Institutional resources as a source of trade union power in Southern Europe

Institutional resources are one of the sources of power available to trade unions, but recent literature has tended to pay less attention to these than to associational and organizational resources. We examine institutional resources in three Southern European countries, Greece, Portugal and Spain, which share many common characteristics. However, the character of institutional resources in Spanish industrial relations is distinctive. We examine the plasticity of industrial relations institutions in Spain in terms of labour market outcomes but argue that institutional security is an essential platform for unions seeking to develop other sources of power.


Introduction
Analysis of the power resources available to trade unions has become increasingly relevant in the difficult environment in which unions in Western Europe operate today. Four forms of trade union power have typically been identified: structural, associational, organizational and institutional (Gumbrell-McCormick and Hyman, 2013). As structural power has declined, Anglo-American literature on trade union renewal has tended to focus on associational and organizational power (Bryson et al., 2011). Discussion of institutional power resources, while stressing their importance for trade union security, has tended to be limited and to emphasize their potentially problematic nature.
In this article, we examine institutional resources as a source of union power in Spain, Portugal and Greece. The three countries share common features. Their industrial relations systems were initially constructed in the relatively recent past as part of the post-dictatorship institutional settlements. The new democracies displayed a high degree of state intervention, with a regulatory framework supportive of the trade unions (Barr and Mackie, 2013). The economic crises recently affecting all three countries have resulted in calls from international actors for these settlements to be unravelled. Their economies also display similar characteristics, with a preponderance of small and medium-sized enterprises (SMEs), late and incomplete industrialization, a dependence on seasonal sectors such as tourism and agriculture and an important role for the informal economy. These common characteristics facilitate a stronger focus on the particular configurations of institutional resources in the three countries; doing so in a period of crisis provides an opportunity to interrogate in more detail the potential and limits associated with institutional power.
We first review literature on institutional resources as a source of trade union power, outline the institutional resources which trade unions have historically enjoyed in the three countries and consider the impact of the economic crises on these resources. After discussing the reasons for differences in their resilience, we reflect on the implications of this article for our understanding of institutional resources as a source of union power.
Our principal propositions are that the institutional resources of the Spanish unions have proved the most resilient during the period of economic crisis and that this resilience remains a source of strength for the Spanish unions, notwithstanding the limitations in the substantive outcomes. We argue that this resilience can be attributed to two factors. The first is what we term the Durkheimian qualities of the institutions, which do not rely on self-interest and are considered normatively binding. Durkheimian institutions impose obligations which are enforced by a third party, typically governmental. They can be distinguished from coordinating institutions constructed by market actors to increase the efficiency of exchanges (Streeck, 2009). The second is the reluctance of employers to challenge these institutions because they provide resources compatible with their own long-term interests and they possess alternative avenues to achieve flexibility in a segmented labour market.

Institutional resources as a source of trade union power
Institutions, defined as the norms and rules and resulting constructs which shape the interaction of economic actors and their associated interests, are usually considered in industrial relations literature as key factors conditioning trade union influence. There is little disagreement that it is difficult for unions to operate without a degree of institutional security, which depends to a considerable extent on the rights generated by governments, making inevitable the engagement of unions in politics (Gumbrell-McCormick and Hyman, 2013;Masters and Delaney, 2005). The importance of particular types of institutions has been emphasized. Thus, it is argued only a system of multi-employer bargaining can cover most employees and ensure the involvement of the unions in matters of economic and social policy (Traxler, 2007).
In the later 20th century, academic research placed considerable emphasis on the role of institutions in conditioning industrial relations outcomes. For example, Clegg (1976) related outcomes such as union membership density and strike behaviour to the institutional configurations of national industrial relations systems. Hyman (2001), reviewing this literature, makes the distinction between 'old' institutionalism, involving a high degree of structural determinism, and neo-institutionalism which leaves more space for actor impact and choice.
More recent Anglophone research on trade unions has been less interested in institutional analysis. The deregulation and institutional displacement associated with neoliberalism in the United Kingdom and the United States and the resulting 'deliberate neglect' of institutions (Streeck and Thelen, 2005) have encouraged a focus on union renewal which emphasizes actor autonomy in strategies such as organizing and social movement unionism, fitting more closely the traditional culture of voluntarism in those two countries. Bryson et al. (2011) question the relevance of this research for unions beyond the Anglophone world, as many European trade unions face a different set of choices, embedded as they are in social, political and economic institutions.
However, political economists often analyse industrial relations systems as an important dimension of discussion, as in the varieties of capitalism literature (Hall and Soskice, 2001). Relevant in this context is the contrast by Streeck (2009) between Durkheimian and Williamsonian institutions. The former do not rely on self-interest and are normative and obligational, often enforced by government and constitutional processes; the latter are voluntary and are constructed from below by self-interested market participants to increase the efficiency of their exchanges, making cooperation possible in their longterm interests. Streeck discusses the transformation of German industrial relations institutions from a Durkheimian to Williamsonian model via a process of disorganization from below, which has involved decentralization, fragmentation, increased flexibility and voluntary employer participation. Examples of this process are the establishment of opening clauses in collective agreements and of sections in employer associations for companies not wanting to be bound by collective agreements. However, it is not always easy to classify institutions in these terms. It can be difficult to identify actor motivations, and both normative and self-interested action can, at the same time, be the basis of institutional resilience (Etienne and Schnyder, 2014).
Researchers who have considered the role of institutional resources in supporting trade union influence have often pointed to the problems of relying on such resources. For example, achieving and maintaining institutional resources typically involve political alliances. This can result in unions and workers generally concentrating too much on political routes to achieve their objectives, to the neglect of their role in resolving issues in the workplace and developing organization, reducing their visibility on a day-to-day basis. The history of French industrial relations provides an example (Parsons, 2013). Unions can become too dependent on particular political allies, particularly if they become organizationally weak. The difficulties of such dependence have been emphasized by the rightward movement of social-democratic parties, weakening their ties with the unions, which have become less attractive (and necessary) as allies as they have become weaker (Masters and Delaney, 2005). Even where union ties with social-democratic parties are still important, they have become less significant as the strength of these parties has been eroded, as in Sweden (Gumbrell-McCormick and Hyman, 2013).
There are also dangers in placing too much emphasis on institutions themselves rather than what they deliver. In the German context, Kinderman (2005) stresses that continuity of structure should not be confused with continuity of content, pointing to the ways in which employer strategies are changing the effects of the structures of the industrial relations system and weakening the traditional works council role. Baccaro and Howell (2011) similarly argue that there has been institutional change on an incremental basis (institutional plasticity), with the result that the form of the institution may remain but the content is changing. As evidence, they point to the de facto decentralization process taking place in France and the growth of firm-level agreements.
For other writers, dependence on traditional institutional power resources can create conservative and inward-looking trade unions. Hassel (2007) sees institutional security as a liability for German unions: the institutionally secure environment and lack of competition enjoyed by the manufacturing-based industrial unions give them little incentive to recruit new groups of workers. Institutional resources are a double-edged sword, stabilizing union organization but stopping unions from recruiting new groups by establishing monopolies, diminishing competition and guaranteeing workplace representation. Similarly, Thelen (2012) suggests that the solidarity of German institutions has declined and 'manufacturing firms and their workers can be expected to jointly defend traditional institutions and practices for themselves while on the unorganized periphery new … patterns of employment emerge' (p. 154). This division into insiders and outsiders means divergent goals emerge for different segments of workers. Institutions need to adapt to changed circumstances, embrace vulnerable groups and accommodate wider coalitions, ensuring that greater flexibility is balanced by compensatory social policies. This may involve a wider campaigning role for the unions, building alliances with collectives outside their traditional constituencies and thus 'reasserting their identity as "sword of justice"' in order to 'help construct a new type of politics … engaging with campaigning and protest movements' (Gumbrell-McCormick and Hyman, 2013: 203-205).
Finally, too great a reliance on institutional power resources alone makes it difficult for unions to translate formal agreements to the workplace. Thus, the impact of extensive bargaining coverage resulting from multi-employer agreements and extension mechanisms will be limited unless it is supported by a level of membership density and organization which can make the bargaining effective (Vernon, 2006).

The institutional resources of unions in Southern Europe before the economic crises
We now examine the development of the institutional resources of the unions in three areas: the regulatory framework, the collective institutions of bargaining and representation, and corporatist institutions of social dialogue.
The Spanish unions sought the establishment of a strong regulatory framework in which the union role was assured after the death of Franco. A fragmented, SMEdominated economy with a high unemployment rate made it difficult for unions to exploit structural power resources or rely on associational and organizational strength. New organizations had to be established rapidly at a time when the new democracy was by no means secure, and expectations by the unions were high.
As in Greece and Portugal, the new regime inherited from the dictatorship a strong regulatory culture for industrial relations, typical of Mediterranean models of capitalism (Amable, 2003). However, the political position of the Spanish unions was strong enough to ensure their interests were reflected in the new institutional and legal framework -the Estatuto de los Trabajadores (Workers' Statute) of 1980 -helped by the institutional vacuum and the weakness of the early democratic governments (Hamann, 2012). The unions, having played an important role in opposition to the dictatorship, were seen as a major part of the political settlement. In Greece and Portugal, the position of the unions during the transition period was different.
In Greece, the unions, handicapped by their lack of political independence and by political rivalry, were not as directly involved in the transition process, which did not result in a complete break with the previous regulatory regime but the continuance of centralized bargaining, with provision for state intervention through arbitration. Similarly, in Portugal, the trade unions were not as directly involved in regime change: the transition was more abrupt than the negotiated process in Spain. The Portuguese trade union movement, which had before the end of the corporatist regime achieved official status and a significant role in sectoral bargaining, concentrated in the early transition years on demanding more government intervention to legislate on working conditions, rather than constructing industrial relations institutions.
The regulatory framework which emerged in Spain provided for works committees (comités de empresa) with full bargaining powers and the right to call strike action in companies with 50 or more employees and for delegates to be elected in smaller companies. Unions which obtained 10 percent of the elected delegates in a sector were recognized as 'representative' for negotiating purposes. The status of 'representative' union at national level was allocated to those union confederations achieving 10 percent of works committee delegates, with attendant advantages of state financial support and involvement in national social dialogue. Legislation provided for the automatic extension of collective agreements to all relevant employees, and when agreements expired, their terms were extended until replaced by a new agreement. Most employees' terms and conditions were covered by multi-employer agreements, an exception being those employed in large companies which completed enterprise agreements. Labour courts were available for the unions to enforce these rights and wider obligations, such as the need for employers to bargain in good faith. They also adjudicated issues arising from the complex system of employment protection.
The new Greek and Portuguese regulatory models had general similarities to those in Spain, but there were important differences of detail. Portuguese legislation provided for worker representative bodies elected at company level (comissões de trabalhadores), but these had more limited bargaining rights than their Spanish counterparts and did not provide the basis for union recognition in bargaining and social dialogue (Pinto, 1990). The extension of collective agreements was not automatic but had to be activated by ministerial decree. In Greece, legislation in 1988 provided for works councils (symvoúlia ergazoménon) to be established in enterprises with more than 50 employees, but their powers were limited to consultation and they did not develop a strong role or a basis for union recognition outside the workplace. As in Portugal, there was provision for the extension of collective agreements but dependent on ministerial decree and an agreement already being binding for companies employing 51 percent of workers in a sector.
The collective bargaining institutions which developed in all three countries showed similarities, with an emphasis on multi-employer bargaining and relatively high levels of bargaining coverage, helped by extension provisions. In Spain, the coverage rate fluctuated between 70 and 80 percent (Pitxer i Campos and Sánchez Velasco, 2008). Pre-crisis bargaining coverage for Greece and Portugal has been estimated at 65 and 60-70 percent, respectively (Fulton, 2013;Schulten et al., 2015). In Greece, a significant role was played by third-party intervention. Initially, compulsory arbitration was used to resolve disputes, but even after its abolition in 1990 and the establishment of an independent mediation body (OMED), 25 percent of agreements between 1992 and 2008 were a product of an arbitration process favourable to the unions (Koukiadaki and Kokkinou, 2016).
While worker representation systems at enterprise level played a limited role in the Greek and Portuguese industrial relations systems, in Spain their importance was far greater. Workplace elections were dominated by the Comisiones Obreras (CCOO) and Unión General de Trabajadores (UGT), giving them a decisive role in both collective bargaining and social dialogue institutions. Electoral participation was typically high, around two-thirds (Jodar et al., 2012), giving Spain the fourth highest level of worker representation in Europe (Kankaraš and Van Houten, 2015). All three countries had low levels of union density immediately before the economic crises (Greece, 23 percent; Portugal, 21 percent; and Spain, 17 percent; OECD, 2014). The emphasis of the unions on the works committee system contributed to the low membership in Spain: CCOO (2016) has typically spent 10 times more resources on winning workplace elections than on gaining new members.
A third source of institutional resources in the three countries has been involvement in national social dialogue. In the 1980s, peak-level agreements between the Spanish social partners covered employment legislation and pay, but social dialogue on pay was weakened by the rivalry between socialist-and communist-aligned confederations. However, from the 1990s, bipartite agreements on pay signed by both confederations became the norm, and the unions consolidated their role in a number of corporatist institutions. The Pacto de Toledo in 1995 established an all-party agreement on how key decisions on the social security system including pensions should be taken, setting in motion a number of tripartite pacts on pensions. Consultation took place on key labour market issues such as the minimum wage. Bipartite bodies funded by the state were developed in such areas as training, collective and individual conflict resolution. The development of social dialogue at national level was mirrored by similar institutions in the autonomous regions. The limits of corporatism were evidenced by periodic crises and occasional 1-day general strikes, but the fragmentation of corporatist relations into different negotiating arenas meant that a blockage in one did not necessarily paralyse dialogue in others. Unity among the actors helped give some stability to social dialogue. There has only been one major organization of employers since the transition, the Confederación Española de Organizaciones Empresariales (CEOE), and from the 1990s, unity of action also developed between the two major union confederations as they distanced themselves from political alignments (which, in turn, facilitated dialogue with governments of different political hues). Further stability was also given to social dialogue by the central authority of the confederations.
As in Spain, national social dialogue gained importance in Portugal after the dictatorship in the form of tripartite pacts on incomes policy. However, there was a difference in the stability and reach of the dialogue. Convergent union perspectives resulted in both Spanish confederations signing agreements from the 1990s, but in Portugal the two main confederations often went their different ways, retaining stronger attachments to their initial ideological references than their Spanish counterparts. The communist-oriented Confederação Geral dos Trabalhadores Portugueses (GGTP), the largest confederation, signed only 6 of the 17 pacts up to 2008.
Greek social dialogue yielded disappointing results, despite efforts to promote it (Tsarouhas, 2008). Apart from the national collective labour agreement between union and employer bodies, which established the minimum wage, there was little evidence of concertation. The politicization of the trade union movement and its division into different factions, reflecting party loyalties, coupled with frequent changes of government, did not provide the conditions for the development of a mutual trust which could sustain social dialogue. The structure of the Greek union confederation, the General Confederation of Greek Workers (GSEE), with some 100 autonomous affiliates, also weakened its credibility as an organization able to deliver the results of social dialogue. Employers were similarly fragmented, with the main federation the Hellenic Federation of Enterprises (SEV) divided between individual companies and local and sectoral organizations. Hence, although the dominant role of state intervention, typical of the early years of democracy, was modified, social dialogue too often 'rubber-stamped' decisions taken at state level which were only communicated to the social partners.

Institutional resources and economic crises
All three countries had to respond to the demands of external agencies, in return for aid, as a result of economic crises beginning in 2008-2009. Greece and Portugal were subject to a government debt bailout and Spain to a bank bailout; it is difficult to assess the implications of this difference since Spain was forced to make commitments in secret letters in return for the bailout and the purchase of government bonds by the European Central Bank. All the countries had to sign memoranda of understanding with the Troika laying out specific economic and social reforms to be implemented.
These all involved increased flexibility and decentralization, with limitations on the validity of collective agreements after expiry, a priority for company over multi-employer agreements, the extension of opening clauses enabling companies to deviate from sectoral agreements under certain conditions and the reduction of protective employment legislation. However, there were differences. Restrictions on the extension of collective agreements were confined to Greece and Portugal. No changes were introduced to the worker representative system in Spain, while in Greece the arbitration system which had favoured the unions was modified (Koukiadaki and Kokkinou, 2016).
Union density remained low in all three countries during the years of crisis. In 2013, it remained at 17 percent in Spain and declined by 2 percent in Greece and Portugal to 21 and 19 percent, respectively (OECD, 2014). However, when we consider other dimensions of industrial relations, differences emerge. In contrast to the stability in Spain, Greece and Portugal were among the countries experiencing the heaviest losses in bargaining coverage during the period of economic crisis (Visser, 2016). In Greece, coverage declined to 40 percent. Although the national collective agreement establishing minimum pay survived, there was evidence of a rapid decline of sectoral agreements (see Table 1). The degree of loss of coverage in Portugal is disputed (Addison et al., 2015;Schulten et al., 2015;Távora and González, 2016). However, even accepting claims that dramatic falls in coverage to under 10 percent are misleading because they fail to take into account previous agreements still in effect, indices do suggest a significant decline in coverage. Table 1 indicates the decline in multiemployer collective agreements published by the Ministry of Labour during the crisis. There has not been a commensurate increase in company agreements to compensate for this. Since 2011, the number of annual extension decrees has varied between 9 and 17, compared to an annual average of 170 between 2000 and 2010 (Schulten et al., 2015).
Formally, the involvement of Spanish trade unions in social dialogue at national level and in the autonomous regions has not been dramatically affected by the crisis. National agreements recommending pay increases have still been negotiated throughout the crisis years. The unions are still represented in the same bipartite and tripartite institutions as before, and there has been continuing social dialogue with the government and other social partners. Inevitably, the economic crisis has had an impact on the results of this social dialogue. For example, the funding allocated to the occupational training system Source: Addison et al. (2015), Koukiadaki and Kokkinou (2016) and Ministerio de Empleo y Seguridad Social (2017).
managed by the social partners has been cut, although the unions have not been excluded from seeking funding to organize training. The state funding directly allocated to the representative unions was reduced by 44 percent during the crisis (together with that to political parties). It is difficult to point to examples where social dialogue with the government has had much impact, with trade union initiatives tending to be disappointing in their results, as with the relatively low take-up of union-inspired government initiatives on financial aid to the long-term unemployed. In Greece, the already fragile nature of social dialogue was accentuated during the years of crisis. The politicization and fragility of the Greek unions which had inhibited national social dialogue continued to make social pacts difficult to broker during the crisis. This played a part in the social partners being sidelined by the Troika more than in other countries and being consulted less about labour market reform (Koukiadaki and Kokkinou, 2016). As a result, crisis negotiations were monopolized by the state, which further entrenched its dominant role in industrial relations. Evidence of this was the pressure it applied in the national collective agreement to restrict increases negotiated to minimum pay rates (Voskeritsian and Kornelakis, 2011).
Portugal had a more established history of social pacts before the crisis, as noted earlier. National dialogue has continued, but union divisions continued to affect its efficacy. The Confederação Geral dos Trabalhadores Portugueses (CGTP) rejected the intervention of the Troika and the associated government proposals, concentrating more on political and social mobilization. Therefore, dialogue tended to be monopolized by the smaller União Geral de Trabalhadores (UGT), although its patience was tested and on occasion it mobilized in opposition to the government, participating for the first time in a general strike against a socialist government in 2010 (Campos Lima and Martin Artiles, 2011). Nonetheless, despite these divisions, the input of the social partners into government decision-making during the crisis was greater than in Greece, and this, it has been argued, contributed to a less severe impact on bargaining (Távora and González, 2016).
The role of institutional resources during the periods of economic crisis in the three countries must be considered not just in respect of continuity but also in relation to labour market outcomes. This is inevitably a complicated process. Table 2, comparing changes in four indices, shows a mixed picture. Although the level and purchasing power of the minimum wage and the performance of real pay has been stronger in the case of Spain, labour market segmentation is a greater problem, with proportionately more workers employed on temporary contracts. It is, therefore, difficult to argue that the stability of institutional resources in Spain has been translated into better labour market outcomes overall. Positive comparisons in relation to pay have to be balanced against poorer results in respect of precarious employment.

Discussion
Our comparison of the three crisis-affected countries has indicated a greater stability of institutional resources in Spain (less regulatory change, more continuity in collective institutions and in social dialogue). This greater stability can, in part, be related to their Durkheimian characteristics. The crisis left intact key obligational elements of the regulatory framework: the automatic extension of multi-employer agreements, the worker representation system and 'representative' unions. The democratic connection between works committees, the worker representative system, representative unions and the collective bargaining structures and their links to elements of the Spanish Constitution made any fundamental change to the rules of bargaining both complex and problematic. The low take-up by employers of the provision to opt out of sectoral agreements exemplifies this: employers were discouraged by the way the labour courts interpreted the need to justify and negotiate opt-outs with worker representatives and to negotiate in 'good faith', as defined by the constitution.
Normative elements of the post-dictatorship settlement exercised an influence on the government and the judiciary. Thus, the conservative Partido Popular government, notwithstanding its overall majority and external pressures, held back from radical action, leaving intact key elements of the 1980 Estatuto. With determination, it could have drafted legislation more tightly, particularly in the face of unfavourable court decisions (Gómez, 2015). Subsequently, the government has continued to resist demands from the Troika to introduce more flexibility into collective bargaining, even claiming credit for the continuing vigour of the bargaining system. Its restraint can be attributed to the continuing elements of political capital located in the post-Franco settlement. In respect of the judiciary, the 'denseness' of the existing regulatory framework, as well as the judges' traditional inclination to be favourable to worker interests, enabled unions to dilute the impact of the reform package in the labour courts (Gómez, 2015). There was considerable opportunity for legal challenge and for individual labour courts to arrive at their own interpretations of legislation, practising a form of judicial populism (Ruiz Soroa, 2016). Thus, expiring collective agreements were allowed to continue on the grounds that they were regulated by previous legislation, an interpretation eventually upheld by the Supreme Court (Durán López, 2015).
However, the Durkheimian character of institutional resources in Spain does not alone adequately explain their resilience; this also reflects the material interests of both unions and employers. The unions wished to maintain an industrial relations system which ameliorated the impact of an economy dominated by SMEs. During the crisis, the unions have continued to display the negotiating culture which they had embraced since the late 1980s, maintaining dialogue even at times of conflict. More explicit concession bargaining has been pursued to ensure continuity. Recent interviews with worker representatives have emphasized that their bargaining priority was to preserve their collective agreements,  Busch et al. (2013), Pavelka (2013). even at the cost of a devaluation of their contents, for example, through salary moderation or freezes (Lago Peñas and Cruces Aguilera, 2016).
The commitment of employers to existing institutional resources and their lack of interest in pursuing a process of disorganization merits more detailed consideration. A considerable part of the purpose of Spanish employer organizations lies in their role in collective institutions. They have always been ambivalent about the decentralization of collective bargaining, being more concerned about flexibility in relation to dismissal (Hamann, 1998). Thus, despite the opportunities offered by the new legislation, employers' organizations reached agreements with the unions to extend the life of agreements (Fernández Rodriguez et al., 2016).
At enterprise level, Spanish firms have traditionally sought to seek to minimize uncertainties in the labour market by maintaining multi-employer bargaining (Royo, 2007). For the SMEs which make up most of the economy, bargaining decentralization could be a disadvantage. They lack the resources to devote to local bargaining and the worker representatives with whom to negotiate (Malo, 2015). Sectoral agreements avoid the transaction costs of negotiating employment conditions company by company but do not disturb paternalistic management practices (Meardi, 2014). When asked why they had not instituted a company agreement, 73 percent of employers said that the sectoral agreement met their needs, 15 percent that they were unable to do so because no worker representatives existed in the company and 10 percent that they did not know how to (MESS, 2014). However, the willingness of employers to support the collective system needs also to be understood in the context of the numerical flexibility and structural power offered by the segmented Spanish labour market. Spanish employers could support Durkheimian institutions because the labour market offers them an alternative flexibility.
In Greece and Portugal, the institutional resources available to the trade unions were not sustained to the same degree. In both countries, the resources were less obligational, particularly in respect of collective bargaining extension and the role of worker representation. There was little evidence for the normative reach of the resources, and actor commitment was weaker. Union disunity contributed to this, but employers did not show the same desire to maintain collective institutions exhibited by their Spanish counterparts. The removal of the extension mechanism in Greece meant that many employers stopped participating in bargaining, worrying they would lose a competitive advantage against smaller firms paying less and using undeclared work (Koukiadaki and Kokkinou, 2016). Zambarloukou (2010) shows that in Greek banking, employers sought to break up sectoral negotiations, and in telecommunications, a newly privatized and fragmented sector, they had little interest in developing them. In Portugal, Tavora and González (2016) discuss the difficulties unions have had in reaching sectoral agreements with employers' associations in chemicals and textiles and footwear.

Conclusion
We have shown a greater degree of continuity in the institutional resources available to trade unions in Spain than in Greece and Portugal. We have argued that this is in part a result of the framing of these institutional resources in Spain and their Durkheimian qualities. The linking of a robust worker representation system to the actors in collective bargaining establishes their democratic credentials and limits the political scope for reform (as well as facilitating unity of action and mobilizing capacity). The obligational character of institutions, such as extension provisions, limits disorganization. Therefore, the detail of institutional resources is important, particularly when, as in the Spanish case, it is based on a settlement which continues to exercise some normative influence on key actors such as the government and judiciary.
However, the continuity of the Spanish institutions also owes much to their ability to satisfy the Williamsonian self-interest of united trade unions and employers for collectivist institutions, emphasizing that the same institutions can be based on both normative and self-interested action. As a result, there has not been an appetite in Spain for a process of disorganization. Equally, a highly segmented labour market, established initially in the 1980s, has met employer needs for flexibility, reducing any pressure to modify drastically the system of multi-employer bargaining.
Our analysis inevitably invites the question of the value to the Spanish unions of institutional power resources which, although resilient in structure, have failed to prevent a high level of precarious employment and a significant reduction in real pay, emphasizing the need to distinguish between continuity of structures and their outcomes (Baccaro and Howell, 2011;Kinderman, 2005). In response, we argue that institutional continuity has been beneficial for the Spanish unions for the following reasons.
First, the maintenance of these structures, particularly multi-employer bargaining, has been essential to provide the Spanish unions with a platform from which to operate. Union presence would quite possibly have withered away in the absence of multiemployer agreements. Continuity has meant some, albeit limited, amelioration of the impact of the economic crisis in respect of salary devaluation and job loss. In the event of an economic recovery, signs of which are beginning to emerge, disjuncture and discontinuity in bargaining institutions would make it more difficult for the unions to recover ground lost during the recession. A comparative study of the 'hard discounter' Lidl showed how worker representatives in Spain are able to make use of sectoral agreements to counter management power more effectively (Geppert et al., 2015).
Second, from a trade union perspective, the positive impact of institutional continuity on employment relations is wider than the collective bargaining system. The maintenance of the worker representative system, the consolidation of the autonomous system of third-party intervention in labour conflicts, the representation of workers' interests in the judicial system and union influence in national corporatist institutions are all practical consequences of continuity. Thus, the institutional support for workplace consultation and union recognition results in Spanish subsidiaries of multinational corporations being able to exercise more autonomy in the management of employment relations (Belizón et al., 2014).
Third, the continuity of institutional resources as a platform for union action has significance beyond the industrial sphere. Traxler (2004Traxler ( , 2007 has pointed out that the welfare state is likely to decline if multi-employer bargaining fades away and that organized business and labour are only likely to have significant rights of participation in public policy where an inclusive bargaining system exists. The kind of compensatory social policies Thelen (2012) discusses is unlikely to emerge without civil society pressure in most European countries, given the current weakness of social-democratic parties. By some degree, trade unions are the most important force in civil society. Their legitimacy to play this wider role would be called into question without their role in collective bargaining and worker representation.
We have shown that the 'right' institutional resources can play a role in embedding industrial relations actors in collectivism, demonstrating the importance of beleaguered trade union movements campaigning for and cherishing such resources. However, we also show that relying on institutional power alone to maintain trade union influence is inadequate. Institutional resources can help to establish a platform for union influence, but to push on from this, other power resources need developing. The nature of these other power resources will be related to the context of industrial relations. Thus, in Spain, the structure of the economy is always likely to make it difficult for unions to develop a high level of associational and organizational power. It is, therefore, important for them to focus on the development of moral and collaborative power resources (Gumbrell-McCormick and Hyman, 2013), building on a wider campaigning role, in which the Spanish unions have traditionally engaged, not only to renew and revise the existing institutional configuration but also to 're-embed the economic system in society with new mechanisms of regulation, social protection and redistribution' (Standing, 2014: 132-133).