Abstract
We examine the effects of State-led decentralization of collective bargaining on workplace employment relations. After exploring the development of company-level collective bargaining in France in the broader European context, we analyse intra-national variations in social dialogue practices across companies and workplaces, and propose a classification of four workplace clusters. As the decentralization process varies in its extent and in nature between types of workplace, we argue that the diversity of workplace practices is best explained by contrasting reactions to legal requirements. Nevertheless, these requirements have standardizing effects on certain bargaining practices, which might in turn lead to the impoverishment of the collective bargaining process.
Introduction
Pressure to decentralize collective bargaining has increased in Europe since the 2008 economic crisis (Leonardi and Pedersini, 2018), with ‘incremental corrosion’ in Northern countries and ‘frontal assault on multi-employer bargaining arrangements’ in Southern ones (Marginson, 2015). The differences lie not only in the extent of the changes but also in the way they were introduced, on a ‘concerted’, ‘contested’ or ‘imposed’ basis. The fact that France is one of the countries where changes have been concerted between employers and trade unions should not mislead us, however. As Marginson (2015) argues, ‘influence of the state in promoting further decentralization is apparent’ (p. 103). The State-led nature of French decentralization is a well-known feature (Howell, 2009). France is thus a pertinent case for analysing the implications of this specific form of decentralization for regulating employment relations at workplace level.
As Ilsøe et al. (2017) note in their study of the Australian and Danish cases, decentralization has taken two broadly different paths, either through state intervention resulting in increased legislative or regulatory control over both the processes and substantive outcomes of agreement making, or through the more agreement-based framework for change typical of the Nordic countries. (p. 14)
Successive governments in France have presented the ‘Danish model’ as an example when undertaking labour market and labour law reforms (Rey, 2011), but the similarities with the Australian case seem greater. In contrast to Denmark, where ‘parties themselves articulate the framework within which local negotiations occur’, in Australia ‘the legislative approach to decentralization is articulated by a third party, the state’, entailing a risk that this approach ‘to workplace bargaining can undermine the very dynamics of the process’ (Ilsøe et al., 2017: 15). Our study shows that the French ‘legally regulated’ shift towards more decentralized collective bargaining has increased the control over the process and outcomes of negotiating practices, making it more difficult for local social partners to be fully involved in collective bargaining. At the micro level, they have been equipped with new and innovative instruments for bargaining (Jobert, 2013), while experiencing an increasingly limited autonomy inasmuch as economic pressures have tended to restrict choices (Koukiadaki et al., 2016) and legal directions have paved the way.
Against this background, the article examines the rise of company-level collective bargaining to question how far the decentralization process has really gone and how it is shaped by its State-led nature. Acknowledging the ‘need for a workplace perspective’ (Amossé et al., 2016), it draws evidence from a study based on the combined analysis of two sets of data (Béthoux and Mias, 2014): a quantitative survey conducted in 2010-2011 by the French Ministry of Labour involving individual employees, management and employee representatives from more than 4,000 establishments, and 15 in-depth case studies that were carried out at the workplace level in various socio-productive contexts between 2012 and 2014.
We start by examining the general patterns of negotiations in contemporary France in the broader European context. Then, after presenting the data and research design, we investigate the effects of the State-led decentralization of collective bargaining, making a twofold contribution. First, we analyse intra-national variations in social dialogue practices across companies and workplaces, agreeing that ‘the idea of homogeneous national industrial relations systems was always in important respects misleading, even in countries with extensive legal regulation such as France and Germany’ (Gumbrell-McCormick and Hyman, 2006: 488). Using the theoretical framework offered by legal consciousness studies, we argue that because of contrasting reactions to legal requirements, the very existence of extensive legal regulation is a key to understanding the diversity of negotiating practices across companies and workplaces. Second, we show how the legalistic framework might also lead to the impoverishment of the bargaining process and its outcomes, thus making the decentralization process it supports incomplete and somewhat ineffective.
Decentralizing collective bargaining
We now explore the development of company-level collective bargaining in France in the broader European context. Briefly reviewing the process over time, we recall the pivotal role of the State in promoting the decentralization process, showing that the growing number of agreements concluded at company level is not merely a quantitative process: it is closely related to a more complex shift in regulating employment relations and defining labour rules.
The rise of company-level bargaining
Has France experienced a ‘steady decentralization’ as in Austria, Finland and the Benelux (Voss et al., 2015), or an ‘escalating process’ (Rehfeldt and Vincent, 2018)? The last decade has certainly seen an accelerating pace of reforms aimed at redefining the rules of the game, with major legislation in 2008, 2013, 2015, 2016, and 2017. The 2016 El Khomri law has attracted most attention because of the vehement and multi-faceted reactions it provoked (Béroud, 2018). The merits of further decentralization of collective bargaining are under close scrutiny, as are the ways to manage and coordinate the process.
However, the rise of company bargaining is nothing new. From the early 1980s, the State has played a pivotal role in promoting company and workplace agreements on an increasing range of issues. Decentralization was first given a strong legal impetus in 1982 when the Auroux laws required employers to negotiate annually with union representatives on wages and the organization of working hours (Eyraud and Tchobanian, 1985). Reducing the maximum legal working week at the turn of the 21st century represented a second major step in the spread of company bargaining, with agreements being reached on how the organization of the new working week should be implemented (Boulin et al., 2004). Since then, compulsory bargaining at workplace or company level has been regularly extended to address a variety of qualitative issues, such as health benefits (2000), equal rights for men and women (2001), saving schemes (2001), employment of disabled people (2005), strategic workforce planning (2005), equal pay for men and women (2006), employment for older workers (2008) and saving schemes (2015).
As a consequence, the number of company agreements has risen sharply throughout the past 30 years (see Table 1), now covering over 60 percent of the workforce. Thus, France now has a higher diffusion of decentralized bargaining than in similar European multi-tier bargaining systems which were traditionally organized around the sectoral level: by contrast Germany and Italy have a medium level of diffusion and Spain a low one (Leonardi and Pedersini, 2018: 27).
Agreements concluded at company level in France (private sector).
Source: Ministère du Travail.
Furthermore, negotiations now normally result in an agreement (Pignoni and Raynaud, 2013). At company level, the propensity of unions to sign agreements now ranges from 84 percent for the CGT to 94 percent for the CFDT. This is particularly relevant for issues such as saving schemes or supplementary social protection (for which financial incentives are important), while it is less so for working conditions or technological and organizational change (which are not priority targets of the legal incentives to negotiate).
The rise in the number of company agreements thus sits alongside the increasing range of issues that are dealt with at this level. But not only are the themes more numerous, they are also increasingly linked to each other in single package agreements. The development of ‘strategic workforce planning agreements’, that must be negotiated every 3 years in companies with over 300 employees, reflects the growing complexity involved in negotiating this new type of agreement (Oiry, 2013). In particular, the difficulties lie in the fact that it is less easy to assess what is gained and what is lost: as a consequence, explaining the trade-offs to employees becomes difficult for employee representatives. As the Danish case illustrates, even in the ‘best organized sectors … shop stewards … are faced by significant challenges of intra-organizational bargaining, when bargaining activity increases at company level’ (Ilsøe, 2012: 775–777).
Redefining the rules of the game
Promoting company-level bargaining is not only a question of widening the bargaining agenda through increasing mandatory bargaining; there are three other important features of this State-led process. First, since the late 1990s, collective bargaining itself has been made an instrument of public action in the French war against unemployment (Gautié, 2015). Second, to strengthen social partners’ legitimacy at company and workplace levels, the regulation of negotiations and agreements has been redefined. The 2008 law on union representativeness clearly reflects a decision to anchor union activity more firmly in the workplace (Bevort, 2008), in particular restricting the right to participate in company bargaining to unions which obtain at least 10 percent of the votes in workplace elections. To be valid, agreements must be signed by one or more unions gaining at least 50 percent of the combined votes at the elections. Yet in practice, collective bargaining remains weak in smaller companies where unions are not present (which is the case for 96 percent of those with fewer than 50 employees). The situation is regularly addressed in government reforms, for instance, allowing smaller employers to negotiate with non-union employee representatives or to submit agreements directly to an employee referendum.
Third and most importantly, reforms have gradually moved away from the ‘favourability principle’ once at the core of the French industrial relations system. Thanks to the extension mechanism, sectoral agreements had long been considered as defining rules that encompass all aspects of the labour contract: hence collective bargaining coverage in France ranks among the highest in Europe (over 90 percent), despite one of the lowest unionization rates (around 11 percent). In the 1990s, however, some employers argued that sectoral regulation restricted the autonomy of company bargaining and its capacity to respond to the new challenges of competitiveness (Sisson, 2013). Their lobbying first resulted in the 2004 Fillon law, which introduced the possibility of derogation from sectoral agreements, except for negotiations over minimum wages, job classifications, supplementary social protection and vocational training. This law made France a forerunner, given that in Europe ‘until the onset of the crisis, formal change in the hierarchy implied by the favourability principle was the exception rather than the rule’ (Marginson, 2015: 99). Through the 2000s, French governments introduced further provisions to exempt company bargaining from important aspects of the law or sectoral agreements (Amable, 2016). As a result, many sectoral agreements are no longer ‘specifying the main substantive standards’ but are ‘incomplete framework agreements … predicated on substantive variation between companies’ (Marginson, 2015: 101–102). The latest reforms, the 2017 Macron ordinances, introduce a new way of articulating the different bargaining levels: shifting further away from the traditional hierarchical structure by giving each level responsibility for a specific set of issues (Jobert et al., 2017); decentralization will no longer operate through ‘delegation’ but rather through ‘demarcation’ (Marginson and Galetto, 2016). In this three-tier system, company bargaining gains new ground, but it does so in a context where union representatives are less widespread at company level than in the Nordic countries in which demarcated arrangements prevail.
The State-led decentralization process has been seen as oriented towards the liberalization of employment relations institutions (Baccaro and Howell, 2017): ‘the core of the state’s strategy … was to create legal obligations inside the firm that would have the effect of generating autonomous and self-sustaining social dialogue that would in turn permit deregulation of the labour market’ (Howell, 2015: 9). However, there is ‘no direct link between reforms and outcomes’, as underlined by Leonardi and Pedersini (2018) in reference to the French, Italian and Spanish cases (pp. 25, 34): how the new instruments are used is not predefined. In a context of international economic competition, this is particularly evident regarding restructuring issues, which have attracted much attention in bargaining processes in many European countries (Pulignano and Stewart, 2012). In France in particular, such issues have fuelled the ‘exemption trend’ over the past 15 years. In some cases, employers took advantage to veer away from legal rules and keep strict control over both the bargaining and the restructuring processes, thus leading to clear instances of ‘bargaining in adversity’ (Gumbrell-McCormick and Hyman, 2014). In others, however, these new possibilities fostered innovative discussions at workplace level, and employee representatives were better equipped to anticipate and participate in the management of restructuring (Dupuis, 2017).
How to account for such differences? If the institutional framework in which collective bargaining takes place has been profoundly reshaped, the actual impact of the State-led decentralization process on workplace practices has yet to be understood.
Investigating workplace social dialogue practices
We draw on research examining social dialogue at workplace level and evaluating the influence of information, consultation and collective bargaining practices on job management and working conditions. Evidence comes from the 2010-2011 REPONSE survey, which amassed data for the period 2008-2010, and from a qualitative inquiry involving 15 case studies with interviews between early 2012 and early 2014.
Data and research design
Like the WERS surveys in the UK, REPONSE (Relations professionnelles et négociations d’entreprise) provides regular data on the interactions between management, employee representatives and employees at workplace level (Amossé et al., 2016; DARES, 2015). We draw on the 2010-2011 survey, the fourth in a series starting back in 1992 (the most recent was conducted in 2017). It covers activities in metropolitan France for private and semi-public sectors (with the exception of agriculture), civil service, social protection organizations and individual employers and was conducted in workplaces with 11 or more employees (the threshold was 20 employees in the previous surveys). Questions focus on the presence and nature of employee representative bodies, the nature and output of negotiation and discussion, the existence of collective conflicts, HRM policies, work organization, economic strategies and perceptions of the social climate. The 2010-2011 survey was conducted in 4,023 workplaces, interviewing 4,023 managers, 2,433 employee representatives and 11,378 employees.
To measure how social dialogue practices develop at workplace level, we used the data to build nine synthetic indicators related to employment, training, working conditions, technological and organizational change, activity of employee representative bodies, conflicts and social climate, satisfaction at work, management networking and union networking (Béthoux and Mias, 2014). The first four indicators measure the intensity of social dialogue and the importance of each theme in discussions between employer and employee representatives. Our decision to measure social dialogue intensity rather than collective bargaining dynamics as such is explained by the fact that in the REPONSE surveys, managers, employee representatives and employees were asked about ‘discussions or negotiations’ more broadly, not just about ‘negotiations’. It also reflects the increasing interconnection between information, consultation and negotiation practices in French workplaces. The ‘activity of employee representative bodies’ roughly corresponds to the time spent by representatives in their various activities, especially in contact with employees (gathering demands, welcoming new employees, campaigning with casual workers, providing legal assistance), but also in their discussions with the employer. The ‘conflicts and social climate’ indicator includes estimations of the social climate and its evolution by each actor, and a record of conflicts in the workplace in the last 3 years. The ‘satisfaction at work’ indicator combines questions addressed to employees on working conditions and job security. The last two indicators can be interpreted as proxies of the external resources that each actor can use for social dialogue (European, national, sectoral or local union structures for instance in the case of employee representatives).
Since the statistical analysis is based on questions selected in all three questionnaires (management representatives, employee representatives and individual employees), it was conducted on a sample of 2,286 workplaces in which at least one worker representative and one employee answered the questionnaire (there were no worker representatives interviewed in 1,590 workplaces and no employees in 147 workplaces). In order to give each indicator a mean score, the answers to these questions were coded into a dichotomous variable (‘yes’ or ‘no’). Thanks to k-means clustering, the 2,286 workplaces were then partitioned into four clusters of workplaces (each workplace belongs to the cluster with the nearest mean).
We followed two objectives in conducting the quantitative analysis. First, to assess the diverse ways of regulating local employment relations, considering that the decentralization process has often been viewed as expanding scope for local determination or variation. Second, to identify the 15 workplaces where case studies could represent the largest diversity of practices. These studies involved face-to-face interviews with management and employee representatives (union delegates, works committee members, health and safety committee members), as well as labour experts and labour inspectors in some cases. Combining the analysis of quantitative data and the interview material helped us to interpret the workplace classification that had first arisen from the statistical data.
The workplace classification
Four clusters of workplaces were identified according to their average scores for each indicator (see Table 2). There is first a sharp contrast between Clusters 1 and 2: the first combines intense social dialogue for all four items with the highest levels of industrial relations activity and conflicts, an average level of satisfaction at work and dense networks for both management and unions. The second displays low activity in terms of negotiation and conflicts and loose networks, but higher satisfaction. Cluster 1 includes a high proportion of large workplaces (46 percent have 500 or more employees), often belonging to large companies in manufacturing, network industries and financial services. By contrast, more than half of the workplaces in Cluster 2 have fewer than 50 employees and are independent companies operating in construction and commerce.
Scores for nine synthetic indicators, attributed to four clusters of workplaces.
Source: REPONSE Survey 2010-2011.
The picture for Clusters 3 and 4 is more complex. Social dialogue in Cluster 3 seems rather intense, especially for employment, as are the activity of representative bodies and the level of conflicts. Networking is denser for unions than management, and the level of satisfaction at work is the lowest. The highest satisfaction, associated with few conflicts, is in Cluster 4. Networking is here much denser for management than unions, and the intensity of social dialogue is contradictory: highest for training, high also for technological and organizational change, but lower for working conditions and for employment. Clusters 3 and 4 are quite well-balanced in terms of workforce size. Half the workplaces in Cluster 4 belong to a group of companies, and two-thirds in Cluster 3.
Four ways of regulating employment relations at workplace level
The classification of workplaces into four contrasting clusters confirms that the decentralization process varies in its extent and in nature between workplaces. The case studies helped us to define more accurately the four clusters, which encompass four different ways of regulating employment relations at workplace level.
Social dialogue is intense in Cluster 1 workplaces for various reasons: there is a dense network of employee representative institutions and a mutual recognition of social partners (in particular, employers recognize the positive role and functions that trade unions play at workplace level) and a good understanding by employee representatives of how information, consultation and negotiation operate. In certain circumstances, a general framework agreement brings coherence to the multitude of existing agreements. It is also significant that these workplaces have met their obligations under the different laws that enforce compulsory bargaining at company level. They are the ‘model pupils’ of State-supported social dialogue: actors in those workplaces see the law as a resource to be used to meet their interests and achieve their goals. Nevertheless, our Cluster 1 case studies reveal a lessening of discussions in recent years, with reduced mutual trust and a growing scepticism about certain achievements and practices. This is particularly so with employment issues. Whereas the possibilities of negotiating such issues at company level have been augmented over the years in France, as well as in other European countries (Freyssinet, 2015), we can observe an increasing difficulty in discussing or negotiating them, especially when management has opted for individual over collective procedures as traditionally discussed through works committees: as with the rupture conventionnelle procedure introduced in 2008 whereby an employer and an employee can mutually agree to end an employee’s contract. Several employee representatives therefore admitted that their relationship with management had deteriorated recently as unions were progressively sidelined in discussions over employment prospects. Where dialogue on employment has been hindered, there is sometimes a parallel shift towards more discussion and bargaining on working conditions, particularly psychosocial risks: this can be seen as a strategic move for employee representatives who are in search of issues on which they can have a say (Béthoux et al., 2015).
In contrast to Cluster 1, the paucity of social dialogue in Cluster 2 reflects the absence of employee representative institutions in a large number of workplaces: the legal thresholds explain the absence of work committees and délégués in companies with fewer than 50 employees. State-supported social dialogue has almost no impact in such workplaces. Beyond this size effect, three particular features distinguish social dialogue in Cluster 2. First, employee representatives tend to define themselves as a ‘bridge’ between management and employees, insisting on their intermediary position and their capacity to link the two parties. They facilitate communication in both directions by articulating what employees have to say, but also by transmitting managerial messages they consider legitimate. Second, management and employee representatives view collective organizations with suspicion, which explains the very low scores for management and union networking. Third, they share a core value as regards laws and the ‘excessive’ formalization of employment relations they are said to create. Formal social dialogue is viewed as an administrative requirement, too far removed from the usual concerns. Law is thus mainly considered as a constraint that actors try to avoid. Conversely, management and employee representatives are committed to informal and interpersonal arrangements. In such cases, authenticity is frequently at odds with legality. As a result, discussions in representative institutions mainly focus on individual cases, and neither bargaining nor collective deliberation is actually possible.
The statistical analysis gives Cluster 3 workplaces an average position compared with the scores of the first two clusters. This might suggest that Cluster 3 brings together average cases along a continuum of intensity in social dialogue and conflicts. Our analysis shows that this is not the case. Case studies reveal a deteriorating social climate, which rarely leads to forms of conflict (strikes, petitions, demonstrations) being recorded in the REPONSE survey. Tensions flare up, but conflicts and disagreements are not openly expressed. Workplaces display latent conflict and a correlative incapacity to resolve problems and difficulties. This deterioration of labour relations manifests itself in different ways: increased absenteeism, legal action, tensions and lack of confidence in employer-employee relations, but also in relations between trade unions and between employees themselves (more employees declared that they encountered tensions with other employees). In the workplaces we studied, this latent conflict can be interpreted as the collective reaction to the disruption of a previous order. Greater budget constraints, economic recession or an expected merger introduce a significant disturbance in the former balance and in some situations lead to a destabilization of the internal labour markets. If actors are convinced that their usual routines are inadequate in the face of this uncertain future, tensions grow until a profound change causes previous practices and representations to be abandoned. This latent conflict is also reflected in a not-so-shared vision of the situation, which in turn does not facilitate the bargaining process. In that context, employee representative institutions operate very formally: social partners conclude agreements without conviction or expectation. Legal requirements are met but law acts as a screen behind which conflicts are hidden and remain unspoken.
The last cluster of workplaces has an atypical profile, with a low level of conflict activity and a relatively intense social dialogue, in particular regarding training and technological and organizational change. Cluster 4 displays a strong commitment from employee representatives to make company institutions operate effectively. From this perspective, law is considered a necessary but not sufficient framework. The qualitative study reveals a particular feature of employment relations in this cluster, which we refer to as ‘company trade unionism’. As they deliberately distance themselves from external trade union structures and resources, employee representatives consider themselves in the service of the company which they tend to perceive as a common good or as a community of interest. In consequence, they tend to be ‘responsible’ in their use of regulatory instruments and to look for problem resolution and compromise (including informal discussion with managers) rather than choosing voice. They are also committed to maintaining close links with employees in order to articulate a genuinely collective point of view.
Each cluster can thus be defined by combining two main dimensions: the nature and the density of social dialogue practices. The nature of social dialogue practices differ across a spectrum that goes from workplaces where union rights are highly developed, industrial relations actors are well-trained and experienced, representative bodies are systematically established and legal instruments are regularly used (Clusters 1 and 3), to workplaces with much more informal (Cluster 2) or interpersonal (Cluster 4) workplace relations and which keep legal regulation in the background. The density of social dialogue practices encapsulates the frequency and effectiveness of discussions and negotiations, as well as the quality of their outcomes according to the interviewees. It contrasts a dense (Cluster 1) and productive (Cluster 4) social dialogue with a minimalist (Cluster 3) and deficient (Cluster 2) social dialogue. In reference to their positions related to these two dimensions, we name the clusters proactive, a-legalistic, formalistic and locally focused (see Table 3).
The four clusters and legal consciousness.
The legal structuring of workplace bargaining practices
How can these differences in bargaining practices be explained? What makes a workplace more likely to develop one particular way of regulating employment than another? While a combination of explanatory factors should be considered, we argue that the law and the way it is perceived by workplace actors play a decisive role. Factors such as the size of establishments, the economic context or the workforce composition explain some contrasting practices, but none can fully account for the differences between all four clusters of workplaces. By contrast, in a context based on extensive legal regulation, the local appropriation of legal rules acts as a structuring element. While it is not the only explanatory factor, how individuals relate to the law provides an understanding of the contrasting ways in which employment relations are regulated at workplace level. Variations in social dialogue practices are thus best explained by contrasting reactions to legal requirements as we move from one cluster of workplaces to the next. The case study analysis shows, however, that while legal incentives and obligations to negotiate provoke differentiated reactions, they also have standardizing effects on certain bargaining practices, which might lead to the impoverishment of the collective bargaining process.
The differentiating effect of legal consciousness
Among institutional factors, the existence of legal thresholds is commonly advanced to explain differences in bargaining practices. In France, the absence of certain employee representative bodies in establishments with fewer than 50 employees considerably limits the possibilities for negotiation and reduces the density of social dialogue in the same way. But the size effect only applies to Clusters 1 and 2. It is not relevant for understanding how social dialogue works in the two other clusters. Moreover, employment relations in small enterprises cannot be interpreted solely in terms of what is lacking: small enterprises are not ‘small large enterprises’ (Holten and Crouch, 2014). As shown above, in such cases social dialogue takes an original form, which cannot be interpreted only as a deficit in relation to the ‘norm’ in larger companies.
Socio-economic factors may explain some of the differences in our typology. The establishments belonging to Clusters 1 and 3 are experiencing deteriorated economic situations, while the number of employees is rather stable in Cluster 2 and is growing in the majority of the establishments in Cluster 4. The frequency and extent of redundancies explain certain features of employment relations in Cluster 1 workplaces, such as the intensity of social dialogue on employment, the high activity of employee representative bodies and the high level of conflict. In the four Cluster 3 cases we studied, the latent conflict can be linked to the deteriorating socio-economic environment in which the companies operate. In the two manufacturing companies (metal sector and petrochemicals), employment is falling and the establishments suffer from the restructuring strategies of the groups in which they are integrated. The two other establishments operate in the medical and medico-social sectors. Their managements struggle to cope with the constraints imposed by the redefinition of public health policies and cannot build coherent HR policies to meet them. Another socio-economic factor at play is the workforce composition. For instance engineers and managers are relatively more present in Cluster 4 workplaces, probably explaining the lower level of conflict and the higher intensity of discussions on training (which in France is oriented disproportionately to those who are already highly qualified). How important they may be, these socio-economic factors explain only some aspects of the employment relations observed in each cluster. They do not reflect the idiosyncratic consistency of each four of them.
Ewick and Silbey (1998) use the concept of ‘legal consciousness’ to analyse the understandings and meanings of law among social actors. ‘Law is experienced and interpreted by specific individuals as they engage, avoid, or resist the law and legal meanings’ (Silbey, 2008: 696). Likewise, we argue that the way management, employee representatives and employees regard the legal and institutional context is a key element in understanding the type of employment relations they develop at workplace level: all the more so in a context that has long involved strong State involvement with extensive legal regulation playing a structuring role. Cross-company variations in the perceptions of local bargaining actors ‘shape the extent to which [they] can leverage institutional resources in a strategic way’ (Paolucci, 2017: 16). The place given to law and the way it is used (or not) in employment relations vary considerably from one cluster of workplaces to another. Legal consciousness studies provide understandings of compliance, with legal requirements by focusing on the way actors perceive, define and construct their legal environment (Silbey, 2005). Drawing on this literature, we assess four ways of conceiving and acting, which in turn shape contrasting social dialogue practices at workplace level.
We suggested above that the four ways of regulating employment relations at the workplace were related to four distinct ways of perceiving law. In Cluster 1, actors see the law as a resource to be used to meet their interests, to achieve their goals and to gain power in the employment relation. It corresponds to the manner of constructing legality that Ewick and Silbey (1998) call ‘with the law’. In Cluster 2, law is considered a constraint that actors wish to avoid. Employment relations appear here constructed ‘against the law’, as local actors feel imprisoned by the law and caught up in an excessive formalization of human relations. In Cluster 3, where employee representative institutions operate very formally, legal requirements are met but law acts as a screen behind which conflicts are hidden. These cases could be related to an attitude ‘before the law’ insofar as the use of law is perceived as detached from everyday life at work, which is one of tensions and worries. In Cluster 4, law is considered a frame: necessary as it ensures a regular and normal functioning of the social dialogue; but insufficient to face the company’s challenges and to allow its socially harmonious development. Completing Ewick and Silbey’s threefold typology, based on our own material we describe this fourth attitude as being ‘beyond the law’ (see Table 3).
Alternatively a resource, a constraint, a screen and a frame, law is defined in very different ways when we move from one cluster of workplaces to another. Therefore, depending on how they are perceived, legal requirements and incentives to negotiate at workplace level have extremely different effects, from intensive mobilization of new room for manoeuvre to the avoidance of an additional bureaucratic constraint. Thus, our study shows that the State-led decentralization of collective bargaining is fully effective only in some workplaces, mostly in Cluster 1 and, to a lesser extent, in Cluster 4.
The standardizing effect of legal requirements
The recent State-led developments of company and workplace negotiations have also had standardizing effects on collective bargaining practices, as all establishments in Clusters 1, 3 and 4 and some in Cluster 2 are committed to complying with legal requirements. This carries demands to which local actors respond with a simplification of complexity and a mimetism that can be understood as a form of ‘coercive isomorphism’ (DiMaggio and Powell, 1983). Our case studies shed light on three connected elements resulting from this process. Under most circumstances, they limit the involvement of local actors in bargaining practices.
The first element relates to the guidelines imposed by the law in terms of bargaining schedule, content and outcomes. The bargaining process is bound by a strict framework, which may risk its impoverishment: compliance with the law takes precedence over other objectives (Mias, 2014). Negotiators admit that most agreements concluded in this way remain rather formal. This was true of most of our 15 case studies but was particularly significant in the ‘formalistic’ workplaces in Cluster 3. The statistics may indicate that social dialogue is quite intense in these cases (see Table 2), but fieldwork reveals that it is mainly formal and rather inefficient: issues are discussed and agreements are concluded, but in a degraded social environment with management and employee representatives neither believing in nor embracing what they are doing.
It would probably be a different kind of negotiation if the issue was emerging from the company’s situation; it would be treated differently. … I tend to think that all these compulsory negotiations have a counterproductive effect on social dialogue in the end. Because of financial penalties in particular, employers want to negotiate on these issues. But they do so at a forced march because they always launch the process just two weeks before the deadline. No one is really prepared for it. (Labour inspector involved in Cluster 3 companies)
This is reinforced by the second element, arising from an expanded bargaining agenda. With a greater number of compulsory bargaining rounds to complete, some interviewees felt that they did not have sufficient time, skills or activist resources to complete them all properly. In order to comply with legal requirements and to meet tight deadlines, they are tempted merely to duplicate the terms of the law or those defined by ‘good practice’ or found in ‘toolboxes’ provided by the law itself. In such instances, negotiation and standardization go hand in hand (Charpenel et al., 2017).
This is really the minimum. We are well aware of it, but we were not very motivated. As unionists, we have mostly followed what the employer proposed … It is my fault too because I guess I did not spend enough time preparing this negotiation [on young workers’ employment], discussing it with other colleagues to see what they had done and how we could make a more thorough proposal. (Union delegate, Urban planning consulting firm, Cluster 4)
The tendency to stick to formal or minimal agreements is particularly true of agreements regarding older workers’ employment (Farvaque, 2011). It is less so however for ‘strategic workforce planning’ agreements, for which we observe a greater variety of bargaining practices that seem quite creative in certain cases. Facing an extended bargaining agenda thus leads to uneven involvement in bargaining rounds: interviewees confess that some subjects are less worth fighting for than others.
A third element is at play. Most of these negotiations are oriented towards general aims such as maintaining older people’s employment, promoting equal rights for men and women or managing uncertain careers. Therefore, it is first necessary for bargainers to share a common understanding of the situation in their own workplace or company. Where a shared understanding is not reached, as is often the case in Cluster 3 cases and more and more so in Cluster 1, there is a risk that such agreements will lose any substance and simply become ‘empty shells’, as shown with equal opportunities policies in the UK (Hoque and Noon, 2004).
However, in some cases in the ‘proactive’ Cluster 1 and the ‘locally focused’ Cluster 4, ‘social dialogue has already broken free of the confines of traditional forms of negotiation, bringing a deliberative component’ into the picture that could help renew bargaining practices at workplace level (De Munck et al., 2012: 15). This rests in particular on more continuous exchanges between negotiation on the one hand and information and consultation on the other. Thanks to a dense network of long-established institutions at the workplace (Cluster 1) or to the strong internal commitment from employee representatives (Cluster 4), what is addressed by union delegates and by elected representatives, such as works committee or health and safety committee members, tend to overlap. This is reinforced by the fact that most legally underpinned agreements define new schemes that require follow-up procedures in which union signatories and elected delegates are jointly involved. This fosters more collaborative work between the various employee representatives in a given workplace, thus increasing their political capacity (Didry, 2013). However, the recent labour law reforms have also redefined the role and status of workplace employee representatives through the imposed merger of délégués, works committees and health and safety committees. This might hinder this emerging capacity and reinforce the standardizing effect.
Conclusion
The French case illustrates the ambiguities of the decentralization trend in a context marked not only by high unemployment, but also by important institutional reforms that challenge the autonomy of industrial relations actors. The State-led decentralization process that started in the early 1980s has been clearly reinforced over the last decade. Indisputably, it led to a downwards shift in the ‘actual level of collective bargaining’, but counteracted at the same time by changes in different directions. The intensification of national social dialogue since 2008 is one of these, through the negotiation of a series of national cross-sectoral agreements (accords nationaux interprofessionnels) at the centralized level. In this perspective, France stands, like Spain, as an exception to ‘the dominant trend of associational regulation decline’ in Europe (Meardi, 2018: 647), making it difficult to argue that it has fully embraced a common European neoliberal trajectory (Baccaro and Howell, 2017). More recent data from the 2017 REPONSE survey show for instance that sectoral agreements still act as a major reference in company and workplace bargaining rounds, especially for wage-setting (Daniel, 2019). This signals a limited impact of disorganization and deregulation, in spite of ‘fiercer north-north competition’ (comparison with Germany being a significant driver of change in France) and ‘east-north destabilization’ (Dølvik and Marginson, 2018). Furthermore, if we move from processes to outcomes, recent studies have shown that ‘decentralization “by decree” was … ineffective’, considering that ‘there are no signs of lasting growth with regard to decentralized agreements in France, Italy and Spain, nor of significant use of the opportunities for derogations’ (Leonardi and Pedersini, 2018: 34). Looking at how the new institutional context diversely affects workplace employment relations, our own contribution helps to understand why this is so. In a context of extensive legal regulation, our main argument is that the diversity of workplace practices is best explained by contrasting reactions to legal requirements. Taking legal consciousness narratives into account (Ewick and Silbey, 1998), we show that where local employment relations are developed ‘against the law’ or ‘before the law’, the decentralization process is likely to remain largely ineffective. Where local industrial relations actors see themselves as acting ‘with the law’ or ‘beyond the law’, the standardizing effect of the legally regulated decentralization process might in turn limit its actual impact.
The introduction of legal consciousness among the explanatory factors should not be restricted to the French case. Based on our own data and fieldwork material, the four ways of regulating workplace employment relations that we have identified as being with, against, before and beyond the law have been linked to various socio-economic profiles that are both geographically and historically situated. But they could be relevant in different national and sectoral contexts. Cluster 2 ‘a-legalistic’ workplaces for instance, considering the law as a constraint, mainly involve SMEs with a stable workforce in the French case. But employment relations in some US multinationals might fit this type too, leading to an interesting comparison between small and big companies. In France, Cluster 4 ‘locally focused’ workplaces, standing ‘beyond the law’, comprise companies with a relatively highly skilled and growing workforce. Yet recent developments in social economy or third sector companies, with a much less skilled workforce, share some characteristics with this type of workplace employment relations. In that sense, our article is also an invitation to discuss the interest of introducing legal consciousness studies in future comparative labour research.
Footnotes
Acknowledgements
We are grateful to our colleagues who participated in the initial research project: Guillaume Blache, Camille Dupuy, Annette Jobert, Tommaso Pardi, Jean-Louis Renoux, Catherine Spieser, Michèle Tallard and Catherine Vincent. Earlier versions were presented at the International and Comparative Labor seminar, ILR School, Cornell University, May 2018, and at the ILERA World Congress, Seoul, July 2018. We thank the audiences for their helpful comments. Many thanks also to the anonymous referees and the Editor.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The article is based on a research project funded by DARES, as part of the programme Post-enquêtes à l’enquête REPONSE 2010-2011.