Higher education policy in Canada and Germany: Assessing multi-level and multi-actor coordination bodies for policy-making in federal systems

Abstract Modern governments are increasingly faced with problems of policy coordination. However, coordination does not come naturally to organizations as it demands overcoming institutionalized working modes. Thus, countries have to find ways to tackle these problems to ensure efficient provision of public services. This contribution focuses on a specific and complex case, namely policy coordination for higher education policy in federal systems. The main research interest is to analyse the way in which coordination bodies responsible for higher education policy in two federal countries, Canada and Germany, organize their activities. Through this the study contributes to the understanding of the relevance of policy coordination in multi-level and multi-actor policy-making environments as well as the particular institutions that are dedicated to this task. Both coordination bodies are found to have many commonalities. However, the persisting differences, which can be traced to constitutional surroundings, also stress the importance of local conditions.


Introduction
Modern governments increasingly face problems of policy coordination, meaning 'the extent to which organizations attempt to ensure that their activities take into account those of other organizations' (Hall, Clark, Giordano, Johnson, & Roekel, 1977, p. 459). While they might be rational from an efficiency point of view, coordination activities do not come naturally to organizations as they demand to overcome institutionalized working modes. Even though coordination is among the oldest problems for public sectors, it is becoming highly relevant today as policy-making is increasingly complex, characterized by multi-issue, multi-actor and multi-level dynamics (Peters, 2015). Thus, countries have to find ways to tackle these problems to ensure an efficient provision of public services. This contribution focuses on a specific case of these activities, namely policy coordination in federal systems, which can be regarded as being especially complex due to the involvement of multiple actors and different levels of policy-making (Bakvis & Brown, 2010).
Federal systems allow for subsidiarity of decision-making and thus decentralize authority for (some) policy areas. Therefore, they face increased policy coordination complexity given the existence and relevance of additional actors linked to sub-national policy-making (Benz & Broschek, 2013;Hueglin, 2013). In response to these coordination problems, many systems have developed adaptive mechanisms, such as specific coordination bodies, that use different strategies including information exchange, consultation and policy negotiation between different sub-national units (Bolleyer, 2013;Peters, 2015). Thus, the key puzzle motivating this study is how coordination bodies function and what kind of mechanisms they use to enhance policy-making.
To investigate coordination challenges in federal countries this contribution will focus on higher education (HE) policy. Given the growing importance of HE policies for modern societies following the massification of HE and the emergence of knowledge economies (Altbach, Reisberg, & Rumbley, 2009;Maassen & Stensaker, 2011), the challenge to coordinate HE policy grew with its political relevance. As educational activities are often regarded to be sensitive local topics in federal countries, they are more likely to be within the competences of sub-national governments. Therefore, these federal countries face increased coordination challenges for HE policy, as they need to find a way to coordinate the different sub-national governments in an attempt to create some form of national harmonization. In this sense, HE policy can be seen as an exemplary case for other policy areas that are of national importance but within the competence of sub-national entities.
Based on these considerations, this contribution focuses on the way in which the coordination bodies responsible for HE policy in two federal countries, Canada and Germany, organize their activities that aim to harmonize the sub-national units' HE policy-making. Through this approach, the study aims to contribute to the understanding of the relevance of policy coordination in multi-level and multi-actor policy-making environments as well as the particular institutions that are dedicated to this task.
The analysis is based on conceptual approaches that focus on different ways to ensure policy coordination (Peters, 2015) as well as processes of voluntary coordination in federal systems (Bolleyer, 2013). The data for the analysis consist of academic literature, archival material, policy documents as well as a set of extensive semi-structured interviews with experts from the two countries. The following section will present the conceptual understanding of policy coordination. Afterwards the methods and data used for the analysis are introduced. This is followed by the presentation of the way in which the Canadian and the German coordination body in HE policy function as well as the mechanisms used to perform their tasks. The final section compares the working modes of the two coordination bodies and presents some overall findings.

Policy coordination
Problems of policy coordination are a common phenomenon in modern states. As governments today are inherently multi-organizational institutions and since public sectors evolve, the potential for incoherence in policies increases, or as Peters (2015, p. 1) formulated it: 'It seems that wherever there are governments, there will be conflicts between the specializations and programs and the coordination of all the programs in the policy area' .
Using a policy coordination lens has not only the advantage of putting this key challenge for modern policy-making in the centre of the analysis. It also enables the researcher to move away from a 'unitary' view on policy-making, in the sense that one actor (e.g. the government) decides about policies. Instead the coordination approach highlights the more complex interactions and negotiations between multiple loci of power (e.g. multiple ministries) within a national government. This is especially helpful in federal systems with their subsidiarity and regional decision-making authority, as their political structure makes it necessary to analyse the interactions of all governments on different levels of government. Thus, coordination of policy is a major challenge, especially in federal countries if they want to assure some form of nationally harmonized policy outcomes or public services (Bakvis & Brown, 2010;Macmahon, 1972). Federations have developed varying approaches to this challenge, partly relying on specific formal institutions or informal arrangements, whose level of institutionalization can vary from country to country (Bolleyer, 2006).
Among these different institutions, coordination bodies represent arenas that have been specifically created and tasked to provide a meeting place where different actors negotiate and coordinate policy (Bolleyer, 2013). Thus, they are the institution in which the policy, which is observed when 'only' looking at the nation as a whole and assuming a unitary policy-making actor, is being created. Therefore, it can be expected that the modus operandi of these venues influences policy coordination and in turn the outcome of policy-making processes in federal countries. Consequently, a focus on policy coordination helps to get a more nuanced understanding of the complex reality of policy-making in federations, especially in relation to policy sectors that are divided among the constituent entities.

Five dimensions of policy coordination
Policy coordination can be analysed along several dimensions. These dimensions do not necessarily correlate with one another and specific national arrangements can combine elements from several dimensions. However, they provide a helpful structure in thinking about different strategies of policy coordination. First, one can differentiate between horizontal and vertical coordination. While the former refers to coordinating activities between actors or institutions on the same governance level, the latter addresses coordination between different levels of government (Peters, 2015, p. 11f). Second, coordination can take place through some kind of force in hierarchical arrangements or voluntarily between equal entities (Bolleyer, 2013). While the former one is more common in vertical arrangements, the latter one applies more to horizontal ones. Third, one can further differentiate between positive and negative coordination. While negative coordination describes an agreement that the involved actors will not harm each other's activities, positive coordination is more encompassing and aims at creating coherence among policies (Peters, 2015, p. 9f;Scharpf, 1997). Thus, positive coordination can in general be described to be harder to reach than negative. Fourth, policy coordination can take place on different levels in the policy-making process, varying between the level of policy-makers or the level of administrators. While the former describes a top-down approach, which is based on the assumption that designing sufficiently well-coordinated policies will ensure the desired outcomes, the latter highlights a bottom-up perspective that stresses the necessity to include implementing agents (Peters, 2015, p. 13f;Sabatier, 1986). The final relevant differentiation addresses time and distinguishes between ad hoc short-term coordination and long-term institutionalized forms of coordination (Peters, 2015, p. 18f;Selznick, 1957). This dimension is especially relevant for coordination bodies, as the level of institutionalization of these bodies influences their ability to coordinate policy. In general, this means that the more these bodies are seen as institutionalized parts of the policy-making process, the better equipped they are to fulfil their tasks and the more their arena is respected by the different policy actors (Bolleyer, 2013;Meyer & Rowan, 1977). In this, key indicators for the level of institutionalization of these bodies are the existence of stable organizational structures and a staffed inter-governmental secretariat that operates continuously on policy coordination, as well as decision-making structures that go beyond unanimity decisions thus signalling that participating actors cede some part of their veto-power to majority agreements reached in the process of coordination (Bolleyer, 2013).

Potential barriers to coordination
Coordination is not only a multi-dimensional and thus complex activity, it also has to overcome certain challenges that are linked to the core problems hindering coordination in the first place. These barriers address both predispositions of actors and structural questions of the policy-making process.
The first of these barriers is ignorance of involved actors. Sharing information about policy-making activities is the most important prerequisite for coordination, as it is not possible for policy-makers to work together if they are not aware of one another's activities (Peters, 2015, p. 27f). This is the most fundamental challenge and applies to all dimensions of coordination. The second barrier is linked to the organization of the state following New Public Management reforms (Dahlström, Peters, & Pierre, 2011;Peters, 2015, p. 30f). These reforms stressed de-centralization, organizational autonomy as well as independence. This created little incentives, and in some cases even disincentives, for actors to work together and reduced the hierarchical power of central governments. Thus, especially vertical, hierarchical, top-down and also positive coordination are more challenged through this.
Another set of barriers are the tendencies of public sector organizations to protect their areas of influence and avoid policy risks (Hood, 2011;Peters, 2015, p. 31f). Here, the tendencies of bureaucrats and politicians to protect their influence and 'policy heartland' against intrusions from outside is the main force that might prevent coordination. Furthermore, 'sectoral ideology' , meaning the political beliefs of policy actors about what is considered as a good policy in their area of responsibility, can act as a barrier for coordination (Campbell, 1998;Peters, 2015, p. 37). These are bigger challenges in cases of horizontal, voluntary or positive coordination due to the lack of hierarchical power or the desire for more encompassing coordination results. Partisan politics can also present a barrier for policy coordination; parties with differing ideological backgrounds might favour contradicting policy solutions, and especially in situations of coalition governments or multi-level co-decision-making, partisan differences can prevent policy coordination (Döring, 1995;Peters, 2015, p. 35f).

Four strategies for coordination
To overcome these barriers, four main strategies can be used. However, not all strategies will be available in all policy-making environments or in relation to specific dimensions of coordination, and often one can find that countries use combinations of strategies (Peters, 2015). First, hierarchy is maybe the most familiar approach and based on vertical decision-making processes using rules and internal authority to ensure coordination (Peters, 2015, p. 46f;Scharpf, 1997). This strategy is most common in vertical, hierarchical, and top-down coordination activities. Additionally, due to the lack of negotiation hierarchy is also less suited to achieve positive coordination.
The second strategy relies on markets and a rational choice approach to coordination. This strategy, which is very much in line with New Public Management ideas, relies on the self-interest of actors in a quasi-market (Peters, 2014(Peters, , 2015. In this view competition, resource dependency and incentives lead to policy coordination. This strategy functions well in voluntary and horizontal coordination and it can be argued that due to its focus on competition it is not well-suited to achieve positive coordination as this demands more collaborative approaches. The third strategy relies on the use of networks to facilitate cooperation between policy actors (Marsh & Rhodes, 1992;Peters, 2015, p. 57f). In this, networks can be formal or informal structures that link actors, be it individuals or organizations, with the means to reach collective goals based on a communicative logic. Therefore, the network approach relies on the willingness of the involved actors to produce coordination. The fourth strategy is related to the network approach but is based more on the attitudes of the participants instead of their structural relationships. It highlights collaboration as a joint activity of multiple policy actors to increase the public value of their tasks through cooperation (Peters, 2015, p. 65f). Both the network and the collaboration approach are well-suited to achieve positive as well as negative coordination in both horizontal and vertical settings.

Coordination bodies as tools for coordination
These four abstract strategies can be combined with a wide variety of tools and institutions that help to overcome barriers to coordination (for an extensive discussion see: Peters, 2015, chapter 4). As the focus of this study is on one specific type of these tools, namely horizontal coordination bodies, we will restrict our discussion to their characteristics as well as how they offer possibilities to realize coordination strategies.
As outlined above, federal countries experience unique coordination challenges as they need to cope with the division of competences between the federal and the sub-national entities. This division leads to more complex policy-making structures as it is necessary to balance de-centralized competencies with the desire for coordinated national policies (Benz, 2013;Bolleyer, 2013). Since in these settings there is no entity that could use vertical, hierarchical influence to exert pressure, the sub-national entities have to agree among themselves how they want to arrange policy coordination. The instrument commonly used is some form of intergovernmental institution that serves as an arena of negotiation for the sub-national entities.
The focus of these coordination bodies is on horizontal coordination. However, many of them also include some form of exchange, be it formal or informal, with the federal level. While this exchange gives the possibility for a certain level of vertical coordination mainly through information exchange, this is not part of the core mandate of the body. Moreover, since the federal level only participates as a guest without voting rights in these contexts, the vertical dimension only plays a very minor role. 1 Therefore, these bodies work mainly on 1 in addition, some countries have additional bodies with the specific task of vertical coordination in which also the federal level participates as a full member.
the horizontal, voluntary self-coordination of the sub-national entities, and the success of it depends on the entities' willingness and capability to come to agreements (Bolleyer, 2013). As these bodies are mostly intergovernmental, they often rely on top-down coordination as the different governments usually participate in these bodies through members of their cabinets and thus the political leadership. Therefore, their ability to coordinate is also linked to their level of control over both the administration and public sector that delivers the services linked to the coordinated policies (e.g. universities).
Coordination bodies are also specialized organizations in the sense that their raison d' être is to facilitate policy coordination among the participants. Therefore, coordination through these bodies aims to be long-term as the creation of these institutions clearly signals them to be somewhat stable arenas.
Since the success of coordination activities in these intergovernmental institutions depends on the willingness of the sub-national governments to find agreements, the type of government creates differing incentive structures that influence coordination and which shape the environment in which coordination bodies work (Bolleyer, 2013). Powerconcentrating majoritarian governments have a tendency to maintain autonomy and flexibility and thus avoid binding agreements such as highly institutionalized inter-governmental institutions that facilitate coordination. These governments' tendency to rely on their autonomy makes them focus more on negative coordination and less prone to achieve the positive one. In addition, they are more used to hierarchical modes of coordination which are hard to realize in coordination bodies.
Power-sharing governments to the contrary are already used to negotiations among different parties forming the government and thus are more willing to also enter highly institutionalized coordination arenas even if that might entail a certain loss of autonomy. These governments are more likely to also agree to positive coordination and rely more often also on voluntary coordination due to them being used to negotiations and the need to find agreements with others.
To assess how coordination bodies perform their key task, we will examine in how far they cover the different dimensions of coordination, whether they address specific barriers to coordination, and finally which coordination strategies they use. We will study three clusters of attributes, which are summarized in Table 1 and which provide insights into the different dimensions of coordination.
In addition, we will assess the level of institutionalization of the two bodies. As argued above, the level of institutionalization influences their ability to coordinate policy and is therefore linked to their capacity to overcome barriers and successfully apply strategies for coordination. In this, key indicators for the level of institutionalization are linked to the attributes in Table 1 and include the existence of stable organizational structures, a secretariat as well as decision-making structures that go beyond unanimity decisions.

Methods, data and case selection
This analysis employs a comparative case study approach (Lijphart, 1971) to investigate policy coordination in federal countries. The two cases of federal systems selected are Canada and Germany, and the empirical focus will be on specific governance bodies that both countries created to coordinate HE policy. The two countries were chosen based on a logic following the method of difference. They are similar in the sense that they are large federal OECD democracies with established HE systems and sub-national entities that are responsible for HE policy. However, with regard to the structure of the government, which is expected to influence the level of institutionalization of voluntary coordination bodies, the two countries show significant differences, as Canada has power-concentrating governments while Germany has power-sharing governments (Bolleyer, 2013).
The cases are developed based on document descriptions of the policy coordination process in each country. The approach for each case is to describe, explain, compare and to analyse the function and mechanisms of the coordination body. The document analysis is supported by a number of extensive expert-interviews from both countries. A total of seven in-depth interviews have been conducted with policy actors in the two coordination bodies focusing on the working processes and coordination activities. To ensure consistency in the study's exploration and to allow for follow-up questions, a semi-structured interview approach was used.
The focus on federal countries and their coordination bodies stems from the interest to investigate the role of these bodies for policy coordination, especially in a policy area where there is no vertical coordination as the federal level lacks the necessary constitutional responsibilities, but where the countries still have to provide coordinated policies. This is especially relevant as governments have an interest in HE due to its pivotal importance for modern societies. Thus, a well-functioning HE system is politically desirable creating the need to coordinate the actions of different state governments in federal systems (Capano, 2015, p. 106).

Canadian policy context
Both constitutional arrangements and institutions of government shape HE policy processes in Canada. Within the two orders of government, Canada's sub-national legislatures have exclusive jurisdiction over education, a constitutional condition of confederation largely responding to concerns of religious and linguistic rights of francophones (Council of Ministers of Education, Canada [CMEC], 2003). Using Colino's (2013) typology, Canada's federalism is characterized as segmented federalism, given constitutional design, culturally diverse communities and intergovernmental institutions (Capano, 2015). In this highly decentralized federation, Canada lacks a federal department of education, a national body responsible for institutional accreditation and difficult federal-provincial relationships; HE is the one policy area within social policy where federal-provincial dialogue is 'virtually non-existent' (Meekison, 2013, p. 257). The federal government's main instrument has been to exercise its spending power (Meekison, 2013). However, in contrast to other federations, the bulk of fiscal transfers from the federal to provincial governments are unconditional, including transfers for HE (Bakvis, 2013, p. 209). The primary policy coordination mechanism emerging in this policy context is the CMEC.
The federal government does not have a formal relationship with nor is a member of CMEC, and, institutionally, CMEC's formal links with federal government are 'somewhat limited' (Meekison, 2013). While CMEC interacts operationally with a variety of federal departments, there is no formal representation to the federal government. At the programme level, CMEC administers national official languages programs and assists Canada in meeting its international treaty obligations, the latter providing a considerable impetus for coordination. There is a formal protocol established to coordinate communication and activities with the federal department of Foreign Affairs regarding participation in education related international conferences and meetings as well as administrative arrangements (CMEC, 1977). This relationship and the overall conditions of intergovernmental relations present obstacles to formal multilateral policy coordination, in spite of numerous calls from various stakeholders for improvement, including CMEC itself (CMEC, 1993(CMEC, , 2003.

Council of Ministers of Education, Canada
CMEC is a voluntary national body created in 1967 by the provincial education ministers, providing a single framework for inter-provincial and territorial cooperation in all levels of education. It is the oldest continuing interprovincial organization in Canada, and unique within the Canadian federal system; until 2003, it was the only pan-Canadian interprovincial and territorial structure in a policy field governed through a memorandum of agreement, with a Secretariat, a mandate and a budget (Meekison, 2013). CMEC's public aim is to 'promote discussion of issues of common interest as well as collaboration and consultation among provinces and territories, providing liaison with the federal government, and representing Canada internationally in the education field' (CMEC, 2001, p. 16). It supports 'open, non-obligatory non-restricting discussion' on education policy matters (Interview CMEC 2).

Legal mandate and structure
CMEC was created as 'an instrument of the provinces' (CMEC, 2001), through which they consult and cooperate with national education organizations and the federal government (CMEC, 1998). As a body of 'diplomacy' (Allison, 2007) supporting a type of 'treaty' (Interview CMEC 2), CMEC contributes to the 'exercise of the exclusive jurisdiction of provinces and territories over education' (CMEC n.d.), in response to major political concerns to guard against federal 'intrusion' into provincial affairs (Interview CMEC 3). This mandate was reaffirmed in the Joint Ministerial Declaration in 1999: We, the ministers responsible for education, unanimously reaffirm our responsibility for providing leadership in education at the pan-Canadian level through the CMEC. While the provinces and territories remain responsible for education in their jurisdictions, there continues to be a need for joint action. We believe that our collective will to work together will create a synergy that will benefit each province and territory. (CMEC, 2001, pp. 16, 17) This statement highlights that CMEC focuses on horizontal, voluntary coordination and operates top-down through the involvement of provincial ministers. At that time, CMEC adopted a number of priorities for collaboration and pan-Canadian joint action: focusing on education outcomes, sharing information on best practices, collaborating on curriculum initiatives, promoting policy-related research, strengthening the postsecondary sector and increasing access, supporting international activities, promoting mobility and enhancing CMEC as a forum for effective and fruitful cooperation with the federal government (CMEC, 2001, p. 17.) CMEC initiatives are divided into two categories: pan-Canadian strategy activities, in which all jurisdictions participate, and consortium activities, which provinces and territories choose to participate in and fund, according to their interests (CMEC, 1998).
CMEC's founding document established several different components that facilitate policy coordination (see Table 2).

Working processes and agenda setting
CMEC has a permanent secretariat, with an annual budget of $38 million CDN (2014CDN ( /2015, and approximately 50 employees. The Secretariat provides the significant administrative effort required to organize meetings and maintain a fully bilingual organization. While CMEC does have some policy-analytic capacity, policy is formally the purview of the

Body
Characteristics council • comprised of up to two Ministers from each of the provinces and territories, representing each provincial Ministry of K-12 education and He • is a highly transient body due to continuity problems with changing Ministers and different sub-national electoral cycles • Meets at least twice each year, and elects a chair and Vice-chair of the council biennially director general • appointed chief executive officer, an ex officio a member of each committee or task force of the council • Provides liaison among the Members and may, when requested, represent the council in negotiations with the federal government, liaise with educational and professional organizations and represent the council at international education events advisory committee • comprised of the deputy ministers of the members of the council and the director General, and other senior sub-national officials • Makes recommendations and reviews matters to be presented to the council, prepares the agenda for the council meetings, and provides oversight to the administrative matters of cMec committees, task forces and working groups • established by the council or advisory committee, and may include provisions for regional representation of officials and participation of non-Members, if deemed appropriate • Focus on three different types of activities: programs and projects undertaken or managed by the organization (such as the official languages programs, Programme for international Student assessment), programs and projects for which the organization plays a coordinating role for the federal-provincial interface (student financial aid, national statistics, international education), and policy topics or interest areas in which new policy work might be generated provincial elected officials, and CMEC does not have the mandate nor funding to perform as a policy institute. Available research is disseminated and information is exchanged through regular meetings. Network development is a valued component of CMEC, with on-going and informal information exchange between and among policy-makers and officials supported through CMEC 'social networks' (Interview CMEC 3). CMEC focuses on issues arising from members and is generally responsive to the sub-national First Ministers, who coordinate the overall agenda through another interprovincial body, the Council of the Federation. Hug (2011, p. 102) found that the country's largest four provinces tend to dominate agenda-setting and have closely coordinated policy positions including resisting federal involvement in HE and limiting policy coordination with the federal government: 'to this day … provincial governments have refused to give the federal government a permanent seat and to transform CMEC into an institution of intergovernmental cooperation ' (2011, p. 112). CMEC's decisions are made through reaching consensus, thus de facto the principle of unanimity.

Limitations
There are several specific Canadian challenges to coordinate within the federation, given the political dynamics arising from regional and constitutional concerns. First, some key HE policies involve many government departments in addition to education, reaching well beyond CMEC's scope for coordination. This multi-issue challenges results in a jurisdictional gap. For example, science and research policy provides federal funding to university researchers and is a major source of university revenue, and yet in many provinces and territories, science and research is not in the education portfolio and therefore those ministers are not part of CMEC. In these areas, policy coordination can only be achieved through multilateral voluntary cooperation outside the institutionalized realm of CMEC.
Second, the relationship between universities, their associations and the federal government is a challenge to policy coordination. In the absence of intergovernmental dialogue, there exists a 'surrogate or alternative process' (Meekison, 2013, p. 273) which provides opportunity for 'resolution through other venues' (Interview CMEC 5). The universities and their associations have developed strong and highly successful lobby relationships directly with the federal government, in part due to their 'somewhat jaundiced view of federalism … [universities] generally see themselves as national institutions serving national, if not universal, values and interests' (Cameron, 2002, p. 154f). This alternative policy process offers federal policy-makers an easier venue for influence. However, this policy success is gained without the benefit of multilateral policy coordination, and has occasionally given rise to a 'fractious relationship' between CMEC and these associations (Interview CMEC 5).
Further, provinces do have other venues to coordinate activities, which might undermine the centrality of CMEC to national policy coordination. Finally, CMEC may suffer from insufficient senior policy-maker attention, therefore reducing the effectiveness of the venue and the ability to have provinces rise above individual jurisdictional concerns in the pursuit of common objectives (Hug, 2011).

German policy context
Given the central role of the Länder for HE policy, Germany is often characterized to have de facto 16 independent sub-national HE systems (Capano, 2015, p. 118). Thus, while the German system in general can be described as part of the unitary federalism (Colino, 2013), in HE policy it shows signs of balanced federalism due to the high degree of autonomy of the states. Therefore, one can separate three political arenas: the states responsible for their own HE system, the cooperation between the states and the federal level, and the coordination among the states (Pasternack, 2011b, p. 17). The latter arena is institutionalized in the form of a coordination body, which in general is described as the third level of German federalism and where the federal government participates as a regular guest (Wolf, 2008, p. 22).
A key factor in favour of coordination is the constitutional call for the creation and assurance of comparable living conditions throughout the federation (Braband, 2005), which partly contradicts the constitutional autonomy of the Länder in education. Additionally, Germany is a rather homogenous society with an understanding of a unified (higher) education area (Braband, 2005;Erk, 2003). Recent changes in coordination of HE policy in Germany came through the re-organization of the federal system that started in 2003. The reform strengthened the responsibilities of the states in HE and weakened the federal level thus increasing the need for coordination between the Länder (Griessen & Braun, 2010).

The standing conference of the Ministers of Education and Cultural Affairs of the Länder of the Federal Republic of Germany
The coordination body that is in charge of HE policy is the 'Standing Conference of the Ministers of Education and Cultural Affairs of the Länder of the Federal Republic of Germany' (KMK). It was created in 1948 even before the federal republic, and has its permanent secretariat in Bonn as well as in Berlin (Erk, 2003). It can thus be seen as working towards long-term coordination and until today the KMK has created a tight net of more than 1000 single agreements between the Länder on matters linked to education, HE and culture (Wolf, 2008, p. 21). The work in the KMK is based on a legal agreement between the Länder as well as policy-making traditions (Pasternack, 2011a). This means that if one Land would decide to opt out of the agreement that created the KMK it would cease to exist. So while it focuses on voluntary and horizontal coordination, KMK's work has a certain level of compulsion, given the public sentiment towards a unified German HE system. At the same time, it helps to maintain a German intellectual unity, or as Erk (2003, p. 304) formulates it: 'the co-operation and co-ordination carried out through KMK has allowed a level of standardization in educational policy often associated with unitary political systems' .

Legal mandate and structure
The legal basis for the KMK is an interstate contract from 1948 that has been revised in 1991 (KMK, 1991). In this contract, the Länder agree to fund and create a secretariat for the KMK that is supposed to facilitate its activities. Thus, the work of the KMK is mainly based on cemented policy-making traditions (KMK, 2014). The main policies in HE addressed by the KMK include access and admission, formal structures of degree programs and the framework for quality assurance in HE.
Members of the KMK are all 16 Länder represented by their ministers for education, HE, research and culture, and the KMK has several bodies that facilitate policy coordination (Table 3).

Working processes and agenda setting
The working processes of the KMK are based on initiatives from the Länder, and the starting point is always 'a mutual need for information' (Interview KMK 1). 2 The proposal can be brought up at any of the working levels, and is usually referred to an ad hoc commission to prepare a statement that then moves through the structure. Decisions in the plenary of ministers and state secretaries were traditionally based on the principle of unanimity. In 2005, the KMK changed their working mode and limited unanimity decisions to areas that are necessary to create unity and mobility in the German area of education, decisions that have a fiscal effect or decisions that affect the KMK as an organization (Wolf, 2008). All other decisions are taken with a majority of 13 votes in favour. Overall, 2 the interviews at KMK have been conducted in German, all quotes are translated by one of the authors. The secretariat is a 'service provider' (Interview KMK 1) that facilitates information exchange between the Länder. However, it is 'not allowed to act on its own and bring up new topics' (Interview KMK 2). Furthermore, the discussions taking place on the different levels in the KMK facilitate direct exchange of information and peer-learning between the involved policy-makers and administrators.

Limitations
Limitations of the work of the KMK are linked to the complexity of interactions in HE policy as some policies are in need of two forms of horizontal coordination: between the different Länder and between different ministries in the Länder (Edler & Kuhlmann, 2008). Additionally, there are potential conflict lines among the Länder that can limit the influence of the KMK, which are based on: (1) partisan competition between state governments led by social democrats or Christian conservatives, (2) east vs. west German Länder, (3) city states vs. larger states and (4) rich vs. poor states (Pasternack, 2011a). Another factor limiting the influence of the KMK is the growing tendency among all Bundesländer to increase institutional autonomy of HE institutions. Thus, while the Länder gained competencies after the 2006 federal reforms, the parallel process of implementing New Public Management reforms in HE limits the potential for horizontal coordination as the Länder Governments have less direct influence on their universities (Griessen & Braun, 2010). Furthermore, the KMK 'lacks any ability to sanction Länder that ignore agreements' (Interview KMK 2), highlighting its reliance on voluntary coordination.
There are three main barriers to coordination that emerge from the set-up and working mode of the KMK. First, the principle of unanimity with regard to the most important decisions makes policy-making rather slow. Second, the formally non-binding character of agreements reached in the KMK and the fact that all agreements need to be transferred into law by the Länder parliaments is another potential barrier. While governments can usually rely on their parliamentary majority to pass laws, especially in situations where HE policy is politically salient and where there might be local demands that contradict the agreement, parliaments can decide to deviate from coordinated policies. Third, some decisions of the KMK that touch the responsibilities of other ministries need support from the respective ministers' conference. In cases where the negotiated agreement from the KMK is not supported by the other ministers, the conference of the heads of state of the Länder has to take the final decision.

Policy coordination in the two countries
Based on the above presentation of the two coordination bodies, it is now possible to compare them with regard to the level of institutionalization, the different dimensions of coordination they address, and barriers they help to overcome, as well as coordination strategies they use.

Institutionalization of policy coordination
Both CMEC and KMK are bodies that moved beyond weak levels of institutionalization as they both have their own dedicated secretariat and staff that works independent of the entities that they coordinate (Bolleyer, 2013). The staff provides CMEC and KMK with continuity in their work and offers expertise that can be used to moderate conflicts. While CMEC only reaches a level of medium institutionalization as they de facto still rely on the principle of unanimity in their decision-making, KMK reaches a higher level of institutionalization as they opened up for majority decisions on several of the coordinated issues.

Dimensions of and barriers to coordination
Positive vs. negative Both coordination bodies focus on negative coordination as they try to provide a forum for information exchange on issues that need coordination to prevent negative effects for the national HE sector. However, due to the constitutional task to assure comparable living standards the work of the KMK also has some characteristics of positive coordination, since the Länder actively pursue coordination in an attempt to ensure better national policies.

Top-down vs. bottom-up
CMEC and KMK focus on a top-down approach to policy coordination as the main decisions are taken by the ministers. However, both bodies also incorporate civil servants from sub-national units in their structures. Thus, they facilitate exchange both on the policy and administrative level.

Short-term vs. long-term
While CMEC and KMK are clear examples of long-term coordination, they also allow for ad hoc coordination through the possibility to create commissions. This organizational flexibility might be a factor that supports their stability over time.

Barriers to coordination
The activities of CMEC and KMK help both countries to overcome some of the barriers for coordination, while others seem to create persistent challenges. Both coordination bodies have information exchange as a central part of their mandate and enable policy debates between both politicians and administrators. Thus, they work actively against a lack of information.
At the same time, the level of autonomy of universities both in Germany and in Canada limits the scope of influence of the sub-national governments and therefore also the ability of the coordination bodies to do their work. While this is not a problem unique to federal countries it can be argued to have a stronger effect on horizontal coordination bodies, as it limits there room to manoeuvre due to the diminished direct influence of the participating governments.

Strategies for coordination
As neither KMK nor CMEC have the ability to enforce their decisions and both depend on the sub-national policy-making bodies to follow up on their agreements, they cannot use hierarchy as a strategy of coordination. Also a market based model of coordination does not properly describe the work of KMK and CMEC, since both highlight information exchange and the desire to learn from one another with regard to policy-making instead of focusing on competition. Given the stable organizational character of CMEC and KMK and their explicit mandate to facilitate coordination they focus on a network strategy for coordination, and especially KMK due to its openness towards positive coordination can be described to also employ a collaboration approach. All in all, both KMK and CMEC have surpassed weak levels institutionalization as they have secretariats with dedicated staff. However, KMK has a higher level of institutionalization as they allow for majority decisions on some of the coordinated issues (Bolleyer, 2013). Additionally, KMK and CMEC have a strong focus on information exchange and policy debates among sub-national states, and they include policy-makers as well as civil servants in their different structures, thus enabling coordination at multiple governance levels. Finally, both organizations have created a structure that allows tackling also new issues through the creation of ad hoc commissions giving them a certain level of flexibility in handling new developments in the policy area.

Conclusion
This study contributed to the understanding of the relevance of policy coordination in multi-level and multi-actor policy-making environments as well as the particular institutions that are dedicated to this task by analysing how the coordination bodies responsible for HE policy in Canada and Germany function. The analysis focused on three clusters of attributes: (1) mandate, structure and organization, (2) working processes and agenda setting and (3) limitations.
Both bodies analysed in this paper are stable forums for policy coordination, responding to the constitutional logics of their political systems and coordinating HE policy between the sub-national units in their respective countries.
However, CMEC and KMK also showed some differences, which can be explained as a reaction to their constitutional and federal environment. First, CMEC has a strong purpose to resist any form of federal participation in education policy, while in Germany the relationship between the two levels is more fluid and object of constitutional debates. Second, CMEC plays an important role in international policy processes. Given that the federal government does not have constitutional standing on education policy, CMEC represents Canada's sub-national jurisdictions when Canada is a participant in international venues. Finally, the difference in levels of institutionalization and intensity of coordination between CMEC and KMK is in line with the already described expectations in the literature concerning the influence of the structure of government (Bolleyer, 2013). Canada's power-concentrating governments rather avoid constrains by highly institutionalized coordination arrangements due to the governments' tendency to maintain autonomy making them focus more on negative coordination and less prone to achieve positive one. Contrary to that Germany's power-sharing governments are already used to negotiations and thus also more open to these kinds of arrangements. The results of the analysis show that both coordination bodies have many commonalities in their mechanisms, coordination strategies and also limitations they face. This hints towards the possibility that there is a set of typical characteristics of voluntary coordination bodies in federal systems. Following institutional theory, one could say that these bodies fulfil similar tasks in their contexts and thus rely on similar organizational configurations and behaviours to perform these tasks (Meyer & Rowan, 1977). However, the differences between CMEC and KMK, which can be traced to their constitutional environments, also stress the importance of local conditions that demand certain adoptions of organizational configurations and practices. In this sense, the procedures of coordination bodies as organizations have to be in line with the views of its constituencies (the members) and legitimate in the context of the political system, highlighting the embeddedness in the social and state context as a key factor (Meyer & Rowan, 1991).
As both coordination bodies represent policy-making arenas in which negotiation is necessary and where each participant could in theory opt out at any given point in time, the fact that both bodies manage to create policy coordination supports an institutional understanding of policy-making rather than rational choice arguments. It seems that an institutionalized policy culture that involves cooperation and coordination can create strong incentives towards coordinated policies even without hierarchical and coercive tools (see also : Bakvis, 2013;Bakvis & Brown, 2010).
The results of this comparison open several venues for further research. First, it would be relevant to broaden the empirical focus and assess how similar coordination bodies in other federations operate. Second, a detailed assessment of the parts of HE policy, which are within the mandate of different coordination bodies, would be a relevant study, as this would shed light on the question, which issues are regarded as nationally critical and thus need to be coordinated, and which issues might be too disruptive to be included in these processes. Finally, while this study presented a detailed assessment of the functioning of coordination bodies, the link between their activities and policy outcomes is so far unexplored and demands further attention. This also includes more detailed investigations into the role of the staff of these coordination bodies for policy-making and policy outcomes.

Disclosure statement
No potential conflict of interest was reported by the authors.