with labour agency in GVCs/GPNs in turn closely engage with work from Labour Geographies, a sub-discipline within economic geography that is concerned with highlighting the active role of workers in shaping economic landscapes (Herod 1997, 2001a; Castree et al. 2004; Castree 2007; Coe and Jordhus-Lier 2011).

It can probably be attributed to the different intellectual traditions of these strands that hitherto interaction between studies on labour control and on labour agency in GVCs/GPNs has been rather limited.¹ I argue here that a closer engagement between both schools can be fruitful for advancing our understanding of the constraining and enabling conditions for local union agency in garment-producing countries. Analysing institutionalised dynamics of labour control can help us to better understand the conditions that constrain collective worker organisation and processes of building union bargaining power at a particular node within the garment GPN. Moreover, an enhanced understanding of labour control dynamics can help us to unveil the specific practices of capital and state actors that (re-)produce labour exploitation in GPNs—and therefore allow us to identify potential target points for labour action. In turn, a focus on labour agency—i.e. on the practices and actions of workers and unions—is useful to identify strategic approaches that enable workers and unions to build sustained bargaining power, allowing them to successfully contest practices of labour control.

With this taken into consideration, this book contributes to the ‘Marxist Political Economy’—strand of research on labour in GVCs/GPNs by linking dynamics of labour control and strategies for labour agency. To this end, this study draws on central ideas and concepts from both research on labour control and labour agency in GVCs/GPNs and develops a relational approach to labour control regimes and labour agency in GVCs/GPNs. However, before doing so in Chapter 3, the next section first provides an overview of the conceptual and empirical contributions, and of the shortcomings of existing studies on labour control (Sect. 2.3) and labour agency (Sect. 2.4).

3 Research on Labour Control in GVCs/GPNs

Labour control emerged as a popular research subject within social sciences in the 1970s and 1980s, with the inception of Labour Process Theory (LPT) by critical industrial sociologists. The research agenda of LPT can be summarised in most general terms as explaining the “nature and transformation of labour power under capitalism” (Thompson 2010). LPT highlights labour control as a central condition for the transformation of labour power in the labour process, following the assumption that market mechanisms alone cannot hedge the ‘indeterminacy of labour’ (see Sect. 2.2.2). As a result, there is a ‘control imperative’ which compels capital to implement management systems to reduce the ‘indeterminacy gap’ (Thompson 2010: 10).

One of the most influential and founding contributions to LPT, which has also been influential in debates on labour control in GVCs/GPN, has been made by Michael Burawoy’s (1979, 1985) typology of different ‘factory regimes’. In this typology, Burawoy (1985) links control dynamics of the labour process with external factors such as dynamics of inter-firm competition, mode of reproduction of labour power and forms of state intervention. He distinguishes between two basic types of factory regimes based on two distinct modes of including workers into the labour process: despotic factory regimes based on coercive work and hegemonic factory regimes based on consenting work. Despotic factory regimes tend to exist within political economic systems characterised by ‘market despotism’, in which capital-labour relations are predominantly mediated through labour markets, with the state being absent as a regulating instance. Hegemonic factory regimes, in contrast, emerge under political economic systems characterised by ‘hegemonic regimes’, in which the state plays an active role in mediating capital-labour relations through the provision of welfare and social security, labour rights and legislation for collective bargaining. As a result, in hegemonic regimes, capital is compelled to coordinate its interests with those of labour and to take measures to persuade workers to take part in the labour process and consent to their own exploitation. Whereas historically under capitalism, despotic factory regimes prevailed during the period of industrialisation, hegemonic factory regimes prevailed under Fordism. For the current period of globalised capitalism, Burawoy argues that a third, new type of factory regime characterised by ‘hegemonic despotism’ is likely to emerge, which is characterised by hybrid elements of coercion and consent under new, harsh market conditions (Burawoy 1985: 122–129). This new factory regime under globalised capitalism is hegemonic in so far as that consent is more dominant than coercion in the labour process. However, it is at the same time despotic since capital uses its relatively higher mobility to extract concessions from relatively immobile labour. The capitalist period of ‘hegemonic despotism’ is thus characterised by an increasing shift of production to developing and newly industrialising countries as well as by an accompanying process of undermining and undercutting of labour standards (Kelly 2001: 3).

Contrary to Burawoy’s early attempts to connect the dynamics of the labour process to the wider political economy, subsequent studies in LPT have, however, adopted a rather narrow focus on dynamics of control, consent and resistance at the point of production, while neglecting dynamics outside of the workplace. This narrow focus on the workplace in mainstream contemporary LPT can be attributed to the paradigm of the ‘relative autonomy of the labour process’ underpinning contemporary LPT studies. This paradigm is based on the assumption that “similar external situations can produce different internal labour process outcomes because of the distinctiveness and peculiarities of particular workplaces” (Taylor et al. 2015: 4; see also Edwards 1990). As a result, contemporary LPT has for a long time been perceived as “less equipped to address, […] the varieties of (often informal or unwaged) types of work, [and] temporal and spatial dimensions” of labour control (Thompson and Smith 2009: 917).

Over the past decade, however, a dialogue between LPT and GVC/GPN literature that is swiftly gaining more traction has addressed these shortcomings, while also sharpening the attention of GVC/GPN studies for the social relations of production (Cumbers et al. 2008; Newsome et al. 2015; Rainnie et al. 2011; Selwyn 2013). Marxist economic geographers and LPT theorists have repeatedly called for a greater focus on labour process dynamics as crucial for the understanding of the structure and functioning of GPNs (Cumbers et al. 2008: 371f.; Hammer and Riisgaard 2015: 89; Rainnie et al. 2011: 160; Selwyn 2013: 87). In this line, Cumbers et al. (2008), for example, point out that the entire rationale for capital restructuring in the form of outsourcing and setting up GPNs is the need for capital to overcome the indeterminacy of labour:

In the abstract, capital restructuring is always […] about being ‘in flight from labour’ or rather is a response to the problems capital comes up against in extracting surplus value through exploitation of labour in production. Whether through the imposition of new technical or spatial fix (Harvey, 1982), capital is viewed as responding to the problem of labour control. (Cumbers et al. 2008: 372)

This revived interest in labour control from GVC/GPN scholars has given rise to a significant body of work analysing mechanisms, dynamics and frameworks of labour control in the context of GVCs/GPNs over the past ten years. In particular, work by economic geographers drawing on the concept of (local) labour control regimes (Jonas 1996) has contributed to broadening the notion of labour control: Whereas LPT-informed studies tended to focus on intersections between value chain dynamics and workplace labour control dynamics, studies of labour control regimes shifted the analytical focus to capital and state strategies directed at securing the broader conditions for capital accumulation at the local and national level (see e.g. Baglioni 2018; Neethi 2012; Smith et al. 2018). In this vein, Baglioni (2018) has stressed that all GPNs are underpinned by complex ‘architectures of labour control’ that emerge from the interplay between dynamics of exploitation and disciplining at various levels.

In the light of the diversification of analytical perspectives and concepts within the broader research field on labour control in GVCs/GPNs, I propose that we can distinguish between three sub-strands: (1) a research strand at the intersection of GVC analysis and LPT that is concerned with how value chain dynamics shape labour control in the workplace; (2) a research strand drawing on the concept of labour control regimes that is concerned with how the relation between ‘global capital’ and ‘local labour’ is mediated by local and national actors; and (3) a research strand that combines the concept of labour control regimes with the multi-scalar perspective of the GPN framework. Figure 2.3 provides a graphic representation of these three research strands on labour control in GVCs/GPNs.

Fig. 2.3

Source Author

Research strands on labour control in GVCs/GPNs.

The following sections introduce the theoretical frameworks and assumptions of each research strand and point out their main empirical insights and limitations regarding their potential for identifying constraints and potential target areas for labour agency in the context of GVCs/GPNs.

3.1 Studies from Labour Process Theory: Approaching Labour Control in GVCs/GPNs with a Focus on the Workplace

The first research strand, which is concerned with labour process and control dynamics in the workplace and how these are shaped by broader dynamics of GVC governance, is primarily informed by the assumptions and research agenda of Labour Process Theory (LPT). As aforementioned, LPT has traditionally focussed on the dynamics of the labour process and of control, consent and resistance at the workplace. In this vein, the rather young strand of LPT studies focussing on labour control in GVCs/GPNs aims to reveal how managerial control practices and labour processes in specific sectors are shaped by dynamics of value chain restructuring in the face of new competitive pressures resulting from globalisation.

In principle, it can be said that LPT-informed studies share two basic assumptions about the role of the labour process under globalised capitalism: First, to remain competitive under globalised capitalism, firms must organise and coordinate labour processes at the different stages of the value chain in a way that ensures maximum surplus value (Hammer and Riisgaard 2015: 97). Second, the restructuring of GVC governance and inter-firm relations has led to greater competition among suppliers and generated increased cost pressure in many sectors. As a result, GVC restructuring has affected how relations of production and control over the labour process are coordinated around the globe (Hammer and Riisgaard 2015: 84). In particular, LPT-informed studies have highlighted three interrelated processes that shape the relations of production and the organisation of the labour process under globalised capitalism: First, the increased fragmentation of production on a global scale has led to a reordering of how different labour processes are linked and compete with each other. Second, new competitive challenges for firms resulting from outsourcing and upgrading processes lead to a re-segmentation of the workforce along lines of employment status, type of contract, etc. As a result, we can, third, observe an increased ‘tiering’ of the workforce whereby workers performing essentially equivalent tasks are divided by a range of different employment statuses (Hammer and Riisgaard 2015: 90). From the perspective of LPT, this fragmentation and segmentation of the labour process and of the relations of production represent capital strategies that ensure continued value extraction from ‘living labour’ under changing structural conditions in the era of global capitalism (Bair and Werner 2015: 131).

With this in mind, LPT-informed studies of labour control in GVCs/GPNs have made three important contributions to advancing our understanding of labour control dynamics in GVCs/GPNs at the workplace scale. First, LPT-informed studies have produced valuable insights into the effects of lead firms’ practices and strategies of outsourcing, off-shoring and subcontracting on employment relations and labour processes at supplier firms. Particularly in captive value chains, lead firms frequently exercise significant price pressure over suppliers or subcontractors, which in turn leads to the rationalisation and flexibilisation of labour processes (Flecker and Meil 2011). Haidinger and Flecker (2015) point out in this regard that increased workforce segmentation resulting from the combination of various types of outsourcing—including subcontracting firms, temporary employment agencies or self-employed work—also serves as a disciplining mechanism since it hampers collective worker organisation.

With regard to empirical insights into the influence of lead firms in the garment GVC/GPN on labour processes and employment relations at suppliers, a particularly comprehensive study has been provided by Anner (2019). He illustrates how fashion retailers’ ‘predatory purchasing practices’—including ‘price squeeze’, demands for shorter lead times, fluctuations in order volumes and changes to product specifications at short notice—lead to an ‘employment relations squeeze’ in the Indian export-garment industry. To respond to cost pressures and fluctuations in demand, Indian garment manufacturers rely on several practices of ‘squeezing’ workers, including informal employment and piece-rate work, gender-based forms of exploitation and ‘wage theft’ practices (Anner 2019: 707f.).

The second contribution of LPT-informed studies of labour control in GVCs/GPNs lies in their description of the ‘new factory regimes’ and practices of managerial control in the labour-intensive export-manufacturing sectors of the Global South. This contribution has been centrally informed by Burawoy’s (1985: 263ff.) argument that many countries of the periphery are characterised by “political orders which would nurture repressive factory regimes”. Hence, according to Burawoy, labour control in peripheral countries tends to be exercised through “brutal coercion at the point of production”—as opposed to the mix of consent and coercion prevalent under ‘hegemonic despotism’ in industrialised countries (Burawoy 1985: 265). Drawing from this argument, various LPT-informed studies of labour control in GVCs/GPNs have drawn attention to ‘new factory regimes’ in labour-intensive export-manufacturing of the Global South. These new factory regimes are characterised by “strategies of control [..] that go far beyond the fairly regulated terrain of the workplace and the employment relationships in the Global North” (Hammer and Riisgaard 2015: 91; Anner 2015b; Jenkins and Blyton 2017).

With regard to ‘new factory regimes’ in garment-producing countries, Anner (2015a) develops a three-fold typology of factory regimes in the garment industry (which he calls ‘labour control regimes’²) that distinguishes between state, market and employer regimes. In state labour control regimes, labour is controlled by a system of legal and extra-legal mechanisms that prevent or curtail worker organisation and collective action. State labour control regimes can, thus, be found particularly in countries with authoritarian governments, such as China or Vietnam. In market labour control regimes, in turn, labour is disciplined by unfavourable market conditions installing fear of job loss and resulting un- or underemployment in workers, as is the case in countries with a large ‘reserve army’ of labour, such as India, Bangladesh or Indonesia. Lastly, employer labour control regimes are characterised by highly repressive employer actions against workers, including the use or threat of violence. It is important to note, however, that although Anner proposes this typology to distinguish between different labour control regimes, he also stresses that the three forms of control (i.e. state, market and employer control) are not mutually exclusive or static. Rather all countries have elements of each system.

As a third contribution, LPT-informed studies concerned with labour control in GVCs/GPNs have drawn attention to the subjectivity of labour exploitation and disciplining (Burawoy 1985). The subjectivity of labour exploitation and disciplining results from their intersection with other elements of workers’ identity, such as gender, age, religion or ethnicity. Employers can strategically deploy these identity features to duplicate disciplinary structures rooted in wider social relationships within the workplace. In this light, several LPT studies have sought to expose how “the creation of value from heterogeneous living labour depends upon […] [the] ideological power that is effected through constructions […] of gender, […] racialisation, heteronormativity, and other forms of social difference” (Werner 2012: 408; see also McGrath 2013).

Concerning the subjectivity of labour exploitation and disciplining in the export-garment industry, LPT-informed studies have pointed in particular at the intersections between gender and exploitation in the labour process. In this vein, Werner (2012), for example, illustrates how the upgrading from assembly to full-package production in a Dominican garment factory is linked to a restructuring of the labour process, which gives rise to a new segmentation of the workforce along intersecting lines of gender, skill- and pay levels. Jenkins (2015) adds to Werner’s observations by highlighting how employers in the South Indian export-garment industry base their competitive strategies on the use of ‘dis-empowerment’ as a mechanism of labour control. Employers hire predominantly women because they are associated with greater distance from access to employment and representation rights and are thus perceived to be less likely to organise or to cause unrest at the workplace.

In summary, LPT-informed studies have brought the workplace back into GVC/GPN analysis and thereby advanced our understanding of labour control dynamics in GVCs/GPNs in three ways. Firstly, LPT-informed studies have opened up analytical space for exploring the interrelations between lead firm strategies under globalised capitalism and new work organisation and managerial control forms. Second, LPT-informed studies have sharpened our understanding of the coercive managerial strategies that characterise ‘new factory regimes’ in the export industries of the Global South ‘at the bottom’ of buyer-driven GVCs. Lastly, LPT-informed studies drawing from newer feminist and anthropological perspectives have highlighted the intersections of exploitation processes at the point of production with wider social relations of gender or race.

Regarding our understanding of the constraining factors and potential target areas for labour agency in garment-producing countries, LPT-informed studies hence have two important implications. First, findings from LPT-informed studies highlighting the influence of lead firms’ practices on suppliers’ labour processes imply that workers and unions at specific nodes of a GPN need to simultaneously target factory managers’ and lead firms’ practices to achieve long-lasting changes in the labour process. Second, by highlighting the diverse disciplining mechanisms that hinder collective labour organisation in garment-producing countries—including labour market pressures, employer and state repression and managers’ gendered ‘dis-empowerment’ strategies—LPT-informed studies have provided important insights into the factors that constrain labour power in garment-producing countries.

Notwithstanding these critical contributions, LPT-informed studies show two crucial limitations when it comes to understanding the nature of the ‘labour control architectures’ underpinning GPNs. First, due to their rather exclusive focus on value chain and workplace dynamics, LPT-informed studies have neglected an important dimension of labour control in GPNs. Labour control not only encompasses employers’ and lead firms’ exercise of direct control over labour, it also encompasses capital and state strategies directed at securing the broader conditions that allow lead firms and domestic firms to reproduce exploitative labour processes (c.f. Neethi 2012: 1241). Considering this, it seems necessary to broaden the analysis of labour control in GVCs/GPNs to account for the role of capital and state actors at the horizontal dimension, who secure the broader conditions and social relations for capital accumulation at specific nodes of a GPN. Second, LPT-informed studies of labour control in GVCs/GPNs have paid little attention to the spatial characteristics of labour control. For example, LPT studies have provided little insight into how labour control dynamics at the workplace may vary across various nodes of GPNs due to the territorial embeddedness of labour processes into place-specific social relations and regulatory frameworks.

To address these shortcomings, I argue that a more geographical and therefore spatially sensitive approach to labour control in GVCs/GPNs can be helpful. Such an approach has been developed by the second strand of work concerned with labour control in GVCs/GPNs. This strand builds on the concept of labour control regimes as a heuristic for analysing the architectures of labour control underpinning GPNs. It will be introduced in the next section.

3.2 Spatial Approaches to Labour Control: National and Local Labour Control Regimes

The second research strand on labour control in GVCs/GPNs draws on the conceptual framework of labour control regimes to explore territorially embedded, institutionalised frameworks for capital accumulation at specific nodes of a GPN (Azmeh 2014; Neethi 2012; Padmanabhan 2012; Smith and Pun 2006). The theoretical concept of labour control regimes has been originally developed by Jonas (1996, 2009) and Peck (1992, 1996). In the most general manner, labour control regimes can be defined as stable institutional frameworks for accumulation and labour regulation constructed around national and local labour market reciprocities (Jonas 1996: 323).

Studies concerned with labour control regimes in the context of GPNs share two central assumptions. The first assumption is that—despite the ability of global capital to move to (and between) countries in the periphery—global capital still depends on ‘spatial fixes’ to realise value extraction from labour (Harvey 1982, 2001). These spatial fixes can only be realised through the engagement of global capital with actors in production countries, who ensure labour supply, regulation and disciplining (Kelly 2001: 3). The relation between ‘global capital’ and ‘local labour’ hence needs to be understood as enabled and mediated through various actors, relationships and institutions at different levels (Kelly 2001: 2). Against this backdrop, the concept of labour control regimes was also introduced explicitly in response to Burawoy’s (1985) thesis that coercive labour control in the periphery is enabled by repressive state environments. Kelly (2001: 3) points out in this regard that “rather than simply being sites of oppression and coercion, […] new destinations for global capital require a new and more or less stable regime of social regulation for labour control to be put in place”. These regimes are not solely based on providing “the cheapest and most unregulated economic environment” for lead firms and suppliers, but also encompass strategies of active regulation of the labour market, for example through measures directed at promoting “productivity enhancement, skill development and innovation” (ibid.).

The second shared assumption of studies of labour control regimes in GPNs is that labour control is an “irretrievably […] spatial process” (Jonas 1996: 328) for two reasons. On the one hand, labour control is spatial because it is territorially embedded in labour control regimes that emerge in place-specific form from reciprocities constructed between the spheres of production, regulation and reproduction in a specific locale (Jonas 1996: 325). As a result, labour control regime literature rejects Burawoy’s (1985) thesis of a universal regime of ‘despotic hegemonism’ under globalised capitalism. Instead, studies of labour control regimes emphasise that “there is no one grand institutional fix to the problem of labour control but rather multiple fixes constructed in different ways in different places (and at different scales) by different agencies” (Jonas 1996: 331). On the other hand, labour control is also an irretrievably spatial process because the control of space is a central tool to control labour under capitalism (Kelly 2001: 1). Since workers are ‘free’ to sell their labour power to the employer of their choice, regulating and restricting the mobility of labour in space becomes a central instrument of labour control. In this context, broad attention has been given by studies of labour control regimes in GPNs to the role of capital and state-controlled migration flows in producing spatial and institutional fixes for ensuring labour supply (see e.g. Azmeh 2014; Mezzadri 2008, 2017; Padmanabhan 2012).

Hence, whereas LPT-informed studies of labour control focus on the intersection of lead firm and workplace control dynamics, studies of labour control regimes shift the analytical focus from workplaces to ‘work-places’. They focus on local and national actor networks and institutional frameworks securing the broader conditions for capital accumulation at specific nodes of a GPN (Jonas 2009: 64). Figure 2.4 illustrates the two analytical emphases underpinning by LPT-informed studies of labour control in GVCs on the one hand, and studies of labour control regimes in GPNs on the other.

Fig. 2.4

Source Author

Analytical emphases of LPT-influenced literature and labour control regimes literature.

Studies of labour control regimes can be further divided into two sub-strands according to their main scale of analysis and the features of labour control that they emphasise. The first sub-strand of work on labour control regimes is inspired by Peck’s (1989, 1992, 1996) work on national labour market regulation and takes on a national regulatory approach to labour control regimes. This strand emphasises the role of national regulatory frameworks and institutions in ensuring the broader conditions for capital accumulation. In this light, studies adopting a regulatory approach to labour control regimes tend to pay particular attention to the state’s role in balancing labour supply and demand, e.g. through training programmes, employment measures and welfare systems. Whereas these regulatory mechanisms usually function at a national level, Peck, however, stresses that concrete local labour market structures may show variations in different places. Nevertheless, the heuristic entry point for regulationist studies of labour control regimes typically consists in scrutinising the wider institutionalised regulatory frameworks at the national (or international) level and how these materialise in specific (work)places.

The second sub-strand of work on labour control regimes in GPNs, in turn, draws on Jonas’ (1996) concept of the ‘local labour control regime’, which—as the name suggests—takes informal practices and networks at the local level as heuristic entry point for analysing institutionalised labour control frameworks. As opposed to Peck’s focus on national formal regulatory institutions, Jonas emphasises the concrete localised, informal practices and relationships constructed around specific workplaces or local industries as constitutive of local labour control regimes. According to Jonas (2009: 61), “there is a tendency for labour control to stabilise around place-specific social practices, which affect the social integration of labour inside the workplace but influence conditions outside it as well. Whether firm-specific or industry-wide, these practices are locally constructed and become routinised and institutionalised in time and space”. As a result of this tendency for labour control to stabilise around place-specific social practices, local labour control regimes emerge in form of “historically contingent and territorially embedded set[s] of mechanisms which coordinate the time-space reciprocities between production, work, consumption and labour reproduction within a local market” (Jonas 1996: 325). The analytical movement of studies concerned with local labour control regimes thus starts at the workplace or local level. It extends the analysis from there to the wider sphere of the labour market.

The remainder of this section outlines the contributions of the two sub-strands on national and local labour control regimes in GPNs.

3.2.1 ‘Regulationist’ Studies of National Labour Control Regimes

The main empirical contribution of regulationist studies of national labour control regimes has been to challenge the widespread assumption that the intensification of transnational production under global capitalism has been driven by an increasing ‘rollback’ of the state, especially in developing countries. Contradicting this assumption, literature on national labour control regimes has revealed the fundamental “role of the state, local cultures, and specific classes of employers, managers and workers” (Ngai and Smith 2007: 29) in producing fragmented working classes and flexible labour markets, which in turn enable lead firms to outsource labour-intensive production to developing countries. In this context, Mezzadri (2008: 603ff.) points out that whereas global capital may impose a ‘general diktat’ of ‘cheapness’ on labour, this diktat is realised through territorially embedded, place-specific social regulatory mechanisms.

In this context, regulationist analyses of labour control regimes have reiterated the emergence of national migrant labour regimes in major garment-producing countries. Ngai and Smith (2007), for example, illustrate how the competitiveness of the export-manufacturing sector in China is rooted in a specific type of national migrant labour regime, which they designate the ‘dormitory labour regime’. This labour regime captures the labour force of rural migrants for short-term use by Chinese and foreign export factories located in Special Economic Zones as well as in urban industrial areas. Ngai and Smith highlight the role of the Chinese state in regulating the mobility of rural migrants by providing housing and accommodation in the form of state-owned dormitories, which factory-owners can rent for their migrant workers. Under the agency of the state and local capitalists who actively promote, regulate and channel labour mobility from rural to industrial zones, “a hybrid, transient workforce is created, circulating between factory and countryside, dominated by employers’ control over housing needs and state controls over residency permits” (Ngai and Smith 2007: 31).

In her extensive work on the Indian export-garment industry, Mezzadri (2010, 2016, 2017), in turn, highlights the role of the state and local capital in ensuring international competitiveness by promoting the fragmentation and informalisation of labour relations. According to Mezzadri (2017), the increasing informalisation of production has been actively facilitated and bolstered by the Indian state through two (historical) regulation mechanisms. Firstly, by allocating quotas for garment production to (initially only small and medium) individual enterprises, the state has actively promoted the regionalisation and fragmentation of production. As a result, the Indian export-garment production is now scattered across multiple localised networks of small firms and subcontracting units. Second, while formally maintaining relatively strong labour laws restricting, for example, the use of contract work or retrenchments, the state has since the 1970s increasingly allowed capital to circumvent those laws in practice, thereby actively promoting the informalisation of labour relations.

Altogether, regulationist studies of labour control regimes have made an important contribution to our understanding of labour control in garment-producing countries (and other export industries ‘at the bottom’ of GVCs/GPNs) by highlighting the role of the state and of local capitalists in mediating the relationship between ‘global capital’ and ‘local labour’. They have emphasised how labour exploitation and control at the workplace is shaped by broader regimes of labour regulation at the national level. In particular, they have highlighted the important role of migrant labour regimes and the informalisation of employment relationships as broader conditions for reproducing exploitative labour processes in garment-producing countries. By exposing the active role of the state in producing the conditions for labour exploitation, regulationist studies of labour control regimes have highlighted state policies as a field of contestation for workers and unions in garment-producing countries. Moreover, by showcasing the spatial fragmentation of the workforce caused by circulatory migration and informal, home-based work, regulationist studies of labour control regimes have exposed barriers for collective labour organisation in garment-producing countries beyond direct repression by employers and state actors.

However, critics have argued that due to their focus on formal labour regulation and their analytical priority for dynamics at the national level, regulationist studies of labour control regimes have paid too little attention to the highly localised informal relationships and practices that are “often contrary to the formal provisions of national-level labour regulations” (Kelly 2001: 21). This criticism might not be entirely justified when reviewing recent literature on national labour control regimes, which has studied informal exploitation practices of local employers (see e.g. Mezzadri 2016, 2017). Nevertheless, it remains true that the second sub-strand of work on labour control regimes in GPNs drawing on Jonas’ concept of local labour control regimes has shown a much more explicit focus on informal networks and practices. The following section provides an overview of the empirical insights produced with this analytical focus.

3.2.2 Studies of Local Labour Control Regimes

As opposed to regulationist studies of labour control regimes, which predominantly choose the national level as analytical entry point, literature concerned with local labour control regimes focuses on institutionalised frameworks of labour control at the sub-national level. In doing so, studies of local labour control regimes have made two key contributions to advancing our understanding of the ‘architectures of labour control’ underpinning GPNs. First, studies of local labour control regimes have highlighted the importance of everyday informal social and cultural practices (as opposed to formal regulatory mechanisms) for the constitution of institutionalised frameworks of labour control at specific nodes of a GPN (Jonas 1996: 327). Thereby, studies of local labour control regimes have opened analytical space to account for the role of discursive practices in shaping and justifying labour exploitation and disciplining (Coe and Kelly 2002).

Besides highlighting the role of informal cultural practices as constituent parts of labour control regimes, studies of local labour control regimes have, second, underlined the relational nature of local labour control regimes which are conceptualised as emerging from the ‘collective interaction’ of a wide variety of actors (Kelly 2001: 7). As a result, local labour control regimes are at the same time relatively stable, yet open to contestation and change over time (Jonas 1996: 328f.). The concept of local labour control regimes is hence underpinned by a deeply relational understanding of ‘the local’, which considers the “material, social and political form” of the local as “determined by its interactions with local and wider power geometries” (Jonas 1996: 328; see also Massey 1992, 1993).

Several studies have provided empirical insights into these characteristics of local labour control regimes, with particular attention on export industrial clusters in newly industrialising regions in Asian countries (Coe and Kelly 2002; Kelly 2001, 2002; Neethi 2012; Padmanabhan 2012).

With regard to local labour control regimes in the export-garment industry, Padmanabhan (2012) provides insights into the direct practices and indirect mechanisms constituting the local labour control regime in an export-garment park in Kerala, South India. The ‘direct practices’ of labour control in the workplace include inter alia restrictions for leave-taking as well as a strict monitoring of production rates. The ‘indirect methods’ of labour control consist of tailoring recruitment practices to the specific characteristics of the local labour market. Employers exploit the labour market dependence of young, less educated women from economically vulnerable families in rural areas by specifically targeting recruitment strategies at this segment of labour. Workers stay in park-run hostels, allowing employers to dispose flexibly over labour’s time even beyond the official working hours. Padmanabhan also highlights informal networks extending into workers’ sites of reproduction as indirect mechanism of labour control. For garment firms in Kerala’s export promoting parks, fostering relationships with local leaderships of surrounding villages is crucial for ensuring labour supply since in most cases communities collectively take the decision to send young women to garment export parks.

In a nutshell, whether adopting a focus on the formal regulation of labour markets or on informal relationships and practices as means of labour control, studies of labour control regimes have advanced our understanding of labour control in GPNs by highlighting the role of regulatory frameworks, actors, processes and relationships at the national and local level as mediators between ‘global capital’ and ‘local labour’. Moreover, literature on labour control regimes has refined our understanding of labour control dynamics in GPNs by highlighting the geographical variations between labour control regimes constructed around different national and local labour markets, and by exposing the control of space as an essential tool of labour control under capitalism. By highlighting the place-specific dynamics and networks at the horizontal dimension of the GPN, studies of labour control regimes complement studies at the intersection of GVC analysis and LPT, which have tended to neglect the broader political economies into which workplaces are embedded. Regarding the constraining factors and the territory for the agency of local unions in garment-producing countries, the insights from studies on (local) labour control regimes suggest that we should take a more critical stance towards formal labour law as potential power source for local unions. In reality, a variety of informal interactions and practices of state actors and employers tend to make these frameworks ineffective in practice (see e.g. Kelly 2001; Mezzadri 2010). Taking this into consideration, the state appears—besides employers—as a relevant target for local union action.

Despite these important contributions, there are also some shortcomings with studies of national and local labour control regimes in GPNs. Particularly with regard to the originally relational concept of local labour control regimes as embedded within and shaped by ‘wider power geometries’ (Jonas 1996: 329), empirical studies of labour control regimes have not yet lived up to this conception. Despite the explicit framing of (local) labour control regimes as mediating institutions between ‘global capital’ and ‘local labour’, labour control studies have paid little to no attention to how value chain dynamics or lead firm practices shape national or local labour control regimes. Instead, global capitalist dynamics remain in the background of these studies as abstract structural forces that impose the ‘diktat’ to produce cheap and flexible labour on national, regional and local actors (Mezzadri 2008: 614).

Addressing this gap, over the past five years, a still small number of studies has sought to integrate the labour control regimes framework more explicitly with the multi-level heuristic of the GPN framework (Baglioni 2018; López Ayala 2018; Pattenden 2016; Smith et al. 2018; Wickramasingha and Coe 2021). The following section gives an overview of this relatively new branch of research promoting an explicitly ‘multi-scalar’ approach to labour control regimes.

3.3 Multi-Scalar Approaches to Labour Control Regimes

Multi-scalar approaches to labour control regimes in GPNs have emerged since 2016 in an attempt by economic geographers to explore the role of dynamics, actors and practices at various levels in constructing labour control regimes at specific nodes of a GPN in a more in-depth manner (Baglioni 2018; López Ayala 2018; Pattenden 2016; Smith et al. 2018; Wickramasingha and Coe 2021). The first scholar to propose an explicitly multi-scalar framework for analysing labour control regimes was Pattenden (2016: 1813), who suggested a ‘three way approach to labour control regimes’. Pattenden’s approach distinguishes between the scales of the macro-labour control regime, the local labour control regime and the labour process. The macro-labour regime—comprising social relations, the state agenda and formal regulations at the national level—shapes the material conditions of the local labour control regime as well as of the labour process through ‘pro-capital state practices that create and maintain informal flexible working places, dangerous labour processes and poor living conditions’ (Pattenden 2016: 1823).

While Pattenden’s ‘three way approach’ stops at the national level, Baglioni (2018), Smith et al. (2018) and Wickramasingha and Coe (2021) go one step further by integrating the ‘global level’ into their approaches to LCRs in GPNs. In Baglioni’s (2018) concept of the LCR, global dynamics of capital restructuring provide the broader context within which states and transnational firms construct labour control regimes at the national and local level. Driven by global competitive pressures, national states construct regulatory frameworks and development agendas, which in turn pave the way for the agency of transnational firms. Transnational firms, in return, construct locally specific labour control regimes by linking labour exploitation at the firm level with strategies of spatial and social disciplining extending to the reproductive sphere (Baglioni 2018: 113ff.).

In contrast, Smith et al.’s (2018) ‘nested scalar approach’ to labour control regimes in GPNs takes labour control dynamics at the workplace as an analytical starting point. Smith et al. then suggest analysing how the workplace labour control regime—including the labour process, wage relations and forms of worker representation—is shaped on the one hand by dynamics of the ‘regional and national political economy of labour control’ and by global lead firm dynamics, on the other. Similar to Jonas’ (1996) conception of the local labour control regime, Smith et al. (2018: 558) conceptualise the ‘regional political economy of labour control’ as comprising dynamics of social reproduction and local labour markets as well as local political contexts. The ‘national political economy of labour control’ is constituted through formal regulatory mechanisms such as national labour laws, labour inspectorates and state policies relating to labour standards as well as through ‘the balance of class forces’ (ibid.). The workplace labour regime and the regional and national political economy of labour control are in turn shaped by ‘lead firm dynamics’ comprising supply chain pressures and governance dynamics as well as lead firms’ codes of conduct and lobbying practices (ibid.) Fig. 2.5 visualises the different levels of labour control and their articulations in Smith et al.’s ‘nested scalar approach’ to labour control regimes.

Fig. 2.5

Source Adapted from Smith et al. (2018: 558)

‘Nested scalar approach’ to labour control regimes by Smith et al. (2018).

From an empirical point of view, Smith et al.’s (2018) work has contributed to our understanding of the complex, multi-scalar constitution of labour control regimes in garment-producing countries. Applying their conceptual framework to the Moldovan export-garment industry, Smith et al. illustrate how workplace regimes in Moldovan garment factories are characterised by ‘poverty wages’, extensive and regular use of overtime, intensification of the labour process and employer resistance to union formation. Smith et al. then identify how these workplace regimes are shaped by lead firms’ demands at the global level and by regional and national dynamics and institutions of labour regulation. On the one hand, lead firms’ demands for lower prices, higher product flexibility and shorter production times require manufacturers to maximise productivity while keeping labour costs down. On the other hand, poor workplace conditions are shaped by the weakening of labour inspectorates and low national minimum wages due to the asymmetric power balance between employer associations, the state and labour. With their analysis, Smith et al. provide a good understanding of how more general ‘network dynamics’ intersect with place-specific labour control dynamics at the national and workplace level.

In summary, multi-scalar conceptions of labour control regimes have embedded workplace control dynamics within broader dynamics of labour control at the regional, national and global levels. Moreover, they have promoted a spatial understanding of labour control: labour control structures in GPNs are territorially embedded into and shaped by place-specific social relations and power structures on the one hand, and by global dynamics of capital accumulation on the other hand. Thereby, multi-scalar approaches to LCRs have enhanced our understanding of the multiple processes and dynamics at different scales involved in constructing labour control regimes at specific nodes of the GPN. Moreover, particularly Smith et al.’s (2018) analysis has provided some initial insights into the links between labour process dynamics, lead firm practices and broader political economic dynamics of labour control at the regional and national level. Nevertheless, only recently a pioneering study by Wickramasingha and Coe (2021) has highlighted the links between dynamics at various levels as central for the constitution of labour control regimes. Hence, a more systematic engagement with the relational nature of labour control regimes at specific nodes of the garment GPN is necessary.

3.4 Interim Conclusion: Contributions and Shortcomings of Existing Analytical Approaches to Labour Control in GPNs

This section has introduced three different strands of research on labour control in GPNs and highlighted the insights gained from these research strands as well as their shortcomings: (1) an LPT-informed research strand focussing on the intersection between value chain and workplace control dynamics; (2) a research strand focussing on national and local labour control regimes; and (3) multi-scalar approaches to labour control regimes at specific nodes of a GPN.

In a nutshell, these research strands have enhanced our understanding of the constraining conditions and the terrain for worker and union agency in global production countries in two regards. First, existing research on labour control in GVCs/GPNs has highlighted the interrelations between global value chain dynamics and workplace exploiting and disciplining dynamics. In particular, past studies have illustrated how lead firm pressures for lower prices and shorter production times give rise to workplace regimes characterised by informalised employment relations, low wages, gender-based exploitation and violations of workers’ collective rights.

Second, in particular studies concerned with local, national or multi-scalar labour control regimes at specific nodes of a GPN have drawn attention to the fact that structures of exploitation and disciplining are not only produced by lead firms at the global scale and by employers in the workplace. They are also produced by various other actors at multiple scales who fulfil central functions in reproducing the broader social relations that secure capital accumulation at specific nodes of a GPN. Hence, to understand the terrain and constraining factors for the agency of local unions in garment-producing countries, we need to look at the practices of these actors and scrutinise their implications for worker and union agency—either concerning the specific constraints they pose for building local unions´ bargaining power or as potential target areas for union interventions.

Despite these critical overall contributions to the literature on labour control in GVCs/GPNs, I argue that three aspects of the ‘labour architectures underpinning GPNs’ remain underdeveloped in existing studies of labour control (regimes) in GPNs. First, while existing research on labour control in GPNs has revealed the diversity of actors and dynamics at multiple scales involved in the production of labour control regimes, the interrelations and interdependences between labour control dynamics at different levels have so far remained under-researched. I hold that this analytical gap results at least in part from the scalar heuristic underpinning past studies of labour control regimes: The focus on a priori-defined scalar categories creates an artificial analytical separation between processes and dynamics conceptualised as located at different levels (c.f. Marston et al. 2005: 442). However, as Wickramasingha and Coe (2021: 6) have recently stressed, place-specific labour (control) regimes are in reality constituted through “a mix of geographically distant and proximate relations across different scales”. In this light, I argue that to gain a more nuanced understanding of the ‘architectures of labour control’ underpinning specific GPNs and of the implications for labour agency, a relational analytical perspective that emphasises connectivity rather than a priori defined scalar scaffolds can be beneficial.

Second, scalar studies of labour control regimes in GPNs tend to assume a universal socio-spatial order as characterising labour control regimes in GPNs (c.f. Latham 2002: 138). This assumption is also closely intertwined with the pre-supposed ‘nested multi-scalar’ structure of labour control regimes (c.f. Smith et al. 2018). I argue that such a pre-given assumption of a universal socio-spatial order of labour control regimes constrains the analytical space for empirically carving out the place-specific articulations of geographically more proximate and more distant relationships that constitute labour control regimes at specific nodes of a GPN (c.f. López 2021; Wickramasingha and Coe 2021).

Lastly, the nested multi-scalar perspective informing existing studies of labour control regimes in GPNs has tended to reproduce a hierarchical ‘top-down’ conceptualisation of ‘global/local’ dynamics, in which labour control dynamics at higher levels are unilaterally imposed on local labour (Hastings and MacKinnon 2017: 105). Such a ‘top-down’ understanding, however, completely ignores the role of workers’ as sentient social actors who actively shape economic landscapes, including labour control regimes (Herod 1997; Wickramasingha and Coe 2021). Under repressive state and employer regimes, workers and unions may not have the power to challenge institutionalised frameworks for labour control in their totality; however, workers and unions may still succeed in shaping and transforming specific elements of the labour control regime by stopping selected practices of exploitation. Thus, I argue that to gain a better understanding of how workers and unions in garment-producing countries can achieve lasting improvements for workers, we need an enhanced understanding of the conditions that enable workers to achieve even ‘small transformations’ of labour control regimes at specific nodes of the GPN (c.f. Latham 2002).

Table 2.2 on the next page provides an overview of the central research questions, contributions and shortcomings of the three research strands/approaches introduced in this section.

Table 2.2 Overview of different research strands on labour control in GVCs/GPNs introduced in this section

To tackle these shortcomings in existing research on labour control in GPNs, and to produce a more nuanced understanding of labour control regimes as relational-structural contexts for worker and union agency in garment-producing countries, I develop a relational analytical approach to labour control regimes in GPNs (see Chapter 3). This approach conceptualises labour control regimes at specific nodes of a GPN as emerging from the place-specific articulations of multiple processual relations of labour control that stretch across various distances and link actors through practice relations. I argue that such a practice-oriented, relational approach to labour control regimes in GPNs bears two central benefits for generating a better understanding of the conditions that constrain and enable the agency of workers and unions in garment-producing countries: First, a relational approach conceptualises labour control regimes by means of networked practices and relationships and therefore allows for a closer examination of the complex interdependencies and interrelations between labour control dynamics at various scales. Thereby, a relational approach is well equipped to shed light on the challenges for union agency resulting from the complex networks of practices and relations that constitute the labour control regime. Second, the relational approach developed in this book highlights that the labour control regime is primarily constituted through practices (as opposed to abstract capitalist forces or mechanisms). As a result, it is able to shed light on the ‘small transformations’ of labour control structures achieved by workers and unions.

Before proceeding to develop my relational approach to labour control regimes in Chapter 3, in the next section, I first turn to review a second body of literature concerned with the agency of workers and unions in GPNs. While research on labour control in GPNs has provided important insights into the constraining conditions and the broader terrain for the agency of labour in garment-producing countries, it has told us little about the conditions that enable workers and unions to build the capacities and power to actively challenge labour control structures. Therefore, in the next section, I introduce a second research thread that explores the strategies and enabling conditions for the agency of workers and unions in GPNs.

4 Research on Labour Agency in GVCs/GPNs

Labour agency first became popular as a research subject among Marxist geographers at the end of the 1990s with the emergence of labour geography as a geographical sub-discipline (Lier 2007; Coe and Jordhus-Lier 2011: 213). Labour geography emerged as a critique of the until then predominant conceptualisations of labour in neoclassical and Marxist economic geography, which tended to conceptualise labour either as a mere production factor or as a passive victim of the strategies of capital (1997: 1ff.). In response, early labour geographers such as Andrew Herod (1997, 2001a) and Jane Wills (2005) sought to shift the attention to “working class people as sentient social beings who both intentionally and unintentionally produce economic geographies through their actions” (Herod 1997: 3).

The research objective of labour geography in its initial phase was to highlight how workers and their organisations shape the geographies of capitalism. Early studies in labour geography tended to conceptualise ‘labour agency’ in narrow terms as the organised, collective agency exercised by trade unions in manufacturing sectors in developed countries (Coe and Jordhus-Lier 2011: 213). Consequently, early studies in labour geography were predominantly characterised by narrative accounts of “isolated success stories of workers with strong capacities to act and enhance their position vis-à-vis capital” (ibid.). As a result, the theoretical concept of agency initially remained underdeveloped. As Taylor et al. (2015: 10) put it: labour agency “had simply come to mean any meaningful manifestation of collective worker activity” (ibid., emphasis in original).

Responding to these shortcomings, labour geographers have, over the last decade, sought to ‘re-embed’ worker agency within the broader structures and social relations that shape the terrain for the agency of unions and workers, particularly in GPNs (Coe and Jordhus-Lier 2010, 2011; Cumbers et al. 2008; Riisgaard and Hammer 2011). These studies have stressed workers’ and unions’ embeddedness into a set of intersecting ‘structural forces’ at the vertical and horizontal dimensions of the GPN that enable and constrain collective labour strategies (Arnold 2013). At the vertical dimension, these ‘structural forces’ comprise inter-firm power relations and dynamics of value chain governance, including the CSR and sourcing practices of lead firms (Lund-Thomsen and Coe 2013: 277). At the horizontal dimension, these structural forces entail “the formations of capital, the state, the community and the labour market in which workers are incontrovertibly yet variably embedded” (Coe and Jordhus-Lier 2011: 214). The terrain for the agency of workers and unions at a specific node of the GPN is hence shaped by value chain governance dynamics on the one hand and by territorially embedded social relations (including those constituting local labour control regimes) on the other hand. Figure 2.6 illustrates the intersecting structural forces and relations at the vertical and the horizontal level that shape the terrain for worker and union agency in GPNs.

Fig. 2.6

Intersecting structural forces and relations at the vertical and the horizontal level that shape the terrain for worker and union agency in GPNs

In this context, Coe and Jordhus-Lier (2011) have coined the notion of ‘constrained agency’ to highlight the paradigmatic shift in labour geography from affirmative research telling ‘isolated success stories’ of worker struggles towards a more relational and holistic study of the complex socioeconomic practices and relationships shaping the terrain for labour agency in different places and industries.

As mentioned earlier, the central objective of this book is to develop a more nuanced understanding of the conditions that enable and constrain union agency in garment-producing countries, following the assumption that only by shifting the local capital-labour balance sustained improvements for workers can be achieved. With this in mind, the remainder of this section gives an overview of the various concepts and analytical approaches adopted so far by studies at the intersection of labour geography and GVC/GPN analysis to study the collective agency strategies of workers and unions in GPNs. It is important to note that, in this book, collective worker agency or union agency are by no means limited to the agency of formal, multi-level trade union organisations that prevail in the Global North. The focus of this study is on worker organisations in garment-producing countries—and specifically in India. As a result, this book adopts a broader notion of collective worker agency as has been forwarded in debates on ‘social movement unionism’. These debates have highlighted the variety of collective worker organisations in the Global South, which also encompass politically independent grassroots unions or worker-led community organisations (see e.g. Fairbrother 2008; Fairbrother and Webster 2008; Moody 1997; Nowak 2017). Within these organisations, boundaries between workplace organising and community organising and between class and other social categories are increasingly blurred (Castree et al. 2004: 225). The notion of labour agency adopted in this study hence includes various forms of collective worker organising and is sensitive to the intersection of class with other lines of social differentiation, such as gender or race (c.f. Valentine 2007).

In the remainder of this section, I review existing studies on collective worker and union agency in GPNs. I classify existing studies on labour agency in GPNs into three main research strands working with different conceptual and analytical approaches: (1) a research strand drawing centrally on the concept of worker or union power resources; (2) a research strand adopting a scalar approach to labour agency; and (3) a research strand concerned with the constitution and structural effects of ‘Networks of Labour Activism’ (Fig. 2.7).

Fig. 2.7

Source Author

Research strands on labour agency in GPNs.

In the following, I will introduce the major conceptual and empirical contributions made by each strand of work and point out research gaps, before explaining how this book contributes to closing these gaps.

4.1 Approaching Union Agency Through the Lens of Worker and Union Power Resources

The ‘power resources approach’ has been widely used by labour geographers as a conceptual tool to assess the conditions that constrain and enable labour and union agency in GPNs (Schmalz et al. 2018; Webster et al. 2008; Webster 2015). The power resources approach is based on the assumption that labour can successfully advance its own interests vis-à-vis capital by collectively and strategically mobilising different types of power resources (Schmalz et al. 2018: 114). ‘Power’ is generally defined in the context of the power resource approach as “the ability of actor A to make actor B do something that B would not otherwise do” (Zajak 2017: 1012; Knight 1992). The groundwork for this conceptual approach was laid by Wright’s (2000) distinction between structural power deriving from workers’ position within the economic system and associational power deriving from workers’ capacities to form collective organisations. This primary distinction has subsequently been refined and expanded by other labour scholars. Silver (2003), for example, has proposed to distinguish between marketplace bargaining power and workplace bargaining power as two subtypes of structural power. Brookes (2013) and Schmalz and Dörre (2014) have introduced the notion of institutional power to refer to worker and union power resources that are derived from formal and informal rules, regulations and mechanisms, such as labour laws or collective bargaining mechanisms.

Debates on strategies for trade union renewal in the Global North (Dörre et al. 2009) as well as debates on ‘social movement unionism’ in newly industrialising countries of the Global South (Moody 1997) have further added to the concept of worker and union power resources. In this context, scholars have pointed at several ‘new’ trade union power resources (Webster 2015), which might be able to compensate for workers’ decreasing structural and associational power under globalised capitalism. For example, authors like Brookes (2013) or Schmalz et al. (2018) have pointed at coalitional power as a broader form of associational power that workers and unions may activate through building relationships of external solidarity with community organisations, social movements or consumer organisations. Lastly, several authors have pointed to new resources of moral power, which unions may activate by creating awareness for labour rights violations or struggles in public and media discourses (Webster et al. 2008; Schmalz et al. 2018). Webster (2015) has proposed to subsume coalitional and moral power under the category of ‘societal power’, since both derive from workers’ or unions’ collaboration with other societal groups. Figure 2.8 gives an overview of the different types of power resources that workers and unions may potentially activate in GPNs.

Fig. 2.8

Source Author

Worker and union power resources.

In the remainder of this section, I will introduce each power resource in more detail and summarise findings from existing studies regarding the potential of different power resources to help local unions build sustained bargaining power vis-à-vis employers.

4.1.1 Structural Power: Marketplace and Workplace Bargaining Power

Structural power derives from the specific position of workers within the economic system (Wright 2000: 962). When analysing the structural power of workers in GPNs, we can distinguish between marketplace and workplace bargaining power as two sub-forms of structural power (c.f. Silver 2003). Marketplace bargaining power accrues to workers positioned in labour markets characterised by low unemployment, who possess scarce skills, or who have access to savings, social security or other income enabling them to withdraw from the labour market (Silver 2003: 3). The exercise of marketplace bargaining power is thus not necessarily tied to collective forms of agency, but instead it could also be exercised by individual workers. In contrast, workplace bargaining power accrues to workers “who are enmeshed in tightly integrated production processes, where a localised work stoppage in a key node can cause disruptions on a much wider scale than the stoppage itself” (Silver 2003: 13). Whether or not workers and unions at a specific node of a GPN are able to exercise structural power hence depends on their position within intersecting value chain dynamics at the vertical dimension and labour market dynamics at the horizontal dimension of the GPN.

Past studies have stressed, first and foremost, the constraints for workers in garment-producing countries to leverage structural power resources. On the one hand, workers’ workplace bargaining power tends to be severely curbed by the geographical fragmentation of production networks and retailers’ ability to relocate their sourcing activities in reaction to wage increases or strikes (López 2021). On the other hand, workers in the export-garment industries of the Global South generally possess low marketplace bargaining power due to the little skill and knowledge intensity of the larger share of tasks in garment production (c.f. López et al. 2021). However, Kumar (2019a, 2019b) argues that, in more recent years, the emergence of ‘oligopolistic suppliers’ in garment GPNs has endowed garment workers with new structural power resources. These new structural power resources result from lead firms’ increasing dependence on a smaller number of large ‘oligopolistic’ tier one full-package suppliers. These full-package suppliers increasingly take on strategic logistics and inventory management functions for lead firms (see also Azmeh and Nadvi 2014; Merk 2014). Consequently, Kumar (2019a: 355) argues that for workers and unions at these tier one suppliers’ “structural power (…) increases alongside the degree of market spatial inflexibility”. Since these large suppliers can capture greater value shares, unions at these suppliers are more likely to succeed in pressuring employers into collective bargaining through work stoppages and other forms of collective action.

The emphasis on disruptions to the production process through work stoppages inherent to the notion of ‘structural power’, makes it clear that to mobilise structural power resources, workers need the strategic capacities to plan and perform collective action. Therefore, ‘structural power’ cannot be thought independent from the second basic form of workers’ power, which is associational power.

4.1.2 Associational and Organisational Power

Associational power accrues to workers from their capacities to form collective organisations and to act collectively (Wright 2000: 962). The mobilisation of associational power resources thus requires organising and training processes directed to build workers’ capacities to develop and execute collective action strategies (Silver 2003: 13ff.). Since associational power is tied to workers’ abilities to act collectively, trade unions and other forms of collective worker organisations are a central driving force for the exercise of associational power. However, it is important to note that the formal existence of a trade union is not enough for workers to exercise associational power. Unions as organisations need to foster specific types of relationships and capacities for agency among their members (c.f. Hauf 2017: 1003; Lévesque and Murray 2002). In particular, unions need to cultivate and synthesise the ‘social capital’ of their members “so that they identify themselves as part of a collectivity and support its purpose and its policies” (Gumbrell-McCormick and Hyman 2013: 30).

This qualitative understanding of associational power has led Gumbrell-McCormick and Hyman (2013) to introduce the notion of organisational power as a sub-form of associational power. Organisational power refers to the capacities of a union to build the social capital and strategic capacities of its members and functionaries as well as to foster a ‘cohesive collective identity’ among their members. Hence, to build organisational power, unions need to foster a solidary mindset among union members that goes beyond individual motives of obtaining personal benefits or protection through the membership. This solidary mindset is achieved through stimulating lively communication and relationships among members and through actively involving members in union life (Lévesque and Murray 2010: 336f.). Moreover, unions need to foster internal democratic structures and practices to develop their members’ strategic capacities. To this end, Gumbrell-McCormick and Hyman (2013: 30) stress that unions need to nurture “a culture favouring discussion between rank and file and officials and [through] educational work to ensure that policies are well understood and reflect the conditions experienced on the ground”.

Regarding the ability and opportunities of workers and unions in garment-producing countries for building associational and organisational power, past research has highlighted the various constraints resulting from repressive employer and state regimes (Anner 2015a, b; López 2021; Ruwanpura 2015). Given the price pressure by lead firms in global garment value chains, employers in export-garment sectors of the Global South frequently rely on disciplining practices to suppress collective worker organising. These include threatening, dismissing or co-opting union activists. Likewise, public rallies or demonstrations for higher wages are not seldom shattered by the police (e.g. Anner 2015a, b; Padmanabhan 2012). In addition, various studies have highlighted constraints for building workers’ associational power—particularly in Asian countries—resulting from the internally fragmented nature of labour movements (see e.g. Arnold 2013; Hauf 2017). Established trade union federations in Asian countries are often affiliated to political parties and dominated by rent-seeking union leaders (ibid.). As a result, over the last decades, in many Asian countries, an increasing number of ‘independent’ trade unions and worker organisations have emerged (Kumar 2014; Jenkins 2015). This proliferation of worker organisations often hampers the ability of workers in garment-producing countries to exercise associational power through coordinated joint action in the political sphere, for example in the context of national minimum wage negotiations (Arnold 2013).

4.1.3 Institutional Power

Next to associational and structural power, institutional power represents the third ‘traditional’ power resource that workers in GPNs may be able to draw on. Institutional power accrues to workers from their ability to invoke formal and informal rules and mechanisms that structure capital-labour relationships. Sources of institutional worker power are represented, for example, in national labour legislation and in institutionalised dispute settling, wage setting or collective bargaining mechanisms (Brookes 2013: 188). However, in the context of global capitalism, ‘traditional’ sources of workers’ institutional power are dwindling due to neoliberal state policies of labour market flexibilisation (see e.g. Cumbers et al. 2016; Fairbrother and Webster 2008; Webster et al. 2008). This is particularly true for many garment-producing countries, where states have pushed global market integration through regulatory reforms that have severely curtailed institutionalised labour representation. Governments of garment-producing countries have, for example, frequently set up so-called Special Economic Zones for export industries, which are exempt from national minimum wage regulations or from the right to collective bargaining and freedom of association (FoA) (Ruwanpura 2015).

Against this backdrop, studies concerned with (potential) power resources for unions in the garment GPN have pointed at transnational private or public-private social regulation mechanisms, such as codes of conduct (CoC), multi-stakeholder initiatives (MSI) or Global Framework Agreements as potential ‘new’ institutional power sources for workers and unions in garment-producing countries (see e.g. Anner 2012; Lund-Thomsen and Coe 2013; Zajak 2017). In response to sustained pressure from anti-sweatshop movements, almost all international garment brands and retailers have introduced corporate CoC that set supplier labour standards. In addition, many retailers and brands are members in MSI, which include NGOs as independent monitoring organisations to increase firms’ accountability (Fütterer and López Ayala 2018: 15ff.). It is important to note, however, that—as opposed to national labour laws and institutionalised bargaining mechanisms—(semi-)private transnational regulatory mechanisms are not the outcome of prior labour struggles, in which workers successfully shifted the capital-labour power balance through exercising associational and structural power. (Semi-)private transnational regulatory mechanisms, such as CoC or MSI, are voluntary initiatives by retailers, which are usually introduced as a response to consumer criticisms from the Global North (Hauf 2017: 1003). As a result, rather than institutionalising labour power, CoC and MSI seek to institutionalise ‘business models of ethics’ (Scheper 2017: 1086) based on discourses of corporate responsibility and conflict management.

In view of this, several empirical studies pointed at the limits of CoC, and related auditing and complaint mechanisms to be used as sources of institutional power by workers and unions in garment-producing countries. For example, Anner (2012) and Egels-Zandén and Merk (2014) studied two prominent transnational MSIs in the garment GPN: the Fair Labour Association and the Fair Wear Foundation. They found that local unions’ complaints about garment manufacturers’ union-busting practices in most cases had no effect. Lund-Thomsen and Coe (2013) in turn highlight that CoC may represent an effective institutional power resource for unions in garment-producing countries when used as an organising tool or as a reference in public campaigns. They illustrate how local unions in Pakistan achieved improvements in working conditions by using Nike’s CoC as a reference in worker organising campaigns and in public media campaigns conducted in collaboration with several national NGOs. Hence, as opposed to ‘traditional’ institutional power sources such as labour laws, which workers and unions can deploy autonomously, to deploy private regulatory mechanisms as a power source, workers and unions depend in part on other societal actors to advocate for their rights.

In the light of the limitations for workers to leverage traditional institutional power resources, several scholar have argued that workers’ abilities to build coalitions with other societal actors and to influence public discourses are central for leveraging new ‘societal power resources’ (see e.g. Webster et al. 2008; Webster 2015).

4.1.4 Societal Power: Coalitional and Moral Power

Webster (2015: 1) employs the term ‘societal power’ to refer to workers’ and unions’ capacities to build coalitions with social movements and to influence public and media discourses. Societal power can be further divided into coalitional power and moral power as two sub-forms. Coalitional power is defined by Schmalz et al. (2018: 122) as the ability of unions to build relationships of external solidarity with other social actors. These social actors may be mobilised for public campaigns or they may provide other types of support (e.g. financial support, capacity-building). Moral power, in turn, refers to workers’ and unions’ ability to discursively frame labour issues and solutions “in line with prevailing views of morality” and therefore able to attract public support (Schmalz et al. 2018: 123).³ This framing may be achieved, for example, by invoking notions “of the struggle of ‘right’ against ‘wrong’, providing a basis for an appeal to both, the public and politicians, as well as to allies in civil society” (Webster et al. 2008: 12).

Research on labour agency in garment-producing countries has debated in particular the possibilities for unions in garment-producing countries to deploy coalitional and moral power resources to effectively compensate for the lack of structural and institutional power resources (Zajak et al. 2017: 907). This debate has been motivated by the widespread argument in labour studies and social movement literature that workers and unions ‘at the bottom’ of GPNs can shift the local power balance by harnessing the so-called boomerang effect (Wells 2009; Merk 2009). The term ‘boomerang effect’ refers to generating “Northern pressure to support workers’ rights in the South”, usually through transnational consumer campaigns (Wells 2009: 571). However, empirical studies of union agency in the garment GPN have found that moral power through consumer campaigns remains limited when unions lack a strong local associational and structural power base (Anner 2015b; Kumar 2014, 2019b; Zajak 2017). Anner (2015b), for example, found that the transnational campaigns carried out by local unions at college apparel suppliers in Honduras and El Salvador in collaboration with student anti-sweatshop movements in the US were only effective in cases of particularly crude and violent labour rights violations. In these cases, “the very extreme nature of the threat of bodily harm” allowed anti-sweatshop activists to frame labour rights violations “in a way that resonates with larger values of human decency” and hence to mobilise broad public support (Anner 2015b: 165). In his research on union agency in the Indian garment industry, Kumar (2014, 2019a) similarly finds that unions without a strong associational power base have only occasionally achieved some ‘isolated victories’ by activating moral power resources through transnational consumer campaigns. According to Kumar, unions were more successful in achieving lasting improvements for workers when they used moral power from transnational consumer campaigns to strategically support workplace organising and collective bargaining processes (Kumar 2019a: 360ff.).

In summary, existing empirical studies on how unions in garment-producing countries use ‘new’ societal power resources have shown that “only if strength is established domestically does a positive reinforcement effect across different power sources become possible” (Zajak 2017: 1008f.). However, it remains a central issue for empirical studies to develop a more nuanced understanding of how specifically unions can use moral and coalitional power resources to support the building of associational power in the workplace. I argue that to generate such an understanding, we still need more insights into the interplay between different power resources as well as into the structural effects that transnational consumer campaigns have on local garment unions’ organising practices and internal relations. First insights by Hauf (2017) and Zajak (2017), for example, illustrate how close cooperation with transnational consumer campaigning networks can also constrain unions’ capacities for building associational and organisational power: Relying on the ‘borrowed’ moral power of consumer campaigning networks can, for example, decrease unions’ investments into members’ strategic capacities and into fostering internal union democracy (Zajak 2017: 1019f.; Hauf 2017: 997).

Motivated by these observations, this study aims to better understand how different forms of mobilising coalitional and moral power resources may enhance or constrain associational power building. I argue that we can benefit from adopting a relational perspective on union agency to gain such an understanding. Taking on such a perspective, we can conceptualise the agency strategies of local unions in garment-producing countries as emerging at the intersection of various relational spaces of labour agency that workers and unionists construct themselves. These spaces for labour agency comprise the union itself as a relational space as well as networks with other unions and NGOs at various levels. In this context, it is essential to pay attention to the spatial and scalar dynamics shaping unions´ interactions with other actors (Coe 2015). In the next section, I therefore introduce a second branch of work that approaches agency strategies of unions in GPNs through the lens of scale as a central spatial analytical category.

Before doing so, to conclude this section, Table 2.3 summarises the definitions of the various power resources and their sub-forms (marked with →) that have been introduced in this section.

Table 2.3 Union and worker power resources and their definitions

4.2 ‘Scaling’ Worker and Union Agency in GPNs: ‘Scale-Jumping’ and ‘Up-Scaling’

While the concept of worker and union power resources has been progressed primarily by labour sociologists and industrial relations scholars, labour geographers have highlighted workers’ and unions’ scalar strategies. According to Coe and Jordhus-Lier (2011: 219), “scale is particularly useful [as an analytical concept] as that it captures the double nature of the spatiality of labour agency: not only do the actions of labour play out in complex social geographies, but they can be understood as spatial phenomena in themselves. In other words, both the conditions and the strategies of labour agency are spatial”. However, scale as a theoretical concept has been scarcely theorised in research on labour agency in GPNs. Generally, Coe and Jordhus-Lier (2011) argue that we can distinguish between three analytical moments in which the concept of scale has been employed in research on labour agency in GPNs: to define the level at which bargaining takes place, to refer to the group of workers for whom decisions are made, and to demarcate the territories across which solidarity is being sought.

The engagement of research on labour agency in GPNs with the concepts of space and scale hence reflects the approach to these concepts adopted in labour geography more generally. As mentioned before, research on labour agency in GPNs has developed at the intersection of labour geography and GPN analysis. It has been driven to a large extent by labour geographers’ concern for understanding the spatial conditions and strategies of labour agency in the context of globalised production. Originally, labour geographers’ concern with scale has been informed by the empirical observation that capital—while seeking to ‘up-scale’ its own operations—is simultaneously interested in ‘downscaling’ labour relations: capital aims to contain the bargaining of wages and working conditions at the local or workplace level as a means to play workers in different locations off against each other (Merk 2009: 603). Hence, labour geographers have frequently pointed out that ‘jumping scale’ and ‘bridging space’ are important strategies for labour agency under global capitalism since they allow workers to match capital’s organisational scales (Castree et al. 2004; Herod 2001a; Merk 2009).

Labour geographers’ concern for workers’ scalar and spatial strategies has been informed by the political assertion that “workers actively produce economic spaces and scales in particular ways” and hence “shape the location of economic activity and the economic geography of capitalism” (Herod 1997: 24f.). In this vein, labour geographers have explicitly distinguished their theoretical perspective from the one forwarded by regulatory international political economy, which has traditionally been dominant in GPN analysis. Studies from the field of regulatory political economy conceptualise the “spatial scales at which social life is organised” as constructed and shaped predominantly by capital and state actors (Herod 2007: 29). Herod (2007) argues that such a perspective leads to a view of local, national and international scales for labour action as “areal containers of discrete absolute spaces” that are premade by capital and state actors and which contain workers’ activities. According to Herod (2007), this perspective “denies [labour] actors the social agency to construct the geography of global capitalism in different and varied ways, for it suggests that the global scale of capital organisation is something that simply exists, waiting to be discovered and used, rather than something that had to be made and is constantly remade through the actions of diverse social actors”.

Against this background, it has been a central concern of research at the intersection of GPN analysis and labour geography to enhance our understanding of the different scalar strategies that workers and unions employ to “reconfigure political landscapes and [to] renegotiate social hierarchies in ways which are more beneficial to the interests of workers” (Coe and Jordhus-Lier 2011: 219). In this light, various studies from labour geography have analysed the ‘multi-scalar strategies’ deployed by unions in globalised industries or service sectors (see e.g. Alford et al. 2017; Anderson 2009; Tufts 2007; Wills 2002). These studies have enhanced our understanding of workers’ scalar agency in GPNs in two important ways. First, these studies have highlighted that while transnational alliances can help to increase workers’ leverage vis-à-vis globally organised capital, there is no substitute for local organising (Herod 2001a; Wills 2002). Second, these studies have highlighted the diverse scales of labour organising and alliance building within GPNs, which range between workplaces, communities, cities, regions or the globe (Alford et al. 2017).

Which scales of organising and alliance building are most apt for workers at a specific node of a GPN depends on the spatial configurations of capital organisation and on workers’ and unions’ positionality within the broader production system. Particularly for workers in the industrial export sectors of the Global South, where transnational lead firms shape working conditions, literature has stressed the importance for workers to engage in transnational organising and campaigning to exercise leverage on geographically distant lead firms (Anner 2015b; Hale and Wills 2005; Kumar 2014; Merk 2009). With this in mind, Merk (2009), for example, illustrates how local unions in garment-producing countries can use the ‘urgent appeal system’ provided by the Clean Clothes Campaign (CCC)—a campaigning network led by European consumer organisations—as tool for ‘up-scaling’ workplace struggles. The urgent appeal system allows workers or unions in garment-producing countries to submit a complaint about labour rights violations in garment factories to the CCC together with a request for action. The CCC member organisations then engage in various actions to harness brands’ leverage over local garment manufacturers. Actions carried out by the CCC range from sending letters or emails to brands’ management teams, to planning and carrying out full public campaigns. This includes publishing fact-finding reports and the naming and shaming of brands embroiled with the factory where the labour rights violations occurred (Merk 2009: 607). According to Merk, the CCC’s urgent appeal mechanism “provides a grassroots based system to build labour solidarity across space, which may help [workers in garment-producing countries] to regain leverage over capital” (Merk 2009: 599).

However, studies focussing on the ‘up-scaling’ of workplace conflicts through transnational consumer campaigning have, at the same time, received criticism from various scholars. These scholars argue that the narrow focus on ‘up-scaling’ through transnational campaigning has shifted the analytical focus from the agency of workers to the agency of consumer groups or NGOs in the Global North. As a result, the latter are perceived and presented as the main agents of change (see e.g. Hauf 2017: 989; Wells 2009). Contradicting this notion, Wells (2009: 568) stresses that the mobilisation of extra-local leverage through transnational campaigns merely represents one of many strategic actions that workers and unions in the Global South usually employ when leading a labour struggle. To achieve sustained improvements in working conditions, he argues, workers’ and unions’ strategic actions at the workplace or at the community level are equally important (Wells 2009: 577). Similarly, Wills (2002) and Tufts (2007) emphasise that ‘up-scaling’ specific labour struggles should not be seen as an isolated strategy but as part of a broader ensemble of networks and institutional arrangements with which unions engage. They, therefore, stress that analyses of ‘multi-scalar’ agency strategies should not be limited to identifying the most appropriate scale for action. Instead, they argue, studies of ‘multi-scalar’ labour agency strategies in GPNs should explore the different types of resources that unions can leverage through actions at multiple scales (Tufts 2007: 2387).

In summary, scalar approaches to labour agency in GPNs have contributed to enhancing our understanding of the diverse spatial strategies at various scales that workers and unions in GPNs may employ. However, we still need a better understanding of how workers and unions can strategically link actions at various levels in the context of specific struggles. I argue that we need a better understanding in particular of the following three aspects of workers’ and unions’ multi-scalar strategies. First, we need a better grasp of the different power resources that unions can activate through building relationships at different scales—from intra-union relationships over building alliances with local community organisations to engaging with transnational consumer campaigning networks (c.f. Nicholls 2009; Tufts 2007). Second, I contend that we need a better understanding of the different resources and capacities that unions may access and develop in different relationships and networks at the transnational scale. Studies of the ‘multi-scalar’ agency of workers and unions in GPNs have tended to conflate all transnational networks and relationships constructed by local unions under the notion of ‘up-scaling’, be it relationships with consumer campaigning networks or with international worker networks. However, this conflation has obscured the fact that different kinds of transnational networks may be constituted through very different practices, therefore enabling unions and workers to access very different types of resources (c.f. Lohmeyer et al. 2018; Fütterer and López Ayala 2018; López and Fütterer 2019). Third, we need a better understanding of how unions’ strategies and networks at different scales not only complement but also shape and influence each other. Understanding the interplay of unions’ actions at various scales is important particularly in the light of the findings of studies working with the power resources approach. These studies found that unions’ practices of engaging in transnational alliances can directly impact unions’ internal organisational practices and workplace organising strategies (see Sect.  2.4.2).

To tackle these gaps, in this book, I develop a practice-oriented, relational approach to union agency in GPNs that emphasises practices and relationships instead of a priori-defined scalar categories (see Sect. 3.3). To this end, I build on a third strand of work within research on labour agency in GPNs focussing on ‘Networks of Labour Activism’, which is introduced in the next section.

4.3 Network Approaches to Labour Agency in GPN: ‘Networks of Labour Activism’

The ‘Networks of Labour Activism’ (NOLA) approach was first introduced by a group of interdisciplinary scholars in 2017. The approach centres explicitly on the practices, actions and relationships of workers and labour organisations in the Global South positioned “on the bottom rungs of the globally networked economy” (Zajak et al. 2017: 916; see also contributions to Forum Debate 2017 of Development and Change Vol. 48, Nr. 5). The NOLA approach is motivated by the observation that labour-intensive export sectors in Asian countries “have also become the testing ground for new forms of networked worker agency and activism” (Zajak et al. 2017: 900). These new forms of networked worker agency are based on workers’ construction of solidarity networks and relationships with a broad range of actors in varying geographical distances. Actors with whom workers construct solidarity networks range from local community organisations over national union coalitions to transnational consumer activist networks (Zajak et al. 2017: 901).

Contrary to scalar approaches to labour agency in GPNs, which have focussed rather one-sidedly on workers’ transnational ‘up-scaling’-strategies, the NOLA approach conceptualises workers’ agency in GPNs as multi-directional and multi-layered. The agency of workers and unions is multi-directional since it is performed in and through relationships with a variety of actors, including potential targets as well as potential allies. In addition, workers’ agency is multi-layered since it is performed through relationships that stretch across various distances—i.e. relationships at multiple scales (Zajak et al. 2017: 904). Nevertheless, the NOLA approach still emphasises ‘cross-border strategising’ as a central feature of networked labour agency in GPNs (Zajak et al. 2017: 905). Cross-border strategising is essential for workers ‘at the bottom’ of GPNs due to the structural context of supply chain capitalism within which they are embedded. Supply chain capitalism refers to global value chains as a system of capital accumulation that is based on Northern lead firms’ practices of dis-embedding the labour-intensive production steps from regulated, unionised environments through outsourcing (Tsing 2009). Therefore, to tackle geographically distant lead firms, which ultimately determine labour conditions along the supply chain, workers need to engage in cross-border strategising. Cross-border strategising refers here to strategic practices of building solidary relationships with potential allies in different countries on the one hand, and to strategic practices of targeting actors or institutions located in a foreign country on the other hand (Zajak et al. 2017: 905).

It needs to be noted here that a central difference to ‘up-scaling’ literature—which has also been concerned with workers’ cross-border strategies—consists in the explicitly constructivist perspective of the NOLA approach. Whereas ‘up-scaling literature’ has put emphasis on how workers and unions use the leverage of transnational alliances to enhance their position vis-à-vis local capital (or the state), the NOLA approach shifts the attention to the practices and relationships through which workers and unions construct these alliances in the first place (Zajak et al. 2017: 903). It is within these alliances that workers and unions develop specific strategic capacities and power resources which may in turn contribute to strengthening workers’ domestic associational power base (Zajak et al. 2017: 1009). At the same time, the NOLA approach recognises that ‘networks of labour activism’ become “their own structural forces” (Zajak et al. 2017: 1020). In other words, the interactions and learning processes within these networks shape the behaviour of the actors involved in it.

Empirical studies using the NOLA approach have provided a range of important insights on the structural effects of various types of solidarity networks on the practices of unions in garment-producing countries. Hauf (2017), for example, has studied the effect of local unions’ engagement with the Play Fair Campaign in Indonesia, an MSI led by Northern NGOs and global trade union federations. Hauf found that participation of Indonesian unions in the transnational Play Fair campaign helped unions to forge new networks with other local garment unions, since they had an incentive to develop a common position. However, he also finds that the potential of these new solidary networks at the local and national level for building sustained associational power was limited. Practices of collaboration between unions were strictly confined to interactions within the framework of the Play Fair campaign. Consequently, these interactions did little to bridge the political and ideological differences between local unions in the Indonesian garment sector and thus to establish the foundation for collaborations on other issues (Hauf 2017: 988).

Hauf’s findings are further supported by Zajak (2017) in her study of Bangladeshi garment unions’ engagement with the Bangladesh Accord for Fire and Building Safety.⁴ She finds that local unions’ interactions with the transnational stakeholder network of the Bangladesh Accord have, on the one hand, strengthened local unions’ organisational power. For example, trainings conducted by these stakeholders have enabled local unionists to develop important strategic capacities for reframing workplace issues so that they fall under the mandate of the Accord (Zajak 2017: 1017). But, on the other hand, she finds that local unions’ increased engagement with international organisations and Northern NGOs has also led to new divisions and relations of competition between local unions and hence to a weakening of associational power at the national level (Zajak 2017: 1019).

Whereas Hauf (2017) and Zajak (2017) have focussed on the practices through which local garment unions engage with transnational multi-stakeholder networks, Lohmeyer et al. (2018) have analysed how Asian garment unions engage with transnational labour networks. In their study of the TIE ExChains network, which links Asian garment workers with German fashion retail workers, they illustrate how solidary relationships within the network are constructed through practices of facilitating shared experiences and of developing a common interpretation of labour rights violations along the chain. At the same time, Lohmeyer et al. (2018) illustrate potential tensions between practices of performing cross-border solidarity campaigns within the network and practices or local workplace organising. These tensions arise because cross-border solidarity campaigns are usually centred on specific issues and therefore have a rather short-term outlook. Workplace-organising practices are, in turn, usually embedded within a unions’ long-term strategy. Against this background, Lohmeyer et al. (2018) stress that a careful strategic alignment of transnational campaigning practices with unions’ long-term strategic goals and organising strategies is necessary to ensure that transnational campaigning strengthens unions’ local associational power base instead of undermining it (Lohmeyer et al. 2018: 417).

In summary, network-centred approaches to worker and union agency have enhanced our understanding of union agency in GPNs in three ways. First, networked approaches to labour agency have contributed to breaking up the ‘black box’ of unions as collective actors. This has been achieved by making visible how “union strategies evolve through contested socio-spatial relations both within unions themselves and with other social actors” (Cumbers et al. 2008: 369). Second, through engaging with the power resources approach, studies of workers’ networked agency have opened up analytical space for assessing the potential of different ‘networks of labour activism’ for strengthening local unions’ associational power base. Third, network-centred studies of labour agency in GPNs have enhanced our understanding of how workers and unions forge their own relational geographies and add new layers to existing GPNs.

Notwithstanding these critical insights into the agency of local workers and unions in garment-producing countries, a central dimension of labour agency in GPNs has hitherto remained underexplored in NOLA studies: the spatial dimension of labour’s agency. The neglect of space as an analytical category is reflected, first, in Zajak et al.’s (2017) proposal of ‘cross-border strategising’ a central characteristic of networks of labour action. ‘Cross-border’ refers, on the one hand, to networks that connect actors across borders. On the other hand, it also refers to a situation in which workers build alliances with other social groups at the local or national level to target actors located in a different country. However, this subsumption of local, national and transnational networks under the same category of ‘cross-border action’ ignores that very different practices and mechanisms are involved in constructing networks at different scales. Second, subsuming networks that span across different territorial extensions under the same category of ‘cross-border strategising’ neglects the fact that networks forged within particular places and networks forged across great distance may play distinct but complementary roles in building local union power (c.f. Nicholls 2009). Last but not least, the neglect of the spatial dimension of labour agency is reflected in the little attention paid by studies working with the NOLA approach to the place-specific structures of labour control that constrain labour’s spaces and options for agency.

Therefore, in this book, I build on the network perspective of the NOLA approach and integrate relational perspectives from economic geography on the one hand, and with literature on labour control regimes on the other. Drawing on these different bodies of literature, I develop a relational approach to labour control and labour agency in GPNs. Before doing so, in the next section I first summarise the contributions and limitations of existing research on labour agency in GPNs.

4.4 Interim Conclusion: Contributions and Shortcomings of Existing Analytical Approaches to Labour Agency in GPNs

This section has introduced three different research strands on labour agency in GPNs: (1) a research strand drawing on the concept of worker and unions power resources; (2) a research strand adopting a scalar approach with a focus on workers’ ‘up-scaling’ strategies; and (3) a research strand that analyses labour agency in GPNs through the lens of ‘Networks of Labour Activism’. Table 2.4 on the next page provides a synthesised overview of the central research questions, main insights and shortcomings of each research strand.

Table 2.4 Overview of different research strands on labour agency in GVCs/GPNs

In a nutshell, I argue that we can learn three critical lessons from existing research on labour agency in GPNs regarding the conditions that constrain and enable the agency of workers and unions in garment-producing countries:

First, by re-embedding the agency of workers and unions into structural sets of social and economic relations—encompassing value chain dynamics at the vertical dimension and state, labour market and cultural relations at the horizontal dimension—existing research on labour agency in GPNs has highlighted potential structural constraints for building local union power ‘at the bottom of GPNs’. The analysis of different worker and union power resources in combination with a focus on ‘constrained’ labour agency has contributed to a more fine-grained understanding of the conditions that curb local unions’ opportunities to build sustained bargaining power vis-à-vis employers. In this regard, past studies have highlighted how value chain governance dynamics at the vertical dimension of the GPN curtail the structural power of workers and unions in garment-producing countries. Moreover, repressive state and employer labour control regimes at the horizontal level significantly constrain workers’ opportunities for collective organising and building associational power resources, as well as workers’ opportunities for leveraging institutional power sources, such as legal frameworks.

Second, existing research on labour agency in GPNs has also provided insights into potential enabling conditions for building local union power in garment-producing countries. Since geographically distant lead firms represent the locus of power in these GPNs, existing research has stressed opportunities for local unions in garment-producing countries to ‘up-scale’ local labour struggles to the transnational level. ‘Up-scaling’ labour struggles can enable unions to leverage coalitional and moral power resources through consumer campaigns in the Global North and thereby to exert pressure over lead firms and manufacturers. Nevertheless, existing research has also pointed out that activating coalitional and moral power resources through ‘up-scaling’ strategies cannot substitute associational power resources on the ground. In this line, past studies of labour agency in GPNs have highlighted that transnational consumer campaigns only contribute to building local union power, when local unions strategically deploy them to create spaces for collective organising and bargaining on the ground.

Third, in particular the NOLA approach has contributed to our understanding of the constraining and enabling conditions for building local union power in garment-producing countries. It has highlighted that not only capital, state and labour market relations but also the networks constructed by workers and unions themselves represent enabling and/or constraining contexts for labour agency. On the one hand, workers and unions can develop and build capacities and power resources through and within these networks. On the other hand, unions’ and workers’ embeddedness within ‘networks of labour activism’ can also pose constraints for unions’ capacities of building associational power. In this line, studies using the NOLA approach have highlighted how the participation of local unions in transnational NGO-led multi-stakeholder networks can hamper internal union democracy and create inter-union divisions within garment-producing countries.

Notwithstanding these important contributions of past research on labour agency in GPNs, three important aspects of union agency remain underdeveloped and require further research. First, for a more nuanced understanding of how local unions in garment-producing countries can build sustained bargaining power, we still need a better understanding of how unions can link actions at various scales to exercise leverage over employers. In this context, labour geographers’ argument that workers and unions can activate and develop different power resources through actions at different scales requires further exploration. Labour geographers have stressed that scales of labour action traditionally encompassed the workplace and the international level and also multiple scales in between, such as the community, the city or the national level (c.f. Anderson 2009; Tufts 2007; Wills 2002). In particular when studying unions in the Global South, where non-traditional forms of ‘social movement unionism’ have emerged, a stronger analytical focus on the implications of unions’ engagement with community actors for building bargaining power vis-à-vis employers is necessary.

Second, past research on labour agency in GPNs has tended to conflate the variety of practices through which local unions engage with transnational networks under the same notions of ‘up-scaling’ or leveraging ‘coalitional power’. However, this conflation has concealed essential differences in the everyday practices and power relations that constitute local unions’ relationships with different types of transnational networks. There are important differences, for example, between the practices through which consumer campaigning networks exercise solidarity compared to grassroots worker networks (see Lohmeyer et al. 2018). Consequently, if we take the lesson from NOLA studies seriously that also the networks constructed by unions themselves have structural effects on unions’ associational power resources, we need to be more sensitive to the different practices and relationships that constitute different ‘networks of labour activism’. Only an analytical perspective that is sensitive to such differences will allow us to evaluate to which extent the engagement in different transnational networks allows local unions to build strategic capacities and power resources.

Third and last, studies concerned with the agency of labour in GPNs have only superficially engaged with studies of labour control (regimes) in GPNs. Studies on labour agency in GPNs have generally highlighted that repressive labour control regimes in the export industries of the Global South constrain workers’ and unions’ opportunities for building and leveraging associational and institutional power resources. However, these studies have not yet provided a more fine-grained analysis of the specific capital and state practices that constrain the agency of workers and unions. Moreover, studies on labour agency in GPNs have so far neglected the role of workers and local unions ‘at the bottom’ of GPNs in co-shaping and potentially transforming labour regimes at specific nodes of a GPN. Therefore, to fully understand the dialectical relationship between labour’s agency and labour control regimes as structural contexts within GPNs, we need to look closely at how unions tackle and transform specific state and capital practices that constitute labour control regimes.

To tackle these shortcomings and to further develop our understanding of the constraining and enabling factors for building sustained union power in garment-producing countries, in the next section, I develop a relational approach for studying labour control regimes and labour agency in GPNs.