EU Committee Governance and the Emerging Community Administration Jarle Trondal Centre for European Studies Agder University College Servicebox 422 Phone: +47 38 14 15 61 Fax: +47 38 14 10 29 Email: jarle.trondal@hia.no Web-site: http://www.hia.no/oksam/ces/ Forthcoming in Herwig Hofmann and Alexander Türk Introduction This chapter argues that an important fabric of EU administrative governance is the emerging multilevel community administration. This community administration integrates essential parts of national government institutions and important segments of the EU institutions. The study demonstrates that the domestic components of the community administration include the lower middle levels of the member-states' ministries and agencies/directorates. At the EU- level the community administration covers the lower echelons of the Commission services, the Commission agencies, the Commission expert committees, the Council working parties, the comitology committees, and finally the national civil servants seconded to the Commission for shorter periods of time (Trondal 2004). Together, this community administration meshes and integrates executive institutions at two levels of government charged with responsibilities for policy initiation, policy decisions and policy implementation (Heywood and Wright 1997). This study highlights the development of a community administration within the three classes of EU committees referred to above. These committees integrate national civil servants into EU decision-making processes. They pose, however, different challenges to the decision- making behaviour, roles and loyalties among the officials attending these committee meetings. The survey data presented reveals that the institutional autonomy of the community administration is stronger within the European Commission than within the Council and the comitology setting. Hence, the picture of one unified community administration has to be sacrificed for the model of a multifaceted community administration balancing intergovernmental, functional and supranational dynamics. The data thus seriously challenges sweeping generalisations of administrative fusion and bureaucratic engrenage (Wessels 1998). This chapter also demonstrates that deliberative dynamics are not omnipotent within the comitology committees as asserted by Joerges and Neyer (1997). EU committees represent adequate laboratories for studying what happens when contrasting decision-making dynamics meet because such committees embody civil servants from different layers of governance. Arguably, EU committees embody primarily three decision making dynamics: 1) defending nation-state preferences (intergovernmentalism), 2) providing neutral expertise (functionalism), and 3) defending the "common European good" (supranationalism). The study poses the following question: to what extent do domestic civil servants evoke intergovernmental, supranational or functional role perceptions and decision- making behaviour when attending EU committees? This study demonstrates how these three decision-making dynamics are balanced within the community administration by analysing the role orientations and decision-making behaviour of domestic EU committee participants. This question pertains to the problematic nature of the transformative power of international organisations in general and in EU institutions in particular (Barnett and Finnemore 1999). EU committees are important laboratories for studying transformational change that transcends the territorial dynamics of Westphalia because they have codified and institutionalised the bringing together of national and community officials in day-to-day decision-making. Assuming that functional and supranational dynamics transcend the existing territorial nation-state order, intergovernmental dynamics are treated as our null hypothesis. The study covers the following EU committees: The Commission expert committees (ECs), preparing decisions for the Commission, the Council working parties (CWPs), preparing decisions for the Council of Ministers, and the so-called comitology (CCs) committees, assisting the Council in controlling delegated powers to the European Commission. Decision-making within EU committees is about initiating and drafting optimal and effective policy proposals and aggregating interests towards stabile equilibrium, creating preferences, meaning, identities and roles, and transforming these into integrated regulative, cognitive, normative and symbolic arrangements (Barnett and Finnemore 1999; March and Olsen 1994: 265). Committees are 'generic features of modern political life', important venues for regulative decision making in the EU and important arenas where national and supranational decision-makers meet, interact, persuade, argue, bargain, adapt, learn and re-socialise (Christiansen and Kirchner 2000: 12). Hence, decision-making within EU committees also pertains to the less acknowledged aspects of actor socialisation and re-socialisation, identity change and role-play (Checkel 2004). The identities and roles activated by individual EU committee decision-makers are moulded institutionally through processes of matching perceptions of appropriateness to particular situations and through systematic and routinised allocation of attention. Departing from a 'sociological' and a 'cognitive' institutionalist perspective, this study argues for middle-range institutionalist theorising on EU committee governance. The endeavour is to unravel important institutional conditions under which national decision-makers evoke particular roles and decision-making behaviour when attending EU committees. Studies of European integration and domestic change have increasingly transcended the sui generis distinction between intergovernmentalism and neo-functionalism that dominated earlier theorising of European integration. Several bridge-building exercises between intergovernmental and neo-functional approaches have been suggested during recent years, notably, deliberative perspectives (e.g. Eriksen and Fossum 2000), social constructivist accounts (e.g. Checkel 2004; Risse 2002), institutional and organisation theory approaches (Egeberg 2004; Olsen 2003), and multilevel governance and network approaches (e.g. Kohler-Koch 2003). This list is not intended to be exhaustive (see Jupille, Caporaso and Checkel 2003). This chapter suggests that the conceptual divide between intergovernmentalism and neo-functionalism may be surmounted and bridged by outlining middle-range institutionalist approaches that explicate measurable hypotheses. This bridge- building exercise is done by specifying the institutional conditions under which different decision-making dynamics prevail within EU committees, be they intergovernmental, functional and/or supranational dynamics. The study outlines two institutionalist arguments on institutional 'mattering', one cognitive and one "sociological". By combining these institutionalist accounts, a multilevel institutionalist approach is outlined emphasising the causal impact of domestic and EU institutions. The study buys into the multilevel governance approach as a descriptive device but not an explanatory toolkit. Accordingly, the role perceptions and decision-making behaviour evoked by EU committee participants is explained by reference to their multiple institutional embeddedness. Multiply embedded civil servants are Janus-faced and are constantly challenged by competing expectations and perceptions regarding what roles to play and what decision-making behaviour to emphasise in different situations (Egeberg 1999; Trondal 2000). To substantiate this claim, the study specifies the conditional validity of six institutional mechanisms. Assuming that a mix of different institutions moulds civil servants' roles and behaviour, six ceteris paribus clauses are outlined that explicate the causal arrows and the expected empirical implications. The highlighted hypotheses are suggestive, although not exhaustive. The chapter proceeds in the following three main steps. The next section outlines a middle-range institutionalist approach and outlines six operational hypotheses on how domestic and EU institutions 'matter'. The next section introduces the data and methodology. We then proceed to an empirical illustration of the hypotheses. The empirical analysis benefits from an extensive survey that includes national civil servants from 14 member-states with experiences from EU committees. A Middle-range Institutionalist Approach The roles and identities evoked by individuals are causally salient because they provide cognitive, normative and ethical priors and generalised prescriptions of how to act in different situations. Roles and identities evoked by incumbents in public administration are foremost the product of primary and secondary institutional socialisation processes (Egeberg 2004). This study focuses on roles and decision-making behaviour as institutionally constructed at the crossroads of national civil services and the EU committees. To account for how domestic and EU institutions frame role and behaviour the study employs a 'sociological' and a 'cognitive' institutionalist approach. Whereas the sociological perspective underscores roles and decision-making behaviour as products of the logic of appropriateness, the cognitive perspective explains officials' role perceptions and behavioural pattern by referring to their bounded rational choices (March and Olsen 1995; Simon 1957). The 'pure' rational choice institutionalist account is left out of the analysis basically because it neglects the constitutive role that both domestic and EU institutions have on actors' roles and behaviour and reduce it to a question of optimalisation of expected utility. The cognitive argument advocates that organisations serve as buffers to information and thereby simplify the relevant decision-making stimuli exposed to incumbents. The bounded and local rationality of decision-makers is subsequently aggregated systematically into organisational rationality by this buffer function. Consequently, the systematic selection of information and relevant premises for decision making and role enactment mould how organizational members think, feel and act (Simon 1957). According to this cognitive perspective, decision-making dynamics are conditioned by the formal organisation of public administration, both domestically and at the EU level. The 'sociological' argument stresses that institutions are more than formal organisations and that the roles and behavioural patterns unfolding within institutions are 'driven by rules of appropriate or exemplary behavior, organized into institutions' (March and Olsen 2004:2). Whereas a cognitive perspective focuses on the allocation of scarce attention among bounded rational actors, the sociological perspective focuses on how actors match identities and roles to particular situations, and that this matching process takes time, is inefficient and path- dependent (March and Olsen 1995). 'The central logic is that of matching one's repertoire of identities and roles to specific (institutional) situations' (Trondal 2001: 15). In our context, EU committee participants may be seen as maintaining a repertoire of national, functional and supranational roles and identities, 'each providing rules of appropriate behaviour in situations for which they are relevant' (March and Olsen 2004:4). Whereas a cognitive perspective emphasises the historical efficiency of organisations and the causal importance of formal organisations to understand incumbents' roles and decision- making behaviour (roles and behaviour adapt efficiently to organisational contingencies), the sociological approach highlights the historical inefficiency of institutions (roles and behaviour adapt slowly and imperfectly to institutional contingencies). Moreover, whereas the cognitive perspective departs from a bounded rational perspective on human computational capacities, the sociological approach views actors as socialised humans that have normatively internalised 'patterns of behaviour and role expectations which characterize the groups in which they interact' (Alderson 2001: 416). Normatively internalised behaviour and roles are taken for granted and thus go beyond mere behavioural compliance. Accordingly, the length and intensity of interaction within groups and the length and intensity of exposure towards decision-making situations affect processes of socialisation of in-group behaviour and roles. 'Identities can be seen as arising from a process of socialization into socially defined relationships and roles' (March 1994: 62). Accordingly, '?t?he longer one is exposed to particular stimuli; the more one is likely to absorb these influences' (Hooghe 2001: 15). This is particularly relevant for explaining supranational roles (see H6 below). According to both the cognitive and the sociological institutional approaches, organisational and institutional borders, respectively, are causally relevant because they create relevance criteria for particular roles and behavioural patterns (Barnett and Finnemore 1999). From this premise it follows that civil servants, having several organisational and institutional affiliations, are exposed to multiple buffers to attention, logics of appropriateness, classification schemes, temporal rhythms, physical structures and symbolic arrangements (Biddle 1986:73). Moreover, single organisations may provide more or less consistent and conflicting challenges to the organisational members and more or less ambiguous cues for action. However, most of the time organisational dynamics are triggered when organisational borders are challenged by internal and external actors criss-crossing them. Civil servants who are multiply embedded tend to cross organisational borders fairly frequently and evoke multiple cognitive priors and perceptions of appropriate behaviour. Relevant to ask is what happens 'when the state [and its civil servants] is embedded in more than one institution and each institution demands a different role and set of behavioral actions' (Barnett 1993: 273)? After all, "[d]emocratic governance involves balancing the enduring tensions between different logics of action…" (March and Olsen 2004:17). The emerging multilevel community administration analysed here activates several partly colliding governance dynamics. Arguably, inter-organisational activities reduce the perceptional barriers to create new roles and behaviour because the actors are exposed to new experiences and information (March and Olsen 2004:13). The empirical data presented below reveals that functional roles and behavioural patterns are activated among officials attending ECs. By contrast, officials attending the CWPs and CCs evoke the role as a national representative more strongly. This difference between EC, CWP and CC participants are due to the organisational components underneath the Commission and the Council of Ministers (see H5 below). National officials attending EU committee participants, however, have their primary institutional affiliations back home. Hence, the domestic ministry and agency to which they are employed is likely to mould their roles and behaviour more strongly than the EU committees. The concept of multiple roles goes beyond an either-or style of representation, as seen in the 'free agency versus mandate' emphasised by intergovernmentalists and neo-functionalists (Eulau et al. 1959: 746). It is commonly assumed that civil servants evoke multiple roles (Biddle 1986). However, multiple roles may be differently related; they may be hierarchically nested, crosscutting and/or meshed and blended (Wendt 1994: 385). I argue that roles often are crosscutting on different dimensions or cleavages (Radaelli 1999: 34). For example, civil servants in domestic sector ministries are accustomed to the role of an independent neutral expert. Diplomats at the Foreign Office, however, are used to combining the sector expert role along with the role as a national representative. These roles are relevant on a sectoral – territorial cleavage but irrelevant on another cleavage: the national – supranational cleavage. Accordingly, civil servants evoking a supranational role, perceiving oneself as a community official, may at the same time evoke a territorial role, perceiving oneself as a national representative, because these roles are cross-cutting. Moreover, particular roles may be more or less important to civil servants in particular situations. A role that is perceived as marginal to a civil servant is not likely to outweigh a role deemed salient. Hence, civil servants tend to live with role conflicts because they attach different weight to them, and because some of them crosscut each other. Finally, roles may be evoked sequentially in different institutional contexts. Hence, role conflicts may be coped with by sequential attention. The next two sub-sections outline six hypotheses that each specifies conditions under which intergovernmental, functional and supranational role perceptions and decision-making behaviour are evoked by national officials attending EU committees. Hypotheses on domestic institutional 'mattering' Domestic civil servants are primarily affiliated to their national government institution, portfolio and role. These affiliations tend to mould decision-making behaviour and role perceptions because '?t?he routines of attention allocation tend to give priority to those things that are immediate, specific, operational and doable' (March and Olsen 1979: 50). For example, domestic institutions are likely to influence domestic civil servants' roles and behaviour more strongly than EU institutions. The following hypotheses are based on the sociological and cognitive perspectives on organizations and draws on fairly generic principles in organisational life. H1 Bureaucrats used to specialised vertical organizational structures are more likely to adopt sectoral and supranational roles and behaviour than those used to vertically non-specialised organisational structures Generally, integrated and coherent polities are more robust against external demands and penetration than fragmented and disintegrated polities. H1 postulates that vertically specialised structures are conducive to the emergence of sectoral and supranational roles and identities among civil servant (Egeberg 2001). Sectoral and functional roles may reflect the fact that vertically organised government institutions buffer and filter political signals and demands downward in the hierarchy. Civil servants may therefore act as independent and neutral sectoral experts within EU committees because they have the leeway to do so. Moreover, the lack of hierarchical control and steering within vertically specialised government structures enhances the leeway to act independently and evoke new supranational allegiances (Beyers and Trondal 2004). Civil servants from lower echelons in the domestic hierarchy are accustomed to professional autonomy and the appropriateness of role flexibility and freedom of manoeuvre, and are thereby inclined to adopt new supranational roles when entering the hallways of the EU committees. H2 Bureaucrats from sectoral ministries are more likely to adopt sectoral and supranational roles and behaviour than diplomats from the Foreign Office and the Permanent Representation in Brussels Sectoral organisational structures are conducive to the emergence of sectoral roles and behavioural patterns because civil servants are exposed to sectoralised information, considerations, criteria for success, cues for action, etcetera. Moreover, lack of co-ordination from the Foreign Office may accompany the enactment of supranational roles among EU committee participants. 'Not being reminded of their "national missions" on a daily basis in Brussels, the actors can easily lose sight of the nation-state as their primary locus of loyalty' (Trondal 2002: 473). Bureaucrats attached to the Foreign Office in the capitals and diplomats at the Permanent Representations in Brussels are more used to think in territorial terms as they represent their country (territory) and not a specific policy sector. The Council infrastructure is largely compatible with the territorial principle of organisation these officials are used to (See H4). As these officials become primarily involved in Council decision-making during the later stages of COREPER-deliberations, they are - compared to bureaucrats at the lower echelons of working groups - likely to consider issues in terms of 'national interests'. H3 The greater the degrees of organisational and institutional misfit between domestic institutions and EU committees, the more likely that officials change role and behavioural patterns H3 argues that organisational and institutional misfit across levels of governance accompanies a perceived adaptational pressure among civil servants towards role shift and behavioural adjustments (Egeberg 2004). Arguably, '?t?he impact of institutional compatibility reflects the compatibility of cognitive shortcuts and scripts' (Trondal 2000: 316) and similar logics of appropriateness (March and Olsen 1995). Organisational and institutional misfits accompany role conflicts, notably between national, functional and supranational roles. "The concurrent appearance of two or more incompatible expectations for the behaviour of a person" is likely to challenge existing roles and established behavioural practices among civil servants because they are exposed to new sets of information and new situations that leave existing roles and identities less relevant (Biddle 1986:82). Accordingly, organisational and institutional misfit is conducive to supranationalism among domestic civil servants attending EU committees. This argument may explain the contra-intuitive observation made below that some EU committees tend to reinforce intergovernmentalism. This is the case among national officials coming from the Foreign Office attending the CWPs and the CCs. Hypotheses on EU institutional 'mattering' H4 Officials attending the ECs are more likely to evoke sectoral roles and decision-making behaviour than civil servants participating in the CWPs and the CCs r Parallel to domestic sector ministries, the Commission ECs are primarily organised according to a sector principle, as opposed to the CWPs and the CCs, which are primarily organised according to a territorial principle, mirroring the domestic Foreign Office (see H2) (Egeberg and Trondal 1999). Accordingly, we assume that government officials attending the ECs evoke a sectoral role perception and decision-making behaviour. On the contrary, participation in the CWPs and CCs is primarily conducive to the emergence of a national role perception. Moreover, all government institutions, including EU committees, are organised according to several principles, and thus pose additional challenges to civil servants to evoke several partially conflicting roles. However, we expect EC participants to mainly evoke an expert role, and CWP and CC participants to mainly emphasise a national role (Egeberg and Trondal 1999). H5 The longer domestic officials have participated in EU committees, the more likely that they evoke a supranational role According to the sociological institutionalist perspective institutional 'mattering' is subject to time lags. When domestic civil servants first attend EU committees they are not likely to become instantly re-socialised. They are more likely to evoke roles that have recently been evoked (March 1994: 70). As civil servants interact with officials of other nationalities and with Commission officials over time they are likely to become slowly re-socialised into community-minded supranational agents. Accordingly, the longer civil servants have participated in EU committees, the more likely that they have internalised a supranational role. Research on small groups indicates that the development of in-group consensus is associated with the duration of interaction among the same actors, and subsequently of the age of the group (Biddle 1986:77). H6 The more intense degrees of participation and interaction within EU committees, the more likely that officials take on a supranational role Whereas H5 emphasises the causal effect of time, H6 highlights the causal weight of intensity of interaction to the understanding of supranationalism. Intensive interaction in our context refers to the number of EU committees attended, the number of formal and informal sessions joined, the degrees of active involvement during discussions, joining work lunches, etcetera. Contact theory and small group theory emphasise the causal importance of face-to-face interaction and contact repetition to understand role and identity change (Hart et al. 1997; Pettigrew 1998). According to Checkel (1999: 549), '?s?ocial learning is more likely where a group meets repeatedly, and there is a high density of interaction among participants'. Similarly, March (1999: 29) argues that, ?t?he interactive character of decision making extends over time so that the development of beliefs, rules, and expectations in one organization is intertwined with their development in others'. Data and Method Since 1995 the European Institute of Public Administration (EIPA) in Maastricht has organised seminars for member-state officials on the role of committees in the EU political process. In the spring of 1997 we started to distribute a questionnaire to those participants in the seminars who had been involved in one or several committees at the EU level. The questionnaire was designed to give an overview of the experience of member-state officials in EU committees: In what kind and how many committees they were involved, how frequently meetings were taking place, how long they lasted, what languages were used, how committee meetings were co-ordinated, etcetera. The major part of the questionnaire focussed on the question of how member-state officials viewed the roles they performed in these committees, how they perceived the roles performed by other participants and how well they were co- ordinated and prepared before meetings. During the first day of the seminar, those participants who had been involved in EU committees were asked to complete the questionnaire. By distributing the questionnaires at the first day of the Seminar, we minimised potential influences or "noise" from the seminar as such. Participation in the seminars in Maastricht was very unevenly distributed between different member-states. There were very few participants from the Southern member-states, but regular participation from central European member-states, the U.K. and Ireland. In addition to the seminars in Maastricht, EIPA organised a number of "Comitology seminars" in the member-states, particularly those that had joined the EU during the last wave of enlargement in 1995. Unquestionably, this led to a very unbalanced sample towards the new member-states. In order to correct this, an effort was made in early 1999 to contact the permanent representation of all the member-states from which we had a very small number of respondents (N), asking them to help to increase the number of completed questionnaires from these member-states. This effort was very successful in the case of Belgium and Spain, but did not result in many additional completed questionnaires from the other member-states. The composition of the sample, by member-state, is summarised in Table 1. The Table also shows the type of ministry the respondents came from, differentiating between the foreign ministry, other ministries, agencies and the member-state's permanent representation in Brussels. Table 1 about here This sample cannot claim to be representative neither with respect to the member-states included, nor with respect to the type of committees which member-state officials participate in. From the total sample, 132 respondents participated in ECs, 134 participated in CWPs and 76 in CCs. Not unexpectedly, 61 respondents participated in at least two types of committees and 31 in all three types. Moreover, the officials studied here are mostly employed within ministries other than the Foreign Ministry and in medium or lower rank positions. Moreover, our data (not presented in Table 1) show that EC participants are mostly recruited from sectoral ministries and agencies and less from Permanent Representations. CWP participants, in contrast, are recruited to a larger extent from Permanent Representations and sectoral ministries. Like in all written questionnaires, there was a considerable number of missing items - respondents who did not complete all of the questions, even if – as was the case in our questionnaire – for most of the questions multiple choice answers were provided for. For this reason the N will vary between tables in the following sections. The Emerging Community Administration Despite Commission efforts to reduce the number of ECs and simplify the CC procedures, the sheer number of committees and the total complexity of the EU committee system seem to gradually increase over time (Larsson 2003:15; Schaefer 2002). For member-state officials, participation in EU committees means consumption of scarce resources like time, time that will not be available for national concerns. Table 2 shows that time spent on EU matters varies with the place in the hierarchy of a respondent. Table 2 about here As could be expected, the major burden of committee work is carried by head of sections, senior advisers and advisers, the middle and lower middle level of member-states' administrations. Nearly two-thirds of the respondents belong to this group. These observations are empirically supported by the studies of Egeberg (1999), Trondal (2001) and Trondal and Veggeland (2003). Surprising is the relatively large proportion (20 percent) who come from the Director General or Deputy Director General level. This can possibly be explained by the fact that it is common practice that, on important issues, the top level of member-states' administrations will attend committee meetings in Brussels, often accompanied by lower level officials. It may also be taken as an indicator of the importance assigned by member-states' administrations to EU matters. The fact that more than 60 percent of this top-level group spends almost a day or more of their weekly working time on EU matters supports this conclusion. Moreover, CWP participants report that they seldom attend committee meetings alone. Most of the time officials go together with colleagues from their own ministry or from the Permanent Representations. Hence, the community administration includes large proportions of the domestic administrative fabric, notably the lower middle level of experts civil servants. Member-state officials' roles and identities. Civil servants often evoke multiple preferences, interests, roles and identities due to their multiple institutional embeddedness. Civil servants are multiple selves with several non- hierarchical interests and allegiances (Elster 1986; Fouilleux, Smith and Maillard 2002; Risse 2002). The evocation of one particular interest or identity does not necessarily trump another. By attending different institutions at different levels of governance officials learn to wear Janus-faces and to live with diversity and partially conflicting interests and loyalties (Lewis 1998). Hayes-Renshaw and Wallace (1997) picture a "continuous tension between the home affiliation and the pull of the collective forum". However, particular roles, identities and modes of decision-making behaviour tend to be evoked in some situations more than in others (March and Olsen 1995; Simon 1957). National officials attending EU committees spend most of their time and energy in national administrations . Accordingly, we expect their dominant institutional allegiances and identifications to be national when entering EU committees. However, "membership" in EU committees imposes additional obligations on officials, although for most of a secondary character. They are exposed to new agendas and actors, and are expected to look for common solutions (Egeberg 1999). According to Christiansen and Kirchner (2000), "committees permit national officials to familiarise themselves with the nature of the EU's administrative system". However, officials participating in CWPs and in CCs may be expected to behave more like government representatives than officials attending Commission ECs. The main reason for this is the basically territorial principle of organisation underlying both Council and comitology groups. In the Commission ECs, on the other hand, participants are expected to behave more like independent experts. Thus, professional allegiances and sectoral role conceptions are likely to be enacted fairly strongly among the latter (H4). Table 3 shows that national officials who attend different EU committees express more allegiance towards their own national government institutions than towards the EU committees on which they participate (H1 and H2). Thus as expected, EU-level loyalties seem to be secondary to national allegiances. However, some officials feel considerable responsibility towards EU level entities, particularly the CWP participants (Lewis 1998). This is mostly due to the high degree of intensity of day-to-day participation within the CWP meetings (H6). Hence, a certain kind of 'system allegiance' seems to be stronger among CWP officials than among EC and CC participants. Intergovernmentalism and EU-level loyalty thus do not seem to conflict but to complement each other (Risse 2002) (H3). Moreover, the vast majority of the committee participants have positive attitudes towards European integration generally and within their "own" policy/issue area particularly. However, relatively few officials change attitudes in this regard due to committee participation (Egeberg, Schaefer and Trondal 2003: 25). Table 3 about here Also as expected, those in CWPs tend to assign more weight to their relationship to their own government than those attending the Commission ECs, although the difference is not very big. A remarkably large proportion of CWP participants identify themselves with their own sector administration, policy arena or professional background. This pattern is probably due to the high degree of functional specialisation that accompanies participation in the basically intergovernmentally arranged Council structure. Hence, national officials attending EU committees evoke a complex role repertoire indeed (H3). The respondents were further asked to indicate how they perceived the roles of their fellow colleagues within EU committees. Table 4 about here Table 4 reveals that civil servants who attend CWPs and CCs tend to consider other colleagues mainly as government representatives (Fouilleux, Smith and Maillard 2002; Schaefer 2002). Commission EC participants, on the other hand, tend to perceive other colleagues as having more mixed roles (H4). Hence, only a minority (45 percent) find that their counterparts behave mainly as government representatives. Thus, although role conceptions are highly multi-faceted across types of committee (cf. Table 3), actual behaviour seems to mirror more clearly the prevalent organisational features of the various arenas (H4). Next, the respondents were asked to asses how much consideration they put on proposals, statements and arguments from different actors and institutions when attending EU committees. Table 5 about here First, almost no major differences can be observed between officials attending different EU committees as far as the above considerations are concerned. Second, as to the relative priority given to the proposals, statements and arguments of other actors, one consideration seems to be more important than others: Officials attending EU committees pay most attention to what their colleagues and experts from their own country have to say. This observation underscores the tendency already indicated in Tables 3 and 4 on the primacy of national allegiances among EU committee participants (H1). Participants, however, also emphasise the points of view of colleagues from other member-states who have demonstrated considerably expertise on the subject matter at hand. Officials give considerably less attention to arguments from colleagues from large member-states as such, and colleagues from member-states within their own region. In support of the deliberative supranationalist account (e.g. Joerges and Neyer 1997), the quality of the argument presented by other committee participants is considered more important than the sheer size and geopolitical location of the member-states they represent. Moreover, the EU Commission is also considered more important than large member-states and member-states within their own region. Finally, interest groups and firms are deemed considerably less important than colleagues from other member-states. By comparison, however, interest groups and firms from their own country are considered much more important than EU level interest groups and firms. This observation underscores the general tendency apparent in Table 5, namely that national officials attending EU committees pay more heed to national institutions than to supranational ones (H1). Hence, the community administration has a strong intergovernmental dynamic, particularly within the CWPs and the CCs. In sum, what we see is that arguing, not only bargaining, is a salient feature of the emerging community administration (Lewis 1998). Hence, the intergovernmental perspective, picturing national actors entering EU arenas with predetermined and fixed preferences has to be slightly modified. Obviously, deliberation is taking place among actors in which interests may be moved and reshaped on the basis of expert knowledge, however, not primarily in the CCs as assumed by Joerges and Neyer (1997). Moreover, there is obviously also a good deal of trust in the Commission, as further underpinned by Table 6. Table 6 about here National officials attending different EU committees seem to agree on the relative independence of Commission officials from particular national interests. Only a very small minority, mostly among the CWP participants, reports that Commission officials act more in the interest of their country of origin. Hence, there is obviously a good deal of trust in the Commission as an independent supranational executive. Thus, participation in EU committees tends to affect the institutional allegiances and role perceptions of the participants. Nonetheless civil servants largely retain their national and sectoral identities when attending EU committees. An element of EU-level loyalty does, however, supplement such pre-existing allegiances to some extent (H1 and H6). The emerging community administration thus seems to uphold a strong intergovernmental dynamic that weakens its institutional independence from the member-state administrations. As expected, however, the institutional autonomy is stronger among the ECs than among the CWPs and the CCs. The co-ordination behaviour of member-state officials attending EU committees. In the last section we have demonstrated that national officials attending Commission ECs are probably behaving more like independent experts than when attending CWPs and CCs. In contrast, when attending CWPs and CCs, national officials perceive themselves and their colleagues from other member-states more as government representatives. The different role and identity perceptions of national government officials attending different EU committees may partly reflect different co-ordination processes at the national level (Trondal 2002). One difference may be expected between officials attending Commission ECs on the one hand, and officials participating in CWPs and CCs on the other. Officials attending Commission ECs are expected to be less subject to national co-ordination. Officials attending CWPs and CCs, on the other hand, are more likely to participate in committee meetings with clearly co- ordinated 'positions' from their respective national governments (Trondal 2000). The reasons for this difference are twofold: The formal organisation of the committees and the voting practices within them. First, the Commission ECs are mainly organised according to sectoral and functional principles. The CWPs and the CCs, although sectorally and functionally specialised, have a stronger territorial component in their organisational structures (H4). Arguably, committees organised by territory accompany stronger co- ordination pressure on the participants than committees organised by sector and function (Egeberg and Trondal 1999). Secondly, voting focuses the attention of decision-makers. Voting also signals expectations from the principals towards the agents with respect to representing agreed-on and often written "positions". In contrast to CCs, the ECs and CWPs do not vote in any formal sense (Mattila and Lane 2001; Tuerk and Schaefer 2002). CWPs are, however, located more clearly in the "shadow of the vote" than Commission ECs (Golub 1999; Tuerk and Schaefer 2002). Whereas EC participants are not expected to reach any agreements or formal decisions during most committee meetings, officials attending the CWPs and the CCs are expected to reach compromises, majority decisions and often consensus at the end of meetings (Lewis 1998). Table 7 reveals different modes of policy co-ordination behaviour amongst EU committee participants. As expected, participants in Commission ECs seem less co-ordinated nationally than officials participating in CWPs and CCs (H2 and H4). Officials attending CCs seem to be even better co-ordinated nationally than officials attending CWPs, though the difference is not very large. By comparison, officials in Commission ECs tend to take 'positions' that are less strongly co-ordinated back home (H1 and H2). Still, when asked whether national interests or professional considerations are deemed vital when deciding what 'positions' to pursue, no major differences are observed between officials participating in different EU committees (H4). CWP participants seem, however, to pay more attention to national interests than do ECs and CC participants (Fouilleux, Smith and Maillard 2002). These differences are marginal, however. The most significant observation is that in Commission ECs, participants have much more leeway to follow "their" own position than in the CWPs and the CCs. Table 7 about here Conclusion The study of European integration has increasingly shifted focus from the horizontal spill- over processes at the EU level and the 'grand bargains' struck between the strong EU member-states towards the vertical blurring of governance levels across the EU – nation-state interface. This article has focused on the emerging community administration where government levels interact and affect each other. In the empirical analysis we observed that many national officials spend a considerable amount of time and energy on EU committee work. In fact almost one third of our respondents use at least half of their working hours on preparation, co-ordination and participation in EU committees. CWPs are more demanding in this respect than other EU committees. Officials from small member-states seem to attend meetings more frequently than their counterparts from larger countries. This is due to the smaller size of their administrations. Moreover, as could be expected given the primary institutional affiliation of national officials, national allegiances are more clearly expressed than EU-level identities. However, a considerable proportion also feels loyalty to the committee(s) in which they participate. A clear majority expresses considerable trust in the Commission in the sense that they acknowledge its independence from particular national interests. Commission officials are among their most important interlocutors. Sheer intergovernmentalism is also transcended in the sense that the quality of the arguments seems more important than the kind of country the speaker originates from. The multiple roles and identities evoked by our respondents also point beyond a pure intergovernmental logic. In all kinds of committees they identify themselves heavily with sectoral and functional administrations and policy arenas. The government representative role is most clearly expressed in the CWP and CC settings. It is also in these settings that their positions and mandates are most clearly co-ordinated and instructed back home. As already said, our sample cannot claim to be a representative one. However, our main findings are clearly substantiated by studies based on other sources (Egeberg 1999; Trondal 2001; Trondal and Veggeland 2003). Recent literature argues that EU committees are sites of vertical and horizontal fusion of administrative systems and policy instruments, often described as Europeanisation (Egeberg 1999, Maurer and Larsson 2002; Schaefer 2002; Trondal 2001). This study has pictured this phenomenon as the emergence of a multilevel community administration. The observations presented demonstrate that EU committees are indeed an important part of the emerging community administration that cross-cuts existing borders of the member-states and the EU institutional apparatus. The study demonstrates that the attention, energy, contacts, co- ordination behaviour and loyalties of national civil servants are to a considerable extent directed towards the Brussels committee system. Hence, the decision-making and agenda- setting processes within national governments are integrated into the EU agenda setting phase (see Larsson and Trondal in this volume). However, this study also indicates that the re- socialising and transformative powers of the EU committees are heavily filtered and biased by the national institutions embedding the EU committee participants. Last, but not least, the data reveals that the institutional autonomy of this multilevel community administration is stronger within the Commission than within the Council and the comitology setting. Hence, the picture of one unified community administration has to be sacrificed for the model of a multifaceted community administration balancing intergovernmental, functional and supranational dynamics. References Alderson, K. (2001) 'Making Sense of State Socialization', Review of International Studies, 27 (issue 3): 415-33. Barnett, M. (1993) 'Institutions, Roles, and Disorder: The Case of the Arab States System', International Studies Quarterly, Vol. 37: 271-96. Barnett, M. and M. Finnemore (1999) 'The Politics, Power, and Pathologies of International Organizations', International Organization, Vol. 53, No. 4: 699-733. Beyers, J. and J. Trondal (2004) 'How Nation-States "Hit" Europe. Ambiguity and Representation in the European Union', West European Politics, Vol. 27, No. 4 (forthcoming). Biddle, B.J. (1986) 'Recent Development in Role Theory', Annual Review of Sociology, Vol. 12: 67-92. Checkel, J.T. (1999) 'Social Construction and Integration', Journal of European Public Policy, Vo. 6, No. 4: 545-60. Checkel, J.T. (2004) 'Social Mechanisms and the Quality of Cooperation: Are Europe and the EU Really All That Different?', ARENA working paper, No. 8. Christiansen, T. and E. Kirchner (2000) Europe in Change. Committee Governance in the European Union. Manchester: Manchester University Press. Egeberg, M. (1999) 'Transcending Intergovernmentalism? Identity and role perceptions of national officials in EU decision-making', Journal of European Public Policy, Vol. 6, No. 3: 456-474. Egeberg, M. (2001) 'How Federal? The Organisational Dimension of Integration in the EU (and Elsewhere)', Journal of European Public Policy, Vol. 8, No. 5: 728-46. Egeberg, M. (2004) 'An Organisational Approach to European Integration: Outline of a Complementary Perspective', European Journal of Political Research, Vol. 43, No. 2: 199- 219 Egeberg, M., G.F. Schaefer and J. Trondal (2003) 'The Many Faces of EU Committee Governance, West European Politics, Vol. 26, No. 3: 19-40. Egeberg, M. and J. Trondal (1999) 'Differentiated Integration in Europe: The Case of the EEA Country, Norway', Journal of Common Market Studies, vol. 37, No. 1: 133-142. Elster, J. (1986) The Multiple Self. Cambridge: Cambridge University Press. Eriksen, E.O. and J.E. Fossum (eds.) (2000) Democracy in the European Union. Integration through Deliberation?. London: Routledge. Eulau, H. et al. (1959) 'The Role of the Representative: Some Empirical Observations on the Theory of Edmund Burke', American Political Science Review, Vol. 53: 742-56. Fouilleux, E., A. Smith and J. de Maillard (2002) 'Council Working Groups: Their role in the production of European problems and policies', in G.F. Schaefer (ed.) Governance by Committee. The Role of Committees in European Policy-Making and Policy Implementation. Maastricht: EIPA. Golub, J. (1999) 'In the shadow of the vote? Decision-making in the European Community, International Organization, Vol. 53, No. 4: 733-764. Hart, P. et al. (1997) Beyond Groupthink. Ann Arbor: The University of Michigan Press. Hayes-Renshaw, F. and H. Wallace (1997) The Council of Ministers. New York: St. Martin's Press. Heywood, P. and V. Wright (1997) 'Executives, Bureaucracies and Decisin-making', in M. Rhodes, P. Heywood and V. Wright (eds.) Developments in West European Politics. New York: St. Martin's Press. Hooghe, L. (2001) The European Commission and the Integration of Europe. Images of Governance. Cambridge: Cambridge University Press. Institut für Europäische Politik (1987) Comitology – Characteristics, Performance and Options. Preliminary Final Report: Bonn. Joerges, C. and J. Neyer (1997) 'From Intergovernmental Bargaining to Deliberative Political Processes: The Constitutionalisation of Comitology', European Law Journal, Vol. 3, No. 3: 273-99. Jupille, J., J.A. Caporaso and J.T. Checkel (2003) 'Integrating Institutions - Rationalism, Constructivism, and the Study of the European Union, Comparative Political Studies, Vol. 36, No. 1-2: 7-40 Kohler-Koch, B. (2003) Linking EU and National Governance. Oxford: Oxford University Press. Larsson, T. (2003) Precooking in the European Union. The World of Expert Groups. A Report to the Expert Group on Public Finance. Ministry of Finance, Sweden, Ds 2003:16. Lewis, J. (1998) 'Is the 'Hard Bargaining' Image of the Council Misleading? The Committee of Permanent Representatives and the Local Elections Directive', Journal of Common Market Studies, Vol. 36, No. 4: 479-504. Maurer, A. and T. Larsson (2002) 'Democratic legitimacy in EU politics – no way out for committees', in G.F. Schaefer (ed.) Governance by Committee, the Role of Committees in European Policy-Making and Policy Implementation. Maastricht: EIPA. March, J.G. (1994) A Primer on Decision Making. How Decisions Happen. New York: The Free Press. March, J.G. (1999) The Pursuit of Organizational Intelligence. Massachusetts: Blackwell Business. March, J.G. and J.P. Olsen (1979) Ambiguity and Choice in Organizations. 2nd ed., Oslo: Universitetsforlaget. March, J.G. and J.P. Olsen (1994) 'Institutional Perspectives on Governance', in Systemrationalität und Partialinteresse: Festschrift für Renate Mayntz / Hans-Ulrich Derlien. Baden-Baden: Nomos. March, J.G. and J.P. Olsen (1995) Democratic Governance. New York: The Free Press. March, J.G. and J.P. Olsen (2004) 'The logic of appropriateness', ARENA working paper, No. 9. Mattila, M. and J.-E. Lane (2001) 'Why Unanimity in the Council? A Roll Call Analysis of Council Voting', European Union Politics, Vol. 2, No. 1: 31-52. Olsen, J.P. (2003) 'Reforming European Institutions of Governance', in J.H.H. Weiler, I. Begg and J. Peterson (eds.) Integration in an Expanding European Union. Reassessing the Fundamentals. Oxford: Blackwell. Pettigrew, T.F. (1998) 'Intergroup Contact Theory', Annual Review of Psychology, Vol. 49, No. 1: 65-85. Radaelli, C.M. (1999) Technocracy in the European Union. London: Longman. Risse, T. (2002) 'An Emerging European Identity? What we know and how to make sense of it', in Europe Transformed? The European Union and Collective Identity Change. ARENA/IDNET International Policy Conference, 11 October 2002. Schaefer, G.F. (ed.) (2002) Governance by Committee, the Role of Committees in European Policy-Making and Policy Implementation. Maastricht: EIPA. Schaefer, G.F. et al (2002) 'How do Comitology Committees work: an insider perspective', in G.F. Schaefer (ed.) Governance by Committee. The Role of Committees in European Policy- Making and Policy Implementation. Maastricht: EIPA. Simon, H. (1957) Administrative Behaviour. 2 ed., New York: Macmillan. Trondal, J. (2001) Administrative Integration Across Levels of Governance. Integration through Participation in EU Committees. ARENA Report, No. 7. Trondal, J. (2000) 'Multiple Institutional Embeddedness in Europe. The Case of Danish, Norwegian and Swedish Government Officials', Scandinavian Political Studies, Vol. 23, No. 4: 311-341. Trondal, J. (2002) 'Beyond the EU membership – non-membership dichotomy? Supranational identities among national EU decision-makers', Journal of European Public Policy, Vol. 9, No. 3: 468-487. Trondal, J. (2004) 'Political Dynamics of the Parallel Administration of the European Commission', in A. Smith (ed.) Policies and the European Commission. Actors, interdependences, legitimacy. London: Routledge. Trondal and F. Veggeland (2003) 'Access, Voice and Loyalty. The Representation of Domestic Civil Servants in EU Committees', Journal of European Public Policy, Vol. 10, No. 1: 59-77. Tuerk, A. and G.F. Schaefer (2002) 'Legislation and Implementation: theoretical considerations and empirical findings', in G.F. Schaefer (ed.) Governance by Committee. The Role of Committees in European Policy-Making and Policy Implementation. Maastricht: EIPA. Wessels, W. (1998) 'Comitology: Fusion in Action. Politico-Administrative Trends in the EU System', Journal of European Public Policy, Vol. 5, No. 2: 209-34. Notes This study has been financially supported by Sørlandets kompetansefond, EIPA and ARENA. The emergence of multiple roles and identities may reflect ill-coordinated processes. New roles and identities may be adopted without removing old ones, old identities and roles may be removed without adopting new ones, or the mix of and balance between existing identities and roles may simply be reorganized. From a sociological institutionalist perspective it may be easier to adopt new roles and identities than to remove old ones. According to the cognitive perspective, however, the mix of roles and identities are likely to be institutionally contingent and efficient. The questionnaire was jointly developed by Morten Egeberg, Jarle Trondal and Guenther F. Schaefer together with the „Comitology team" at EIPA. By the end of 1999, 232 questionnaires had been completed. Of these, 8 were Norwegians, and in 6 cases it was impossible to identify clearly the Member State affiliation of the respondents. Both these categories of respondents are excluded from this analysis. This article is thus based on 218 completed questionnaires as indicated in table I. It could be argued that this may be the result of sampling. The top level of the Member State administrations can not usually be expected to attend three-day seminars. In fact, this top level may well be over-presented in our sample since it hardly constitutes 20% of a Member State's administration. See also Institut für Europäische Politik (1987). Almost 30% of the respondents reported, however, that they spent 50% or more of their working time on EU matters. See Table 2. TABLES Table 1 Composition of the sample, by member-state and institutional affiliation Member-State Ministry or Institution Total Foreign Ministry Other Ministries Agencies etc Permanent Representation AUSTRIA 14 3 17 BELGIUM 2 20 7 29 DENMARK 1 5 1 7 FINLAND 2 17 2 21 FRANCE 3 1 4 GERMANY 7 3 1 11 GREECE 1 1 2 IRELAND 1 2 4a LUXEMBOURG 1 1 NETHERLANDS 2 10 1 13 PORTUGAL 5 3 1 9 SPAIN 55 5 60 SWEDEN 2 23 9 34 UNITED KINGDOM 1 4 1 6 TOTAL N 16 163 34 4 218a a) One respondent did not answer the question about institutional affiliation. Table 2 Working time consumed on EU committee work, by position (%) Position Working time consumed on EU committees Director general, Deputy dir. general Head/Deputy of unit/division Head of section, Senior advisor, Advisor Total 15% or less 37 26 24 27 15-50% 43 44 44 44 50% or more 20 30 32 29 Total % 100 100 100 100 N 40 27 131 198 Table 3 Percentage who to a great extenta feel allegiance to (identify with or feel responsible to) the following when participating in EU committees EC CWP CC My own government 65 76 69 My own ministry, department or agency 74 81 60 The requirements of the policy arena in which I am working 58 65 58 My own professional background and expertise 60 65 60 The committee or group in which I participate 39 57 44 Total N 106 109 58 a) Values 1 and 2 combined on the following five-point scale: to a very great extent (value 1), to a fairly great extent (2), both/and (3), to a fairly small extent (4), to a very small extent (5). Table 4 Officials' perception of the role of colleagues from other countries when participating in EU committees (%) EC CWP CC Mainly independent experts 33 11 6 Mixed roles 22 12 20 Mainly government representative 45 77 74 Total % 100 100 100 N 113 122 66 Table 5 Percentage who give much considerationa to proposals, statements and arguments from the following when participating in EU committees EC CWP CC Colleagues and experts from my own Member-State 87 84 81 Colleagues from other member-states who have demonstrated considerably expertise on the subject matter at hand 73 70 69 Colleagues from large member-states 38 38 30 Colleagues from member-states from my own region 42 46 48 Colleagues from member-states who share a similar position 61 71 68 Representatives from the Commission 57 60 57 Interest groups and firms I know from my Member-State 26 32 44 Interest groups and firms I know or have contact with at the European level 17 11 13 Total N 113 121 66 a) Values 1 and 2 combined on the following five-point scale: very much consideration (value 1), fairly much consideration (2), both/and (3), fairly little consideration (4), very little consideration (5). Table 6 National officials' perceptions of Commission officials' independence of particular national interests when participating in EU committees (%) EC CWP CC Mainly independent 81 70 79 Mixed roles 13 18 16 Mainly dependent 6 12 5 Total % 100 100 100 N 109 112 63 Table 7 Percentage of officials who co-ordinate their "position" most of the timea before participating in EU committee meetings EC CWP CC I have to co-ordinate with the Foreign Office or another central co-ordinating body 20 47 43 My "position" has in fact been co-ordinated with all relevant ministries 28 47 53 My "position" has been co-ordinated with all relevant departments in my own ministry 38 55 59 I have clear instructions about the "position" I should take 28 35 46 I take the "position" I think is in the best interest of my country 63 72 66 I take the "position" I think is best on the basis of my professional expertise 43 43 34 If I have no instructions, or if the question is not important for my country, I take the "position" I think is the best for the member-states as a group 52 46 46 Total N 110 119 62 a) Value 1 on the following three-point scale: always or most of the time (value 1), about half of the time (2), rarely or never (3). 1