1. Introduction and theoretical framework

Strengthening the governance of skills system is a central policy challenge across many OECD countries. This introductory chapter develops an analytical framework that informs the following case studies of skills policy in Estonia, Germany, Korea, Norway, Portugal and the United States. The framework identifies four main challenges in governing skills systems: promoting co-ordination, co-operation and collaboration across the whole of government; engaging with stakeholders throughout the policy cycle; building integrated information systems; and aligning and co-ordinating financing arrangements. The chapter discusses each of these challenges in greater detail. It also includes a short discussion of the methodological approach of this study, in particular the reasoning behind selecting cases for this study and the conduct of semi-structured interviews with experts and stakeholders in the country cases.

    

Background and motivation

The governance of skills systems is complex. On the one hand, the development of skills and their effective use in labour markets follows the logic of the “life course”, where individuals acquire and make use of skills as they move through the different stages of their educational and employment careers. On the other hand, each of these different stages may be governed by different rules and regulations. Therefore, the governance of skills policy does not necessarily follow the life-course logic, and remains fragmented across different levels of government, as well as different public ministries and agencies.

This report tackles the fundamental challenge of strengthening and improving the governance of skills systems. Issues related to the development and effective use of skills remain at the forefront of debates about the transformation of industrial economies to post-industrial knowledge economies. Investing in skills and ensuring access to different forms of learning is crucial to prevent the social exclusion of disadvantaged students, learners and workers, as well as to effectively mobilise existing skills in the economy.

For the purposes of this report, a skills system can be broadly defined as covering all institutions and individuals, as well as policies, laws and regulations, concerned with the development and use of skills in the economy. Therefore, skills policies are at the intersection between various sectors of the education system, including early childhood education and care (ECEC); schools; vocational education and training (VET); adult learning and higher education; labour market policies, especially those that aim to make use of existing skills such as active labour market and training policies; policies that support the adoption of high performance workplace practices; and certain immigration policies.

The core challenge in governing complex skills systems is that policy-making responsibilities for the various policies are usually widely distributed across levels of government, as well as ministries and public agencies. Furthermore, there is often little active co-ordination between these institutions, which hampers the effectiveness of strategies to mobilise the use of skills. The guiding question of this report is therefore: as OECD countries transform into post-industrial knowledge economies, how can the governance of skills systems be strengthened and improved in order to promote the development and effective use of skills?

This question is related to and motivated by the OECD Skills Strategy 2019 (OECD, 2019[1]). The skills strategy framework has three dimensions: 1) developing relevant skills over the life course; 2) using skills effectively in work and society; and 3) strengthening the governance of skills systems. While this report focuses on the third of these dimensions, it also contributes and refers to the other two key dimensions as they are inherently related. The first dimension centres on lifelong learning, for example by promoting policies that make lifelong learning visible, rewarding, affordable, accessible and relevant. The second dimension focuses on the use of skills in the economy, for instance via policies that reduce skill imbalances, stimulate the demand for high-level skills, expand the pool of available talent, or promote social and labour market participation.

Strengthening the governance dimension of these and related policies raises four main challenges:

1. 1. Promoting co-ordination, co-operation and collaboration across the whole of government.

2. 2. Engaging stakeholders throughout the policy cycle.

3. 3. Building integrated information systems.

4. 4. Aligning and co-ordinating financing arrangements.

The goal of this report is to derive concrete policy recommendations about how governments can tackle these challenges. It presents findings from six case studies in selected OECD countries: Estonia (Chapter 2), Germany (Chapter 3), Korea (Chapter 4), Norway (Chapter 5), Portugal (Chapter 6) and the United States (Chapter 7). These countries are selected to cover a broad range of cases with different welfare state and labour market regimes, as well as different political institutions and cultures. Each case study focuses on a particular policy or sector in the country’s skills system, the study of which holds important lessons for other countries.

It is important to emphasise that country case studies are not simply highlighting “best-practice” examples, since there is rarely a case which performs “best” on all dimensions under consideration. Rather, each case study contributes insights regarding how policy makers have responded to one or (usually) several of the four challenges mentioned above. In the final chapter of the report, these different insights are integrated into a comprehensive set of policy recommendations on how to strengthen the governance of skills systems in a broad range of cases. The different policy recommendations are subsequently translated into a practical self-assessment tool for policy makers and stakeholders (see Annex A). Based on a number of “yes/no” questions, the tool allows all interested parties to assess the strength of their own country’s arrangements for governing skills policies.

In developing policy recommendations, either for a particular country or for a larger set of cases, it is important to recognise both the idiosyncratic characteristics of the case as well as more broadly held characteristics that are – under certain conditions – generalizable to a larger set of cases. At times, policy reports veer between the two extremes of arguing that “every country is different” or implying that “one size fits all”. This report aims to tread a middle ground between these two extremes by devising policy recommendations that are applicable and relevant for more than one country case, but at the same time identifying the conditions under which these recommendations can be transferred to other cases.

Adopting an “institutionalist” perspective is helpful in this regard. This perspective argues that countries can be classified and grouped into different kinds of institutional regimes, for example welfare state regimes (Esping-Andersen, 1990[2]), varieties of capitalism (VoC) (Hall and Soskice, 2001[3]) or – more relevant for this report – skill formation regimes (Busemeyer and Trampusch, 2012[4]). Institutional regimes are (more or less) coherent configurations of institutions covering different spheres of the economy that incentivise and shape the behaviour of actors in a particular system or country. Hence, rather than devising recommendations for individual countries, this report aims to identify governance practices that can be transferable to a larger set of cases, although not necessarily all cases in the OECD.

The remainder of this introductory chapter is structured as follows: the next section provides additional background on the four main challenges; this is followed by a description of how the cases were selected; and finally there are remarks on methodological aspects.

Core challenges in the governance of skills systems

Policy makers are confronted with a range of issues regarding skills policies that can be considered under four main challenges: 1) promoting co-ordination, co-operation and collaboration across the whole of government; 2) engaging with stakeholders throughout the policy cycle; 3) building integrated information systems; and 4) aligning and co-ordinating financing arrangements. The following section discusses each of the four challenges in greater detail. Corresponding policy recommendations on how to deal with these challenges are presented in the final chapter of this report, following the case studies.

Promoting co-ordination, co-operation and collaboration across the whole of government

Legal competencies and decision-making powers in skills policy are often widely distributed in both the vertical and horizontal dimensions of policy making. The vertical dimension concerns different levels of government, from local governments to regional and other subnational governments to the federal level. OECD countries differ widely regarding how exactly competencies are distributed along the vertical dimension, but broadly speaking, centralised countries, such as France or Japan, can be distinguished from decentralised countries, which delegate significant decision-making powers to lower levels of government. Among decentralised countries, there is a further distinction between countries with a formally federalist constitution (e.g. Australia, Belgium, Canada, Germany and the United States) and those without (e.g. the Netherlands, Scandinavian countries and the United Kingdom). In both groups, the involvement of the national government in skills policies is more limited than in purely centralised systems. However, among countries with systems that are formally unitary, but de facto decentralised, local governments and authorities may have a considerably higher degree of autonomy than in formally federalist countries, in which decision-making powers are often concentrated at the intermediate level of subnational governments (such as Canadian provinces, or German Länder).

Regarding the horizontal dimension of co-ordination, skills policies often require collaboration between different governmental departments. According to the conventional division of labour, education ministries and agencies are in charge of policies related to skills development, whereas labour market ministries and agencies are more concerned with policies that maximise the effective use of these policies by promoting further training opportunities and labour market activation measures. Ministries of economy and/or industry may also play a key role in skills policy by developing and promoting policies and strategies for regional, sectoral or general economic development, in which skills often occupy a central place. Ministries of finance are responsible for ensuring that policy and governance decisions are backed up with the required financial resources, and for aligning financial incentives to maximise the effectiveness of skills policies.

This seemingly neat division of labour, however, underestimates the real-world complexities of educational and employment careers, which are increasingly characterised by intermittent employment and non-linear trajectories, and where the movement between phases of skill development and usage becomes a permanent condition rather than a one-time event in the early stages of the life course. Disregarding these complexities risks promoting the social exclusion of disadvantaged workers and learners who fall between the gaps at the transitory edges of the various subsystems.

The dimension of horizontal co-ordination stretches beyond the scope of government in a narrow sense as the two dimensions of promoting co-ordination across the whole of government and stakeholder involvement are strongly connected. Societal stakeholders such as trade unions, employers’ associations, individual businesses and workers, as well as training providers, research institutes and non-profit organisations, are all deeply involved and engaged in the governance of skills systems. In OECD countries, decisions on the allocation of labour are ultimately market decisions, in which employer’s demand for certain types of labour is somehow matched by the supply of workers with particular skills. The role of government policy in this area is to set and further develop the regulatory framework that ensures an effective match between skills demand and supply, while also preventing the emergence of market failures, such as the systematic exclusion of disadvantaged groups and individuals. Where exactly to draw the line between government regulation and market autonomy, and how to balance the potentially competing interests of actors involved, is a matter of political decision and political majorities. For the purpose of this report, it is important to acknowledge that societal stakeholders are regularly involved in governance decisions and therefore need to be taken into account in efforts to improve the co-ordination of decision-making processes in skills policy (there is more on the issue of stakeholder involvement in the next section).

In addressing the issue of co-ordination and collaboration along the vertical and horizontal dimensions, this report subscribes to the whole-of-government approach (Christensen and Laegreid, 2007[5]; UNESCO and ILO, 2018[6]). This approach is based on the idea that the adoption of administrative practices according to the new public management (NPM) model, which has been influential in many OECD countries and beyond, contributes to fragmentation in the development and implementation of policies. This is because the NPM approach recommends the adoption of market-oriented steering modes to the world of public administration, which promotes competition rather than collaboration between actors and institutions. The whole-of-government approach, instead, adopts a holistic perspective that aims to maximise the payoffs of collaborative decision making, while maintaining the autonomy of the actors involved. As observed by Christensen and Laegreid (2007[5]): “WG [whole-of-government] activities may span any or all levels of government and involve groups outside government. It is about joining up at the top, but also about joining up at the base, enhancing local level integration, and involving public-private partnerships.”

Thus, in an ideal setting, the whole-of-government approach amounts to a collective and well-co-ordinated policy response to pressing problems. In a more realistic setting, the implementation of policies according to the whole-of-government model alone is challenging. Given the multitude of actors with different interests, achieving a consensus between a broad range of stakeholders within and outside government is difficult, maybe impossible if distributive conflicts (“who gets what from whom?”) are dominant. Furthermore, implementing the whole-of-government approach requires actors to adopt a certain kind of “interaction orientation” (Scharpf, 1997[7]), i.e. they would need to participate in collective decision making with a motivation to contribute to joint problem solving rather than particularistic bargaining. This kind of interaction orientation needs to be present in most, if not all actors in the system, as there is a danger that even actors motivated by problem solving may resort to particularistic bargaining if they realise that others are doing so. Finally, there is a danger that the whole-of-government approach may deteriorate into some form of centralised super-bureaucracy, in which the autonomy of stakeholders outside of government are compromised. Hence, the involvement of non-governmental stakeholders in governance decisions is crucial (see next section for more on this point).

The broader challenge of policy co-ordination along the horizontal and vertical dimensions can be summarised in three main challenges:

1. 1. Putting skills policies at the top of the government agenda: Maximising the potential of skills policies to contribute to economic well-being and social inclusion implies that these policies have to be put at (or close to) the top of the government’s policy-making agenda. As legal competencies and decision-making powers are distributed across levels of government and governmental departments (see above), prioritising skills policies is more challenging compared to other policies, since there is no single agency or actor that is fully responsible – besides the government as a whole. Skills policies therefore compete with a multitude of other policy priorities for the top spot in the government agenda. Some of these competing policies may be better placed to occupy the central place in the government agenda as they promise concrete benefits to policy makers (and their voters) in the present, whereas skills policies often require policy makers to adopt a long-term oriented perspective. Furthermore, even though policies that support skills investment are broadly popular among citizens, they are still often politically contested and debated (Busemeyer et al., 2018[8]). Citizens (and political parties representing their interests) may disagree regarding the relative distribution of fiscal resources across different policy areas, or on the relative emphasis to be put on concerns related to social inclusion or economic growth. Furthermore, citizens may, and often do, disagree regarding the design of skills policies, for instance, the role of the state versus private providers. For policy makers, the challenge is to integrate these competing interests into a coherent whole that can serve as a guiding model for policy reforms, while keeping skills policy at the top of the agenda.

2. 2. Identifying and engaging with relevant stakeholders: As argued above, adopting a whole-of-government approach entails the notion of pursuing collaborative partnerships between state and non-state actors. A first challenge to overcome in this regard is to identify who the relevant stakeholders actually are. Depending on the complexity and size of a particular skills system, policy makers at the central level may simply not be aware of the multitude of non-state actors involved in governance decisions at different levels of government. Related to this is how to distinguish between relevant and non-relevant actors and stakeholders, even when all of them are known. There are clearly strong incentives for non-state actors to get involved in governmental decision making; at the same time, not all of those who could potentially be involved should necessarily be involved, not least because some might pursue a particularistic or idiosyncratic agenda. Ideally, governmental actors should have at their disposal a general rule, recognised as legitimate and effective by those concerned, that helps them to distinguish between relevant and non-relevant actors. Ultimately, decisions about which stakeholders to involve and how strongly depends on the prevailing political context. Once stakeholders are properly identified, governmental actors still need to find ways to effectively engage and communicate with them, which will be discussed in greater detail in the next section. The central challenge for stakeholder engagement with the whole-of-government approach is how to create commitment among stakeholders to adopt a joint orientation towards problem solving, rather than particularistic bargaining, and to follow and take seriously a strategic agenda at the same time.

3. 3. Encouraging co-ordination between central and subnational authorities: Achieving collaboration and co-ordination across levels of government (vertical dimension of co-ordination) is at least as challenging as engaging with non-state stakeholders (horizontal dimension of co-ordination). Compared with other policy areas, subnational governments usually have more competencies and decision-making powers in skills policies, which makes subnational governments particularly sensitive regarding efforts to re-balance or challenge the existing division of labour between central and lower levels of government. Furthermore, depending on the country context, competencies for different parts of the skills system are distributed unevenly across levels of government. For instance, labour market or lifelong learning policies are often in the hands of central ministries or agencies as they are directly related to national labour markets, whereas responsibility for the provision and financing of education, in particular schooling but also higher education, is delegated to subnational governments. In some cases, local governments are strongly involved in education and labour market policies, whereas in others, these responsibilities are divided between local authorities and local representatives of national agencies. To sum up, the distribution of governance competencies across levels of government is likely to vary significantly across different parts of the skills system, which is a challenge, and debates about re-organising the co-ordination between central and subnational authorities are likely to accompany political conflicts about the proper division of labour.

Engaging stakeholders throughout the policy cycle

Engaging with non-state stakeholders can be an effective instrument to support policy makers in dealing with the inherent complexities of skills policies. However, the process of engagement needs to be well-balanced and grounded in sound strategy to avoid the involvement of a multitude of stakeholders contributing to, rather than mitigating, further complexity.

There are various benefits of stakeholder engagement from the perspective of policy makers. First, stakeholders contribute valuable information to the policy-making process that is difficult for policy makers to access themselves. For instance, non-state stakeholders have experience regarding the real-world effects of policies and regulations, which might be quite different from the steering effects that governmental policy makers initially intended. This is also because non-state stakeholders have a certain leeway to influence the implementation of policies themselves. Information from non-state stakeholders is a valuable resource as it provides governmental decision makers with a sense of “what works and what doesn’t”.

Informational input is particularly helpful for the implementation phase of policies, but also during the phase of policy formulation and design. The notion of the “policy cycle”, which is commonly used as a point of reference in policy studies, implies that there is a constant feedback process between the phases of policy implementation and (re-)design as policies are revised according to implementation experiences and, ideally, based on input from research. Implementation experiences from stakeholders are therefore a valuable input into governmental decision making not only during the implementation phase, when governments try to improve the implementation of existing policies, but also during the policy design phase, when they attempt to design “better” policies for the next loop of the policy cycle.

The second benefit to stakeholder engagement is that the involvement of non-state stakeholders generates political legitimacy, which can itself be an important resource during the implementation phase. In order to be fully accepted and supported by those concerned, governance decisions should ideally be based on a broad consensus between involved actors. This will increase the likelihood that policy decisions will be implemented according to the initial intentions of decision makers, and thus improve the overall effectiveness of the decision-making process.

However, this ideal setting depends on a number of preconditions, some of which may be difficult to achieve. If the issue in question does not have significant distributional implications (i.e. if it is not immediately related to the distribution or redistribution of resources), and if stakeholders in principle share similar goals, then consensus among stakeholders is easier to achieve. Furthermore, the quality of decisions taken is more likely to surpass the low threshold of “least common denominator policies” if different stakeholders contribute their particular informational inputs to jointly solving a commonly perceived problem. In contrast, if the policy issue under debate is related to distributional conflicts and if the actors involved do not share similar goals, the decision-making process is likely to be more contentious. This can lead to either gridlock, when no decisions are taken, or low-quality decisions (“lowest common denominator”) where contentious, but probably important issues are systematically excluded from the agenda.

To some extent, government decision makers can mitigate some of these problems. For instance, they can attempt to promote a shared understanding of the underlying problems and goals of policy making. They could also make available additional resources to lessen distributional conflicts, i.e. to transform them from re-distributional conflicts (“who gets what from whom?”) into merely distributional conflicts (“who gets what from the central government’s budget?”), but these efforts have inherent limitations. Thus, a first precondition for stakeholder engagement to work is that the policy issue under debate is not too contentious and burdened with distributional implications.

A second important precondition is that actors involved in governance decisions adopt a constructive interaction orientation (Scharpf, 1997[7]) that values joint problem solving rather than particularistic bargaining and “strategising”. Independent of the nature of the issue under debate (see previous point), actors involved in the process of decision making need to be open to joint efforts of problem solving and not regard their involvement in the process as merely another opportunity to extract resources from the central government or other stakeholders. To some extent actors always weigh the costs and benefits of involvement based on their individual assessment, and joint problem-solving actions need to create some tangible pay-offs for the actors involved in order to be sustainable and legitimate in the long run. However, joint problem solving only works if the actors involved at least partly adopt a problem-solving interaction orientation. Otherwise, stakeholder engagement boils down to “pork barrel politics”1 in the worst case, which would eventually undermine the overall legitimacy of engagement with non-state stakeholders. Joint problem solving also implies that involved stakeholders need to share a certain understanding about the distribution of authority and decision-making powers within the group, in particular, which actors have the capacity and legitimacy to set the overall strategy.

A third precondition for stakeholder involvement to create political legitimacy is that stakeholders represent the diversity of positions within society, rather than particular segments, in order to prevent the “capture” of decision-making processes by special interests. This may be the most challenging precondition to be met as it is directly connected to the contested role of the state (or government) in mediating, channelling and influencing the distribution of power and influence across societal stakeholders and interest groups. As a consequence, there are widely different practices and philosophies across countries, which need to be taken into account when developing concrete policy recommendations.

It is possible to distinguish between a “pluralist” and a “corporatist” perspective regarding the legitimacy of stakeholder involvement and the role of the state (Lehmbruch, 1979[9]; Schmitter, 1979[10]; Streeck and Schmitter, 1985[11]). According to the pluralist view, the diversity of interests in society is best represented by associations and when the state refrains from intervening in the free competition of associations for members and influence. Similar to the notion of free competition in the marketplace, competition between associations and interest groups will eventually ensure that the prevailing groups are those that enjoy the largest degree of support and legitimacy among citizens. The pluralist view implies a more distant “at arms’ length” relationship between the state and interest groups in order to prevent the state from tilting the scales in favour one particular group. Pluralist practices of stakeholder engagement are commonly applied in the “liberal” welfare states, market economies and skills formation regimes of Anglo-Saxon countries.

Critics of the pluralist view say that the supposedly free competition between interest groups leads to power asymmetries and inequalities, as some types of interest are easier to mobilise and organise than others (Olson, 1965[12]; Offe and Wiesenthal, 1980[13]). Olson’s theory of collective action, for instance, highlights the fact that the broad, diffuse interests that often go along with collective concerns are more difficult to organise and mobilise than particularistic “special interests”. Thus, in the worst case, the supposedly free competition of interests and interest groups may lead to the dominance of special interests as they have more resources at their disposal than interest groups representing diffuse interests and the “common good”.

Corporatist practices of interest mediation try to mitigate and neutralise real or perceived imbalances of power between stakeholders by setting up decision-making bodies that put representatives of opposing interests on a level playing field that is to some extent independent of the previous real distribution of power and influence. The most common forms of corporatist decision making concern the involvement of business associations and trade unions, as these represent opposing interests in the arguably most important political cleavage in OECD countries – the conflict between business and labour interests. In countries where corporatist practices of stakeholder involvement have a long tradition, such as Austria, Belgium, Germany, the Netherlands, Scandinavian countries, Switzerland and – to a lesser extent – Southern European countries, unions and employers are regularly involved in policy decisions through established decision-making bodies, such as the Socio-Economic Council (SER) in the Netherlands or the Federal Institute for Vocational Education and Training (BIBB) in Germany.

An important advantage of corporatist decision making is that stakeholder involvement throughout the different stages of the policy cycle enhances the commitment of non-state actors to joint decisions, which facilitates the implementation of policy decisions later on. Furthermore, government actors can draw on the expertise and informational input of non-state actors during the policy design stage. However, compared to pluralism there is a higher risk that the continued involvement of the same stakeholders through the same decision-making bodies could lead to the emergence of an “insider-outsider cleavage”, where newly emerging interests have a harder time accessing decision making than established stakeholders. Furthermore, corporatist decision making does not fully prevent the risk of capture of governmental decision making by special interests. On the one hand, state actors are more likely to reach out to groups that tend to be under-represented in a free competition of interests (e.g. consumer or environmental groups) due to their lack of resources in corporatist settings, but on the other hand, these kinds of groups may become particularly dependent on state support for their survival, which affects their strategic positioning.

A final precondition that needs to be fulfilled in order to mobilise the potential of stakeholder engagement to create political legitimacy is that the involvement of non-state actors has to be meaningful and consequential, and not merely “window dressing”. This requires a genuine willingness on the part of policy makers to share decision-making powers, which ultimately depends on a certain degree of trust between non-state and governmental actors. Policy makers need to trust non-state actors to consider collective concerns rather than pursuing particularistic goals single-mindedly, and non-state actors need to trust the government to take their input seriously before they commit significant resources to getting involved in governance decisions.

Meaningful engagement with stakeholders is more likely when they have a formally pre-defined role in governance decisions and decision-making bodies. For instance, the government could make the passing of a particular policy reform conditional on stakeholders achieving consensus. In contrast, meaningful engagement is less likely, and commensurate claims are less credible, when stakeholder involvement happens in a more ad hoc and informal manner, such as when stakeholders are merely invited to provide input at a particular point in time during the policy process (e.g. by submitting opinions via a website or issuing position papers), but are not involved in a continuous manner throughout the whole process. In this sense, pluralist practices may be more susceptible to superficial stakeholder engagement than corporatist practices. If the goal is to maximise the number of stakeholders involved, the multitude of opinions voiced may become so large that governmental actors retain the ultimate decision-making powers. In corporatist settings, the number of actors may be smaller and limited to the most “relevant” actors, but the influence of these actors on decision making is likely to be larger, which contributes to a more significant involvement of non-state actors.

Meaningful stakeholder involvement throughout the policy cycle can be summarised in four key challenges:

1. 1. Building stakeholders’ trust: Building mutual trust between governmental and societal actors is crucial to ensure meaningful stakeholder involvement. Lack of trust is often a significant obstacle to expanding the role of stakeholders in governance decisions, particularly in countries without a long tradition of social partnership and stakeholder involvement through corporatist decision-making bodies. On the one hand, government actors need to develop trust in the capacity and willingness of stakeholders to contribute constructively to joint decision making rather than maximising their own self-interests. On the other hand, stakeholders need to trust governmental actors that their input is meaningful and valued in decision making. Both sides need to actively contribute to overcoming this mutual lack of trust. Trust building can and should be supported by efforts to design the process of stakeholder involvement in particular ways. For instance, stakeholder engagement should be meaningful and continuous rather than superficial and fleeting. Meaningful engagement is more likely when stakeholders are involved through pre-defined and at least partly formalised channels and bodies rather than in an ad hoc manner. Furthermore, governmental actors in charge of the overall process should be attentive to the overall balance of interests involved in the project in order to prevent the emergence of stark power asymmetries or the systematic exclusion of particular groups from the process. These efforts also increase the overall legitimacy of the engagement process.

2. 2. Engaging stakeholders takes time: There are several reasons for stakeholder engagement taking a significant amount of time. First, the initial step of stakeholder engagement is to identify all relevant (as well as potentially non-relevant) actors. Depending on the size and complexity of the skills system, this may take a significant amount of time. Furthermore, the process of separating “relevant” from “non-relevant” actors could be politically contentious and therefore, also time consuming. Mapping skills systems in terms of relevant actors and institutions before a particular process of stakeholder involvement begins is helpful to save time. Second, building trust also takes time, but is crucial to ensure the long-term sustainability and effectiveness of the governance system. Building trust is based on repeated interactions between the actors involved and cannot be created quickly. Third, the process of stakeholder involvement often goes along with the presumption that resulting decisions are based on a broad consensus between the actors involved. Organising and achieving such a consensus takes time if the underlying issues are contentious and involve distributional conflicts. Furthermore, some actors may use and abuse the process of stakeholder involvement to pressure others into supporting their particularistic goals, as consensus-oriented decision-making grants significant “hold-up” power to individual actors. In order to deal with this particular challenge, governmental actors need to retain some degree of unilateral decision-making power, which they can activate in the case of gridlock. Scharpf (1997[7]) introduced the term “shadow of hierarchy” to describe this situation. Governments need to find a way to stay responsive to demands from a general public that may demand immediate action in response to salient problems, at the same time as allowing meaningful stakeholder engagement to develop and unfold over a longer period of time.

3. 3. Resourcing adequately: Meaningful stakeholder involvement depends on the availability of sufficient and adequate resources. Providing these resources is challenging for policy makers. Besides the general scarcity of fiscal resources, a particular challenge in this regard is that providing fiscal support to non-state actors may be perceived as privileging special interests in the general public. Nevertheless, it is crucial in order to achieve meaningful stakeholder engagement. Resources need to be made available to bodies and institutions in charge of organising the process of stakeholder involvement if they are formally established as independent entities. Providing adequate resources to advisory bodies of this kind puts stakeholder involvement on a more secure, long-term foundation, and could also be invested in research capacities so that non-state actors have access to information and knowledge acquired independently of governmental bodies and agencies. Expanding the research capacities of independent advisory bodies that involve stakeholders would also support the creation of mutual trust and a joint problem-solving orientation among stakeholders as they can refer to a common, yet government-independent, source of information. Providing adequate resources for stakeholder engagement could also involve supporting the organisation and mobilisation of stakeholders themselves, in particular groups that represent diffuse rather than special interests and that therefore find it more difficult to mobilise resources. Providing direct support for stakeholders and groups should be contingent on these actors meeting certain requirements, such as democratic procedures in electing key personnel.

4. 4. Resolving conflicts of interest: The notion of stakeholder involvement usually goes along with the idea that governance decisions made in co-operation with stakeholders are based on a broad consensus shared by all actors involved in the process. However, in real-world settings, there are issues relating to conflicts about (re-)distribution, underlying values or power relations. If the nature of the issue under discussion falls into the latter category, the outcome of consensus-oriented processes involving a larger number of stakeholders may be unsatisfactory, amounting to “lowest common denominator” rather than encompassing policy solutions. If decision making requires the agreement of all involved then each individual actor has veto power, which means that significant departures from the status quo are less likely, particularly if some actors would have to accept losses. Hence, taking up the challenge of meaningful stakeholder involvement not only requires governments to be able to organise consensus-oriented solutions, but also to have at their disposal strategies for conflict resolution. Providing a sound evidence base to policy deliberations is one potential instrument in this regard, as it promotes interaction orientations among actors that are geared towards joint problem solving rather than particularistic bargaining and ideological in-fighting.

Building integrated information systems

Making good governance decisions depends on the availability of high-quality data and information. In recent decades, the OECD and other organisations have made available a range of data sources on educational attainment, adult skills and labour market outcomes. This wealth of data has triggered and made possible a wave of new research on the determinants of learning and employment trajectories, transition processes, social exclusion and inclusion, the influence of socio-economic background on educational performance, and many other issues of relevance for the governance of skills systems. Before data of this kind became available, political debates about skills policies were often strongly driven by ideological conflicts, with limited input from neutral and objective data sources. Political-ideological conflict is an inherent component of any political decision-making process, as actors subscribe to different values and represent different sectors of society with different material interests. However, before the rise of evidence-based policy making, ideological conflicts were also more likely to materialise in the form of differing opinions on the causal relationship between two (or more) variables.

For instance, before data became widely available, some might have argued that parental background obviously strongly shapes the educational attainment of children, whereas others might have questioned the strength of this association. As data on educational attainment and parental background became available, it became possible to determine (more or less) exactly the strength of the association between parental background and educational attainment. Individuals subscribing to different ideologies might still disagree about the policy responses that should be taken in response to this fact (i.e. more or less state involvement, more or less tracking in schools, etc.), but at least ideological debates are provided with a layer of objective knowledge that can help to defuse ideological conflicts to some extent (see previous section).

The real world of policy making is still far from an ideal governance model in which decisions are based on and informed by input from high-quality data and information management systems; however, improving the informational foundations of governance decisions could significantly contribute to strengthening and improving these decisions.

Building integrated information systems can be summarised in three concrete challenges:

1. 1. The multiplicity of data sources: Existing data sources on skills outcomes and policies are often fragmented, covering only parts of the skills system. In decentralised systems, data collection may also be fragmented across levels of government, as individual regions or municipalities collect their own data that may not be immediately comparable to data from other regions. Furthermore, most large-scale assessments at the international or national level are only available in the form of cross-sectional data rather than longitudinal or panel data, which makes it difficult to trace learning and employment paths over time. Of particular importance are the transition points between different sectors of the skills system, i.e. from secondary to post-secondary education or from education to employment. Hence, a comprehensive approach to building integrated information systems could aim to generate data on how individuals proceed through the various stages of their education and employment careers, from primary to secondary and higher education to employment. This implies the importance not only of information about skills and educational outcomes, but also skills assessment and anticipation exercises, labour market conditions, and learning opportunities. Collecting data of this magnitude would likely trigger serious concerns about data protection and usage among the general public, which need to be taken seriously when developing these systems. Information systems will only be accepted and used when there is a commensurate “user culture” in a particular country context. Furthermore, quantitative data from surveys and assessments should be complemented with qualitative data from other sources such as self-assessments from educators and the outcomes of school review processes.

2. 2. The multiplicity of end users: Once established, skills, learning and labour market information systems will be used by a multiplicity of users such as parents, teachers, students, researchers, policy makers and stakeholders. Different users will pose different demands to the design of information systems. Parents and students may value accessibility and transparency, whereas researchers would put stronger emphasis on data quality and availability. Policy makers and stakeholders involved in governance decisions will likely require data to be provided in a way that makes it immediately relevant for governance decisions. Regional and local governments will expect information systems to provide data for their particular zone of influence. This multitude of users and demands poses a significant challenge to policy makers when establishing and fine-tuning information management systems. If information systems are perceived as not creating sufficient user value (while being expensive), it is unlikely that enough public support can be created for their establishment and continued maintenance.

3. 3. Management of complexities: The management of large-scale information systems poses significant challenges. On the technical side, information systems need to be able to deal with the multiplicity of data sources and potentially huge amounts of data. This requires significant resources for infrastructural investments and maintenance, as well as special attention to privacy concerns among the general public related to data storage. The size of information system varies with the size of the population of a particular country. Estonia, with a population of 1.3 million, has been a pioneer in setting up a comprehensive education information system (see case study below), but scaling up these efforts to populations the size of Germany (83 million inhabitants) or even the United States (328 million inhabitants) seems challenging, if not impossible. However, if regional solutions have to be sought for larger countries, the ensuing challenge is how to harmonise them across regions. On the political side, the use value of information systems for policy making depends on their contribution to solving concrete governance problems. Hence, the output from information systems needs to be systematically connected as input to decision-making processes in the governance of skills systems. However, establishing such a connection creates new challenges as actors adopt their strategies and priorities in order to reflect the system’s parameters (“teaching to the test”), which creates negative and/or unintended side effects. Establishing information systems that rely on a multitude of data sources, including both quantitative and qualitative data, might mitigate these unintended side effects, but will also increase the complexity of information management systems.

Aligning and co-ordinating financing arrangements

Decisions about the governance of skills systems (or any other policy field) are inherently linked to questions about financing. Decisions such as the expansion or retrenchment of programmes and personnel, the building up of infrastructure such as complex information management systems, or the (re-)design of the policy-making process by involving stakeholders require or affect decisions about the allocation, distribution and re-distribution of fiscal resources. Although some government policy decisions are mainly regulatory in nature, there will often be fiscal implications.

As briefly mentioned above, the central role of skills and skills policies in the socio-economic transformation of industrial to post-industrial knowledge economies is widely recognised in OECD countries. Furthermore, plans to increase investments in education and training generally receive popular support among citizens and voters (Busemeyer et al., 2018[8]). At the same time, political calls to invest in skills and skill development are constantly at risk of being superseded by more pressing and short-term oriented demands.

This is partly due to the fact that although educational investments in schools and universities may create short-term benefits for parents, students, teachers and professors, the bulk of the pay-off in terms of economic growth, employment opportunities and social inclusion will only materialise at some distant point in the future when the current beneficiaries of educational investments are in the labour market. Therefore, other more pressing concerns such as fighting unemployment, increasing pensions or expanding healthcare with clear and immediate pay-offs in the present are likely to outdo long-term concerns for investment (Streeck and Mertens, 2011[14]). Governments keen on promoting skills policies therefore must find ways to ensure a sufficient level of investment in education and skills, while also taking seriously the short-term concerns of citizens and voters.

Investments in skills may also lose out to other policy areas in terms of fiscal resources because the benefits of education and skills are shared between a multitude of stakeholders, and the incentives for investing in skills are often not well-aligned between these stakeholders. Educational investments create significant public benefits such as a well-educated citizenry that participates actively in political and societal life, which boosts employment and wage growth. At the same time, educational investments generate significant private benefits for those at the receiving end of these benefits, primarily in the form of higher wages and better employment prospects. It is very difficult to identify exactly how public and private benefits for a given amount of educational investments are distributed. The predominant incentive structure, however, is not geared at maximising the total sum of investments in education, but rather the opposite. From the private perspective, the financing contribution of the public side should be as large as possible in order to minimise individual costs, whereas from the public perspective, which is often under pressure to constrain growth in expenditure, there is significant interest in increasing the financing contribution of private actors, such as through fees or taxes, and in reducing overall costs.

The distribution of financing responsibilities across levels of government also needs to be considered. This is particularly important in formally federalist countries, but also in formally unitary, but de facto decentralised, systems. There are many examples where financing responsibilities and incentive structures are not well-aligned. For instance, when the legal competencies for raising taxes and other sources of revenue mainly lie in the hands of the central government, but the administration and implementation of policies is left to lower levels of government (as in Germany, for example), the central government does not have strong incentives to maximise potentially available fiscal resources. Furthermore, if lower levels of government do not have to worry about raising revenue, they might overspend or misallocate resources to politically salient sectors of the skills system.

Incentive structures in the private sector can also be misaligned, for example, individuals could underinvest in skills development if education aspirations compete with short-term pressing needs. This is particularly important in the case of adult learning, which remains a relatively weakly institutionalised sector of skills systems in many OECD countries. In this case, governmental policies, such as statutory rights for lifelong learning, the provision of education credits or vouchers, counselling, and the set-up of a supporting infrastructure, can help to partly shift the underlying incentives to maximise the potential of skills policies. Individual incentives can also be misaligned when individual educational choices constituting the supply of skills on the labour market do not match the needs of labour market actors (the demand side). Philosophies vary across different countries on how to deal with the problem of skills (mis-)match, for example by tying educational choices to market forces through high tuition fees, or by adjusting the number of available study places to the skills needs of the economy. Each of these has advantages and disadvantages that need to be carefully balanced and interpreted in light of country-specific political contexts.

Incentives to invest in skills may also be misaligned in the case of business actors. As Becker (1993[15]) already highlighted in his human capital theory, business actors have strong incentives to invest in the development of specific skills among their workers that have immediate pay-offs in terms of productivity, but little incentive to invest in broader, more general, transferable and multi-faceted (Streeck, 1996[16]) skills that may have the greatest overall benefit on employment and growth, as labour market demands are constantly adapting to a changing environment. Hence, government policies need to make sure that business actors overcome short-term urges to invest in specific skills by devising policies that reward or push actors into providing more general, widely usable skills.

In countries with a strong collective skills policy regime (Busemeyer and Trampusch, 2012[4]), these incentives are provided in the form of collective wage bargaining arrangements, corporatist decision-making bodies that include different and multiple stakeholders in decision making, and a regulatory policy environment that prevents (for better or worse) the fast turnover of workers in the short term. Liberal skill regimes rely on market forces to nudge employers into investing in skills, but some, such as the United Kingdom and Ireland, have experimented recently with levy-grant schemes in the domain of vocational training. These schemes can be an effective instrument to align the incentives of business actors with the needs of skills development: Firms pay a certain levy, depending on the size of their payroll, into a joint fund. The created revenue is then used to finance training measures in individual firms. This way, training firms can offset their levy payments with training subsidies received, which sets strong incentives to offer and participate in training.

The main challenges in the domain of fiscal incentives and resources can be summarised as follows:

1. 1. Diversifying sources of funding: The payoffs of skills investments are distributed between the public, individuals (households) and businesses. Therefore, there is a case to be made that each of these three sectors should contribute to the financing of skills investment. However, countries are likely to differ regarding how exactly the fiscal burden should be shared, depending on political conditions, historical developments and culture. Hence, re-shifting the boundary in the financing of skills policies between public and private actors is likely to be contentious as it is associated with redistributive conflicts. Nevertheless, a certain diversification of funding sources could help maximise the potential of skills policies to impact future growth and employment, and avoid or at least mitigate misaligned incentive structures. For instance, levy-grant schemes can spur employers to invest more in skill development, whereas education credits and vouchers can help individuals avoid underinvestment in education.

2. 2. Finding appropriate resource allocation and budgeting mechanisms: Besides raising and securing the required funds for skills investment, a significant challenge remains in how to distribute and allocate available fiscal resources in an efficient and fair manner across different levels of government and sectors of the skills system. The core issue is to avoid incentive structures that decrease the amount of resources available overall, and support incentive structures that maximise the amount of available resources. A first priority in this regard is to avoid financial arrangements where responsibilities for raising revenues and for spending are located at different levels of government, which is likely to lead to either under or overinvestment. A second priority is to tie fiscal strategies in the domain of skills policies to long-term goals and multi-year planning processes of the government, which should decrease the likelihood that investment priorities and goals fall prey to short-term urges and pressures. Finally, a third priority is to make sure that the use of financial resources is clearly tied to accountability mechanisms and some form of performance measurement. This helps both the effectiveness (do investments have the desired effect?) and efficiency (how much needs to be spent to achieve a particular goal?) of spending, which in turn improves the evidence base for decisions regarding the prioritisation of policy programmes and instruments.

3. 3. Ensuring equity in funding considerations: Policy makers are increasingly concerned about equity-related aspects in the distribution of funding. Research has found that educational investments can contribute to mitigating inequality in the long term (Busemeyer, 2015[17]; Huber and Stephens, 2014[18]; Solga, 2014[19]). However, the contribution of skills policies to lowering inequality depends on an equitable distribution of educational resources in the first place to ensure equal chances of opportunity and low barriers of access. If left to market forces alone, decisions about the allocation of fiscal resources to different actors and sectors of the skills system may aggravate rather than mitigate inequalities, as well-performing regions or sectors usually have more resources at their disposal. Policy makers are therefore confronted with the challenge of allocating resources in a way that helps weak performers to improve, while also rewarding strong sectors and regions to maintain their incentives to perform well. Addressing this challenge can be supported by a clear and widely communicated strategy of how to balance equity with other concerns. The exact balance between these different concerns likely depends on the political context.

4. 4. Providing adequate resources: Previous work by the OECD in the context of the Skills Strategy projects has often revealed a certain imbalance between policy responsibilities and resource allocation. If policy making or implementation responsibilities are delegated from the central to the subnational level, or from government actors to semi-public or private agencies and institutions, decision makers need to make sure that the provided resources match these responsibilities. For instance, if agencies are tasked with involving stakeholders or with conducting research they need to have resources at their disposal that allow for engagement with stakeholders and researchers to be meaningful. Furthermore, the matching of resources to responsibilities needs to be undertaken in a way that enables incentive structures that support the formation of joint problem-solving perspectives, rather than encouraging particularistic bargaining. The former is more likely if public-private partnerships and stakeholder involvement are based on mutual trust rather than overly detailed and onerous accountability mechanisms.

Methodological approach

Based on the theoretical framework introduced later on in this chapter, this report provides a series of case studies from Estonia, Germany, Korea, Norway, Portugal and the United States. It is important to emphasise that the case studies do not aim to describe these countries in their entirety, but instead focus on the particular policy reforms, issues or sectors of the skills system likely to be most relevant from an international comparative perspective as they highlight successful examples of innovation in skills policies. At the same time, the country case studies should not be interpreted as “best-practice” examples that can be easily transferred to other countries. Each case has strengths and weaknesses that need to be discussed, and each skills system has idiosyncratic characteristics that affect the possibilities for policy transfer. However, taken together the six case studies show how countries have addressed the four challenges identified in this report (Table 1.1). By studying these cases in detail, it is possible to derive general statements about a broader set of OECD member and partner economies in order to identify institutional arrangements that underpin the effective governance of skills system.

Table 1.1. How the case studies correspond to skills system governance challenges
Governance challenges
	Promoting co-ordination across the whole of government	Engaging stakeholders throughout the policy cycle	Building integrated information systems	Aligning and co-ordinating financing arrangements
Estonia (Estonian Education Information System- EHIS)		X	X
Germany (The Alliance for Initial and Further Training)	X	X		X
Korea (Lifelong learning)	X	X		X
Norway (Skills Policy Council and Future Skills Needs Committee)	X	X	X
Portugal (National Agency for Qualification and Vocational Education and Training - ANQEP)	X	X		X
United States (Massachusetts‘ Early Warning Indicator System)	X		X

In order to allow researchers to draw generalizable conclusions from a relatively small set of cases, the cases needed to be carefully selected based on a sound strategy (Gerring, 2006[20]). When undertaking such a task, simple comparisons (“A is different from B”) can (and should) be avoided as long as the case selection is theory-driven. Thick descriptions of cases were welcome and important. However, case studies also needed to have an analytical focus to provide insights regarding their broader relevance and contribution. For the purpose of this report, a “case” is defined as a particular institution that is part of a larger skills system. As such, cases of this kind are always embedded in broader institutional arrangements of skills development and use in the particular countries. The case selection strategy used in this report follows the approach of selecting the “most diverse cases” in terms of both the dependent (skills policies) and the independent variable (prevailing institutional context) - for introductions to the case selection literature, see Gerring (2006[20]); Seawright and Gerring (2008[21]).

Regarding the diversity of institutional contexts, this report focuses on two aspects: 1) the institutional set-up of the system for skills development and use (“skills regime”), which encompasses the intersection between education and training institutions, as well as labour markets and industrial relations; and, 2) the degree of centralisation/decentralisation in skills regimes, which has implications for the co-ordination of policy-making activities across different levels of government.

Skills regimes and varieties of capitalism

The pertinent literature regarding the diversity of skills regimes distinguishes between different types depending on the public commitment to, and employer involvement in, vocational education and training (Busemeyer and Trampusch, 2012[4]), as well as different “varieties of capitalism” (Hall and Soskice, 2001[3]). The core argument of scholarship in this tradition is that decisions by businesses, students, parents and households are strongly conditioned by the prevailing institutional context. As a consequence of actors adapting their preferences and strategies to the status quo, the probability of radical policy change decreases over time, whereas the likelihood of institutional stability increases. The literature broadly distinguishes between two different types of capitalism: liberal market economies (LMEs) and co-ordinated market economies (CMEs).

In LMEs, co-ordination between and among employers occurs primarily via open market mechanisms, which contributes to confrontational relationships between employers and trade unions, and a rather distant (“arm’s length”) relationship between employers and state actors. Typical examples for LMEs are Australia, Canada, the United Kingdom and the United States. This report focuses on the United States.

In CMEs, in contrast, relationships between actors and stakeholders often rely on non-market forms of co-ordination, primarily via intermediate associations such as employers’ associations, trade unions and local chambers of industry and commerce. Many of these institutions also exist in LMEs, but in CMEs they play a much more important role in the governance of skills regimes due to the dominance of the “corporatist” model of interest mediation in CMEs. The notion of “corporatism” (Lehmbruch, 1979[9]; Schmitter, 1979[10]) implies that stakeholders and representatives of the interests of different sections of society are continuously “incorporated” into governmental decision making, as well as during the implementation process of government policies.

The advantage of systematically incorporating societal interests into decision-making and implementation processes, which are often consensus-oriented in corporatist countries, is that decisions reflect broader societal needs rather than particular urges of the current government majority (Lijphart, 1999[22]). Furthermore, if broad consensus between the major stakeholders in society is reached during the decision-making process, there is likely to be less conflict and opposition during the implementation phase, which contributes to a more stable, sustainable and reliable institutional framework in the long term. However, corporatist decision making also has downsides. For example, since decision-making processes aim to achieve consensual solutions, they may be slow and cumbersome, particularly when the issues at hand have distributional implications. These processes are also prone to individual actors exercising veto power for individualistic, particularistic reasons. Corporatism therefore requires all participants to have a particular “interaction orientation” (Scharpf, 1997[7]), i.e. a certain willingness to forego particularistic goals in order to achieve collectively binding and welfare-improving solutions.

Although early literature on VoC only distinguished between LMEs and CMEs, more recent contributions have identified the diversity of institutional regimes within the large group of CMEs (Amable, 2003[23]; Anderson and Hassel, 2013[24]; Busemeyer, 2009[25]). For instance, researchers have identified an Asian model of capitalism (Amable, 2003[23]) in which large corporations (and their internal labour markets) play a decisive role in the development and use of skills. Rather than investing in the development of skills for broader labour market needs, employers and firms in the Asian model tend to focus on particular “segments” of the labour market, which are defined by the concrete skills needs of large, multinational corporations. As a consequence, this model has been named the “segmentalist” skills regime in the literature (Busemeyer and Trampusch, 2012[4]). In this model, employers are strongly committed to the development of skills, but trade unions and worker representatives are much less involved in the governance of the skills regime. This report studies Korea as a representative of this regime type.

A contrasting model can be found in Scandinavian countries, which have a long tradition of corporatist decision making going back to the late 19th and early 20th century (Martin and Swank, 2012[26]). In these countries, organisational representatives of employers and workers (as well as other societal interests) are regularly included in governance decisions. Due to the historical power of trade unions and social democratic government parties, the influence of worker interests is much higher than in other countries, which is also mirrored by a higher level of union density (i.e. the share of workers who are union members). The historical dominance of unions and social democrats in government has led to a partial marginalisation of employers in the provision and financing of training, as public financing and provision has crowded out the private involvement of employers (Lundahl et al., 2010[27]). In this “statist” skills regime (Busemeyer and Trampusch, 2012[4]), governmental policies need to pay particular attention to the challenge of mobilising employers (as well as workers) to engage in skills development at the workplace. This report studies the case of skills councils in Norway as an example of such a regime.

Corporatist decision making is also widespread in Continental Europe, with countries such as, Austria, Belgium, Germany, the Netherlands and Switzerland often regarded as typical representatives of consensus-oriented decision making in formal or informal “grand coalitions”, with a strong involvement of interest associations that represent different sections of society. With a stronger focus on skills development and usage, the regimes within this subset of Continental European countries are commonly referred to as “collective” skills regimes (Busemeyer and Trampusch, 2012[4]). Similar to the “statist” skill regimes in Scandinavian countries, business interests and workers’ groups are strongly involved in the governance of the skills regime. However, in contrast to Scandinavian countries, the influence of employer associations (and employers themselves) is stronger. In concrete terms, this implies that employers and firms have a greater leeway in setting the content of workplace training, which increases the commitment of employers to provide such training. Different from the “segmentalist” skills regime, unions and workers’ representatives are stronger, which means that the leeway for individual firms to design training content and implement particular skills policies is more constrained. In short, the governance regime encourages and requires actors, stakeholders and representatives of different societal interests to work together, representing a “collective” approach to skills development and usage. The collective skills regime has many strengths, such as a strong involvement of employers and strong public commitment of governmental actors to skills policies. However, a significant downside is that decision-making processes can take a long time and lead to “lowest-common-denominator” solutions if the agreement of all actors is necessary to move forward. This report studies the case of Germany as a typical representative of a collective skills regime.

Further discussions in the literature on skills regimes and VoC centre on the question of whether Southern European countries constitute a model of their own, or whether they should be regarded as a variant of the other regime types discussed above (Amable, 2003[23]; Ferrera, 1996[28]). Although corporatist practices are also common in Southern Europe, the intermediary associations of employers and workers are less encompassing than in the Northern European cases discussed above, and therefore more prone to representing particular segments of society rather than “workers” and “employers” as a whole. Furthermore, the reach of these intermediate associations into their respective domains, i.e. their ability to support the government in implementing policies, is much more limited than their Northern European counterparts. Therefore, in Southern European countries more effort is devoted to establishing the institutional, organisational and societal foundations for corporatist decision making, as well as the durable and lasting involvement of societal stakeholders in the governance of skills policies. This report explores the case of Portugal as an example of this regime, in particular the establishment of the National Agency of Qualification and Vocational Education and Training (ANQEP).

Similar discussions to those concerning Southern European have occurred in the literature regarding the diversity of models of capitalism (and skills policies) in Eastern European countries (Bohle and Greskovits, 2012[29]). Even though Eastern European countries have a common history, the variety of capitalist models within these countries seems to be larger than their commonalities. For instance, the Baltic countries (Estonia, Latvia, Lithuania) tend to follow a neoliberal trajectory, with some Scandinavian elements added in (Toots and Lauri, 2017[30]), whereas the Visegrad countries (the Czech Republic, Hungary, Poland and the Slovak Republic) pursue a more statist and increasingly centrally controlled approach to economic, social and skills policies. This report focuses on the case of Estonia and the newly established Estonian Education Information System. Estonia has been at the forefront of developments regarding the implementation of digitalisation strategies in the public sector (Kattel and Mergel, 2018[31]). As argued above, the establishment of information management systems can effectively contribute to promoting collaboration across levels of government, as well as across governmental departments. As few countries have undertaken significant activities regarding the digitalisation of educational governance, Estonia is particularly interesting for further study. This case study will also provide insights into other governance dimensions, such as stakeholder involvement.

Centralisation and decentralisation

The degree of centralisation/decentralisation in a particular country was important in selecting cases for further study in this report. The degree of decentralisation has important implications for the challenge of co-ordinating policies across different levels of government, as discussed above. To some extent, the degree of decentralisation depends on country size, as a country such as the United States naturally faces different challenges in co-ordinating policies across levels of government than a country such as Estonia. However, the correlation between country size and decentralisation is not deterministic, and is largely dependent on a country’s historical and cultural background. There are large countries with a relatively high degree of centralisation, such as France or Italy, but others of similar size with a more decentralised structure, such as Germany. Likewise, small countries may be federalist and decentralised (i.e. Belgium) or unitary and centralised (such as the Scandinavian countries, albeit with a strong role for municipal governments).

There are different dimensions of decentralisation: administrative, fiscal and political. These dimensions often correlate with each other, but the correlation is not deterministic. Administrative decentralisation refers to the delegation of administrative tasks to lower levels of government, which may (or may not) go along with the respective fiscal capacities and political decision-making powers. Fiscal decentralisation captures the degree to which the authority to tax and spend is delegated to lower levels of government. Again, these two aspects are related, but there may be instances where the central government delegates spending authority to lower levels of government, but restricts the leeway for subnational governments to raise their own taxes (or change levels of taxation). Political decentralisation implies that formal decision- and policy-making powers are delegated from the central to lower levels of government. In formally federalist countries, such as Canada, Germany or the United States, subnational governments and entities formally retain some aspects of independent statehood and are usually represented via secondary parliamentary chambers in the federal decision-making process. However, even in formally unitary (i.e. non-federalist) countries, local levels of government may have considerable leeway in making decisions about the provision of local benefits and services.

Broadly speaking there has been a general trend towards decentralisation in the provision and financing of welfare services, in particular education, across OECD countries (Gingrich, 2011[32]; Mons, 2004[33]). This trend has been particularly strong in countries that previously had relatively centralised forms of decision making in education, such as Sweden (Arreman and Holm, 2011[34]; Lundahl, 2002[35]) or the United Kingdom (Hudson and Lidström, 2002[36]). This can be regarded as a reaction to citizen demands for a more differentiated and regionalised, and less centralised, approach to education policy. Paradoxically, the degree of decentralisation at the bottom level, for example as measured through degree of school autonomy, may be significantly lower in formally federalist countries, such as Austria and Germany, than in formally unitary (i.e. non-federalist) countries. This is because the bulk of decision-making and administrative competencies are concentrated at the mid-level of subnational governments, which are reluctant to delegate competencies to even lower levels of government. In short, the degree of centralisation and decentralisation across the three dimensions (administrative, fiscal, political) varies significantly across countries.

Considering country-specific context conditions regarding the degree of decentralisation is important in the analysis of skills regimes, and the resulting policy recommendations. On the one hand, a higher level of decentralisation may enhance the overall flexibility of the skills regime and allow subnational governments to better adjust to particular regional needs in terms of skills and local labour markets. Decentralisation may also enhance possibilities and opportunities for stakeholders to get involved in local and regional decision making.

On the other hand, a more decentralised governance structure raises the challenges of co-ordinating policy activities across different levels of government, as well as across the variety of subnational governments. For this reason, decentralisation measures are often accompanied by efforts to recentralise other aspects of decision making, or by the setting up of new governance structures that ensure the accountability of local and subnational policy makers to nationally set standards and goals. In the case of education policy, for instance, decentralisation in educational governance has been intertwined with debates about shifting from input-oriented modes of steering (i.e. focusing on the input factors to the educational process, such as the training of teachers and investments in school infrastructure) to output-oriented modes (e.g. measuring the relative performance of individual schools, teachers and pupils in the form of comparative assessments).

Developing measures and policy recommendations that strengthen the governance of skills systems – as is the goal of this report – must consider institutional context conditions such as the degree of decentralisation in individual countries. In highly centralised countries, a certain delegation of decision-making and implementation competencies to lower levels of government might be sensible in order to improve the overall flexibility of the system and its responsiveness to local needs. In contrast, highly decentralised countries might benefit from a certain degree of recentralisation and a strengthening of accountability mechanisms across levels of government and between subnational governments in order to more effectively pursue a “whole-of-government” and co-ordinated approach to skills policy.

In order to study the variety of governance challenges related to decentralisation, this report includes and selects country cases with different degrees of decentralisation along the three dimensions mentioned above (administrative, fiscal, political). Germany and the United States are both formally federalist countries, which implies that subnational governments (German Länder and USA states) have a high degree of autonomy in setting policies within their respective domains. This poses particular challenges for co-ordination across levels of government, as subnational governments demand strong involvement in decision-making processes at the federal level. In the German case, the Alliance for Initial and Further Training is studied. This alliance is an attempt to bring together stakeholders from different levels of government in order to devise a joint strategy on skills policies. The United States case focuses on the local and regional levels and explores the leeway that subnational governments have in devising their own governance arrangements within a framework of governance that is significantly influenced by federal regulations.

In contrast, Korea and Portugal are formally unitary (i.e. non-federalist) and relatively centralised countries. Therefore, they in particular face the challenge of how to involve local and regional stakeholders in governance decisions at the national level. If the mid-level subnational/regional level of government is absent or weakly developed, the obvious problem is how to effectively reach out and include local stakeholders, which are usually more fragmented and numerous than subnational governments at the regional level. Both case studies explore how central governments and their agencies have managed to partly deal with this challenge.

In the final two case studies from Estonia and Norway, the challenge of co-ordinating across levels of government is less daunting, either because these countries are relatively small in size, and/or because they have a long tradition of autonomous local government.

Empirical approach to case studies

The following case studies apply standards and procedures that are common in qualitative case study research in empirical social science. Each case study first provides some background information on the historical and contemporary developments in skills policy in the particular country of interest. This background analysis is based on insights from desk-based research and the careful scanning of primary source material such as government reports, policy papers and occasionally newspaper reports, complemented with insights from scholarly literature where available.

The project team visited each country for a period of five to seven working days to conduct first-hand interviews with experts, stakeholders and representatives from different institutions and organisations involved in the governance of skills policies. For each country, one particular example, such as a particular institution, organisation or policy reform, was selected to be studied in greater detail, rather than assessing the skills regime of a country as a whole. This more focused approach is better suited to highlight and identify concrete successes and challenges in policies related to skills development and usage, and leads to more concrete policy recommendations. As stated, it is important to highlight that these examples should not be understood as “best-practice” examples, but as differentiated case studies that show what does and does not work. For each case, the interviews were transcribed and analysed following the country mission and serve as background material for the analyses and the development of policy recommendations.

Each country chapter in the report contains an extensive analysis related to the four challenges identified in Table 1.1. All chapters also include a number of country-specific policy recommendations, equally connected to the four main challenges in strengthening the governance of skills systems and summarised in Table 1.2.

Table 1.2. Overview of country-specific policy recommendations
Case study	Governance challenges	Recommendations
Estonia: The Estonian Education Information System (EHIS)	Building integrated information systems Engaging stakeholders throughout the policy cycle	Use questions as drivers of data collection Support innovative data analyses Put the data into perspective Streamline and automate the data collection process Promote the use of EHIS data among schools Increase user friendliness
Germany: The Alliance for Initial and Further Training	Promoting co-ordination acrossthe whole of government Engaging stakeholders throughout the policy cycle Aligning and co-ordinating financing arrangements	Make better use of the potential of the Alliance to promote innovative skills policies Prevent multiplication of governance bodies Link duration of Alliance agreements with election cycles Allow sufficient time for meaningful stakeholder engagement
Korea: Lifelong Learning in Korea	Promoting co-ordination across the whole of government Engaging stakeholders throughout the policy cycle Aligning and co-ordinating financing arrangements	Institutionalise co-operation between the Ministry of Education and the Ministry of Employment and Labour Strengthen vertical co-ordination Establish a comprehensive database on lifelong learning Introduce quality control mechanisms Engage with employers and employees in order to develop employment relevant lifelong learning Educate employers and employees about the importance of lifelong learning
Norway: Norway’s Skills Policy Council and Future Skills Needs Committee	Promoting co-ordination across the whole of government Engaging stakeholders throughout the policy cycle Building integrated information systems	Give new governance arrangements enough time Improve format of meetings and agenda setting process within the Skills Policy Council Improve communication and demonstrate to Skills Policy Council members the impact of their advice Clearly define the mandate and reporting line of the Skills Policy Council in relation to other tripartite bodies Further improve horizontal and vertical co-operation between governance bodies Use full potential of social partners’ expertise in the Future Skills Needs Committee Refine the mandate of the Future Skills Needs Committee concerning targeting individuals
Portugal: Portugal’s National Agency for Qualification and Vocational Education and Training (ANQEP)	Promoting co-ordination across the whole of government Engaging stakeholders throughout the policy cycle Aligning and co-ordinating financing arrangements	Stabilise ANQEPs role in the Portuguese skills system Stabilise the financial basis of the Portuguese VET system Engage the General Board in a meaningful way Support the sectoral councils so that they can meet their duties
United States: Massachusetts’ Early Warning Indicator System (EWIS)	Promoting co-ordination across the whole of government Building integrated information systems	Establish a monitoring system to constantly evaluate EWIS Empower schools and districts to work with information management systems Improve the usability of EWIS

The recommendations identified in this report as being relevant to a broad range of contexts (and which are described in greater detail in the final chapter of the report) are summarised in Table 1.3.

Table 1.3. Overview of general policy recommendations
Governance challenge	Recommendations
Promoting co-ordination, co-operation and collaboration across the whole of government	Establish co-ordinating committees with a meaningful mandate and clear internal governance structures. Define a clear mandate Define mandate with real substance Establish strong internal governance structures
Engaging with stakeholders throughout the policy cycle	Promote the involvement and commitment of non-governmental stakeholders, while managing the risk of undue influence from special interests. Promote meaningful stakeholder involvement Limit the overall number of stakeholders involved Prevent gridlock and manage the risk of undue influence by special interests Allow stakeholder engagement to develop over time
Building integrated information systems	Support the establishment of information management systems, but ensure that they provide usable and relevant information to stakeholders and policy makers. Involve stakeholders in the design of integrated information systems Use information management systems to inform rather than automate governance decisions Make use of different kinds of data
Aligning and co-ordinating financing arrangements	Invest and commit the diversified resources needed to strengthen skills policies. Provide adequate resources by setting long-term budgetary goals Tap into multiple sources of funding while keeping equity concerns in mind

References

[23] Amable, B. (2003), The Diversity of Moden Capitalism, Oxford University Press, Oxford, New York.

[24] Anderson, K. and A. Hassel (2013), “Pathways of change in CMEs: Training regimes in Germany and the Netherlands”, in Wren, A. (ed.), The Political Economy of the Service Transition, Oxford University Press, Oxford, New York.

[34] Arreman, I. and A. Holm (2011), “Privatisation of public education? The emergence of independent upper secondary schools in Sweden”, Journal of Education Policy, Vol. 26/2, pp. 225-243.

[15] Becker, G. (1993), Human Capital: A Theoretical and Empirical Analysis with Special Reference to Education, University of Chicago Press, Chicago, London.

[29] Bohle, D. and B. Greskovits (2012), Capitalist Diversity on Europe’s Periphery, Cornell University Press, Ithaca, London.

[17] Busemeyer, M. (2015), Skills and Inequality: The Political Economy of Education and Training Reforms in Western Welfare States, Cambridge University Press, Cambridge, New York.

[25] Busemeyer, M. (2009), Wandel trotz Reformstau: Die Politik der beruflichen Bildung seit 1970 (Change despite a backlog of reforms: The policy of vocational training since 1970), Campus Verlag, Frankfurt/New York, http://www.mpi-fg-koeln.mpg.de/pu/mpifg_book/mpifg_bd_65.pdf (accessed on 14 January 2019).

[8] Busemeyer, M. et al. (2018), “Investing in education in Europe: Evidence from a new survey of public opinion”, Journal of European Social Policy, Vol. 28/1, pp. 34-54, https://doi.org/10.1177/0958928717700562.

[4] Busemeyer, M. and C. Trampusch (2012), The Political Economy of Collective Skill Formation, Oxford University Press, Oxford, https://global.oup.com/academic/product/the-political-economy-of-collective-skill-formation-9780199599431?cc=fr&lang=en&# (accessed on 14 January 2019).

[5] Christensen, T. and P. Laegreid (2007), “The Whole-of-Government Approach to Public Sector Reform”, Public Administration Review, Vol. 67/6, pp. 1059-1066.

[2] Esping-Andersen, G. (1990), The Three Worlds of Welfare Capitalism, Polity Press, Cambridge.

[28] Ferrera, M. (1996), “The ’Southern Model’ of Welfare in Social Europe”, Journal of European Social Policy, Vol. 6/1, pp. 17-37.

[20] Gerring, J. (2006), Case Study Research: Principles and Practices, Cambridge University Press, Cambridge, New York.

[32] Gingrich, J. (2011), Making Markets in the Welfare State: The Politics of Varying Market Reforms, Cambridge University Press, Cambridge, New York.

[3] Hall, P. and D. Soskice (2001), “An Introduction to Varieties of Capitalism”, in Hall, P. and D. Soskice (eds.), Varieties of Capitalism: The Institutional Foundations of Comparative Advantage, Oxford University Press, Oxford, New York.

[37] Hay, C., M. Lister and D. Marsh (2006), The State: Theories and Issues, Red Globe Press, London.

[18] Huber, E. and J. Stephens (2014), “Income inequality and redistribution in post-industrial democracies: Demographic, economic and political determinants”, Socio-Economic Review, Vol. 12/2, pp. 245-267.

[36] Hudson, C. and A. Lidström (2002), “National School Policy Changes in Britain and Sweden”, in Hudson, C. and A. Lidström (eds.), Local Education Policies: Comparing Sweden and Britain, Palgrave, Houndmills, Basingstoke, Hampshire, New York.

[31] Kattel, R. and I. Mergel (2018), “Estonia’s Digital Transformation: Mission Mystique and the Hiding Hand”, UCL Institute for Innovation and Public Purpose Working Paper, Vol. 2018-09.

[9] Lehmbruch, G. (1979), “Consociational Democracy, Class Conflict, and the New Corporatism”, in Schmitter, P. and G. Lehmbruch (eds.), Trends Toward Corporatist Intermediation, Sage, Beverly Hills, London.

[22] Lijphart, A. (1999), Patterns of Democracy, Yale University Press, New Haven, CT.

[35] Lundahl, L. (2002), “From Centralisation to Decentralisation: Governance of Education in Sweden”, European Educational Research Journal, Vol. 1/4, pp. 625-636, https://doi.org/10.2304/eerj.2002.1.4.2.

[27] Lundahl, L. et al. (2010), “Setting Things Right? Swedish Upper Secondary School Reform in a 40-Year Perspective”, European Journal of Education, Vol. 45/1, pp. 46-59.

[26] Martin, C. and D. Swank (2012), The Political Construction of Business Interests: Coordination, Growth and Equality, Cambridge University Press, Cambridge, New York.

[33] Mons, N. (2004), “Politiques de décentralisation en éducation: Diversité internationale, légitimations théoriques et justifications empiriques”, Revue française de pédagogie, Vol. 146, pp. 41-52.

[1] OECD (2019), OECD Skills Strategy 2019: Skills to Shape a Better Future, OECD Publishing, Paris, https://doi.org/10.1787/9789264313835-en.

[13] Offe, C. and H. Wiesenthal (1980), “Two Logics of Collective Action: Theoretical Notes on Social Class and Organizational Form”, Political Power and Social Theory, Vol. 1, pp. 67-115.

[12] Olson, M. (1965), The Logic of Collective Action: Public Goods and the Theory of Groups, Harvard University Press, Cambridge, MA.

[7] Scharpf, F. (1997), Games Real Actors Play: Actor-centered Institutionalism in Policy Research, Weastview, Boulder, CO.

[10] Schmitter, P. (1979), “Still the Century of Corporatism?”, in Schmitter, P. and G. Lehmbruch (eds.), Trends Toward Corporatist Intermediation, Sage, Beverly Hills, London.

[21] Seawright, J. and J. Gerring (2008), “Case Selection Techniques in Case Study Research: A Menu of Qualitative and Quantitative Options”, Political Research Quarterly, Vol. 61/2, pp. 294-308.

[19] Solga, H. (2014), “Education, economic inequality and the promises of the social investment state”, Socio-Economic Review, Vol. 12/2, pp. 269-297.

[16] Streeck, W. (1996), “Lean Production in the German Automobile Industry: A Test Case for Convergence Theory”, in Dore, R. and S. Berger (eds.), National Diversity and Global Capitalism, Cornell University Press, Ithaca, London.

[14] Streeck, W. and D. Mertens (2011), “Fiscal Austerity and Public Investment: Is the Possible the Enemy of the Necessary?”, MPIfG Discussion Paper, Vol. 11/12.

[11] Streeck, W. and P. Schmitter (1985), “Community, market, state – and associations? The prospective contribution of interest governance to social order”, in Streeck, W. and P. Schmitter (eds.), Private Interest Government: Beyond Market and State, Sage, London, Beverly Hills, New Delhi.

[30] Toots, A. and T. Lauri (2017), “Varieties of Social Investment Policies on Two Sides of the Baltic Sea: Explaining Routes to Endurance”, Social Policy & Administration, Vol. 51/4, pp. 550-576.

[6] UNESCO and ILO (2018), Taking a whole of government approach to skills development, UNESCO, ILO.

Note

← 1. Pork barrel politics refers to policy makers supporting a piece of legislation specifically intended to benefit a particular group/constituency in return for that group’s subsequent support (Hay, Lister and Marsh, 2006[37]).