Introduction

International framework agreements (IFAs) are a new generation of agreements between multinational companies (MNCs) and global union federations (GUFs). While still an emerging phenomenon most typically associated with MNCs that are already engaged in organized social dialogue at national and transnational levels, notably for MNCs originating in the European Union, they represent an approach to corporate social responsibility (CSR) whereby the MNC commits to respect worker rights in its operations and often in its supply chain. As is increasingly expected of CSR policies and reflected in international trade and investment agreements, IFAs usually refer to compliance with core International Labour Organization (ILO) conventions, including Conventions 87 and 98 on the right to associate, to organize and to bargain collectively (ILO 2016; Papadakis 2008; Bourque 2005; Connor 2004).

Several studies have highlighted the limited impact of IFAs on worker rights (Williams et al. 2015; Riisgaard and Hammer 2011), but others have shown that these agreements can have a concrete impact on worker capacity to organize and engage in collective bargaining (Cotton and Royle 2014; Bourque and Hennebert 2011; Miller 2004; Papadakis 2011; Robinson 2011). This is significant because it shows CSR in a different light, one that is often neglected. Our aim in this paper is to explore the conditions under which IFAs can enhance worker and trade union rights through an analysis of a successful union organizing campaign in Colombia in 2012.

This campaign can be considered as a critical case for three reasons. First, the campaign focused on a large French MNC in the service sector, probably among the more difficult industries to organize. Second, this particular MNC was in a difficult financial situation, suggesting if one follows the standard interpretation of CSR that it would be less likely to engage in socially responsible activities because it had fewer resources (Orlitzky et al. 2003; Waddock and Graves 1997). Third, the case was in many ways a litmus test of the effectiveness of IFAs because the workers concerned were in Colombia, which has long been considered to be one of the most dangerous countries in the world for the exercise of trade union rights (ILO 2009) and where violations of the freedom of association and union rights in Colombia had become a cause celebre in the international trade union movement.

Our focus is on the way that the IFA was used in this particular campaign. Our basic argument is that management strategies towards CSR and social dialogue influence the impact of IFAs on worker rights but that this relationship is mediated by the capacity of managers and worker representatives at multiple levels to mobilize their capabilities. We explore how the success of the organizing campaign in Colombia was related to the dissemination of capabilities across a multilevel coordination structure and, most importantly, their complementarities at various levels. This approach focuses on the dynamic between trade unions and with management representatives at various levels. It also illustrates the pivotal role of various sources of power in understanding the effectiveness of IFAs.

The paper contributes to the literature on both CSR and multilevel governance. First, in exploring the role of institutionalized dialogue on CSR, the paper builds on institutional theories of CSR (Bartley 2007; Campbell 2007) to highlight how institutions beyond the market are not only necessary to ensure that firms behave in socially responsible ways but how firms need to engage in institutionalized dialogue with local managers and worker representatives to reinforce worker rights (Fichter et al. 2011; Niforou 2012). Second, the paper contributes to the literature on multilevel governance. Originating in political science (Jessop 2004; Marks et al. 1996; Peters and Pierre 2001, 2004), particularly in accounts of European political integration, multilevel governance approaches have been extended to other fields of inquiry such as environmental studies (Amundsen et al. 2010), industrial relations (Marginson and Sisson 1996) and international labor rights (Dehnen and Pries 2014; Helfen and Fichter 2013). We contribute to this literature in two ways. The paper shows how levels of regulation (local, national, regional, and supranational) across different spaces and the strategies of dominant actors (HQ, subsidiaries managers, GUF, workplace, and national workers representatives) are interconnected. It also highlights the crucial role of actors’ strategic capabilities, in particular their capacity to create a shared meaning and vision of CSR and social dialogue among the stake holders (framing capabilities), to link and connect actors and institutions across various levels (bridging capabilities), and to arbitrate between and reconcile conflicting interests (intermediating capabilities) (Garriga 2010; Lévesque and Murray 2010a; Manning and Roessler 2014; Tengblad and Ohlsson 2010).

The first three sections review the research on codes of conduct and IFAs and the factors that favor the implementation of IFAs, notably with regard to MNC and trade union capabilities. The consideration of these factors allows us to set out the key analytical factors that inform this study of the effectiveness of IFAs in the context of CSR. The fourth and the fifth sections outline the research methods and the case study background. The sixth and following sections outline the campaign undertaken to organize Colombian workers and analyze trade union and management strategies and the relationships between the various dimensions of power and their influence on the effectiveness of IFAs in promoting worker rights. The discussion contrasts our results with previous research on IFAs, while the conclusion explores the consequences of our findings for the literature on CSR and multilevel governance.

Corporate Social Responsibilities: from Codes of Conduct to IFAs

CSR encompasses various fields including the environment, sustainable development, ethics, and worker protection. Since the early 1990s, the development of codes of conduct to improve working conditions within MNCs and across global value chains (GVCs) has gathered pace and it continues to be a subject of debate (Evans 2015; Williams et al. 2015). The many criticisms leveled against such codes have been tempered by various empirical studies underscoring their potential to improve working conditions, occupational health and safety, working hours, and other outcomes (Egels-Zandén and Merk 2014; Barrientos and Smith 2007; Zadek 2004; Frenkel 2001). However, there appears to be a consensus that codes of conduct have a limited impact on the promotion of union rights such as freedom of association and the right to bargain collectively. In a recent comprehensive study of private compliance programs in global value chains, Locke and colleagues show that while these programs may have some impact on working conditions, notably health and safety, they do not deliver in terms of reinforcing labor standards and rights, particularly the freedom of association (Locke 2013). Anner (2012) comes to the same conclusion after reviewing the audits conducted between 2002 and 2010 by the Fair Labor Association. The extensive review of Egels-Zanden and Merk (2014) on codes of conduct similarly shows that these tools of regulation have a limited potential to improve trade union rights.

Given the limits of codes of conduct as a means of ensuring compliance with core ILO conventions, which is a key objective of CSR policies, there has been considerable interest in the role of IFAs to achieve this objective (ILO 2016; Papadakis 2008; Bourque 2005; Connor 2004). IFAs have been promoted chiefly by the GUFs (global union federations) and, although typically not legally binding, are seen to represent a higher level of commitment on the part of MNCs because, unlike corporate codes of conduct, they are the outcome of bilateral negotiations between MNCs and trade union representatives. Although IFAs may differ greatly in their content, they usually incorporate ILO core labor rights (freedom of association and the right to collective bargaining, abolition of child and forced labor, elimination of employment discrimination) and as such make explicit reference to other arenas of regulation, including ILO and host country legislation (Niforou 2012). They are the outcome of a voluntary governance approach, without government intervention, in the protection of labor rights (Hammer 2005).

IFAs have been associated with new forms of global, transnational or multilevel governance (Egels-Zandén 2009; Fitchter et al. 2011; Niforou 2012). The concept of multilevel governance has been widely adopted and helps to understand the more complex forms of interdependence between different levels of regulation. Peters and Pierre suggest that multilevel governance “refers to negotiated, non-hierarchical exchange, between institutions at the transnational, national, regional and local level” (2001, p. 131). They emphasize the vertical layering of governance while Craig suggests that for advocates of multilevel governance, “political arenas are interconnected rather than nested” (Craig 2003, p. 22). Marks et al. express the same view and add that multilevel governance involves a sharing of decision-making competencies at different levels (1996, p. 346).

A key question concerns whether actors adopt a centric or multilevel approach (Jessop 2004; Marks and Hooghe 2004) towards IFAs. Research on IFAs highlights that both of these approaches have been followed (Fichter et al. 2011; Helfen and Fichter 2013; Hennebert and Bourque 2013). In some cases, the negotiation and implementation of an IFA are conducted by a small number of union leaders, managing the dialogue with the senior management of an MNC and facilitating exchanges between unions (Hennebert and Bourque 2013). In other cases, the process is characterized by the existence of mechanisms at multiple levels, which ensure the links and coordination between various stakeholders, including management, trade union representatives and NGO across various levels (Cotton and Royle 2014; Fichter and McCallum 2015). It is our contention that the development of a multilevel approach towards IFAs requires the sharing of decision-making competencies at different levels and the dissemination of management and trade union representative’s capabilities across levels.

The Role of Management and Trade Unions Capabilities

We draw on two strands of literature to disentangle the role of actor capabilities in the implementation of IFAs: micro-politics within MNCs and transnational trade union perspectives.

The micro-politics perspective takes an actor centric approach and examines both strategies and capabilities developed by managers at various levels (Becker-Ritterspach et al. 2016). This perspective is particularly relevant since it explores the transfer of practices from HQs to subsidiaries (Fenton O’Creevy et al. 2011; Ferner et al. 2012; Morgan and Kristensen 2006). Transfer of practices to the local level is complex because MNCs are contested arenas characterized by tensions between competing actors (Edwards and Bélanger 2009). Far from monolithic, they are built around coalitions of actors pursuing different goals and deploying their power to attain them. HQ actors can take advantage of their position at the interface of multiple institutional terrains to engage in what Morgan et al. (2007) label institutional arbitrage or bricolage. HQ actors have the possibility to pull together elements from various institutional contexts and can act as rule-makers in respect of host institutional contexts and modify local norms and practices. Actors at subsidiary level can challenge and resist such transfers by asserting the incompatibility between the transfer of HQ policy and practice with the sources of competitive advantage, the normative frameworks and traditional practices in the host country (Ferner et al. 2012; Dörrenbächer and Gammelgaard 2011; Bouquet and Birkinshaw 2008; Morgan and Kristensen 2006). In their influential study of different national subsidiaries of the same MNC, Kristensen and Zetlin (2005) show how local actors strategize in order to reinforce the position of their subsidiaries within the corporation.

The micro-politics perspective has contributed significantly to the identification of key management capabilities, in particular bridging capabilities, capabilities for collaborative action, and capabilities to manage meaning. In their analysis of the micro-politics of resistance to the transfer of knowledge-based practices across multiples sites of an MNC, Fenton O’Creevy et al. (2011) show the importance of bridging skills and brokering collective knowledge. Bridging capabilities entails linkages between actors and institutions across scale and space. For Kristensen and Zetlin (2005), it is not only a question of boundary spanning but also of managing conflict and cooperation in a context where there are multiple interests at play. In their study, they highlight the crucial role of the capability for collaborative action (2005: 206). In particular, they emphasize the importance for actors to develop collaborative relations across various levels (within the site, within the local economy and within the MNC), to mediate conflict, and to balance competition and cooperation. Ferner et al. (2012) in their analysis of transfer of employment practices in MNCs insist that the capability of HQ and subsidiaries actors to shape the cognitive and normative frame is especially crucial (2012, p. 166). They contend that transfer of practices reveals the exercise of agency because hitherto taken-for-granted cognitive and normative frames become visible.

The transnational trade union perspective is predicated on the asymmetrical distribution of power between employers, who are increasingly international in scale and scope in the context of globalization, and trade unions, which remain largely embedded in local or national contexts (Anner 2012; Martínez Lucio 2010; Webster et al. 2008). The development of production networks and the ability of MNCs to transfer production between countries reinforce management bargaining power vis-à-vis workplace unions. MNCs use coercive comparisons and threats of relocation to obtain wage concessions and more flexible work and employment arrangements. The extension of global value chains (GVCs) in Eastern Europe, South America and Asia accelerates this process and puts unions operating in quite different economic and institutional contexts in competition with one another. It is therefore argued that trade unions must extend their practices into transnational space as a counterweight to MNC power. The extension of trade union action into transnational space opens up the development of multiple types of linkage between trade unions at the local, national and supranational levels (Fairbrother et al. 2013; Croucher and Cotton 2009). Such linkages include exchanges, participation in forums, and campaigns and can take a variety of forms. Each level of regulation remains relatively autonomous and is of variable scope but these different levels are interconnected in rather complex ways across both scales and spaces.

Several scholars have attempted to capture the evolving forms and models of transnational unionism (Anner 2013; Martínez Lucio 2010; Munck 2004). Others have focused on the new sets of resources and capabilities that trade unions deploy in their transnational actions (Fairbrother 2013; Lévesque and Murray 2010a, 2010b; Turnbull 2006). Several studies (Fairbrother 2013; Turnbull 2006; Wills 2002) examine how union leaders bridge issues at the local, national and international levels. Such bridging activity is often refined and developed over time, reflecting the changing capacities of unions. For example, Turnbull shows that, through bridging, trade union leaders “could draw on a much wider set of social resources to defend their members’ interests” (2006, p. 321). Several scholars emphasize the importance of intermediating capabilities such as mediating between contending interests and fostering collaborative action (Lévesque and Murray 2010a). Bronfenbrenner (2007) points to the tensions that can arise in coordinating actions across countries and shows that the success of these actions rests on the ability to mediate between both the varied interests pursued by workers and the contrasting repertoires of action and mobilization strategies developed by trade unions. Finally, several studies demonstrate how the trade union capability to frame a narrative plays a pivotal role. Narratives consist of the range of values, shared understandings, stories, and visions that aggregate identities and interests and translate and inform motives (Lévesque and Murray 2013). They constitute a body of interpretative and action frames that can be mobilized to explain new situations and new contexts. Fairbrother (2013) shows how trade unions’ framing capabilities play a pivotal role in integrating the tools of transnational regulation into their repertoires of action.

Despite some terminological differences between these two strands of literature, both the micro-politics within MNCs perspective and the transnational trade union perspective emphasize actor agency and stress the importance of actor capabilities, in particular their capacity to frame a shared vision among various stakeholders (framing capabilities or managing meaning), to link and connect institutions and actors (bridging capabilities) and to mediate between conflicting demands (intermediating capabilities and capabilities for collaborative action). In the following section, we examine research on IFAs through the lens of these two perspectives.

Understanding the Implementation of IFAs

The strategies pursued by HQs and subsidiaries in the implementation of IFAs are quite variable. While some HQs put much pressure on their subsidiaries to implement IFAs (Dehnen and Pries 2014), others only “recommend” their implementation, giving much leeway to local actors (William et al. 2015; Niforou 2014). For instance, Niforou (2012) stresses that the HQs of the MNCs Endesa and Telephonica demonstrated neither the willingness nor the capacity to impose the implementation of their IFAs in their Latin American subsidiaries, instead giving precedence to local laws and preferences. Subsidiaries can also adopt various approaches to IFAs. Some avoid or only engage in ritual compliance (Williams et al. 2015; Rissgaard and Hammer 2011); others are more likely to internalize the practices related to IFAs and apply the agreement (Cotton and Royle 2014; Fichter et al. 2011).

The implementation of IFAs seems to be related to the capacity of HQ and subsidiary actors to shape the normative and cognitive understanding of IFAs and of the role of the various stakeholders (framing or managing meaning capabilities). HQ actors try to control cognitive frames through their vision and value statements about IFAs and social responsibility, while subsidiary actors may try to challenge this frame on the basis of their interpretation of national and local practices (Tengblad and Ohlsson 2010; Campbell 2007). Several studies of IFAs highlight this dynamic process (Dehnen and Pries 2014; Hammer 2013). For instance, subsidiary actors can escape compliance by highlighting the particularity of their context and its impact on subsidiary performance (Dehnen 2013). Hence, subsidiaries often draw on the particularity of their host institutional context to resist the implementation of IFAs. For example, Hammer (2013) has shown how local managers draw on permissive institutional arrangements in India and Russia to prevent the implementation of IFAs in the construction industry. Niforou (2012) observes the same phenomenon in Latin America. Trade unions are also active on this front. According to Cotton and Royle (2014), the success of the campaign to organize contract mineworkers in Colombia is related to the enlargement of an action frame (a sense of injustice regarding the treatment of these workers within the industry) that changed the dominant narrative regarding the difficulty of organizing these workers.

The effective implementation of IFAs also appears to be associated with mechanisms at multiple levels to ensure the links between actors (bridging capabilities) and enhance their participation. Fichter and McCallum (2015) argue that the involvement of workers and local trade unions explains the success of the UNI organizing campaign of G4S in South Africa and India. Cotton and Royle (2014) also emphasize the capacity of trade unions to link and coordinate actions across multiple scales and spaces. They highlight how the relationship with the community contributed to the success of the ICEM organizing campaign of contract workers in the mining industry in Colombia. Helfen and Fichter (2013) point to the decisive role played by coordinators in bridging actors across transnational networks. The absence of such bridging capabilities appears to be detrimental to the implementation of IFAs. Williams et al. (2015) link the lack of involvement of local actors, particularly managers, to the limited effectiveness of IFAs as regards the promotion of worker rights. Niforou (2012) comes to the same conclusion in her study of Endesa and Telephonica in Latin America.

Finally, several studies suggest that the success of IFAs is related to the capacity of actors to mediate between conflicting interests and to foster collaborative action (intermediating capabilities). Dehnen and Pries (2014) consider that, given the multiple interests at play, the capacity of actors to mediate between conflicting demands is paramount. Fichter et al. (2011) highlight that the existence of conflict resolution mechanisms is an essential element for effective transfer and implementation of IFAs. Fichter and McCallum (2015) also stress the importance of balancing conflicting and collaborative approaches. While a social dialogue approach can foster the negotiation of IFAs, such an approach is often less useful for ensuring its implementation. Their case studies show that an approach that fosters both dialogue and battle, in a form of “conflictual partnership,” is more effective for the implementation of IFAs.

It is our contention that, in order to understand the effectiveness of IFAs in securing worker rights, we need to map out how these various management and trade unions capabilities play out across scales and spaces. The case considered in this paper offers an opportunity to do so.

Research Method

The research method is based on a case study of a French multinational (MNC) in the service sector, hereinafter FrenchCorp. It draws on multiple data sources including observation of meetings, extensive field-notes, taped and transcribed interviews and multiple internal documents. In 2010 and 2012, research team members attended company and trade union official meetings. Both of these meetings lasted two days and involved two types of activities. The first included FrenchCorp management team and trade unions from UNI and the affiliates from all the unionized workplaces in Europe, Latin America, and Asia. During this first day, management representatives reported on the financial situation of the firm and its future plan, and also shared with trade union representatives and managers from subsidiaries its vision and approach towards CSR and social dialogue. A period was open for trade union representatives to engage with management team and put forward specific queries and problems. On the second day, trade union representatives met alone, i.e., without managers from FrenchCorp. UNI trade union representatives reported on the situation at FrenchCorp and highlighted the priorities of the network. Trade union representatives exchanged information on working conditions and collective bargaining in their respective workplaces and shared their vision about their expectation from the network. Between 2010 and 2014, we also conducted 30 interviews in European and Latin American with senior managers and trade union representatives at various levels (local, national, regional, and international) and from various countries in Latin America (Colombia and Argentina) and in Europe (France and Belgium).

We draw on this rich qualitative data on multiple units and levels of analysis to construct a “realistic tale” or a “thick description” (Flyvbjerg 2001) of the organizing campaign in Colombia. We adopt a “contextualist” perspective (Langley 1999) that emphasizes both events and contexts, with a particular focus in our case on what the actors do in these contexts and how they activate their capabilities to resolve the problems and tensions they face.

Case Study Background

FrenchCorp is not only strongly present in its original national market but has also developed significant operations in the rest of Europe, Asia, and Latin America. FrenchCorp’s internationalization strategy was greatly intensified from the mid-1990s onwards with acquisitions in more than twenty countries. It was during this period that FrenchCorp began its operations in Colombia, first establishing a site in the capital Bogotá, followed by expansion in the rest of the country, notably in the working class neighborhoods of medium-sized towns that its competitors had tended to shun.

The early 2000s saw increasing competitive and financial pressures on FrenchCorp, including the loss of market share to its competitors in its home market and a decline in its share price. In a major strategic shift, management sought to rationalize and concentrate its activities on markets in which it held a dominant competitive position, withdrawing from countries where it was not among the top three competitors. The company’s activities in the remaining countries were, for the most part, also downsized and a franchise model developed that resulted in worse working conditions for employees than in non-franchised stores. However, following other shifts in strategic direction and in response to shareholder pressures, it also strengthened its presence in some national markets, even entering a couple of new national markets.

The major players in FrenchCorp’s core industry are characterized by a high degree of geographical mobility. For example, the variations in FrenchCorp’s internationalization strategy, including its decision to refocus on its country of origin market, show the extent to which it can readily invest in new markets and, at the same time, relocate its activities or withdraw from other national markets. The configuration of the value chains of these large companies also makes them quite resistant to pressure from local unions because of the high number of establishments, including those of their competitors, in countries where they are located, and because of the diversity of their products. These companies also employ unskilled workers, often in precarious working conditions, with a high rate of employee turnover, which further undermines worker bargaining power.

The labor movement in Colombia is relatively weak with an estimated rate of unionization of 5% (Castano 2012). In their authoritative study of Latin America, Collier and Collier (2002) find that the Colombian labor movement is the weakest among the eight countries examined. They attribute this weakness to the small size of the labor movement, its division among several union confederations, and the presence of numerous independent unions (2007, p. 677–682). These independent unions have very limited resources to assist workers in core union functions such as organizing and bargaining. The weakness of the labor movement is also related to the labor law framework. First, Colombian labor law permits the coexistence of various forms of labor contracts, notably collective pacts as opposed to collective agreements (Castano 2012). If a union represents more than a third of employees in a company, then all workers are covered by a collective agreement. Otherwise, if the union does not meet this threshold, then only the unionized workers are covered by the collective agreement and the employer can establish a collective pact covering the non-unionized workers. Castano (2012, p. 13 ) argues that employers use these collective pacts to avoid trade unions. Second, Colombian labour law is permissive in that it is weakly enforced. Although freedom of association is constitutionally protected and unionized workers enjoy legal protections that prevent them from being fired simply for taking part in union activities, Colombia is often cited in reports by the ILO’s Freedom of Association Committee and by the International Trade Union Confederation (ITUC) as one of the countries in the world where trade union rights are most often violated.Footnote 1

Like a number of other French MNCs, FrenchCorp was an early signatory of an international framework agreement (IFA). In a symbolic exercise with the UNI Global Union, and in the presence of ILO leaders, FrenchCorp signed an IFA. While a very slight document, this IFA undertakes to apply ILO Conventions Nos. 87 to 98 and 135, meaning that workers are entitled to join the labor union of their choice, that they have the right to engage in collective bargaining, and that they and their representatives are to be protected against any act of discrimination infringing on their freedom to organize. FrenchCorp also undertook to respect union rights, to recognize fundamental worker rights and to condemn child labor so as to prevent slavery and forced labor (an obligation further extended to the company’s suppliers).

Since the early 2000s, UNI has focused on the implementation of this IFA in FrenchCorp subsidiaries. In 2009, it created an international alliance that included all national trade unions representing FrenchCorp workers in Europe, Latin America, and parts of Asia. The initial activities of the alliance were described as “modest” or even “stagnant” by most of the union representatives. Multiple reports confirm that inter-union conflicts in the home country (France) made this initiative quite dysfunctional from its inception. Domestic conflicts were simply transposed to another forum. According to one union observer:

the first year nothing happened, at least not much… We were hamstrung by the feuding between the French unions. We were never able to set up a steering committee for the alliance. Because we wanted to choose a few people. The French never managed to agree with each other, and that completely paralyzed the process.

In fact, the tensions between the French unions opened up space for the involvement of union representatives from Europe and Latin America. Despite the problems related to the alliance’s internal structure and its mode of operation, the UNI representatives were able to manage and contain internal conflicts, organize actions to support their members, and coordinate the involvement of several national unions in a spirit of transnational solidarity. The alliance organized two meetings, one in Europe in 2010 and, most importantly, the other in Brazil in 2011 when the participation of the French unions was more limited. These meetings marked a significant shift in the leadership of the alliance. A new “steering committee” was elected and included representatives of both UNI and the main unions of Italy, France, Spain, and Belgium. Decisions were taken to enhance the steering committee and to make it more active and more representative of the trade union members of the alliance. The participants laid the groundwork for an effective structure of inter-union dialogue at the international level, and establish the alliance’s short-term agenda—enabling the company’s Colombian workers to organize.

FrenchCorp Organizing Campaign in Colombia

Despite FrenchCorp’s IFA, Colombian workers could not exercise their rights to organize and engage in collective bargaining. No fewer than three attempts to organize a trade union had been led by workers but each attempt had failed, with workers active in the organizing campaigns being fired. UNI was not directly involved in these efforts to form a union. Aware of these failed attempts UNI established an organizing campaign that targeted multiple actions at different levels. In 2010, workers from FrenchCorp initiated another organizing campaign to unionize the thousands of FrenchCorp employees in Colombia under one independent national union.

This new campaign was difficult notably because FrenchCorp’s facilities were widely dispersed across Colombia. As expressed by one local organizer:

There are… [dozens of sites] in the country, so it was really hard to see how we were going to be able to get together to create a single union. It was really challenging because we’re in places that are very far removed from each other. If you look at a map of Colombia, we have one [site] in the north, another in the south, another in the east and another in the west, so, to create a single union, we needed support and a strong organization.

The Escuela nacional sindical (ENS) played a pivotal role in this organizing campaign. The ENS is an NGO (non-governmental organization) that works with Colombian trade unions and acts primarily as a research center, providing training for union representatives and supporting workers in their efforts to organize. ENS organized many meetings with workers, which contributed to the creation of a network of activists across the company. One UNI Americas representative acknowledged the contribution made by the ENS as follows:

You can’t forget the tremendous work accomplished by the Colombians themselves. The ENS was really behind all that. UNI came in afterwards… you have to remember that the ENS prepared the ground by identifying a potential network of unionists within FrenchCorp.

The organizing campaign gained momentum when FrenchCorp’s management in Colombia signed an agreement with UNI acknowledging the right of workers to establish an “autonomous and democratic” union organization and to freely join the union. According to one Colombian union representative, this agreement provided the framework and the protections needed to give FrenchCorp workers in Colombia real access to unionization:

The agreement is actually a national-level agreement. It fits into the international framework agreement signed between the company and UNI but it deals more specifically with the situation and obligations of the employer in Colombia. It was because of this national agreement that we were able to form a union in Colombia. (…) You must not forget that this is an anti-union country. (…) Mobilizing to create a single union among […] workers was no easy task!

The national agreement on the right to organize thus provided union activists with a narrative to frame positively the organizing campaign and to convince the workers to join the union. Unionization was not defined as a battle against the employer but rather as a process of social dialogue between workers and management. The agreement also favored the direct participation of various international and national trade unionists involved in the UNI international alliance. In particular, an Argentinian service sector union played a significant role. Representatives from this union traveled to Colombia on several occasions, sometimes for extended periods of time, to help coordinate the organizing campaign on the ground. One trade unionist from FrenchCorp in Argentina described the experience as follows:

I was involved in the entire organizing campaign in Colombia and I travelled all across the country. You have to remember that we were in places where unions are considered to be bad and where unionists are often assassinated. The discussions led by UNI and some European unions with the management of FrenchCorp opened the door for us in the worksites, and we were able to communicate the message to workers that unions give workers a “voice.”

This organizing campaign involved intense work on the ground by activists from the ENS and Argentina and from FrenchCorp workers in Colombia. This group of activists organized 180 meetings with workers and 348 visits to various facilities of FrenchCorp across Colombia. Women played a vital role in the organizing campaign. The trade union was led by a woman working as an office staffer in the HR department of FrenchCorp in Colombia. Moreover, the trade union was structured to achieve gender parity: at the national level and in each municipality, the ten-member executive committee of the trade union was distributed equally between men and women.

During the campaign, another event galvanized Colombian activists. At a meeting of the UNI international alliance held in Brazil, the Brazilian workers of FrenchCorp postponed a demonstration against the employer that might have jeopardized the agreement in Colombia and the process of social dialogue. The employer threatened to halt the process of social dialogue with Colombian workers if the Brazilian workers went ahead with the demonstration. This movement of solidarity, which sparked considerable debate within the Brazilian union, exerted a positive influence on the campaign as highlighted by this Colombian union representative:

I was in Brazil and I brought back all the information and I told my Colombian colleagues, “We have the support of friends, Brazilian colleagues from FrenchCorp-Brazil. They cancelled a demonstration to make unionization possible in Colombia.” We have colleagues who don’t even know us, who’ve never even met us, but they’re there and they’re supporting us. This gave the workers a boost in the sense that we could say, “We’re not alone. There are other people behind us.”

In addition to this support from the international union network, the Colombian campaign was also able to secure support from political leaders in Colombia. In particular, a former leader of the Central Unitaria de Trabajadores (CUT) in the 1980s who held an important position within the government publicly supported the workers in their quest to create a national trade union and to negotiate a collective agreement with FrenchCorp.

Ultimately, more than a third of the workers concerned joined the union. According to Colombian labor law, meeting this threshold meant that the union had the right to engage in collective bargaining with the employer and to apply the agreement to all the workers of FrenchCorp. The ENS continued to support the trade union by providing several training sessions on collective bargaining. Representatives from ENS and from UNI were present at the bargaining table, mainly as observers, to support the Colombian union executive and ensure that the process ran smoothly. Management and trade union representatives concluded the collective agreement in early 2012 after a month and half of negotiations. This first collective agreement improved working conditions, in particular with regard to working hours and overtime. Perhaps most importantly in the view of the leaders, the contract secured the position of the union through union leave and union recognition provisions. In 57 of 72 workplaces, the union had at least one officially recognized union representative. The number of delegates varied according to the number of workers: from one delegate in workplaces with less than 50 workers to three delegates in workplaces with more than 100 workers. According to the trade union, this translated into 180 union leaders with trade union immunity. As explained by one trade unionist, “Union immunity provides job protection to these workers in their workplaces.”

During the negotiations, the number of union members fell below the threshold of a third of the workforce. This was attributed to the high turnover rate at FrenchCorp establishments. In relation to one of the features of Colombian labor law highlighted above, this meant that the collective contract only applied to unionized workers and the non-unionized were covered by a collective pact established unilaterally by the employer. However, Colombian labor law further stipulates that such pacts cannot establish working conditions superior to those negotiated with the union. This loss in membership had a significant effect on union capacity but the Colombian union nonetheless remains strongly present in FrenchCorp’s Colombian operations. Further, according to a Colombian unionist, this presence of the union provides voice to workers:

Before, workers didn’t have any say in anything. You couldn’t speak up or you’d have to leave the company. Now, employees speak up and the employer has to listen.

Several factors account for the success of this organizing campaign. First, the Colombian activists showed a clear sense of cohesion and built a grass roots network of delegates, predicated on the principle of gender parity, that was grounded in worker realities and problems. Second, these activists embraced a narrative frame that emphasized social dialogue, worker voice, and international solidarity. They were able to show workers that they were “not alone.” The action undertaken by the Brazilian workers to support their cause acted as a “fortifying myth,” exerting a positive influence and reinforcing their perception of being part of a larger movement. By presenting themselves as an independent union affiliated with an international union, the Colombian union representatives were able to use their ties with an international actor to strengthen their legitimacy domestically. Third, Colombian activists had the capabilities to tap into and mobilize their networks by seeking the support of local actors. In particular, ENS, the domestic NGO, played a crucial role by identifying a network of union activists within the company while also providing training in collective bargaining to Colombian union leaders. These leaders were able to establish relations at multiple levels while remaining rooted at the local level. Fourth, the UNI international alliance played a decisive role. UNI created a Solidarity Fund dedicated exclusively to this campaign and set up a coordination structure that engaged actors at various levels. As will become apparent in the following section, UNI leaders were able to frame a new narrative and to mediate conflicting demands and expectations, and to articulate actions across space and scale.

UNI International Alliance Strategy

The UNI approach to FrenchCorp is built around the creative tension between two somewhat different union approaches, one centered on social dialogue and the other on trade union organizing campaigns. The emphasis on social dialogue has a strong tradition among UNI representatives from Northern Europe and is championed by leaders more inclined to engage in an “inside game” with FrenchCorp managers. They draw on their proximity and close personal relations to get things done, quite often to the exclusion of national union affiliates. Indeed, the inception of FrenchCorp’s IFA can be attributed to such an approach. Moreover, international union representatives appeared conscious of the importance of maintaining good relations with FrenchCorp because of its willingness to engage in social dialogue with union representatives and of the ways that social dialogue is integral to the firm’s overall social relations strategy. It is difficult for international union representatives to make the argument that transnational social dialogue is the “new normal” if they cannot point to MNCs like FrenchCorp as illustrative of this new web of social relations. The second approach emphasizes the importance of “organizing campaigns” in which national unions need to be networked and mobilized to exert pressure on the employer, be it to organize unions or to enhance local bargaining power. Whereas the signature of a FrenchCorp’s IFA was clearly associated with the first approach, the evolution of the international network of FrenchCorp delegates was increasingly associated with the second. UNI representatives had to reconcile these two approaches and to alleviate some of the sharper conflicts among contending unions within the alliance regarding its strategic orientations. These two approaches are not mutually exclusive but rather formed a backdrop to ongoing strategic discussions within and between UNI affiliates about how to conduct relations with FrenchCorp.

The creative tension between these two philosophical approaches was readily apparent during the organizing campaign in Colombia. UNI representatives used their privileged contacts with FrenchCorp HQ to put pressure on Colombian managers to apply the IFA and to respect the worker rights to freedom of association. This was seen to be a building block in the creation of social dialogue between workers and managers in Colombia, a discourse that was completely internalized by Colombian trade unionists. However, during the 2011 international union alliance meeting in Brazil, workers were eager to organize a public demonstration to protest against their working conditions and to exert pressure on FrenchCorp local managers in Brazil. HQ managers threatened to withdraw from the discussions in Colombia if Brazilian workers organized a protest movement targeting the FrenchCorp activities in Brazil. Given the extensive efforts that had been invested in organizing Colombian workers, several representatives from UNI at various levels urged the Brazilian union to postpone the protest action. Representatives from UNI Global and UNI America had to navigate around these conflicts and mediate between conflicting demands while at the same time pursuing social dialogue with FrenchCorp’s management. A representative from UNI Americas hailed this gesture of inter-union cooperation on the part of the Brazilian union while nevertheless referring to the paradox involved:

So we had to convince the Brazilians not to hold the demonstration and that was a very tense moment. It was the first time I’d had to convince unionists not to protest. That’s the position I had to defend. (…) So we “succeeded,” so to speak, and the protest didn’t go ahead. It was a victory for labor too—being able to convince our Brazilian colleagues not to organize a demonstration against FrenchCorp on the day before their Congress.

This episode also highlights the capacity of UNI to establish channels of communication not only between unions in different countries and at various levels (international, national, local) but also with management at the HQ and national levels. According to a UNI representative, during this campaign they set up an operational structure where the actions of trade union representatives located at different scales were interconnected:

These structures were completely interconnected and each one played its role perfectly. UNI Americas on the ground, close to the Colombian workers, UNI Europa vis-à-vis the European Works Council, with the strengths of European unions, and UNI Global, which established contact with the national management in Colombia. So each actor played its role (…) All the regional and international structures were operational and coordinated.

On top of funding and setting up a coordination structure involving a processes of deliberation and discussion between unions representing FrenchCorp workers, UNI developed a narrative that integrated the respect of worker rights and social dialogue with the employer. UNI representatives were able to convince the members of the alliance of the need to support the struggle of Colombian workers who wished to form a union and of the importance of making this goal a collective project for the alliance. The particular irony of this project was that it provided a degree of unity to a number of home country unions otherwise torn by internecine conflicts. This narrative was also embraced by FrenchCorp Colombian workers, making transnational solidarity a key vector for the renewal and strengthening of worker capacity at the local level.

During this organizing campaign, trade unionists at various levels learned through experimentation. Union representatives became familiar with new repertoires of action, learned to coordinate action at various levels and to better manage the tensions and conflicts between members. Learning and deploying these strategic capabilities constituted important levers in the union organizing campaign in Colombia.

FrenchCorp Strategy Towards IFAs and CSR

The interventions of FrenchCorp HQ are considered as decisive moments in the organizing campaign of Colombian workers. The HR director from FrenchCorp HQ and UNI representatives met several times with Colombian managers to support the implementation of the IFA provisions with regard to freedom of association. A UNI representative acknowledges that:

the role of the human resources director from the FrenchCorp group, who also travelled to Colombia on several occasions, was necessary, both to reassure the local management in Colombia, and to ensure that they were really listening to us.

These interventions gave much legitimacy to the organizing campaign of Colombian workers and underscored the importance of the IFA for HQ managers. The IFA is indeed considered by them as a useful pedagogical device to advance their conception of CSR and social dialogue in a context where subsidiary managers are otherwise relatively autonomous. HQ can hardly use sanctions to compel local managers to apply the IFA. One previous joint intervention by HQ managers and UNI representatives in 2010 in Eastern Europe to apply the IFA had been unsuccessful. Managers at subsidiaries felt that they were accountable from an economic point of view. They had financial goals to achieve but they felt HR issues and worker rights had to be dealt with locally not at the HQ level. Another joint intervention in 2011 elsewhere in Eastern Europe was more successful. FrenchCorp HQ management and UNI representative were able to convince subsidiary managers to apply the IFA and allow workers to set up a trade union. Regarding this particular joint intervention, a UNI representative stresses:

The IFA is important for FrenchCorp’s management at the global level to be able to “sell” a number of things. Headquarters management sometimes has to convince national leaders of certain things, and the way to do it is with the framework agreement, saying, “Look, we committed as a company, we have to meet our commitments.”

These examples highlight how HQ managers at FrenchCorp cannot impose the implementation of the IFA but rather have to obtain the consent of local managers through persuasion. In this process of persuasion, the approach of management to CSR and social dialogue is paramount. FrenchCorp has specialist functions in CSR and the company has long argued that it is a good corporate citizen and that CSR is integral to its company philosophy. HQ managers also emphasize the importance of setting benchmarks for CSR. Over the years, it has invested in protecting the bio-diversity, in building partnerships with suppliers, and in protecting human rights and the environment. For example, in Colombia, FrenchCorp has been involved in programs facilitating access to education and in projects encouraging farmers to implement sustainable ecological agriculture methods. More generally, the efforts to protect and promote the reputation of its supply chain activities are seen as integral to its brand.

FrenchCorp is also seen as a leader in social dialogue. This tradition is legally embedded in institutional arrangements in its home market, with multiple levels of legally enacted social dialogue at its sites and for the company as a whole in France. Moreover, this approach has been transposed to supranational levels as the company expanded into other national markets. A particular example is that FrenchCorp created mechanisms for European-level social dialogue [a European Information and Consultation Committee (CICE)] before the EU directive established a legal requirement to establish a European Works Council (EWC).

FrenchCorp’s approach to social dialogue is also part of a paternalistic tradition. This is highlighted by the influence of pro-company unions within the structure of social dialogue. This is a source of bitter divisions not only between French unions (notably between FO and the CGT) but also between Spanish unions and some Latin American unions. The transition from a CICE to a EWC also illustrates this point. While FrenchCorp was an early proponent of European social dialogue, it really preferred to share information, as opposed to consult or (beyond EWC requirements) negotiate. In establishing FrenchCorp’s structure for social dialogue at European level, UNI and other union representatives were involved in a trade-off: the CICE was a venue for information-sharing and social dialogue, in response to management preference; but only UNI affiliates would be present, thus ensuring a strong union presence and effectively sidelining some pro-company unions. This insider deal, which appeared to suit both senior union leaders and FrenchCorp management, remained in place for many years until a change in UNI’s approach opened up the possibility of consultation, however weak, according to EWC standards. It came at the cost of opening up participation to a wider range of union and non-union representatives, albeit in a fairly constrained role where UNI representatives continued to drive the dialogue. FrenchCorp representatives lamented their inability to use the EWC as a kind of in-house, informal consultation forum. One employer representative noted:

What was originally interesting about how our CICE functioned was that it created a place for exchange that was more informal than anything else. Today, the obligation to consult implies several things, including sending files, mobilizing experts, disseminating opinions, etc. What added value does a delegate from Cyprus provide when we’re discussing reorganization in Spain? The employer is supposed to take into account a viewpoint when many delegates are very far removed from that local reality.

It is probably fair to say that despite FrenchCorp management’s marked preference for a previous model of social dialogue, it was not unwelcome to have a venue, at a distance from the bitter inter-union rivalries within its home market, where it was possible to discuss with FrenchCorp unions. One international union representative noted that management could effectively use its influence and its union allies in the turf wars in France to “padlock” EWC discussions. It should also be emphasized that FrenchCorp greatly facilitated the creation of the international union alliance, supporting the participation of other national unions in a meeting held in tandem with the annual EWC meeting. Moreover, FrenchCorp managers actually participated in a segment of the alliance meetings and regretted that they did not participate in the entire meeting.

Social dialogue and CSR are closely coupled and shape the cognitive frame used by HQ managers to influence both local managers’ practices and trade union strategies. FrenchCorp managers had already proved adept at using social dialogue to pursue their objectives at the European level, so it was not a stretch to pursue these objectives beyond the confines of legislated institutional obligations in France and in the EU. The threat to withdraw from the discussion in Colombia unless the Brazilian union cancelled its protest action highlights how the HQ uses IFA mechanisms to define and engage in social dialogue on its own terms. HQ has also used the renegotiation of the IFA as a possible bargaining chip in national-level negotiations in France. FrenchCorp maintained that the unwillingness of one of its French unions to sign a national agreement was a sufficient basis for no longer engaging in the re-negotiation of the IFA. FrenchCorp’s strategic use of the IFA demonstrates how these instruments of social dialogue can be subject to local and national bargaining on a continuing basis.

The employer’s strategy with regard to social dialogue thus appears to be overlaid with a variety of cognitive and strategic considerations. The company had certainly demonstrated a degree of vitality in the area of social responsibility, in particular, being one of the first companies to sign an IFA agreement and engaging in social dialogue with many different stakeholders about the activities of its supply chain. This cognitive frame is mobilized by various actors to shape the terms of their relationships and reinforce the effectiveness of the IFA. FrenchCorp managers have also engaged in institutional arbitrage to influence trade unions strategies at national and transnational levels using both persuasion and threats to pursue their objectives. Their capacity to play with institutional processes enabled them to reinforce social dialogue through the implementation of the IFA.

Discussion

As argued above, the Colombian unionization campaign presented in this paper represents a critical case for a number of reasons. The campaign focused on a large French MNC in the service sector, which is a difficult industry to organize. Workers had weak bargaining power because their low levels of skills and poor working conditions meant that employee turnover was high. Moreover, Colombia is considered by international observers as one of the worst countries in the world for labor rights, often coming at tragic cost for those seeking to exercise their basic rights to associate, organize, and bargain collectively. To add to this rather bleak picture, facing declining market share its home market, FrenchCorp was in the midst of major restructuring strategies, which could have weakened its engagement to CSR. So how then might we explain the success of this organizing campaign and the implementation of FrenchCorp’s IFA with regard to its Colombian workers?

First our findings point to the decisive role played by management capabilities (Fenton-O’Creevy et al. 2011; Ferner et al. 2012; Kristensen and Zetlin 2005). Previous research has highlighted the importance of management approaches (Williams et al. 2015; Niforou 2014). Niforou (2014) emphasizes the commitment of managers and the difficulty for HQs to impose compliance with IFAs in Latin America. Our results add to this literature by showing how HQ managers can counter this resistance through the mobilization of three sets of capabilities: framing, bridging, and the capability for collaborative action.

For FrenchCorp, the IFA represented a strategic opportunity, which was fully integrated into its cognitive frame on CSR and on the importance of social dialogue. HQ managers used this cognitive frame to convince subsidiary managers to apply the IFA. The IFA is not only defined as a pedagogical tool but also as a transfer mechanism to apply best CSR and social dialogue practices, as defined by HQ managers, beyond home country and EU borders. However, this is a contested terrain where subsidiary managers can also resist the implementation of an IFA. Indeed our 2012 critical case was in fact the third attempt by Colombian workers to form a union. The same HQ managers have encountered both failure and success in persuading subsidiary managers from East European countries to adhere to their vision of CSR and social dialogue. HQ managers also used their bridging capabilities to intervene in the decision-making process in Colombia and influence the outcome and the process of negotiation between management and workers. This intervention had a considerable impact on the organizing campaign. HQ managers not only used their strategic position to influence subsidiary manager behavior with regard to workers rights but also to shape trade union strategies. Hence HQ managers, themselves embedded in a web of institutional arrangements in France and in the EU, used their strategic position at the interface of these multiple institutional terrains to engage in arbitrage and force the union to negotiate on HQ management terms. Finally, HQ managers were able to meditate conflicts of interest and balance cooperation and conflict. They threatened to halt the discussion in Colombia if Brazilian workers went ahead with their means of pressure. They also refuse to renegotiate the IFA in an unsuccessful effort to force trade unions in France to change their approach to domestic bargaining.

Second, the mobilization of trade union capabilities at different levels (Fairbrother 2013; Lévesque and Murray 2010a, 2010b; Turnbull 2006) was instrumental in implementing the IFA and building a cohesive local union organization in Colombia. Our findings are consistent with an emerging literature that points to the development of capabilities, in particular bridging (Cotton and Royle 2014; Helfen and Fichter 2013), intermediating (Fichter et al. 2011) and framing capabilities (Fichter and McCallum 2015; Dehnen and Pries 2014). Our results support these findings.

UNI Global offered an inclusive narrative frame, linking worker rights to both organizing and social dialogue. This narrative brought together members of the alliance that otherwise would not have worked together. UNI America contributed to the development of the narrative frame by insisting on the importance of building international solidarity to increase Latin America trade union capacity. The Colombian activists were also able to convey a mobilizing narrative to their colleagues by putting forward the idea of obtaining a “voice” in the workplace in a context of social dialogue with the employer. It thus integrated the narrative frame of UNI Global but adapted it to the reality of Colombian workers. UNI Global also played a key bridging role both in creating channels of communication between unions in different countries and at various levels (international, national, local) and in managing dialogue with the FrenchCorp’s HQ management in order to convince them to intervene in national subsidiaries at a critical time. UNI America also played an important brokerage role between UNI and the Colombian union throughout the organizing campaign. Colombian activists were also able to build relations within the community and secure political allies. By presenting themselves as an independent union affiliated with an international union, these activists were able to use their ties with an international actor to strengthen their legitimacy domestically. They established relations at multiple levels while remaining rooted at the local level. The UNI representatives also had to use their intermediating capabilities, seeking to alleviate some of the sharper conflicts among contending unions within the alliance regarding its strategic orientation. Furthermore, the representatives from both UNI America and UNI Europa were very active in mediating conflicting demands notably, at the behest of an understanding between UNI leadership and FrenchCorp management, by convincing the Brazilian union to postpone its collective action in order to support the Colombian workers organizing campaign. Colombian activists with the support of ENS, the Colombian NGO, also demonstrated their capacity to mediate tensions between workers and with local managers.

Third, the establishment of multilevel approach to foster the implementation of IFAs needs to be supported by the deployment of capabilities across all levels. Much of the literature on IFAs is quite critical about the centric or top-down approach favored by several GUFs in the negotiation and implementation of IFAs (Cotton and Gumbrell-McCormick 2012; Williams et al. 2015). The literature points to the importance of having a bottom-up approach involving a strong participation from local actors (Cotton and Royle 2014; Helfen and Fichter 2013; Fichter et al. 2011). Our findings point in the same direction and show the vital role played by local actors. However, the strategy observed is not simply a bottom-up approach but rather a combination of both bottom-up and top-down approaches. The actions of management and trade unions representatives at various levels were interconnected in complex ways. Our findings further suggest that development of a multilevel approach towards IFAs requires the sharing of decision-making competencies at different levels and the dissemination of management and trade union capabilities across levels. These capabilities are intertwined and reinforce each other.

Conclusion

Subject to the limits of our case-study methodology, this study adds to a growing body of research showing how in countries with permissive national labor laws the mobilization of IFAs can ensure the recognition of worker rights to form a union and negotiate a collective contract. In such cases, the existence of an IFA is a necessary but not a sufficient condition for the respect of worker rights. Management and worker representatives at multiple levels must also mobilize their sources of power and deploy their capabilities to ensure the effective implementation of IFAs. That is the key to understanding whether or not an IFA is effective.

This paper contributes to the literature on CSR by analyzing how actors engage in institutionalized dialogue to reinforce worker rights (Campbell 2007). It shows that institutions beyond the market are not only necessary to ensure that firms behave in socially responsible ways but that MNCs need to engage in institutionalized dialogue with local managers and worker representatives. The common cognitive frame put forward by both FrenchCorp management and UNI representatives started from social dialogue and the promotion of worker rights. According to this narrative, worker rights and social dialogue are bound together and constitute the interpretative and action frames that both management and workers representatives mobilize to define and legitimate their repertoires of action on CSR. This does not mean that there are not any conflicts or tensions between social dialogue and the promotion of worker rights. Our results highlight several points of tension within management, within trade unions and between management and trade unions. The crucial point is that this narrative frame becomes a strategic compass through which actors assess patterns of behavior and relationships.

This paper also contributes to the literature on multilevel governance in the fields of international worker rights and industrial relations by highlighting how levels of regulation (local, national, regional, and supranational) and the strategies of actors are interconnected. The distinction between a bottom-up and top-down strategy does not capture the nature of the relationships between actors. It is a combination of both where, under certain circumstances and at some points, the dominant actor can be located at different levels. In our study, HQ managers sometimes played a decisive role, while at other times it was either a Colombian NGO or UNI America or UNI Global. Such linkages require the dissemination of capabilities across an entire governance structure at multiple levels.

Finally, this paper highlights the conditions for the emergence of institutionalized social dialogue on CSR. The presence of both managers and trade union representatives with a set of capabilities appears to be an inescapable ingredient for the emergence of such dialogue. Our findings stress the crucial role of actors’ strategic capabilities in framing a new shared narrative that pulls together CSR and social dialogue, in connecting actors and institutions across various levels (bridging capabilities), and in arbitrating and reconciling conflicting interests (intermediating capabilities). It is the combination of these capabilities, on the part of both management and trade unions representatives, that provides the conditions for institutionalized social dialogue on CSR and, in the context of IFAs at least, the effective implementation of worker rights.