Introduction

While, in the last 50 years, Europe has occasionally been running, jumping, walking and stumbling in the direction of ever closer union, directionless moves have been made during large periods. Responsible politicians have engaged in extended sur place exercises and sometimes they have hit the ground. Outcries against further steps toward European integration have been heard. At the same time the kernel of European cooperation that started in the early 1950s among six countries at the Western side of the postwar East-West divide, has tremendously grown in its geographical extension and in other dimensions. In this paper we look at this process as a series of steps that contributed to the formation of a novel polity. (Galtung 1973 for an early effort at critical analysis; a broad historical-geographical narrative providing opportunities for historical comparison is Heffernan (1998).

We are in this instance primarily interested in the basic political mechanisms of the current process that result in: the variable geographical extension in the practice of a European scale, the authority shifts primarily across scales that bring a European polity into being and the emerging practices that translate the construction of a European scale level into a newly networked system of multilevel governance (see in the first place Marks 1993; Marks et al. 1996; Hooghe 1996; Hooghe and Marks 2003). In this paper we do not aim at an exhaustive review of the literature, nor do we present a rounded analysis of a clearly demarcated, freshly collected data-set. It is a conceptual paper providing an empirically grounded theoretical interpretation organized around some of the currently central notions in the discussion on European cooperation from a geographical perspective.

Viewing the institutionalization of a European scale from the perspective of the scale debate in geography (especially Swyngedouw 1997; Brenner 1998, 2001; Delaney and Leitner 1997; Cox 1998; Marston 2000; Paasi 2002; Herod and Wright 2002, Howitt 2003; Leitner 2004; Mamadouh et al 2004; Sheppard and McMaster 2004; Jessop 2005; Marston et al. 2005), it is key to acknowledge that the establishment of a European level by definition impacts the configuration of all existing scales in that part of the world. The European level is not only a new scale of government, it includes new relations between scale levels, a system often referred to as multi-level governance. The emergence of this new system of rule in Europe represents a shift from the modern or Westphalian state system to a state system plus something extra whereby interstate, suprastate and transnational cooperation affect and change traditional state authority in different ways. The general direction of this historical transformation is widely accepted and debated, but where are we now? What do we know about the functioning of the new multi-scalar configuration of governance known as the European Union?

This paper focuses on the European level of governance in this configuration and its relations to other scales. What could we mean when we refer to a ‘European level’ in the context of European governance? Most obviously there is always some geographical extension implied but Europe lacks a generally accepted delimitation.

To spot its European level, we have to look for aspects of governance that show up as European by way of impact range or representation. There may be some dispute as to the governance relevance of some actions compared to others. There can be hardly any doubt that the realm of rights and obligations based on authoritative decision-making forms the hard core of the European governance level while a lot of unsteadily framed activities are of more peripheral importance.

Different types of authoritative European decision-making are based on different geographical extensions. This fundamentally differs from a system of rule embodied in states. The European level brings to life a set of new authorities, and therefore also a set of new state transgressing networks, and a set of new frames of reference. Europeanization, the process that makes Europe relevant at the other scale levels, has several meanings. It has also different directional dynamics (up/download) depending on the actor pushing Europeanization in a specific situation: local or national actors can push the Europeanization of an issue—bottom up—if they think they will be more successful defending their interest or their ideas in the European arena. Likewise supranational actors can promote Europeanization from above to homogenize conditions for different players in the European arena across the different lower level scales (for example the harmonization policies that flanked the realization of the internal market “1992”). As Europeanization progresses, networking across scales is stimulated and the initially neat distinction between scale levels becomes fuzzy.

We distinguish three main dimensions of the construction of the European level and the Europeanization of governance: time, space (as jurisdiction) and sectors. These three dimensions are both obvious and poorly researched. The sectoral segmentation of governance is a striking feature of European governance as differences between policy sectors can be extremely large when it comes to the competencies of the European level or the role of the differently scaled actors in the decision making process. Less obvious and certainly worthy of much more empirical research are the interconnections between the sectors and their dynamics. This implies an interest in the sequence and timing of these processes.

Regarding the time dimension more generally, the obvious shifts across (many) sectors at some critical junctures need to be reconstructed: were they planned from above or the result of spontaneous aggregations? Did they involve directed changes or evolutionary transformations like spillovers from one sector to the other?

Finally the spatial dimension provokes an obvious demand for further analysis. The varying territorial domains of European governance add up to a confusing array of temporary territorialities that differs from the typical hard-edged integral territoriality that is so characteristic of the modern nation-state. We have moved from a series of territorially homogenous systems of rule to a still impermanent assemblage of territories, authorities and rights as drawn up by Sassen for a number of other cases (Sassen 2006).

This article aims at assessing the evolution of the kernel of EU governance by considering first the construction of a European level per se and then some aspects of the processes of Europeanization that connect scale levels in a new fashion It aims at offering a more subtle analysis of the emerging scalar configuration within Europe than those that decide too quickly for a sequence of successive stages of integral integration as a result of the transfer of competences to a European level. This refers e.g. to both sides in the intergovernmentalism/supranationalism debate where opinions only differ as regards levels and rates of overall integration. To achieve this aim the paper discloses the complexity of the new scalar arrangements by reviewing and comparing shifting arrangements in different policy sectors based on the important classificatory efforts of earlier scholars plus some indications for the transgression of traditional levels that has accompanied the construction of a European level per se.

The article first explores the successive and sometimes sectorally differentiated geographical domains of European cooperation, often jurisdictions, and emphasizes the different and always changing geographical extension that this round of European cooperation has continuously demonstrated. In the next section we look at the transfer of competencies to the EU level (the making of a jurisdiction) and the prescribed role of supranational actors in the ensuing processes of European decision- making. Building upon Börzel’s work on the width and the depth of integration as a meaningful way to interpret differences between policy sectors (Börzel 2005), we show how integration evolves and increases resulting in an ever more distinctive European level, that nonetheless remains of variable importance among policy sectors. In the following section we attempt to go beyond an assessment based on the outcomes of disputes between major contenders concerning competencies and the implied procedural roles at critical junctures. Building upon the work by Thomson et al. (2006) on recent instances of ordinary decision making in the EU, we discuss the apparent culture of decision-making that has emerged within the EU that goes far beyond the formal players and procedures prescribed in the ultimate stages of those decisions. To elaborate on this, we then turn to the administrative backstage of the political decision making process. Very few studies scrutinize the administrative wheels of the EU machinery. The collection of essays edited by Hofmann and Türk (2006a, b, c) is an example (see also Nugent 2001; Stevens 2001; Wakefield 2007). They show the importance and the actual practice of the administrations in the different stages of decision-making and they strongly emphasize the intertwining of the EU-administrations at the European level and those at originally other levels transforming the procedures at all these levels in a new integrated European administration. Finally we reflect on our results in terms of general processes of institutional development.

The seemingly obvious European level: geographical extensions

The European level is unlike most established scales—the national certainly, the local often—not taken for granted in social and political life. While its legitimacy is often openly disputed, its existence is often assumed as a distant but real and paradoxally vague entity. One reason for this lack of salience, though certainly not the only one, is the changing and variable extension of Europe in the context of European governance against the background of general, rarely well articulated, notions of essential Europeanness.

Demarcating the European level in geographical terms has always been difficult. Jacques Lévy has stated that still at the present time the European level can be determined in a number of ways between the European urban core and the globe. Thinking about European space provokes a permanent gymnastics to look for the proper units (Lévy 1997, p. 263). Even if we narrow the search for the European level to those forms of cooperation emanating from the modes of European cooperation that go under the aegis of the current European Union and its immediate predecessors, there is a wide variation in possible demarcations. This refers not merely to the successive geographical extensions of this type of European cooperation from the initial six countries of 1951–1957 to the extensions of 1973, 1981, 1986, 1995, 2004, 2007 (Map 1). It also refers to the various demarcations of vital ingredients of that cooperation which also implies a heterogeneous basis of the ruling authority/ies (Mamadouh 2001). Map 2 draws some of the current differences. They have also changed over time. There have been recent changes in the adherents of the Schengen agreement (that even covers non-member states) concerning border traffic of persons and in the countries that are part of the Eurozone.

Map 1
figure 1

The varying geographical dimension of the European level: The continuing extension from 6 member states in 1952 to 27 in 2007. Note: The territory of the former German Democratic Republic became part of the EC12 in 1990 with German reunification

Full size image
Map 2
figure 2

The currently varying geographical dimensions of the European level: Coexisting extensions of EU27, Eurozone and Schengenland. Note: Most EU member states are part of the Eurozone and Schengenland (including older and newer ones such as Slovenia and Malta), Cyprus too, but the implementation of Schengen is not yet completed; Ireland is part of the Eurozone but not Schengenland. Another large group of EU member states are part of Schengen but do not yet use the Euro (all joined in 1995 or 2004); the two most recent member states have not yet implemented Schengen and do not yet use the Euro; and the UK does not participate in any of them. Finally two outsiders are part of Schengen and two are about to implement it (Switzerland and Liechtenstein)

Full size image

So the European level is a jurisdiction of variable dimension. As a political actor in the wider world the EU is also variable, as it can speak in the name of a variable collection of member states (for 27 in WTO, in environmental negotiations and as a donor in development programmes but for a smaller though in fact hazy number in military matters, and for 15 in monetary matters on the basis of a jointly accepted euro). The EU is represented on these occasions by different spokes-persons, that are not always clearly coordinated in a formal fashion (the President of the Commission, the President of the European Council, the High Representative for the CFSP, the President of the European Central Bank, the Commissioner for Trade). The EU level is a playground for diverse players of which only some have a specifically European territorial range, such as the treaty-based European institutions. The other players include territorially institutionalised actors as states and political and administrative regions (including cross-border regions and states external to the EU); nationally regionally or locally organised interest groups, political parties and NGOs; transnational ones; as well as non-territorially organised actors such as corporations and epistemic communities. The EU also comes to life as an arena of variable dimension or perhaps as a set of arenas. The players have their encounters in various places epitomized by the establishment of the major EU institutions in three different cities (Brussels, Luxemburg and Strasbourg). But there are many more arenas where EU encounters take shape, e.g. at places where agencies settle and engage with other parties and at temporary meeting places e.g. of the European Council.

All in all, the territory under the jurisdiction of the European level has dramatically increased from the early 1950s till the present, expanding from a core area of six relatively similar Western European states that in this way overcame their recent murderous adversities to a common organization of almost all the states of the continent, with all the others (Belarus excepted) involved in some kind of agreement with the European Union. The repeatedly shifting shape of this Europe is at least slightly confusing. The simultaneous coexistence of diverging territorial shapes for different policies is even more difficult to communicate and to legitimate, for its fuzzy geography and as it results in simultaneous different configurations of players. For example, the UK is a major player in EU decision-making, it is one of the two key players of military Europe but it is at the same time self excluded from decisions that pertain to the Eurozone. Not only does the European level consist of a bewildering array of players in different configurations, they also congregate in a whole series of arenas. This also does not contribute to the salience and finally the authority of the European level. While the European Union is often portrayed and perceived as an unproblematic if somewhat fuzzy entity, it is in fact appropriate to see it as an intangible social unit that largely lacks permanence and struggles with a number of simultaneously different geographical delimitations. The union is still to a considerable extent a project. As a union, it is in the circles that support it an imagined fact of life.

The authority shifts that contributed to European level formation

In the course of time several arrangements for decision-making have been developed at the European level. In each case the European level is the arena, the place from where effects of actors’ practices reach all parts of a perceived European territory, a place also that supposedly embodies the representation of Europe. The current set of arrangements is as follows:

  • The so-called community method (with the EU as jurisdiction, the EU as arena, and EU institutions as the Commission, the Council, the Parliament, and the Court of Justice as key actors);

  • The intergovernmental procedure (with the EU with slight exceptions of inclusion and exclusion as jurisdiction, and also as arena and the member states as key actors);

  • The open method of coordination (OMC) (with the member states as jurisdictions, the EU as arena, and the member states as key actors) (see Caporaso and Wittenbrinck 2006, p. 474; Borrás and Jacobson 2004; Duina and Raunio 2007).

The community method defines the original core of the European level. It assumes the major EU institutions and prescribes their action spaces through the consultation and later the co-decision procedures that are distinguished by the increasing legislative role of the European Parliament. In the intergovernmental procedure a modern version of traditional congress diplomacy should result in new agreements for the EU area. Their modern elements are the involvement of the Commission as an extremely well prepared and expert neutral party. The open method of coordination engages member states in processes of mutual assessment and learning after they have commonly agreed on a far reaching goal (like collectively becoming the most productive economy in the world). In the intergovernmental procedure and the open method of coordination the European organs and particularly the Commission at its centre try to perform as coach and honest broker.

Europe’s system of multilevel governance with its newly developed European level has taken its current shape over the last half century and it still evolves. At every point in time it functions as a result of repeatedly renewed basic treaty negotiations in a series of Intergovernmental Conferences. These have resulted in the slow, but steady expansion and refinement of the community method across the different policy areas of European cooperation. This implies ever adapted official roles, a new attribution of competences and authorities and revised rules of the game.

The various stages of the construction of the European level of the European governance system have been conveniently summarized by Börzel (2005). She describes the changes in the different policy sectors as regards the distribution of competences among the different levels as changing levels of integration. On the basis of an earlier table produced by Hix (2005) eighteen policy sectors were characterized in terms of the width and depth of integration at different moments in time. Width refers to the proportion of competencies in a policy sector to be treated at European Community/Union level, depth refers to the involvement of supranational bodies (commission, parliament, court) and the voting rules in the council as spelled out in the treaties. All policy sectors are scored on a 0–5 scale on these dimensions of integration following the information in the respective treaties on the following dates: 1958 (Rome), 1987 (SEA), 1993 (Maastricht), 1999 (Amsterdam), 2003 (Nice) and some undisclosed date in the not too distant future when the Lisbon Treaty enters into force. In fact the calculations have been based on the Constitutional Treaty text of 2004. As far as we are aware the Lisbon Treaty is not different in the aspects that are relevant for these calculations.

Figure 1 shows the evolving overall formal integration (the relative concentration of authority at the European level) along these two dimensions by summing all scores per dimension and converting to a scale 0–100. Apart from the obvious difficulties in scoring all these policy sectors in a comparative fashion, two additional problems should be stressed. The demarcation of policy sectors changes over time and we cannot be quite sure that the same policy field is covered consistently. In addition, it may well be the case that not all potential policy sectors have thus been covered (e.g. much of what goes under the label of spatial planning and housing in the Netherlands cannot be properly positioned). In other words, a score of 100 does not necessarily mean complete integration across the entire range of governmental competences. But as one moves through the list, the impression is nonetheless that a lot of ground has been covered.

Fig. 1
figure 3

Width and depth of European integration among 18 policy sectors: Percentage of maximum possible overall score. Source: Calculation based on Börzel’s assessment of 19 policy sectors (Börzel 2005, pp. 222–223)

Full size image

From Fig. 1 one can draw a number of conclusions. In both dimensions, integration increases over time. The process may have slowed down on several occasions; so far, it has never gone in the opposite direction. At the time of the Rome Treaty nearly a third of the entire integration process in terms of width had been covered in the initial step with the depth trailing behind. In other words, relatively speaking, the institutional provisions that are necessary to deal with the tasks have only been established in a piecemeal fashion. This is largely made good at the time of the Single European Act. From that time until the Maastricht Treaty the integration process moves forward at an accelerated rate along these two dimensions simultaneously. During this period the creation of the necessary institutional provisions in fact takes already a slight advance over the extension of the policy field covered. This difference becomes more pronounced at the time of the Nice Treaty and even slightly more in the Constitutional/Lisbon Treaty. After Maastricht the increase in integration in terms of the extension of the terrain covered, has been small. But one should keep in mind that it was even smaller during the long period of time from the Rome Treaty to the Single European Act. In the meantime European integration has (in a much larger Union) progressed significantly. The emphasis has been on deepening the integration in fields that had already been brought under the aegis of the European level. For the policy issues covered, the process has deepened to more than three quarters of the whole range, a fourfold increase over the level at the initial stage.

In fact the overall increase in integration in two dimensions consists of 18 times two processes in the different policy sectors. What about the similarities and differences among these sectors? Do we have early innovators that remain in an advanced degree of integration over time, mainstream sectors that follow suit and laggards staying behind which would suggest emulation and spillovers in variations of a diffusion process, in short some kind of overall system; or is this a less tidy set of processes, each policy sector driven by its own dynamics? In order to reduce the variation in the data all scores were put in three categories: high (over 3,5), middle (over 1,5 until 3,5) and low (0–1,5). Positions at the time of the Rome Treaty were then compared to those that resulted from the abortive Constitutional treaty of 2004, that is as far as we know in these respects equal to the results of the upcoming Lisbon Treaty. The shift that we portray is a process of half a century. As there turn out to be no policy sectors in which integration on either of these two dimensions decreases over time, we get six different trajectories for width and six for depth (LL, LM, LH, MM, MH, HH). In Tables 1 and 2 we show the positions of the 18 different policy areas for the two dimensions of integration at the start and at the end of our period.

Table 1 Levels of integration in width dimension for 18 policy sectors in 1958 and in 200y
Full size table
Table 2 Levels of integration in depth dimension for 18 policy areas in 1958 and 200y
Full size table

The slightly lower initial level of the depth of integration compared to the width and the quicker pace of the increase of depth does show up again in these figures. A shift of one class upwards over this time period is more or less the ‘normal’ pace of integration during this period in the width dimension. It applies to 11 out of 18 cases (10 moved one category upward, while one, that started already in the highest category could not rise higher and can therefore also be considered as ‘normal’). However, in the depth dimension only six cases move one category upward. By far the largest single class moves two classes up from a low position at the time of the Rome Treaty. Integration in four policy sectors in the width dimension stagnates, and three sectors show a steep growth rate. Integration in two sectors in the depth dimension stagnates and 10 show a steep growth rate.

Apparently the presence of an overall trend suggests a systemic process of integration. There is not a single one among these 36 cases where integration over this half century has diminished. At the same time initial differences among policy areas are by no means all maintained over time. Agriculture has been the permanent frontrunner. It was the only policy sector that had a high score on the width dimension right from 1958 and the depth of integration moved from a middle score in 1958 that it then only shared with three other policy sectors in the domain of Economic Affairs to a high score half a century later. Currently four more policy sectors have reached a high score in the width and depth dimension simultaneously. Together with Agriculture they are now the frontrunners of integration. They are Environment and Consumer Protection, Occupational Health and Safety Standards, Territorial, Economic and Social Cohesion and Monetary Policy. They originate from different initiatives in the history of European cooperation and have apparently acquired this frontrunner position for different reasons.

Territorial, Economic and Social Cohesion is the policy area most relevant to spatial planning and housing which is an issue area of central concern in other papers in this collection (notably those by Gore, Johnson and Dukes). It is now among the few most integrated policy areas. However, sustained efforts to bring spatial planning up to the European level (Faludi and Waterhout 2002) have so far only met with moderate success. At the same time within the field of housing there is in the Netherlands a dispute about the position of housing associations vis à vis European rules of unfettered competition. In short, integration levels and the demarcation of policy areas should be looked at carefully for what they imply.

At the other extreme one policy area has remained a laggard in both dimensions: Tax The tax systems as basic mainstays of the state’s existence are still largely out of bounds for authorities transgressing the state level in the EU. It is interesting that this does not apply to other existential attributes of the state’s existence that have often been put under the label ‘high politics’. This distinction again introduced by Börzel (2005) to account for a major part of the differences among the policy sectors turns out to be less relevant. They have been addressed under the second and the third pillar in the Maastricht Treaty but do not show up as a specific class of policy areas in this analysis in terms of the integration trajectories covered. In other words, the basic logic of the treaties in terms of the three pillars does not translate in actual classes of integration trajectories based on the more detailed contents of these same treaties. As it turns out, the dynamics of each policy sector contribute to a considerable extent to the course of their integration process despite the systemic characteristics of the overall process.

The European level and the processes of Europeanization

As European cooperation incrementally developed over the years, every new institutional compromise prescribed a renewed general plan for the day-to-day decision-making of the European institutions following the community method in terms of rules and procedures. At the same time the repertory of decision-making mechanisms widened. The number of intergovernmentalist instances of decision-making increased under the second and third pillar and these were thus also prescribed under treaty rules. Finally, the Open Method of Coordination became a new form of interaction involving the Economic Affairs and Finance (ECOFIN) and Employment, Social Policy, Health and Consumer Affairs Councils (EPSCO) plus the Commission and its services (Neuhold and Radulova 2006, pp. 60–65).

A recent study (Thomson et al. 2006, chapters by Achen 2006, pp. 264–298, and by Schneider et al. 2006, pp. 299–316 in particular) provides some important insights concerning the process of decision-making in ordinary non-constitutional cases in the EU following the community method during 1999–2001, the period after the Amsterdam Treaty had been agreed but before Nice and the Constitutional Treaty episodes. The EU had 15 member states at the time. Consequently, this research refers to a stage in the evolution of the European institutions when the different institutions and their mutual relations had considerably matured, but before the major extension of 2004 had occurred.

The study considered 66 Commission proposals that in a reasonably short time ended in a final decision by the Council or the Council and the European Parliament (in case they had to be treated under the co-decision rules). It is a pity that we do not know from this study why some proposals come to a final decision and others do not. What the typical blockages were at the time when the system was not yet put under renewed strain by the massive entry of new member states in 2004, remains hidden. The proposals that were studied, were distributed across all the relevant policy sectors. The book consists of the construction and testing of decision-making models, all variants of the rational choice institutional approach: actors pursue their preferences as best they can within the constraints of the institutional environment in which they find themselves. What were the outcomes, which models performed better than others in terms of proper forecasts of the final decisions, how did the EU (if it followed the community method) decide?

Models that only take into account the procedures to be followed and consider the decision as a vote where all participants try to make maximum gains in every individual decision, do badly. This is not how the EU following the community method works. In fact decisions reflect each member’s preferences, and those of more powerful members often somewhat more. This apparently results from cooperative bargaining. Irrespective of the voting rules in the particular case (qualified majority voting for instance) there is a strong orientation to end up with a unanimous vote by a decision in which every participant finds enough positive points to support it. Exchanges among parties within the context of the initial proposal, smaller scale coalitions and exchanges of advantages across proposals increase the support for proposals. In addition, numerous negotiations before and during the decision-making by various factors may smoothen its path. Non-cooperative styles of bargaining where threats are used are unhelpful in accounting for actual EU-decision-making. Finally, it is interesting that there are not only no differences in this regard between the areas where there are procedural differences, but also no differences between decisions in the different issue areas. The EU-level preference for maximal support in cases where the community method applies, is indeed an EU-level finding.

In the years around 2000 cooperative bargaining kept the flow of new EU rules running. This pattern was probably set during an extended previous period in which the oft repeated games that had to result in new legislation had to be played again and again. This resulted in informal common norms and expectations in each case, but also in the possibilities of informal exchanges during successive games. The earlier rounds of new accessions that brought ever new participants to the table, had apparently been successfully accomodated in this respect. This must at least to some extent have resulted from the relatively stable administrative background that was part of the policy process. These administrations had long developed working procedures and effectively integrated newcomers despite growing diversity reflecting the increasingly varied political cultures from which the participants emanated.

Vital administrative institutions like the top functionaries in the directorates-general, the cabinets in the Commissioner’ s offices and the legal services division of the Commission plus the Secretary General and his staff of the Council plus the COREPER gatherings of permanent representatives of the member states in Brussels that prepared the decisions of the Council in its different compositions provided the necessary continuous capacity for wheeling and dealing that enabled formal decisions. As the European Parliament has more emphatically entered the decision-making loop with the extension of the co-decision procedure the Commission has engaged in informal ‘trialogues’ with representatives of the Parliament and the Council to find common ground for its policies at an early stage before an official proposal is put forward.

The results of the study suggest that the most important requirement in preserving the capacity to act after the accession of so many new entrants was perhaps less the procedural changes that got all the limelight in Nice and the Constitution but perhaps more the ability to play these cooperative games that seem to be the essence of successful decision processes in the EU. This is not merely a question of providing rules and procedures that allow for these outcomes, but just as much a cultural question about ordering preferences and the normative preference for certain courses of action over others. In addition, some very practical assets have always been needed. The so-called ‘diplomatic sherpas’ (not only professional diplomats but also administrators with diplomatic skills) are a vital ingredient of all decision-making in this heterogeneous and ever widening arena where European players intermingle with actors with a more limited geographical range and with actors whose ranges transgress European boundaries, to maintain and strengthen the European level jurisdiction(s). If these pre-conditions suffice under ever enlarged actor numbers remains to be seen.

Hofmann and Türk (2006a), in concentrating on the evolution of the EU administrative apparatus and the national level administrators involved in European business over time, provide some further stepping stones for assessing the strength of the EU level and its resilience to withstand new strains as member state numbers increase. In their collection of essays the ‘sherpas’ get profiled as the widening body of administrators at the European and the national level that plays an essential part in keeping European cooperation working on a day to day basis. They call it an integrated administration, be it of a novel kind. They do not pay any attention to the lower administrative levels within member states where European orientations have gradually also become more prominent (Dukes 2007, Chaps. 6 and 7 in particular).

The size of the administrative manpower at the European level is still modest, not larger than the civil service of a sizeable city within the Union. It started from truly humble beginnings. A comparison of numbers of employees in the different European institutions in the period 1970–2003 according to data from the Financial Report 2003 of the EU is shown in Table 3.

Table 3 Staff of the community institutions and bodies: comparison 1970–2003
Full size table

These numbers have to be used with caution. They are probably in fact somewhat larger, but the trend is unmistakable. With the increase of the competences at the European level, the supporting manpower has considerably grown. The most impressive increases have been with the institutions that not so much set the agenda, make policy and implement it, but with those primarily charged with control, assessment and evaluation: the Court of Justice and the Auditors on the one side and the European Parliament on the other. Generally at a more gradual growth path the support of the Council shows a steeper rate of increase than that of the Commission. The Economic and Social Committee and the Committee of the Regions launched midway in this period (both advisory) have seen their administrative support grow as that of the Council.

The limited size of the administrations at the European level have long been explained as the consequence of the doctrine of executive federalism that supposedly guided the development of the administrative system of European cooperation. Following this doctrine policy issues that were deemed to be European in nature, should be developed at the European level and result in legally binding rules that were then to be implemented by the administrations of the member states. Consequently the European level itself would not or hardly dispose of capacities for implementation. This doctrine backed up by a powerful early arrest of the European Court of Justice (the Meroni-doctrine, 1958) has survived in law but not in actual practice.

In fact, national administrations and those at the European level have become increasingly intertwined and are constantly engaged in joint deliberations that aim to produce supranational government in the context of a growing body of accepted rules and practices within the EU. This results in an integrated administration (resulting from these continuous deliberations) across levels (European and national) that remains fragmented (between the levels but also between policy sectors and between the different parts of the policy process) and heterarchic (no single hierarchy, various sources of authority). This integrated administration has an evolutionary nature (early stages condition the next ones, practices in adjacent policy sectors may be transferred or spill over, change is perpetual). The intertwining of the European and national levels of administration results in an ongoing process of Europeanization at the lower level, necessarily also passed on to those parts that are not directly involved in relations with the European level. At the same time, the continuous interaction of the European level with the lower levels results in its turn in a steady process of adaptation of the European character of the European level (Hofmann and Türk 2006c, pp. 573–596).

Policies within any polity supposedly all pass through a policy cycle consisting of a stage of agenda setting, of decision-making and of implementation. In the case of European cooperation the contribution of administrators is vitally important in all the stages of the policy cycle. And in all these stages the European level and the national level are in a process of ever more intimate interaction.

At the agenda-setting stage the Commission is still a predominant player. It is supported by the directorates-general that form the backbone of its services. But long before the drafting of proposals has started, issues have been discussed in expert groups that the Commissioners install for the express purpose to come up with new ideas. And provisional suggestions have been circulated among lobbyists who make their annotations. Lobbyists obviously may also come forward with policy ideas that the Commissioners may then take into account and pass on for further consideration. The expert groups are to a considerable extent populated by national representatives with or without a mandate. Directorate General Research is particularly strong in mobilizing expert groups. Nobody knows how many expert groups are in existence at a given moment and the numbers obviously fluctuate. Estimates of current numbers are 800–1300 (Larsson and Trondal 2006, p. 19). They have grown with the increase of the Union’s competences, but it would be interesting to know their rate of increase over time. Unfortunately we do not.

What happens during the period of decision-making (that may well take years, but may also be achieved in matter of months) depends very much on the role of the different institutions in the different procedural paths that can be taken. In all procedures under the community method the Council is of the utmost importance in decision-making. There is a permanent EU supporting staff, there are the civil servants of the member states that are part of the Brussels offices of the Permanent Representatives and there are the members of the national bureaucracies that serve part time at the approximately 250 working parties to prepare the decisions of the two COREPER bodies (Larsson and Trondal 2006, p. 26; Neuhold and Radulova 2006, p. 50). They are the portals of the Council under the community method or act as precursors for the intergovernmental forms of decision-making under the second and the third pillar. Under the Open Method of Coordination four differently structured preparatory bodies are at work. Interestingly, their membership does not merely cover nationals permanently based in Brussels and delegates from national bureaucracies but also civil servants who are part of the Commission services. In fact, not only are the differences between national and supranational parts of the administration completely blurred, this also applies to their origins within the treaty-based institutions that supposedly play completely different and separate roles in the making of policy (Neuhold and Radulova 2006, pp. 60–65). In the growing number of cases where Parliament is involved in decision-making its largely increased support staff helps Parliamentary Commissions where much of the actual work occurs and it helps rapporteurs in drafting reports and proposals. These administrators are also active in the conciliation procedures where the views of the Council and the Parliament have to be brought in line with the Commission proposals. To smoothen that process the Commission has more recently engaged in informal trialogues at the beginning of the decision-making process in order to clarify the difficulties of the decision-making process for all the partners from the outset.

Despite the reluctance and opposition to the European level becoming engaged in implementation, this has nonetheless happened in a number of fields. As European rules have increased, so has their implementation at a European wide scale. But again, institutional features at the European level have become narrowly intertwined with actors originating at the national level and still permanently engaged there as well (Hofmann and Türk 2006b, pp. 74–112, particularly p. 75). This process has started with the emergence of the Common Agricultural Policy early in the life of European cooperation. After decisions had been taken, there was a need for detailed implementation measures e.g. for every individual product. This was done by committees of national experts who jointly worked for the Commission. That is how the nebula of comitology came about that then spread to other policy areas. One prominent example were the financial services where the Lamfalussy report created the basis for a similar methodology. Over an extended period of time an increasing number of agencies has additionally been formed to implement EU policies in various areas (a useful overview is Versluis 2007, pp. 191–204). The management and board of these agencies is composed of representatives of member states and of the commission. In actual practice national services narrowly cooperate with agency personnel to provide the required implementation. Finally there is also the occasional delegation of implementation to private agencies.

In all these different forms of intertwining there are different combinations of national, supranational and functional orientations of the people concerned (Trondal 2006, pp. 391–416). They all play their role in the deliberations that finally result in the different stages of the European policy cycle. It clearly does not mean that all policy-making is completely turned into an exercise of Europeanized policy-making, nor is the European perspective completely submerged in the professional and national loyalties that are at stake. It is an ever differently coloured European polity that slowly takes shape by unexpected turns and in sometimes surprising directions. At the same time the traditional role play between the institutions at the European level is transformed by additional networks, restructuring the European level in the process.

Conclusion

From a geographical perspective the European Union is by no means a clearcut entity. In a number of important policy sectors some member states are not involved. Non-participants vary between sectors. In some policy sectors non-member states take full part in the proceedings. Apart from the repeated increase in the number of member states, there is therefore a continuously different number of participants in various important policy sectors and consequently no salient demarcation of what the union is supposed to be at any moment in time. Obviously we know who the member states are. They have signed the basic treaties but this legal formula does only very partially cover the daily reality of the EU as a meaningful social unit to be called a Union. In the literature the unproblematic existence of the EU as a clear-cut entity is too much taken for granted.

The political integration process formally establishing a European level in the system of governance within the European state system has nonetheless moved forward for more than a half century incorporating this increasing collection of member states. The process that knew a fairly slow start in the 1950s and later repeatedly stalled for considerable periods, knew a quickening pace during the late 1980s and early 1990s. Integration levels have always differed considerably between sectors. With some exceptions due to relative growth spurts of individual sectors, rank orders in integration levels of sectors have been maintained, agriculture at the forefront from the beginning and taxation permanently trailing behind. Current integration levels do not completely coincide with positions in the first, second or third pillar according to the treaties. In no policy sector has a reverse process of des-integration ever occurred. The integration process (the transfer of competences to the European level) is apparently partly based on system wide dynamics, but additionally each sector is subjected to specific conditions that cause additional dynamics that are specific for that sector. In addition, these may cause spill-overs in adjacent policy sectors.

European decision-making in the confines of the first pillar involving European level actors and others is based on repeated games with well established procedures that result in pretty effective cooperation through continuous cooperative bargaining. There is a vital role for ‘diplomatic sherpas’, originating from circles far beyond the traditional diplomatic corps, that shuttle between decision-making arenas at different levels. In the process they form trans-scalar networks that bridge the original discrete scale levels and provide new often informal arenas for decision-making. Very often such networks are policy sector specific. On the other hand the initially clear distinctions between the operative rules of decision-making in the various pillars tend to get blurred over time. It remains to be seen to what extent intersector coordination either at the level of national states or at the European level will be effective or even feasible in the future.