Trajectories of liberalisation on the European industrial relations systems A comparative study

The main hypothesis of this article is that labour liberalisation in the post-Fordist period has affected the three principal areas of collective bargaining coordination (coverage, dominance and control) in two different ways: by attacking the core (deregulation) and the margins (dualisation) of industrial relations. Due to differences in the institutional structures of European countries, these processes could have different effects in each country, resulting in differentiated and gradual path-dependent transformations. This hypothesis is tested by studying the institutional frameworks of industrial relations in several representative European economies: France, Germany, Greece, Italy, Poland, Spain, Sweden and the United Kingdom. Finally, we conclude that there has been a substantial modification of the role, activity and weight of collective actors in the economic system of each country as a consequence of the effect of the different forms of liberalisation on the coordination of collective bargaining. As a corollary, a shift in European Union labour policies towards a further strengthening of collective bargaining spaces is needed to give social partners greater capacity to seek coordinated solutions to contemporary economic and social problems.


Introduction
The industrial relations system not only affects the productive sphere, personnel management and working conditions (increasing efficiency and productivity and avoiding conflicts), but also influences the distributive sphere, where the wage ratio and income distribution are determined (Dunlop, 1958;Clegg, 1976;Hyman, 1989;Crouch, 1993;Kelly, 1998;Gumbrell-McCormick & Hyman, 2013). It therefore includes aspects such as the regulation of labour, how the relative power of capital and labour are socially determined and the interactions between employers' organisations, trade unions and the state. The field of 'comparative industrial relations' (Clegg, 1976;Blyton, et al., 2008;Hauptmeier & Vidal;Wilkinson et al., 2014) has studied the important differences between institutional structures in advanced economies, which ultimately reflect the different forms of coordination between economic actors in the labour market. However, the outcome of cooperation depends essentially on the structural and associative power resources of each economic agent (Ibsen, 2015;Cárdenas & Arribas, 2022).
Our brief review of the literature includes three main debates. First, there is the debate whether there are different typologies of institutional structures of industrial relations and whether these typologies are useful for classifying and comparing macroeconomic performance in a comparative perspective. Second, there is the debate on whether these institutional structures are relatively stable or have trajectories of change over time. Third, there is a debate about how these have been affected by the Great Recession and the austerity policies promoted by the EU. These policies have put protective labour institutions under considerable pressure.
In the context of the first debate, that of the existence of different typologies, the work of Calmfors and Driffill (1988) predicts the behaviour of nominal wages, prices and unemployment according to the centralisation of wage bargaining, an analysis which has subsequently given rise to a whole set of interesting reflections. In contrast to the neo-corporatism literature (Goldthorpe, 1984), they argue that there are two equally efficient forms of collective bargaining for economic performance.
The first of these, the inverted-U (hump-shaped) hypothesis, holds that wage bargaining generates lower real wage growth when it is fully decentralised or fully centralised, while in intermediate situations its growth tends to be higher. This is due to the coverage of the union (number of workers it represents, from one union per industry to a single union across the whole economy) and the effect of a nominal wage increase on prices, considering in turn the impact of prices on workers' purchasing power. This model has important limitations but remains interesting because it denies that there is a one-to-one relationship between wage bargaining centralisation and macroeconomic performance. It thus qualifies the corporatist hypothesis (Molina & Rhodes, 2002) on the pre-eminence of centralised collective bargaining led by large actors. This was mainly inspired by the labour market institutions of the Nordic economies and Austria, whose economic performance after the oil shocks of the 1970s was markedly better than that of other European countries.
Second, there is a broad consensus that Calmfors and Driffill's (1988) indicators are excessively formal, as coordination attempts to capture the actual wage-setting behaviour of actors and the extent to which economic agents with less bargaining influence follow the benchmarks set by the more influential ones. Several authors have differentiated between the two concepts (Kenworthy, 2001;Traxler, Blaschke & Kittel, 2001;Aidt & Tzannatos, 2008;Visser, 2013): coordination in its vertical dimension (the scope of bargaining) and the horizontal dimension (the interrelation with other sectors). In this sense, Traxler et al. (2001:105-111) point to the importance of 'governance capacity' , defined as the effectiveness of vertical coordination mechanisms that ensure that lower levels comply with the agreements made at higher levels. This depends on the legal enforceability provisions along the dimensions of coverage, domain and control. Here, coverage refers to the degree of multi-employer coverage and collective bargaining; domain refers to the degree of organisation of the social partners; and control refers to the degree of compliance with the collective agreement at the workplace.
Third, another strand of the literature has pointed out that the highly centralised coordination systems, characteristic of economies such as the Austrian or Swedish economies in which large organisations have reached agreements for many branches of activity, have deteriorated to the point of disappearing. In coordinated European economies, the predominant form is pattern bargaining. This is the most decentralised version of coordination: a bargaining unit at branch or company level sets the tone for the other units, without a formal central role and without apex organisations with a leadership role, being a non-hierarchical mode of coordination because the bargaining units are not subordinated to a coordinating authority. There are two types of pattern bargaining: intra-industry and inter-industry (Traxler, Blandl & Glassner, 2008).
However, both theses ('hump-shaped' and 'U-shaped') have something in common, namely the existence of institutional frameworks that make negotiators (trade unions and employers) internalise the externalities of their demands (wages and prices of goods and services, respectively). Bechter, Brandl and Meardi (2012) point out that to a large extent, the same industrial relations patterns can be observed in the same sector in different countries, so it follows that not only the national but also the sectoral level has to be considered, and the combination of the two seems to be paramount. Thus, the sectoral structure of the economy also plays a role in conditioning the type of collective bargaining coordination.
As to the second debate mentioned above, whether institutional structures are relatively stable or change over time, different interpretations have been made. First, for the varieties of capitalism (VoC) school (Hall & Soskice, 2001), the institutional structure tends to perpetuate itself over time, given the perfect complementarity between institutional spheres (interfirm relations, labour relations, governance and financing of firms, and the education system). In this model, there are self-reinforcing dynamics between institutions and the choices made by actors. In contrast, this idea of resilient institutional frameworks has been criticised in the light of the significant dynamics of institutional change and the types of decisions made by actors in some coordinated economies. In particular, in Germany, the archetypal example, it has been found that the functional adjustment thesis could not account for normative and regulatory changes that have shown a significant trend towards liberalisation of those areas where strategic interactions dominated, and a retreat from state intervention (Streeck, 2010;Hassel, 2014). This line of the common liberalisation trajectory (Baccaro & Howell, 2017) leads, as a fundamental consequence, to the breakdown of the large collective actors (unions and employers) and their capacity to act as referents in collective bargaining (Visser, 2016).
In addition, other authors have argued that, although the process of liberalisation has eroded the institutional framework, important differences remain, because this erosion has affected only certain segments of the labour market (Thelen, 2014), thus distinguishing between the institutional displacement of deregulation and the stratification of dualisation. An example of this is the way that disintegration (vertical or horizontal) reduces the capacity of control by trade unions, through the creation of new intermediate markets in a production process previously integrated into the value chain (Doellgast & Greer, 2007). These practices include subcontracting or outsourcing production to a specialised company (e.g. to multi-service companies), the creation of independent subsidiaries (at home or abroad), the use of temporary employment agencies for staffing, or directly market-based relationships (e.g. hiring self-employed workers).
The third debate concerns the impact of the Great Recession and the austerity measures and support for the liberalisation of European labour markets by European institutions (Marginson, 2015;Meardi 2018). Although competitive wage moderation policies have been implemented in Europe in recent decades, generally based on nominal wage restraint in the protected sector (domestic services) for the sake of price competitiveness, the pressure to implement such policies increased markedly during the debt crisis (Armingeon & Baccaro, 2012). Arnholtz, Meardi & Oldervoll (2018) analyse liberalisation pressures in different Northern European countries, concluding that outcomes vary according to the strategies of entrepreneurs and the power relationship between social actors. In a similar vein, Dolvik and Marginson (2018) compare the Nordic countries (Sweden, Denmark, Norway and Finland) with Germany, concluding that they have responded differently to liberalising pressures, with the Nordic countries showing greater institutional resistance (although with differences between them) and reorienting towards a bargaining pattern led by the manufacturing industry (as the exposed sector). Finally, Leonardi and Pedersini (2018) studied five countries (Germany, France, Spain, Italy and Belgium) analysing different responses to the decentralisation of collective bargaining.
From a sectoral approach, several studies show the different responses to EU integration of the same sector in different countries. Benassi, Doellgast and Sarmiento-Mirwaldt (2016) studied the case of the telecommunications sector and how in Austria and Sweden coordination in collective bargaining was maintained while in Germany and Denmark it was not because employers took advantage of loopholes in the legal framework. Hassel, Steen Knudsen and Wagner (2016) studied the low-wage and meat sectors in Germany and Denmark and concluded that the response was different in the two countries. While in Germany (as a dual market) the integration of migrant labour was in the secondary market, in Denmark the inclusion of these workers and procurement from German suppliers was the priority.
In short, based on these three debates in the literature, we can state that the type of institutional structures and how they are being transformed is a subject of study in the field of industrial relations. Our main contribution is to study whether different typologies of institutional structures of industrial relations exist, whether these institutional structures have trajectories of change over time, how they have been affected by the Great Recession, and thus to what extent bargaining institutions are being transformed in several selected European countries. This is a relevant question as the findings can be used to derive recommendations for labour policy and legislative changes required to strengthen the capacity of national social partners to reach agreements and improve the governance of the industrial relations system. This fact is relevant for policy-makers at European level, as the European Commission has emphasised in recent years the need for greater involvement of the social partners in public policy-making, especially in the European Semester and in the country-specific recommendations (CSRs). This requires a better understanding of how the various institutional changes, many of which were the result of the very policies previously promoted by the EU, have affected industrial relations systems in each country and to what extent policies and legislation can be formulated to strengthen the capacity of national social partners to reach agreements and improve the governance of the system. Accordingly, we put forward the following hypotheses: • • Hypothesis 1: There are differences between the level of indicators reflecting institutional structures and the liberalising trends that have operated in recent decades. On the governance capacity, they can be grouped into: • Hypothesis 1.1: Decentralisation of collective bargaining implies a shift from higher levels (sectoral or national) to lower levels (firm or workplace level), leading to a reduction in collective bargaining coverage and union presence.
• Hypothesis 1.2: Disorganisation affects the dominance and control of collective bargaining due to the fragmentation of trade union organisations and the existence of heterogeneous interests, for example, due to the rise of professional organisations or trade unions representing specific groups of workers rather than class-oriented confederations. • • Hypothesis 2: The liberalising trend is common, especially driven in the austerity phase, but it has affected each country differently due to the different starting points of the institutional framework and the different power resources in the collective action of social actors. This means that both deregulation and dualisation can be distinguished.
In order to test these hypotheses, we studied the institutional frameworks of industrial relations in several representative European economies. The UK as an Anglo-Saxon economy, Germany and France as continental economies, Sweden as a Nordic economy, several Mediterranean economies (such as Spain, Italy and Greece) and Poland representing Eastern European economies. We presumed that institutional structures, especially the modalities of coordination of economic agents, cannot be measured solely through variables that include formal institutional elements, but must be combined with others of a qualitative nature, which refer to aspects that transcend the formal and better capture the character of interaction of economic agents. Applying this methodology, first, we would reject the first hypothesis if there has been no decentralisation of collective bargaining (levels of coverage of sectoral or national collective agreements are maintained) or if trade union fragmentation is maintained or reduced in the countries analysed. Second, we would reject the second hypothesis if the results of liberalising trends were similar across countries and, therefore, previous differences and their different power resources were not relevant.
In the next section, we use indicators on the information collected in the OECD/ AIAS database on Institutional Characteristics of Trade Unions, Wage Setting, State Intervention and Social Pacts (ICTWSS), where these particularities can be observed. This is followed by an analysis of institutional developments in each of the countries in the following section. Finally, the last section is dedicated to the conclusions.

Quantitative analysis of industrial relations
First, looking at collective bargaining coverage, membership of unions and employers' associations has been as an indicator of labour market coordination, although in most European economies, discrimination in wage settlements between unionised and nonunionised workers is not allowed (Visser, 2013), with union agreements extending to a much larger proportion of the workforce. Table 1 shows the disparities between the eight economies. The progressive breakdown of the major collective actors has been a constant in the European economies since the 1980s. This is reflected in the trend in the loss of union membership.
In the period under study, loss in union membership has been particularly strong in Germany and Sweden. The sharp fall in union membership in Germany can be attributed to reunification, when unions in the West absorbed those in the East, where union membership was compulsory, leading to a sharp increase in membership and in the membership rate in 1991 (Jackson & Sorge, 2010). In Sweden, another cause pointed out by the literature is the weakening of the Ghent system. Through this system, unemployment benefits are managed by the union as a tool to increase union membership, especially in periods of low unemployment (Lind, 2009).
A more appropriate measure to approximate the coverage of the system is the number of workers covered by collective agreements, since in some countries membership rates are very low (as in France or Spain, where bargaining coverage is high because it is guaranteed by law, on the erga omnes principle). Collective bargaining coverage thus indicates the proportion of employees who are covered by a collective agreement (whether company, sectoral, regional or national), and makes it possible to approximate the share of workers whose working conditions are regulated on a non-individual basis and through factors not directly related to the market. The rate of collective bargaining coverage has tended to decline since the 1980s in European countries and continued to do so during the period under study. There are also differences across countries, with Sweden having the highest coverage, Spain and Italy with a share of around 80%, followed by Germany, below 60% from 2010 onwards, and the UK in last place, where, in addition, most of the coverage is due to company-level agreements.
However, the variation has been very different in the countries studied: coverage in Greece, where the economy collapsed the most, almost halved between the two periods. Greece is followed by Poland (26.1 points difference), Germany (more than 11.3 points between the two periods), the UK, which already had the lowest coverage, with a fall of 6.6 points and Spain, where it fell by 4.3 points. The smallest drop was in Sweden,  which already had the highest coverage and only fell by 3 points on average between the two periods. In contrast, in France and Italy, coverage remained almost constant over the years studied. Logically, the differences are even more pronounced if we take the extreme values between 1995 and 2018. However, as we are interested in the structure, the more stable components, in this case the averages, are more representative. The degree of coordination of collective actors in wage setting has been measured in different ways. Table 1 also shows the dominance of collective bargaining through centralisation. Except for the United Kingdom, where bargaining normally takes place at the company or establishment level, all other economies are dominated by sectoral bargaining. However, centralised wage bargaining does not necessarily lead to a high degree of coordination, as mentioned above. Thus, the centralisation indicator is based on the dominant bargaining level but adjusted for the incidence and control of additional bargaining at the firm level. Another factor is when and how many central or sectoral agreements allow for additional bargaining at lower levels and the varying extent to which agreements can be avoided by using 'opt-outs' that allow companylevel negotiators to bypass sectoral agreements for coordination purposes. This approach also considers the possibilities for decentralisation, that is the frequency with which what is agreed at the sectoral or national level is renegotiated at the firm level. Additionally, the existence of disconnection clauses or the possibility to opt out of wage agreements and rework them at the firm level, makes it possible to achieve freedom to negotiate more in line with their strategies and needs, with less state intervention.
We find that France, Germany, Italy, Spain and Sweden are the economies with the highest level of centralisation, where bargaining is generally predominantly at sector or industry level, but with the possibility of alternating between sector and company level bargaining due to the above-mentioned clauses. In contrast, Greece, which started from one of the highest levels of centralisation, generally between sectoral and industry bargaining, during the period 1995-2008 and during the post-Great Recession phase is one of the countries with a bargaining level closer to the company level. A similar situation is found in the United Kingdom and Poland, where bargaining is generally at the company level.
Thus, in static terms, most countries have a sectoral level of bargaining (conditioned by the possibilities of application of such an agreement) and only two of them bargain at company level. In dynamic terms, the average value of the index for countries with sectoral bargaining in the post-crisis period is slightly lower in all countries, indicating tendencies towards disorganisation, but there is only a clear change of pattern in the case of Greece. Thus, according to this indicator, disorganisation exists, but the institutional structures of industrial relations show some resistance to change in the level of bargaining (not necessarily in its content).
Consequently, although pressures for institutional change have been directed towards greater decentralisation (Meardi, 2018), there are different collective bargaining structures in Europe that should be taken into account in policy formation, especially by the European institutions.
Finally, the coordination indicator (which captures the level at which wage bargaining most frequently takes place, together with the limits or ability to circumvent what is set at that level through other agreements in firms and workplaces) is based on Traxler et al. (2001). This coordination indicator reflects the mechanism through which coordination of wage bargaining behaviour takes place, which is based on the behavioural patterns or activities of the main actors (unions, employers and governments) involved in wage setting, and is, therefore, an indicator of the collective bargaining process, not of its outcomes.
However, the coordination index scores are not very sensitive to institutional changes. This indicator is constructed based on the expectations generated by the existence of a set of institutional features, that is the mere presence of a set of formal and informal institutions is what allows us to approximate the degree of coordination. As long as these institutions remain present, despite the emergence of dynamics of change, the indicator score will not change. It is, therefore, important to combine the indicator score with a more detailed examination of the facts, which are discussed in the following section.

Institutional developments
In France, the coordination of collective bargaining is carried out through government signals (minimum wage increase, reference wages in the public sector, etc.). France has traditionally had one of the lowest union membership rates in Europe, probably influenced by the fact that collective agreements negotiated by trade unions apply to all employees (erga omnes), regardless of whether they are union members. Thus, workers have little personal incentive to join a union, which focuses mainly on bargaining advocacy rather than service delivery and elections of union representatives in companies and has the capacity to successfully mobilise workers (Pernot, 2017).
There are five national trade union confederations and three existing employers' confederations. The fragmentation and decentralisation (Vincent, 2019) induced over recent decades has led to a collective bargaining system in which a significant part of sectoral collective agreements take the minimum wage as a reference (SMIC) and sectoral agreements are in turn the benchmark for agreements negotiated at company level (approximately 40% of all agreements). This leads to a high dependence on public sector signals and takes away the autonomy of the social partners. However, this industrial relations framework has been transformed by the flexibility-oriented legislative changes introduced in 2016 and 2017. In short, this reform is clearly part of a deregulation process that aims to increase the discretionary power of the company.
In Germany, wage setting has traditionally been coordinated informally through the economy-wide benchmarks set by the agreement between IG Metall and Gesamtmetall in the Baden-Wuttenmberg region, although there are also openness clauses and the possibility to renegotiate wages set in workplace agreements. Changes in the model have been profound since the early 1990s, but important elements of coordination remain. The existence of a dual level of representation, based on sectoral collective agreements negotiated by trade unions on the one hand and works councils on the other, favours coordination in wage bargaining.
In parallel, especially since the Hartz reforms 1 in 2003, a segmented or dual labour market has been consolidated in Germany, with the emergence of a sector of temporary workers and workers not covered by collective bargaining agreements, both with low wages and worse employment conditions, something that affects both the service sector and manufacturing industry (Doellgast & Greer, 2007;Benassi, 2016). This dualisation has favoured the recovery of employment without seriously attacking the traditional institutional core, but rather creating a range of new possibilities for hiring and 1 The labour reforms in Germany, named Hartz I-IV after Peter Hartz, executive director of human resources at Volkswagen and head of the commission that advised the government, brought about profound changes in the German labour market. The management of the public employment service was comprehensively transformed, including the creation of public temporary employment agencies, the provision of individualised care for the unemployed, and the introduction of a new training system. Stricter rules for receiving unemployment benefit were introduced, the period of receipt of contributory benefit was limited and, for the long-term unemployed, a new lump sum benefit (not linked to the last income) was created, which is combined with social assistance payments. On the other hand, social contributions payable by the employee are eliminated for wages below 400 euros (currently 450 euros). For salaries between 400 and 800 euros an increasing scale of contributions was established (mini-jobs are generated: part-time, low-skilled jobs). Among other things, this has led to an increase in the risk of poverty for both unemployed workers and those in part-time jobs. remuneration (Palier & Thelen, 2010), which has led to an erosion of the German model (Hassel, 2014). Thus, while coordination is maintained at the institutional core, the expansion of a low-wage sector may have contributed to the deterioration of economy-wide coordination and the loss of the overall logic of the institutional framework, although the indicator does not capture this dynamic. This process of dualisation means that a growing group of workers is left out of collective bargaining, and other instruments are needed to reduce inequalities between groups of workers.
A process of labour market re-regulation has followed dualisation. The main milestone in this development was the introduction of the minimum wage in 2015. Despite counteracting the above trend, the German institutional structure still retains some of its characteristic features: a dual labour market and industrial relations framework, in which coordination and social concertation mechanisms for a set of insiders, who enjoy greater employment protection and comparatively high wages, coexist with a large segment of workers employed under atypical forms of contracts and who are on the margins of industrial relations. Moreover, the actors retain the decisionmaking capacity for social concertation, using coordination tools at sectoral and workplace level (Herrero, 2021). However, external workers are not covered by these agreements and external flexibility is concentrated on them. It can therefore be concluded that the German institutional structure remains within a process of dualisation.
In Greece, coordination has also been affected as a consequence of the Great Recession, as have all institutional areas and industrial relations. Coordination has shifted from centralised bargaining of the 'inter-association by maximum associations' type to one based on government signals (Katsaroumpas & Koukiadaki, 2019). The majority of trade unions in Greece are members of the confederations, GSEE (General Confederation of Greek Workers), which represents private sector workers, and ADEDY (Confederation of Public Employees), which represents public sector workers. Trade unions are organised at three levels: autonomous trade unions at company level; sectoral or professional federations or regional organisations; and confederations at national level. Employers, on the other hand, are organised in three major organisations: SEV (Hellenic Federation Enterprise), which represents large enterprises, GSEVEE (Hellenic Confederation of Professionals, Craftsmen & Merchants) and ESEE (National Confederation of Hellenic Commerce), which represent SMEs (Kornelakis & Voskeritsian, 2014).
Before the Great Recession, the system was based on the national agreement at the highest level of the General National Collective Agreement (EGSSE), which served as a reference for all lower-level agreements. This implied a wide degree of collective bargaining coverage (up to 100 per cent) fostered by the right of extension and the important role of the Mediation and Arbitration Service (OMED). Following the reforms imposed by the Memorandums of Understanding (MoU) of 2010 and 2012 (Theodoropoulou, 2016), measures were introduced such as the elimination of the application of the most favourable clause, compulsory decentralisation to company level through the primacy of company agreements over sectoral agreements, the possibility for non-union representatives to sign agreements and the forced expiry of sectoral agreements.
These measures eliminated the possibility of EGSSE continuing to act as a centralised coordination system. The establishment of a statutory minimum wage, instead of being signed in the EGSSE, was accompanied by the creation of a tripartite National Social Dialogue Committee. However, decision making was led by the state, conditioned by the supranational Troika 2 authorities, which in practice has led to legally determined wage variations being taken as a reference in the wage-setting mechanism (Koukiadaki & Kokkinou, 2016). The transformation of collective bargaining into a state-led system in the post-Great Recession period can only operate in the direction of wage restraint (nominal and real) given the context of the Troika's intervention. In conclusion, this is a profound deregulation of industrial relations in Greece, which has led to a reduced degree of governance capacity of the system because of the weakening of social actors and the absence of coordination mechanisms outside the market.
In Italy, tripartite bargaining based on social concertation has been very present, with the government playing a central role in setting rules that condition the content of sectoral agreements, but the model has been transformed over time (Pulignano, Carrieri & Baccaro, 2018). There are three major trade union confederations (CGIL, CISL and UIL), but a wide range of employers' associations, the largest being Confindustria (representing large enterprises in manufacturing), Confapi (representing SMEs), CNA and Confartigianato, among others 3 .
Since 1993, the coordination of collective bargaining has been intra-association and takes place at two levels, industry and company, with the clear intention of controlling the evolution of prices. Multi-year sectoral agreements linked wage increases to the expected behaviour of prices to contain inflation, while at company level, wages were linked to the company's productivity growth (Damiani, Pompei, & Ricci, 2020). However, the bargaining role at firm level was weak, due to low union membership rates (Pedersini, 2019). Traditionally, sectoral agreements were governed by the erga omnes principle, even if there was no formal obligation to do so, hence partly one of the reasons for the low relevance of bargaining at company level.
Subsequently, after successive attempts by the main employers' organisation, Cofindustria, to restructure wage negotiations, and failed attempts at tripartite negotiations, reforms were carried out in 2009 with the participation of the trade union confederations, CISL and UIL, while CGIL dissociated itself from the negotiations. This led to the signing of different collective agreements with the various unions in some sectors. In 2011, the path of decentralisation continued, with a key milestone being the company agreement reached by FIAT (Leonardi, Ambra, & Ciarini, 2018) with CISL and UIL, under threat of offshoring, detaching itself from the sectoral agreement and thus breaking the erga omnes principle. In 2011, a tripartite agreement was reached allowing companies to opt out of sectoral agreements if more than 50% of the workers supported this.
Since the economic crisis of the Great Recession, the policies (the Fornero reforms) have continued this trend, representing a disorganisation of the industrial relations system by allowing bargaining units to separate from the common approach, which hampers both horizontal and vertical coordination. At the same time, the coverage of other social benefits (Picot & Tassinari, 2017), such as unemployment benefits and pensions (indirect wages) was maintained or increased. All in all, the liberalisation of industrial relations in Italy can be seen as a response to a dualisation in so far as the Italian labour market is segmented.
In the case of Poland, coordination is of the same type as in the UK, based solely on company-level bargaining and with no notable changes during the period in the scope of coordination. With the labour law reform of 1996, a collective bargaining mechanism was introduced without state influence, which is limited to establishing common minimum standards in labour law. Despite the almost non-existence of sectoral or national agreements, a considerable number of the numerous company unions are affiliated to two trade union confederations: the Pan-Polish Trade Union Alliance (OPZZ) and Solidarnosc. It is important to note that these two confederations are the result of very different dynamics: while the former represents Soviet-era trade unionism, the latter led the movement for political change (Czarzasty, 2019).
Poland has implemented liberalising labour market reforms that have resulted in a drastic increase in temporary employment, generating a dual labour market, as in other European economies (Maciejewska, Mrozowicki & Piasna, 2016). This completely decentralised industrial relations system and fragmented labour market, with the stated aim of providing flexible labour to attract foreign direct investment, has resulted in a highly fragmented and politicised model with state-led social dialogue (Bernaciak, 2017). Institutional change in Poland can therefore be seen as liberalisation by deregulation. The coordination mechanism has been undermined by highly fragmented bargaining units and state pressure for wage moderation.
In Spain, the type of coordination has been transformed from intra-association to inter-association by associative peaks, based on the coordination of employer's confederations and trade unions and equivalent to that of central-level bargaining. National agreements, with strong state intervention (renewed at variable intervals), and essentially with four partners, play a major role. These four partners are the employers' association of large companies (CEOE), the employers' association of small companies (CEPYME) and the two major trade unions: the Trade Union Confederation of Workers' Commissions (CCOO) and the General Workers' Union (UGT). As in the French case, the erga omnes principle applies, which creates fewer incentives for union membership and focuses union activity on representation and elections (where participation is relatively high).
Nonell, Alós-Moner, Artiles and Molins (2006) analysed vertical and horizontal coordination before the economic crisis of the Great Recession and concluded that governance in the industrial relations system rests on specific legal and cultural elements. However, wage bargaining mechanisms have been conditioned by competitive social concertation (Fernández, Ibañez & Martínez, 2016), according to which wage moderation was a necessary condition for achieving macroeconomic stability.
After the Great Recession, with the reforms of 2010 and 2012, the level of centralisation decreased drastically by generalising the possibility of non-application of the sectoral agreement through much more lax criteria than those that existed previously (Molina, 2014). After 2012, non-application became a unilateral decision of the company which establishes that an agreement that is not renewed cannot have its conditions extended beyond one year. Thus, if the agreement is not renewed, a higher agreement will be applied, even applying the basic legal minima (Workers' Statute), and, therefore, harsher conditions than those that apply in company agreements (longer working hours, lower wages, etc.). All this has led to a significant weakening of working conditions and a marked devaluation of wages as a result of the reduction of workers' bargaining power. In 2022, a reform was approved that modified some of the previous changes, which points to a re-regulation process, but it is too early to see its effects. In conclusion, Spain's trajectory is therefore one of deregulation of industrial relations.
In Sweden, the system of nationally centralised wage bargaining arising from the Saltsjobaden Agreement, which allowed for equal wage growth with reduced wage drift to avoid inflationary pressures, and which had characterised the Golden Age era, was transformed in the wake of the 1991 crisis (Pontusson, 1992). Thus, in the early 1990s, a sectoral collective bargaining system combining greater decentralisation with the restoration of coordination was gradually established, and the principles of wage restraint and equality characteristic of the model continue to be respected, albeit less rigidly. This is reflected in the transition from an intra-organisational model, where top-level associations do not enter centralised negotiations, but negotiations take place at a lower level, to one of employer bargaining. As in Germany, wage setting takes place at a double level: at sectoral negotiations, led by national-sectoral trade unions and employers' organisations, and at local level, with local branches of trade unions and individual employers or on an individual worker-employer basis. Union representation is organised by type of occupation, separating blue-collar workers (represented by the Swedish Trade Union Confederation -LO), white-collar workers (represented by The Swedish Confederation of Professional Employees -TCO) and professional workers (represented by the Swedish Confederation of Professional Associations -Saco), who sign their own local agreements in each workplace.
Wage bargaining follows a wage review model that obliges workers and employers to discuss regularly before wages are set, and establishes new roles for management, the local union and the worker. Managers are responsible for setting wages and ensuring that wages create incentives to work; local unions have a role in supporting workers prior to the wage review; and individuals must have the opportunity to discuss wages and other benefits (Granqvist & Regnér, 2008). The role of the welfare state as a counterweight to corporate power has also deteriorated, which has changed the power resources of workers and employers (Buendía & Palazuelos, 2014). Occupational shifts away from manufacturing and increasing wage dispersion have reduced the hegemonic position of trade unions within the Swedish trade union structure. As a result, the solidarity-based nature of the Swedish labour market has been severely limited and the differences between unionised and non-unionised workers have deepened, resulting in a process of dualisation of collective bargaining.
In the United Kingdom, from the 1980s onwards, following the wave of deregulation promoted by the Thatcher governments, sectoral bargaining and collective agreements (as well as regulatory institutions such as Wages Councils) were abandoned and a highly individualised and marketised wage bargaining system was established (Howell, 2005). Thus, there was no longer any organised coordination mechanism in the institutional structure of industrial relations, but one based on company-level bargaining. This was accompanied by a fall in union membership and collective bargaining coverage rate (from 70% to about 25%) in the private sector; in contrast, the public sector maintains high union coverage and membership. Despite this decline in collective bargaining participants, there has been a trend towards a merger of organisations representing workers (with the creation of UNITE 4 , for example). Large employers are represented by the Confederation of British Industry (CBI).
Wage bargaining is therefore carried out in a fragmented way at the workplace and with a single employer, which explains the low coverage rates. Existing collective agreements are signed at company and workplace level, and this is where trade unions play a more prominent role; however, these agreements are not binding and do not have to be enforced and, therefore, do not necessarily have to be complied with (Visser, 2013). Moreover, isolated workplace or individual wage agreements are not coordinated between the different bargaining units, among which market relations predominate and cooperation is residual. This leads to a pattern of greater wage and income inequality in the absence of institutional mechanisms to compensate for market wage differentials.
Facing decentralisation, the Labour government established in 1999 a national minimum wage, which must be respected. It resulted in lower wages growing significantly, compensating for the effect of less coordination. This can be linked to the fact that there are informal mechanisms of horizontal coordination in wage increases based on the benchmarking system in which firms tend to imitate each other's wage increases (Schmidt & Dworschak, 2006). After 2008, further changes were introduced in the industrial relations framework, when the passing of the Trade Union Act in 2016 enacted a new wave of restrictions on the right to strike. All in all, the liberalisation of the UK labour market is mainly about deregulation.

Conclusions
In conclusion, combining the above quantitative indicators with the qualitative aspects pointed out in the literature, it can be affirmed that the first hypothesis put forward is confirmed. First, in the economies studied, important differences are observed between institutional structures, which ultimately reflect the different forms of coordination between economic actors in the labour market.
In this case, Sweden and Germany appear as the economies with the highest levels of coordination in wage setting, based on pattern bargaining models. Italy and Spain base their coordination on relatively centralised sectoral agreements, but these have been eroded by the latest labour reforms since the Great Recession. In contrast, in the other 4 Unite is a British and Irish trade union that was formed in 2007 through the merger of Amicus and the Transport and General Workers Union (TGWU). Unite is the second largest union in the UK, with over 1.2 million members in the construction, manufacturing, transport, logistics and other sectors. countries, company-level bargaining predominates. On the one hand, in France and Greece, company-level bargaining is coordinated by the state (albeit in different ways) and, on the other hand, the traditional Anglo-Saxon model of individualised, marketoriented industrial relations is evident in the UK and Poland. Thus, the institutional structure of industrial relations coordination differs significantly from country to country, but it is also relatively stable in most countries. This points to the need to identify well the existing institutional frameworks when designing public policies.
Second, while differences between countries remain, there have also been profound transformations in the ability of collective actors to coordinate and agree on wages and working conditions as a result of the labour market liberalisation of the post-Fordist period. However, each of these countries has had a different liberalising trend through decentralisation and disorganisation, constituting elements that affect the coverage, dominance and control capacity of collective bargaining.
On hypothesis 1.2, the disorganisation of collective bargaining, the average value of the index for countries with sectoral bargaining in the post-crisis period is slightly lower in all countries, indicating that there have been trends towards disorganisation, but only a clear change in the pattern in the case of Greece. Thus, according to this indicator, disorganisation exists, but the institutional structures of industrial relations show some resistance to change in the level of bargaining (not necessarily in its content).
This is an important conclusion as it helps to point out that institutional change has increased the disorganisation of collective bargaining, which makes the industrial relations system less governable. This is particularly relevant in the face of major economic challenges such as the COVID-19 crisis.
Regarding the second hypothesis, it can be stated that liberalisation is unquestionable and has a common orientation in all the economies studied, as Baccaro and Howell (2017) have argued, with declining coverage and dominance in all the economies analysed. However, relevant differences are also found in the forms of liberalisation. Thus, countries such as France, Greece, Poland, Spain and the UK have a deregulation trajectory at the core. This implies that liberalisation has affected central aspects of their industrial relations system. The most dramatic case is Greece, which has had deep deregulation in all areas. In Poland and the UK, although they already had levels of enterprise bargaining, important aspects such as the right to strike, restrictions on funding and trade union resources have been modified. Finally, in France and Spain the effect has centered on the reduction of employment protection, especially severance pay and dismissal procedures, which have eroded the core of stable employment with greater union coverage. This suggests that the weakening of collective bargaining institutions has meant greater decision-making power for employers, but it has also eroded the capacity for horizontal and vertical coordination among employers themselves.
By contrast, in Germany, Italy and Sweden liberalisation has led to a stratification of industrial relations. In the case of Germany and Sweden, usually identified with models of greater coordination, the erosion of their industrial relations systems has come mainly from the reduction of trade union membership (and to a lesser extent also from the fall in collective bargaining coverage). In the case of Italy, although various changes, such as the Fornero reforms 5 , have affected important aspects of the industrial relations system, these have been minor and concentrated on aspects at the margin. The existence of workers with relatively stable collective bargaining mechanisms combined with a growing group of workers exposed to the market makes it difficult for bargaining outcomes to be generalised to a majority, resulting in a segmented pattern.
In sum, this analysis of comparative industrial relations shows that the transformation of the major collective actors, especially the trade unions, and their ability to act as referents in collective bargaining as it has been practised in the past, is a fundamental consequence of liberalisation in whatever form. In all European countries, to a greater or lesser extent, liberalising institutional reforms have taken place and this has weakened the bargaining position of trade unions. These have also affected the very form of coordination of the collective bargaining mechanism, as the different ways of untying the sectoral agreement have weakened sectoral bargaining in favour of company bargaining.
Our conclusions point to the need to strengthen the institutional frameworks in which collective bargaining can take place in order to bring together the different interests of trade unions, employers and policy makers. The relevant contribution is that there are a number of institutional requirements that are necessary for the industrial relations system to be governable. This is why it is necessary to introduce legislative changes that provide stable bargaining spaces for the social partners, because in their absence, it is much more difficult or even impossible to bring together the different interests both within and between the different groups.