One firm, two countries, one workplace model? The case of Foxconn’s internationalisation

Abstract

This article aims to provide insight into the employment relations in China-based multinational companies internationalising to Europe, a still relatively unexplored topic. We investigate the transfer of work and employment practices from Foxconn’s manufacturing headquarters in mainland China to its subsidiaries in Czechia and the factors that influence the firm’s internationalisation of production. By drawing upon original ethnographic fieldwork, the study makes a two-fold contribution. First, it shows the analytical inadequacy of the ‘latecomer’ model which assumes that the Chinese firm is an asset seeker. Second, it illustrates the relevance of diversity of labour and non-institutionalised forms of workers’ agency for theorisation of internationalisation. These topics are still insufficiently addressed by the literature that favours managerial agency and the model of distinctive and stable national labour forces. The study contributes to the theoretical debates on internationalisation by illustrating the limits of the national institutionalist perspective, the importance of considering a multi-scalar analytical framework and the relevance of labour composition in shaping multinational employment relations.

Introduction

The upsurge of Chinese investments and their geographical coverage and diversification have given rise to the fear that Europe will be taken over by Chinese firms (cf. Burgoon and Raess, 2014; Meunier, 2012). Chinese foreign direct investment (FDI) has diversified from manufacturing and heavy industries to real estate, food and agriculture, transport and infrastructure, and financial institutions. Although the majority of FDI still flows into Western Europe, the economies of Central and Eastern Europe (CEE) have boosted their share of Chinese investments and in November 2016 formalised their partnerships with Chinese investors through a €10 billion China–CEE Investment Cooperation Fund (Reuters, 2016). The increased presence of Chinese multinational corporations (MNCs) in Europe has led to a rapid growth in studies on the outward expansion of Chinese firms. These studies focus on the microeconomic determinants of Chinese FDI such as choice of location, modes of entry and motivation for FDI (Cozza et al., 2015; Rios-Morales and Brennan, 2010). However, these studies do not explore the link between employment relations (ERs) and internationalisation, despite ample evidence that the diffusion of ER is critical for the globalising strategies of MNCs (Almond and Ferner, 2006; Dörrenbächer and Geppert, 2011).

To date, very few studies have been conducted on the relevance of ER for the overseas expansion of Chinese MNCs operating in Europe. These studies are limited both in terms of topic and geographical coverage. With the exception of a study on the Chinese takeover of a motorcycle company in Italy (Zhu and Wei, 2014), most research is UK-focused and examines the ways in which firms adopt UK’s advanced human relations (HRs) practices and transfer these to China (Shen and Edwards, 2004; Zhang and Edwards, 2007). Moreover, theoretical development remains narrowly focused on the reverse diffusion of HR intended as the transfer of ‘best practices’ from MNCs’ subsidiaries to the parent company in the home country. As MNCs from China are ‘latecomers’, much research explores firms’ low technical and managerial competence and ‘catch up’ strategies.

This article proposes that studying the internationalisation of MNCs from China through the prism of the latecomer model has negatively affected the research field in two ways. First, the focus on home versus host country effects has overemphasised the homogeneity of distinct national models. Second, the studies on how internationalisation is operationalised have placed managers at the central stage and overlooked the role of labour in shaping the transfer of ER. This study begins to fill this gap in the literature through a case study of Foxconn. We address two interlinked questions. First, as per the latecomer model, are MNCs originating from a developing economy less likely to export the management systems from their country of origin to their overseas subsidiaries in a developed country? Second, to what extent the transfer and labour relations in the subsidiaries are shaped by the agency and heterogeneity of the workforce?

To answer these questions, the article draws on an in-depth case study of Foxconn’s subsidiaries in Czechia over the years 2012–2018. The Czech subsidiaries, in Pardubice and Kutna Hora, are Foxconn’s first and most important European plants; moreover, it is Foxconn’s manufacturing headquarters for the Europe, the Middle East and Africa (EMEA) region. Foxconn is the third-largest private employer in the world and the world’s largest electronics contract manufacturer. Its core business is contract electronics manufacturing, mainly for international buyers. Foxconn’s largest buyer by far is Apple, but its clients include Samsung, Microsoft, Google, Intel, HP, IBM, Dell, Cisco, Amazon, Sony and Nintendo. Although the firm is Taiwanese-owned, its manufacturing headquarters as well as the bulk of its factories are located in mainland China. While we recognise the differences between Taiwan and mainland China, Foxconn’s case fits the latecomer model as both Taiwan and China are categorised as emerging economies (Chang et al., 2009: 77), and Czechia, a welfare state with the continental European social model and advanced high income, is categorised as a developed economy.

By investigating the ER that Foxconn applies in its Czech subsidiaries, this article exposes a greater variety within the latecomer model than previous studies have posited. This variety is most visible, we suggest, when attention is paid to the industrial context and to labour agency. Furthermore, highlighting the limits of the institutional perspective, we suggest analysing ER on the basis of global production networks (GPNs). We build on the GPN literature as it allows an examination of the relevance of sectoral specificity for internationalisation, as well as the role of labour in shaping ER and production systems in MNCs. The study will, therefore, update the recent literature on the ER of Chinese firms in Europe and revisit the transfer debate that so far has paid limited attention to labour agency in its theorisation of internationalisation.

Internationalisation of Chinese MNCs and employment practices

Chinese MNCs that have expanded overseas have been labelled ‘latecomers’ (Hay and Milelli, 2011), a term indicating late arrival on the industrial stage. Chinese MNCs have been expanding their operations overseas only since the late 1990s, following an earlier inward internationalisation process. Scholars suggest that as latecomers, MNCs from developing countries located in advanced economies will aim to acquire management practices of the advanced economies to compete in international markets and compensate for their latecomer disadvantage (Zhang and Edwards, 2007). The subsidiaries, therefore, play a key role in learning and absorbing advanced management practices. Zhang and Edwards (2007) found that Chinese MNCs in the UK displayed a high level of adaptation to the host environment and used reverse diffusion to transfer the advanced management practices and expertise to China. Subsidiaries also function as a ‘learning channel’ (Cooke, 2009) through which Chinese managers acquired knowledge about industrial relations (IRs) strategies. Despite the considerable labour disputes and unrest in China during the last decade (Pringle, 2013), Chinese MNCs seem to underestimate the significance of labour relations in host countries and misjudge the nature and function of foreign trade unions. Based on the advanced management practices employed by subsidiaries, Chinese managers could develop strategies such as taking advantage of HR or applying a passive conflict avoidance approach to reduce conflict and initiate interaction with organised trade unions (Bian and Emons, 2017). Overall, scholars have found very little evidence of Chinese MNCs transferring their management practices to Europe. Therefore, a general agreement in the literature is that MNCs from China apply a ‘light touch approach’ (Miedtank, 2017) to their overseas units and use the subsidiaries as learning channels.

The idea of lateness based on the national stages of industrial development limits our understanding of internationalisation owing to the narrowness of a research agenda that relies on the home versus host country effects embedded within the distinctive National Business System (NBS). Institutionalism, in particular its comparative strand, has demonstrated that no single coherent institutional framework exists. Moreover, the transfer of organisational practices is a ‘political and contested process’ (Ferner and Tempel, 2006: 29) that needs to be understood in relation to the interactions of institutional domains at local, national and supranational levels (Gamble and Huang, 2009).

Based on the radical transformation of CEE economies since 1989, scholars discuss whether institutional deregulation is producing a convergence with one of the two classic models (Baccaro and Howell, 2011; Howell and Kolins Givan, 2011) or is best understood as a new model of capitalism (Nölke and Vliegenthart, 2009). The pressure on the CEE countries to liberalise their labour markets, open their economy to FDI inflows and privatise (Meardi, 2012), led to a fierce competition among them. This resulted in the growing fragmentation of national state spaces (Drahokoupil, 2009) and favouring of a production archipelago on the basis of FDI. At the same time, this process shaped various arrangements that do not fit with the existing models of capitalism. Bohle and Greskovits (2007) suggest that CEE countries can be grouped according three models, neo-corporatist (Slovenia), liberal market (Baltic State, Bulgaria and Romania) and intermediate position (Visegrad countries such as Czechia, Hungary, Poland and Slovakia), with competitive export platforms anchored in FDI and welfare institutions. Similarly, Nölke and Vliegenthart (2009) posit Visegrad countries as a third variety of capitalism, classified as a Dependent Market Economies (DMEs). The DME categorisation is relevant because it takes into account transnational influences and overcomes traditional methodological nationalism, typically of the Varieties of Capitalism (VoC) approach that is based on national states ‘as the natural institutional containers’ (Brenner et al., 2010: 187). In a similar vein, studies on the Chinese model of capitalist development suggest that work and employment in China are characterised by a heterogeneity of business models (Lüthje et al., 2013; Smith and Liu, 2016). In addition, scholars stress the substantial degree of regional heterogeneity that challenges national institutional coherence and suggest that ‘Chinese capitalism’ is ‘embedded in national regime and constitutively connected to offshore economies and networks’ (Zhang and Peck, 2016: 57).

We analyse Foxconn’s overseas expansion, building on the above scholarship to examine the limitations of the latecomer model based on the notion of NBS. In addition, we propose a multi-scalar analytical framework, whereby we suggest investigating how internationalisation is deeply networked into a range of institutional scales, as well as into GPNs. Studies of internationalisation from an institutional perspective privilege managerial agency (Cooke, 2009), perceive the workforce as nationally bound and acknowledge labour’s active role only when institutionally organised as trade unions (Aguzzoli and Geary, 2014). On the contrary, the GPN analysis highlights the importance of labour as an analytical category to explain the global geography of production (Newsome et al., 2015). To achieve this, scholars have stressed the necessity of creating analytical space for individual and collective action, paying attention to the extent to which the workforce is ‘variegated’ and ensuring that the strategies of labour agency are spatial (Coe and Jordhus-Lier, 2011). This approach has helped to better understand the role of workforce variegation and the social embeddedness of labour in the mode of insertion of a certain location within a certain type of GPN (Newsome et al., 2015).

Based on studies of GPN relating to how labour dynamics impose location-specific production and labour-control regimes, we focus on the transnational composition of labour and non-institutionalised forms of workers’ agency. We examine work stoppages, turnover and exit as practices through which workers have pressured MNCs to alter their ER. In positing labour as a determinant of internationalisation, the article contributes to revealing the role of labour in driving the adjustment of both MNCs’ ER and trade unions’ bargaining strategy. This approach is not aimed at showing the power of workers to disrupt capital but rather at indicating how the diverse composition of labour and its ‘labour mobility capacity’ configure the processes of local capitalist development (Smith, 2006).

Methodology

A major weakness in the existing literature on Chinese investment in Europe is that it is dominated by quantitative data, statistical methods and desk research. Studies of ER in the overseas Chinese MNCs are often based on a ‘snapshot’ perspective for problems linked to distance, time, costs and access (Gamble and Huang, 2009). To counter this weakness, qualitative and longitudinal methods are adopted that are particularly effective for understanding the transformation over time (Gamble and Huang, 2009) and observing the changes in workforce composition. Accordingly, our fieldwork was conducted over an extended period of time, and since 2012, the site was visited each year for several weeks. During the fieldwork, 70 interviews were conducted in Pardubice, Kutná Hora and Prague: 12 with privileged witnesses (heads of the public placement office and local labour inspectorate, job agency managers, a labour ministry delegate, union officials and local NGO members), and the other 58 with workers directly employed by Foxconn (24 with Czech, Mongolian, Vietnamese, Ukrainian, Bulgarian and Slovak workers) as well as migrant agency workers (34 with Romanian, Bulgarian, Slovak and Polish workers).

To show the variety of viewpoints within the firm and obtain a more complete picture of the interactions among different actors, this article presents quotes from managers, shop stewards, directly employed workers and agency workers. For data analysis, the material was first manually coded to identify the main overarching themes. Next, qualitative software was used to create more detailed labels for each theme. This process was repeated for each batch of newly gathered material allowing comparison with the material collected in the previous year and identification of the continuity and/or variations in the management’s ER, unionisation strategy and workforce composition. Since 2010, the University Research Group on Foxconn and the Hong Kong-based campaign group Students and Scholars Against Corporate Misbehaviour (SACOM) have gathered data through participant observations, questionnaires and interviews on Foxconn’s employment practices in China (Pun et al., 2014). We contrast the material gathered in Czechia with the existing data on Foxconn’s labour regime in China to examine the transfer of ER from the parent company to the Czech subsidiary.

Foxconn’s work and employment practices in Czechia

The transformation of work and employment practices in Foxconn’s Czech subsidiary is best understood when diachronically viewed. In this section, we outline three distinct phases, 2000–2004, 2005–2009 and 2010–2016, to illustrate the changes in Foxconn’s ER.

2000–2004: establishment phase

Foxconn entered Czechia in 2000 following the acquisition of Tesla in Pardubice, an important electronics manufacturer in socialist Czechoslovakia, producing TV sets, computers, tape recorders and radio-locators for the military. The national and local governments welcomed Foxconn’s investment because of its promise to create jobs and reduce local unemployment. Following the acquisition, Foxconn started with a workforce of 200 on full-time employment contracts. Some of these were former Tesla workers while the others were local unemployed Czechs recruited through the local employment office. The monthly wage was between 8000 and 10,000 Czech Koruna (Kč) (€225–€281); with overtime, the monthly wage could reach up to 12,000 Kč (€337). However, Foxconn experienced difficulties in recruiting the local workforce owing to the nature of assembly line production, which was until then unknown to the workers in the region, as well as the speed, boredom, irregularity of shifts and compulsory overtime. Expatriate managers made most of the decisions and tensions were rife:

Foxconn managers arrived and tried to implement the way they do things in China: before every shift, they lined us up, inspected us and then shouted commands. When on the assembly line, they did not let us take breaks. Our Czech manager did not want to put up with this: he stood up for the Czech workers and as a consequence got fired. (Czech directly employed assembly worker, Pardubice, February 2012)

Tensions escalated when workers signed two petitions against management and addressed them to the firm as well as at the Labour Office, Ministry of Labour and Social Affairs, and Mayor’s Office of Pardubice (Čaněk, 2016: 103).

Three major changes followed to resolve the labour shortage and conflict between management and labour. First, instead of recruiting local Czech workers on full-time contracts, management hired Slovak, Vietnamese and Mongolian workers on fixed-term contracts. The state facilitated the recruitment of migrant workers through the existing agreement on free movement for Slovak citizens and labour recruitment and educational schemes for Vietnamese and Mongolian nationals (Drbohlav, 2003: 197). Second, Foxconn changed its management structure and replaced the Chinese and Taiwanese senior and middle managers with Scots and Czechs. The Czech managers underwent training for 3 months at Foxconn’s Longhua plant in mainland China and were assigned at the middle-management positions on their return to Pardubice. Scottish managers were recruited after HP’s merger with Compaq in Erskine, Scotland, and took up executive management positions in the Czech subsidiaries. Third, the municipality of Pardubice mediated the industrial conflict, leading to Foxconn’s acceptance of the existing trade union and a new collective agreement.

2005–2009: steady expansion phase

In this phase, Foxconn enlarged its workforce from 2500 to 7200 workers. Contrary to the first phase, wherein all of the workers were directly employed, in the second phase, a proportion of the workers were directly employed while the rest (~3000–3500) were hired through Temporary Work Agencies (TWAs). The state enabled the recruitment of agency workers through the introduction of agency employment approved in 2004 in the context of European Union (EU) enlargement. Having initially recruited workers from the local labour market, the unwillingness of Czech workers to do repetitive work for low wages and the fluctuating production demand compelled Foxconn to enlarge its workforce by deploying agencies to recruit workers from Czechia’s traditional pool countries of labour migration: Slovakia, Poland, Mongolia, Vietnam and Ukraine:

Our employer [Foxconn] has an interest in employing foreign [agency] workers. A year or two ago, we wanted to employ Czechs, but they are not stable. They don’t like the job and they quit. The Labour Office was pushing us to employ Czechs. However, this is not how things work here. We are in a business where we might need 60 workers for manual positions at an hour’s notice. So we work with agencies. (Interview no. 13, HR manager, Pardubice, September 2012)

The shift in the size and composition of the workforce brought about significant changes to the ER regime. It entailed a radical change in recruitment practices because agency workers needed to be recruited (selected and tested) abroad, moved to the host country and housed. To accommodate their workers, the agencies had recourse to dormitories. The presence of agency workers also required a different form of management. Within the factory, agency workers were under a ‘dual control system’ (Gottfried, 1992); they were supervised by both Foxconn and the agency’s line managers, who were responsible for work coordination and control inside the plant. The wages were also differentiated: directly employed production workers earned 11,500–13,000 Kč (€407–€460), while indirectly employed staff earned 9000–11,000 Kč (€318–€390). The national trade union, Metalworkers Federation (KOVO), reacted to the introduction of agency employment by pressuring the government to limit migration. On the plant level, the union opted for a narrow unionisation strategy that excluded agency workers from union representation. As a consequence, the union’s collective bargaining was mainly directed towards wage increases for directly employed workers.

2010–2016: reconsideration phase

The economic crisis of 2008 resulted in a decrease and greater variability in orders in the electronics industry. Foxconn significantly reduced both the number of its directly (~2500) and indirectly employed workers (~1500–2000). Among the agency workers, the number of Mongolians, Ukrainians and Vietnamese significantly dropped. This was attributable to the fact that since 1 January 2012, the government had barred TWAs from hiring non-EU workers, who now needed to secure a direct contract with the employer. In response, Foxconn widened its recruitment boundaries and recruited, again through TWAs, workers from Romania and Bulgaria, including older workers well into their 50s and 60s. Importantly, Foxconn’s increased deployment of TWAs was also a strategy to deal with workers’ protests. In 2011, two protests took place. The first was in the former Apple division at the Kutná Hora plant, where workers threatened a strike and demanded the payment of bonuses and better working conditions. The second was at the Pardubice plant, where a number of Vietnamese workers put in place a work stoppage in response to a supervisor’s demands for ever higher productivity levels.

Now, in addition to the turnover of directly employed workers (25%–30%), subsidiaries had to manage the turnover of agency workers (40%–50%) who viewed the jobs at Foxconn as temporary:

I am neither a migrant nor a foreigner! Because this is the EU, I am first a citizen of the EU, then a Slovak and then from [Slovak town] … On the one hand, it is good to be mobile like this, as it provides new opportunities to meet new people and learn how they live, get a job and learn new skills. If I found a permanent job here, I would stay here as I don’t necessarily wish to return to Slovakia. I may stay here or go to Germany, who knows? (Interview no. 39, Slovakian agency worker, Pardubice, September 2012)

In the light of the heterogeneity of the workforce, Foxconn introduced different labour management strategies for workers employed directly and via agencies, in an effort to limit turnover and simultaneously guarantee flexibility. Agency workers were hired on an initial contract of 300 hours, established by collective agreement, which in periods of low production took 3 months for completion. During the first month, workers were not paid full wages but were given an advance of 1000 Kč (€40) per week to deter them from leaving. Directly employed workers were managed though an ‘hour-bank’, which required workers to work a total of 930 hours over 6 months. The hour-bank system permitted Foxconn to have higher flexibility and cut costs because all hours were paid at the same flat rate:

Those who work 12 hour shifts have to work 930 hours within 26 weeks. The people who had more hours, were sent home … or they could take a vacation. Law allows this. If one has more hours in the bank, they will be paid for these hours at the end of this balancing period. But if one does not have them because there was no work, they will have these hours paid anyway. If the employer doesn’t have work for that person and doesn’t dismiss him, the employer has to pay him every month as he was at work … This is what unions negotiated; average for the whole year is exactly 163.05 hours [a month], so we agreed that wage will be counted for workers based on this average. So, although one does 180 hours in one month and only 50 hours in another month, the employer pays him every month for 163.05 hours. (Interview no. 56, Czech direct worker, Pardubice, September 2012)

In its continuous attempt to reduce labour costs and turnover, the subsidiary engaged in a process of substituting some of the agency workers with Mongolian workers recruited directly in their country of origin. After having recruited about 170 workers in 2016, its management set the overall recruitment target for 2017 to 1000 workers:

The arrival of Mongolian workers improved the fluctuation of production because before this, the retention of workers who stayed over three months was 37%, which was really bad. The fluctuation of workers is very important for the company as those who pass the probation period of three months get the contract. The Mongolian workers are also good workers as they arrive without their families, and do not have their family, children, friends and gardens to keep them occupied. They are ideal workers as they can fully commit to work. There is now the question of what will happen after they have been here for 6 months, as then they can legally bring their children over here. So far, a proportion of those workers tended to return to Mongolia. They just disappeared. (Interview no. 69, Czech shop steward, March 2017)

From 2012 management’s position towards the union changed. Managers’ interactions with shop stewards became more frequent, beginning the process of strengthening bargaining that led to an increase in the wages of directly employed assembly workers. In 2016, a core operator could earn 16,000 Kč net (€600); a group leader about 20,000 Kč (€740); and a supervisor about 24,000–25,000 Kč (€890–€925). At the same time, the widened recruitment practices changed the ratio of directly employed and agency workers to 60:40. This ratio was reversed during periods of high production, with agency workers making up majority of the workforce (Čaněk, 2016). Despite these changes, the plant-based union continued to bargain solely for the directly employed workers – over wage increase, shortening of the 12-hour shift and benefits packages.

Revisiting the transfer debate

In this section, we discuss the shortcomings of the latecomer model and identify the limitations posed by the national frame of analysis. We discuss the relevance of the multi-scalar institutional environment in the host country, sectoral specificity and labour agency for understanding both the firm and trade union behaviours.

Legacy and innovation in employment arrangements

The flexible work regime is typical of Foxconn’s plants in both China and Czechia. In Czechia, the flexible work regime is attained through agency workers. On the contrary, in China, Foxconn hires rural migrant workers with fixed short-term contracts (1-year and 3-year) and employee-students from vocational schools, who make up 15% of the total workforce during peak production (Smith and Chan, 2015). In China, workers depend on extensive overtime and personal performance bonuses to attain a large proportion of their wage because the base wage provides only 50% of their regular income (Lüthje and Butollo, 2016). In Czechia, the flexibility of the workforce is both internal and external. External flexibility is achieved through the use of agency workers, and internal flexibility is achieved using the hour-bank system for directly employed workers and decreasing the wage security of agency workers, whose contracts do not specify a fixed number of weekly working hours.

Hence, while the flexible work regime is typical of Foxconn’s operations in both countries, the practices through which the regime is implemented in Czechia are determined by the host country’s institutional environment. Since the 1990s, CEE governments have not only put in place what Drahokoupil (2008) calls ‘investment-promotion machines’ (p. 207) to foster competition through pro-growth coalitions composed of national, regional and local state actors but also made use of the EU directives, following the 2004 EU enlargement, to make the labour market flexible (Meardi, 2012). Such an institutional and regulatory framework in the host country enabled Foxconn to achieve an extremely high level of labour flexibility.

Flexible work and employment arrangements pivot on collective dormitories. In both China and Czechia, Foxconn relies on dormitories to house and control migrant workers. In China, as millions of migrant workers arrived in industrial towns and cities, the provision of collective dormitories for these workers became a necessity for firms. The arrangement of dormitories allows the fast and flexible organisation of production as well as compulsory overtime (Chan et al., 2013). At the plants in China, men and women live in separate areas and are not allowed to visit one another’s rooms; workers are barred from cooking in the name of fire safety and from washing clothes so as to save water and energy through the factory cleaning service. Dormitories prohibit family life, and workers are assigned a single bed in a room shared with other workers. In Czechia, living spaces are shared between men and women, with cooking facilities and showers within rooms, or shared kitchenettes and showers on every floor. Control over workers lies with agencies and private labour supply firms and is less obsessive than in China. Importantly, the Czech subsidiary made use of dormitories to reduce turnover and extract additional work from agency workers. During periods of low production in particular, dormitories allowed Foxconn to avail itself of a stationary workforce pool by standing workers down for several days without risking a shortage of labour.

Dormitories are therefore pivotal to Foxconn’s flexible organisation of production: they allow the firm to facilitate just-in-time, high-speed production by compressing the workplace and living space, lengthening the working day, and extending management’s control over labour outside the workplace. The systemic use of dormitories is not distinctive of today’s China. It has a long history in the UK, South Africa, Latin America and Asia, where dormitories were used during industrialisation (Smith, 2003).

Similarly, in former Czechoslovakia, the state instituted dormitories with the aim of concentrating industry in the major urban centres but keeping a large proportion of the population in rural areas (Fuchs and Demko, 1978: 178). Hence, as with flexible labour use, the reliance on dormitories in Czech plants should not be seen simply as Foxconn applying management practices from China to its factories abroad. Labour dormitories are best viewed as a systemic practice of capital to manage and control migrant labour. The difference today, compared with previous historical processes of industrialisation, is that dormitories are no longer confined to male workers and are not geared towards securing a workforce for long-term employment. Today’s dormitories accommodate single male and female workers for short-term employment. This form of ER, defined by Pun and Smith (2007) as the ‘dormitory labour regime’, reorganises work and residence for global capital use and is pivotal for delivering flexible and high-speed production within globally integrated circuits of production.

Such a high level of labour flexibility is intrinsically linked to the ‘just in time’ (JIT) production model dominant in the electronics industry. The electronics industry is characterised by vertical disintegration between product development and manufacturing. This gives leading firms such as Apple, HP, Cisco, Dell, HTC, LG and Sony power over subcontractors such as Foxconn because manufacturers need to compete for contracts from leading firms. This supply chain facilitates the polarisation of profit opportunities, and results in low margins with intense pressure on costs for contract manufacturers, particularly labour costs (Chan et al., 2013: 105). In Czechia, core operators earn €520–€600 and agency workers earn €450–€520 per month. This is slightly higher than the national minimum wage (€407 in 2017) but well below the general Czech median wage of €880 (ISPV, 2016). In China, wages are linked to the regions: in Chongqing an operator earns €300–€400 per month including overtime, which is above the minimum wage (€235 or 1800 RMB in 2017), but in Shenzhen, wages can be up to €500, almost equal to agency workers’ wages in Czechia.

The problem of low wages, shared by both countries, arises owing to the state’s intervention wherein differential regional wages are set in China and temporary agency workers are supported in Czechia. The problem is also attributable to vertical disintegration and specialisation in the electronics supply chain. Hence, low wages are embedded in the industry’s uneven distribution of power and profit. In other words, in the electronics industry, sectoral specificity outweighs national differences.

As we can observe, contrary to what scholars of Chinese MNCs in Europe have suggested so far, the parent company under analysis did not use the subsidiaries so much as a learning channel, but instead transferred several of its home employment practices to the subsidiary. In addition, rather than the home firm being provided with training about host country practices (Shen and Edwards, 2004; Zhang and Edwards, 2007), local managers underwent a month-long training period in mainland China to learn about the parent firm’s production and work organisation practices.

Given the limitations of the latecomer model in explaining Foxconn’s adherence to the ER of the parent MNC, we find quite pertinent the argument that adoption and diffusion of management practices are influenced by a range of factors including sectoral specificity (Sayim, 2011). Sectoral characteristics and economies of scale might give latecomer firms a comparative advantage in specific industries. The identified discrepancies suggest a variation of transfer modalities among latecomer MNCs and raise the question of whether, with regard to MNCs from China, there is indeed a single national model that has transnational applicability (Smith and Zheng, 2016). This analysis shows that Foxconn’s expansion to Czechia is an example of what Yeung (2004) calls a hybrid form of economic organisation, in that it reveals the influence of GPNs and reflects supranational flows of capital and people.

Labour-driven changes in employment and IRs

Foxconn’s workforce has changed over time, from a majority of directly employed Czech workers (and some Slovaks, Mongolians and Vietnamese) to a more heterogeneous workforce including EU migrant workers from Slovakia, Romania, Poland and Bulgaria indirectly employed by TWAs. The change in recruitment strategies was aimed at reducing the labour costs and contractual responsibilities of direct employment. In addition, it was a response to collective worker actions, labour shortages and high worker turnover rates. However, although relying on TWAs might have externalised the risks of recruiting and managing a large temporary workforce pool, the problems of turnover were exacerbated. Similarly, in China, high labour turnover is one of the main problems confronted by Foxconn because workers quickly move from one factory-cum-dormitory to another (Pun and Smith, 2007).

Scholars continue to regard exit as inferior to voice because it arises from individual acts rather than collective mobilisation (Smith, 2006). However, the exit of migrant workers is engendered by shared ‘migrant capital’ (Ryan et al., 2015), namely, a knowledge of the labour market, experiences of migration and social networks. The EU agency workers are a ‘hypermobile’ workforce (Berntsen and Lillie, 2016) that weighs its labour opportunities within the EU-wide labour market owing to a regulated freedom of movement (Andrijasevic and Sacchetto, 2016). In China, the relaxing of the hukou regime (household registration system) permitted migrant workers to move across different plants or leave the industry for the service sector. In other words, the migrant capital that workers possess leads them to view their jobs as temporary and develop exit strategies.

While some scholars interpret the segmentation of the workforce as an example of the firm applying structural conditions under which workers’ bargaining power is reduced, we suggest that workers may accumulate ‘labour mobility power’ (Smith, 2006). The fact that workers in the EU as well as in China weight their opportunities in a large labour market puts these workers in a stronger bargaining position against an employer that uses labour market manipulation with the aim of recreating dependency. Therefore, as studies of GPNs have shown, a direct link exists between the demands of the lead firms in the global electronics supply chain and the methods of production and labour control (Chan et al., 2013). In other words, the heterogeneity and mobility of labour have shaped the subsidiaries’ ER practices. The hour-bank system and postponement of wages and dormitories in Czechia are forms of control developed by the subsidiary at different stages of internationalisation in response to the heterogeneity and mobility of the workforce. These management practices aim to curtail workers’ exit and re-create dependency, so as to enable the subsidiary to cope with conditions of uncertainty and fluctuating demand in a highly competitive buyer-driven value chain. Building on the approach of GPN scholars, our case study contributes to correcting the ‘labour deficit’ (Newsome et al., 2015) construct in the internationalisation literature and shows the importance of the ‘agential’ impact of migrant labour on global firms and in the internationalisation of production.

The mobility of labour brings challenges for trade unions. The electronics sector is characterised by managerial hostility towards trade unions. However, in Czechia, while the trade unions are weak, they are not one and the same with the management as in mainland China, where the commonality of interests between management and unions severely undermines the capacity of enterprise unions to represent the workers (Pringle, 2013). Foxconn in Czechia did not enforce an aggressive anti-union strategy. When Foxconn acquired the Tesla plant, it inherited the trade union that was active within the plant. After initially ignoring the union, Foxconn put in place a more careful cooperation strategy that resulted in several collective bargaining agreements with the union. This might be due to the weak trade unions in Czechia and the declining union density in the post-1989 period that characterised all CEE countries (Żuk and Żuk, 2018).

We suggest, however, that the explanation to why management accepts organised labour is not to be found as per traditional views of internationalisation, by solely scrutinising the constraints imposed by the national regulatory framework (Jürgens and Krzywdzinski, 2009; Zhu et al., 2014). The subsidiary’s management tolerates the union and engages in collective bargaining because the union follows a narrow unionisation strategy and does not represent agency workers. Directly employed workers benefit from the presence of migrant agency workers as the latter absorb the impact of the fluctuating demand for labour. In doing so, the union favours directly employed domestic workers over migrant agency workers and upholds the management’s workplace segmentation.

This overlap between the interests of the management and union has benefitted Foxconn as well as the directly employed workers who, in contrast with the agency workers, receive a vast range of benefits such as training and development opportunities, leisure activities, subsidised catering, travel allowances, discounted mobile phone tariffs and language courses for themselves and their families (Foxconn Technology Group, 2013). We suggest that the prevailing views explaining the varieties of IRs (Baccaro and Howell, 2017) overemphasise the relationship between national regulation and trade unions and do not take account of the relationship between the trade union and workforce.

Considering the role of migrant workers in the analysis of IRs allows us to understand the interdependencies and interconnections that arise from the mobility of capital and labour. In addition, identifying the heterogeneity of labour relations within the firm is important because it allows us to understand how MNCs adjust their ER to respond to the highly mobile workforce. Furthermore, it enables examination of the implications for organised labour when MNCs’ employment practices are underpinned by the conditions of a transnational labour market.

Conclusion

In this article, we show that the plants in China and Czechia share several similarities in terms of work and employment practices, such as flexible labour arrangements, low wages, use of dormitories and weak trade union representation. These similarities are not simply an outcome of the importation of a regime from mainland China, but rather are a result of several overlapping factors such as institutional environment in the host country, the sectoral specificity of the electronics industry and the social composition of the workforce.

The relevance of our findings is two-fold. First, the article contributes to the emerging scholarship on ER in Chinese firms expanding to Europe. While we acknowledge the limits of generalising findings from a single case study in one sector, our analysis suggests that the ‘latecomers’ model is inadequate to explain the variation of ER in Chinese MNCs overseas. This model certainly exploits the stark social divide between urban and rural citizens in China and between nationals and EU migrants in Czechia. However, we suggest that it is important not to confuse the application of practices by latecomers with utilisation of practices already present within the host country. By showing that the latter are an outcome of multi-scalar institutional arrangements, we suggest that national distinctiveness is diminishing. Similarly, with regard to Chinese MNCs as latecomers, our results challenge the national business distinctiveness thesis and raise the question of whether a single national model having transnational applicability indeed exists.

Second, the article makes a theoretical contribution to internationalisation debates by challenging the established viewpoint on labour. We show that the ER of MNCs respond to and develop in relation to workers’ agency as well as to the specific composition of the workforce. The conceptual relevance of a close examination of the composition of the workforce lies in the fact that both the MNCs’ ER and trade union strategies are influenced by this composition. We challenge the conventional view that labour actively partakes in the process of internationalisation only when it exhibits institutionalised forms of organised agency. Rather, our analysis of the shifting composition of the workforce and mobility of labour demonstrates the relevance of non-institutionalised forms of worker agency for the theorisation of internationalisation. Further research is required to explore the tensions between the spatial embeddedness of trade unions and the mobility of labour, and to interrogate trade unions’ tendency to defend ‘place rather than class’ (Herod, 2012: 343) in response to the increasing heterogeneity of labour.