Migrant workers trapped between individualism and collectivism: The formation of union-based workplace collectivism

Abstract

Transnational labour migration challenges collectivism as well as migrant workers’ labour market rights, due to employers’ strategies such as segregating workers, and the migrant workers’ individualistic strategies. This article, arguing that there are no intrinsic impediments to creating (instrumental) collectivist solutions encompassing both migrants and host-country workers, develops a dynamic, conceptual framework of four preconditions – workers’ closeness, feeling of unity, shared problem perception and reference groups – which are necessary for migrant workers to develop collective labour market strategies. The article then utilises the framework in three empirical cases to illustrate how the framework, combined with union strategies, helps explain the different degrees of migrant workers’ labour market inclusion and to help understand why gaps between migrant workers and host-country workers may form.

Keywords

Collective interests collectivism instrumentalism migrant workers unions

Introduction

Transnational European migrant workers face severe issues of precarious and even exploitative working conditions and exclusion from labour market regulation and workplace representation. National institutions are often stalemated in complex transnational employment configurations with multiple actors and fast-shifting settings, where some employers actively seek to exploit various loopholes (such as posting of workers) to circumvent national regulations when employing low-wage migrant workers (Arnholtz and Lillie, 2020; Berntsen and Lillie, 2016; Heindlmaier and Blauberger, 2017; Lillie, 2016; Lillie and Greer, 2007). Additionally, migrant workers are typically concentrated in certain, highly exposed sectors, are highly dependent on their employers, and are furthermore often segregated from native¹ workers and other ethnic groups of workers at the workplace (Holgate, 2005; Milkman, 2006). Many migrant workers furthermore identify with workers and working conditions in their home-country rather than with the host-country conditions, wherefore they are more inclined to accept inferior wages and working conditions (cf. reference group theory; Merton, 1968 [1957]; Piore, 1979). These structural factors (including the type of work contract and citizenship status) and employer-driven factors are reinforced by migrants’ low incentive to engage with local unions (Berntsen, 2016), or even by feelings of resignation over their position in the local labour market (Lever and Milbourne, 2017). Altogether, this means that migrant workers often rely on individualistic labour market strategies, which also influences their attitudes towards unions (Alberti, 2014; Berntsen, 2016; Caro et al., 2015; Kall et al., 2018).

Migrant workers often have an instrumental approach to labour market organisation; however workers’ instrumental orientation is not a new phenomenon in labour history, and by no means unique to migrant workers. While some workers have egalitarian and altruistic values that lend unconditional support to union-based collectivism, most workers in general take an instrumental approach (Crouch, 1982; Lockwood, 1966; Waddington and Whitston, 1997; Webb and Webb, 1897). Hence migrant workers’ instrumental attitudes are neither new nor highly divergent from those of native workers. But out of workers’ instrumental approach can materialise collectivistic or individualistic strategies. Many Nordic workers share an instrumental approach to work organisation and unions, as opposed to an ideological approach (Caraker et al., 2014; Kevätsalo, 1995; Rantala, 2013), but due to societal norms and workplace social customs, societal efficiency and the legitimacy created by the outcome (Andersen et al., 2014), unions and worker collectivism remain imperative to most Nordic workers’ labour market strategy (Hasle and Sørensen, 2013). Collectivism thus serves the economic interest of the workers, and is not merely an ideological or altruistic phenomenon, and much collective action can hence be considered ‘instrumental collectivism’ as opposed to ‘ideological collectivism’ (Crouch, 1982; Fox, 1985; Hyman, 1992; Salmela, 2011). Most intra-European migrant workers however have little or no understanding of collectivism’s potentially positive impact – their potential gain from collective action vis-a-vis employers – or they find themselves in a situation where the connection is less straightforward (Berntsen, 2016; Kall et al., 2018).

Nordic unions have traditionally been very powerful and remain comparatively so (Andersen et al., 2014); the Nordic countries therefore serve as a good setting for investigating how collectivism forms among migrant workers when influenced by strong unions. Institutions that reinforce the unions’ societal position remain highly embedded in the Nordic societies and rooted in national compromises between labour and capital that are being constantly reconstructed (Kettunen, 2012). Their strong position makes the unions relatively efficient at safeguarding working conditions – e.g. by enforcing inclusive institutions such as collective agreements (Andersen et al., 2014) – particularly in comparison with the deteriorating position of unions in most other countries (Gumbrell-McCormick and Hyman, 2013).

Nonetheless, creating and maintaining strong worker collectivism that includes migrant workers is challenging for unions and workers, both at the sector and workplace level (Caro et al., 2015). By comparing three case studies from Denmark and Finland, we illustrate how the theoretical framework we develop of workers’ closeness, unity, shared problem perception and reference groups, in combination with union strategies, can help explain the inclusion (or exclusion) of migrant workers through workplace collectivism. We thus seek to answer the following question: Given the apparent contradiction between traditional collectivism (unionism) and the typical individualistic orientation of migrant workers, under what conditions can worker collectivism form among migrant workers? Secondly, we discuss how unions can actively seek to overcome this contradiction.

The article is structured as follows. First, we elucidate migrants’ labour market strategies and the employers’ strategies towards migrants, with a particular focus on employer-driven ethnic separation. Then we provide our theoretical framework based on four fundamental preconditions for worker collectivism among migrant workers. Next we discuss the unions’ role in and contribution to enhancing migrant collectivism. The data, method and analysis are then explained in the ‘research setting’ section, before the case studies are introduced. An analysis of the three cases based on the framework follows to illustrate the usefulness of the framework in explaining the presence or absence of collectivism. Finally the overall discussion and conclusion engages with the existing literature to move forward our overall knowledge on migrant workers, collectivism and unionism.

Migrant workers’ individual strategies and employers’ segregation strategies

The act of labour migration is conditioned by structural and social factors (Piore, 1979), employer behaviour (MacKenzie and Forde, 2009), migrant networks and kinship (Ciupijus et al., 2020; Piore, 1979), and the interplay between different regulatory spaces and actors (MacKenzie and Martinez Lucio, 2019). Nevertheless, migrant workers are also active labour market actors seeking to improve their life and living conditions (Andrijasevic and Sacchetto, 2016; Berntsen, 2016), so the migration decision remains largely individual (or family-based), which may influence migrants’ labour market behaviour (cf. Stark, 1991). We therefore discuss how various structural, institutional and social factors shape (and reinforce) the individualism of many migrant workers. For some migrant workers employment in the receiving countries is conditioned on their acceptance of the breaching of labour market regulations (Arnholtz and Refslund, 2019), or at least on accepting below-average standards (Castles and Kosack, 1973; Piore, 1979; Waldinger and Lichter, 2003), meaning that migrants may even be hostile towards union activity (Berntsen, 2016; Lillie, 2016). Posting of workers plays furthermore a critical role in the European migrant workers’ case with strong implications for the workers’ (and employers’) strategy. The posting regimes are often designed to create dual labour market structures and reduce the workers’ collective action (Arnholtz and Lillie, 2020). Employers may further hamper collectivistic approaches through workforce segregation, give little or no information to the migrants on their rights, and constantly change the workforce composition for instance via posting. This means migrant workers find themselves in a weaker labour market position, which is further compounded by their tendency to see their home-country working conditions and workers – instead of those of the host country – as their natural ‘reference group’, which Piore (1979) refers to as ‘dual frames of reference’. Overall, this makes migrant workers much more predisposed to accept inferior working conditions than are host-country workers.

As the title of Piore’s classic work Birds of Passage implies, he emphasised explicitly the transitional dimension of labour migration (Piore, 1979: 3); but migration in our case is not only transitional, nor is it a linear process (as Piore also argued; see also Andrijasevic and Sacchetto, 2016). Piore further stated that the longer the intended stay the more likely the migrants will change their reference group to that of the host country, something later research has also shown both in terms of reference groups and industrial relations approaches including unionism (Cam, 2014; Krings, 2014; Quinlan and Lever-Tracy, 1988; Refslund, 2018; Stark, 1991). Hence it is important to acknowledge the different types of migrant workers and their status. Engbersen et al. (2013) suggest a labour migrant typology of ‘footloose’, circular, bi-national and settler migrants, which affects the migrant workers’ integration and interaction with national industrial relations systems. Such typology has likewise implications for the theoretical framework we propose here.

The overwhelming majority of the migrants in our case studies originate from Eastern European post-communist countries, which implies a particular pattern of ‘learned’ individualist behaviour (Sippola, 2013), and the changes following the breakdown of the communist regimes strengthened the individual’s responsibility for coping with the risks and contradictions caused by the social and economic restructuring (Kalmus and Vihalemm, 2006). The Eastern European workers’ weak incentive towards collectivistic strategies and union activity is further worsened by their often negative – or total lack of – experiences with unions, and a cultural and historical background of negative perceptions of unions (Crowley, 2004).

When migrant workers are hired in sectors like construction and agriculture, the work processes are often structured to promote the atomisation and separation of workers, typically by ethnicity (Caro et al., 2015; Haakestad and Friberg, 2020; Lillie and Sippola, 2011). Ethnic divisions are often further fostered by employers to reinforce the labour market segregation through e.g. job competition staged between different migrant groups or by interchanging them to maintain a regime of cheap labour (Lever and Milbourne, 2017; MacKenzie and Forde, 2009; Piore, 1979). This tends to amplify ethnic differences over cross-group traits, which restricts collectivism across migrant groups (Caro et al., 2015; Castles and Kosack, 1973: 480; Piore, 1979), with negative implications for the migrant workers’ perception of collectivism (Andrijasevic and Sacchetto, 2016).

In light of the structural factors and the migrant workers’ background in our study, the often-found individualistic strategies and instrumental approach are not surprising. Our core argument however is not that migrant workers will always have individualistic strategies, but that it depends on the preconditions identified in our framework as well as on the context. In different settings migrant workers may even be highly inclined towards collective approaches, as for instance Ruth Milkman’s classic study (2006) of mainly Hispanic migrant workers and unionism in Los Angeles precisely illustrates.

Workplace collectivism and migrant workers – four preconditions

We argue collectivism and class consciousness are not omnipresent objective phenomena, but rather the dynamic and contingent outcome of processes and social interaction embedded in the social context (cf. Fantasia, 1988; Hyman, 1999; Simms, 2012; Thompson, 1968). While there are varying definitions of worker collectivism in the literature, we define it here simply as when the (majority of) workers define their labour market interests or the means to achieving them in collective terms as opposed to individual terms. We do not hence equate collectivism with unionism although these are often closely aligned in the context we study; nevertheless these two might exist independently or even in contradiction, as originally pointed out by Lysgaard (1967 [1961]). Nor do we equate collectivism with collective action per se, though collectivism (at least momentarily) is a necessary condition for collective action. The workplace, its organisation and not least the social interaction taking place there are important for the formation of collectivism, since this is grounded in real-life experiences rather than normative viewpoints (Fantasia, 1988; Hyman, 1999; Lysgaard, 1967 [1961]; Thompson, 1968). Working-class consciousness has additionally been declining for decades (Hyman, 1999; Simms, 2012) and the importance of other identities – beyond that of the industrial worker – is growing (Peetz, 2010), which is particularly important for migrant workers (Alberti et al., 2013; Holgate, 2005).

In the analysis of the formation of workplace collectivism we build a theoretical framework with four preconditions for collectivism among migrant workers. The first three preconditions draw upon Norwegian sociologist Sverre Lysgaard’s classic analysis (1967 [1961]) of the ‘workers’ collective’.² Lysgaard’s work, although not well known outside Scandinavia, has been highly influential in the Scandinavian sociology of work and working life studies (Hasle and Sørensen, 2013; Skorstad and Karlsson, 2020). The three preconditions are: workers’ closeness to co-workers; workers’ perception of unity; and workers’ concession to opposing interests vis-a-vis the employer (Lysgaard, 1967 [1961]; Skorstad and Karlsson, 2020). We supplement this with a fourth precondition of specific importance to migrant workers: the lack of a unified frame of reference as implied by Piore’s (1979) dual frame of reference. An important aspect of human behaviour is the recognition of reference groups (Merton, 1968 [1957]), which features prominently in many studies on labour migration (Piore, 1979; Waldinger and Lichter, 2003). Migrant workers often see home-country workers rather than host-country workers as their reference group, and hence accept lower conditions – often in a secondary labour market. This leads to dual labour market structures and thus becomes a barrier to collectivism, since migrant workers have higher earnings than possible in the home country, which reduces their incentive to engage collectively.

Turning to the three preconditions derived from Lysgaard, he argues that ‘workers’ collectives’ are created through social relations and interactions at the workplace level, and are organised informally and mainly with a defensive character to safeguard the workers against the ‘techno-economic’ system (the company, its goal – mainly increased profit – and the resulting work organisation). He further stresses the formative aspects of ‘workers’ collectives’; however, they are not formed automatically (Lysgaard, 1967 [1961]: 143–149), and the three conditions highlighted must be met.

Firstly, the workers need to acknowledge a shared problem vis-a-vis the techno-economic system, e.g. pressure from the company. Many migrant workers are dependent on the employer for continuing the work relation, and often also staying in the country, which makes them less prone to act upon disagreement with the employers. Additionally the higher wages migrant workers can earn in the host country may reduce the recognition of their contrary interests vis-a-vis the employers. Secondly, the workers need to interact for ‘workers’ collectives’ to thrive, since collectivism is based on social interactions, e.g. the workers can share concerns and information on wages and working conditions. Where company size and hence closeness among workers is restricted, e.g. at fragmented, small or remote worksites, creating collectivism is much more challenging because of the lacking interaction (Fox, 1971). Thirdly, the workers need to have a shared feeling of unity (Lysgaard, 1967 [1961]: 146–147). In Lysgaard’s original work, unity includes, among other things, not to have ‘managerial aspirations’ and hence to take pride in their work and the ‘workers’ collective’, and unity thus has some resemblance with ‘working class consciousness’, or with what other scholars more generally discuss as ‘social identity’ (e.g. Cregan et al., 2009). However, for migrant workers to be fully included in national industrial relations systems based on unionism there also needs to be some cross-group unity, which is by far the most difficult to establish. Often ethnic divides amplify the lack of cross-(ethnic)-group unity, since the potential unity is typically more related to kinship or ethnic groups for the migrant workers rather than abstract categories of workers more generally.

These three conditions of problem-perception, unity and closeness of workers are often hampered among migrant workers by ethnic separation and employer behaviour. Caraker (2011) further argues that the more flexible the mode of production, the more difficult it becomes to organise ‘workers’ collectives’, which closely follows Lysgaard’s original claim. The workplace size also strengthens the impact of social customs, and workers on larger production sites are more likely to have adversarial attitudes towards employers (Fox, 1971), which strengthens collectivism.

Importantly, we do not ascribe any mechanistic causality to the framework, but rather argue that these necessary preconditions must to some degree be met before collectivism among migrant workers can arise. Nevertheless, other particularistic factors may also influence collectivism, such as the role of leadership (Cregan et al., 2009; Fantasia, 1988; Kelly, 1998) or salient events like abrupt changes in wages or working conditions, strikes (Fantasia, 1988) or what Kelly (1998) in general terms ‘injustice’, which is the key trigger in his mobilisation theory. Nonetheless, for collectivism to result in labour market inclusion in our high union-density setting, it must be aligned with union strategies, so we therefore now turn to the unions’ role.

Unions, migrant workers and worker collectivism

Although the unions’ position has weakened in many European countries (Gumbrell-McCormick and Hyman, 2013), they remain in most countries the main foundation of worker inclusion, and therefore also migrant worker inclusion (Doellgast et al., 2018; Kall et al., 2018; Marino et al., 2015; Pulignano et al., 2015; Refslund, 2018). Strong labour and union representation is often a necessary condition for creating more inclusive settings for migrant workers, but even in countries with long traditions of unionism, unions per se may not be sufficient, as they have not been able to accommodate migrant workers (Holgate, 2005). Moreover, migrant workers have lower unionisation rates when compared with native workers (Cam, 2014; Gorodzeisky and Richards, 2013).³ Collectivism may additionally emerge from sources other than unions, such as civic communities (Holgate, 2015; McBride and Lucio, 2011); however this is often found where unions have withered away or were never present.

Unions have in the past been blamed for excluding migrant workers or even reinforcing labour market dualities, which in some cases has been based arguably on unions’ xenophobic attitudes towards migrant workers, partly due to native workers’ hostility towards migrants (Castles and Kosack, 1973; Penninx and Roosblad, 2000). However, later accounts illustrate how unions often have shifted their policy towards including and organising rather than excluding migrant workers (arguably also after learning that migration could not be reversed) (Adler et al., 2014; Haus, 2002).

Migrant vulnerability is reinforced by the intersectionality of various dimensions such as gender, age, ethnicity and class (Alberti et al., 2013), and unions must therefore adapt to a more diverse workforce than previously, and not simply apply a traditional unionisation approach, but instead apply elements of activist-led organising strategies or ‘social movement unionism’ (Cam, 2014; Holgate, 2005) – something that has proven very challenging for unions in many countries.

Overall unions, in particular the Nordic unions in our study, remain essential in bringing migrant workers into the industrial relations system. The unions may facilitate the creation of collectivism through fostering a feeling of unity and workers’ voice, creating room for dialogue between different workers by pointing out the shared issues they face, even where contact is restricted by employers’ segregation of workers. In strongly institutionalised settings the unions can further utilise formal voice mechanisms such as shop steward systems and collective bargaining to promote collectivism. This implies that collectivist behaviour of workers is enhanced by union presence at the workplace level and by union strategies, particularly in the Nordic countries we study, where well-established local union representation remains the norm.

Methodology and research setting

After developing the conceptual framework we apply a cross-national approach with three empirical in-depth case studies conducted in Finland and Denmark (one in Finland and two in Denmark) to illustrate how our framework can explain the presence or absence of collectivism among migrant workers. The comparative design enables us to take into account the national and institutional variation, while including different industries (slaughterhouses and construction) and provide different contexts for applying our conceptual framework. We advocate the notion of ‘slow comparativism’ in research on work and industrial relations as suggested by Almond and Connolly (2020), which allows for an alternative research approach to faster but ‘thinner’ comparisons that are often little contextualised. We have long-term research interests in the industries and sites being studied in both countries, allowing us to ‘immerse’ in the case studies and gain in-depth, ethnographic knowledge beyond the face value of, for instance, the interviews conducted (Almond and Connolly, 2020). Thus we make contextualised comparisons of the three cases reflecting on our conceptual framework, highlighting the conceptual and thematic ‘unity across countries’. Doing so we are still producing in-depth knowledge of each case and the context, and assessing this in-depth knowledge as significant rather than the comparison itself (Almond and Connolly, 2020: 67).

Our case studies (see Table 1) involved collecting multiple forms of data, with the main data in all three cases being semi-structured interviews, which were supplemented by fieldwork notes based on observations, two media surveys and other available sources, which all enabled us to deepen and broaden our understanding of migrant workers, collectivism and unions. In the Finnish case the data were collected between 2006 and 2014 in an Academy of Finland-funded project concerning labour relations at the Olkiluoto 3 nuclear power plant construction in 2006–2008 (including two follow-up interviews in 2014). There were 25 interviewees altogether, of which 10 were union officials, five civil servants, five site workers, three representatives of the main contractors and two subcontractors. Additionally, 10 interviews were conducted with representatives of trade union confederations and a representative of the European Migrant Workers’ Union (EMWU). These data were complemented by a media survey of five Finnish newspapers from 2005 to 2008. The Danish data were collected between 2011 and 2017. In the slaughterhouse case, 17 interviews were conducted with unionists (at national and local level), company management and employers’ organisations representatives. In the Copenhagen metro case, 18 interviews with 22 interviewees were conducted with union organisers and officials, labour inspectors, employer association representatives, representatives from the Metro Company and management in different companies on the project. In the metro case the results were supplemented by a media survey.

Table 1.

Overview of the case studies.

	Size of the site	Number of migrant workers	Work organisation	Trade union site access
Copenhagen metro project	Fluctuating with a total of 4–5000 workers involved (peaking with up to 3000 in 2015/16)	Fluctuating, but above 80% – peaking with 2500+ migrant workers in 2015/16	Fragmented along the production chain, large share of posted staffing company workers	Full, but somewhat limited initially
Danish slaughterhouses^a	Varying from 100 to >1000	Half of the workforce in many sites	Tayloristic	Full and integrated
Olkiluoto 3 construction site	Fluctuating and peaking at 4500 in 2009	2/3 of the workers	Hierarchically organised contracting chains, mostly based on posting of workers	Limited (more pronounced in the beginning)

For more in-depth information on the sector see Navrbjerg and Krogh (2018), Steffensen and Ebsen (2017), Wagner and Refslund (2016).

Via these data, the preconditions for migrants’ collective action and union strategy were traced. While the case studies were not conducted specifically for cross-national comparative studies, they share many features, which makes the comparison pertinent. The commonalties of the studies include: tracing preconditions for collectivism among migrant workers; case-study method; and data mainly collected via interviews. Although there are some time differences in the studies’ data collection, we do not expect the preconditions for collectivism to change so swiftly that it undermines our results (cf. Almond and Connolly, 2020). Nevertheless the time dimension may partly explain why the Danish unions were more successful than their Finnish counterparts, as we examined these at a later stage, where they had had more time to develop their strategy.

An advantage of our design is that the countries share commonalties in the industrial relations regime, with both Denmark and Finland being part of the ‘Nordic industrial relations family’ (Andersen et al., 2014), and sharing relatively recent experiences in receiving migrant workers via the EU’s free movement of labour. Another common feature in the Nordic countries, including Finland and Denmark, is the particular notion of social citizenship based on the idea of a virtuous circle between the divergent interests of ‘labour market parties’ as well as some established symmetry between workers and employers (Kettunen, 2012). Both countries additionally have had a relatively homogeneous workforce in terms of ethnicity, but the accession of the new EU member states in 2004 caused some divergent tendencies. Denmark adopted a transitional regime not restricting incoming labour migration while demanding wages and labour conditions for migrants to resemble those in the collective agreements, while Finland imposed a restricted access to the country for EU8 (accessed member states) citizens for two years (Dølvik and Eldring, 2006). Further differences that add nuance to our overall ‘most-similar design’ is that there is less labour market legislation in Denmark, where worker rights mainly stem from collective agreements. Collective agreements are also important in Finland, but are supplemented by legislation and occasionally corporatist agreements on wage development, and the legal extension of the collective agreements, which never occurs in Denmark.

Our cases are in highly male-dominated blue-collar industries, and our findings are therefore characteristic of male workers. The results most likely would be different if women migrant workers were included in the study. The male-dominant workplaces may contribute to the nature of social relations at the workplace, as masculinity can provide the basis for (collective) identity (Ness, 2012). Almost all the migrant workers in our case studies are intra-European migrants who have special characteristics and are highly influenced by social and historical contexts, which may limit the generalisability to other migrant groups. Still, by and large we suggest that the four preconditions in our framework could be applied as an analytical framework to worker collectivism in other migration contexts as well.

Contextualising the case studies

Before discussing the framework in relation to the three case studies, we provide an overview of the empirical findings and the setting of the three cases including union strategies, which potentially can further facilitate worker collectivism. (See Table 2 for an overview.)

Table 2.

Ethnic separation strategies, collectivism and union strategies in the different cases.

	Ethnic separation strategies	Worker collectivism	Union strategies
Copenhagen metro project	Inherent to the work processes	Weak among migrant workers initially, but strengthens over time, in particular for some workers through union action	Proactive in enforcing and politicising labour relations and organising migrant workers
Danish slaughterhouses	None, due to high union power	Mechanical, due to social structures and institutionalisation of collectivism within union structure	Somewhat mechanical; based on strong traditions
Olkiluoto 3 construction site	Prominent (from the beginning to end), e.g. use of posted workers, not allowing direct communication between workers of different subcontractors	Weak, but with impromptu collective action by Poles	First reactive, but later more proactive

Olkiluoto 3 nuclear power plant construction

At the Olkiluoto site in Finland, the French-German main contractor Areva-Siemens sought to exempt itself from Finland’s consensual industrial relations system in three aspects. First, Areva-Siemens did not establish any cooperation with the Finnish trade unions, but sought to keep unions off the construction site and undermine union activities. Second, the work organisation was hierarchical and formal compared to typical Finnish sites, which inhibited effective functioning of the shop steward network. Third, the contractor used mostly a posted workforce, which is somewhat uncommon in Finland: two-thirds of the workforce was from abroad and mostly posted, whereas only one-quarter of the workforce in construction was at that time foreign (Alho, 2013). Because the construction union (Rakennusliitto) faced these challenges and did not have the necessary critical number of members employed by the foreign contractors, its traditional secondary boycott tactics were useless (Lillie and Sippola, 2011).

The unions’ recruitment of migrants was meagre at the beginning of the project, but increased toward the end by building up greater trust with the migrant workers (Danaj and Sippola, 2015). Initially, the shop steward network was exclusively Finnish, which was problematic in relation to the multi-ethnic workforce. Interpreters were needed, but as these were provided by the employer, a trustful relationship between union representatives and workers was still difficult to establish (Lillie and Sippola, 2011). Union officials had restricted access to the site and therefore depended on the monitoring reports of labour inspectors. Such a reliance upon secondary evidence was contrary to Rakennusliitto’s usual thinking (Lillie and Sippola, 2011). There were, however, manifestations of worker collectivism in the form of Polish workers’ impromptu industrial action when Rakennusliitto could not help with their complaints about long working hours, unpaid wages, holiday pay and other benefits. The Poles increasingly identified with the other workers at the site, demanding equal conditions and defining their interests as conflicting with those of management. The Poles even appealed to the European Migrant Workers’ Union (EMWU) to defend their case, but did not get redress by this means. A main obstacle to their union enrolment at the site was the short-term posting (like most foreign workers). This was a case of worker collectivism arising independently of – or even in contradiction to – union structures. Here however, at a later stage, another Finnish union (the Finnish Electrical Workers Union, Sähköliitto), having learned from previous union experiences on the site, assumed a more proactive role and defended successfully the case of the Polish workers.

The Copenhagen metro ring project

The Copenhagen metro ring was a large construction project involving many transnational migrant workers (Arnholtz and Refslund, 2019), where the Danish construction unions were (comparatively) rather successful in achieving high labour regulation compliance and improved unionisation of migrant workers, although several issues persisted through the construction process. The project was from the outset dominated by adverse labour relations with the many participating foreign companies and subcontractors, high labour process disintegration, and prevalent ethnic segregation of the workforce, all of which reinforced the migrant workers’ individualist strategies. From the outset the highly transnational workforce was often used to circumvent national regulations, typically through working time schemes breaching the collective agreements, but often imposed with the migrant workers’ consent, which made traditional means of union enforcement and dialogue very difficult.

To overcome these challenges, the unions initiated long-term organising efforts directed towards the migrant workers, with particular emphasis on the migrants’ needs beyond traditional industrial relations, such as help with taxes, housing, social insurance and other administrative issues. As explained by a union organiser: ‘I organise for all issues’ (interview, Danish union metro organiser with Italian background), and another union organiser further stated: ‘Taxation, health insurance, resident’s permit . . . things Danes would take care of themselves. It was an unusually high level of service we provided and it was very costly in time and resources, but it was really important for turning things around’ (interview, Danish union metro organiser).

This change in approach helped gradually build trustful relations with the migrant workers (Arnholtz and Refslund, 2019). Combined with strategic on-the-ground enforcement of collective agreements, this helped overcome many of the labour issues. Nevertheless, the shift was also helped by changes in the public owner’s approach, including a stricter application of labour clauses. While comparatively the case is a success story, issues of non-compliance remained (but were often exposed by the union), and the migrant workers’ unionisation rates remained well below those of native construction workers, and issues of disputes between the union and the migrant workers did also occur (Mathiassen, 2018).

The union strategy nevertheless strengthened worker collectivism across multiple worksites through the organisational effort to increase migrant worker awareness of the common problems across the various ethnic groups. To advance such awareness, the union furthermore held courses for migrant workers, but also to bridge some of the cleavage created by the ethnic separation. The union also helped establish a transnational union club (‘United Metro Workers’) that was to be run by the workers themselves, although this was not particularly successful and the inter-ethnic group dialogue remained weak. The union further actively sought to help align the migrant workers’ growing collective strategies with union strategies. Through strategic enforcement of the existing collective agreements they further showed the migrant workers that collectivism could make a difference. In particular, one large labour dispute helped align the workers and the unions. In the Danish industrial relations system, the union is the legal partner in the collective agreement, which meant in this dispute that the union was entitled to keep a large financial penalty (almost 2 million euros) that a company was fined in the labour court for underpayment of wages, despite the union not having any members working for the company. Yet the union decided to pay the money to the migrant workers if they joined the union post hoc, which greatly improved the migrant workers’ perception of the union.

Danish slaughterhouses

While the construction projects were characterised by workforce fragmentation, the opposite is evident in the Danish slaughterhouses, which feature highly Tayloristic work processes concentrated at large production sites (Wagner and Refslund, 2016). This setting constitutes the backbone of historically strong and resilient norms and traditions of collectivism and unionisation. The union and the shop stewards are highly important at the workplace and closely aligned with the ‘workers’ collectives’, which put substantial pressure on co-workers to join the union. In this vigorous worker culture many native slaughterhouse workers have a more instrumental approach to union organisation, acknowledging the need for strong collectivism without necessarily sharing any normative ideals about workplace organisation (interviews, union officers, Danish slaughterhouse union). The union density in the Danish pig slaughterhouse industry remains de facto at 100%, at least for all major production sites, despite significant shares of migrant workers, estimated at up to half of the workforce at several production sites and around a one-quarter in pig processing in general (Navrbjerg and Krogh, 2018). Overall the workforce setting can be described as ‘institutionalised worker collectivism’ – to the degree that most companies do not hire non-union workers, since that would lead to labour unrest. Hence migrant workers are mechanically enrolled in the union.

The workplace setting and workers’ collectivism prevent the employers from trying to use ethnic separation to lower wages, since all workers have equal wages and terms, which remains pivotal to the workers and at the same time eliminates their motivation for individualistic strategies. When migrant (slaughterhouse) workers feel unfairly treated by their employers, they feel confident in the power of the union, and some of the most militant union members are often migrants – in particular Poles (interview, union officer, Danish slaughterhouse union). Overall, the setting – despite being rather mechanically contrived – still results in credible and efficient collectivism while barring ethnic segregation from the workplace.

Four preconditions for worker collectivism: Empirical relevance based on the comparative case studies

In this section we illustrate the analytical relevance of the four theoretical preconditions to strong worker collectivism emphasised in our conceptual framework: the identification with reference groups (Piore, 1979); closeness to co-workers; perception of unity; and workers’ recognition of opposing employer interests (Lysgaard, 1967 [1961]). We further discuss how the unions can influence these preconditions to create union-based collectivism.

The identification with reference groups

Piore (1979) argued that migrant workers initially have a dual frame of reference, and therefore often accept below-average working conditions. However, a recent contribution by Clibborn (2018) raises the question whether one should instead speak of ‘multiple frames of reference’. Clibborn (2018) finds that international students, who often perform underpaid (migrant) work, tend to refer to a ‘peer frame of reference’ comprising international students rather than home- or host-country reference groups. Clibborn’s approach could theoretically be pertinent for instance to categories of ‘transnational or migrant workers’, of which we found some elements in our data, for example, in Waldemar’s⁴ (a Polish construction worker at Olkiluoto) narrative, in which he increasingly considered himself a European. Nevertheless, he still sees the home-country conditions as his benchmark as regards wages, thus corroborating Piore’s original argument:

I was never interested in money. . . . If they [other people] get thousands . . . let them have [it]. I only know that I have more money than I could get in Poland [laughs]. This is, that is the reason that I am here.

Clibborn (2018) further underlines the importance of fluid categories, suggesting the dichotomy of home- and host-country references may be too static, since reference groups may change, particularly over time, as originally argued by Piore (1979). We found the intentions and length of stay (and hence the typology of migrant workers) affect the migrant workers’ reference groups: short-term and posted workers are much less likely to perceive host-country workers as their reference group, and workers who stay for longer periods tend to change their reference group over time (cf. Piore, 1979; Quinlan and Lever-Tracy, 1988; Stark, 1991), albeit while often having elements of both host- and home-country conditions as reference group. Many of the migrant workers do not have a long-term plan for their stay, in particular in construction where the work often is temporary. Posting of workers – the dominant staffing strategy at the Olkiluoto site, and widely used at the Copenhagen site as well – also fits poorly with familiarising themselves with host-country conditions due to the short duration of stay. At the Copenhagen metro the workers initially identified with their own ethnic group and were mainly oriented to working conditions in their home country, which meant that many actively consented to evading the Danish IR system. However, this changed for some of the migrant workers over time, so that their reference group became their Danish colleagues instead. This change was further stimulated by union officials. The institutional setting in the Danish slaughterhouses, on the other hand, by design ensures that workers align to the one reference group, that of the home-country workers – at least in terms of wages and working conditions. Nonetheless, the dominating but mechanical collectivism there does not automatically result in the migrant workers (or for that matter native workers) becoming socially integrated in unions or ‘workers’ collectives’, as shown by a research project conducted for the Danish slaughterhouse union (Steffensen and Ebsen, 2017). Thus the shared reference group largely reflects the industrial worker only.

Taking the fluidity and time-dimension into consideration it can still be stated that (temporary) migrant workers’ (initial) reference group is home-country workers rather than host-country workers, although this tendency is likely to change during the migrant worker’s stay. The difference in reference group is thus a barrier to inter-group collectivism, as it may reinforce the dual labour market understanding (Piore, 1979) and reduce the migrant workers’ drive to improve their working conditions and limit their willingness to engage with local workers and unions (Andrijasevic and Sacchetto, 2016). Hence the diverging reference groups remain important for the possibility of forming solid ties to local workers and unions; in particular, unity among workers is more easily undermined if the workers have different reference groups. Furthermore, migrant workers with a strong home-country frame of reference are in any case less inclined to openly define their interests as opposed to the employers’, also because they often are highly dependent on their employers. The migrant worker reference group is nevertheless a dimension difficult for the unions to influence, in particular those of short-term workers (e.g. posted workers in construction) (cf. Greer et al., 2013), although the union can expose differences in pay and working conditions across groups of workers.

Interaction among workers

Because workers’ interaction is an important factor in the emergence of robust collectivism among migrant workers, segregation and inter-group competition is often actively imposed by employers to prevent workers from interacting. This employer strategy is often used in construction (Caro et al., 2015; Haakestad and Friberg, 2020), and was also applied by employers at the Copenhagen metro project and the Olkiluoto project. Both projects involved large numbers of subcontractors and migrant workers from more than 25 countries. For example, at Olkiluoto the communication policy was very rigid and hierarchical by Finnish standards, and workers of different subcontractors were advised not to address other workers directly. A Rakennusliitto shop steward who worked for a subcontractor was given a warning for communicating directly to the main contractor without consulting first his employer:

I delivered an invitation [from Rakennusliitto to dialogue] to [the main contractor], . . . and brought it also to [the employer]. . . . I got an email back that if I . . . again deliver any letters to [the main contractor] without . . . permission, I’ll get a written reprimand. (Interview, shop steward, Olkiluoto site)

In both construction cases the lack of interaction reduced the potential for collectivism across ethnic groups. Again the slaughterhouse case is different, as the workers interact daily, since they are concentrated in the same workplaces and work processes. The closer interaction combined with the strong tradition of unionism makes a segregation strategy unviable for the employers. The union further utilises the concentration to have joint meetings for all workers (more than 1000 workers in some slaughterhouses) discussing working conditions and wage-related topics.

Thus work organisation clearly matters with regard to workers’ interaction. In large construction projects, such as Olkiluoto, the labour process is often characterised by long, divided subcontracting chains and often ethnicity-based posting of workers (Lillie and Sippola, 2011) in which individualisation of migrant workers is much more likely than at a single production site. The labour process is on the contrary not divided in the Danish slaughterhouses; however research from other countries like Germany and the UK shows how ethnically-based separation tactics can be implemented in slaughterhouses if the institutional settings and union strength do not inhibit it, as in the Danish case (Lever and Milbourne, 2017; Wagner and Refslund, 2016).

Shared problem perception vis-a-vis the employer

Workers need a shared problem perception in opposition to that of the company in order for collectivism to thrive (Lysgaard, 1967 [1961]; see also Lockwood, 1966). This is often inhibited in migrant workers because of their expectations and dependency on the employers. In some instances migrant workers even consent to the circumvention of labour market regulation, and do not perceive it as an ‘injustice’ (cf. Kelly, 1998), whereas other workers (in particular local workers) perceive the same arrangement as an injustice. Additionally, the native workers may be sceptical towards the migrant workers, not only in regard to labour market issues, but also broader concerns on e.g. welfare benefits and tax contributions. A shop steward at Olkiluoto expressed his concerns on the posted workers not paying income tax to any country:

. . . concerning those, whose [multinational] employer has not registered in Finland. [The Polish workers] will not become liable to taxation in Finland. They wander around the world in any country except for their own and they won’t become [a] taxpayer anywhere.

This makes the cross-group problem perception very fragile and partly explains unions’ previously negative attitude towards migrant workers (in addition to outright hostility; cf. Castles and Kosack, 1973). Here the length of the migrant workers’ stay may also be influential, so that workers on longer stays are more likely to develop antagonistic sentiments towards the employers, which the findings in our two construction cases tend to illustrate. The culture in the slaughterhouse is based on workers recognising the opposing interests vis-a-vis the employers. While the migrant workers may not initially share this understanding, they are over time socialised into this understanding through the workplace culture, where they engage with native workers. Hence migrant workers’ problem perception may change over time; also for instance if their reference group changes or they face conditions perceived to be unfair. The unions may play a role in ‘constructing’ the workers’ problem perception, often just amplifying and clarifying it for the workers and providing options for action like the enforcement of existing collective agreements.

The problem perception may arise in separate groups of workers at different times, as examples from the Copenhagen metro and Olkiluoto sites demonstrate. Antagonistic sentiments towards their employer were evident for instance in the impromptu industrial action organised by Poles at the Olkiluoto site because of pay deductions, among other complaints. Yet, the Polish workers became frustrated about the handling of the dispute by Rakennusliitto and started cancelling their union memberships. Thus, while the Poles had a clear problem perception, it did not materialise into collective action because of dissonance between the unions and the migrant workers. The union took a reactive stance and was therefore not able to echo the collective sentiments among the Poles. However, this highlights how joint problem perception is not enough per se to create collectivism, especially in a segmented workplace like large construction projects, where particularly the lack of unity among workers can obstruct it.

Perception of unity

The individual labour market strategies of migrant workers reduce unity within migrant groups, but in particular across groups of workers, where it may enhance the vertical split of the working class with migrant workers often constituting a second-tier labour market (cf. Castles and Kosack, 1973: 476–477; Piore, 1979). Often the main frame for migrant workers’ labour market understanding is the micro-level kinship or meso-level ethnic group (Ciupijus et al., 2020). Moreover, unity is difficult to achieve if workers have different reference groups, since they then have different perceptions of the same work, and hence also diverging perceptions of justice, and where employers stage competition between different groups of workers.

This lack of unity can be pronounced at large construction projects where there is little interaction among ethnic groups, as seen at the Copenhagen and Olkiluoto worksites. Many Copenhagen metro project migrant workers felt themselves in opposition to the employer, but they lacked a feeling of unity with other metro workers beyond the ethnic group of co-workers, which obstructed their drive towards seeking collective solutions. In fact we found inter-group competition (to some degree initiated by the employers) between ethnic groups to be more important than cross-traits reference groups such as ‘migrant workers’, ‘metro workers’ or ‘transnational workers’ – at least in the construction cases. The Danish construction union aimed to build unity and collectivistic norms around the transnational category of ‘metro workers’, but it proved challenging to overcome this ‘identity gap’ (cf. Cregan et al., 2009). The migrant workers at the Copenhagen metro still by and large emphasised inter-group differences rather than shared problems, and many identified themselves foremost as Poles or Italians rather than as metro workers. The industrial action therefore remained largely based on the ethnic group, although there were cases comprised of different ethnic groups of workers. At the same time it was challenging for the Danish unions to get the native workers’ acceptance to engage with migrant workers on a larger scale, although they ultimately secured their general acceptance. The perception of unity is also affected by other community-based identities, e.g. at Olkiluoto, dozens of workers gathered at the local Catholic church every second weekend, which constituted an important part of many Poles’ identity (Ciupijus et al., 2020). Piotr, a Polish construction worker, went to the church, maintaining that ‘if you are Catholic you have no chance to be not Catholic’.

Unity may furthermore be absent because of ethnic distrust, distrust in the unions or because of a dual labour market understanding held by the workers. Nevertheless unity is important for collectivism, also from the unionists’ point of view: ‘If we don’t unite, then we don’t have the strength to defend ourselves’ (interview, Danish union metro activist). But as our cases show, although migrant workers may recognise they have opposing interests vis-a-vis the employers, this does not easily translate into unity among and in particular across (ethnic) groups of workers.

Unity is a dynamic phenomenon and can arise for instance in relation to strikes or other industrial actions (Fantasia, 1988; Gallie, 1983). This was seen at Olkiluoto where it seemed initially almost impossible for the unions to foster unity among workers of different ethnicity. Although the Finnish construction union Rakennusliitto made some effort to defend the Polish workers, it preferred to follow its traditional pattern of interest defence, rejecting the EMWU’s proposal to send Polish-speaking staff to help. This along with the union’s general reactionary style hampered the building of a sense of unity among the workers. Again the Danish slaughterhouses stand out as a contrasting case, with a strong worker collectivism across different ethnic groups. As expressed by an interviewee: ‘[the migrants’ perception of] unity, and that we are fighting off the worst [pressure from the company] by supporting the shop steward . . . until now is impeccable’ (interview, union official, Danish slaughterhouse union confederation). Yet, the feeling of unity there remains mainly restricted to concerns of workplace conditions – it does not extend into the social sphere.

Discussion and conclusion

Our conceptual framework, building on Lysgaard’s preconditions of collectivism, accompanied by a ‘unified frame of reference’, a derived precondition from Piore, allows us to scrutinise the preconditions crucial to union-based collectivism to develop among migrant workers despite employers’ aggressive strategies such as workforce segregation and posting of workers. We further argue that the absence or restriction of these preconditions may explain why many migrant workers rely on individual rather than collectivist labour market strategies. Some of the preconditions are somewhat intertwined, as shown above. Nevertheless, we argue that there is an independent analytical impact of each one of them, and for collectivism to arise all four preconditions must – at least to some degree – be fulfilled. Furthermore, the preconditions cannot be seen as static phenomena, but must like collectivism in general be understood dynamically and situationally (cf. Fantasia, 1988; Gallie, 1983; Thompson, 1968), since they can change rather swiftly e.g. during industrial action like strikes. The framework can thus help analytically to overcome the static understanding of migrant workers’ labour market strategies and preferences inherent in the dual labour market approach.

Our cases illustrate how the preconditions can vary across cases that share many institutional features (see Table 3), and with similar labour processes as the two construction cases. Furthermore, the fulfilment of the preconditions changes over time in our cases, in particular the construction cases, as the migrant workers become more familiar with the national system and some change their perspective during their stay (cf. Piore, 1979). While there are elements of a growing collectivism in the construction cases, this remains somewhat sporadic. We find the highest degree of collectivism (although rather mechanistic and instrumental) in the Danish slaughterhouses, which is explained by the degree of interaction and the strong tradition of unionism that influence the problem perception and unity and result in a shared ‘industrial reference group’ in relation to wages and working conditions (thus overcoming the dual labour market problem). The empirical cases further illustrate how the structural setting, in particular the difference between concentrated and fragmented labour processes, has important implications for the preconditions.

Table 3.

Preconditions for worker collectivism.

	Migrant workers’ reference groups	Interaction among workers	Shared problems vis-a-vis the employer	Workers identify with each other (unity)
Copenhagen metro project	Workers identify with home-country workers, but for some this partly changes towards host-country workers	Ethnically segregated, but increasing interaction partly facilitated by the union	Increasing and influenced by the union, but still limited	Unions proactive in creating cross-group identity, but with difficulty
Danish slaughterhouses	Workers identify as slaughterhouse workers (no separate ethnic groups)	High	Strongly articulated	At workplace level there is strong degree of unity
Olkiluoto 3 construction site	Workers identify mainly with their own ethnic group	Very low (restricted by the main contractor)	Several particular occasions, but with no joint effort to unify these	Largely lacking, mainly due to main contractor strategies, but also the union’s reactive role

Returning to two of the main issues identified in the sociological literature on labour migration – the unions’ role and the type of labour migration – we seek to contribute to the further development of these discussions. Previous research has alluded to the importance of migrant workers’ time perspective (not least Piore, 1979); however, we suggest that research on industrial relations and labour migration, in order to fully grasp the dynamic complexity, must encompass and combine reflections on the time dimension with the migrant worker typology (Engbersen et al., 2013), e.g. whether the worker is settled or transient. Theoretical and empirical enquiries seem to suggest that all four of our preconditions are more likely to be met the longer and more permanent the intended stay of the migrant workers in the host country (see e.g. Krings, 2014; Quinlan and Lever-Tracy, 1988; Refslund, 2018; Stark, 1991). Other studies have found that circular or transitional migrants for instance are less likely to act collectively and hence are very difficult to organise for unions (Berntsen, 2016; Berntsen and Lillie, 2016; Greer et al., 2013). Our case studies show likewise how the long duration of the project played a role in the comparative success of the Danish construction unions at the Copenhagen metro. The fact that the migrant workers at the Copenhagen metro and Olkiluoto sites were predominantly posted also played a major role in the potential for collectivism, and the employers were able to use a separation strategy efficiently because of the workers’ short stay. This further underlines the dynamic aspect of worker collectivism, suggesting that previous accounts of migrant worker collectivism (or the lack thereof) may have been too static. Hence, we encourage future research to pay more attention to dynamic and formative aspects of worker collectivism along with national, regional and sectoral differences (Doellgast et al., 2018; Wagner and Refslund, 2016), as well as structural factors such as labour market characteristics (Cam, 2014). Our framework can hence be applied to situational occurrences of collectivism, as in Fantasia’s (1988) seminal work on ‘cultures of solidarity’.

The role of unions in relation to migrant workers has been widely discussed in the literature, and the historical debate occasionally found that unions actively contributed to labour market dualisation (e.g. Castles and Kosack, 1973; Penninx and Roosblad, 2000). Nevertheless, our studies corroborate the argument that unions in most contemporary Western societies seek to include and represent migrant workers rather than exclude them (Adler et al., 2014; Doellgast et al., 2018; Haus, 2002; Marino et al., 2015; Pulignano et al., 2015). Unions can thus play an active role in bridging collectivism across different groups of workers. Yet, our framework can also help explain the oft-occurring disconnect between unions and migrant workers by pointing to the lacking preconditions for collectivism among the migrant workers. Conversely, it has been underlined that unions need to adapt their strategy and behaviour to migrant workers, their needs and often intersectional vulnerabilities, which diverge from those of native workers, something unions have previously not been very good at (Alberti et al., 2013; Holgate, 2005). In our cases, the Poles’ industrial action at Olkiluoto illustrates this: collectivism formed without union involvement and the interaction with the union was problematic. Hence the unions must seek to overcome the language barrier to be able to extend the traditional union structures to a multicultural workforce, and to be able to act upon the migrant workers’ impetus towards collectivism (cf. Holgate, 2005; Milkman, 2006). This typically requires a proactive strategy, as illustrated by the Danish construction unions at the Copenhagen metro project, who changed to a ‘social unionism’ strategy rather than maintain a strong enforcement approach which had proven insufficient. This strategy involves helping the migrant workers with issues beyond the industrial relations sphere, something the Finnish construction unions at Olkiluoto did not do (or at least not to the same extent). By contrast, the slaughterhouses show that in a strong institutional setting with strong traditions of unionism, this can be achieved without the unions changing strategy. The discussion of unions’ role further endorses a dynamic understanding of labour migration and collectivism. Also, incorporating an intersectional approach to the analysis of worker collectivism and analysing collectivism of migrant workers vis-a-vis the rest of the workforce, as has been done with regard to unionisation (Alberti et al., 2013; Cam, 2014), would be potential areas for further research.

The strong institutional setting in our case countries suggests that we should find stronger traits of migrant worker collectivism here; nevertheless several of the preconditions are found to be weak in our cases, in particular in the construction cases, for instance the sense of unity. This shows that a strong institutional setting is not enough to foster collectivism among migrant workers, and we would expect studies in countries with weaker institutions and unions than in Denmark and Finland to have a lower likelihood of meeting the preconditions. This is also something for future research to investigate. This has the empirical implication that future initiatives to overcome migrant worker precarity in countries with weaker institutions should potentially include other policy measures such as public regulation emphasising migrant workers’ legal and social rights and the enforcement of these, rather than relying on the collective action of the migrant workers. In countries with weaker national regulation EU regulation may take on a more principal role. Concentrated efforts are thus needed at the EU level to regulate, for example posting of workers, which is a key element in the precarious situation of many migrant workers (Arnhotz and Lillie, 2020), for higher protection of these workers. EU regulation is also only useful when accompanied by monitoring and enforcement capacities; so these need to be further developed at both national and European level.

Overall, the framework suggested here helps to analytically understand how and why labour market collectivism may form among migrant workers, something the sociological literature on labour migration often has not had much to say about, due to the emphasis on the dual frame of reference found in most classic sociological work on labour migration (e.g. Castles and Kosack, 1973; Piore, 1979; Waldinger and Lichter, 2003). The dual frame of reference may result in a rather static approach, where it is assumed that collective action towards improving working conditions is severely limited among migrant workers, and that migrant workers by and large accept their inferior labour market position (and with unions potentially contributing to reinforcing the dual labour market). Other often-cited strands of the literature like Penninx and Roosblad (2000) have been more heuristic in focusing on whether unions chose to include migrants or not, and only paid secondary attention to the dynamic dimensions of how and why collectivism forms in the first place, in particular at workplace level. By theorising how collective interests are formed among migrant workers and identifying the preconditions for collectivism our framework helps further enlighten research as to what is needed for ‘organising approaches’ (Holgate, 2005; Milkman, 2006) to be successful with migrant workers.

Footnotes

Acknowledgements

The following have provided helpful comments on previous versions: Jens Arnholtz, Nathan Lillie, Lisa Berntsen, Jean-Pierre Durand, Jeremy Waddington and Tesseltje De Lange.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Bjarke Refslund

Notes

Author biographies

Bjarke Refslund is Associate Professor in Sociology at Aalborg University. His main research areas include industrial relations, labour migration and labour market sociology as well as collectivism and unions. His research has feature in journals such as European Journal of Industrial Relations, Journal of Common Market Studies, Work, Employment and Society and Competition and Change.

Markku Sippola is Senior Lecturer at University of Helsinki, Finland. His research interests revolve around industrial relations, migration, social policy as well as societal and labour processes in the Baltic, Nordic and Russian contexts. He has published in Economic and Industrial Democracy, Employee Relations, European Journal of Industrial Relations, Journal of Baltic Studies and Work, Employment and Society.

References

Adler

Tapia

Turner

(2014) Mobilizing against Inequality. Ithaca, NY: Cornell University Press.

Alberti

(2014) Mobility strategies, ‘mobility differentials’ and ‘transnational exit’: The experiences of precarious migrants in London’s hospitality jobs. Work, Employment and Society 28(6): 865–881.

Alberti

Holgate

Tapia

(2013) Organising migrants as workers or as migrant workers? Intersectionality, trade unions and precarious work. The International Journal of Human Resource Management 24(22): 4132–4148.

Alho

(2013) Trade union responses to transnational labour mobility in the Finnish-Estonian context. Nordic Journal of Working Life Studies 3(3): 133–153.

Almond

Connolly

(2020) A manifesto for ‘slow’ comparative research on work and employment. European Journal of Industrial Relations 26(1): 59–74.

Andersen

Dølvik

Ibsen

(2014) Nordic Labour Market Models in Open Markets. Brussels: European Trade Union Institute (ETUI).

Andrijasevic

Sacchetto

(2016) From labour migration to labour mobility? The return of the multinational worker in Europe. Transfer 22(2): 219–231.

Arnholtz

Lillie

(2020) Posted Work in the European Union: The Political Economy of Free Movement. Abingdon: Routledge.

Arnholtz

Refslund

(2019) Active enactment and virtuous circles of employment relations: How Danish unions organised the transnationalised Copenhagen metro construction project. Work, Employment and Society 33(4): 682–699.

10.

Berntsen

(2016) Reworking labour practices: On the agency of unorganized mobile migrant construction workers. Work, Employment and Society 30(3): 472–488.

11.

Berntsen

Lillie

(2016) Hyper-mobile migrant workers and Dutch trade union representation strategies at the Eemshaven construction sites. Economic and Industrial Democracy 37(1): 171–187.

12.

Cam

(2014) Non-unionised migrant workers in Britain: Evidence from the Labour Force Survey for a triple-challenge model. Economic and Industrial Democracy 35(3): 531–556.

13.

Caraker

(2011) Solidaritet og individualisme: Fællesskabsformer blandt danske lønmodtagere. Arbejderhistorie 2: 26–45.

14.

Caraker

Høgedahl

Jørgensen

Møberg

(2014) fællesskab før forskelle. Research report. Aalborg: CARMA.

15.

Caro

Berntsen

Lillie

Wagner

(2015) Posted migration and segregation in the European construction sector. Journal of Ethnic and Migration Studies 41(10): 1600–1620.

16.

Castles

Kosack

(1973) Immigrant Workers and Class Structure in Western Europe. Oxford: Oxford University Press.

17.

Ciupijus

Forde

Mackenzie

(2020) Micro- and meso-regulatory spaces of labour mobility power: The role of ethnic and kinship networks in shaping work-related movements of post-2004 Central Eastern European migrants to the UK. Population, Space and Place 26(3): e2300.

18.

Clibborn

(2018) Multiple frames of reference: Why international student workers in Australia tolerate underpayment. Economic and Industrial Democracy. Epub ahead of print 12 April 2018. DOI: 10.1177/0143831X18765247

19.

Cregan

Bartram

Stanton

(2009) Union organizing as a mobilizing strategy: The impact of social identity and transformational leadership on the collectivism of union members. British Journal of Industrial Relations 47(4): 701–722.

20.

Crouch

(1982) Trade Unions: The Logic of Collective Action. London: Fontana.

21.

Crowley

(2004) Explaining labor weakness in post-communist Europe: Historical legacies and comparative perspective. East European Politics and Societies 18(3): 394–429.

22.

Danaj

Sippola

(2015) Organizing posted workers in the construction sector. In: Drahokoupil

(ed.) The Outsourcing Challenge. Brussels: European Trade Union Institute (ETUI), pp. 217–236.

23.

Doellgast

Lillie

Pulignano

(2018) Reconstructing Solidarity. Oxford: Oxford University Press.

24.

Dølvik

Eldring

(2006) Industrial relations responses to migration and posting of workers after EU enlargement: Nordic trends and differences. Transfer 12(2): 213–230.

25.

Engbersen

Leerkes

Grabowska-Lusinska

, et al. (2013) On the differential attachments of migrants from Central and Eastern Europe: A typology of labour migration. Journal of Ethnic and Migration Studies 39(6): 959–981.

26.

Fantasia

(1988) Cultures of Solidarity. Berkeley: University of California Press.

27.

Fox

(1971) A Sociology of Work in Industry. London: Collier-Macmillan.

28.

Fox

(1985) History and Heritage. London: George Allen & Unwin.

29.

Gallie

(1983) Social Inequality and Class Radicalism in France and Britain. Cambridge: Cambridge University Press.

30.

Gorodzeisky

Richards

(2013) Trade unions and migrant workers in Western Europe. European Journal of Industrial Relations 19(3): 239–254.

31.

Greer

Ciupijus

Lillie

(2013) The European Migrant Workers Union and the barriers to transnational industrial citizenship. European Journal of Industrial Relations 19(1): 5–20.

32.

Gumbrell-McCormick

Hyman

(2013) Trade Unions in Western Europe: Hard Times, Hard Choices. Oxford: Oxford University Press.

33.

Haakestad

Friberg

(2020) Deskilling revisited: Labour migration, neo-Taylorism and the degradation of craft work in the Norwegian construction industry. Economic and Industrial Democracy 41(3): 630–651.

34.

Hasle

Sørensen

(2013) Employees as individually and collectively acting subjects: Key contributions from Nordic working life research. Nordic Journal of Working Life Studies 3(3): 9–29.

35.

Haus

(2002) Unions, Immigration, and Internationalization. New York: Palgrave Macmillan.

36.

Heindlmaier

Blauberger

(2017) Enter at your own risk: Free movement of EU citizens in practice. West European Politics 40(7): 1–20.

37.

Holgate

(2005) Organizing migrant workers: A case study of working conditions and unionization in a London sandwich factory. Work, Employment and Society 19(3): 463–480.

38.

Holgate

(2015) Community organising in the UK: A ‘new’ approach for trade unions? Economic and Industrial Democracy 36(3): 431–455.

39.

Hyman

(1992) Trade unions and the disintegration of the working class. In: Regini

(ed.) The Future of Labour Movements. London: Sage, pp. 150–168.

40.

Hyman

(1999) Imagined solidarities: Can trade unions resist globalization. In: Leisink

(ed.) Globalization and Labour Relations. Cheltenham: Edward Elgar, pp. 94–115.

41.

Kall

Lillie

Sippola

Mankki

(2018) Overcoming barriers to transnational organizing through identity work: Finnish-Estonian trade union cooperation. Work, Employment and Society 33(2): 208–225.

42.

Kalmus

Vihalemm

(2006) Distinct mental structures in transitional culture: An empirical analysis of values and identities in Estonia and Sweden. Journal of Baltic Studies 37(1): 94–123.

43.

Kelly

(1998) Rethinking Industrial Relations. London: Routledge.

44.

Kettunen

(2012) Reinterpreting the historicity of the Nordic model. Nordic Journal of Working Life Studies 2(4): 21–43.

45.

Kevätsalo

(1995) Edut aatteena ja vakuutus solidaarisuutena. Report. Helsinki: The Finnish Metalworkers’ Union.

46.

Krings

(2014) Unorganisable? Migrant workers and trade union membership. In: Industrial Relations in Europe Conference (IREC), Dublin, 10–12 September 2014.

47.

Lever

Milbourne

(2017) The structural invisibility of outsiders: The role of migrant labour in the meat-processing industry. Sociology 51(2): 306–322.

48.

Lillie

(2016) The right not to have rights: Posted worker acquiescence and the European Union labor rights framework. Theoretical Inquiries in Law 17(1): 39–62.

49.

Lillie

Greer

(2007) Industrial relations, migration, and neoliberal politics: The case of the European construction sector. Politics and Society 35(4): 551–581.

50.

Lillie

Sippola

(2011) National unions and transnational workers: The case of Olkiluoto 3, Finland. Work, Employment and Society 25(2): 292–308.

51.

Lockwood

(1966) Sources of variation in working class images of society. The Sociological Review 14(3): 249–267.

52.

Lysgaard

(1967 [1961]) Arbeiderkollektivet. Oslo: Universitetsforlaget.

53.

McBride

Lucio

(2011) Dimensions of collectivism: Occupation, community and the increasing role of memory and personal dynamics in the debate. Work, Employment and Society 25(4): 794–805.

54.

MacKenzie

Forde

(2009) The rhetoric of the ‘good worker’ versus the realities of employers’ use and the experiences of migrant workers. Work, Employment and Society 23(1): 142–159.

55.

MacKenzie

Martinez Lucio

(2019) Regulation, migration and the implications for industrial relations. Journal of Industrial Relations 61(2): 176–197.

56.

Marino

Penninx

Roosblad

(2015) Trade unions, immigration and immigrants in Europe revisited: Unions’ attitudes and actions under new conditions. Comparative Migration Studies 3(1): 1–16.

57.

Mathiassen

(2018) Den faglige indsats på Cityringen #2. Report for 3F/BJMF.

58.

Merton

(1968 [1957]) Social Theory and Social Structure. New York: Free Press.

59.

Milkman

(2006) LA Story. New York: Russell Sage.

60.

Navrbjerg

Krogh

(2018) WP2 National Report – Denmark: Structural Characteristics of the Pork Value Chain. Copenhagen: FAOS, University of Copenhagen.

61.

Ness

(2012) Constructing masculinity in the building trades: ‘Most jobs in the construction industry can be done by women’. Gender, Work and Organization 19(6): 654–676.

62.

Peetz

(2010) Are individualistic attitudes killing collectivism? Transfer 16(3): 383–398.

63.

Penninx

Roosblad

(2000) Trade Unions, Immigration, and Immigrants in Europe, 1960–1993. Oxford: Berghahn.

64.

Piore

(1979) Birds of Passage. Cambridge: Cambridge University Press.

65.

Pulignano

Meardi

Doerflinger

(2015) Trade unions and labour market dualisation: A comparison of policies and attitudes towards agency and migrant workers in Germany and Belgium. Work, Employment and Society 29(5): 808–825.

66.

Quinlan

Lever-Tracy

(1988) Immigrant workers, trade unions and industrial struggle: An overview of the Australian experience, 1945–85. Economic and Industrial Democracy 9(1): 7–41.

67.

Rantala

(2013) Interest defence and commitment to trade unions. Graduate thesis, Department of Social Sciences and Philosophy, University of Jyväskylä.

68.

Refslund

(2018) When strong unions meet precarious migrants: Building trustful relations to unionise labour migrants in a high union-density setting. Economic and Industrial Democracy. Epub ahead of print 13 March 2018. DOI: 10.1177/0143831X18760989

69.

Salmela

(2011) Kollektiiviset tunteet solidaarisuuden liimana. In: Laitinen

Pessi

(eds) Solidaarisuus. Helsinki: Gaudeamus, pp. 61–81.

70.

Simms

(2012) Imagined solidarities: Where is class in union organising? Capital and Class 36(1): 97–115.

71.

Sippola

(2013) The awkward choices facing the Baltic worker: Exit or loyalty. Journal of Baltic Studies 44(4): 451–473.

72.

Skorstad

Karlsson

(2020) The worker collectivity and Anglo-Saxon theories of collectivity. Economic and Industrial Democracy 41(2): 397–418.

73.

Stark

(1991) The Migration of Labor. Cambridge: Basil Blackwell.

74.

Steffensen

Ebsen

(2017) Ind i fællesskabet. En kvalitativ undersøgelse af integrationen af polske medarbejdere på danske slagterier og i fødevareforbundet NNF. Report. København: Political think-tank CEVEA for NNF (Danish Food Workers Union).

75.

Thompson

(1968) The Making of the English Working Class. London: Penguin Books.

76.

Waddington

Whitston

(1997) Why do people join unions in a period of membership decline? British Journal of Industrial Relations 35(4): 515–546.

77.

Wagner

Refslund

(2016) Understanding the diverging trajectories of slaughterhouse work in Denmark and Germany: A power resource approach. European Journal of Industrial Relations 22(4): 335–351.

78.

Waldinger

Lichter

(2003) How the Other Half Works. Berkeley: University of California Press.

79.

Webb

(1897) Industrial Democracy. London: Longmans & Green.