To Extend or Not to Extend: Explaining the Divergent Use of Statutory Bargaining Extensions in the Netherlands and Germany

Employee coverage by multi-employer bargaining declined since the 1980s in many countries, but countries differ in the extent of that decline. These differences are due, in part, to statutory coverage extension. We analyse the use of statutory coverage extension in two countries, Germany and the Netherlands. Agreements are extended frequently in the Netherlands, where coverage remained stable as a result, but sparingly in Germany, where coverage eroded. The article shows that different employer attitudes are the main cause of this difference. These differences in employer attitudes result from (a) different perceptions of the effects of wage competition by non-organized ﬁrms on organized ﬁrms and (b) differences in employer views on the appropriateness of state compulsion.


Introduction
In recent decades, collective bargaining has been on the retreat in many OECD countries.One indicator of this trend is the collective bargaining coverage rate, that is, the share of employees covered by multi-employer collective agreements.While coverage has declined substantially in some countries, in other countries it has remained stable or even increased (cf.Baccaro and Howell 2017: table 3.3).Among continental European countries, Germany has experienced the most severe decline in collective bargaining coverage, from 85 per cent in 1990 to 56 per cent in 2016.In the Netherlands, it increased from 71 per cent to 78.6 per cent during the same period (ILO 2019).Forces common to advanced industrial countries drive the erosion of collective bargaining, in particular the intensification of cost competition in a globalized economy, the rise of the service economy and union decline (cf.Baccaro and Howell 2017;Hassel 1999).However, these common forces did not invariably result in an erosion of collective bargaining coverage, because in some countries state regulations protect bargaining institutions.One major form of state support is the extension of multi-employer collective agreements to non-organized firms (outsiders) by government decree, known as statutory coverage extension.As a result of coverage extension, bargaining coverage can remain high even when employer or union strength declines.
Although extensions often affect only a small share of employees directly, they stabilize collective bargaining by eliminating incentives for firms to leave their associations.Several studies have highlighted the importance of statutory coverage extension for maintaining high levels of collective bargaining coverage (Traxler et al. 2001:194;Schulten 2012;Schulten et al. 2015;Visser 2018).Yet, we do not know why some countries use extensions more frequently than others.
In this article, we aim to answer this question by comparing two countries with very different practices of statutory coverage extension, the Netherlands and Germany.The legal system in both countries allows for the government to extend multi-employer agreements to non-organized firms; the legal requirements for extensions are similar in both countries.However, despite these similar rules, 47 per cent of agreements were extended in the Netherlands in 2017, as compared to only 1.6 per cent in Germany (Visser 2018: table 2.1 for the Netherlands, and; Schulten 2018a: figure 3.7 for Germany).Moreover, the frequency has diverged since the 1990s, with the number of extensions declining in Germany, but not in the Netherlands.As Kathleen Thelen has noted: '[t]he erosion of collective bargaining in Germany is in large measure a function of state inaction and incapacity' (Thelen 2014: 205).Our analysis of the causes of different extension practices in the two countries will contribute to a better understanding of the conditions for state support for wage bargaining.
We argue that the main explanatory factor for the different frequency of extensions in these two countries are historically evolved differences in how employers' associations assess this institution.We show that greater employer support, not differences in union strength or in government willingness to extend, is the main reason why agreements are extended more frequently in the Netherlands.We identify differences in the way organized employers perceive (a) wage competition by outsiders and (b) state compulsion as the two main factors that explain their different assessment of extensions.We argue that these different attitudes cannot be reduced to differences in sectoral or national-level economic structure.Rather, differences in interest construction constitute a causal factor in their own right.
After introducing our argument in the following section, we present our case studies of statutory extension in the Netherlands and Germany.We look into more detail at the German case because conflicts over the use of C 2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd. statutory extension were more intense there.These conflicts played out in part at the sector level, and we thus include studies of two key sectors, retail and construction, for Germany.Developments in the Netherlands were less eventful, since labour market organizations are in agreement about the use of extensions.

Theory: explaining the use of statutory extension
Why do some countries extend sector-level collective agreements more frequently than others?Not all countries have the legal prerequisites to extend agreements to non-organized firms, but even among those countries that do, like Germany and the Netherlands, the frequency of extensions differs considerably.Why is this?
Our explanation builds on the importance of differences in employer attitudes towards (a) wage competition between organized and non-organized firms and (b) state compulsion in industrial relations.We argue that these employer attitudes explain the differences in the use of extensions better than union strength, government positions or the rise of low-productivity services.We show this in two ways: First, we show that employer attitudes are the only variable that co-varies with the outcome, the frequency of extensions, in the expected way.Other variables are either constant or do not co-vary in the expected way.Second, we show through within-case evidence that in Germany employer resistance was the main cause of failure where attempted extensions did not succeed.Going beyond documenting that employer attitudes are the key causal factor, we also develop a theoretical argument for why employer attitudes differ.
Our explanation of the frequency of extensions thus has two levels: First, we demonstrate that employer attitudes are the main reason why the Netherlands extends agreements more frequently than Germany.Second, we analyse why employer attitudes differ between the two countries, drawing on publications by the associations and 18 semi-structured interviews with employer representatives (nine in Germany, nine in the Netherlands).In the remainder of this section, we outline our explanation for these differences in employer attitudes.
Firms engage in multi-employer bargaining to secure industrial peace, reduce transaction costs and keep labour costs predictable and stable.The price they pay consists in potentially higher and more rigid labour costs than those of non-organized firms.The main advantage of extension for organized firms is the prevention of wage under-cutting by non-organized firms.Statutory extension creates a level playing field; organized firms can be sure not to be outcompeted on labour costs; a strategy of labour market governance that Peter Swenson called cartelism (Swenson 2002: 21).The main disadvantage of extensions for organized firms is the removal of wage competition by outsiders as a tool to discipline unions.Extensions may thus drive up wages, as some economists argue (see, for instance, Lesch 2005 However, research in economics on the effects of extensions on wages is scarce and inconclusive (see Visser 2018: 53 for a review of studies).Research in industrial relations on the preferences of employers regarding coverage extension is scarce as well.A recent study by Bulfone and Afonso of employer attitudes towards collective bargaining in Spain, Portugal and Italy found that organized small firms there favour extensions as a way to prevent wage undercutting (Bulfone and Afonso 2019: 8).A recent study by Arnholtz on coverage extension in construction in the Nordic countries finds that employers in Norway support extensions, while employers in Denmark and Sweden oppose them (Arnholtz 2019).Our findings show heterogeneity in employer attitudes as well, though we offer a new interpretation and causal explanation of this heterogeneity.
Studies of employer politics often rely on differences in the characteristics of firms and sectors to explain any observed heterogeneity in the policy preferences and attitudes of employers.Some accounts focus, for instance, on conflicts of interests between trade-exposed and sheltered firms (see Hassel 2014;Pontusson and Swenson 1996 as examples), others on conflicts between small and large firms (see Mares 2003 as an example) or between firms with different skill needs (Estévez-Abe et al. 2001).Since these accounts build on the different characteristics of firms and sectors, they lead us to expect similar patterns of intra-business conflict in all countries.We argue that such accounts do not offer a satisfactory explanation of employer attitudes towards coverage extension, because we find different patterns of within-country preference heterogeneity in the two countries we study: in the Netherlands, employers in all sectors back extensions; in Germany, support varies across sectors, with employers in most sectors being critical towards extensions.
The alternative account we propose builds on historically evolved, countryspecific interest constructions that result in organized employers assessing the appropriateness of extensions to protect collective bargaining differently, even when confronted with comparable challenges, such as wage competition by posted workers.The ambiguity of the pros and cons of extensions for organized firms means that organized firms need to make use of normative and causal ideas to make sense of their interests.We argue that normative and causal understandings of two issues are key to how employers resolve this ambiguity: first, ideas about the desirability of wage competition by outsiders; and, second, ideas about the appropriateness in principle of state compulsion in industrial relations.These ideas are in turn affected by historically evolved experiences with state compulsion and wage competition and have emerged over a period that goes back to the inter-war years.These historically evolved ideas result in organized employers today, confronted with the new challenge of posted workers from low-wage EU countries, assessing the appropriateness of statutory extensions as a response to wage under-cutting differently.
Our argument builds on approaches that identify ideas as a causal factor in shaping employers' policy positions.Idea-centred approaches to politics highlight how normative and causal ideas impact actors' behaviour by shaping their perceptions of what their interests are (see Blyth 2002: 9-10;Hall 2005: C 2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd. [134][135][136].From this perspective, it is thus actors' behaviour that is in need to be explained, rather than what should do the explaining.This perspective is useful also for the study of employer politics, and several studies have proposed idea-centred accounts to analyse how employers construct their policy preferences in situations characterized by uncertainty and interest ambiguity (Culpepper 2008;Ibsen 2016;M ünnich 2011) or to analyse how different ways in which employers are organized affect the way they perceive their interests (Martin and Swank 2012).
Building on the conceptual insights of these studies, our account offers a novel analytic specification of those ideational dimensions that are most relevant in shaping employers' assessment of statutory coverage extension: wage competition and state compulsion.Employers'normative and causal views of wage competition and state compulsion, which are affected by earlier experiences but over time take on a life of their own, shape their stances towards coverage extension today, leading in turn to different institutional practices by either enabling or constraining governments' capacity to extend agreements.At the same time, we show that union strength and government partisanship are less important than employers' views as determinants of institutional practices.

Research design and methods
Our choice of the Netherlands and Germany as cases resembles a most similar systems design, that is, a selection of cases that allows us to keep many relevant factors constant in order to identify those factors that vary between the two cases.The two countries diverge on the outcome -limited and declining use of extensions in Germany, widespread and stable use in the Netherlands.At the same time, many institutional and economic similarities exist between the two countries.Both countries are characterized by sector-coordinated market economies, have similar levels and forms of organization of unions and employers, net union density in both countries is about 17 per cent, the welfare state in both countries follows the Christian Democratic model with earnings-related benefits, both countries are export oriented and both are members of the European Union and thus subject to its rules on the posting of workers from other member states -an issue of relevance to outsider wage competition.We can thus control for and exclude as explanations a number of political, institutional and economic features.
In both countries, the bargaining parties can veto extensions, either formally or informally.In Germany, this veto is enshrined in law; in the Netherlands, it is practiced informally.The established practice in both countries is that the Ministry does not extend against the will of one of the bargaining parties.In Germany, in addition, a committee that consists of representatives appointed by the peak federations of employers and unions, called the 'bargaining committee' (Tarifausschuss), needs to approve extensions.In the Netherlands, employers do not have a formal veto, but in practice, the Ministry does not extend against their will.Where the two countries differ, though, is in the attitude of employers towards extensions, with Dutch employers supporting them and German employers predominantly opposing them.As we are going to show through within-case evidence, these differences in employer attitudes are causally responsible for the differences in how frequently the two countries extend sector-level agreements.
Our within-case analysis builds on semi-structured expert interviews with representatives of employers' associations (nine for each country), data on coverage extension by the national ministries, media reports, survey data, and policy statements by employers' associations, unions, corporatist institutions and political parties.

The Netherlands
As in most other advanced industrial countries, the Netherlands has in recent decades experienced strong union decline, increased international competition and a shift in production from industry to services.In tandem with political decisions that have facilitated the use of non-standard labour, these developments have placed considerable strain on the Dutch collective bargaining system.In recent years, various studies have consequently identified signs of erosion (De Beer and Keune 2018;Ibsen and Keune 2018).In other respects, the Dutch bargaining system has proven to be resilient, however.Unlike in many other advanced industrial societies, the Netherlands has for instance not experienced a shift from industry-level to firm-level bargaining.Collective agreements continue to be of central importance in determining salary and working conditions for covered workers, and the collective bargaining coverage rate has actually increased since the early 1980s (ILO 2019).Moreover, while the Netherlands has experienced a particularly sharp increase in non-standard employment since the 1990s, it has been relatively successful in ensuring that workers on non-standard contracts are given equal treatment with those on regular contracts.In fact, as we shall see, much of the recent increase in the coverage rate of collective bargaining can be attributed to efforts to discourage employers from using non-standard contracts as a tool to lower wage costs and evade collective agreements (Visser 2013).
These and other Dutch government efforts to shore up wage bargaining have received broad support from the employers' associations, which continue to profess their appreciation for sectoral wage bargaining, despite occasionally calling for more flexibility and underscoring that labour unions represent an increasingly smaller proportion of the workforce.Because of its key importance in maintaining high levels of collective bargaining coverage, the employers' associations without exception also continue to support the frequent use of statutory extension of collective agreements.They do so despite sometimes-strong opposition to sectoral-level bargaining by individual firms and growing scepticism towards the principle of statutory extension in the Dutch parliament.The following two sub-sections illustrate that employer support for sectoral-level collective bargaining is broadly shared at both the sectoral and peak levels.In addition, we analyse the reasons for this support.

Statutory Extension in the Netherlands
Similar to their counterparts in many other Western countries, the Dutch employers' associations have not always been major supporters of collective bargaining, let alone of statutory intervention to shore up sectoral-level collective bargaining.The 1937 Act on General Extension and Non-Extension of Collective Bargaining Agreements, which made it possible to extend collective agreements to an entire sector, for instance, came about despite longstanding opposition to its introduction by the peak employers federations at the time (Bruggeman and Camijn 1999).Whereas the confessional peak employers federations eventually accepted the introduction of the Act for partly strategic reasons, the largest peak association at the time, the liberal Association of Dutch Employers (VNW) continued to speak out against this 'unprecedented restraint' on businesses for long after the law's introduction had become unavoidable (Hoefnagels 1966).
Of crucial importance to the VNW's opposition to the 1937 Act was that most of its members did not yet have much experience with collective bargaining at that time.As a result, the association among others feared that a mechanism designed to strengthen collective bargaining would fail to result in a decline in the level of industrial conflict, and only serve to increase labour power and consequently labour costs.In addition, it resented the fact that responsibility for implementation of statutory extension under the Act lay with the Minister of Social Affairs, whom employers viewed as pro-labour (Bruggeman and Camijn 1999).As a result, employer opposition to the Act extended to many employers who could be expected to benefit from its application (Windmuller et al. 1979).
In later years, however, all of the peak employer federations came to voice their appreciation for the Act's role in bringing about industrial stability.There is no evidence that any of them challenged this principle in a major way during the post-war period, despite its rather frequent use: Today, more than 60 per cent of all Dutch workers are covered by agreements that are subject to statutory extension (SER 2007;Visser 2018).While this percentage has not increased in recent decades, it could be argued that the sharp decline in unionization in this period has made sure that the stability of the collective bargaining system is now more dependent on statutory support than during the 1980s.The employers' associations are acutely aware of their central role in the extension process (Interviews NLD1,NLD3 and NLD 4).Their continual support for the principle of statutory extension thus reflects their on-going appreciation of the principle of sectoral collective bargaining.
The importance of this mechanism does not primarily lie in its direct ability to increase collective bargaining coverage, as it extends coverage C 2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd.
directly to only about 7 per cent of all workers (De Beer 2013).Of much greater importance is its indirect effect: The frequent use of the extension mechanism makes it quite difficult for employers to evade collective bargaining commitments by leaving their associations.As a result, the practice of statutory extension also plays an important role in protecting the employers' associations from erosion.Like their counterparts in many other European countries, including Germany, the employers' associations in the Netherlands report that they are gradually finding it more difficult to attract new members because of processes such as societal individualization, sectoral blurring of activities and the increased entry of foreign firms in the Dutch labour market (Interviews NLD 1 to NLD 9).Yet, contrary to in Germany, the problem of active employer 'flight' is much less pressing in the Netherlands.As a result, the organizational density of employers' associations remains relatively high.Estimates for employer density -that is, the share of employees working in organized firms -vary widely, between 60 and 86 per cent, but the numbers are stable (De Beer 2016; Visser 2016).
In addition, the practice of statutory extension makes it more difficult for employers to lower their labour costs by resorting to outsourcing and other forms of non-standard labour.As employers started to make heavier use of these strategies, organized labour has responded by seeking to extend collective bargaining coverage to agency work and to service sectors such as cleaning, catering and security; ensure that employers do not discriminate based on work hours; and extend collective bargaining rights to workers on fixed contracts.These efforts have largely received support from the peak employers federations -often through broad national-level social pacts that were negotiated in close cooperation with subsequent governments (Visser 2013).
The frequent use of extension can in turn partly be attributed to the procedural ease through which it can be applied.The decision to award a request for statutory extension of a collective agreement, for instance, solely lies with the Minister of Social Affairs.The minister usually grants the request automatically when the signatories to the agreement represent a substantial majority, defined as at least 55 per cent of workers employed in the sector.Moreover, where there is opposition to the decision to extend the agreement to the entire sector, the minister merely has to ask the social partners for advice.When the signatories to an agreement represent fewer than 55 per cent of workers in the sector, the minister can still decide to extend the agreement when this is deemed to be in the 'general interest'.In addition, the minister can refuse to extend an agreement if this rejection is in the general interest, even when the parties to the agreement represent a substantial majority.The latter has rarely occurred, however, although centre-right ministers have in the past threatened to use this prerogative to coax the unions into moderating their wage demands (Delsen 2008: 39-43;Meer et al. 2003: 63-65).
As the initiative for statutory extension lies with the 'social partners', it is also of crucial importance that both organized labour and business broadly support its frequent use.Among employers, this support is broadly shared at both the sectoral and peak levels.Contrary to in Germany, where the use of statutory extension, as we will see, has led to various intra-business conflicts in recent decades, most notably in retail, there are no signs that the use of this mechanism has become more controversial at the sectoral level in the Netherlands.Moreover, while all of the peak and sectoral employers' representatives that we interviewed, as noted earlier, acknowledged that they are gradually finding it more difficult to persuade firms to become members, and many felt that that employer's views regarding the employeremployee relationship have become more diverse, none of them viewed these developments as reasons to use statutory extension less frequently (Interviews NLD1 to NLD 9).Instead, several of them explicitly highlighted the importance of statutory extension in maintaining employer membership rates as well as sectoral-level bargaining (Interviews NLD1, NLD3, NLD5, NLD6 and NLD9).
Existing studies on individual employer attitudes towards the use of the mechanism consistently show that the bulk of Dutch firms likewise support the principle of statutory extension, including those employers that operate in sectors that are subject to strong international competition and those that report strong cost pressures (Harteveld 2013;Heijnen and van Rij 2003).The two most commonly cited reasons for their support for the practice of statutory extension by individual employers is that it reduces uncertainty and wage competition (Heijnen and van Rij 2003;Van den Berg and van Rij 2007).Moreover, support for statutory extension is by no means limited to the members of employers' associations: According to a 2007 report on its use, as many as 82 per cent of non-affiliated firms support its use.Only a very small minority of firms believes that there are more disadvantages than advantages to sectoral-level bargaining (Harteveld 2013), and these firms are in no position to disassociate themselves from its consequences through nonmembership or by requesting dispensation.

The Dutch Employer Representatives' Response to the Assault on Statutory Extension
As mentioned earlier, while the initiative for statutory extension of collective agreements has to come from sectoral representatives of business and labour, the decision to award this request lies solely with the Minister of Social Affairs.Contrary to in Germany, the extension of agreements is thus not subject to the approval by a committee made up of peak representatives of labour and business.As a result, the peak employers' federations have no direct influence over individual applications for statutory extension.Situations such as those that have arisen in the German construction sector, where the peak employer association, as we will see, blocked the extension of bargaining agreements despite broad sectoral employer support, can therefore not happen in the Netherlands.One of our interviewees emphasized that even if this would be legally possible, peak associations would be unlikely to engage in such a conflict with sector-level associations (Interview NLD4).After all, they are C 2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd.
in broad agreement with their sectoral affiliates on the merits of sectoral bargaining and the importance of statutory extension in maintaining sectoral bargaining.This support can among others be inferred from a reading of internal documents and from their behaviour.
The largest peak employer federation in the Netherlands, the VNO-NCW, which is often viewed as primarily catering to the interests of large firms, has on many occasions pointed out that statutory extension is of crucial importance for its members because it protects them 'against those who would seek to disassociate themselves from sectoral agreements through nonmembership' (STAR 1989a).Without this mechanism, it has furthermore warned, 'collective agreements will become much less common, because they will be undermined' (SER 2006).In various reports published by the country's two main corporatist bodies, the bipartite Labour Foundation (Stichting van de Arbeid) and tripartite Social Economic Council (Sociaal-Economische Raad), it has in turn underlined its commitment to sectoral-level bargaining, emphasizing that it reduces transaction costs, fosters industrial stability and limits statutory intervention on matters such as schooling, old age pensions and health and safety by enabling the 'social partners' to arrange these matters at the sectoral level instead (SER 2006;SER 2012).The other two peak federations in the Netherlands, MKB Netherlands, which mainly represents small and medium-sized firms, and LTO Netherlands, which represents agricultural interests, have likewise consistently spoken out in favour of the frequent use of the statutory extension mechanism (SER 2006;SER 2012).
These statements have strong political consequences, as the reports of the Labour Foundation and Social-Economic Council, when these reflect a broad industrial consensus, still have a strong influence on parliamentary views (Oude Nijhuis 2018).Moreover, the employer federations wield considerable political influence, in particular over parties on the right side of the political spectrum.This is important, as these parties have grown increasingly sceptical of the frequent use of statutory extension in recent decades.Since the late 1990s, all parties on the right have expressed criticism of the frequent use of the mechanism, proposing that additional criteria be used (including concrete measures to improve labour market participation and worker training), or even called for the outright abolition of the extension mechanism The People's Party for Freedom and Democracy (VVD) and the populist Freedom Party (PVV) have both repeatedly called for its abolition in recent years.The politically more moderate Democrats 66 (D66) and the Christian Democratic Union (CDA) have so far only called for reform and a less frequent use of the mechanism.In recent years, these four parties have frequently held a parliamentary majority.
The fact that parliament's growing scepticism of the statutory extension mechanism has not yet resulted in legislative action can to a large extent be attributed to the peak employer federations' staunch defence of this mechanism.This defence has taken two forms.First, the employer federations have continuously spoken out in favour of statutory extension in direct address to political stakeholders.They have also articulated their views as members of the Labour Foundation and Social Economic Council, which have routinely assessed the statutory extension mechanism's functioning since the late 1980s (STAR 1989a;STAR 2006;STAR 2011;SER 1992b;SER 2006;SER 2012;SER 2013).These assessment reports have among others refuted the claim that the statutory extension mechanism increases labour costs (see also De Ridder and Euwals 2016).As labour's share of national income in the Netherlands has declined by roughly the same amount as in neighbouring countries in recent years, this assessment is not surprising (Van der Velden 2016).
Second, employer representatives have consistently spoken out against making statutory extension dependent on specific criteria.Documentary evidence points to two motivations for this rejection.First, in so far that these criteria served to force the social partners to incorporate specific macroeconomic goals, they feared that such a move would lead to the politicization of collective bargaining, thus undermining its stability.Second, as these additional criteria would make it more difficult to extend agreements, it would also work to decrease levels of coverage, which Dutch employers view to run against their interests (SER 1992a;SZW 2004, and Interview NLD4).
Finally, to prevent such a decrease from occurring, the peak employers' federations also supported a 2007 reform of the Act on statutory extension that made it more difficult for individual employers to obtain dispensation from (certain) provisions of collective agreements that are extended.Prior to the 2007 reform, these dispensations had been granted easily and renewal was often granted on an automatic basis.Up to the early 2000s, these dispensations were not viewed as a major threat to the collective bargaining system.This changed as various firms in the transport sector were accused by the established unions of resorting to yellow unionism tactics to disassociate themselves from sectoral bargaining agreements by setting up their own (less generous) company-level agreements and using these to apply for dispensation from sectoral-level agreements.After the investigative committee provided credence to this accusation, the peak employer federations supported the introduction of dispensation criteria that closed this loophole (STAR 2006).

Non-Standard Contracts and Bargaining Erosion in the Netherlands
The peak employer federations' support for efforts to give equal treatment to workers on non-standard contracts further testifies of their commitment to maintaining high levels of collective bargaining coverage.While the use of non-standard labour has gone further in the Netherlands than in many other European countries, most workers on non-standard contracts have been given equal treatment in terms of wages and benefits in order to discourage employers from opting for non-standard contracts in order to lower labour costs.A good example of this are part-time contracts, which are more popular in the Netherlands than anywhere in the world.One of the reasons for this is that employers may not discriminate based on work hours, which means that part-time workers are entitled to the same hourly wages, wage supplements, C 2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd.
bonuses, holidays and so on.This has been enshrined by law since 1996, and has been part of most collective bargaining agreements long before that.
In addition to granting equal treatment to part-time workers, the social partners and subsequent governments have also taken measures to make sure that workers on other types of non-standard contracts as well as agency workers are entitled to similar treatment on wages and non-wage benefits.By doing so the social partners have made it much more difficult for employers to opt for non-standard contracts to lower labour costs.While the labour unions were undoubtedly the driving force behind equal treatment for part-time labour and other forms of non-standard contracts, the peak employer federations have generally not opposed these demands.For parttime labour, part of the reason for this seems to have been that employers were genuinely concerned that without equal treatment, this form of labour would remain limited to certain low-paid sectors (STAR 1989b;STAR 1993;SER 1993).
In sum, employer support for the use of statutory extension is broad in the Netherlands and is based on an appreciation of its importance for the maintenance of multi-employer bargaining.Moreover, support extends to all sectors and even most non-organized firms.This makes the Netherlands very different from Germany, where most employers are critical of statutory extensions.

Germany
Collective bargaining coverage in Germany declined from 85 per cent of employees in 1995 to 57.6 per cent in 2016 (ILO 2019).Rather than viewing statutory extension as a solution to stop that erosion, most German employers are critical of its use.In the 1990s, the BDA began to use its veto power more frequently to block extensions.In addition, it opposed legislative initiatives designed to make extensions easier.At the sectoral level, several associations also came out in opposition to statutory extension, including in sectors such as retail where they previously had supported it (Behrens 2011: 175-186).Internal divisions exist and some sectors, in particular construction, continue to endorse widespread extensions.
Why did German employers' associations not pursue the same strategy as their Dutch counterparts and use extensions to stop bargaining erosion?In this section, we are going to show how German employers constructed their interests differently, despite the institutional setting being similar to the Netherlands.We show how a more favourable assessment of wage competition and a stronger aversion to state compulsion led German employers' associations to oppose the frequent use of statutory extension.We begin with a brief overview of the operation of statutory coverage extension in Germany, whereupon we discuss employer' attitudes in two sectors, retail and construction, as well as reforms to regulate posted workers and to modernize extension.

Statutory Extension in Germany
The legal tool of statutory coverage extension was introduced in 1918 and reintroduced after the end of the Nazi regime with the collective bargaining law (Tarifvertragsgesetz, TVG) of 1949.The most recent reform of the regulations occurred in 2014 as part of a law intended to strengthen collective bargaining (Tarifautonomiestärkungsgesetz).In order to extend a collective agreement, one of the bargaining parties needs to apply to the state-level Ministry of Labour, and the extension needs to be in the public interest.Like in the Netherlands, the Ministry decides, but cannot take the initiative.As we mentioned earlier, an extension needs to be approved by a committee, known as the 'bargaining committee', which is composed of three members appointed by the peak employer federation (BDA or its state-level unit) and three members appointed by the peak union federation (German Trade Union Confederation, DGB).A majority must support the extension, which means that employers can veto an extension.Until 2014, there was a coverage quorum of 50 per cent, which means that at least 50 per cent of employees in the area covered by an agreement needed to be covered.To make extensions easier, the national parliament abolished this requirement in 2014.However, the BDA kept its veto power.
Over the past two decades, the share of extended agreements in Germany has declined from 5.9 per cent in 1991 to 1.6 per cent in 2017 (Schulten 2018b: figure 3.7).Extensions were traditionally used widely only in a few sectors, in particular in construction, retail and wholesale commerce, cleaning and cosmetics.Agreements are rarely extended in manufacturing, a large segment within the BDA.
In sum, statutory extension has been used more sparingly in Germany than in the Netherlands for the entire post-war period, and its already limited use declined from the 1990s on.The following two subsections illustrate the reasons for and consequences of this decrease by looking at two sectors where the use of statutory extension used to be widespread, namely retail and construction.In retail, the collapse of statutory extension resulted from internal conflicts within the sector.In construction, statutory extension underwent a crisis due to opposition by the BDA, but did not collapse.As the use of statutory extension remains uncontroversial in the Dutch construction and retail sectors, as we emphasized in the previous section, developments in Germany and the Netherlands thus also differ on the sectoral level.

Statutory Extension in Retail
One of the sectors that turned away from extensions is retail.The number of extended agreements there declined from 63 in 1997 to 30 in 2001, and to 3 in 2013 (Wiedemuth 2013).
The erosion of extensions can be traced to conflicts in that sector that escalated in the mid-1990s.Up to about the early 1990s, a consensus among employers and unions in retail existed to extend most agreements.The employer side, not the unions, typically applied for extension.In the 1990s, C 2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd. some retail chains planned to expand their low-price self-service divisions, a goal they tried to achieve in part by under-cutting bargained wages.Starting in 1995, employers in some regions discontinued applying for extensions, which in effect ended the extension practice (Behrens 2011: 176-186).
Developments in North Rhine-Westphalia (NRW) were most significant, because this is the largest land in terms of population and, thus, retail consumers.Employers there turned against extensions in 2000, a decision that, like in the other Länder, followed a period of internal divisions.Since the late 1940s, two associations had organized the German retail sector, known by their German-language abbreviations BAG (Federal Association of Medium-Sized and Large Firms in Retail) and HDE (German Retail Federation), the latter being the larger one.Until 1999, the regional branches of both organizations in NRW supported statutory extensions, whereby only HDE bargained with unions, though with the backing of BAG.
In the late 1990s, the BAG in NRW changed course and decided to negotiate separately with the unions and to allow its members to opt-out from bargaining, which required the association to terminate its support for extensions.This induced the HDE to introduce its own bargaining opt-out in order to avoid losing members to BAG.Both associations consequently vetoed union applications for statutory extension, in cooperation with the employer representatives in the regional bargaining committee for NRW.According to an interview with an HDE-NRW representative conducted by Martin Behrens, the NRW office of the BDA had earlier put pressure on HDE-NRW to limit the use of extensions (Behrens 2011: 182).Like in construction, the BDA thus pushed against extensions (for construction, see the following sub-section).
The dismantling of extension in retail thus resulted from newly emerged competition between the two associations, which previously had cooperated on wage policy.This change in the associations' approach to extensions was due to the expansion of low-cost, low-wage business strategies among large member firms of both associations.Both associations organized small as well as large firms.Among the largest firms organized in the BAG are the department store chain Karstadt-Quelle, the cosmetics retailer Douglas and the clothing retailers Woolworth and Peek & Cloppenburg.The largest group organized in HDE was the Metro-Kaufhof group, which operated, inter alia, department store chains and food wholesale and retail shops.However, 98 per cent of its over 100,000 member firms had fewer than 50 employees.The BAG's members accounted only for about 10 per cent of the sector's sales.The BAG, while being the smaller of the two, was thus able to bring about the termination of extensions in this sector.The BAG was dissolved in 2009, after Karstadt, its largest member, went bankrupt.Most of its members have since joined HDE.
According to media reports, some larger firms support statutory extension (Metro, Lidl, Aldi Nord, Rewe, Kaiser's Tengelmann and Otto).Other large firms are opposed (Peek & Cloppenburg, Edeka) (Wiedemuth 2013;dpa 2017).Lidl in particular campaigns for a return to statutory coverage extension.Lidl's chairman, Klaus Gehring, has pointed to the rise of Amazon as creating a need to return to statutory extension (dpa 2017).In short, the issue of extensions continues to divide retail.

Statutory Extension in Construction
Construction is another sector with a well-established tradition of statutory extension and where extensions came under attack since the 1990s.The conflict in this case took place not within construction, where employer support for extensions continues to be strong, but between construction employers and the peak federation, the BDA.Since the mid-1990s, the BDA has taken an increasingly sceptical stance towards statutory extension, seeking to restrict its application as much as possible (BDA 2017a: 1-2; BDA 2017e).
Similar to retail, construction is a labour-intensive sector where labour costs are an important determinant of competitiveness.In addition, construction is characterized by small firms and a large number of competitors.Germany has about 70,000 construction firms employing on average fewer than 10 employees (Asshoff 2012: 542).Due to the seasonal fluctuations in workload, many firms rely on temporary contract workers, as a result, labour turnover is high.
To address these sector-specific conditions, unions and employers' associations in construction have developed a range of collectively agreed occupational social policies, including supplementary occupational pension plans and collectively agreed funds for the transfer of holiday rights across firms.These occupational social policies rely on statutory extension to prevent under-cutting by non-organized firms.The construction union (IG BAU) and the two construction employers' associations (the Federation of the German Construction Industry and Federation of the German Construction Trades) all support these extensions.
In the mid-1990s, however, extensions in construction came under fire from the BDA.In 1996/1997, the employer side vetoed an extension of an agreement on minimum wages in construction, officially because of a lack of a public interest.The employer representatives in the bargaining committee came from the BDA, metalworking and textiles (Czommer and Worthmann 2005: 6;Schlachter 2010: 28).Other sectors apparently feared that higher minimum wages in construction might spill over into sectors purchasing construction services.This led to a stand-off between construction employers and employers in other sectors.Construction employers backed higher bargained minimum wages to prevent contractors from EU countries using posted workers with lower wages and temporarily left the BDA as a protest.The Minister of Labour, Norbert Bl üm, mediated a compromise (Asshoff 2012;gk 1997).The crisis of extensions in construction was thus the result of conflicts between construction employers and employers in other sectors.
C 2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd.

Reforms of Statutory Coverage Extension
Several reforms intended to facilitate the use of extensions occurred since the 1990s.These reforms focused on two issues.The first related to the intensification of wage competition caused by the effects of the European Union's rules on the posting of workers, which allows foreign contractors to operate in Germany using workers employed in their home country, often at wages that are below those required by German agreements (Wagner 2015).The second issue was the decline of statutory extensions, due in part to an increasing number of agreements failing to meet the required threshold of 50 per cent.
In this section, we analyse the reform initiatives taken to address these two issues.The BDA opposed these reforms.This stance has been in line with the views of most of its members: One survey of association executives shows that 75 per cent of German employers' associations oppose measures to facilitate extensions (Nicklich 2013: 529).The associations supporting statutory coverage extension are thus in a minority within the employer camp.

The Posted Workers' Act
The Posted Workers Directive adopted by the European Union (EU) in December 1996 (Directive 96/71/EC) in response to a ruling by the European Court of Justice (ECJ) allows companies to temporarily post workers to other countries and regulates the rights and working conditions of these workers.These rules led to debates in Germany about their effects on wage competition and on the introduction of a new instrument for the extension of bargained minimum wages, with requirements more relaxed than those that apply under the 1949 law (Menz 2003: 544-546).
In March 1996, the national parliament passed a law regulating the working conditions of posted workers, known as the Posted Workers Act (AEntG).The law was partly a result of lobbying by construction employers, while the BDA opposed it.The Posted Workers Act allows the statutory extension to posted workers of collectively agreed, sector-level minimum wages.Employers and unions can agree on sector-level minimum wages and request that these are extended to all firms, including those using posted workers.The provisions for coverage extension under the Posted Workers Act are more restrictive than those in the Collective Bargaining Law of 1949.First, only minimum wages can be extended, not the entire wage scale.Over time, however, the list of regulations that can be extended was enlarged, including, inter alia, occupational health and safety, holiday pay and maximum working time.Second, extension was initially possible only in construction and constructionrelated sectors.Over time, other sectors were added, such as cleaning, security guards, and care.In 2014, finally, the extension provisions of the Posted Workers Act were made applicable to all sectors.The coverage threshold of 50 per cent and the veto by the BDA applied initially to statutory extension under the Posted Workers Act, but were later abolished.
The BDA strongly opposed the extension of bargained minimum wages, officially because it saw this as 'an attack on the autonomy of wage bargaining'.The executive director of the BDA, Reinhard G öhner, argued in an interview in November 1998 that statutory extension of collectively agreed sector-level minimum wages would drive up average wages (pt 1998).As we have shown earlier, Dutch employers' associations disagree with this view, and do not believe that extensions drive up wages.
Today, extensions under the Posted Workers Act of 1997 play a much larger role than extensions under the Collective Bargaining Law of 1949.In 2012, about 245,000 employees were covered by wage agreements extended under the Collective Bargaining Law, comprising 1.3 per cent of all employees covered by collective agreements.About 3.7 million workers were covered by sector-level minimum wages extended under the Posted Workers Act, comprising about one-fifth of all employees covered by collective agreements (Bispinck 2012: 503).

The Reform of the Collective Bargaining Law in 2014
The erosion of bargaining coverage and the decline of extensions under the Collective Bargaining Law has led to new legislative initiatives to facilitate extensions.These initiatives came from the union federation (DGB), the Social Democrats (SPD), the Greens and the Left Party.The main demand of these actors was the abolition or lowering of the coverage threshold.The BDA opposed these initiatives and was supported by the Free Democrats (FDP).As long as the FDP was part of the federal government (28 October 2009 to 17 December 2013, cabinet Merkel II), no reform was thus feasible.This changed in December 2013 with the formation of a Grand Coalition of CDU/CSU and SPD.
On 3 July 2014, parliament passed the 'Law for the Strengthening of Collective Bargaining' (BGBl I, p.1348), which was supported by the two governing parties, the Christian Democrats and Social Democrats, as well as by the Greens.The Free Democrats voted against, while the Left Party abstained.The law included several measures, including the adoption of a statutory minimum wage (Marx and Starke 2017).The law also made extensions easier.Agreements can now be extended even if fewer than 50 per cent of employees are covered, as long as the extension is considered to be in the public interest and both bargaining parties support the extension.The abolition of the threshold was intended by political parties to facilitate the use of statutory extension.Moreover, collectively agreed minimum wages can now be extended in all sectors.The law also removed the bargaining committee's veto against the extension of collectively agreed minimum wages.The BDA's veto power remains for extensions under the Collective Bargaining Law.
The BDA opposed the reforms.In the opinion it submitted to the federal parliament, the BDA argued that statutory extension violates the negative freedom of association and the autonomy of bargaining, and should therefore be used only in exceptional cases (BDA 2017b BDA 2017a; BDA 2017d; BDA 2010: 1-2).The BDA also explicitly stated that statutory extension should not be used to limit wage competition or to establish a wage floor (BDA 2014: 2-3).One of the associations that spoke out in favour of facilitating the use of statutory coverage extension was the construction federation ZDB (ZDB 2013: 2-3).

Motives for German Employers' Opposition to Statutory Coverage Extension
Our interviews with representatives of German employers' associations and unions point to a principled opposition to state compulsion and to support for wage competition as reasons for employers' sceptical stance towards statutory extension.The BDA representative we interviewed said that he rejected 'the option to use collective agreements to rule out competition' (Interview DE8).Similarly, a study by the Institute of the Germany Economy (IW), a think tank funded by German employers' associations, argues that: 'It is competition from outsiders that ensures the appropriateness of a collective agreement, making the positive and negative freedoms of association equally necessary' (Lesch et al. 2017: 4).In short, the BDA representatives and publications by employer-sponsored organizations present statutory extension as a threat to wage competition, views that contrast with the Dutch employers' more sceptical stance towards wage competition by outsiders.
Metalworking employers expressed a similar perspective in our interviews.A representative of metalworking in NRW, Metall NRW, said that 'statutory extension is not loved by employers' (Interview DE2).A representative of metalworking in Baden W ürttemberg, S üdwestmetall, called statutory extension 'an extremely double-edged sword'.He compared it to 'treating oneself with cortison', arguing that it has positive consequences in the short term, but negative consequences in the long term, the negative consequence being the undermining of voluntarist character of the German model (Interview DE7).Furthermore, the representative of Metall NRW said: We are not just interested in having as many members as possible . . .we are also . . .endowed with a . . .political orientation -. . .which is against state interference, and statutory extension is such an interference.(Interview DE2) Only the representatives of construction articulated support for statutory extension.According to the representative of the ZDB, 'the advantages [of statutory extension as practiced in construction] by far outweigh the disadvantages' (Interview DE1).The HDB representative argued that statutory extension of minimum wages is necessary to prevent wage undercutting by East European firms using posted workers (Interview DE3).Both representatives emphasized that statutory extension was essential for the functioning of occupational social funds, since workers in this sector switch between employers frequently.The representative of the ZDB referred to extensions as 'the corset rods holding the German labour market in construction together' (Interview DE1).However, both construction representatives were sceptical about extending statutory extension beyond its current use, for instance by extending the entire wage scale.
In short, we found through the interviews and our analysis of documents published by associations that German employers use the benefits of wage competition and a principled opposition to state compulsion to justify their scepticism of statutory extension, not so much economic advantages for their members.Employers in construction deviate from this line and support widespread extensions in their sector.

Comparative discussion
We found that differences in employer attitudes are the main reason why multi-employer agreements are extended so much more frequently in the Netherlands than in Germany: German employers' veto applications for extension much more frequently than Dutch associations.Union strength is about the same in both countries, at a net density of 17 per cent, and does therefore not explain the variation.Political parties in both countries take different stances on extensions, but reforms in Germany to facilitate extensions turned out as ineffective, because of the employer veto power.The reason for the limited use of extensions in Germany is thus not lack of government support or weaker unions, but greater employer resistance.
Sectoral differences in employer attitudes exist in Germany, but not in the Netherlands.In Germany, employers in constructions support extensions, and employers in retail used to support extensions up to the 1990s, while attitudes within that sector have later become more heterogeneous.Opposition to extensions by the German peak federation (BDA) also intensified since the 1990s, while in the Netherlands support by the peak federation remains stable.
The further erosion of German employer support for extensions since the 1990s is in part the result of a shift to low-wage business models and intensified cost competition.More importantly, though, an ideational path dependency, characterized by long-standing opposition to state intervention and an embrace of outsider wage competition, locked German employers' associations into a voluntarist model of bargaining in a way that Dutch associations' different causal and normative views of wage competition and state intervention did not.
We identified two ideational factors along which the attitudes by employers' associations on extensions differ: (a) their views on the effects of wage competition by outsiders on organized firms and (b) their views on the desirability of state intervention.First, associations in the two countries differ in how they perceive wage competition: Dutch associations motivate their support for statutory extension among others by referring to the need for a level playing field for organized and non-organized firms, their German counterparts motivate their opposition by referring to the need for wage competition by outsiders in order to ensure that agreements with unions remain reasonable.In other words, they assess the effects and desirability of wage competition differently.Second, associations in the two countries view the appropriateness of state intervention in industrial relations differently.German associations view any form of state intervention as problematic.Dutch associations by contrast only view state involvement as problematic when the state seeks to interfere directly with the content of collective bargaining.

Conclusions
In this article, we have made two contributions: First, we contributed to our understanding of the causes of the stability and erosion collective bargaining institutions by documenting that differences in employer attitudes towards coverage extensions are one major determinant of stability.Second, we have contributed to the study of employer politics by putting forward an explanation for why employer attitudes towards extensions differ.While heterogeneity of preferences is a common finding in the study of employer politics, studies differ, first, in terms of their descriptive characterization of that heterogeneity, and second, regarding the causes they attribute that heterogeneity to.Our ideas-centred account differs from accounts that emphasize country-level or sector-level differences in firm characteristics, such as firm size, skill profiles or mode of coordination, as determinants of employer attitudes.These accounts are unsatisfactory, because they cannot explain why employers in the same sector in two countries take different stances.Neither can differences in union strength explain the differences in employer views between the two countries, since net union density is the same in both countries.Different views of the appropriateness of state intervention and of the effects of wage competition are better able to explain this.These views are rooted in different historical experiences but have over time taken on a life of their own and have turned into independent causal factors.
Clearly, our finding of employer support for coverage extension in the Netherlands, and in construction and retail in Germany, should not be mistaken as evidence for employer opposition to liberalization of collective bargaining, since liberalization does happen also within collective bargaining (cf. for instance, Benassi 2016;Benassi and Dorigatti 2015;Paster 2012: 172-177;Streeck 2009: 38-5;and Thelen 2009: 481-482).Yet, it does suggest that employers' associations may hold liberalizing and de-liberalizing preferences at the same time, depending on what aspect of the bargaining system we look at.
In light of our findings, we see case studies investigating employer attitudes toward statutory extensions in other countries as a promising avenue.As we pointed out in the introduction, we kept a range of features constant through our choice of cases, such as union strength, corporatist traditions, welfare state regimes and EU membership.Clearly, introducing variation on these variables would help to test the robustness of our findings for a larger sample of cases.Moreover, to understand the intensification of employer opposition to extensions in Germany from the 1990s on, future research will have to look at additional sectors.

Thomas
Paster is at the University of Roskilde.Dennie Oude Nijhuis and Maximilian Kiecker are at University of Leiden.C 2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

C
2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd.

C
2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd.

C
2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd.

C
2019 The Authors.British Journal of Industrial Relations published by John Wiley & Sons Ltd.
Final version accepted on 18 November 2019 7. Interview NLD7, Director Social Affairs, Detailhandel Nederland, 10 January 2019 8. Interview NLD8, Policy Advisor Detailhandel Nederland, 7 February 2019 9. Interview NLD9, Policy Advisor Bouwend Nederland, 25 February 2019 Funding Thomas Paster acknowledges that the project leading to this article has received funding from the European Union's Horizon 2020 research programme under the Marie Sklodowska Curie grant agreement No. 705677.