Has the recast made a difference? An examination of the content of European Works Council agreements

Abstract

The 2009 recast of the European Works Council (EWC) directive introduced substantial revisions to provisions on training, reporting back to employees and trade union involvement, together with amendments to the definitions of information, consultation and transnational competence. We use data from almost 1100 EWC agreements to evaluate its effect. We find a combination of direct effects from the recast and a general learning effect in the negotiation and renegotiation of EWC agreements. Which is dominant, depends on the issue under examination. The impact of the law tends to accelerate the development of EWCs as an institution in process.

Keywords

Consultation European Works Councils European Works Council agreement information recast directive training

The 1994 EU directive on information and consultation arrangements for European Works Councils (EWCs) set in train the establishment of more than a thousand representation institutions within multinational companies (MNCs) with operations in Europe. An array of case study and survey evidence demonstrated that the intentions of policy-makers were not realized in the practices of EWCs (Lecher et al., 1999, 2001, 2002; Marginson et al., 2013; Waddington, 2011; Weiler, 2004). In an attempt to address these shortfalls, a ‘recast’ directive was adopted in 2009, with the stated objectives of promoting growth in the number of EWCs and improving their operation (Jagodziński, 2009). We examine the impact of the recast on the content of EWC founding agreements¹ and establish whether the documents on which EWC practice is based have been amended accordingly.

Although the directive provided minimum standards regarding the procedures, competence and composition of an EWC, the detail was to be settled through negotiation between representatives of central management and the workforce. Apart from a few cases where representatives failed to reach an agreement and the fall-back terms of the ‘subsidiary requirements’ were implemented, the vast majority of agreements are the outcome of a negotiated settlement (De Spiegelaere and Jagodziński, 2015). This ‘regulated self-regulation’ (Platzer, 2002) was a contested process. Article 13 of the directive accentuated the self-regulation aspect of the legislation, allowing representatives of central management and the workforce to conclude voluntary agreements between September 1994 and September 1996, devoid of any reference to the terms of the directive. Furthermore, the parties to EWC agreements concluded under Article 13 are exempt from any obligation to comply with changes in legislation, as exemplified by the recast. Under Article 6, EWCs established after 22 September 1996 were required to meet the obligations specified in the directive. Similarly, Article 14(1b) of the recast allowed negotiators to conclude new agreements or to renegotiate extant Article 6 agreements between June 2009 and June 2011 exempt from the terms of the recast, but compliant with Article 6 of the original 1994 directive. Parties to Article 6 EWCs that did not renegotiate the agreement during this period must comply with the recast. EWCs are the outcome of contested negotiation processes where the parties can establish an institution appropriate to its specific corporate circumstances.

Article 15 of the directive required the European Commission to consult Member States and the social partners in reviewing the operation of the directive not later than 22 September 1999. While various reviews were undertaken, it was only in June 2009 that the recast came into force. During the intervening period, the Commission argued that the adoption of legislation on the European Company Statute and information and consultation within Member States was a more pressing priority, while also acknowledging that the information and consultation provisions specified in the ‘soft regulation’ approach adopted for the directive were inadequate (Waddington, 2011: 180–210). The European Parliament supported revision of the directive in two resolutions, on restructuring and employment in 2006 and on strengthening European legislation on information and consultation in 2007.

The social partners contested the terms of the revision of the directive, thereby ensuring that the revision process was politically charged. The European Trade Union Confederation (ETUC) sought to improve the operational effectiveness of EWCs by means of improved definitions of information, consultation and transnational competence; improved rights for training, time off and meetings for EWC representatives; a reduction in the period over which special negotiating bodies (SNBs) could deliberate on establishing an EWC; reductions in the workforce size thresholds, thereby increasing the coverage of the directive; and the introduction of a clear and defined role for trade unions (ETUC, 1998, 1999). The positions of managers responsible for EWCs within MNCs and BusinessEurope (until 2006, UNICE) were notable for their differences in emphasis. While acknowledging some operational difficulties in the introduction of new institutions, managers tended to emphasize that EWCs added value to company decision-making processes, to communication systems and in generating employee commitment (Hume Rothery, 2004; Lamers, 1998; Organization Resources Counselors (ORC) 2007). BusinessEurope, in contrast, took a more political position in opposing the revision of the directive and any further legislation on EWCs. It argued that the proposed recast included ‘such high obstacles to taking decisions quickly’ that it would ‘endanger business in the EU’, that such legislation would impair company flexibility in reinforcing centralized decision-making when decision-making in companies was being decentralized, and that ‘one size fits all’ legislation fails to recognize the diversity of MNCs (BusinessEurope, 2008). At a secondary level, within this framework BusinessEurope argued that increases in the coverage of the directive were unwarranted, that reductions in the SNB negotiation period would lead to inadequately constituted EWCs and that EWCs delayed decision-making when plenary meetings were convened annually (BusinessEurope, 2007, 2008; UNICE, 2004).

Given these differences in approach, it came as no surprise that both the ETUC and BusinessEurope saw the recast as not entirely satisfactory. By extensive lobbying, BusinessEurope was successful in excluding a reduction in the SNB negotiation period and changes to the workforce size thresholds, but was unable to prevent the revision per se. The ETUC secured amendments intended to improve the operation of EWCs. We examine six of these:

Definitions. The directive included no definition of information. The recast provided one and amended the definition of consultation to encourage timeliness, improvements in the quality and the undertaking of ‘in-depth’ impact assessments by EWC representatives. A wide range of evidence had shown inadequacies in the timeliness and quality of information and consultation (Lecher et al., 2001; Waddington, 2011). While the subsidiary requirements of the directive included a definition of the ‘transnational competence’, the recast moved this to the core text. In addition, the recast enlarged the scope of ‘transnational competence’ by referring to matters concerning multiple countries and to matters that are important for all European employees, regardless of the number of countries involved. This revision aimed to overcome the ambiguities inherent in the definition included in the directive, which had led to a series of high-profile legal cases (Blanke, 2010; Lamers, 2010).

Training. The recast stipulated that EWC representatives have a right to training without loss of wages and access to the means to conduct their duties. Such training should enable EWC representatives to fulfil their duties effectively. Two factors prompted this revision. First, only 28 percent of 784 EWC agreements extant in 2006 provided representatives with a right to training (Kerckhofs, 2006: 61). Second, managers responsible for EWCs argued that many representatives did not have the skills necessary for engagement in strategic company decision-making (Lamers, 1998; ORC, 2007).

Reporting back. An objective of the recast was to improve articulation between information and consultation arrangements at European and national levels. To this end, the recast required employee representatives to report at national level the content and outcome of the deliberations of the EWC. Data collected between 2005 and 2008 showed that reporting back was undertaken by a majority of EWC representatives, but this was not consistent, was more challenging for EWC representatives outside the home country of the MNC and was dependent on the presence of appropriate institutions at national level (Waddington, 2011: 123–131).

Access to premises. The rationale of the directive and the recast assumes that both ‘top-down’ and ‘bottom-up’ flows of information will inform the activities of the EWC. In practice, representatives are expected to report back the content of EWC meetings to constituents, while also gathering information from constituents to raise at the EWC. Although the recast is not explicit on access to premises, the requirements to report back and to gather information imply that EWC representatives should receive access to the premises where their constituents are employed (Group of Experts, 2010; Picard, 2010).

Select committee (SC). Article 6e of the recast states that ‘where necessary’ the composition, appointment procedure, functions and procedural rules of an SC should be included in the EWC agreement. Furthermore, the recast emphasizes the coordination role of the SC in promoting EWC effectiveness and the need for regular meetings. The recast thus upgrades the status of the SC. Although every European Trade Union Federation (ETUF) during the 1990s had recommended that an SC should be established within each EWC, only 66 percent of 806 active EWC agreements made such a provision in 2006 (Kerckhofs, 2006: 53).

Trade union organizations. The directive made no reference to trade union organizations; this no doubt contributed to the situation in 2006, where only 27 percent of 806 EWC agreements had an ETUF as a signatory (Kerckhofs, 2006: 70).² The intention of the recast was to stimulate greater trade union involvement in the negotiation of founding agreements and in the functioning of EWCs. To stimulate such involvement, the recast stipulated that European trade union organizations are to be informed when negotiations commence on the creation of an EWC. Specifically, the recast also enables negotiators to seek assistance from ETUF representatives. For the most part, the recast draws no distinction between national trade unions and the ETUFs in seeking greater trade union involvement in the operation of EWCs. We follow the same approach.

The expectation of policy-makers was that the revisions included in the recast would be incorporated into EWC founding agreements, thereby generating a ‘step increase’ in the number of agreements including terms consistent with these revisions. Most agreements negotiated from scratch or renegotiated after the adoption of the recast were expected to include definitions of information, consultation and transnational competence, and clauses on training, reporting back, access to premises, SCs and the representation of trade union organizations. We examine whether this expectation has been realized. In doing so, we refer to a general ‘learning effect’. EWC representatives and negotiators have been shown to learn from their own experiences and from practices elsewhere when new or renegotiated agreements are established (Kerckhofs, 2015; Waddington, 2011; Weiler, 2004). It is thus necessary to evaluate the effect of the recast in the context of this general ‘learning effect’.

Below we first identify the source of our data and explain the methods used in the analysis. Next we present the results, and then examine their implications for European level policy-makers and for the social partners. Our argument is that the recast has had an observable effect on the content of EWC agreements regarding the inclusion of definitions of information, consultation and transnational competence, and clauses on reporting back. Learning effects that are independent of the recast influence the issues of training clauses, access to premises, the representation of trade union organizations and the inclusion of definitions of information and consultation. In combination, these findings imply that as recent and innovative bodies, EWCs constitute an institution in process that is developing irregularly along a learning curve as individual bodies are reformed to meet changing circumstances. The character of this learning curve can be altered by legislative change. The policy challenge is to ensure that more EWCs engage in the developments that constitute the learning curve.

Data specification and method

We use the EWC database compiled by the European Trade Union Institute (ETUI), the most comprehensive source on EWCs. The database comprises two elements: information on the number of EWCs and the full text of agreements. While these two elements are clearly interlinked, the numbers in each are not identical. Each EWC may be attached to more than one agreement, a founding agreement and a renegotiated agreement, for example, and agreements for EWCs that are no longer active are included in the full text section of the database. Our research was undertaken in January 2016 when the number of active EWCs was 1073, the number of agreements was 2003 and the number of agreements used for the study was 1078. The agreements used for the analysis include negotiated agreements, irrespective of whether they were the original or a renegotiated document, signed between 1994 and 2015. Agreements based on the subsidiary requirements are excluded from the analysis, and those concluded before 1994 and after 2015 are also excluded as there are insufficient cases for reliable analysis. Using a scheme devised by the authors, the full text of each agreement was coded, thus enabling an assessment of the changing content of EWC agreements.

The six revisions mentioned above are treated as seven 0–1 dummy variables, where 1 indicates the presence of an explicit definition of information and consultation, a definition of transnational competence, a right to training paid for by management, a clause on reporting back, a guarantee of access to company premises, an SC, and trade union representation at the EWC as observer or expert. In addition to these dependent variables, the database allows identification of six independent variables, of which three are company-related, one based on legislation and two time-related. These are specified as follows.

Three company-related control variables are sector, home country and size. Six sectors are identified: building and woodwork; chemicals; food, hotels and catering; metals; private sector services; and others/unknown.³ Each sector was identified on the basis of the principal activity of the MNC and linked to the coverage of the ETUFs. The assumptions underpinning the inclusion of this variable are that unionization rates differ between sectors, suggesting differences in power relations during negotiations; that EWCs may be influenced by the pursuit of different policy objectives with varying intensity by the sectoral ETUFs; and that managerial objectives towards EWCs may well vary across sectors.

The home country of the MNC is included, because EWC agreements tend to reflect the national industrial relations traditions of the home country (Hertwig, 2015; Kerckhofs, 2015). The purpose of including this variable is to establish whether MNCs with different national origins react differently to the recast. To operationalize the variable, four country clusters are employed, each reflecting particular industrial relations traditions (European Commission, 2012; Hall and Soskice, 2001; Visser, 2008): Nordic, comprising Denmark, Finland, Iceland, Norway and Sweden; coordinated market economies (CME), comprising Austria, Belgium, Germany, Luxemburg, the Netherlands and Switzerland; mixed market economies (MME), comprising France, Greece, Italy, Portugal and Spain; and liberal market economies (LME), comprising Australia, Canada, Ireland, Japan, United Kingdom and United States. The new member states of 2004, 2007 and 2013 are also included in the LME category (Bohle and Greskovits, 2007).

A third company variable is size, measured by the number of employees within the European Economic Area (EEA). Three categories are employed: up to 5000 employees, 5001 to 10,000, and over 10,000. Larger companies that come within the scope of the directive are more likely to establish an EWC than their smaller counterparts (De Spiegelaere and Jagodziński, 2015; Kerckhofs, 2002, 2006) and size is also linked to the characteristics negotiated for the EWC (Kerckhofs, 2015). Our objective is to ascertain whether the rate of adoption of the amendments included in the recast is associated with company size.

A fourth variable concerns the article under which the EWC agreement was concluded. As mentioned above, there are three types of EWC defined by the regulatory framework: voluntary, Article 6 and Article 14(1b). The date on which an agreement is concluded indicates, but does not determine, the regulatory framework adopted by negotiators. A voluntary agreement or an Article 14(1b) agreement concluded under the recast, for example, could comprise the terms of the directive and recast respectively if the negotiators so decide. What is clear, however, is that Article 6 agreements are subject to more legal requirements than voluntary agreements; and there is a direct relationship between the regulatory framework of the EWC agreement and the inclusion of clauses on training, reporting back and SCs (Carley and Marginson, 2000). Our expectation is that there will be differences in the impact of the recast dependent upon the regulatory framework of the agreement, with Article 6 agreements more likely than voluntary agreements to include terms that fully comply with the new regulations.

Two time-related variables are utilized. Time as a continuous variable is introduced on the assumption that EWCs are ‘an institution in process’ (Waddington, 2011: 21–24) insofar as the parties may develop the institution to meet changing situations, irrespective of amendments to the regulatory framework. Learning effects, for example, have been noted as the parties to EWCs develop the institution (Kerckhofs, 2015; Weiler, 2004). The analytical issue is whether learning effects predate or are more influential than legislative change, or vice versa. A second time-related variable identifies four periods in the history of EWC legislation: before 22 September 1996, 22 September 1996 to 4 June 2009, 5 June 2009 to 5 June 2011, and after 5 June 2011. This variable is based on the date on which an EWC agreement was signed rather than the regulatory framework adopted by negotiators, although these periods coincide with the legislation mentioned above. In each of these periods, however, agreements were legitimately concluded that did not conform to the legislation in place at the time. In 2014, for example, renegotiated agreements were concluded under Article 13 of the directive.

Our objective is to estimate the effect on the seven binary dependent variables of a range of primarily categorical independent variables. Four characteristics of the database determine the approach taken here. First, it has a hierarchical structure: a single EWC can negotiate several agreements, for example, a founding agreement that is subsequently renegotiated. These agreements comprise the observations for this study and tend to be similar, as renegotiated agreements tend to include terms from previous agreements. In a standard regression equation, however, it is assumed that all observations are independent. Second, most observations are independent insofar as there is only one founding agreement for each EWC at any one point in time. Third, both continuous and time-discrete time effects are employed to ascertain the overall learning and policy effects. Fourth, the time intervals are unequally spaced. For these reasons, a standard logistic regression analysis would generate biased results. Multi-level logistic regression may provide estimates on the development of agreements within a specific EWC. The purpose here, however, is to examine the overall development of EWC agreements. A method to provide population average estimates is thus required. Generalized estimating equations (GEE) are employed to model relations between the dependent and independent variables. The reported GEE estimation uses an exchangeable working assumption and the empirically corrected standard errors (Molenberghs and Verbeke, 2005: 151–161).

We aim to establish whether the recast had an observable effect on the manner in which the seven dependent variables appeared in EWC founding agreements. Existing research has demonstrated a general ‘learning effect’ on the content of agreements as the signatories developed EWCs to meet current circumstances (Marginson et al., 2004). The potential combinations of recast and learning effects are represented in Figure 1:

Figure 1.

The potential impact of the recast.

Boosted learning occurs when the recast reinforces a pre-existing learning effect.

Simple improvement comprises a policy impact of the recast in the absence of any learning effect.

Simple learning results solely from learning effects without any impact of the recast.

No effect occurs; there is neither a learning nor a recast effect.

Results

The share of agreements including a specific clause or definition varies by sector, country cluster, company size, EWC type and periods. Among the 1078 agreements studied, a large majority include provisions on an SC (88%), on reporting back to the workforce (73%) and provide a definition of transnational issues (66%). More than half have a clause on training (58%). The other variables studied appear more rarely: definitions of information and consultation in 22 percent of agreements, guaranteed access to the company premises in 12 percent, and trade union representation in 12 percent.⁴

Table 1 includes the results of the multivariate GEE analysis. This filters out characteristics that arise from the development of agreements, while controlling for differences that arise from the independent variables. Table 1 provides the odds ratios, the standard errors and significance levels for the different variables included in the model. The odds ratios measure the difference (increase, decrease) in the odds (ratio of the probability that an event occurs to the probability that it does not) of a particular clause being included in an agreement, depending on different variables and by comparison with a given reference group. An odds ratio of more than 1 indicates that agreements with a certain characteristic are more likely to have, for example, a training clause, than agreements from the reference group. An odds ratio of 1 indicates there is no difference and an odds ratio of less than 1 indicates that agreements from that group have lower odds of inclusion than the reference group.

Table 1.

GEE estimation results.

	I & C definition	Transnationality	Training	Reporting back	Access	SC	ETUF
	Odds ratio (SE)	Odds ratio (SE)	Odds ratio (SE)	Odds ratio (SE)	Odds ratio (SE)	Odds ratio (SE)	Odds ratio (SE)
Sector (ref: Metal)	2.88	7.51	8.87	15.2**	17.3**	6.31	33.9***
Chemicals	1.21 (0.31)	1.35 (0.27)	1.02 (0.2)	0.58 (0.12)**	0.37 (0.13)**	0.57 (0.16)*	3.18 (1.09)***
Services	0.91 (0.23)	1.15 (0.25)	0.65 (0.14)*	0.71 (0.16)	0.46 (0.14)**	0.53 (0.17)*	2.56 (0.92)**
Food, HORECA	1.29 (0.39)	1.01 (0.25)	0.63 (0.16)⁺	0.81 (0.22)	0.31 (0.16)*	0.6 (0.2)	7.44 (2.75)***
Building and wood	0.78 (0.28)	0.65 (0.17)⁺	0.91 (0.25)	0.38 (0.11)***	0.93 (0.35)	0.67 (0.27)	4.92 (2)***
Other/unknown	1.18 (0.36)	1.17 (0.29)	1.16 (0.29)	0.68 (0.18)	0.61 (0.21)	0.71 (0.27)	2.47 (1.07)*
Country (ref: CME)	24.15***	65.54***	10.21*	30.62***	29.64***	0.99	22.49***
Nordic	0.46 (0.16)*	0.44 (0.1)***	1.86 (0.42)**	0.78 (0.17)	0.19 (0.08)***	0.8 (0.25)	0.52 (0.18)*
MME	0.85 (0.21)	0.3 (0.07)***	0.85 (0.19)	0.92 (0.2)	0.45 (0.13)**	0.98 (0.33)	1.15 (0.31)
LME	1.97 (0.42)**	1.38 (0.25)⁺	1.17 (0.2)	2.31 (0.44)***	0.3 (0.08)***	0.83 (0.2)	0.3 (0.1)***
Size (ref: large)	4.91	1.87	10.24*	3.01	1.1	7.85*	4
Unknown	0.72 (0.24)	0.78 (0.23)	0.58 (0.17)⁺	0.63 (0.2)	1.09 (0.41)	0.55 (0.23)	0.79 (0.32)
Small	0.66 (0.14)*	0.99 (0.17)	0.63 (0.11)**	0.81 (0.14)	0.94 (0.24)	0.55 (0.13)**	0.65 (0.17)⁺
Medium	0.68 (0.19)	0.81 (0.17)	0.93 (0.19)	0.89 (0.19)	0.75 (0.25)	0.7 (0.21)	0.62 (0.2)
EWC type (ref: Art. 6)	7.57*	16.67***	0.78	5.74⁺	2.41	0.16	1.05
Voluntary	0.44 (0.14)**	0.39 (0.1)***	0.85 (0.21)	0.65 (0.17)⁺	1.18 (0.45)	0.87 (0.34)	1.41 (0.58)
Art. 14	1.12 (0.56)	0.63 (0.37)	1.43 (1.01)	0.35 (0.2)⁺	2.93 (2.18)	1.19 (1.36)	0.74 (0.58)
Year	1.14 (0.04)***	1.01 (0.03)	1.11 (0.04)***	1.01 (0.03)	1.2 (0.05)***	1.07 (0.05)	1.14 (0.05)**
Period (ref: 1996–2009)	14.54**	24.64***	9.29*	24.77***	4.87	2.39	6.47⁺
Pre-1996	1.33 (0.5)	1.93 (0.52)*	0.45 (0.12)**	1.83 (0.52)*	0.86 (0.42)	0.59 (0.23)	2.43 (1.06)*
2009–2011	3.63 (2.07)*	4.57 (3)*	1.4 (0.97)	11.47 (7.88)***	0.22 (0.19)⁺	n.d.	0.44 (0.22)⁺
Post-2011	4.18 (1.83)**	3.92 (1.65)**	1.28 (0.57)	3.44 (1.57)**	0.62 (0.28)	2.24 (1.8)	0.57 (0.31)
Constant	0.01 (0.01)***	0.01 (0.01)	0.01 (0.01)***	0.01 (0.01)	0.01 (0.01)***	0.01 (0.01)	0.01 (0.01)**
N	1078	1078	1078	1078	1078	1078	1078
N groups	1009	1009	1009	1009	1009	1009	1009
Chi sq.	199	97	171	80	118	48	74
Chi sq. df	17	17	17	17	17	16	17

GEE: Generalized estimating equations; SC: select committee; SE: standard error; ETUF: European Trade Union Federation; CME: coordinated market economies; MME: mixed market economies; LME: liberal market economies; EWC: European Works Council.

GEE logit estimation with cluster adjusted standard error.

p < .1; *p < .05; **p < .01; ***p < .001.

The results largely confirm the importance of the company and EWC-related control variables. Of specific interest is the influence of the regulatory status of an EWC. Even when controlling for the country cluster, date of establishment and company size, there are still significant differences between agreements concluded in voluntary, Article 6 and Article 14(1b) EWCs. Compared to Article 6 EWC agreements, those made in voluntary EWCs are significantly less likely to include definitions of information and consultation or transnational competence, and are less likely to include clauses on reporting back, but they are more likely to provide for trade union representation. The regulatory status of the EWC is thus still an important issue and continues to influence the quality of EWC agreements.

Of primary interest are the two time-related variables, which model the overall learning process and the impact of the recast. To facilitate the interpretation of the results, Figure 2 shows the evolution of the predicted probabilities based on the model presented in Table 1, as suggested by Agresti (2007). For reasons of clarity, the exact predicted probabilities for the transition period are marked in grey. This approach was taken because the small number of agreements concluded in this period is necessarily associated with a large error margin and the principal focus of this study is a comparison of the situation before and after the recast. Figure 2, together with the model of Table 1, illustrates how the probability of an agreement including a certain clause changes over time.

Figure 2.

The impact of the recast: model based on predicted probabilities 2000–2015.^a

Definitions of information and consultation (I & C)

The probability that an agreement includes definitions of both information and consultation is subject to a clear learning effect. With every passing year, the share of agreements including both definitions rises. At the time of the recast, the probability further increases, demonstrating a boosted learning effect. The legal challenges encountered in defining information and consultation probably contributed to the general learning effect (Sachs-Durand, 2010), while the recast effect can be explained by the prominence of the new definitions in the recast and the importance attached to the new definitions by European policy-makers and trade unionists. Currently, virtually all new EWC agreements include definitions of information and consultation.

Definition of transnational competence

The model shows a clear effect of the recast, but no general learning effect. This pattern of simple improvement can be explained by reference to the relatively large number of early agreements that included a definition of transnational competence. Such inclusions are not surprising, as agreement on transnational competence is essential in defining the scope of an EWC. As the recast proposed a new definition, relatively more agreements included a definition.

Training clauses

The pattern shows a clear learning effect, but no observable recast effect. Although only a few early agreements included training clauses, the share of agreements providing training accelerated during later years to result in a (predicted) probability of about 85 percent in 2009. The recast, which established a right to training for all EWC representatives, did not affect this development in a significant way. This simple learning pattern results from the acknowledgement by employers and employees that training added value to EWC practices and inclusion of such a provision was not necessarily contentious (ETUC, UNICE). At the time of the recast, the probability of inclusion was already so high that an additional increase was difficult to achieve. The recast in this case confirmed reality, but did not change it.

Reporting back

The pattern is one of simple improvement. Just as with the definition of transnational competence, the probability of inclusion was initially relatively high. Over time, this probability did not change markedly, but increased with the recast. Articulation with the other levels of employee representation is essential for the effective functioning of an EWC, and the need to report back to the workforce was included in the subsidiary requirements. This promoted a relatively high probability of agreements comprising reporting-back provisions in the early period. The mention of this right in the recast further increased the probability of inclusion to almost 90 percent.

Access to premises

This is not explicitly mentioned in the directive or the recast. The model shows that the initial probability of inclusion was low. There is, however, an increasing trend in the probability of inclusion, which is unaffected by the recast. This pattern of simple learning seems to be interrupted during the transition period (2009–2011), when the estimated probability decreased sharply. As the number of agreements negotiated in this period is relatively low, however, this decrease is marginally significant and is reversed by the increase after 2011. Obviously, the implicit reference to access to premises in the legislation did not result in a higher uptake of this practice. Over time, however, the probability of inclusion is increasing, suggesting that the negotiating parties see the practice as instrumental for the effective functioning of an EWC.

Select committees

From the outset, the establishment of an SC was included in the subsidiary requirements, resulting in a high initial estimated probability of inclusion of over 80 percent. Over time, this further increased. The initially high probability of inclusion, however, means that the increase tends to be statistically insignificant. Moving this matter to the core text of the recast did not lead to a significant increase of the probability of inclusion. In summary, there is no effect.

Trade union representation

For the inclusion of trade union representatives as permanent experts or members, the models show a distinct pattern of simple learning: a general increase but no recast effect. The manner in which legislators tried to stimulate trade union involvement in the recast, by indirectly including an obligation to inform the unions at the start of a negotiation, did not result in the desired effect.

In summary, the results show that the recast only affected three of the seven variables under study: definitions of information and consultation, definition of transnational competence and reporting back. Two of the non-affected variables were close to full compliance in recent agreements, leaving almost no room for further improvement: training and SCs. The two remaining variables were only indirectly mentioned in the recast: access to premises and trade union representation. Concurrently, four of the variables under study are subject to general learning effects, three of which in practice made the recast ineffective: training, access to premises and trade union representation. This observation supports the idea of ‘regulated self-regulation’, which relies on the proliferation of best practices. The analysis also demonstrates the limits of ‘regulated self-regulation’, as it did not lead to wide-ranging use of guaranteed access to premises or trade union representation, both of which were clearly intended by European policy-makers.

While our study generates insights regarding the effects of the directive and the recast on the content of EWC agreements, there are limitations. First, we focus on the share of agreements concluded in a specific year and not on all EWCs in operation during that year. Some parties to EWCs have never renegotiated the agreement and thus still function on old regulations. This has a major consequence. Even if almost full compliance for some clauses in recent agreements is observed, this does not mean there is a near full compliance in the population of EWCs. The legislation places no obligation on the parties to renegotiate voluntary EWC agreements, for example, yet these agreements are significantly less likely than other agreements to include definitions of information and consultation and clauses on reporting back. Second, the measures used in this study only refer to the presence or absence of a clause on some topics: their content is not assessed. A clause on reporting back could indicate, hypothetically, that this is a sole responsibility of management and that reporting back should be directed to the local management only, or could give very advanced means to EWC representatives to report back to both local union representatives and the entire workforce. These differences in content will be the subject of future study. We thus examine only the ‘formal quality’ of EWC agreements and the effect of policy on ‘formal quality’. Third, as with all policy evaluations, it is difficult to isolate the effect of the policy intervention from all other possible effects. This study controlled for sector, home country and size, together with a general learning effect. In these circumstances, it is not possible to isolate any policy effects arising from the parties to agreements anticipating regulatory change. These effects are difficult to model and may deflate the estimation of the policy effect.

Conclusion

Our study evaluates the impact of the recast by focusing on how the revised terms were incorporated in EWC agreements. It shows that the recast had a clear and observable effect only on the definitions of information and consultation, that of transnational competence and on reporting back. For the other features examined here, the recast did not influence the content of EWC agreements: because the proportion of agreements that included clauses on SCs and training was already very high, and because the recast simply did not affect reality in the cases of access to premises and trade union representation.

We also examined a general learning effect in EWC agreements. As the parties to EWCs are given considerable latitude in the negotiation of agreements, they are influenced by the experiences and practices of others. We identified learning effects in respect of training, trade union representation, access to premises and definitions of information and consultation, illustrated by the probability of inclusion rising over time and this increase being independent of legal intervention. These findings are consistent with previous research, which demonstrated strong learning effects without observable effects of the recast (Kerckhofs, 2015). The absence of an observable effect does not, however, rule out the possibility of any recast effect. Anticipated learning, in which EWCs already take into account upcoming changes in regulation, might mask part of the effect of the recast. In conclusion, we have identified both the accomplishments and limits of legal action and regulated self-regulation. Both strategies seem effective in some fields and not in others.

These observations have clear implications for practice and policy. Regarding practice, we show that reality is not only shaped by legal intervention. Training provisions, the introduction of an SC and access to company premises, for example, proliferated in EWC agreements through learning and copying practices implemented elsewhere. Collectively, EWCs are thus an institution in process as the parties develop practices appropriate to their particular circumstances. This learning process is, in part, dependent upon the presence of coordination mechanisms. In this context the ETUFs undertake important roles as coordinating hubs and as the means of disseminating innovative practices in the form of guidelines and toolkits. The coverage of some of the measures examined here remains low, suggesting that learning processes are restricted in effect in some areas.

Variation in EWC practice and the content of agreements is also influenced by managerial policy. Research shows that there is a wide range of managerial approaches to EWCs, varying from symbolic EWCs, where managers do little more than comply with the legislation, to participative EWCs, where managers and employee representatives are engaged in joint decision-making (Lecher et al., 2001; Marginson et al., 2013). The findings suggest that managerial pursuit of added value has contributed to this learning process insofar as the establishment of what constitutes best practice and its implementation is a feature of the learning curve of EWCs. In short, EWC reality thus can be influenced by both trade union organizations and managerial policy without reference to protracted political processes.

With regard to policy, our research demonstrates the limits of self-regulation. Better definitions of transnational competence and clearer statements on the rights and competences of EWCs are absent from ‘regulated self-regulation’. Voluntary approaches to EWCs are thus likely to generate disparities so that information and consultation arrangements are not taken out of competition. Similarly, the exclusion of voluntary agreements from the legislative remit ensures that many agreements remain rudimentary and limited in scope. Our study also clearly identifies the shortcomings of the recast. Although advanced union involvement was a clear policy objective, for example, the recast did not make a contribution in this area. Nor did it bring greater access to premises. For these policy objectives, the recast only provided an indirect stimulus, which did not meet the expectations of European policy-makers. A more direct stimulus and explicit statements is required if policy expectations are to be met. It is also apparent that the recast did not result in ‘such high obstacles’ that would impair the operation of companies with an EWC as suggested by BusinessEurope (2008). To the contrary, the recast confirmed extant practice in many cases rather than necessitating major reform. It remains to be seen whether the review of the recast that the European Commission plans to complete during 2017 will incorporate analysis of the practices learnt since 2009 and bring the legislation up to date with these practices.

Footnotes

Acknowledgements

The authors are very grateful to Aline Hoffmann, Romek Jagodziński and the editors of this special issue for their insightful comments on earlier versions of this study.

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was funded by the European Trade Union Institute and the Hans-Böckler-Stiftung.

Notes

Author biographies

Stan De Spiegelaere is a research officer at the European Trade Union Institute (ETUI), Brussels.

Jeremy Waddington was a professor of Industrial Relations at the University of Manchester and is the project coordinator at the ETUI, Brussels.

References

Agresti

(2007) In: An Introduction to Categorical Data Analysis, 2nd edn. Hoboken, NJ: Wiley Interscience, 400 pp.

Blanke

(2010) Germany. In: Dorssemont

Blanke

(eds) The Recast of the European Works Council Directive. Antwerp: Intersentia, pp. 159–192.

Bohle

Greskovits

(2007) Neo-liberalism, embedded neo-liberalism and neo-corporatism: Towards transnational corporatism in Central-Eastern Europe. West European Politics 30(3): 443–466.

BusinessEurope (2007) Visit of BusinessEurope President Ernest-Antione Sellière and BusinessEurope Secretary General Philippe de Buck to Confederation of Swedish Enterprises (SN). Brussels: BusinessEurope.

BusinessEurope (2008) Letter to EU Commissioners’ Heads of Cabinets on EWCs. Brussels: BusinessEurope.

Carley

Marginson

(2000) Negotiating European Works Councils: A Comparative Study of Article 6 and Article 13 Agreements. Dublin: European Foundation.

De Spiegelaere

(2016) Too little, too late? Evaluating the European Works Councils recast directive. ETUI report no. 138. Brussels: ETUI.

De Spiegelaere

Jagodziński

(2015) European Works Councils and SE Works Councils in 2015: Facts and Figures. Brussels: ETUI.

ETUC (1998) European Works Councils. Brussels: ETUC.

10.

ETUC (1999) ETUC Resolutions 1999. Brussels: ETUC.

11.

ETUC, UNICE, UEAPME and CEEP (2005) Lessons Learned on European Works Councils. Brussels: ETUC, UNICE, UEAPME and CEEP.

12.

European Commission (2012) Industrial Relations in Europe 2012. Luxembourg: Office for Official Publications.

13.

Group of Experts (2010) Implementation of Recast Directive 2009/38/EC on European Works Councils. Brussels: European Commission.

14.

Hall

Soskice

(2001) An introduction to varieties of capitalism. In: Hall

Soskice

(eds) Varieties of Capitalism: The Institutional Foundations of Comparative Advantage. Oxford: Oxford University Press, pp. 1–70.

15.

Hertwig

(2015) European Works Councils and the crisis: Change and resistance in cross-border employee representation at Honda and Toyota. British Journal of Industrial Relations 53(2): 326–349.

16.

Hume Rothery

(2004) Implementing the directive: A view from UK business. In: Fitzgerald

Stirling

(eds) European Works Councils: Pessimism of the Intellect, Optimism of the Will? London: Routledge, pp. 80–92.

17.

Jagodziński

(2009) Recast directive on European Works Councils: Cosmetic surgery or substantial progress? Industrial Relations Journal 40(6): 534–545.

18.

Kerckhofs

(2002) European Works Councils: Facts and Figures. Brussels: ETUI.

19.

Kerckhofs

(2006) European Works Councils: Facts and Figures 2006. Brussels: ETUI.

20.

Kerckhofs

(2015) European Works Council Developments before, during and after the Crisis. Luxembourg: Publications Office.

21.

Lamers

(1998) The Added Value of European Works Councils. Haarlem: AWNN.

22.

Lamers

(2010) EWC’s role recast: A European actor? In: Dorssemont

Blanke

(eds) The Recast of the European Works Council Directive. Antwerp: Intersentia, pp. 357–397.

23.

Lecher

Nagel

Platzer

H-W

(1999) The Establishment of European Works Councils. Aldershot: Ashgate.

24.

Lecher

Platzer

H-W

Rüb

Weiner

K-P

(2001) European Works Councils: Developments, Types and Networking. Aldershot: Gower.

25.

Lecher

Platzer

H-W

Rüb

Weiner

K-P

(2002) European Works Councils: Negotiated Europeanisation. Aldershot: Ashgate.

26.

Marginson

Hall

Hoffmann

Müller

(2004) The impact of European Works Councils on management decision-making in UK- and US-owned multinationals. British Journal of Industrial Relations 40(2): 209–234.

27.

Marginson

Lavelle

Quintanilla

Adam

Sánchez-Mangas

(2013) Variation in approaches to European Works Councils in multinational companies. Industrial and Labor Relations Review 66(3): 618–644.

28.

Molenberghs

Verbeke

(2005) Models for Discrete Longitudinal Data. New York: Springer.

29.

ORC (2007) ORC European Works Council Survey 2007. New York: Organization Resources Counselors.

30.

Picard

(2010) European Works Councils: A trade union guide to directive 2009/38/EC. Report no. 114. Brussels: ETUI.

31.

Platzer

H-W

(2002) Europäisierung und Transnationalisierung der Arbeitsbeziehungen in der EU. Internationale Politik und Gesellschaft 2: 103–121.

32.

Sachs-Durand

(2010) Information and consultation in the recast directive. In: Dorssemont

Blanke

(eds) The Recast of the European Works Council Directive. Antwerp: Intersentia, pp. 313–356.

33.

UNICE (2004) First Stage Consultation of the European Social Partners on the Review of the European Works Council Directive: UNICE Answer. Brussels: UNICE.

34.

Visser

(2008) Europe’s industrial relations in global perspective. In: European Commission (ed.) Industrial Relations in Europe 2008. Luxembourg: Office for Official Publications, pp. 19–45.

35.

Waddington

(2011) European Works Councils and Industrial Relations: A Transnational Industrial Relations Institution in the Making. London: Routledge.

36.

Weiler

(2004) European Works Councils in Practice. Dublin: Eurofound.