The Formation of Education and Training Regimes in Historical and Comparative Perspective

Scholarly contributions by Arnold Heidenheimer (Reference Heidenheimer1973, Reference Heidenheimer1981), Margaret Archer (Reference Archer1979, Reference Archer1984, Reference Archer1989), Andy Green (Reference Green1990), Fritz Ringer (Reference Ringer1979) and Abram De Swaan (Reference De Swaan1988) are good examples of an older strand of literature, which is located midway between historical macro-sociology and political science. The aim of these studies was to describe and explain differences in the institutional set-up of the educational system. Methodologically, these studies relied on qualitative and historical methods and small-N comparisons.

Heidenheimer (Reference Heidenheimer1973, Reference Heidenheimer1981), for instance, was mostly concerned with explaining how education had developed differently in the United States compared with how it had developed in European countries. He emphasized the importance of the relative timing of macro-social processes such as bureaucratization and democratization, although he also studied the influence of trade unions and professional interest groups. For Heidenheimer, the intense competition between local, state, private and religious educational institutions for students and resources fuelled the early expansion of higher education institutions in the United States, whereas in Europe, higher education institutions were important instruments to limit access to elite positions in powerful state bureaucracies (Heidenheimer Reference Heidenheimer1981: 272).

The topic of decentralization also features prominently in the work of Archer (Reference Archer1979, Reference Archer1984, Reference Archer1989). Archer draws a fundamental distinction between the centralized and decentralized education systems that are associated with different types of policy making: in centralized systems, policy reforms have to be initiated through the centre of political authority, whereas in decentralized systems, institutional change can unfold below the radar of legislative politics because of local innovation (Archer Reference Archer1989: 259–60). Most importantly, Archer criticizes contemporary educational sociology for not paying enough attention to both the ‘cross-national’ perspective and politics in general. The first task in developing a more comprehensive approach to the analysis of education systems is to study ‘the social origins of educational systems; to account, in short, for why historically different nations generated varying structures of systems’ (Archer Reference Archer1989: 255). For Archer, politics and political decision making play an important role in this enterprise:

Unfortunately, Archer's calls for a more historically oriented and politically sensitive sociology of education seem to have had only a minor impact on the further development of the discipline.Footnote 2

The seminal work of Kathleen Thelen addressed an important research gap by exploring the development of vocational training systems in the United States, Britain, Japan and Germany (Thelen Reference Thelen2003, Reference Thelen2004; Thelen & Kume Reference Thelen and Kume1999). Thelen's work is thus an important complement to the earlier work mentioned above, which focused more on the development of general (primary and secondary) and higher education and tended to neglect vocational education and training. With regard to the theoretical framework used, Thelen's work is rooted in historical institutionalism, industrial relations and comparative political economy (see below). In order to explain why vocational training systems developed in different directions, Thelen emphasizes the importance of the balance of power between different groups of actors. In particular, the preponderance of industrial unions over craft-based unionism and the politically supported continued viability of the crafts sector explain why firms remained committed to vocational training in countries such as Germany, whereas in others (Britain and the United States) apprenticeship training faced inexorable decline (Thelen Reference Thelen2004: 20–2).

An important common feature of the contributions made in this field of scholarship is the emphasis put on the role of time and timing (Archer Reference Archer1979: 8; Heidenheimer Reference Heidenheimer1981: 270; Thelen Reference Thelen2004: 31–3). Early events matter more than later events, because they have a strong impact on the selection of development paths (Pierson Reference Pierson2000). Power struggles mediated by institutions (such as the level of decentralization in education systems) at critical junctures in history determine the fate of education systems. What these studies neglect, however, is the role of political actors in the narrow sense, i.e. political parties. These play an important role in the field of scholarship to be introduced next.

Quantitative Studies of Education Policies and Spending

In addition to the historically oriented, small-N studies introduced in the previous section, we find an emerging literature on the political, institutional and socio-economic determinants of policy output in education. Scholarship in this field differs from the previously introduced work on an important point: instead of identifying all the causes that led to the emergence of different education systems (‘causes-of-effects’ approach), it is more interested in assessing the impact of specific independent variables on education policy output (‘effects-of-causes’ approach, cf. Mahoney Reference Mahoney2010). A significant part of this literature uses quantitative methods and large-N research designs, sometimes combined with qualitative case studies. Mirroring a similar approach taken in comparative welfare state research, a number of studies in this field are concerned with identifying the determinants of education spending, i.e. public investments in human capital formation. These studies also borrow their theoretical framework from comparative welfare state research, i.e. questions are often asked such as ‘Do parties matter?’ (Schmidt Reference Schmidt1996) or ‘Do institutions matter?’ (Weaver & Rockmann Reference Weaver and Rockmann1993) in terms of their effect on policy output and outcomes. It is our intention, first, to present an overview of quantitative studies on education, especially on education spending. Subsequently, we will discuss qualitative work in this field of scholarship, which deals with explaining differences in policy making as well as in the governance of educational institutions.

A sizeable literature analyses the statistical determinants of investments in education. Early work in this tradition (Castles Reference Castles1982; Verner Reference Verner1979) revolved around the question of whether political variables were needed at all to explain education policy output or whether output was mostly driven by socio-economic forces. These early studies were hindered by the lack of internationally comparable data on the finance of education.Footnote 3 With the exception of the work of Francis G. Castles (Reference Castles1989, Reference Castles1998), the study of education policy output remained an unexplored subject for a long time. This is why, in recent years, some exploratory studies have rediscovered the field (Ansell Reference Ansell2008a, Reference Ansellb; Boix Reference Boix1997, Reference Boix1998; Busemeyer Reference Busemeyer2007; Iversen & Stephens Reference Iversen and Stephens2008; Schmidt Reference Schmidt2007). In an important contribution, Iversen and Stephens (Reference Iversen and Stephens2008) highlight the elective affinities between well-known welfare state typologies (Esping-Andersen Reference Esping-Andersen1990) and the ‘three worlds of human capital formation’ and call for a reintegration of education and training systems into comparative welfare state research (Iversen & Stephens Reference Iversen and Stephens2008: 602).Footnote 4

Since Castles (Reference Castles1982), the role of government parties as a determinant of education spending has remained a prominent research question. Compared to the more classical objects of study in partisan theory (such as welfare state spending, as in Huber & Stephens Reference Huber and Stephens2001), education is a peculiar case because the redistributive implications of increased investment in education, in particular higher education, are not as clear-cut as in the case of other social policies. What is more, skill-biased technological change has been identified as an important force mitigating the equalizing effects of investments in education, although, as Mosher points out, the balance of power between business and labour continues to influence the distribution of resources on the labour market (Mosher Reference Mosher2007).

On the one hand, parties of the left are expected to expand educational opportunities because unskilled or less skilled workers, being the core electorate of social democratic parties, have an interest in enhancing their human capital (Busemeyer Reference Busemeyer2007, Reference Busemeyer2009b; Castles Reference Castles1982, Reference Castles1989, Reference Castles1998; Schmidt Reference Schmidt2007). Carles Boix (Reference Boix1997, Reference Boix1998) has made a very important contribution to the literature by highlighting the fact that social democrats’ preference for investing in education should be seen as a consequence of changes in the economic context. Since the forces of economic globalization have made Keynesian demand-side management policies unfeasible, social democrats increase public investment in human and physical capital formation as their preferred supply-side strategy.

On the other hand, scholars have questioned whether investment in education is in fact a primary policy goal for leftist parties. A core finding of the sociological literature is that the upper classes have privileged access to education, particularly higher education (Stevens et al. Reference Stevens, Armstrong and Arum2008). Hence, the redistributive potential of education is lower than that of more traditional transfer policies, so that parties on the left should not be expected to expand education at any cost (Ansell Reference Ansell2008b, Reference Ansell2010; Jensen Reference Jensen2011). Busemeyer (Reference Busemeyer2009b), however, finds that the participation of social democrats in government is associated with increases in public spending on higher education – a result that he explains as attempts by leftist parties to reach out to new voting groups in the middle classes.

Other intensively studied topics in this field concern the division of labour between public and private sources of finance for education (Wolf Reference Wolf2009; Wolf & Zohlnhöfer Reference Wolf and Zohlnhöfer2009), the impact of population ageing on education spending (Poterba Reference Poterba1997; Wolter & Cattaneo Reference Wolter and Cattaneo2009) and the relative impact of age and class on individual-level spending preferences (Busemeyer et al. Reference Busemeyer, Goerres and Weschle2009). Tepe and Vanhuysse (Reference Tepe and Vanhuysse2009) provide evidence for the existence of political business cycle effects in the hiring policies for teachers in German Länder, i.e. governments hire more teachers when elections are imminent. Busemeyer (Reference Busemeyer2008) shows that education spending, in contrast to other types of social spending, is positively associated with fiscal decentralization.

Qualitative Studies on Education Policies and Governance of Educational Institutions

In contrast to the increasing number of quantitative studies, there are surprisingly few studies that compare education policy output from a more qualitative perspective. Of course, scholarship in the neighbouring disciplines such as the comparative education sciences provides abundant empirical material on contemporary education systems, but these kinds of studies often focus more on describing the differences between education systems rather than explaining them. If they engage in the latter, theoretical frameworks tend to be borrowed from education sociology, organization theory or management studies rather than political science. Furthermore, comparative education science and sociology are more interested in the effects of education policies on social indicators and societal outcomes (see later) rather than the political and institutional factors that explain their enactment. In any case, the kind of medium-N studies on processes of policy change prevalent in comparative welfare state research (e.g. Huber & Stephens Reference Huber and Stephens2001; Pierson Reference Pierson2001) are lacking so far.

However, we find a few exceptions to this general rule. Following in the footsteps of Heidenheimer, scholars such as Ansell (Reference Ansell2008a, Reference Ansell2010) and Lindert (Reference Lindert2004) have emphasized the deep connection between education and democracy. Ansell (Reference Ansell2008a, Reference Ansell2010) relies on quantitative and qualitative analyses of educational politics and policies, engaging in an impressive tour d'horizon spanning the world of both developing and developed countries and examining both historical and contemporary developments. In addition to partisan politics (see above), Ansell finds that democracy and democratization are positive driving forces of educational expansion, not just historically but also in contemporary developing countries. Lindert (Reference Lindert2004: chap. 15) also emphasizes the importance of democratic voting rights as driving factors in the expansion of mass public schooling starting in the late nineteenth century and continuing over the course of the twentieth century. Lindert's work is also remarkable with respect to the way education is treated as an integral part of the welfare state. Another attempt to integrate the study of education with questions of social equality and cohesion is the edited volume by Green, Janmaat and Preston (Reference Green, Janmaat and Preston2006).

In addition to studies on democracy and education, an important area of qualitative scholarship consists of studies on the governance of education institutions. This topic spans different disciplines and seems to be a useful focal point around which scholars from political science, sociology and the education sciences might come together. In contrast to standard approaches in comparative public policy, this line of research is less interested in explaining differences in policies than in assessing the impact of policies on the governance of and distribution of power within educational institutions (Enders Reference Enders2004). To return to the influential book by Burton R. Clark (Reference Clark1983), scholars in this field are interested in comparing and describing cross-national differences in the way educational institutions are organized and govern themselves (for example, Becher & Kogan Reference Becher and Kogan1992; Goedegebuure et al. Reference Goedegebuure, Kaiser and Massen1994). The lasting contribution of Clark has been to provide a triangular set of diverging forces – academic oligarchy, state authority and market forces – that have proved helpful in explaining both the tensions within the governance of higher education as well as cross-national differences in higher education systems. Reflecting the lack of interest in the determinants of policies, Clark (Reference Clark1983: 2–4) actively encouraged the study of governance mechanisms in educational institutions independently of the surrounding political context, because he regarded higher education as a (semi-)autonomous sub-system of society.Footnote 5 Policies (and their reform) feature as exogenous factors in some works, upsetting and transforming the governance of educational institutions (for example, Musselin Reference Musselin1997), but the study of how this transformation of the political context (within and beyond educational institutions) feeds back into the policy-making process has not as yet become an integral part of the research agenda (Enders Reference Enders2004: 373).

More recently, however, there seems to be some movement on that front. In particular, in the wake of widespread privatization and the introduction of quasi-markets in the provision of education, scholars have started to look at the politics behind the introduction of these reforms, which are aimed at transforming the governance of educational institutions. A comprehensive volume edited by Braun and Merrien (Reference Braun and Merrien1999) assesses the impact of new public management reforms on higher education governance in a number of Western democracies. Lundahl (Reference Lundahl2002) provides an extensive overview of the role of government parties in Swedish education reform during recent decades. Ball (Reference Ball1993) and the volumes edited by Ahier and Esland (Reference Ahier and Esland1999) and Flude and Sieminski (Reference Flude and Sieminksi1999) study the case of Britain, distinguishing themselves from the other contributions in that their perspective goes beyond higher education to encompass general and vocational education. Klitgaard (Reference Klitgaard2007, Reference Klitgaard2008) compares the sometimes puzzling and paradoxical politics of school voucher reforms in the United States, Germany and Sweden. In a comprehensive comparison of the politics of institutional change in the higher education systems of Norway, Sweden and England, Kogan, Bauer, Beiklie and Henkel (Reference Kogan, Bauer, Bleiklie and Henkel2006) adopt a theoretical framework that borrows not only from the classical works in governance studies (Clark Reference Clark1983; Becher & Kogan Reference Becher and Kogan1992) but also from historical institutionalism and comparative public policy (cf. Bleiklie & Kogan Reference Bleiklie and Kogan2006: 10–15). Recently, the governance perspective has become a very popular approach in studying the internationalization of education policy (Jakobi et al. Reference Jakobi, Martens and Wolf2009a; see also later).

Open Questions

Open questions and potential issues for future research in the broad area of comparative public policy are the relationship between vocational training and higher education and the association between the public and private provision of education as parts of the preferred policy packages of political parties. For example, it is not yet clear whether social democratic parties in government prefer investing in higher education or vocational training, whether this depends on the specific context of education systems or whether this relationship has changed over time. Similarly, conservative parties are often associated with limiting public investment in education and promoting market-enhancing governance reforms (Lundahl Reference Lundahl2002), but Ansell (Reference Ansell2008b, Reference Ansell2010) and Jensen (Reference Jensen2011) find that conservative parties can, under specific conditions, also favour increased public investment in education because this is less distributive than social transfer spending.

Furthermore, very little is known about the individual-level determinants of policy preferences on education. Whereas, in comparative welfare state research, we find a large literature on individual-level preferences and positions with respect to the social state (e.g. Andress & Heien Reference Andress and Heien2001; Arts & Gelissen Reference Arts and Gelissen2001; Svallfors Reference Svallfors2003), there is still a lack of similar research for the case of education.Footnote 6

Finally, the governance of educational institutions will remain an important research topic for years to come, in particular in relation to the increasing internationalization of education systems (see subsequent section). A fruitful avenue for future research in this field would be to explore the role of policies and policy change in effecting real institutional change in the governance of educational institutions and the feedback effects of these changes on the distribution of political power in education policy making.

In the following section, we discuss studies on the rise of international organizations (IOs) as important players in education policy. This phenomenon has questioned one central assumption of comparative public policy, namely that national education systems and policies are independent from developments in the international political economy. However, political science has reacted quickly, and the issue of the internationalization and Europeanization of education policy has developed as a new research field.

Governance Mechanisms and Policy-Making Processes at the International Level

The internationalization of education policy is associated with the emergence of new governance structures and policy-making mechanisms at the international level. The most recent overview of the role of international organizations in education policy is provided by Leuze et al. (Reference Leuze2008: 5), who analyse and classify the ‘governance instruments’ used by international organizations in education policy making, such as discursive dissemination, standard setting, financial incentives, co-ordination activities (for example, the EU's Open Method of Co-ordination (OMC)) and technical assistance.

We also find a number of in-depth case studies on the education policies of specific IOs such as the GATS or the EU, as well as on particular governance mechanisms such as the OMC. The studies of Scherrer (Reference Scherrer2005, Reference Scherrer2007) and Verger (Reference Verger2009) discuss the role of GATS as a driving force of the privatization of education. Concerning EU education policy, the most comprehensive recent studies are those by Ertl (Reference Ertl2006), Keeling (Reference Keeling2006), Pépin (Reference Pépin2007), Balzer and Rusconi (Reference Balzer and Rusconi2007), and Walkenhorst (Reference Walkenhorst2008). On the basis of a content analysis of EU documents from 1970 to 2006, Walkenhorst reveals a shift in EU education policy from ‘politico-economic’ goals to ‘economic-functional’ goals, and hence to a pro-market orientation. He also argues that, at the EU level, an ‘intensive transnational policy regime’ has emerged which is driven forward by ‘supranational entrepreneurship’ (Walkenhorst Reference Walkenhorst2008: 569). Besides the EU Commission, Walkenhorst names the European Court of Justice (ECJ) as well as the OMC as the key actors and instruments of EU education policy. Balzer and Rusconi (Reference Balzer and Rusconi2007: 71) and Ertl (Reference Ertl2003: 29) concur that the European Commission is the most important player at the EU level. The recent study of Borrás (Reference Borrás2009) shows that, with the re-launching of the Lisbon Strategy in 2005, the role of the Commission changed. Whereas earlier it was confined to an ‘administrative’ function, it is now able to perform an ideological and political leadership function.

Bound by the legal prohibition to harmonize national education systems, the OMC plays an important role as a ‘soft’ governance instrument. However, as Gornitzka (Reference Gornitzka2006) argues, its use still signals a watershed in European education policy. A reason for this might be the fact that, at different levels, states make strategic use of this instrument as a springboard for domestic reform politics (de Ruiter Reference de Ruiter2009).

With regard to the question of why international organizations increasingly become involved in education policy, the studies provide two main answers. On the one hand, the OECD, GATS and EU act as international and supranational policy entrepreneurs and use education policy to drive forward the liberalization and privatization of political economies as well as to expand their competencies (Scherrer Reference Scherrer2005, Reference Scherrer2007). On the other hand, internationalization can also be seen as a consequence of strategic decisions of national governments, public authorities and domestic socio-economic interest groups. In this perspective, the EU and its initiatives in education policy are exploited as a means to overcome reform blockades in domestic politics. The studies of Martens and Wolf (Reference Martens and Wolf2006) and Nagel (Reference Nagel2009) as well as the edited volume of Amaral, Neave, Musselin and Maasen (Reference Amaral, Neave, Musselin and Maasen2009) illustrate this argument for the Bologna Process. Linking the international level with the national level also makes research into the internationalization of education policy of interest for those researchers studying the dynamic of two-level games. The study of the effects of international education policy on national systems is, therefore, the second major issue for research in this field.

Effects of Internationalization on Domestic Policies, Polities and Politics

Within the literature on the effects of internationalization on national education policy, polity and politics, the three major foci are as follows. First, there are studies which explain the effects on the basis of a multi-level perspective. These studies argue that IOs influence national systems, but their effect depends on domestic institutions, veto actors and politics (Martens & Wolf Reference Martens and Wolf2006; Trampusch Reference Trampusch2009). Secondly, another strand of literature understands the effects of internationalization as a process of policy learning and diffusion, emphasizing the role of ideas and processes of norm construction (Balzer & Rusconi Reference Balzer and Rusconi2007; Jakobi & Martens Reference Jakobi and Martens2007; Jakobi Reference Jakobi2009). An instructive overview of these two approaches is provided by Leuze et al. (Reference Leuze2008). While these two strands of thinking are mostly adopted by political scientists, we can add a third perspective, which comes from education science, but which supplements the other two perspectives with an outcome-oriented analysis. Contributions to education science view education systems, occupations, qualifications, certifications, competencies etc. as socially constructed. Thus, they argue that, due to different national trajectories in constructing skills, convergence is unlikely although international organizations are increasingly engaged in education, as is, for example, the EU, with its initiative of establishing a European Qualification Framework (EQF) in the wake of the so-called Copenhagen Process (Brockmann et al. Reference Brockmann, Clarke and Winch2008; Clarke & Winch Reference Clarke and Winch2006). Tellingly, Clarke and Winch (Reference Clarke and Winch2006) ask: ‘A European skills framework? – but what are skills?’

Research into the effects of IOs at the national level mainly benefits from detailed case studies (Trampusch Reference Trampusch2008) or small-N country comparisons (Dobbins & Knill Reference Dobbins and Knill2009; Trampusch Reference Trampusch2009; Witte Reference Witte2006). Witte (Reference Witte2006) analyses the implementation of the Bologna Process in France, Germany, the Netherlands and England. She shows that public authorities have used the Bologna Process to legitimize national reforms in higher education, and that not only domestic institutional veto points but also the domestic problem load are intervening variables in the process of Europeanization. The study of Dobbins and Knill (Reference Dobbins and Knill2009) assesses the degree of policy convergence in higher education policies in Central and Eastern Europe and shows that the Bologna Process has triggered national reforms in higher education in these countries as well. Trampusch (Reference Trampusch2009) analyses the effects of the Copenhagen Process on vocational training policy in Austria and Germany. Embedding the analysis of Europeanization in a more general analysis of institutional change, she concludes that, in Germany, Europeanization occurs as reform policies and politics; in Austria, however, Europeanization occurs more as institutional change by default, and hence without strategically enacted reform initiatives shaped by the EU and as domestic institutional change that occurs anyway.

The literature does not yet provide clear answers as to the impact of economic internationalization on the role of the state in the provision of education. Scherrer discusses the probable effects of including education as a tradable service in the GATS. According to Scherrer (Reference Scherrer2005), the direct effect of the GATS on national education systems is still low, although it helps to ‘lock in liberalization measures’, which are initiated outside the GATS negotiations by other IOs such as the OECD or the EU. In contrast to Scherrer, in his study of the effects of globalization on education, Ansell (Reference Ansell2008a), relying on a quantitative research design, finds that under certain conditions, economic globalization is associated with the expansion of public education spending. Taken together, political and economic globalization have ambiguous effects on the states’ role in education: on the one hand, economic globalization drives states to increase public spending; on the other hand, political globalization and Europeanization in this policy field is used in various countries as a means to a further privatization and liberalization of education (Bologna). To sum up, what are the main achievements of the research into internationalization and Europeanization in education policy, and where do we still find gaps?

A striking feature of this strand of literature is that it shows how political science has responded quickly to the new development of internationalization in education. Education policy is meanwhile an established research object in international political economy (IPE) and in the literature on Europeanization. Furthermore, the studies are empirically grounded, inductive and actor-oriented, exploring the new developments of internationalization and Europeanization by means of case-oriented qualitative methods in order to develop new hypotheses and theories. This kind of research design is a necessary precondition for further studies in this field, which may apply quantitative methods and deductive reasoning.

Nevertheless, three challenges should be mentioned. First, it is surprising that research into Europeanization and internationalization is less multi-disciplinary than that into the effects of education systems or the studies we discussed in the domain of comparative public policy and comparative political economy. The existing studies mainly search for political factors instead of linking their explanations of change to studies in sociology or education science, for example, which could provide our discipline with well-grounded hypotheses on the effects of cultural, socio-economic and institutional factors.

Secondly, the studies mainly focus on the role of strategic reasoning, analysing internationalization and Europeanization as a matter of reform politics or reform ideas or policy learning. However, one may argue that political and socio-economic actors merely react in a trial-and-error fashion. Confronted with multiple international influences and opportunities, and raised in highly fragmented national education systems, political reformers may simply not know what will happen when they react to the OECD, the EU or GATS.

Thirdly, the effects of the internationalization and Europeanization of education policy on national democracies and identities as well as on democratic legitimacy are left by the wayside. This deficiency is astonishing in so far as education policy and education systems are not only linked to national trajectories of identity and sovereignty, but are also the result of historically evolved economic, social and democratic structures. It seems that education policy is no longer an autonomous domain of the nation states, which has far-reaching implications for the development of national democracies.

In sum, the research into internationalization and Europeanization may benefit not only from including a more comprehensive perspective on the growing influence of international organizations on education systems, but also from allowing more diversity into their theories. Moving beyond the assumption of strategic and rational actors, on the one hand, and connecting research into reform policies and politics in the context of internationalization and Europeanization with the analysis and theories of institutional change (Streeck & Thelen Reference Streeck and Thelen2005), on the other, are, in our view, the two main challenges with which this strand of literature is confronted. In this more actor-oriented perspective, questions of timing and sequencing are of interest, as it is still unclear why we observe this growing influence of international organizations increasingly since the 1990s and not earlier.Footnote 7

Industrial Relations, Education and Training

A common element of the early and the later contributions is an analytical perspective rooted in neo-institutionalism. These works posited that the training decisions of firms could not be regarded as being independent of the institutional context that firms and other market participants were embedded in. In an influential contribution in the early stages of this literature, Finegold and Soskice (Reference Finegold and Soskice1988), for example, explained the failure of repeated attempts by the British government to establish a German-style dual apprenticeship training system by pointing out the differences in the socio-economic institutional framework between these countries. Along similar lines, Streeck (Reference Streeck1989) argued that the kinds of institutions that encourage or even force firms to invest in the training of their workforce cannot be created by means of rational decisions taken by the firms themselves, but rest on political and social foundations and are imposed on firms in the form of ‘beneficial constraints’ (Streeck Reference Streeck1992, Reference Streeck2004). Intermediary associations such as employers’ associations, chambers of commerce and trade unions were regarded as an important and integral part of the implementation process of these constraints (Streeck & Schmitter Reference Streeck and Schmitter1985). The constraints proved to be ‘beneficial’ because they forced firms in co-ordinated market economies such as Germany to specialize in high-quality production strategies (‘diversified quality production’, cf. Streeck Reference Streeck1992), which went along with high wages for workers without academic credentials and thus contributed to an overall low level of social inequality.Footnote 9 Streeck's concept of ‘beneficial constraints’ opens up space for political explanations, because ex ante it is an open question who (i.e. which actor) would be willing or able to impose the kinds of constraints on business that might be beneficial in enhancing human capital formation.

In the 1990s, much of the debate centred on identifying and explaining ‘high skills models’ and ‘low skills traps’. The German case featured prominently in these debates as a positive role model (Franz & Soskice Reference Franz and Soskice1995; Soskice Reference Soskice1994, Reference Soskice1999), but Japan and – with reservation, because of the lack of employer involvement in training – the Scandinavian countries were perceived as successful models as well (Ashton & Green Reference Ashton and Green1996; Brown Reference Brown2001; Crouch et al. Reference Crouch, Finegold and Sako1999; Culpepper & Finegold Reference Culpepper and Finegold1999; Green Reference Green2001). More recently, scholars from the comparative education sciences have become increasingly interested in the institutional and ‘international’ context of vocational education and training, which has led to the publication of a number of edited volumes on the international comparison of Vocational Education and Training (VET) systems (Clarke & Winch Reference Clarke and Winch2007; Bosch & Charest Reference Bosch and Charest2010).

These works provide an abundance of empirical material on the contemporary – as well as partly on the historical – development of training systems in the major economies. Scholarship in this field has been very aware of the institutional complementarities between skill formation, industrial relations and labour market institutions, and scholarly interest in studying the connections between education and welfare state policies is increasing, as will be shown in the next section. From a political science perspective, issues such as the role of electoral politics, political institutions and government partisanship as important factors shaping educational institutions and education policy output remain under-researched. Policy change is often treated as an exogenous, independent variable (as when studying the impact of policies on institutional change). Therefore, the application of political science theories on the social and economic determinants of policy change to the empirically rich material provided by this strand of literature seems a very fruitful way to stimulate interdisciplinary discussion.

Skill Formation in Varieties of Capitalism

In the ‘varieties of capitalism’ (VoC) literature, the issue of skill formation or, more specifically, the skill demands of employers in interaction with political and economic institutions is the crucial factor explaining the divergent development paths of welfare state and production regimes (Hall & Soskice Reference Hall and Soskice2001; Iversen Reference Iversen2005; but see also Crouch & Streeck Reference Crouch and Streeck1997; Hollingsworth & Boyer Reference Hollingsworth and Boyer1997; Estévez-Abe et al. Reference Estévez-Abe, Iversen and Soskice2001; Iversen & Soskice Reference Iversen and Soskice2001; Kitschelt et al. Reference Kitschelt, Lange, Marks and Stephens1999; Mares Reference Mares2003; Soskice Reference Soskice1999). In co-ordinated market economies (CMEs), employers rely on non-market forms of co-ordination via strong employers’ associations and corporatist institutions that facilitate cross-class compromises. In this institutionally dense environment, employers invest in the formation of industry, occupation and firm-specific skills and support generous welfare state policies because these policies help to convince workers to co-invest in these specific skills (Estévez-Abe et al. Reference Estévez-Abe, Iversen and Soskice2001; Iversen & Soskice Reference Iversen and Soskice2001; Mares Reference Mares2003). In contrast, employers in liberal market economies (LMEs) (have to) rely on market mechanisms to co-ordinate economic action. This more competitive environment prevents and punishes skill investments by firms, so that skill formation largely takes place in formal education on the one hand and on-the-job training on the other. Instead of fostering cross-class compromise, the issue of skill formation has been contested across the class divide (Thelen Reference Thelen2004: 21) as craft unions aim at controlling access to skilled labour and employers, in turn, try to reduce their reliance on skilled workers by increasing the capital intensiveness of production. The availability of different kinds of skills – specific skills in CMEs, general skills in LMEs – is associated with distinct comparative advantages of firms in international product markets. According to Hall and Soskice (Reference Hall and Soskice2001), firms in CMEs specialize (and are successful) in the production of diversified quality products (Streeck Reference Streeck1992), i.e. specialized niche products in high-end markets, often in manufacturing, chemicals and the car industry that rely on incremental innovation. In contrast, LME firms have a competitive advantage in cutting-edge technology in new economic sectors such as high-end services, information technology (IT) and biotechnology, which require radical innovation and swift adaptation to changing market conditions. The success of the Scandinavian countries in IT, however, does not really fit with the neat classification of countries into LMEs and CMEs, as these countries are usually regarded as CMEs and their training and education systems are geared at combining the provision of general and specific skills at a high level.Footnote 10

The strength of the VoC school is that it yields concrete hypotheses about institutional complementarities between skill formation and the welfare state, which are open to empirical investigation. Strong vocational training institutions and a generous welfare state encourage investments in occupational skills, which enhance the competitiveness of CMEs in diversified quality production. An emphasis on higher education, in preference to vocational training, and deregulated labour markets that produce high wage premiums for college education in LMEs, bolsters competitiveness in radical product innovation. Since the publication of Hall and Soskice (Reference Hall and Soskice2001), these claims have been vetted from different perspectives.

One strand of the literature is concerned with the implications of the VoC arguments on the micro level. Iversen and Soskice (Reference Iversen and Soskice2001) show that skill specificity is a positive determinant of social policy preferences, i.e. workers with specific skills are more in favour of increasing social spending. Cusack, Iversen and Rehm (Reference Cusack, Iversen and Rehm2006) largely corroborate these findings and provide more macro-level evidence. According to their study, countries with ‘specific skill systems’ (i.e. strong vocational training institutions) respond to economic shocks by expanding public spending to a greater extent than countries with weak vocational training systems (Cusack et al. Reference Cusack, Iversen and Rehm2006: 383–4).

However, we also find studies that question the validity of the positive association between skill specificity and social policies. Kitschelt, for example, posits that the status of blue-collar employment is a more important determinant of redistributive preferences than skill specificity as such (Kitschelt Reference Kitschelt2006: 414). Furthermore, Kitschelt argues that those types of social policies that are closer to the employment sphere, such as unemployment insurance, should matter more than other, less directly related types, i.e. pension and health insurance (Kitschelt Reference Kitschelt2006: 414). In addition, we find a number of empirical studies that have analysed the ‘asset specificity’ thesis. Lee (Reference Lee2007), using data for the United States, finds partial support for the VoC thesis and introduces an interesting twist into the argument. According to his interpretation of the empirical evidence, workers with specific skills support unemployment protection because they find themselves trapped in a shrinking labour market due to de-industrialization and lack the necessary general skills required to update their skill portfolio (Lee (Reference Lee2007): 708). In contrast to Lee, Emmenegger (Reference Emmenegger2009) again relies on internationally comparative data, so that his results are comparable to Iversen and Soskice (Reference Iversen and Soskice2001). Furthermore, borrowing from Goldthorpe (Reference Goldthorpe2000), Emmenegger argues that skill specificity might have a different impact on perceived labour market risks than that assumed by Iversen and Soskice (Reference Iversen and Soskice2001). On the one hand, skill specificity increases labour market risks, because workers will not be able to switch easily to a different job using the same skill set. On the other hand, high levels of skill specificity reduce the probability that the employer will dismiss the worker, because such workers are hard to replace. As a consequence, workers with specific skills should worry less about labour market risks and hence demand less unemployment protection. Emmenegger finds stronger support for the latter ‘replaceability thesis’ than for the ‘asset specificity thesis’. Against these empirical studies, Busemeyer (Reference Busemeyer2009a) revisits the concept of ‘asset specificity’ from a theoretical perspective (see also Huo Reference Huo2009). In an attempt to clear the mist surrounding the conceptual foundations of the definition and operationalization of ‘specific skills’, he argues that the variety of skill regimes needs to be classified along two dimensions instead of one: the involvement of firms in skill formation, and the vocational specificity of the education system as such.Footnote 11 Busemeyer (Reference Busemeyer2009a) identifies three distinct skill regimes in co-ordinated market economies that are, in part, reminiscent of the ‘three worlds of human capital formation’ (Iversen & Stephens Reference Iversen and Stephens2008; see also Busemeyer & Nikolai Reference Busemeyer and Nikolai2010).

In defence of Iversen and Soskice, it can be argued that their model includes both the ‘specific’ risk of skill investments as well as more general labour market risks (see also Cusack et al. Reference Cusack, Iversen and Rehm2006), which might be related more to the general level of skills rather than different kinds thereof. Also, as pointed out by Kitschelt, the relationship between skill specificity on the one hand and income and class on the other is more than superficial. Whereas Kitschelt, in line with Goldthorpe, provides a more ‘sociological’ perspective on occupations and classes, Iversen and Soskice clearly adopt a political economy perspective of class. Income and skills are at the heart of this debate, and Iversen and Soskice convincingly show that in addition to the overall level of skills, investments in different kinds of skills (however they might be called) have a strong impact on labour market stratification.

A long-held criticism of the VoC school is that its characterization of LMEs and CMEs is too ‘Platonistic’, disregarding institutional change, while overemphasizing the stability of production regimes, and neglecting the study of the historical and political foundations of skill regimes (Becker Reference Becker2007; Korpi Reference Korpi2006; Stephens Reference Stephens2006; Streeck Reference Streeck2009). Explaining institutional change has by now become an important issue on the research agenda (Streeck Reference Streeck2009; Thelen Reference Thelen2007; Thelen & Busemeyer Reference Thelen and Busemeyer2008), although comparative studies of institutional change in vocational training from a VoC perspective remain rare (Busemeyer & Trampusch Reference Busemeyer and Trampusch2011; Trampusch Reference Trampusch2010a, Reference Trampuschb). The edited volume of Busemeyer and Trampusch (Reference Busemeyer and Trampusch2011) provides in-depth studies of the evolution and contemporary change of collective skill formation systems, not only in a comparative but also in a historical perspective. Trampusch (Reference Trampusch2010b) shows that patterns of transformative and self-preserving change in collective training regimes are influenced by the employers’ constellation and their interaction with elites of the public education administration. Bosch and Charest (Reference Bosch and Charest2008) look at how the ability to adapt to a changing socio-economic environment varies between LMEs and CMEs and argue that differences between the two remain important (see also Anderson & Hassel Reference Anderson and Hassel2007; Bosch & Charest Reference Bosch and Charest2010; Culpepper & Thelen Reference Culpepper and Thelen2008). Toner (Reference Toner2008) studies the ‘survival and decline’ of apprenticeship training in the Australian and British construction industries, respectively. Toner emphasizes the institutional embeddedness of training systems and argues that apprenticeship training has a higher chance of survival in Australia because its institutional setting contains a number of CME elements, although this country is widely considered to be an LME.

Besides the question of institutional change, exploring the origins of contemporary varieties of capitalism has likewise become an important new field of research. The initial VoC contributions (Hall & Soskice Reference Hall and Soskice2001; Estévez-Abe et al. Reference Estévez-Abe, Iversen and Soskice2001) specifically refrained from providing explanations of the origins of different varieties of capitalism and focused instead on understanding the effects of given institutional arrangements. Over time, proponents of the VoC school became more engaged in exploring the historical dimension, and the skill demands of employers remained the crucial independent variable in explaining the emergence of different production and welfare state regimes (Mares Reference Mares2003; Swenson Reference Swenson2002). For Cusack, Iversen and Soskice (Reference Cusack, Iversen and Soskice2007), the introduction of PR systems in countries that were later characterized as CMEs hinged on the support of employers, who viewed PR as an effective means to create the cross-class co-operation required to promote investment in ‘co-specific’ assets (see also Iversen & Soskice Reference Iversen and Soskice2009).

Several authors have questioned the thesis that the skill demands of employers are the driving force behind electoral systems and welfare states. Most prominently, Korpi (Reference Korpi2006) posits that the strength of organized labour remains the crucial variable in explaining the origins of welfare states. In a response to Korpi, Iversen and Soskice (Reference Iversen and Soskice2009) emphasize that their argument about the role of employers does not imply that unions are irrelevant. Rather, the joint investment of labour and business in ‘co-specific assets’ facilitates the emergence of cross-class coalitions, which in turn propagate the shift from majoritarian to proportional (PR) electoral institutions. Further exploring the relationship between political and economic institutions, Martin and Swank (Reference Martin and Swank2008) argue that political factors such as the electoral system explain the divergent strengths of employers’ associations, and this in turn determines whether co-ordinated or uncoordinated economies will form. In a similar vein, Trampusch (Reference Trampusch2010a), relying on case studies of Denmark, the Netherlands and Switzerland from 1870 until 1940, argues that the establishment of unemployment insurance was primarily motivated by the ‘logic of organization building of capital and labour’ rather than the skill demands of employers.

Besides the role of employers, another important issue is the question of whether modes of co-ordination are self-reinforcing equilibria or unstable social arrangements in need of particular institutional and political support. Culpepper (Reference Culpepper2003, Reference Culpepper2007), for instance, emphasizes the fragility of collective vocational training arrangements (see also Culpepper & Thelen Reference Culpepper and Thelen2008; Soskice Reference Soskice1999). By comparing East Germany and France to West Germany, Culpepper (Reference Culpepper2003) finds that the sustained viability of collective forms of skill formation needs the continued political support of political and economic actors and cannot be created from scratch. This finding mirrors earlier arguments by Streeck (Reference Streeck1992), who stressed the importance of associational networks in nurturing a supportive ‘institutional ecology’ for training arrangements. Martin (Reference Martin2000) found that, even in LMEs such as the United States, the embeddedness of firms in business networks and associations has a significant positive impact on their support for social policies that promote the formation of human capital.

Summing up, the variety of capitalism debate has rendered an enormous service to those political scientists who are interested in exploring the institutional complementarities between education, the labour market and other welfare state policies. The original purpose of Hall and Soskice (Reference Hall and Soskice2001: 2) to open up a new research agenda on these issues is certainly being addressed. However, although this has not been discussed here at length, our impression is that despite its prominent place in the theoretical discussion, skill formation and the politics of education and training reform in particular have been addressed to a lesser extent in the ensuing debates than other issues such as labour market policies, financial regulation, corporate governance and industrial relations (Jackson & Deeg Reference Jackson and Deeg2006). Furthermore, following along the lines of Iversen and Soskice (Reference Iversen and Soskice2001), a significant proportion of the research into the connection between skills and social policies has concentrated on the micro level, neglecting theorizing and empirical research at the macro level. Still, the issue of the political and historical foundations of today's education and training regimes needs to be explored further, in particular for those countries beyond the group of ‘usual suspects’, for example the Scandinavian and East Asian countries. Finally, the VoC school assumes positive complementarities between education and other social policies. More attention should be paid to those instances when education and social policy are in conflict with each other, such as in the competition for scarce public funds.