Introduction
What role does the state play in maintaining social inclusion in coordinated models of capitalism? In these economies, cooperation between the state and social partners through corporatist institutions used to function as a bulwark against the inequality witnessed in liberal market economies by enabling members of the working class to access high-quality training and well-paid, stable employment (Streeck, Reference Streeck1992). Yet can coordinated models maintain their distinctive social pedigree in the face of common structural pressures (Baccaro and Howell, Reference Baccaro and Howell2017; Iversen and Soskice, Reference Iversen and Soskice2019)? While earlier interventions emphasized the equality-enhancing effects of employer collective action (most notably Hall and Soskice, Reference Hall and Soskice2001), in recent years, the focus of this literature has shifted to the role of the state (Schmidt, Reference Schmidt2009). Some authors see state intervention as a vehicle for upholding high levels of equality in face of diminishing collective efforts by employers (Thelen, Reference Thelen2014). In this view, high-capacity states will cajole employers into accepting egalitarian policies that enhance social inclusiveness, that is facilitating labour market access for socio-economically disadvantaged individuals. In contrast, recent scholarship on liberalization emphasizes that the pivotal position of employers in capitalist economies enables them to shape the terms of state intervention, with negative effects for social inclusion (Howell, Reference Howell2019).
This paper argues that scholarship on the resilience of coordinated market economies (CMEs) would benefit from more closely examining under which circumstances state intervention promotes social inclusiveness, understood as policy measures aimed at enabling broad access to quality training and employment. Focusing on the coalitional dynamics between employers, unions and political parties, we argue that the degree of social inclusiveness generated by state intervention varies with the kind of coalition partner available to employers – unions or parties. If unions depend on employers for their organizational influence in a policy field, they become strong coalitional partners for employers in fending off demands for inclusiveness originating in the parliamentary arena. However, if unions have influence independent of any coalition with employers, unions and employers are likely to team up with political parties with whom they share preferences, making the level of inclusion more fluctuating since its specific level depends on who currently holds government power.
This argument is set out in the context of collectivist vocational education and training (VET) systems. Characterized by heavy involvement of employers to provide, fund and monitor dual apprenticeship training (Busemeyer and Trampusch, Reference Busemeyer and Trampusch2012), such systems have long been considered the poster child of how employer coordination can produce egalitarian outcomes and increase labour market inclusiveness (Streeck, Reference Streeck1992; Hall and Soskice, Reference Hall and Soskice2001; Thelen, Reference Thelen2014). However, the employer coordination that long shaped and upheld collective skill formation has come under pressure from globalization, deindustrialization processes and the rise of the knowledge economy (Martin and Knudsen, Reference Martin and Knudsen2010; Unterweger, Reference Unterweger2020; Busemeyer et al., Reference Busemeyer, Carstensen and Emmenegger2022). Declining levels of employer coordination have opened space for the state to take on a more proactive role as training provider. Consequently, we focus on state-led apprenticeships that partially replace the employers’ role in providing apprenticeships. State-provided training has important implications for social inclusion, harbouring the potential to facilitate access to high-quality training for disadvantaged students (Bonoli and Wilson, Reference Bonoli and Wilson2019; Durazzi and Geyer, Reference Durazzi and Geyer2021), which might, however, clash with economic efficiency goals favoured by employers (Carstensen and Ibsen, Reference Carstensen and Ibsen2021; Bonoli and Emmenegger, Reference Bonoli and Emmenegger2021). The policy field consequently provides fertile ground for exploring under which conditions – in the context of declining employer coordination – employers and unions can shape the terms of state intervention and what consequences these processes have for maintaining equality in CMEs.
We do so by examining reform trajectories in Danish and Austrian VET. The study employs a most similar systems design, comparing two countries characterized by long corporatist traditions of social partnership and a historically strong affinity between unions and Social Democrats. Instead of focusing on the well-researched collective skill formation systems of Germany and Switzerland, we compare two similar cases with high-capacity states that could be expected to push employers into line and offer inclusiveness-enhancing policies (Martin and Thelen, Reference Martin and Thelen2007). Consequently, these two countries are also ‘least likely’ cases: If social partners in these two cases, individually or together, can set the terms of state intervention, we are likely to see even stronger social partner influence in other collectivist VET systems like Switzerland or Germany, where state capacity is lower.
Despite these similarities between the two cases, we find important variation regarding the extent to which state intervention increases social inclusiveness. In Austria, a conservative welfare state, employer discretion is restricted by shifting ideological commitments of left- and right-leaning governments. In Denmark, a social democratic welfare state, employer discretion is more pronounced, as VET institutions are not politicized and instead ruled through a consensus among the social partners. Based on data from policy documents, media coverage and interviews with key policy actors in the two countries, we demonstrate how the degree of employer discretion is a function of the coalitional partners available to employers. In Denmark, the unions’ role in VET governance is almost solely dependent on employer involvement. Consequently, unions are more likely to join a coalition with employers. This softens opposition from centre-left parties and weakens the focus on inclusion. In Austria, the unions’ main role in the governance of skill formation systems is not dependent on employer involvement. Consequently, employers and unions advance their interests by seeking coalition partners in the parliamentary arena. The emerging left-right coalitional pattern makes the extent of employer discretion more dependent on who holds government power.
Employers, unions and the state in coordinated capitalism
How are coordinated models of capitalism responding to structural pressures variably associated with globalization and the rise of the knowledge economy? More specifically, are they able to remain inclusive in face of disembedding forces? Positions range from Baccaro and Howell (Reference Baccaro and Howell2017) who observe a common neoliberal trajectory that is universally undermining what is left of coordinated capitalism, to Iversen and Soskice (Reference Iversen and Soskice2019) who highlight its continued resilience to structural pressures. While these two positions long enjoyed equal standing in the literature, the argument that structural pressures do not threaten coordinated capitalism has lost ground (Hall and Thelen, Reference Hall and Thelen2009; Emmenegger et al., Reference Emmenegger, Häusermann, Palier and Seeleib-Kaiser2012).
In recent years, attention has increasingly shifted to the role of the state in these developments (Schmidt, Reference Schmidt2009; Marx and Starke, Reference Marx and Starke2017; Rathgeb, Reference Rathgeb2018; Hassel and Palier, Reference Hassel and Palier2021). In an influential contribution, Thelen (Reference Thelen2014) observes that a straightforward neoliberal shift is all but impossible to build a coalition around in CMEs. Instead, she points to the mediating effect of state capacity on structural pressures, producing different reform trajectories. Using skill formation policies as a key example, she identifies two main trajectories. First, dualization, where a privileged manufacturing core still receives high-quality training and enjoys a relatively high level of income and rights, over and beyond what workers in other parts of the economy receive; and second, socially embedded flexibilization, where solidaristic outcomes are maintained through increased state intervention and public policies.
These trajectories are based on a perceived trade-off between upholding employer collective action and maintaining solidaristic outcomes.Footnote 1 That is, either strategic employer coordination is upheld for a privileged core leading to dualization, or declining employer coordination is compensated with state-promoted social programmes designed to ease the adaptation of weaker segments of society to changes in the labour market (e.g., state-provided vocational training). In this view, the choice between the two routes in large part hinges on the state’s capacity to follow an embedding strategy and cajole private sector actors into agreement. Here, state capacity resides in a large and well-organized public sector, which is well placed to shore up macro-corporatist institutions to pursue what Martin and Thelen (Reference Martin and Thelen2007) consider a key interest among state actors, namely high levels of social inclusiveness.
We are broadly sympathetic with the argument that decreasing employer coordination may be associated with more state intervention. However, it is no necessity for strong states to push for more inclusion. Other researchers argue that the state has been instrumental for the universal liberalizing tendency in the last decades (Howell, Reference Howell2016; Rathgeb, Reference Rathgeb2018). This argument is based on an understanding of state action not as autonomous but rather as contingent and strategic relational, ‘heavily influenced by the relative strength, perception of self-interest and organizational form of class actors’ (Howell, Reference Howell2016: 576).
This perspective broadly aligns with distinct but complementary literatures on critical pluralism and power resources theory, which have long established that employers are structurally privileged and thus able to shape the terms of state intervention (Lindblom, Reference Lindblom1977; Korpi, Reference Korpi1983). Due to this privileged position, the level of solidarity resulting from state intervention is dependent on political coalitions involving employers (Rathgeb, Reference Rathgeb2018; Howell, Reference Howell2019). This literature thus takes due notice of the privileged position of business and provides a useful corrective to scholars who consider high-capacity states largely autonomous from business interests. In the following, we employ both approaches to leverage their respective strengths to increase their explanatory potential.
State intervention and employers in vocational education and training
Most of the literature on the reform of coordinated capitalism has so far focused on welfare state reform and labour market policy. Instead, we develop our argument in the context of VET institutions, which have long been recognized as a key sphere of strategic interaction between employers in CMEs (Streeck, Reference Streeck1992; Hall and Soskice, Reference Hall and Soskice2001). With main stakeholders in these systems pursuing not only economic efficiency but also social inclusiveness (Carstensen and Ibsen, Reference Carstensen and Ibsen2021), they provide a useful setting for investigating under which circumstances interactions between employers, unions and political parties lead to inclusive outcomes through state intervention. In the case of CMEs, the literature typically distinguishes between collective skill formation systems, where firms and unions are deeply involved in the provision, monitoring and financing of education, and statist systems where the state has full responsibility for VET.
In recent years, coordinated models of capitalism, and in particular collectivist VET institutions, have been confronted with a process of de-collectivization, as firms increasingly refuse to act collectively in the financing, provision and administration of training (Busemeyer et al., Reference Busemeyer, Carstensen and Emmenegger2022). The likely consequence of de-collectivization is not that the market instead provides training, since creating a market for VET would require deep institutional and economic commitment on part of the state. Rather, the state is likely to intervene in the provision of training (Bonoli and Wilson, Reference Bonoli and Wilson2019; Unterweger, Reference Unterweger2020; Busemeyer et al., Reference Busemeyer, Carstensen and Emmenegger2022; Durazzi and Geyer, Reference Durazzi and Geyer2021; Graf et al., Reference Graf, Strebel and Emmenegger2021). A prominent example are state-led apprenticeships, which are measures to compensate for decreasing training participation of firms in dual VET systems (Durazzi and Geyer, Reference Durazzi and Geyer2020). State-led apprenticeships are an example of de-collectivization as the state takes over the employers´ role in providing apprenticeships. While still counted as dual in many official statistics, state-led apprenticeships are not part of the collective training system anymore because they are not based on employers’ collective action and firm-based training.
The structurally privileged position of employers is well known in collectivist VET where a significant part of training takes place in companies through apprenticeships. Here it is the employer’s possibility to ‘exit’ from offering training contracts that makes them the pivotal actor (Busemeyer, Reference Busemeyer2012). Yet the privileged position of business goes beyond their role as providers of training contracts. As the main providers of employment, employers – and their representatives in employer associations – have also disproportionate discursive power in defining common knowledge (Culpepper, Reference Culpepper2008) on what constitutes necessary skills for future gainful employment and thus the overall adequacy of training systems (Carstensen and Ibsen, Reference Carstensen and Ibsen2021). This power is backed up by legitimacy (as the providers of employment they should know) as well as exit options (as the providers of employment they can refuse to hire). While this power is not absolute, business is in a strong position to get what it wants.
What does this imply for inclusiveness in the context of VET institutions? We may note that the main line of conflict concerns the tension between efficiency – specifically the provision of relevant skills for employers – and inclusion, understood as access to high-quality skill formation for academically weaker students. We should expect employers to be critical of inclusiveness measures imposed by the state, not because they always know what is best for them and therefore always push for ‘efficient’ policies, but rather because they in some cases will perceive the respective interventions to have negative consequences for economic efficiency (Carstensen and Ibsen, Reference Carstensen and Ibsen2021).Footnote 2 Employers and their collective representation consent to an extended role of the state only as long as they do not fear shortages of labour supply on the apprenticeship market due to an increasing take-up rate of statist training measures. For instance, an increased take-up of better qualified apprenticeship seekers in state-provided training could increase labour costs by pushing up wages and working conditions for standard firm-based apprentices. Similarly, increased take-up of weaker candidates in state-provided training could lead to negative reputational effects for VET as a whole and consequently lead to a reduction of better qualified apprenticeship seekers, who will increasingly opt for education outside VET. However, exactly these elements may be associated with more inclusiveness. We expect employers to try to shape such state-sponsored inclusiveness-enhancing measures to align them with their core interests.Footnote 3 Ultimately, this may produce increasingly statist systems that are not more inclusive than the collectivist solutions they replace.
Explaining the varying outcomes of state intervention
Recognizing the potential of state intervention to increase inclusiveness, while also acknowledging that the privileged position of employers may work against the realization of increased social solidarity, opens the way for state intervention to produce varying outcomes in terms of inclusiveness. This in turn raises the question of how we can explain when employers prevail. To this end, we argue that the degree of employer discretion is a function of the coalition partners available to them. Echoing coalitional approaches in comparative political economy (Hall and Thelen, Reference Hall and Thelen2009), if employers face a united front consisting of most of the political parties and the unions, they are unlikely to prevail because too many actors will challenge their discursive and material power for them to successfully define the necessary skills for future gainful employment and thus the training systems’ overall adequacy. What, then, determines the availability of coalition partners, and how do these coalition partners influence the strategies employers pursue to defend their interests?Footnote 4
We believe unions hold the key to answer these questions. Generally, unions can be assumed to favour inclusiveness-enhancing state intervention (Rathgeb, Reference Rathgeb2018; Durazzi and Geyer, Reference Durazzi and Geyer2020; Carstensen and Ibsen, Reference Carstensen and Ibsen2021). However, unions often privilege organizational interests that secure their long-term positions in governance structures over the pursuit of short-term substantive policy gains for their constituencies (Trampusch, Reference Trampusch2010; Davidsson and Emmenegger, Reference Davidsson and Emmenegger2013). Most notably, unions are strongly interested in maintaining a central role in VET governance, not least because it provides an important avenue for recruitment. Yet the role of unions differs across countries. In some countries, the unions’ role is almost entirely a function of employers’ involvement, while in other countries, their role is to some extent independent of employers. This difference has important implications for union strategies. This is not to say that unions will always prioritize organizational interests. However, in periods of dwindling mobilization capacity, such governance roles become increasingly important for unions’ long-term positions of power and have a decisive effect on unions’ strategic considerations (Davidsson and Emmenegger, Reference Davidsson and Emmenegger2013).
This argument aligns with a rich literature in institutionalist scholarship that emphasizes both the stickiness and the power implications of institutions (Mahoney and Thelen, Reference Mahoney, Thelen, Mahoney and Thelen2010). In addition, it builds on recent contributions to the literature on VET systems. Most notably, Durazzi and Geyer (Reference Durazzi and Geyer2020: 110) highlight the role of ‘institutional resources,’ which refer to the unions’ ‘degree of control of training outside of the dual (collective) system comparable to the control they enjoy on the dual (collective) system.’ They argue that Austrian unions – unlike their German peers – supported an inclusive pathway of reform of the training system. They did so because their role in VET governance is not solely dependent on the employers’ involvement in the dual apprenticeship system but is, at least to some extent, independent because of their institutionalized role in the governance of state-sponsored inclusiveness measures and in the Austrian public employment services that finance these measures.Footnote 5
Based on these reflections, we argue that if the unions’ role in VET governance is dependent on employer involvement in the dual apprenticeship system, they are likely to join a coalition with employers, which will combat state intervention if the state crosses the employers’ red lines. In this case, unions lend their support to the employers’ cause, since this places them in the best position to defend their organizational interests of remaining central to VET governance in face of increasing state encroachment on social partner autonomy. The main dividing line structuring the conflict on state intervention and its degree of inclusiveness is between the social partners and the parliamentary arena.
In contrast, if the unions’ main role in the governance of skill formation systems is not dependent on employer involvement in the dual apprenticeship system, employers and unions are likely to advance their interests by instead seeking coalition partners in the parliamentary arena. The reason is that unions’ independent, institutionalized role in state-sponsored training measures makes them less dependent on employers as coalition partners for gaining the benefits of being involved in the skill formation system. Instead, employers will find political partners among the business-friendly parties, while unions will find partners among the more left-leaning parties, which share their preference for social inclusion (Busemeyer, Reference Busemeyer2015). In this case, the main political cleavage is between left-leaning and right-leaning political blocs, with those interests prevailing that currently hold government office, potentially limiting employer influence.
Research design
Below we investigate the state’s ability to maintain egalitarian capitalism through a comparative study of institutional change and reform processes in VET. The dependent variable is the degree to which state intervention increases social inclusion, understood as access to high-quality skill formation for socio-economically disadvantaged and academically weaker students. We focus on state-led apprenticeships, which are measures to compensate for decreasing employer coordination in dual VET systems by providing alternative forms of training to unsuccessful apprenticeship seekers (Durazzi and Geyer, Reference Durazzi and Geyer2021). We examine their introduction in the mid−1990s and subsequent reforms until the late 2010s.
We employ a most similar systems design by looking at reform trajectories in Austria and Denmark. Both countries are characterized by long corporatist traditions of social partnership, with a historically strong affinity between unions and Social Democrats. Furthermore, both have collective VET systems and feature high levels of state capacity and degrees of employer organization, which should facilitate inclusiveness-enhancing state intervention (Martin and Thelen, Reference Martin and Thelen2007). Consequently, the cases also have ‘least likely’ qualities: If social partners in these two cases can set the terms of state intervention, we should observe even stronger social partner influence in other collectivist VET systems like Switzerland or Germany, where state capacity is lower.
Despite these similarities, we observe counterintuitive variation between the two cases in the dynamics of change they exhibit. If anything, we should expect higher levels of inclusiveness in Denmark with its social democratic welfare model and its strong public sector and service sector unions, while Austria with its conservative welfare state, lower levels of union density and stronger reliance on manufacturing might be more prone for low levels of inclusiveness (Thelen, Reference Thelen2014). While we find that state intervention in both Denmark and Austria considered employers’ preferences, Austrian employers prevailed mostly in times of right-wing governments (see section section 3). In contrast, Danish VET is less politicized and primarily governed by the social partners, making employer discretion more pronounced regardless of who is currently in government. Compared to Denmark, employer discretion in Austria is more restricted by shifting ideological commitments of left- and right-leaning governments. We argue that this variation in the level of social inclusion resulting from state intervention depends on employers’ coalition partners (see section 4).
The analysis is based on policy documents, press releases, media coverage and existing research. Thirteen background interviews with state agencies, parties, social partners, independent experts and vocational schools active in the respective reform processes (see online appendix for the list of interviewees) provided additional opportunities for data triangulation and first-hand insights into the respective reform processes.
State intervention to lighten the burden of apprenticeship provision
In the 1990s, both Denmark and Austria started compensating insufficient training participation of employers with more state intervention. New policies allowed students unable to secure apprenticeship contracts to complete their vocational education through state-led apprenticeships that replace the employers’ role in dual VET by providing practical training. In Denmark, these ‘school-based apprenticeships’ (skolepraktik) were accepted as a temporary measure to reach the political target that 95% of a youth cohort will complete upper secondary education. However, in 2013, they ended up as a permanent feature of the new training centres (praktikcentre) (Retsinformation, 2012; Jørgensen, Reference Jørgensen, Michelsen and Stenström2018). Similarly, Austria responded to decreasing employer participation in training by introducing public training workshops in the 1998 JASG law (Jugendausbildungs-Sicherungsgesetz), which were relabelled ÜBA (Überbetriebliche Berufssausbildung) in 2008.Footnote 6 In both countries, take-up of students in state-led apprenticeships hovers between 5–10% of all apprentices (Dornmayr and Nowak, Reference Dornmayr and Nowak2018; Ministry of Education, 2019).Footnote 7 As we show in the following, despite this expanding role of the state, employers’ interests in both countries were considered as they retained their privileged position of influence in the VET system as well as access to skilled labour.
Keeping Danish school-based apprenticeships on employers’ terms
In Denmark, to avoid unions and employers considering school-based apprenticeships a threat by becoming a school-based alternative to the dual route to a certificate, the system was deliberately set up to work against that happening (Dobbins and Busemeyer, Reference Dobbins and Busemeyer2015). Thus, there are significant disincentives for students to seek out the school-based option instead of a firm-based apprenticeship (see Ministry of Education, 2018a). The salary for school-based apprenticeships is significantly lower than the salary received in a regular apprenticeship. Moreover, students that take up school-based apprenticeships are obliged to continue to apply for firm-based apprenticeships and must show regional mobility in the apprenticeship offers they consider. The result is that even though students may start out in school-based apprenticeships, most end up in some form of firm-based training during their education (Jørgensen, Reference Jørgensen2017). Overall, efforts to prevent competition between school-based and standard apprenticeships have worked, since among students the school-based choice is generally considered less prestigious. This is seen for example in that around half of the students, who qualify for enrolment in school-based training, do not accept this option (Jørgensen, Reference Jørgensen2017), and students from school-based apprenticeships have lower employment rates than students from firm-based apprenticeships (Danish Evaluation Institute, 2012).
Furthermore, while this institution was employed in a context where VET increasingly worked as an inclusionary measure for a residual group of weak learners, the risk of it becoming a tool of social policy was mitigated by limiting access to the school-based option. In the 2000s, with support from the bourgeois government and the social democratic-led opposition, the take-up of students in school-based apprenticeships was increasingly capped. The reason could either be a too strong demand for in-firm training in these programmes, or because unemployment levels in certain sectors were too high (Jørgensen and Juul, Reference Jørgensen and Juul2010) – in both cases signifying too weak a demand from employers. In other words, the school track was only allowed to grow in the areas where employers were not able or willing to provide the necessary apprenticeship positions and where school-based apprenticeships would not crowd out firm apprenticeships.
Beginning in the 2000s, Danish unions and employer associations also increasingly pushed for VET to be relieved of its inclusionary role (e.g., Confederation of Danish Employers in Svansø, Reference Svansø2006) and started floating ideas about setting up admission requirements to increase the quality of VET (e.g., Confederation of Danish Industry, 2008). This frustration among unions and employers was a key background factor for the 2015 reform of Danish VET (interviews with union and employer representatives), which was supported by all parties in parliament except the far left-wing Unity List. Grade requirements for admittance to VET were implemented, with the only other options for admittance being a signed apprenticeship contract or the passing of an admission test (Carstensen and Ibsen, Reference Carstensen and Ibsen2021).
The aim was to address the problem of declining numbers of young people seeking admittance to VET directly from primary education. Beforehand, the system witnessed a severe decline from about 30% in 2005 to just under 20% since 2012, signalling a weakening reputation of VET among young people and therefore threatening employers’ provision of skilled labour. In an interview, a representative of the Social Democratic Party argued that this worked as a burning platform for pushing through the ideologically very sensitive policy of grade requirements. The 2015 reform was thus a clear break with the use of VET for social policy purposes, making it impossible for the weakest students to access the system, and consequently also school-based apprenticeships. The message was that students that are not qualified to be in VET should find something else to get ready for further education, in most cases the ‘transition system’ where it is possible to obtain the necessary grade level (Di Maio et al., Reference Di Maio, Graf and Wilson2020). It also clearly signalled that looking out for these students was the responsibility of the state and municipalities, not employers. Once again, policymakers changed the school-based part of vocational education, leaving employer discretion in the apprenticeship system untouched (Nyen and Tønder, Reference Nyen and Tønder2020).
In 2016, the transition system was reformed to deal more effectively with residual groups struggling to progress into further education or employment (Den Forberedende Grunduddannelse). As part of a reform of the transition system, the guiding goal of Danish education policy for more than two decades – that 95% of a youth cohort should at least gain secondary education – was given up and replaced instead with a 90%-goal, and an explicit recognition that, for some, unskilled labour was the right option rather than education (Government, 2017). In exchange, the employers committed to increased apprenticeship provision in an agreement with unions and the state (Ministry of Education, 2016). Hence, the refusal on part of employers that VET continues to be used as social policy does not mean an all-out rejection of equality-enhancing measures. However, it indicates a growing vigilance where policies threaten to undermine employers’ privileged position in accessing skilled labour.
In sum, the Danish VET system is set up in a way that employers are relieved of some of the pressure of providing apprenticeships while at the same time not losing any of the benefits of the apprenticeship system. Notably, employers are not being forced to take on apprentices and they retain the privileged access to future skilled labour if they do. This chimes with a strong tradition in Danish VET policy that the state avoids intervention in apprenticeship training, focusing instead on regulating the school-based aspect of vocational training (Nyen and Tønder, Reference Nyen and Tønder2020).
Employers, right-wing coalitions and the Austrian workshop-based apprenticeships
Also in Austria, the expansion of statist training happened on employers’ terms – at least up to a certain extent. For the Austrian state, it was always essential that individuals in public workshop-based apprenticeships find adequate employment opportunities later, necessitating that public workshop-based training is to a certain extent compatible with employers’ interests. Employers have only consented to an increased role of the state where they did not fear a potential competition to standard, firm-based dual VET. Austria’s main employer association, the Austrian Economic Chamber (Wirtschaftskammer Österreich WKÖ), campaigned against all measures they perceived to make workshop-based training more attractive for individual apprenticeship seekers than a standard apprenticeship, fearing shortages of labour supply on the apprenticeship market due to an increasing take-up rate of statist training measures. Such a shortage could potentially lead to increased wages and better working conditions for apprentices – in other words, increased labour costs. Employers consequently focused on shaping statist measures in ways that create incentives for individuals to join standard apprenticeship training. As an interviewed representative of the employer association argued, ‘the goal can never be to simply put apprenticeship seekers into the workshops but to facilitate their placement into a firm-based apprenticeship’.
These preferences are reflected in the specific public workshop track that Austrian employers favour. As confirmed by multiple interview partners, employers prefer the partially workshop-based track (ÜBA2), where trainees only sign training contracts for the first year of apprenticeship and are subsequently supposed to find a standard apprenticeship place at an employer (see also WKÖ, 2019) – similar to Denmark’s school-based apprenticeship. These tracks also feature extended internship periods at private companies, which workshop candidates must secure before starting workshop-based training (Wieland, Reference Wieland2020). In contrast, employers insist on keeping places in fully workshop-based tracks (ÜBA1) as low as possible, where individuals receive a training contract for all years of apprenticeship, and without securing internship places at private firms. This is not surprising, as the partially workshop-based track pushes workshop trainees back into the standard apprenticeship market, preventing alleged shortages of labour supply.
Employers profit financially from this model (Seitzl and Unterweger, Reference Seitzl, Unterweger, Bonoli and Emmenegger2022). As argued by several of our interview partners, firms benefit by outsourcing the first year of apprenticeship to workshops and hiring the most promising candidates out of these training workshops as firm-based apprentices. This allows companies to avoid the net costs of the first year of apprenticeship, while still profiting from post-training benefits, that is by ‘retaining a higher share of apprentices and savings on future hiring costs’ (Moretti et al., Reference Moretti, Mayerl, Muehlemann, Schlögl and Wolter2019: 229). Training quality in workshops is quite high (Bergmann et al., Reference Bergmann, Lechner, Matt, Riesenfelder, Schelepa and Willsberger2011; Salzmann-Pfleger, Reference Salzmann-Pfleger2016), making them an attractive option for Austrian employers. While employers could not prevent the introduction of fully workshop-based tracks in 1998, the JASG law established more training places in the partially workshop-based variant. Since then, these partially workshop-based tracks have been firmly institutionalized, and not even left-leaning governments tried to abolish them.
However, the specific makeup of workshop-based training does vary with which parties are in government power. Especially in times of Austria’s right-wing governments, employers’ preferences have prevailed in shaping statist training measures. Already in 2000, a coalition government between the populist radical right (FPÖ) and the Christian Democrats (ÖVP) passed a law that downgraded the fully workshop-based track to an option of ‘last resort’ in case trainees are, despite intense efforts, unable to find a standard apprenticeship at an employer (Republik Österreich, 2000). At the same time, only apprenticeship seekers with positive school-leaving certificates were subsequently allowed to join the training workshops, resulting in training workshops only admitting a pre-selection of better qualified unsuccessful apprenticeship seekers, reducing potentially negative reputational effects of workshops for dual VET as a whole. In parallel to the transition system in Denmark, separate shortened or prolonged workshop tracks were created for those without a positive school-leaving certificate and other socio-economically, psychologically or physically disadvantaged groups (formerly Vorlehre). Originally, these tracks did not even allow students to achieve a standard VET degree at a workshop. Only after pressure from the labour camp, this possibility was implemented (see the Chamber of Labour’s statement, AK, 2000). While integrative apprenticeships (Integrative Berufsausbildung) improved labour market inclusion for individuals with physical or psychological handicaps, this reform also prevented all those with negative or no school-leaving certificates, regardless of their individual prospects of success, from visiting standard public training workshops (Heckl et al., Reference Heckl, Dörflinger, Dorr and Klimmer2008).
The fully workshop-based apprenticeship was reintroduced as a standard workshop track in 2008 by a social democratic-led coalition government, but as announced already in the coalition agreement of the new right-wing government (ÖVP and FPÖ, 2017), this option was subsequently transformed back to a measure of ‘last resort’. Since 2020, public training workshops consequently only sign contracts with trainees for the first year of apprenticeship (Rechnungshof Österreich, 2021). Like the Danish school-based apprenticeship, all trainees in Austrian public training workshops consequently need to apply for standard apprenticeship positions in firms after their first year of apprenticeship, although trainees that cannot land an apprenticeship place are still offered a prolongation of their training contract in order to receive full VET qualifications at the workshop. Austrian employers are also keen on keeping wages of trainees at workshops as low as possible to create incentives for individuals to join a standard apprenticeship with a full apprenticeship wage. The 2017–2019 right-wing government therefore cut the wages for workshop trainees, originally as high as standard apprenticeship wages, roughly by half, thereby allegedly making it more attractive for individuals to join standard apprenticeships (Knecht and Bodenstein, Reference Knecht and Bodenstein2019).
In sum, employers managed to keep state-led apprenticeships in both countries – to a certain extent – on their terms. They did not challenge the increased role of the state per se, as they benefit from the system themselves, but only the characteristics of the workshop-based system that they perceived to compete with standard firm-based apprenticeships. However, Austrian employers prevailed mostly in times of right-wing governments, while in Denmark employer discretion is more pronounced regardless of who is currently in government. The next section shows that this surprising variation in the level of inclusiveness resulting from state intervention depends on the coalition partner available to employers.
Coalition partners and variation in state intervention
In the following, we argue that Austrian unions were able to pursue more inclusive policies together with left-wing parties, partially limiting employer discretion when left-wing parties held government office. We show that this alliance emerged because unions have institutional resources outside the governance of the collective VET system. In contrast, the Danish case displays great stability of social partner coalitions, because the unions’ role in VET governance is almost solely dependent on employer involvement in the dual apprenticeship system. This firmly keeps VET on employers’ terms.
Stable employer-union coalitions in Denmark
Although it is not rare for Danish unions to complain about the lack of employer commitment to providing enough apprenticeship positions (see trade union publication Skou, Reference Skou2017, also Juel, Reference Juel2021), strong conflict has not been openly voiced since the late 1980s (Dobbins and Busemeyer, Reference Dobbins and Busemeyer2015). The unions’ stable coalition with employers in the reform of VET is connected to the diminishing political power of Danish unions. In recent decades, unions have moved from being primary agenda setters in the corporatist arena and have increasingly found themselves outside the influence afforded by inclusion in the policymaking process (Andersen et al., Reference Andersen, Dølvik and Ibsen2014; Jørgensen and Schulze, Reference Jørgensen and Schulze2011).
One place where unions have maintained their place at the table is in VET governance, but the recent side-lining of the corporatist arena in labour market and welfare policy indicates that government parties – including the Social Democrats – give influence to the unions only as long as they maintain peace with employers in the VET system. The primary importance of keeping unions represented is a broader bolstering of the ‘flexicurity’ model that has historically played a crucial role in strengthening the economy’s competitiveness (Madsen, Reference Madsen2004). To effect change on the system, then, unions need the support of employers. This cooperative dynamic also plays out at the level of VET schools and consequently school-based apprenticeships. Here both employers and unions are placed in the board of VET schools – and thus play an important role in providing input, particularly about local businesses’ skill needs – but the authority to implement the law is placed at the management level of individual schools, and, ultimately, with the Minister of Education. In this system, the influence of the social partners is dependent on their involvement in governance bodies responsible for the dual VET system as a whole and therefore on their joint efforts to keep collective training alive. Consequently, throughout the reform period under scrutiny here, we witness great stability in the coalition between employers and unions struggling to maintain their influence on the system.
In Denmark, both social partners have remained vigilant that the school-based apprenticeship model does not compete with the regular firm-based apprenticeships, which would threaten employers’ privileged position in the apprenticeship system. In contrast to Austria, both social partners jointly limited the inclusionary approach of consecutive left- and right-wing governments of the 1990s and 2000s with a cap on the take-up rate of students in school-based apprenticeships. Consequently, social partners’ joint influence on which VET programmes feature school-based apprenticeships secured employers’ privileged position in apprenticeship markets, fearing a potential crowding out of firm-based apprenticeships by statist training measures.
Similarly, in 2014, both social partners jointly put their foot down against the use of VET as social policy (Carstensen and Ibsen, Reference Carstensen and Ibsen2021). Although VET was placed centrally in the government’s reform programme, it initially had no plans to implement a grade requirement (Government, 2011). As a response, social partners broke off the work in the respective tripartite commission and instead agreed on a joint statement in favour of grade requirements (LO and DA, 2013), which made the government and opposition parties consent (interviews with union and social democratic party representatives). From a party-political standpoint, the major challenge was to get one of the government partners – the Social Liberal Party – to accept a grade requirement that went fundamentally against the ideology of the party’s education policy. However, according to an interviewed representative of the Social Democratic Party, the key to making the reform was the acceptance of the social partners’ demand for a grade requirement.
From the perspective of both social partners, the collateral damage of implementing a grade requirement was worth sending a clear signal that VET is not a social policy measure but instead set up to provide relevant skills (Carstensen and Ibsen, Reference Carstensen and Ibsen2021). 7% of the students who applied to a VET programme in 2018 were called for an admission test of which just over 40% passed (Ministry of Education, 2018b). But research conducted prior to the reform showed that about 50% of students that did not have the necessary pass grades would successfully complete a VET programme upon being admitted (Hvidtfeldt and Tranæs, Reference Hvidtfeldt and Tranæs2013), and the reasons for dropout in this group of students were rather connected to immaturity or lack of apprenticeship positions (Hetmar, Reference Hetmar2013). Again, to ensure the viability and reputation of the overall VET system, unions deliberately consented to a reform that could be expected to have negative inclusiveness effects, building an alliance with employers to secure their influence in the apprenticeship system.
When the 95%-goal was ultimately given up and instead replaced with a 90%-goal, it was accepted that the last part of the residual group, that had until then been the task of VET to include in the labour market, could instead legitimately remain as unskilled labour. What had been considered a ‘creed’ of Danish education policy since the early 1990s (Expert Group on Youth Education, 2017) was given up without struggle from the left-wing, including unions that had been part of initiating it three decades earlier. The introduction of grade requirements for accessing VET and the abolishment of the 95%-goal was a significant break with the egalitarian ideological stance of the centre-left government and the unions. More generally, Danish left-wing parties and unions have long pushed for furthering access of working-class youth to vocational and higher education and to increase social mobility (Nelson, Reference Nelson, Busemeyer and Trampusch2012; Dobbins and Busemeyer, Reference Dobbins and Busemeyer2015). Thus, the Danish unions’ adamant support for the 2015 reform is a significant break with their policy of the last almost seven decades directed at increasing access for the working-class youth. But considering the unions’ dependence on employer involvement in the dual apprenticeship system, it is far less surprising that in this case the dog (i.e., the unions) did not bark.
Unions, left-wing parties and VET reform in Austria
Coalitional dynamics differ strongly in Austria. Compared to Denmark, the power of employers has more often been limited in VET politics, in all cases with heavy support from the unions. The Chamber of Labour and the trade union federation, as the main representatives of labour, regularly sided with left-wing parties and opposed business interests to pursue more inclusive reforms, while Danish unions defended business interests in the VET system. This led to a clear left-right divide in Austria and shifted the centre of debate to the parliamentary arena, where employers prevailed when right-wing governments were in office, and more inclusive policies were implemented by coalition governments led by left-wing parties. We argue that the Chamber of Labour and unions could side with the Social Democrats (SPÖ) in Austria because their role in public training workshops is to some extent independent of employers’ commitment to collective training (Durazzi and Geyer, Reference Durazzi and Geyer2020). Similar to Denmark, social partners, via the Federal Advisory Board on Apprenticeships, are responsible for workshops as well as standard apprenticeships (Emmenegger and Seitzl, Reference Emmenegger and Seitzl2020). However, the social partners’ role is not, like in Denmark, limited to such governance bodies responsible for the administration of dual VET as a whole and thereby only a function of employers’ commitment to collective training. Rather, the social partners’ role is also institutionalized in bodies that either have been newly set-up for the workshop-based system or are located outside the traditional apprenticeship system and are embedded in Austria’s broader system of social partnership.
First, social partners were strongly involved in the initial set-up of training workshops in the late 1990s via newly created federal-state level ‘project boards’ solely dedicated to the workshop-based system and independent from collective VET governance structures (Lechner et al., Reference Lechner, Nemecek, Riesenfelder, Willsberger, Michenthaler and Brandenburg2004). Second, as the public training workshops were never considered to be ‘schools’ but active labour market policy, they are financed not via the budget of the Ministry of Education but through Austria’s public employment service (Arbeitsmarktservice AMS). The Chamber of Labour and the Economic Chamber have substantial decision-making powers in the AMS (Weishaupt, Reference Weishaupt2011) – independent of their role in the respective governance bodies of the collective VET system. Even as the responsibilities of the abovementioned ‘project boards’ were transferred to the AMS in 2002 (Lechner et al., Reference Lechner, Nemecek, Riesenfelder, Willsberger, Michenthaler and Brandenburg2004), social partner influence was consequently secured. Via the AMS’s administrative board – and not via bodies dedicated for collective VET governance like in Denmark – the social partners are involved in the implementation and control of the workshop-based system, for example by setting quality standards and guidelines (Durazzi and Geyer, Reference Durazzi and Geyer2020). They also have strong influence in sub-national boards of the AMS, which allows them to influence the types and quantities of occupations trained in workshop-based learning (Bergman et al., Reference Bergmann, Lechner, Matt, Riesenfelder, Schelepa and Willsberger2011). Third, the social partners themselves run roughly half of the workshops financed by the AMS, operatively involving unions in the workshops’ implementation on the ground (Frick et al., Reference Frick, Gregoritsch, Holl and Kernbeiß2015).
In other words, the Austrian Chamber of Labour and unions could secure a strong long-term influence in the workshop-based system outside of governance bodies responsible for collective VET. Even if employers further withdraw from their role as apprenticeship providers, and therefore the importance of collective VET governance bodies potentially diminishes, unions would still be involved in the governance of the workshop-based system. Moreover, although social partners are jointly included in the governance of the workshop-based system, unions are more influential compared to standard dual VET (Durazzi and Geyer, Reference Durazzi and Geyer2020), where employers’ control over firm-based training is much higher. All this helps to explain why unions see no necessity to consent to employers’ demands to restrict the role of public training workshops and instead side with left-wing parties to push for more inclusive policies in VET reforms.
For instance, during the introduction of the training workshops in the late 1990s under a social democratic led coalition government (Republik Österreich, 1998) and heavily promoted by unions and the Chamber of Labour (see for example their statements in AK, 1997 and confirmed by our interviews), employers were not able to prevent fully workshop-based apprenticeship tracks that encroached on their control over apprenticeship supply. It is exactly these fully workshop-based tracks that make the system inclusive, as they give individuals who are unable to get an apprenticeship place at an employer the possibility to gain a full, standard VET certificate at the workshop. In combination with the high quality of training, workshops are acknowledged by employers to be roughly equivalent to standard apprenticeship training at a firm, thereby leading to relatively good labour market chances for workshop graduates (Bergmann et al., Reference Bergmann, Lechner, Matt, Riesenfelder, Schelepa and Willsberger2011; Salzmann-Pfleger, Reference Salzmann-Pfleger2016). Due to pressure by the Chamber of Labour, unions and Social Democrats, employers were never able to permanently abolish this possibility to complete an apprenticeship programme at a training workshop but only managed to reduce it to a measure of ‘last resort’ (Lechner et al., Reference Lechner, Nemecek, Riesenfelder, Willsberger, Michenthaler and Brandenburg2004; on the position of the Chamber of Labour, AK, 2000).
The Chamber of Labour, unions and the Social Democrats have also criticized the (from firms’ perspective) financially beneficial outsourcing of the first year of apprenticeship to workshops, as they fear a reduction in training participation of firms in the first apprenticeship year in general and a shift of costs from private firms to the state (see the article by the unionist Trinko, Reference Trinko2013, published by the Chamber of Labour). Consequently, they want to limit such possibilities for firms by offering apprentices the opportunity to fully complete their apprenticeship at the workshop by institutionalizing this option as a regular, ‘standard’ workshop track (interview with representative of the Chamber of Labour). In 2008, a social democratic led government implemented a ‘youth guarantee’, where every individual received the right for education and training until the age of 18 (Salzmann-Pfleger, Reference Salzmann-Pfleger2016). As long demanded by the unions and promoted by Social Democrats (cf. ÖGB, 2002; SPÖ, 2002), the reform reinstated these fully workshop-based tracks as a regular option, without the obligation to search for an apprenticeship place at a private firm – even though this was ultimately reversed again by the 2017–2019 right-wing coalition. Furthermore, even though they could ultimately not prevent it, the Chamber of Labour, unions and Social Democrats heavily opposed the 2017–2019 coalition government’s reduction of wages for workshop trainees, originally as high as standard apprenticeship wages, fearing increased training discontinuations due to the, from an individual’s perspective, increase in the costs of gaining a VET degree (AK Oberösterreich, 2018).
In short, Austrian employers’ power to set the terms of state intervention is restricted by heavy opposition from unions that side with left-wing parties, ultimately making the workshop-based apprenticeship system, in contrast to the Danish system, comparatively inclusive, albeit with fluctuations over time.
Conclusion
Structural pressures have led to a reconfiguration rather than demise of coordinated capitalism. Their new face now often involves a more prominent role for the state. A case in point are the collective skill formation systems that in recent years have seen the share of state-provided training grow. This paper counters the assertion that high-capacity states can maintain high levels of inclusiveness in the face of structural pressures. Instead, we argue that variation in the level of inclusiveness generated by state intervention depends on coalitional dynamics between employers, unions and parties. By examining the introduction and reform of state-led apprenticeships in Austria and Denmark, we show that if unions depend on employers for their organizational influence, they become strong coalitional partners for employers in fending off demands originating in the parliamentary arena. In contrast, if unions have influence independent of any coalition with employers, unions and employers are likely to team up with political parties that they share preferences with, making the level of inclusion dependent on who currently holds government power.
In this way, our paper shows under what circumstances the parliamentary arena and partisan politics become central in VET politics, which have often been considered ‘the wrong places to look for the politics of institutional change in skills systems’ (Culpepper, Reference Culpepper2007: 612). Recent contributions in comparative political economy have tended to emphasize either partisan politics or producer groups. However, it is more plausible to assume that both the parliamentary and the corporatist arenas matter. In our paper, we connect the two arenas. More specifically, we argue that political conflicts are structured either along ideological lines or between arenas – dependent on the extent to which unions rely on employers for influence. These coalitional dynamics can lead to – at first sight – surprising outcomes. For instance, we find that in Austria, employer discretion is more limited by shifting ideological commitments of left- and right-leaning governments than in Denmark. Moreover, available evidence suggests that developments in Germany, the best-known collective skill formation system, are more in line with the coalitional dynamics in Denmark (Busemeyer, Reference Busemeyer2012; Durazzi and Geyer, Reference Durazzi and Geyer2020), which runs counter to the most common typologies in the literature (e.g., Hassel and Palier, Reference Hassel and Palier2021).
Do our arguments travel beyond collective skill formation systems? Clearly, our arguments focus on processes of de-collectivization, where the state is asked to fill the void previously occupied by employer collective action, which is different from areas where governments pursue unilateral liberalization agendas (Rathgeb, Reference Rathgeb2018). In addition, we emphasize the importance of unions’ dependence on employers for exercising their organizational influence. These conditions are likely to be met in many policy fields in CMEs, especially in economic and social matters governed by neo-corporatist institutions. One pertinent example would be the provision of social inclusionary measures through labour market policy, where employers historically have accepted varying degrees of state intervention (Martin, Reference Martin, Coen, Grant and Wilson2010). Importantly, there is no necessity for unions to enter a coalition with employers. Rather, the more important organizational interests are for unions’ long-term positions of influence and the lower/unclear the costs of non-inclusive policies are for unions and their constituencies, the more likely they are going to enter a coalition with employers.
Outside the boundary conditions surrounding the main argument of the paper, and thus leaving it for future research to assess, one could plausibly conjecture that the mechanism analysed also applies in the setting of increased state intervention outside coordinated capitalism. Although the institutional consequences of the economic crises in the 2010s and 2020s have yet to fully materialize, they currently seem likely to deal a serious blow to the neoliberalism of the last four decades and instead bring much greater room for state intervention in the economy. If so, it will be interesting to follow to what extent employers will be able to make such interventions work on their terms, which is likely to depend on the coalition partners available to them and thus differ between policy fields. In sum, then, we believe that our more nuanced and political account of state intervention is likely to apply to a large set of political processes.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1755773922000017.
Acknowledgements
We thank the three anonymous reviewers, the editors of European Political Science Review as well as Leonard Geyer, Lukas Graf and Cathie Jo Martin for helpful comments on earlier versions of this manuscript. The paper was presented at the 27th International Conference of Europeanists, June 22−24, 2021. The authors are grateful for the comments received in this setting. All remaining shortcomings are the authors’ responsibility.
