Global production networks,labour and small firms

Introduction

In this article we are concerned with the dynamics of the economic organisation and the restructuring of the geography of contemporary global capitalism, especially the development of global production networks (GPNs), the position of small firms in such networks, and the implications for worker organisation and resistance of such positioning. In particular, we suggest that some analyses appear to consign workers in small firms in these networks to a necessarily tenuous, dependent and powerless position. By way of contrast, we argue that understanding the way in which small firms are geographically connected to other firms and how they are spatially embedded allows us to see that they and the workers who labour in them can in fact have significant agency. Such a spatial awareness, then, challenges those analyses of present-day globalisation that fail to theorise the importance of capitalism’s geographical organisation to the way it operates as an economic system, and which see this organisation as simply the reflection on the Earth’s surface of the actions of collective capital: that is to say, as a social artefact upon which workers have little influence. (For additional arguments on the importance of taking issues of spatiality more seriously in analyses of work and employment, see Herod et al. 2007; Rainnie et al. 2007; McGrath-Champ et al. 2010; Rainnie, McGrath-Champ and Herod 2010; and Rainnie et al. 2011.)

The paper is divided into five main sections. The first very briefly considers some ongoing changes in the way transnational corporations (TNCs) are organised and some geographical patterns of recent restructuring, thereby providing a context for what follows. The second section discusses GPNs’ geographical embeddedness and introduces conceptual schema concerning workers’ positioning vis-à-vis GPNs. The third section outlines some approaches to GPNs that have endeavoured to address the role of labour, although they have tended to focus on large firms and to ignore small firms and the agency of their workers. The fourth section develops the argument concerning small firms as being integral to the amalgam that constitutes GPNs, rather than simply as being subservient components thereof, by linking small firms’ activities to broader arguments about uneven and combined development. The fifth section links these themes together, arguing that locationally situated analysis of the labour process that includes bona fide conceptualisation of the role of small business is crucial to a robust understanding of GPNs.

Globalisation, transnational corporations and economic restructuring

It is a truism to state that the global economy is becoming ever more interconnected and networked. What Harvey (1989) calls ‘time-space compression’ means that goods, capital, information and people can now move across the Earth’s surface more quickly than ever before in human history. In the process, geographically distant places are becoming ever more linked together. Despite the self-evident nature of this statement, there are some important aspects of the ways in which this interconnectivity is coming about that are worth detailing. In particular, as Dicken (2011: 20) points out, one aspect of the global economy’s mounting interconnectedness is that the growth of trade has been outpacing the growth of manufacturing output in recent years. In turn, the growth of foreign direct investment (FDI) has been outpacing the growth of trade, such that the primary mechanism of bringing about interconnectedness has now shifted from trade to FDI. The common element in these developments is the transnational corporation (TNC). Indeed, TNCs have come to play an increasingly dominant role as geographical integrators of the global economy relative to, say, the types of small firms that have been the focus of classical economic theorising since Adam Smith. For instance, TNCs presently account for two-thirds of world exports of goods and services, of which a significant share is intra-firm trade.

Changes in the way the global economy is organised economically have been associated with notable shifts in how it is organised geographically, although these shifts should not be overstated. For instance, a number of Global South economies have become major producers of manufactured goods that they are now exporting across the planet. Nevertheless, the planetary geographies of production, trade and FDI remain highly spatially uneven and strongly concentrated in particular parts of the global economy, especially in the industrial giants of the Global North (Dicken 2011: 25). Thus, whilst the established developed economies are being challenged by some newly industrialising countries like Brazil and India, 80 per cent of outward FDI stock nevertheless still emanates from just 15 countries, while almost half of all inward FDI is concentrated in five host countries (see Herod 2009 for more details). One key metamorphosis, however, is that 30 per cent of inward FDI is now concentrated in China and Hong Kong. Indeed, it is suggested that the biggest development in East Asia in recent decades has been the economic (re-) emergence of China (Dicken 2011: 31). This is important because, as we have pointed out previously (Rainnie et al. 2011), the rise of giant China-based supplier companies is fundamentally altering the dynamics of a number of important GPNs.

For Dawley (2011: 401), there are three principal competitive dynamics currently shaping the uneven geography of contemporary corporate investment and the restructuring processes that are impacting GPNs: time-based competition; intensified intra-corporate competition; and TNC mergers, acquisitions and restructuring. Dicken (2011: 59) argues that these pressures have been reinforced and accelerated in recent years by the impact of increasingly dominant financialisation. These developments are influencing the organisational structure of many TNCs in some very important ways. For example, a significant response to such competitive pressures has been the growing disintegration of the model of business organisation that dominated for most of the 20th century – that is, the highly vertically integrated corporation – and the emergence of much more networked organisational forms. Indeed, Castells (1996) has suggested that the era of globalisation has been marked by the rise of what he calls a ‘network society’. One consequence of this has been that activities that were previously conducted ‘inside’ the corporation – such as manufacturing the components that would be used in the assembly of automobiles – are increasingly being contracted out. A classic example of this in recent years was the 1999 spinning off by General Motors (GM), the world’s largest automobile producer, of its components manufacturing subsidiary Delphi in the hope that this would make Delphi more competitive, enable it to diversify its customer base, and ultimately lead it to produce parts more cheaply for GM. Equally, other activities which may remain ‘inside’ the corporation – like janitorial activities – have increasingly been benchmarked against outside contractors. These developments have raised myriad issues for workers and their attempts to organise (see Herod 2007).

One of the most noteworthy developments in all of this, though, has been the growing imbrication of large firms and their suppliers. Thus, even as many large firms have subcontracted out activities that they themselves previously did, subsequently they have frequently developed highly integrated relationships with the firms that now do this work. For example, after GM spun off Delphi, both companies purchased stakes in each other and continued to do business with one another, and GM remained Delphi’s largest customer. Furthermore, in the aftermath of the US government’s bailout of GM and Delphi’s problems with bankruptcy, GM provided Delphi with financial support, paid billions of dollars in reorganisation-related charges to Delphi, lowered its prices for Delphi-made parts, provided funding for labour and pension costs, and bought Delphi’s steering division in 2009 (US Department of Commerce 2009). This is but one of many such examples. Such concentration and consolidation, though, are having important implications for supply chains and the nature of work and employment within them (Flecker 2009). In particular, there are increasing demands to meet the conditions of entry to them, as conscious coordination of the chains becomes ever greater with the growing integration of the kinds of subcontracting structures typified by Japanese keiretsu, Korean chaebol, and Wal-Martian business arrangements. Indeed, this coordination has grown to such an extent that Dicken (2011: 121) has suggested that it now makes more sense to think of TNCs not as embodied organisations but as networks of networks (perhaps the classic example being shoe manufacturer Nike, which does not own any of its own factories but merely subcontracts with myriad independently owned plants across Asia). Conceiving of TNCs in such manner raises critical questions about the place of small firms in such networks, and of their agency and that of their employees.

Global production networks as geographically embedded entities

Elsewhere (McGrath-Champ et al. 2010; Rainnie et al. 2011), we have argued for a GPN-theory approach to understanding work and employment within the contemporary global economy. Such an approach, we contend, gives greater emphasis and acknowledgement to the issue of value creation and appropriation than do either ‘global commodity chain’ (GCC) or ‘global value chain’ (GVC) analyses (see also Taylor 2010). Adopting Ernst and Kim’s (2001: 1) definition of a GPN as something that ‘combine[s] concentrated dispersion of the value chain across firm and national boundaries, with a parallel process of integration of hierarchical layers of network participants’, we maintain that a GPN approach has a deeper spatial awareness concerning workers’ lives than do either GCC or GVC analyses. This is because GPN analysis takes more seriously the ways in which workers’ embeddedness in place shapes the possibilities for their action. Following Hudson (2002), and drawing on a mixture of territorial and relational views of places (that is, views that see places as real units circumscribed by political and economic boundaries; and views that see them in terms of their relationships to other places), we argued that places must be seen not simply as inert arenas or boxes in which working life is played out, but rather as actively and continuously remade locations where local and non-local systems of rules, norms, customs, legal structures and regulatory mechanisms intersect to shape and institutionalise the behaviour of workers and employers (Coe et al. 2004). From such a perspective, the trajectory of a place’s development is conceptualised as a dynamic outcome of the complex interaction between its territorialised internal social relations and how it is linked with other places through global production networks, all within the context of ever-changing regional governance structures. Within such a view, GPNs are seen to act as global pipelines between locally based firms or clusters of firms (‘the local buzz’) within regions and selected partners outside the region (Coe and Hess 2011). The world economy therefore consists of tangled webs of production circuits and networks, with TNCs playing a key role in coordinating those networks.

In such an approach, it is important to recognise that transnational corporations are central agents of urban and regional development, interweaving GPNs with urban and city networks through what Yeung (2009) calls their ‘strategic coupling’ with the local business firms and institutions in the communities in which they are located. This concept of strategic coupling is helpful when seeking to understand GPNs’ dynamics, because it invokes processes of corporate layering, conversion and/or recombination. Here, layering refers to the process whereby successive rounds of investment are made in particular regions; conversion involves a change in the nature of the relationship between the region, its assets and TNCs; and recombination means that a region’s assets are being linked to new waves of investment, which itself can involve decoupling in other regions (a TNC may abandon one community to relocate to another, thereby connecting this new community to its broader corporate network). Significantly, all three of these processes – layering, conversion, and recombination – are about remaking the spatial relationships between economic actors and particular places. Their interface, however, is decidedly complex and geographically specific, because whilst states and local economies are essentially territorial, GPNs are not to the same degree (Dicken 2011: 72). GPNs therefore slice through boundaries in highly differentiated ways, influenced in part by regulatory and non-regulatory barriers and in part by local socioeconomic conditions. This has a profound effect on the bargaining power of the actors involved. Consequently, local urban and regional development needs to be seen as a translocal, dynamic process of growth and change, with multiple actors operating at a variety of geographical scales shaping its trajectory. This means that the strategic coupling processes of these actors in different cities and regions constitute the central dynamic of TNCs’ activity as they bring together urban and regional assets and GPN dynamics in a recursive and cumulative process of growth and development (Yeung 2009: 216).

Although much analysis has focused on the actions of TNCs as the creators of commodity chains, MacKinnon (2012) suggests that regional institutions – which vary considerably across space in their constitution due to the particularities of local histories – are central in all of this because they distinctively shape how strategic coupling occurs through their moulding of regional assets to fit the needs of GPNs, a moulding which is geared towards the creation, enhancement and capture of value (Smith et al. 2002). However, despite the centrality of both TNCs and regional institutions, power asymmetries between them can often result in the latter’s ‘corporate capture’ over time, which affects the degrees of freedom they have with regard to their activities (i.e. it affects their levels of agency). Given all this, then, there is a need to identify frictions and tensions within the strategic coupling process, such as uneven value capture, labour exploitation and social and class conflict in different locations along the commodity chain so that we might better understand how worker agency shapes GPNs’ structure (MacKinnon, 2012). This is particularly so because frictions and tensions can, in extremis, result in corporate downgrading or abandonment – that is to say, ‘decoupling’ through the excision of particular nodes/ places from the chain.

Significantly, such concepts of coupling and decoupling bear a strong resemblance to Harvey’s (1982) notion of successive waves of spatial fixes being created and destroyed as capital enters and exits particular places, such that regional development involves both economic path creation and path destruction. This means that regions’ economic trajectories have a certain degree of momentum which is affected by what happens both within them and beyond them, with endogenous and exogenous forces having different pull on regions at different times, thereby creating spatially distinct outcomes in different places. Thus, as MacKinnon (2012: 234) puts it, ‘the evolution of regions is conditioned by their positionality in relation to wider processes of uneven development and institutional regulation’ – in other words, by the geographically specific coupling, decoupling and, importantly, recoupling of TNCs and local institutions. Given that the role of local/regional institutions in regional development has usually focused on enhancing the creation of value (e.g. through training the local workforce), on facilitating knowledge and technology transfer, on aiding processes of industrial upgrading, and, most crucially, on augmenting value capture, we must consider questions of power and contestation within and between localities if we are to understand GPNs and their dynamics. What this all suggests is that, when considering the evolution of commodity production and distribution networks, we must add analytical attention to the layered histories and uneven geographies of capitalist expansion, disinvestment and devaluation in particular places, for capital investment and accumulation in one place can engender through its very logic disinvestment and devaluation in others (Barker 2006: 81). Consequently, we need to connect any analysis of global commodities and the networks through which they travel from place to place to the politics and geography of disinvestment, devaluation, and place- and subject-making which make their production possible (Bair and Werner 2011: 990). Thus we are not only examining uneven development, but also uneven and combined development.

If we have so far focused on the ways in which GPNs are geographically embedded, it is important to recognise that workers are likewise rooted in space (though this does not mean that they are immobile). In this regard, Coe and Jordhus-Lier (2011: 228) have argued that all workers are inevitably but differentially positioned within four sets of social relations or institutional formations, each of which have particular spatialities attached to them: global production networks; state institutions at various scales; the social relations and associational life of the community; and intermediary agents in the labour market. Strangleman (2001), too, has suggested that individual workers frequently find themselves entwined in four sets of networks that affect the character of labour relations in specific locales, all of which likewise have definite, if different, spatialities to them, though he identifies slightly different ones: occupation/work; the connections of the locales in which workers find themselves to particular other places which may be close or far away (sources of raw materials or capital, origin points of labour migration streams, markets and so forth); class relations and how these have developed across space and in particular places; and family and kinship ties. Collectively, these networks within which workers find themselves link what happens outside the workplace with what happens inside it. However, neither Coe and Jordhus-Lier nor Strangleman address in detail questions about the labour process within the workplace. This last step is however taken by Selwyn (2012), who has suggested that the sort of approach we are outlining here pushes GPN analysts to investigate four interconnected things: 1) the material requirements of commodity production, particularly in GPNs; 2) attempts by firms to establish a labour force and structure a labour regime around these requirements, within the context of competitive accumulation, including contests over the institutionalisation of state practices; 3) processes of class formation in (and often beyond) the geographical region of commodity production; and 4) the ways workers’ movements and organisations attempt, sometimes successfully, to structure (through structural and associational power [see below]) the socio-spatial environment in their favour.

Putting these pieces together, then, produces a framework for understanding how the ongoing and ever-changing interrelationship and overlapping of these networks significantly affects the particularities of place and, thus, how any place is experienced by the individuals and organisations located within it. As a result, we can recognise that specific local articulations between workplace and community politics dramatically shape the extent and character of social agency in each place in which the nodes of any commodity chain are located. This fact demands careful consideration of the articulation between the sphere of production and the sphere of reproduction and the way in which the household connects with the workplace, labour market and regulatory spaces of the state (Coe and Jordhus-Lier 2011: 224). Most importantly, though, integrating the Coe and Jordhus-Lier/Strangleman/Selwyn approach(es) with our earlier argument concerning GPNs’ strategic (de)coupling and regions, we would suggest, provides a framework for analysing the position of small firms within GPNs, but also a basis for analysing the patterns and dynamics of the labour process within them, with such dynamics being conditioned by events both internal to, but also beyond, the workplace. Adopting such an approach allows us to recognise that the workers from whom value is extracted and who make the goods/services that travel along GPNs are deeply spatially embedded in particular places and that this has implications for how networks evolve in response to workers’ geographically conditioned behaviour, whether these are workers who toil in very large or quite tiny firms.

Having thus briefly outlined why the places that form the nodes of GPNs, and within which those who produce the commodities which flow along them live and work, need to be conceptualised as dynamic and socially constitutive locales rather than as merely inert points on a map, we now shift our focus to small firms. This is because small firms are typically locked into (or excluded from) these geographically unevenly developed networks by dint of their relationships to larger firms and the state (Barrett and Rainnie 2002; also see below) and this inclusion/ exclusion is generally presumed to have specific consequences for their labour forces in particular places.

Small business, vulnerable labour and dominant firms

Without question, TNCs’ restructuring of their supply chains and their increasing adoption of subcontracting strategies as a response to what is happening in the global marketplace is producing intense pressure on those bottom-tier firms that supply goods and services to global giants. Interestingly, many are responding by engaging in mergers with the firms that they supply so as to gain better negotiating positions in relation to other top-tier firms or bottom-tier rivals. This means that if we examine firms not just as legal entities but also by their sphere of social and spatial influence, far from being hollowed out or downsized, large firms in the top tiers can be seen to have increased enormously in size in recent years through their closer embrace of those smaller second- or third-tier firms that were previously viewed as operating at arm’s length from them. Importantly for the argument here, however, such embracing raises fundamental questions concerning how the agency of workers in such second- or third-tier firms may shape the personality of the large firms viewed as first-tier ones. ‘Embracing’, in other words, is a two-way process in which each party can potentially influence the other’s behaviour.

Opening up to investigation the position of workers in small, second- or third-tier firms is important, we would like to suggest here, precisely because so much about workers in such small firms is often assumed rather than demonstrated. Hence, much early work on the restructuring of supply chains, restructuring largely driven by the demands of just-in-time production, simply supposed that small firms generally find themselves in secondary or tertiary, and thus dependent, positions in GPNs (see Rainnie 1992). In such analyses, large firms were typically seen to be reforming their supply chains by concentrating on a smaller number of much larger primary suppliers, and this reformation was assumed to have negative consequences for workers further down the supply chain. This early formulation has been reinforced by more recent work examining the changing strategies of lead or dominant firms. For instance, Nolan et al. (2008) have outlined how intense competition has driven an unprecedented wave of industrial concentration and consolidation that has, they suggest, generated a ‘cascade effect’, wherein the process of large firm concentration, through the simultaneous de-merger of non-core business and merger of core business, spreads across the value chain at high speed. However, Nolan et al. simply presume that such spread is to the detriment of small firms.

For their part, Christopherson and Clark (2007) have examined dominant firms’ capacity to strategically combine different locational and labour-force options, whilst Morgan (2009: 589) has explored how TNCs in liberal market economies often lock themselves into clusters in particular localities but only weakly connect to these localities’ own institutions. Importantly for the argument here, though, these authors have suggested that networks frequently take the form of isolated hierarchies, and these are perceived as bringing only marginal benefit to less powerful members of the chain and to the regions in which they are located. As a result, regional small-firm clusters are interpreted as being important to TNCs (or other dominant firms) only in so far as they support their international competitiveness. In such analyses, then, the dominant large firms are seen as the system integrators, the ones driving the chain’s form, whilst the localities in which such clusters congregate are deemed to be little more than the bearers of large firms’ actions.

Significantly, in such approaches the growing financialisation of GPNs – that is to say, the rise of the financial sphere made up of institutional investors and executives of large corporations at the top of GCCs/GVCs/GPNs – is often viewed as exacerbating any already existing inequalities in the distribution of value in the supply chain, thereby making large firms even more dominant (Palpacuer 2008). Such shifts in dominant firms’ relationships to one another and to those firms lower in the hierarchy are noteworthy, Scarborough (2000: 14) has argued, because there are important human resource management (HRM) consequences that emerge from an organisation’s positioning within the supply chain. For one thing, they have obvious implications for the possibilities (or otherwise) of upgrading within networks, a subject to which we will return later. (For a more sophisticated HRM-focused analysis, see Rubery et al. 2010.) In contemplating such consequences, though, it is generally assumed (e.g. Sacchetti and Sugden 2003) that the further one ‘descends’ down a supply chain, the smaller is the unit of production, and hence the worse are the working conditions (see Hurley with Miller 2005 for an account of the clothing production network). The result is that, again, there is little room for small firm workers’ agency in the analysis of GPN dynamics and the supposed characteristics of such small firms are instead simply imputed, based on an understanding of where they lie within the hierarchy.

Certainly, not all writers have ignored small firms, and indeed some have hinted at the agency of those employed by small firms. For instance, Rutherford and Holmes (2008) have suggested that it is possible for clusters of innovative small firms to actually ‘embed’ TNCs within themselves, thus reducing power asymmetries between themselves and large firms. This embedding can come about because, at a regional level, both large and small firms in GPNs rely on localised economies of scale (concentrations of specific knowledge and expertise) and economies of scope (intangible assets of learning and cooperative behaviour). Regions can therefore attract GPNs by providing conditions for the creation, enhancement and capture of value that are not replicated elsewhere: for instance, the local state can play a role in promoting training, cooperative labour relations, technology transfer and upgraded innovative capacity in small and medium-sized enterprises (SMEs), with different local states emphasising unique configurations of such activities. However, although they acknowledge that small firms may play an important role (both individually and in clusters) in research and development (R&D) and innovation, often through providing the services previously conducted within the TNCs’ structure but subsequently outsourced, Rutherford and Holmes nonetheless argue that TNCs generally use their control of GPNs to harness innovation in SME networks and to capture regional economies of scale and scope fostered by state policies. That is, they view the power relationship between the two as fairly unidirectional. The overall conclusion to be drawn from all this is that the power of original equipment manufacturers within GPNs steadily erodes the cluster-based advantages of suppliers in general, and SMEs in particular (Rutherford and Holmes 2008: 540).

Likewise, Whitford and Potter (2007) have explored how successful industrial districts and clusters of SMEs are becoming increasingly integrated into global supply chains. Nonetheless, they argue that this may not be a positive experience, as value chain modularity largely precludes the possibility of upgrading the local clusters’ position within any such chain. Again, control and agency are largely seen to flow down the chain, from the top. More positively, if contentiously, Crouch et al. (2009) argue, based on the industrial sectors they examined (the furniture industry in Nordrhein-Westfalen and southern Sweden; the car industry in Eastern Germany and in North-western Hungary; the biopharmaceutical industry around Munich and Cambridge; and new television and film-making in Cologne and London), that what they describe as ‘creative incoherences’ within national varieties of capitalism can allow networks of SMEs to create local innovative structures at subnational levels. Certainly, they do tend to focus on the way in which large firms are ‘rule-takers’ of national institutions (i.e. the way national institutions shape large firms’ behaviour and so steer their actions towards the path ‘prescribed by the encompassing national mode of capitalism’ [p. 673]) but are ‘rule-makers’ of the local system, thereby presenting a picture in which large firms are seen as dominant. Even so, their argument that in the places they studied the ‘governance of the local economy decouples itself from national institutions and develops its own institutional dynamics’, such that ‘local innovation and production regimes are likely to be the “entry gates” for institutional change’, does seem to open the possibility for small firms to have some influence vis-à-vis larger firms.

Whereas the above authors’ approaches are, then, fairly typical of the contemporary literature concerning the assumed deleterious impacts of commodity/value chain and GPN restructuring upon the lives of workers in small firms, it would be unfair to intimate that there have not been challenges to such interpretations. For example, Barrientos et al. (2010), through their ‘Capturing the Gains’ project (www.capturingthegains.org), which has been concerned with promoting decent employment in global production networks, have argued that the expansion of global production in labour-intensive industries has actually been an important source of employment generation for workers who would otherwise not be able to find paid employment. Furthermore, many jobs have been filled by women and migrant workers, who previously had difficulty accessing this type of waged work. In other words, restructuring has not simply sent bad consequences down the GPN, but has benefited some workers at the bottom. Moreover, Barrientos et al. suggest that GPN restructuring can actually bring about three trajectories of work upgrading:

However, they also suggest that employment remains insecure and unprotected.

Whereas the above authors have largely focused on the organisational structure of GCCs/GVCs/GPNs, and on the individual firms that make them up, Starosta (2010), on the other hand, has attempted a Marxist analysis of global commodity chains (GCCs), developing an alternative account of the social determinants underlying their genesis, structure and evolving configuration. Within this formulation, he seeks to focus less upon the individual nodes (i.e. firms) and their places within GCCs and instead to view the chains as a whole, as an expression of the unfolding of the law of value. Building upon the initial critique of GCC analysis laid out by Smith et al. (2002), he asks that theorists rethink the relations between the embedded economic actors and activities within GCCs and the more general global dynamics of capital accumulation. He does so by suggesting that ‘commodity chains are the social form through which certain “normal” (sic) capitals appropriate the surplus value released by small capitals’ (pp. 450-451). For him, then, GCCs are the concrete form taken by the competition between larger organisations over the extra surplus value that ‘escapes the hands’ of small capitals. In this approach, hierarchical command and control relations are seen to characterise the relationship between large and small firms, whereas more horizontal and competitive relationships characterise those between large capitals.

We do not necessarily disagree with Starosta’s argument and agree with Barker (2006: 81) that competition between capitals not only assumes a starting point of unevenness (including size), but also generates unevenness in the shape of the stratification of capitals and the redistribution of capital among participating capitalists (see also Selwyn [2012] for more on the relationship between the stratification of capitals and uneven and combined development). Nevertheless, we contend that there are still problems with how far Starosta takes the analysis. In particular, we maintain that this formulation still contains a rather simplistic characterisation of all small firms as being secondary, competitive and dependent. Part of the reason for this is that there is no actual investigation of the labour process characteristics of organisations at the ‘lower’ end of the chain. Rather, in Starosta’s framework small firms can play one role only – that of handmaidens to larger firms. Indeed, he even goes so far as to argue (2010: 446, emphasis added) that ‘small capitals are by nature incapable of being at the vanguard of technological development’. We disagree, arguing instead that if we mobilise the notion of uneven and combined development (see Rainnie, Herod, and McGrath-Champ 2010) then we can develop a more nuanced approach to the role played by small firms and their workers within GPNs.

Uneven and combined development, the small firm and agency

In our earlier approach to GPNs and uneven development (Rainnie, Herod, and McGrath-Champ 2010), we argued that local economic and social structures are the complex and institutionally mediated outcome of the distinctive role played by localities in successive spatial divisions of labour. For instance, communities that serve as homes to branch plant manufacturers are tied into corporate networks in different ways than are communities that host research-and-development or headquarter functions. This impacts the types of class relations that develop in these different places, but also shapes the kinds of economic development that might subsequently occur in them. Hence, as Massey (1984) has shown, the industrial geography that was laid down in early 19th-century Britain affected patterns of capital investment well into the 1980s. The deindustrialisation of old steel and chemical communities during the 1960s created lots of unemployment and thus cheap labour, which then encouraged light manufacturing to locate in these communities (though, interestingly, such manufacturers frequently hired women rather than men, thereby changing gender dynamics in the country’s traditional industrial heartland). Likewise, as Harvey (1982) has detailed, the way capital organises itself geographically to create particular ‘spatial fixes’ is fundamental to the way it works. Putting all of this together, then, it becomes quite apparent that uneven development is not an accident or a by-product of economic development, but, rather, is vital to the nature of capitalism.

A central element in creating geographically uneven development is the fact that the law of value under capitalism is based on a contradiction between two opposing tendencies within capital itself. As Smith (1984) has argued, on the one hand capital must be embedded in particular places so that accumulation can take place: accumulation does not occur on the head of a pin, but in real communities that have histories, geographies, people and social formations. On the other hand, capital must remain sufficiently potentially mobile that it can readily disembed itself from one place and move elsewhere if opportunities arise. The result of these opposed tendencies is that the economic landscape is produced as a pockmarked one, with some places seeing deindustrialisation at the same time that others enjoy inward capital investment. Furthermore, according to Castree et al. (2004), place differences persist because of, not despite, heightened interplace connectivity. These differences arise partly because of the nature of places’ internal economic and social relations and the local histories, traditions and political institutions that these have helped create; and partly because not all places are connected to the broader world in the same way. In this way, non-local processes combine with existing local differences to produce unique outcomes. However, places are not simply different from one another, but they are unevenly and causally linked: what happens in one place impacts on what happens in those to which it is connected by lines of corporate control, flows of capital investment or streams of migration. Place relations, in other words, construct unevenness in their wake and operate through the pattern of uneven development already laid down. Thus we must talk of uneven and combined development.

Given the above, studying how uneven and combined development evolves is necessarily a form of analysis that starts from the totality of capitalist relationships and the interconnectedness of the different parts of that totality (see Barrett and Rainnie 2002). Such a starting point allows small firms to be analysed relationally, within their concrete organisational and locational circumstances, rather than as discrete politico-economic objects. It provides, in other words, for an approach to understanding industrial/employment relations in small firms that starts with the character of economic relations and capital’s contradictory constituents, as opposed to with the small (or even large) firm itself. In such an approach, instead of seeing small firms as the relics of a bygone age or as the shock troops of economic change à la Piore and Sabel (1984), their function can be seen to be more complex and integrated into current patterns of restructuring and development. In general, then, such an approach argues for viewing the present state of capitalism’s economic landscape as a palimpsest of the vestiges of different historical stages of development, an amalgam of apparently archaic and more contemporary forms of development. Interrogating the relationship between small and large firms from such a conceptual vantage point, rather than simply counterposing them, we suggest, provides a means whereby the heterogeneity, significance and agency of small firms can be accommodated. Such an approach is similar to that of Selwyn (2012), who, in a discussion of the dynamics of uneven and combined development and the role of the state in driving development agendas, points to the role played by institutions in realising the advantages of ‘backwardness’. However, where we differ is that Selwyn, without necessarily describing all small firms as ‘backward’, nonetheless seems to see large organisations as always capable of taking advantage (in a variety of different ways) of a form (the small firm) that has been viewed as the dominant form in an earlier period of capitalist development (i.e. during the period of ‘competitive capitalism’ that dominated from the beginnings of the Industrial Revolution until the era of monopoly capitalism emerged at the end of the 19th century) but which is now no longer envisioned by many as having much puissance. By way of contrast, we do not see large firms as inherently more powerful than small firms.

Drawing on such a perspective, Barrett and Rainnie (2002) developed a heuristic device in which the role of the small firm in a modern economy is seen to be defined in relation to the strategies of large firms, rather than in and of itself. Specifically, at least four categories of small firm can be identified, these being: dependent, dominated, isolated and innovative. In this taxonomy, dependent firms are those which complement and service the interests of large firms through, for example, subcontracting relationships. This group would most closely resemble the classic small firm and its workforce in supply chains. Historically, subcontractors in the clothing industry in the UK have been emblematic of such dependent firms and they continue to play such a role as this industry internationalises, a fact which has important implications for the much-discussed possibilities of upgrading (Hurley with Miller 2005). Dominated firms, however, differ from dependent firms in that they compete with large firms, often through intense exploitation of labour and/or machinery. Workers in this sort of firm are often on the receiving end of inferior pay and conditions of employment, and for this reason such firms often attract/seek employees with limited bargaining power and job opportunities. A classic example here would be small retailers serving highly localised markets, but coming increasingly under the influence of growing and restructuring supermarket chains. Isolated small firms, on the other hand, exist in specialised or geographically discrete markets that remain (if only temporarily) unattractive to large capital, although large-scale capital can access and dominate this sort of market through the development of franchising forms of organisation. This category would include fast-food franchises with very precarious labour, wherein large firms exploit the small firm form (in every sense of the word). Firms in this category, however, do not necessarily offer the lowest terms and conditions of employment – the category would also include doctors, dentists and local law firm practices, for example. Finally, innovative firms operate in markets (which they have themselves often founded and developed) creating new products or services.

For his part, Hadjimichalis (2011) has argued that geographically and sectorally dispersed small producers and service providers are one of three main categories of small firm (the other two being varieties of clusters). Although he argues that, for ideological and political reasons, the innovative capacity of small firms has often been overhyped, he nevertheless recognises that there are in any economy many thousands of small firms that perform in innovative and novel ways – small firms, he notes, are increasingly present in technology-intensive industries and in creative industries, particularly in large urban areas. Similarly, Ó Riain (2011: 22) points to evidence in the USA of a rapid increase in the number of scientists and engineers working in small firms in Silicon Valley and its many imitators, as part of an open system of innovation. Furthermore, the dominance of Fortune 500 company labs in innovative scientific breakthroughs in the 1970s is being increasingly challenged by university and federal labs, and by collaboration between small firms. Certainly, the nature of the innovative small firms’ activities can make them vulnerable to takeover or acquisition by large capital because, having let the small firm take the risk of the innovatory process, large firms are often in a better position to appropriate the results. Nonetheless, this is not an inevitability.

Such a deconstruction of the category of ‘small firm’ allows us to escape the fallacy that small firms (and thus their employment/HRM characteristics) are homogeneous, and so moves us away from simplistic notions of upgrading, wherein small firms are assumed to ascend the supply chain and so improve the jobs of their own workers. But it also locates such firms within the totality of capitalist relationships. The strength of this formulation is that it recognises that – in certain sectors and under particular circumstances – small firms can actually lead innovation (cf. Starosta), even if they often lie open to the predations of larger firms which can dominate the development part of R&D. Flecker (2009: 256) concurs, arguing that the lengthening and increasing complexity of value chains has opened up opportunities for small companies to act as intermediaries, for example organising the value chain for brand owners. Using this form of analysis we can therefore analyse the ways in which small firms, individually or collectively (through clusters), fit into GPNs, and the ways in which they do not. Consequently, we can begin to analyse the role of small firms, individually or severally, in the processes of layering, conversion and recombination. In particular, we can scrutinise how the small firm’s position within the broader GPN might impact the character of localities and the lives of the workers living within them. In turn, given that workers will be socialised differently in different localities and that such localities will be linked together in different ways, depending upon the specificities of particular GPNs, this form of analysis also provides a conceptual entrée to consider how workers in small firms might differentially shape GPNs’ evolution. This form of analysis, in other words, allows for an examination of the creation of value, as well as its appropriation and capture.

Discussion: So what does this mean for the dynamics of GPNs and small firms within them?

Above, we have outlined an approach that can critically re-evaluate the dynamics of global production networks and, in particular, the position of small firms within them. This is done with a view to being better able to analyse the dynamics of labour process development within GPNs, and thus issues of worker action, organisation and resistance. Against the longstanding view of small firms as being necessarily in secondary, subservient and often dependent positions at the ‘bottom end’ of production networks, with the condition of workers being taken as equally weak and vulnerable, we propose a more complex and nuanced approach to the discussion of the possibility of upgrading within GPNs. In particular, we suggest that a concentration on the dynamics of value chains and/or production networks and different forms of governance does not mean that we should simply accept a uniform picture of dominant firms as drivers and systems integrators at the top of the pile, with firms steadily declining in size, power and influence as we descend further down the chain/network. In fact, the scalar language of ‘ascending’ or ‘descending’ is unhelpful in this respect, for it assumes that those firms at the bottom of the ladder are always, by definition, weaker than are those at the top. By way of contrast, here we present a dynamic form of analysis that can allow for developments within and between GPNs, as well as contradictions within and between them. In particular, if we view GPNs as multi-branched network structures, in which intricate linkages form multi-dimensional, multi-layered ‘lattices’ of spatially embedded economic activity, then a more complex picture emerges. The disruption caused by the 2011 massive earthquake and tsunami in Japan well illustrates this point, as several automobile producers and manufacturers of other goods in both Japan and overseas were forced to shut down plants due to supply-chain disruptions caused by the closure of smaller firms ‘lower down’ the GPN hierarchy. Toyota, for instance, closed down production in plants as far away as Turkey, France, Poland and the UK because it could not get components from some of its Japanese suppliers (Herod 2011). Such idlings have highlighted the vulnerability of lean, extended, geographically dispersed supply chains, as has labour unrest in several electronics plants in Japan. Indeed, these disruptions and others like them have led many manufacturers to question the wisdom of single-chain, elongated sourcing strategies to the point where some are now moving towards adopting multiple supply chains that emphasise close connections to regional suppliers (Marsh 2011). This questioning attests, we would argue, to the power that small firms and their workers can exert over a GPN’s structure and dynamics.

Not only, then, do geographically elongated supply chains face problems of coordination over long distances – a factor that highlights the constitutive nature of space in shaping the way they function – but relationships within them are dynamic and often open to challenge by small firms and their workers. For instance, Levy (2007) sees GPNs as contested organisational fields in which the distribution of income amongst the various actors who make up different parts of the chain is determined by struggle. Equally, matters of supply-chain governance involve not just logistics but also technical standards, quality and design. Levy therefore puts forward an alternative definition of GPNs as ‘integrated, economic, political and discursive strategies with a degree of structural stability, but subject to challenge by strategic actors’ (Levy 2007: 26-27). Taking a neo-Gramscian point of view, he views contested GPNs as highlighting the multiple dimensions of hegemonic strategies and the potential for subordinate groups to develop coordinated strategies that take advantage of tensions and points of divergence in complex socioeconomic systems (Levy 2007: 30). In a similar vein, Coe and Jordhus-Lier (2011: 221) argue that, in seeking to reveal the fragmented yet tightly coordinated organisation of capital at the global scale, GPN analysis can simultaneously serve to reveal the variegated landscape for agency potential across different sectors. This kind of analysis can reveal weak spots, potential lines of solidarity and potentially deleterious impacts of worker action on other groups and the GPN itself. Coe and Jordhus-Lier (2011: 222) conclude that a GPN-informed perspective allows the potential for labour agency to be interpreted and evaluated in the context of the wider global system of which workers are a part. This, it is contended, has great political as well as analytical significance. It is also important in so far as it militates against a view of a vertically structured hierarchy either organisationally or in terms of worker orientation, organisation and action.

The approach presented above, then, has important implications for workers in SMEs far beyond simply having their working conditions improved by external upgrading processes. Here, the move away from homogeneity and determinism in approaching the analysis of the labour process in small firms allows for greater room for the possibility of worker organisation and activity. By linking the labour process in some small firms into structures such as GPNs and current rounds of restructuring, possibilities open up for labour agency. Yet, even in cases where small firms have not been able to associate themselves with larger ones so as to exercise some direct influence in how GPNs are shaped, workers are not necessarily powerless because of what Selwyn (2007) calls the ‘bullwhip effect’, whereby a small disturbance at one node in the supply chain leads to increasingly large disruptions further up or down the chain. Indeed, drawing upon Wright’s (2000) distinction between workers’ associational and structural power, wherein associational power derives from collective organisation and structural power accrues simply from workers’ position/location in the system, with this latter being further subdivided into marketplace bargaining power (derived from changes in supply and demand) and workplace power (derived from workers’ strategic location in the production process), Selwyn shows that in the distribution, automobile and export horticulture sectors, labour’s political or associational disorganisation does not necessarily weaken it structurally. In contrast, the growth of ever shorter lead times and lean production actually offers workers new sources of structural power that may not have been present in previous epochs. Hence, as disputes at United Parcel Service (UPS) (Coleman and Jennings 1998), General Motors (Herod 2000) and by grape workers in North East Brazil (Selwyn 2007) undeniably highlight, relatively small groups of workers can have swift and massively disruptive impacts upon chains when taking action in global structures that are dependent on just-in-time delivery systems, lean production and global logistics (see also Rainnie, McGrath-Champ, and Herod 2010). Such reality needs to be recognised in the theorising of GPNs.

What all this suggests, we would argue, is that any investigation of the conditions of labour within small firms under capitalism cannot start from governance structures of commodity chains, managerial strategies, or similar concerns. Rather, it must instead start from an analysis of the organisation and control of the labour process in concrete settings. The dynamics in these concrete settings vary tremendously, depending on how workers and firms are spatially embedded within them and how they are spatially linked to other communities near and far. However, whilst the outcomes of local struggles might shape the kind of local economy that emerges (as we have argued), these struggles are frequently about the consequences of responding to pressures beyond the immediate locality: that is, to world market pressures, aka globalisation. Given this, we cannot make assumptions about the conditions and structural power of workers in small firms based on their presumed location at the bottom of vertically organised GPN hierarchies. Instead, we must examine them directly.