ARENA Working Papers
WP 99/32

Multiple Institutional Embeddedness in Europe
The Case of Danish, Norwegian and Swedish Government Officials*

Jarle Trondal**
ARENA, Department of Political Science, University of Oslo

Abstract

The question posed in the current study is whether, and how, participation within Commission expert committees and Council working parties affects the co-ordination behaviour of the participants. Based on organizational theory arguments, the co-ordination role of the Ministry of Foreign Affairs is arguably weakened by institutional dynamics existing within Commission expert committees. The opposite is argued to be the case within Council working parties. Empirically, this study is based on 160 questionnaires and 47 face-to-face interviews with Danish, Norwegian and Swedish government officials attending EU committees. Being excluded from attending Council working parties, Norwegian civil servants participating within Commission expert committees are shown to co-ordinate considerably less with the Foreign Ministry than their Danish and Swedish counterparts. Still, notwithstanding these observations, this study also reveals the pivotal role of domestic government institutions as to affect co-ordination behaviour amongst civil servants. In addition to show how EU committees affect co-ordination behaviour amongst the participants, the current analysis also shows how responses to integration requirements are filtered – and even conditioned – by a prior state of affairs at the domestic level of governance.

Introduction

The EU bureaucracy is made up of permanent Eurocrats together with “travelling national civil servants” (Nedergaard 1995:26). Van Schendelen (1996) estimates that about 50 000 domestic officials are temporary engaged in the administrative work of the Union per year. The Commission organized about 5 500 committee-meetings in 1989 and approximately 64 000 days were spent on them. “Since about 1963 the Commission's own work has become dominated by that of these various intermediary committees of national representatives” (Coombes 1970:243). Thus, “committees function as hybrids between EU governance and the organisations represented” (van Schendelen and Pedler 1998:288).

Past and more recent literature on EU committees have in various ways measured the degree of cross-level participation through EU committees (e.g. Institut f�r Europ�ische Politik 1989; Wessels 1998). One question less frequently addressed is how such cross-level participation affects the actors themselves – their identities, role conceptions and modes of acting. The current study tries to grapple with this question through analysing how participation within EU committees affects domestic co-ordination processes. We ask whether the gate-keeping role of the Foreign Ministry (MFA), being the last stronghold of the nation state, is affected by the institutional dynamics existing within EU committees.

Past and contemporary literature pictures EU committee-meetings as generally business-like, depoliticized, consensual and technocratic; emphasising the `technocratic collusion' at the Community level (Dehousse 1997, 48; Wessels 1998; 1999, 265). Current studies on EU committees largely ignore the differentiated institutional character of the EU committees, paying heed to their commonality as a symptom of administrative integration and engrenage more broadly (Joerges and Neyer 1997; Neyer 1999). The current study argues that the way EU committees are formally organized may affect the co-ordination behaviour of the participants. Thus, the co-ordination role of the MFA is arguably affected differently by the institutional dynamics within Commission expert committees and Council working parties.

Studies of actual decision behaviour within domestic governance institutions show that the formal organization of the bureaucracy is important, together with the professional background of the officials (L�greid and Olsen 1984; Egeberg 1999b). The formal build-up of the central governmental apparatus is also logically revealed to impact upon actual decision behaviour amongst government officials (Hammond 1986). This study argues basically that we have to unpack the organizational structure of the EU system in order to derive how domestic co-ordination processes are likely to be affected by it. In addition, we have to carve up the bureaucratic machinery of the nation state in order to unravel the dual institutional affiliations of domestic government officials, and ultimately to determine how different organizational structures are likely to affect modes of policy co-ordination. Organizational structures are seen as mobilizing particular modes of behaviour, identities and role conceptions. Particular models of man and particular models of society embedding social interaction are perceived of as moulded by particular principles of organization. Applying organization theory suggests going beyond a narrow sui generis view of the EU system, providing understanding of “the unfamiliar with the help of familiar conceptual lenses” (J�nsson et al. 1998, 320; cf. Stone Sweet and Sandholtz 1998).

Based upon 160 questionnaires and 47 face-to-face interviews with Danish, Norwegian and Swedish government officials attending EU committees, the current study shows that officials attending Council working parties tend to co-ordinate more frequently inter-sectorally compared to officials attending Commission expert committees. ^[1] I argue that these dynamics are likely to be fuelled by the sectoral organizational principles underpinning the EU Commission in general, and the Commission expert committees in particular. One implication is that the Ministry of Foreign Affairs are largely by-passed as a vital co-ordinating unit at the domestic level. On the other hand, I argue that this mode of sectoral penetration of sector ministries and agencies may be filtered and modified to some extent if the MFA is percieved of as a central co-ordinating body as far as domestic EU politics and policies are concerned. I argue that the Council of Ministers in general, and the Council working parties in particular, may contribute to the empowerment of the MFA as a vital co-ordinating unit. ^[2] However, the current study also shows how “responses to integration requirements are filtered – and even conditioned – by a prior state of affairs” at the domestic level of governance (Spanou 1998:469).

In the following, some theoretical reflections are addressed in order to render different modes of co-ordination intelligible from a multiple institutional embeddedness point of view. Secondly, empirical data is provided in section three, drawing on survey data from Danish, Norwegian, and Swedish civil servants participating within Commission expert committees (EC's) and/or Council working parties (WP's). These data covers officials employed within domestic ministries and agencies, but also officials at the permanent representation to the EU of these Scandinavian countries.

A Theoretical Account on Multiple Institutional Embeddedness

The study of European integration and `Europeanization' of domestic bureaucracies has to an increasing extent been influenced by organization theory perspectives (e.g. Bulmer 1993; Cram 1997; Egeberg and Trondal 1999; Hix 1998; Olsen 1998). Organizational theory has posed two major questions in past and more recent literature: Firstly, whether institutions matter, and secondly how institutions matter. This study goes largely beyond the question of whether institutions do matter (Weaver and Rockman 1993). Aiming at understanding how organizational structures affect co-ordination processes, in the following will be added some microfoundations to this end, based upon cognitive organization theory (Trondal 1999b).

Cognitive theory, developed within social psychology and introduced to organizational theory largely by Cyert and March (1992), March and Simon (1993) and Simon (1997), perceives of organizations and their boundaries as mechanisms of simplification. The underlying assumption is that of bounded rationality: The possibility for attending everything simultaneously is difficult. The cognitive perspective perceives of organizational and professional borders as mechanisms for coupling and decoupling actors, problems, solutions and consequences. Organizational and professional borderlines are seen as buffers to attention thereby biasing the information exposed to the decision-makers (March and Olsen 1995; Tenbrunsel et al. 1996). “Cognitive structures simplify when there is too much, and they thus allow the perceiver to reduce an enormously complex environment to a manageable number of meaningful categories” (Markus and Zajonc 1984, 143). Within a cognitive notion, organizational boundaries are seen as cognitive buffers to attention and information. Organizational and professional boundaries render it possible to decompose complex tasks into sub-tasks that can be carried out within relatively independent units of governance. Hence, organizational and professional structures contribute to the development of `cognitive short cuts' (Johnson 1987, 45). These short cuts contribute to the creation of cognitive categories and representations of world phenomena to the individual. As a result, these phenomena are taken for granted as “the way we do these things” (Scott 1995, 44). Cognitive scripts provide “guidelines for sense-making and choosing meaningful action” (Scott 1995, 44). Organizational and professional boundaries thus affect modes of behaviour because these properties simplify cognitive search-processes (March 1994; Johnson 1987).

Regarding the institutional embeddedness' of decision-makers, two pivotal arguments may be outlined on the basis of this cognitive organization theory perspective. Firstly and foremost, decision behaviour may be moulded and re-moulded on the basis of organizing and reorganizing organizational boundaries (March and Olsen 1989; Nkomo and Cox jr. 1996). Reorganizing from one principle of organization to another will alter the flow of information within organizations, ultimately changing the flow of information exposed to each decision-maker. Hence, when organizational members change organizational affiliations, they tend to choose new ways of acting subsequent to being exposed towards new sets of information. Stated otherwise, certain stimuli tend to produce certain responses. Having internalized a multitude of codes of conduct, certain stimuli - like organizational boundaries - will evoke or activate only a limited proportion of this repertoire of responses (Sev�n 1996). Consequently, reorganizing the set of stimuli being exposed to the actor, the responses are likely to alter in systematic manners.

Domestic civil servants attending EU committees have multilevel institutional affiliations. These officials are employed within domestic ministries and agencies, but at the same time participate within EU committees on the other. Further, these officials also have professional affiliations towards different disciplines. Hence, domestic decision-makers tend to be penetrated by multiple, partially contending, sets of information, premises and considerations simultaneously. One important research task has to do with identifying the conditions under which decision behaviour is affected by different institutional affiliations. Studying government officials being members of government institutions at different levels of governance, I propose to emphasise organizational linkages between government institutions as one vital scope condition affecting modes of co-ordination at the domestic level of governance. One central feature attached to multiple institutional affiliations relates to the formal linkages between organizations, and to the intensity and length of such linkages.

On Principles of Organization
Different modes of policy co-ordination may reflect the basic organizing principles underpinning institutional linkages between formal organizations. One may argue that the EU system of governance in general, and the web of EU committees and working parties in particular, are organized according to two basic and general principles in administrative life. Parallel to the sectoral and spatial institutional build-up of the domestic political-administrative apparatus (sector ministries and agencies versus Foreign Ministries), the EU Commission and the preparatory expert committees underlying it may be seen as exhibiting sectoral and functional principles. ^[3] Conversely, the Council of the European Union and the web of working parties organized under it may be seen as exhibiting spatial and territorial principles of organization (Egeberg and Trondal 1999). Notwithstanding several contending organizing principles do exist within the Commission and the Council of Ministers, I argue that the dominating and uppermost principles are mentioned above. Sectoral principles tend to bias and skew co-ordination behaviour in sectoral directions. Being exposed to institutions organized by sector, the organizational members are systematically exposed to sectoral information and stimuli. The area principle, on the other hand, are arguably more likely to activate co-ordination behaviour being more inter-sectoral in character.

Co-ordination within and between government institutions may be either written or oral, or both. I argue that oral co-ordination is more likely when only few actors are involved in the process, or if the actors share some fairly general interests, problems or world-views. Co-ordination through written documents are more likely when more actors are involved, where more conflicting interests are to be accommodated, and when more problems and world-views have to be taken into consideration. The first scenario is more likely to materialize within, and between government institutions. Within government institutions officials tend to be fewer, and the number of problems, solutions and identities tend to be fewer and more easily accommodated than at the inter-ministerial arena. This argument may also be applied to EU institutions: The EU Commission are sectorally organized, whilst the Council are organized according to an area principle, implying that different sectors more frequently are confronted. Thus, the use of written co-ordination is more likely to result from decision processes within the Council than within the EU Commission.

Taken together, government officials attending Commission expert committees are likely to evoke intra-sectoral and oral modes of co-ordinating, while officials participating within Council working parties are more likely to evoke inter-sectoral and written modes of co-ordinating. Domestic civil servants participating within both expert committees and working parties are likely to evoke both these modes of co-ordination. Owing to the EEA agreement, Norwegian civil servants have access solely to Commission expert committees, not to Council working parties. In the current study, Sweden and Denmark represent countries having participatory rights within both expert committees and working parties. Thus, whereas inter-sectoral, and written modes of policy co-ordination are likely to take primacy amongst Swedish and Danish officials, more intra-sectoral and oral modes of co-ordination are likely to be evoked by Norwegian government officials. Still, emphasising how different organizational principles within different EU committees may affect co-ordination behaviour differently, also Swedish and Danish EC-participants are likely to evoke intra-sectoral and oral modes of co-ordination.

Moreover, formal voting rights may foster co-ordination behaviour more intensively than the lack of formal voting rights, both within expert committees and working parties. ^[4] As a result of Norwegian government officials not having formal voting power within Commission expert committees, as do Danish and Swedish officials, the pressure towards co-ordinating EU `positions' may be generally lower within the Norwegian bureaucracy compared to the Danish and Swedish bureaucracy.

Several conflicting organizational principles are often built into organization structures: Within the EU Commission a geographical principle runs parallel with the sectoral principle, and within the Council a sectoral principle of organization is present, supplementing the area principle. Hence, both inter- and intra-sectoral modes of co-ordination may accompany institutional dynamics within both these sets of EU committees. Still, these additional, and partially contending principles of organizations, may be biased by, and embedded within, other uppermost principles of organization (Gulick 1937). Hence, certain institutional dynamics are likely to take precedence over other dynamics.

Next, when studying government officials having multiple institutional affiliations at different levels of governance, one has to pay heed to how these levels interact. Hence, we need to trace how administrative life is formally organized at both levels of governance, and how each level of governance are formally linked together. I argue that co-ordination processes within domestic government institutions may be affected by EU committees, conditional upon: (i) the degree of compatibility in organizational structures across levels of governance, and upon (ii) individual and institutional seniority at both levels of governance.

On Institutional Compatibility
The impact stemming from organizational principles existing at different levels of governance may be conditioned by the degree of institutional compatibility across these levels of governance (Coporaso, Cowles and Risse 1999; Knill and Lenschaw 1998; Knill and Lehmkuhl 1999; March and Olsen 1995). The impact fuelled by institutional compatibility owes to the compatibility in cognitive short cuts and scripts being systematically presented to the decision-makers. I argue that institutional interpenetration – and thus, the potential for domestic government officials being affected by institutional dynamics within EU institutions - is more likely to emerge if government apparatuses at each level of governance are rather similar. I have argued elsewhere that the Commission expert committees on the one hand, and domestic sector ministries and agencies on the other are, in general, organized along the same sectoral lines. Similarly, Council working parties and domestic Foreign Ministries may be seen as organized according to an area principle (Trondal 1999b). Thus, the co-ordination behaviour chosen by government officials, employed within sectoral ministries and agencies, are likely to reflect institutional dynamics within expert committees more strongly than institutional dynamics within working parties. The inverse relationship is likely amongst officials employed within the MFA attending expert committees. Instead of studying officials from the MFA, however, this study covers officials at the permanent representation to the EU. Permanent representations are formally organized under the auspices of the MFA, thus, embodying territorial principles of organization.

Domestic government officials are only part-time participants at the EU level of governance, and thus likely to be pre-socialized to a great extent before entering EU committees (Beyers 1998). This article considers several primary institutional affiliations: Most importantly, civil servants are employed within domestic ministries and agencies. Ceteris paribus, owing to the vertical de-specialized ministry-agency structure in Norway, Sweden and Denmark, I argue that government officials employed at the agency level are more likely to evoke intra-sectoral modes of co-ordination compared to officials employed at the ministry level. Conversely, officials at the ministry level are more likely to evoke territorial and inter-sectoral modes of co-ordinating. Further, officials at the permanent representations to the EU are likely to evoke inter-sectoral modes of co-ordination more frequently than are officials within the `home administration' – that is, officials employed within domestic ministries and agencies. This owes largely to the fact that the permanent representations are formally organized under the auspices of the domestic MFA – thus embodying territorial principles of organization. Moreover, most permanent representations are internally organized in sectoral and diplomatic realms. I argue that inter-sectoral modes of co-ordination are arguably more likely to be enacted amongst officials in the later than in the former institutional realms. This owes to the fact that different organizational principles are uppermost within different realms at the permanent representations (Trondal 1999b). Further, officials in top rank positions are more likely to enact inter-sectoral modes of co-ordination compared to officials in medium rank positions. Finally, officials are likely to be pre-socialized from professional training even before entering the domestic bureaucracy. The current study consider lawyers, economists, social scientists and officials having more technical training – like biology, physics, chemistry, etc. Generally, I argue that law and social sciences are more cross-sectoral in character than are economy and technical disciplines. Thus, intra-sectoral modes of co-ordination are likely to accompany training in economy and technical disciplines, whereas more inter-sectoral modes of co-ordination may accompany training in law and different social sciences.

On Length and Intensity in Cross-Level Participation
Protracted and intensive participation within EU committees and working parties may add to the effect of (i) organizational principles, and (ii) institutional compatibility. Ceteris paribus, senior participants within EU institutions are more likely to be affected by the institutional dynamics dominating these committees compared to officials being newcomers at the EU arena (Beyers 1998). Length and intensity of attending EU committees arguably affect the extent to which the participants are affected by the organizational principles underpinning these committees. Theoretically, seniority and intensity of participation within EU committees may affect ways of acting through cognitive processes whereby civil servants are systematically exposed towards particular sets of information. Moreover, intensity and length are treated as intervening variables in the current study, conditioning the impact fuelled by principles of organization. The impact fostered by particular organizational principles may be further strengthened owing to intensive and protracted exposure towards these principles. Thus, domestic government officials who generally participate intensively within Union bodies, may be relatively easily affected by institutional dynamics within EU committees, that is, their way of co-ordinating being partially a reflex of institutional dynamics within EU committees (Lindberg 1963:287). Hence, individual seniority at the EU level is likely to condition the extent to which domestic co-ordination processes are affected by the institutional dynamics within EU committees. Individual seniority has to do with length of service within particular institutions. These propositions are ceteris paribus in character. ^[5] Furthermore, within each nation state, different government institutions have participated at the EU level for different periods of time. Institutional seniority has to do with the length to which one institutiona has interacted with another. Hence, officials employed within government institutions being senior participants at the EU arena are more likely to become Europeanized than officials within institutions being newcomers at the European arena. Similarly, officials from old EU member states are likely to be Europeanized more easily than are officials from new member states (Beyers 1998). Henceforth, individual and institutional seniority at the EU level may impinge upon the extent to which processes of Europeanization may occur.

As to sum up the main arguments put forward in this section, the following propositions may be highlighted: Inter-sectoral and written modes of co-ordination are most likely to be evoked amongst domestic government officials, educated in law or social sciences, participating within Council working parties, who have participated for a long time, who generally participate intensively, who are employed at the permanent representation to the EU - especially within top rank positions and within the diplomatic realm of this institution -, and who are senior officials within these institutions. Conversely, intra-sectoral and oral modes of co-ordinating are more likely to accompany domestic government officials, educated in economy or technical disciplines, participating within Commission expert committees, who participate intensively, who have participated for a long time, who are employed within domestic sector ministries or agencies - especially within medium rank positions -, and who have only shorter length of service within these institutions.

On Methodology and Data
In order to illuminate how domestic co-ordination processes are affected by institutional dynamics within Commission expert committees and Council working parties, data from countries having different forms of affiliation towards the EU have been systematically selected. Moreover, in order to minimize noise stemming from possible intervening variables, countries being `sufficiently' similar have been selected (Anchar 1993, 118): Denmark, Norway and Sweden. Norwegian government officials have participatory rights within approximately 50 percent of the Commission expert committees. Norwegian civil servants are, however, formally excluded from participating within Council working parties. Their Danish and Swedish counterparts, on the other hand, have additional participatory rights within the Council working parties. Secondly, in order to study the impact fuelled by individual and institutional seniority within EU committees, both Denmark and Sweden are included in this study despite having almost identical forms of affiliation to the EU: Danish officials may be expected to have, on average, longer tenure within EU committees than their Norwegian and Swedish colleagues (cf. Trondal 1999c).

In order to control for the potential impact of policy sector affiliation, and also aiming at limiting the empirical scope of this study, standardized questionnaires have been distributed to civil servants at the ministry and agency level within the environmental sector, and within the health and safety at workplace sector in the three Scandinavian countries. This survey was conducted from summer 1998 until spring 1999. The response rate is 79 %, giving 160 answers. Additionally, almost similar questionnaires were distributed to all government officials at the permanent representations to the EU of these three Scandinavian countries. This study was conducted from fall 1998 until spring 1999. The response rate of this survey is 53 %, giving 49 answers. Officials from the permanent representations to the EU are included into this study partially owing to a priori expectations as to the intensity of attending EU committees, and partially because the permanent representations formally embody territorial principles of organization. In addition, 47 face-to-face qualitative interviews were carried out during the spring of 1999. Officials to these interviews were selected from the survey sample above.

Certain biases do always accompany data based upon systematic selection of samples. In our data set officials at medium rank positions loom large compared to officials in top rank positions. Moreover, officials from the agency level are also well represented in the sample compared to officials from the ministry level. Similarly, officials from the sectoral realms at the permanent representations are `over-represented' compared to officials from the diplomatic realms. Still, we do not have any universe against which to compare our samples. Yet, available empirical findings indicate that our sample do not deviate significantly from any tentative universe (Trondal 1999c). Moreover, I have aimed at including all relevant officials within the two policy sectors to this study. On this basis, these data sets are not likely to be significantly biased in any meaningful sence. Overall, empirical generalizations from this study have to rest on our theoretical arguments addressed above.

The Empirical Record: Europeanization of Domestic Co-ordination Processes

The current section aims at illuminating how different institutional linkages across levels of governance are likely to affect domestic modes of co-ordinating EU related dossiers – as perceived by the civil servants themselves. The first part of this section reveals how processes of policy co-ordination within the domestic government apparatus reflect different forms of affiliations towards EU committees. The next part of this section aims at controlling statistically for the length and intensity of participation within EU committees. The final section shows how co-ordination processes reflect primary institutional affiliations at the domestic level of governance.

On Secondary Institutional Embeddedness: Forms of Affiliation towards the EU.
Past and contemporary research have uncovered that domestic co-ordination processes are at least two-dimensional in character: intra- and inter-sectoral (Egeberg and Trondal 1999; Hanf and Soetendorp 1998; Harmsen 1999; Metcalfe 1994; Rometsch and Wessels 1996; Trondal 1998). Whereas some scholars observe policy fragmentation and sectorization of domestic EU policy processes (Burnham and Maor 1994; Dehousse 1997; Dimitrakopoulos 1995; Hopkins 1976; Siedentopf and Ziller 1998), other scholars report cross-sectoral co-ordination processes as the central mode characterising EU policy making at the domestic level of governance (Bulmer and Burch 1998; Rometsch and Wessels 1996; Spence 1993). As regard the three Scandinavian countries, intra-sectoral modes of co-ordination are observed in Norway after the EEA-affiliation in 1994 (Egeberg and Trondal 1997 and 1999; Trondal 1996, 1998 and 1999d), whereas co-ordination processes have tended to oscillate between intra- and inter-sectoral modes more prominently within the Danish and Swedish central administrations, subsequent to their EU membership (Dosenrode 1993; Jacobsson 1999; Nedergaard 1994; SOU 1996:6; Statskontoret 1996:7; Trondal 1996). In the following, I reveal that different modes of co-ordination reflect institutional affiliations towards different EU committees.

I have argued that the general amount of policy co-ordination is likely to be greater amongst civil servants participating within Council working parties (WP's) compared to officials attending Commission expert committees (EC's). Moreover, I expect officials solely participating within EC's engaging into co-ordination processes more seldom than do officials also participating within WP's. Table one provides an overview over general co-ordination efforts, paying particularly heeds to different techniques applied to this end.

Table 1: Proportion of officials engaged into co-ordination efforts. (%)

1. This variable, and the following two variables, involves officials doing clearances fairly much, or more.
2. This variable, and the following two variables, regards officials who draw up these documents fairly often, or more.
3. This variable regards officials reporting these documents governing their positions to a fairly great extent, or more.

First, table one reveals that the usage of clearances between governmental institutions is more frequent amongst participants within WP's compared to officials participating within EC's. Still, the general amount of inter-sectoral co-ordination is also fairly high amongst officials participating within EC's. This may partially owe to the contending principles of organization within EC's, but may also reflect primary institutional affiliations at the domestic level of governance (cf. the next section). Finally, this may owe to the great usage of informal clearances amongst EC-participants. Supporting this interpretation, regarding techniques utilized for co-ordination, table one shows that participants within WP's utilizing formal co-ordination arrangements more frequently than do participants within EC's. Conversely, participants within EC's use informal co-ordination arrangements more frequently than do participants within WP's. Similarly, table one reveals that written instructions are used more frequently amongst officials within WP's than amongst officials attending EC's. These findings are also supported in qualitative interviews: Whereas officials negotiate under rather clear, written instructions within WP meetings, more unclear, often implicit mandates tend to accompany negotiations within EC meetings (Source: interviews; cf. Statskonsult 1999:6). However, regarding the extent to which written instructions, problem notes or frame notes tend to affect the positions taken by participants during committee-meetings, no significant difference is identified between the two classes of committee-participants. Still, the relative frequency of applying written problem notes, frame notes, and instructions vary immensely between the two classes of committee-participants.

When comparing officials solely participating within expert committees and officials also attending working parties, the same tendencies are revealed: ^[6]

Table 2: Proportion of officials engaged into co-ordination efforts. (%)

1. This variable, and the following two variables, regards officials doing clearances fairly much, or more.
2. This variable, and the following two variables, regards officials drawing up these documents fairly often, or more.
3. This variable regards officials reporting these documents governing their positions to a fairly great extent, or more.

The distinction made in table two corresponds to a great extent to the distinction between Norwegian officials on the on hand, and Danish and Swedish officials on the other. ^[7] Yet, the only striking difference between table one and table two regards the first variable: It seems that officials solely taking part within EC's make inter-organizational clearances more frequently than do officials also attending WP's. This difference may reflect the frequent usage of informal clearances amongst participants within EC's. Still, the overall pattern revealed by table two indicates significant differences between officials solely attending EC's compared to those also taking part within WP's. These findings are also supported by qualitative interviews. When interpreting table two, however, it is important to note that the dependent variables relate to civil servants' engagement within EC's, not to their participation within WP's. Thus, when participating within EC's, government officials may anticipate the importance of conducting cross-sectoral co-ordination when later entering the Council WP's. Officials participating within both EC's and WP's are exposed to other sources of information than are officials solely participating within EC's.

The above findings are clearly supported by table three: Officials participating within WP's perceive of their instructions as being considerably more clear than those participating within EC's:

Table 3: Proportion of officials who always or fairly often agree on the following assertions. (%)

Table three also reveals that the general amount of perceived leeway, or discretion relating to modes of proceeding, `positions' to follow, roles to evoke, etc, is significantly greater amongst officials participating within EC's compared to officials participating within WP's (cf. table 11). However, every participant works under some kind of national mandates. Still, as seen in the qualitative interviews, the amount of behavioural discretion available to the participants seems far more extensive within the EC's than within the WP's. <sup>[8]</sup> At the permanent representations to the EU, the same difference is revealed as regards modes of co-ordinating. Moreover, going beyond table 3, officials participating within EC's report their `positions' to be based upon professional expertise more extensively than do their colleagues participating within WP's. Furthermore, `national interests', as a catalyst for determining their `positions', are ascribed greater importance amongst officials attending WP's than are professional expertise. The inverse pattern is revealed amongst participants within EC's. Hence, the general amount of inter-organizational co-ordination processes seems to be affected significantly by the institutional embeddedness' within EU committees.

In the reminding part of this section, more systematic attention is devoted towards how co-ordination behaviour is moulded differently within EC's and WP's, thus paying more thorough heed towards intra- and inter-sectoral modes of co-ordination. Past and contemporary research on processes of Europeanization of domestic institutions and policy processes disagree on how domestic co-ordination processes are generally affected by the EU integration process (e.g. Wright 1996). The following section, however, reveals that different modes of policy co-ordination may reflect institutional affiliations towards different EU committees amongst domestic civil servants.

One initial empirical expectation, derived from the above theoretical discussion, would be that inter-sectoral modes of co-ordinating may reflect institutional dynamics within Council WP's, whereas intra-sectoral modes of co-ordinating to a greater extent is likely to accompany dynamics within the Commission EC's. Intra-sectoral modes of co-ordination, however, are also likely to be affected by institutional dynamics within Council WP's: Intra-sectoral modes of co-ordination may, to some extent, be seen as conditioning inter-sectoral co-ordination efforts: when co-ordinating across policy areas, officials may need to co-ordinate intra-sectorally in order to allow for cross-sectoral co-ordination efforts. Hence, institutional dynamics within WP's are likely to foster domestic co-ordination efforts more broadly – be it intra- or inter-sectorally – to a larger extent than institutional dynamics within EC's. A second important distinction to be made is that institutional dynamics within EC's are likely to foster intra-sectoral modes of co-ordination to a far greater extent than inter-sectoral modes of co-ordination. Finally, written modes of co-ordination are expected to accompany the territorial principles of organization underpinning WP's, whereas oral modes of co-ordination are more likely to reflect sectoral principles embedded within EC's

The following two tables (each divided into a and b) show the relative proportion of officials conducting inter-organizational clearances with the MFA, and the proportion of officials reporting intra-organizational clearances with other departments within their own governmental institution:

Table 4a: Proportion of officials, using somewhat or more of their time participating within EU committees, doing clearances with the Foreign Ministry. (%)

Table 4b: Domestic bureaucracies:

Table 5a: Proportion of officials conducting clearances with other departments within own government institution before entering EU committees. (%)¹

Table 5b: Domestic bureaucracies:^a

1) These tables include officials doing intra-organizational clearances fairly often, or more.

The figures presented in table 4a and 4b are not perfectly comparable to the figures presented in table 5a and 5b. Still, they reveal some interesting differences: As expected, participants within WP's utilize clearances more frequently than do participants within EC's – both intra-sectorally (table 5a) and inter-sectorally (table 4a). Still, those participating within EC's conduct clearances more frequently intra-sectorally than inter-sectorally. L�greid (1999, 25) also reveals tendencies of intra-sectoral modes of policy co-ordination to be more widespread within the Norwegian central administration than inter-sectoral co-ordination efforts – owing to Norwegian officials being solely affiliated to the Commission EC's. In Denmark and Sweden, inter-sectoral modes of co-ordination are reported to be more widespread (L�greid 1999; Nedergaard 1994; SOU 1996:6; Statskontoret 1996:7). This difference between Norwegian officials at the one hand, and Danish and Swedish officials on the other, owes largely to the additional institutional embeddedness' within Council WP's amongst the later. Still, Danish and Swedish officials, participating solely within EC's, report a general low frequency of inter-sectoral co-ordination efforts – especially towards the MFA – compared to Danish and Swedish officials attending WP's (Source: interviews). ^[9] Moreover, consulting table 4b, the correlations from table 4a is even more striking: 25 percentages of those officials solely participating within EC's make clearances with the MFA, whereas 56 percentages of those also participating within WP's make such clearances (cf. table 4b). Finally, table 4a shows that clearances with the MFA are taking place more frequently amongst officials at the permanent representations than amongst officials within the `home administrations'. This may owe to the fact that permanent representatives embody territorial, and thus cross-sectoral, roles more fully than do officials within the `home administration' (cf. table 4).

Co-ordination more generally involves contact patterns with other governmental institutions. The following table (a and b) reveals how participation within different EU committees may be reflected in contact patterns within the domestic administrative fabric:

Table 6a: Proportion of officials having contacts with the following units. (%)¹

Table 6b: Domestic bureaucracies:¹

1. These tables contain officials having contacts fairly often or more with the government institutions listed above.
2. This regards contacts towards the political leadership in the capital.
3. This variable regards officials employed at the ministry level within the domestic bureaucracy. This limitation does, however, not involve officials at the permanent representations to the EU.
4. This variable regards officials employed at the agency level within the domestic bureaucracy. This limitation does not involve officials at the permanent representations to the EU.

The upper halves of table 6a and 6b regards inter-sectoral modes of co-ordination. Consistent with prior observations within this section, one of the most striking observations from table six relates to the differences between the two classes of committees: Inter-sectoral contact patterns are more frequent amongst officials attending WP's compared to officials participating within EC's. Most clearly is this pattern revealed regarding contact patterns with the MFA, being the ministry formally embodying territorial principles of co-ordination. Officials participating within EC's have more frequent contacts with other sectoral ministries than with the MFA, save the domestic Parliament, or the executive political leadership. L�greid (1999, 20-21) observes the same tendency: Inter-sectoral contact patterns and co-ordination efforts are more widespread within the Danish and Swedish government apparatus than within the Norwegian bureaucracy, subsequent to their different form of affiliation towards the EU (cf. Sundstr�m 1999). When adding intra-sectoral contact patterns to this picture, the tendency from table five is repeated in table six: Those participating within EC's make relatively more frequent use of contacts with other government institutions within their own policy sector than with government institutions in other policy areas. Still, officials participating within WP's also have more frequent intra-sectoral contacts compared to EC-participants – that is, with hierarchical subordinated agencies, and with hierarchical superior ministries within the same policy area (cf. Trondal 1999d). When also comparing officials solely participating within EC's and officials also attending WP's, the same tendencies from table 6a are repeated in table 6b to a great extent. Still, some modifications are present, owing largely to the figures in table 6b relating to participation within EC's – not to their engagement within WP's (see above).

Finally, modes of policy co-ordination may be measured as how civil servants conceive of the relative importance of different government institutions. Contact patterns do not provide any understanding of the perceived importance of these contacts. Table seven (a and b) reveals how civil servants conceive of the relative importance of different domestic government institutions – both inter- and intra-sectorally:

Table 7a: Proportion of officials assigning weight to the following government institutions when important decisions are reached. (%)¹

Table 7b: Domestic bureaucracies:¹

1. These tables involve officials attaching fairly much weight, or more to the government institutions listed above.
2. This variable regards officials employed at the agency level within the domestic bureaucracy. This limitation does, however, not involve officials at the permanent representations to the EU.
3. This variable regards officials employed at the ministry level within the domestic bureaucracy. This limitation does not involve officials at the permanent representations to the EU.
4. Cf. b)
5. Cf. c)

Consistent with the figures in table 6a and 6b, table 7a and 7b show that the relative importance attached to different government institutions reflect diverse institutional embeddedness' within EU committees. This difference, however, materializes most significantly on the importance assigned to government institutions inter-sectorally. Intra-sectorally, modes of assigning weigh to government institutions seem not significantly affected by institutional dynamics within EU committees. Moreover, officials solely participating within EC's seem to attach greater importance towards government institutions within their own policy sector, and considerably less importance to institutions in other policy fields. Still, officials also participating within WP's tend to pay considerable heed towards i.e. `own' ministry and `own' domestic agencies (cf. table 7b).

When interpreting the figures presented in the above section, one should pay more heed towards tables 1, 3, 4a, 5a, 6a and 7a, than towards tables 2, 4b, 5b, 6b and 7b. This owes to the fact that the dependent variables in the last five tables do not distinguish between EC's and WP's. The former five tables include this distinction, thus to a larger extent contributing to correspondence between the independent variables (institutional affiliation) and the dependent variables.

The analysis so far seems, thus, to confirm our main hypothesis emphasising that the organizational principles underpinning EU committees affect the actors within them. Most importantly, however, this analysis reveals how diverse institutional embeddedness within the EU system of governance affects decision processes within domestic bureaucracies differently. Still, the above analysis does not provide any understanding of to what extent EU committees contribute to resocialise domestic decision-makers. To this end, the next section aims at analysing the relative impact fuelled by the length and intensity of participation within these committees.

On Length and Intensity of Participation across Levels of Governance.
“Negotiations within EC's and WP's sometimes last for years and take place among almost the same set of delegates...” (Neyer 1998, 159). On average, Danish civil servants have participated for a longer period of time within EU committees compared to their Norwegian and Swedish counterparts. ^[10] Whereas Denmark joined the EU in 1973, Sweden became EEA member in 1994, and full-fledged EU member in 1995. Similarly, Norway has been formally affiliated to the EU, through the EEA agreement, since 1994. On this basis, the data sets have been divided into two parts: on the one hand those officials having participated since before 1994, and on the other hand those officials who entered EU committees after 1994. Based upon this division, table eight reveals significant correlations between modes of co-ordination and length of participation within EU committees. ^[11] In the following analysis, not every correlation is presented. Yet, the general tendencies are highlighted.

Table 8: The impact of length of participation within EU committees. (Pearson's r)¹

*) p .05 **) p .01

1. This table compares officials having entered EU committees for the first time after 1994 (coded 1) and officials having participated since before 1994 (coded 2).
2. This variable, and the next variable, regards officials employed at the ministry level within the domestic bureaucracy. This limitation, however, does not involve officials at the permanent representations to the EU.

Generally, the usage of written instructions and problem notes as co-ordinating tools seem to relate negatively to the sheer length of participation within EC's and WP's: Being senior participants within EU committees decrease the likelihood of using binding, written mandates. On the other hand, those officials having participated within WP's since before 1994 seem to co-ordinate to a lesser extent intra-sectorally compared to officials being newcomers within these committees. Similarly, officials at the permanent representations, having participated for longer periods of time within EC's, seem to have less frequent contacts with the political leadership at the domestic level compared to those having participated for only shorter periods of time. Hence, they are more likely to `go native' as regards contact patterns as they become senior participants. Thus, as expected, when exposed towards different EU committees for longer periods of time, officials become more easily affected by the uppermost principles of organization present within these committees. Hence, length of participation seems to strengthen and further the impact fuelled by the uppermost principle of organization present.

In addition to the effect of individual and institutional seniority within EU institutions, one may add the impact fuelled by the sheer intensity to which domestic officials participate within EU committees: the number of committees to which they attend, the number of formal and informal meetings chaired, the degree of activism as regards giving oral presentations within these meetings, etc. Generally, one may assume intensity to correlate possitively with the degree to which officials are affected by the organizational principles being uppermost within these committees. One important finding is that modes of co-ordination seem most strongly affected by the number of committees to which officials are attending. The other variables listed above do not correlate that strongly with modes of co-ordination. ^[12]

Table 9: Impact of the number of committees in which officials are participating. (Pearson's r)¹

*) p .05 **) p .01

1. This table compares officials having participated within maximum 2 EU committees (coded 1) and officials having participated in more than two EU committees (coded 2).
2. This variable, and the next variable, regards officials employed at the agency level. This limitation does, however, not involve officials within the permanent representations to the EU.
3. This variable, and the next variable, regards officials employed at the ministry level. This limitation does, however, not involve officials within the permanent representations to the EU.

This table clearly indicates how the impact fuelled by principles of organziation within EU committees seems strengthened and furthered by the sheer intensity to which officials are exposed towards these principles: Officials participating within many EC's seem to co-ordinate more rarely inter-sectorally than do officials participating in fewer EC's. Similarly, officials participating within many WP's seem to co-ordinate less intra-sectorally than do officials participating within fewer working parties.

Furthermore, when controlling for the sheer number of formal meetings in which these officials take part, the same tendency is repeated: going beyond table nine, the importance of co-ordinating with the MFA is low amongst those officials taking part in many EC-meetings (4 meetings, or more per year) compared to those participating in few formal EC-meetings (less than 4 meetings per year). When also controlling for the number of informal meetings, to which these officials attends, the same tendency is repeated ones again. Hence, our theoretical propositions seem to be confirmed to a great extent: The length and intensity of being exposed towards certain organizational structures – even though these structures are of secondary nature and embedded within collegial arrangements at the EU level of governance – seems to further and strengthen the impact fuelled by the organizational principles present. This section has shown that, (i) secondary institutional embeddedness' do indeed have impact upon modes of acting amongst the participants, and (ii) the intensity and length of participating at the EU level has independent effect on the extent to which co-ordination behaviour is affected by these secondary institutional embeddedness'. The next section of this article aims at uncovering how the relationships presented so far are filtered and modified by the primary institutional embeddedness' of these officials. The officials analysed so far have in fact multiple institutional memberships at different levels of governance.

On Primary Institutional Embeddedness'

Generally, primary institutional affiliations are likely to affect modes of co-ordination fairly strongly (Egeberg 1999a; Knill and Lehmkuhl 1999). The following section shows how domestic institutional embeddedness' affect modes of policy co-ordination. More precisely, the following section illuminates the independent (controlled) effect of domestic institutional affiliations. Thus, the control-variable in the following section is form of affiliation towards EU committees.

Past research on decision-making behaviour amongst Norwegian civil servants reveal that the ministry-agency nexus may be significant for understanding modes of policy co-ordination (L�greid and Olsen 1984; Christensen and Egeberg 1997). I have argued above that cross-sectoral modes of co-ordination are more likely to be evoked amongst officials at the ministry level than amongst officials at the agency level. Moreover, within the permanent representations to the EU, formally embassies under the auspices of the MFA, systematic attention towards cross-sectoral co-ordination is formally institutionalized. Still, officials within the diplomatic realm of these representations are more likely to enact cross-sectoral modes of co-ordination than are officials within the sectoral realms – the later being delegates from domestic sector ministries. Additionally, I have argued that institutional compatibility across levels of governance is likely to condition these effects to some extent: Owing to institutional compatibility, officials within domestic agencies are likely to be affected by the institutional dynamics within EC's more strongly than are officials at the ministry level, save officials within the diplomatic realms at the permanent representations. Similarly, officials at the ministry level, and especially officials within the diplomatic realms at the permanent representations, are likely to be affected by decision impulses from the WP's to a greater extent than are officials within the sectoral realms of the permanent representations, save officials within domestic agencies. The following table largely confirms the main bulk of these propositions:

Table 10: Impact of primary institutional affiliations. (Pearson's r)¹

*) p .05 **) p .01

1) This table compares officials within the domestic bureaucracy being employed at the ministry level (coded 1) versus the agency level (coded 2). At the permanent representations to the EU this table compares officials from the diplomatic realms (coded 1) versus officials within the sectoral realms (coded 2).

Consistent with the above propositions, table 10 reveals that cross-sectoral modes of co-ordination – especially through the MFA ^[13] – are more frequent at the ministry level than at the agency level. Table 10 also shows that this effect is present under different statistical conditions – that is, when controlling for EU affiliation. Still, the first variable in the above table clearly indicates the dual role expectations juxtaposed by these officials: Owing to contending principles of organization present within both EC's and WP's, officials at the ministry level are expected to act like `government officials' and `expert representatives' within both EC's and WP's. Notwithstanding these conflicting role expectations, table 10 clearly indicates how cross-sectoral modes of co-ordination are strengthened as one move from the agency level towards the ministry level. Similarly, when controlling for the EU committee-affiliation, the impact of domestic institutional affiliation turns out significantly negative on whether positions are “based on professional expertice”. Consistent with these observations, officials within the diplomatic realms at the permanent representations to the EU evoke cross-sectoral modes of co-ordination more frequently than do their colleagues within the sectoral realms at the permanent representations. Moreover, the above table identifies positive relationships between domestic institutional affiliations and modes of co-ordination, notwithstanding diverse institutional embeddedness' of civil servants at the EU level. The diverse co-ordinating dynamics fuelled by different EU committees seem to be filtered and modified to some extent by primary institutional affiliations. Still, the last figure in table 10, regarding whether instructions, problem notes and frame notes govern the positions to be followed within EU committee-meetings, is primarily reflecting the EU-affiliation of these officials and not domestic institutional affiliations (r. - 46**). Moreover, effects stemming from institutional compatibility are not clearly reflected in the data: When controlling for European Union affiliation, the sign of the relationship between ministry- and agency-affiliation on the one hand, and co-ordination behaviour on the other, remains the same. ^[14] Support for the compatibility thesis would require significant co-variation between (i) EU affiliation, and (ii) domestic institutional affiliations. Such co-variation is not revealed in the data (cf. bellow). Hence, some of the observations in the first part of this article, revealing how different committees at the EU level affect co-ordination behaviour differently, may be spurious to some extent. Still, it is neither possible nor necessary to test for spuriousness in the current analysis. It is not possible because, in the questionnaire, the EU affiliation-variable is merged with each dependent variable. ^[15] Secondly, testing the relative explanatory power of form of EU-affiliation on the one hand, and domestic institutional affiliations on the other is not necessary because these independent variables are not strongly inter-correlated (the ministry-agency variable correlate only r -.05 with EC-participation, while the ministry-agency variable correlates r .16* with WP-participation). This may indicate that the two independent variables of this study have independent explanatory potential on domestic co-ordination behaviour.

When controlling for rank, the same tendency as in table 10 is repeated: Officials in top rank positions seem to engage in cross-sectoral modes of co-ordination more frequently than do officials in medium rank positions, independent of their diverse European institutional embeddedness'. Additionally, one may expect officials having much behavioural discretion to be more easily affected by their participation within EU committees than officials who's behaviour are highly regulated by formal and informal rules and practices. The following table indicates such relations:

Table 11: Impact of the degree of discretion available to the officials. (Pearson's r)^a

*) p .05 **) p .01

1. This table compares officials who's behaviour are perceived of as being fairly discretionary or more (coded 1) versus officials who's behaviour are perceived of as governed by rules and norms to a fairly great extent, or more (coded 2).

Clearly, officials perceiving of their behaviour to be fairly regulated by norms and rules report their `positions' to be clear. These officials also seem to co-ordinate more frequently – both intra- and inter-sectorally – than do officials having `much' behavioural discretion. Conversely, officials, perceiving of their behaviour as discretionary `often choose which positions to follow'. Officials at the EU representations report that rule following accompany high awareness towards the interests of their home country. Officials reporting their decision behaviour to be governed by rules and practises to a fairly great extent, or more, also tend to co-ordinate actively with the MFA compared to officials who's behaviour being perceived of as fairly discretionary, or more. These tables, thus, indicate the major role played by domestic government institutions as to mould co-ordination behaviour amongst domestic government officials. Moreover, this effect is also significant when controlling statistically for EU committee-affiliation.

Moreover, government officials are highly pre-socialized even before entering formal positions within ministries and agencies. I have argued above that prior professional training may continue to affect modes of acting also after being assigned formal positions within highly formalized organizations like the government apparatus. Table 12 takes into consideration prior professional training amongst civil servants – be it, law, economy, different social sciences or more technical disciplines like biology, physics, etc. Relating to the sectoral-territorial dimension of co-ordination, in the following table are law and social sciences perceived of as representing cross-sectoral disciplines (coded 1), whereas economy and technical disciplines are seen as representing more sectoral professions (coded 2): ^[16]

Table 12: Impact of formal education. (Pearson's r)

*) p .05 **) p .01

Table 12 shows, foremost, that educational background and professional affiliations do have impact upon modes of policy co-ordination: Lawyers and social scientists tend to co-ordinate more frequently cross-sectorally, irrespective of the EU committee to which they attend than do economists and technicians. Conversely, economists and technicians tend to a larger extent to co-ordinate intra-sectorally. ^[17] In example, lawyers and social scientists at the permanent representations, participating within EC's, tend to co-ordinate cross-sectorally (applying national `positions', paying heed to signals from the political leadership), whereas economists and technicians, irrespective of which committees they attend, are more geared towards intra-sectoral co-ordination behaviour (contacts with other ministries and domestic agencies, importance attached towards other domestic agencies). Thus, modes of policy co-ordination tend to vary according to educational background as much as according to committee-affiliation. However, the upper half of table twelve also gives support to a compatibility thesis: Officials participating within WP's tend to enact cross-sectoral modes of co-ordination more extensively if their educational background are compatible with the organizational principles within the WP's – that is, being cross-sectoral in nature (law and social sciences).

Since we are left with some statistically significant bivariate relationships, among which some of the independent variables are related, multiple regression analysis are provided in an effort to unveil spuriousness. I have shown above that the relative explanatory potential of EU committee affiliation is not necessary to unravel because this variable do not inter-correlate with domestic institutional variables (e.g. the ministry-agency variable). Still, in table 13-19 we measure the relative explanatory potential of (i) intensity and length of participation within EU committees, and (ii) primary institutional affiliations.

Table 13: Factors related to the following assertion: “I have clear instructions as to which position to follow”. Pearson's r and multivariate regression (beta). Officials from the capital, attending EC's.

*) p .05 **) p .01

Table 14: Factors related to the importance assigned to the Foreign Ministry. Pearson's r and multivariate regression (beta). Officials from the capital, attending WP's.

*) p .05 **) p .01

Table 15: Factors related to contacts with the domestic political leadership. Pearson's r and multivariate regression (beta). Officials from the capital, attending WP's.

*) p .05 **) p .01

Table 16: Factors related to contacts with other ministries. Pearson's r and multivariate regression (beta). Officials from the capital, attending WP's.

*) p .05 **) p .01

Table 17: Factors related to the importance assigned to the domestic political leadership. Pearson's r and multiple regression (beta). Officials from the capital, attending EC's.

*) p .05 **) p .01

Table 18: Factors related to the following assertion: “I have to co-ordinate with the Ministry of Foreign Affairs, or with other central co-ordinating units”. Pearson's r and multivariate regression (beta). Officials from the capital, attending WP's.

*) p .05 **) p .01

Table 19: Factors related to the following assertion: “My position has been co-ordinated with all relevant ministries”. Pearson's r and multivariate regression (beta). Officials from the capital, attending EC's.

*) p .05 **) p .01

Firstly, comparing the explanatory power of the number of committees to which officials attend on the one hand, and the degree of discretion available at the domestic level on the other, they turn out equally significant. Being the only variables indicating the relative impact of primary and secondary institutional affiliations, the tables 13 and 14 reveal equal explanatory value. Still, the relative explanatory potential of the EC-WP variable is not identified in the current study (see above). Going to tables 15-19, studying the relative explanatory potential of different primary institutional affiliations, the ministry-agency nexus turns out most important, closely followed by the degree of behavioural discretion. Formal education is shown to unveil more spuriousness than the former variables.

Conclusion

Studies of public administration have contributed only scarcely to our understanding of how different institutional affiliations may affect decision behaviour (Egeberg 1999b). This study suggests ways of filling this gap. The current study goes beyond the nation state as the only locus of analysing institutional dynamics, applying a multilevel approach.

The initial argument addressed within this study emphasises primary institutional affiliations as vital cues for affecting decision behaviour within government organizations. However, this analysis also identifies conditions under which secondary institutional affiliations at the EU level of governance may affect domestic co-ordination processes. I argue, basically, that under conditions of (i) institutional compatibility, and (ii) intensive and protracted cross-level participation, may domestic co-ordination processes reflect secondary institutional embeddedness' amongst domestic civil servants. Moreover, I argue that different organizational principles – domestically as well as within EU committees – are likely to affect modes of co-ordination differently.

The empirical analysis supports these arguments to a considerable extent: Regarding principles of organization, inter-sectoral modes of co-ordination are evoked more frequently amongst officials participating within Council WP's than amongst officials attending Commission EC's. Moreover, these differences between EC's and WP's are even furthered and strengthened subsequent to officials participating intensively and for protracted periods of time within EU committees. However, notwithstanding these observations, this study also indicates the pivotal role played by domestic institutions as to affect co-ordination behaviour amongst civil servants. The current study reveals that domestic institutional affiliations have an independent casual impact upon domestic co-ordination processes, even when statistically controlling for EU committee-affiliation. When controlling for the EC-WP embeddedness, Pearson's r tends to vary according to the domestic institutional affiliations amongst the officials. In the same vein, this observation goes largely counter to the institutional compatibility thesis. The empirical support of the compatibility thesis is, thus, rather weak: Variations in the strength of some relationships are observed. Still, the very sign of the relationships are equal along the EC-WP axis. The multiple regression analysis, however, indicate that the relative explanatory value of intensity and length of participation within EU committees on the one hand, and domestic institutional affiliations on the other, are fairly equal. Thus, despite providing clear indications of the transcendence of intergovernmentalism, the current analysis also shows how “responses to integration requirements are filtered – and even conditioned – by a prior state of affairs” at the domestic level of governance (Spanou 1998:469).

References

Anchar, D. 1993. “Comparative Research in the Nordic Countries: Overcoming Ethnocentrism?”, Scandinavian Political Studies 16, 2, 107-126.

Bartolini, S. 1997. “Exit Options, Boundary Building, Political Structuring”, unpublished manuscript, EUI.

Beyers, J. 1998. “How Supranational is Supranationalism? National and European Socialization of Negotiators in the Council of Ministers”, Acta Sociologica 33, 4, 378-408.

Bulmer, S. J. 1993. “The Governance of the European Union: A New Institutional Approach”, Journal of Public Policy 13, 351-380.

Bulmer, S. J. & Burch, M. 1998. “Organizing for Europe: Whitehall, The British State and European Union”, Public Administration 76, 601-628.

Burnham, J. & Maor, M. 1994. “Converging Administrative Systems. Recruitment and Training in EC Member-States”, LSE Working paper.

Christensen, T. & Egeberg, M. 1997. “Sentraladministrasjonen – en oversikt over trekk vedDepartementer og direktorater”, in Christensen, T. & Egeberg. M., eds., Forvaltningskunnskap. Oslo: TANO Aschehoug.

Coombes, D. 1970. Politics and Bureaucracy in the European Community. A Portrait of the Commission of the E.E.C.. London: George Allen and Unwin Ltd.

Coparaso, J., Cowles, M. G. & Risse, T. 1999. “Europeanization and Domestic Change”, unpublished manuscript, University of Washington, Georgetown University and European University Institute.

Cram, L. 1997. Policy-Making in the European Union: Conceptual Lenses and the Integration Process. London: Routledge.

Cyert, R. M. & March, J. G. 1992.A behavioral theory of the firm. Second edition. Cambridge: Blackwell.

Dehousse, R. 1997. “European Integration and the Nation-State”, in Rhodes, M. , Heywood, P. & Wright, V., eds., Development in West European Politics. Houndmills: Macmillan Press.

Dimitrakopoulos, D. 1995. “The Development of The Policy Formulation Process on ECAffairs in Greece: Political and Administrative Considerations”, paper presented at the second Pan-European Conference on International Relations, Paris, 13-16 September, 1995.

Dosenrode, S. Z. 1993. “Den optimale minimall�sning: Danmarks administrative tilpasning til EF”, Nordisk Administrativt Tidsskrift 74, 4, 454-465.

Egeberg, M. 1999a. “Transcending intergovernmentalism? Identity and role perceptions of National officials in EU decision-making”, Journal of European Public Policy 6, 3: 456-474.

Egeberg, M. 1999b. “The impact of bureaucratic structure on policy-making”, Public Administration 77, 1, 155-170.

Egeberg, M. & Trondal, J. 1997. “Innenriksforvaltningens og den offentlige politikkensInternasjonalisering”, in Christensen, T. & Egeberg, M., eds., Forvaltningskunnskap. Oslo: Tano Aschehoug.

Egeberg, M. & Trondal, J. 1999. “Differentiated integration in Europe: the case of the EEA country Norway”, Journal of Common Market Studies 37, 1, 133-142.

Gulick, L. 1937 “Notes on The Theory of Organizations. With Special References to Government in The United States”, in Gulick, L. & Urwick, D., eds., Papers on the Science of Administration. New York: Institute of Public Administration, Columbia University.

Hammond, T. H. 1986. “Agenda Control, Organizational Structure, and Bureaucratic Politics”, American Journal of Political Science 30, 1, 379-420.

Hanf, K. & Soetendorp, B. eds., 1998. Adapting to European Integration. Small States and The European Union. London: Longman.

Harmsen, R. 1999. “The Europeanization of National Administrations: A Comparative Study of France and the Netherlands”, Governance 12, 1, 81-113.

Haverland, M. 1999. “National Adaptations to European Integration: The Importance of Institutional Veto points”, paper presented at the Sixth Biennial International Conference of the ECSA, Pittsburgh, June 2-5, 1999.

Hayes-Renshaw, F. & Wallace, H. 1997. The Council of Ministers. New York: St. Martin's Press.

Hix, S. 1998. “The study of the European Union II: the `new governance' agenda and its Rivals”, Journal of European Public Policy 5, 1, 38-65.

Hopkins, R. F. 1976. “The international role of “domestic” bureaucracy”, International Organzation 30, 3, 405-432.

Institut f�r Europ�ische Politik 1989. Comitology – Characteristics, Performance and Options, Preliminary Final Report, Bonn.

Jachtenfuchs, M. 1997. “Conceptualizing European Governence”, in J�rgensen, K. J., ed., Reflective Approaches to European Governance. Houndmills: Macmillan Press.

Jacobsson, B. 1999. “Europeiseringen och statens omvandling”, in Goldmann, K. et al., eds., Politikens internationalisering. Lund: Studentlitteratur.

Joerges, C. & Neyer, J. 1997. “From Intergovernmental Bargaining to Deliberative Political Processes: The Constitutionalisation of Comitology”, European Law Journal 3, 3, 273-299.

Johnson, G. 1987. Strategic Change and the Management Process. New York: Bail Blackwell.

J�nsson, C. et al. 1998. “Negotiations in Networks in the European Union”, International Negotiations 3: 319-344.

Knill, C. & Lehmkuhl, D. 1999. “How Europe Matters. Different Mechanisms of Europeanization”, European Integration Online Papers (EIOP) 3, 7.

Knill, C. & Lenschaw, A. 1998. “Coping with Europe: the impact of British and German Administrations on the implementation of EU environmental policy”, Journal of European Public Policy 5, 4, 595-614.

Lindberg, L. N. 1963. The Political Dynamics of European Economic Integration. Stanford: Stanford University Press.

L�greid, P. 1999. “Forvaltningens Europatilpaning i Norden”, LOS-senter Notat, 99/31.

L�greid, P. & Olsen, J. P. 1984. “Top civil servants in Norway: key players - on different Teams”, in Suleiman, E. N., ed., Bureaucrats and policy-making. New York: Holmes & Meier.

March, J. G. 1994. A Primer on Decision Making. How Decisions Happen. New York: The Free Press.

March, J. G. & Olsen, J. P. 1989. Rediscovering Institutions. The Organizational Basis of Politics. New York: The Free Press.

March, J. G. & Olsen, J. P. 1995. Democratic Governance. New York: The Free Press.

March, J. G. & Olsen, J. P. 1998. “The Institutional Dynamics of International Political Orders”, International Organizations 52, 4, 943-969.

March, J. G. & Simon, H. A. 1993. Organizations. Second Edition. Cambridge: Blackwell.

Markus, H. & Zajonc, R. B. 1985. “The Cognitive Perspective in Social Psychology”, in Lindzey, G. & Aronsen, E., eds., Handbook of Social Psychology. Volume I: Theory and Method. New York: Random House.

Metcalfe, L. 1994. “International policy co-ordination and public management reform”, International Review of Administrative Sciences 60, 271-290.

Nedergaard, P. 1994. Organiseringen af Den europ�iske Union. K�benhavn: Handelsh�jskolens Forlag.

Neyer, J. 1998. “The Standing Committee for Foodstuffs: Arguing and Bargaining in Comitology”, in Van Schendelen, M. P. C. M., ed., EU Committees as Influential Policymakers. Aldershot: Ashgate.

Neyer, J. 1999. “Justifying Comitology: The Promise of Deliberation”, paper presented at the ECSA Sixth Biennial International Conference, June 2-5, 1999, Pittsburgh, Pennsylvania.

Nkomo, S. M. & Cox Jr., T. 1996. “Diverse Identities in Organizations”, in Clegg, S. R., Hardy, C. & Nord, W. R., eds., Handbook of Organization Studies. London: SAGE.

Olsen, J. P. 1988. “Representativitet og politisk organisering”, in Olsen, J. P., Statsstyre og Institusjonsutforming. Oslo: Universitetsforlaget.

Olsen, J. P. 1998. “Institutional Design in Democratic Contexts”, in Brunsson, N. & Olsen, J. P., eds., Organizing Organizations. Oslo: Fagbokforlaget.

Risse-Kappen, T. 1996. “Exploring the Nature of the Beast: International Relations Theory And Comparative Analysis Meet the European Union”, Journal of Common Market Studies 34, 1, 53-80.

Rometsch, D. & Wessels, W. eds., 1996. The European Union and member states. Towards institutional fusion?. Manchester: Manchester University Press.

Sbragia, S. M. 1993. “The European Community: A Balancing Act”, Publius 27, 3, 23-38.

Sch�fer, G. F. 1996. “Committees in the EC Policy Process: A First Step Towards Developing a Conceptual Framework”, in Pedler, R. H. & Schaefer, G. F., eds., Shaping European Law and Policy. The Role of Committees and Comitology in the Political Process. Maastricht: EIPA.

Scott, W. R. 1995. Institutions and Organizations. Thousand Oaks: SAGE Publications.

Sev�n, G. 1996. “Organizational Imitation in Identity Transformation”, in Czarniawska, B. & Sev�n. G., eds., Translating Organizational Change. Berlin: de Gruyter.

Siedentopf, H. & Ziller, J. 1998. Making European Policies Work. The Implementation ofCommunity Legislation in the Member States. Volume I: Comparative Syntheses. Bruylant: SAGE.Simon, H. 1997. Administrative Behavior. Fourth Edition. New York: The Free Press.

SOU 1996:6. Ett �r med EU. Svenska statstj�nstem�ns erfarenheter av arbete i EU. Stockholm.

Spanou, C. 1998. “European Integration in Administrative Terms: A Framework for Analysis And the Greek Case”, Journal of European Public Policy 5, 3, 467-484.

Spence, D 1993. “The Role of the National Civil Service in European Lobbying: The British Case”, in Mazey, S. & Richardson, J., eds., Lobbying in the European Community. Oxford: Oxford University Press.

Statskonsult 1999:6. Norsk deltakelse i EU-komiteer. En oversikt over trekk ved Forvaltningens deltakelse i komiteer og ekspertgrupper under Europakommisjonen. Oslo.

Statskontoret 1996:7. EU-medlemskapets effekter p� svensk statsf�rvaltning. –Samordning, organisation och arbetsformer i statsf�rvaltningens EU-arbete. Stockholm.

Stone Sweet, A. & Sandholtz, W. 1998. “Integration, Supranational Governance, and the Institutionalization of the European Polity”, in Sandholtz, W. & Stone Sweet, A., eds., European Integration and Supranational Governance. New York: Oxford University Press.

Sundstr�m, G. 1999. Att tala med en r�st. En studie av hur EU-medlemskapet p�verkar Samordningen inom Regeringskansliet. Thesis, department of political science, University of Stockholm

Tenbrunsel, A. E. et al. 1996. “Cognitions in Organizations”, in Clegg, S. R., Hardy, C. & Nord, W. R., eds., Handbook of Organizational Studies. London: SAGE.

Trondal, J. 1996. “Tilknytningsformer til EU og nasjonale samordningsprosesser. En studie av norske og danske departementer”, ARENA working paper 15.

Trondal, J. 1998. “Byr�kratisk integrasjon: EU-kommisjonens ekspertkomiteer somtransformative enheter”, Nordisk administrativt tidsskrift 79, 3, 274-301.

Trondal, J. 1999a. “Seizing a middle ground between liberal intergovernmentalism and neo-Functionalism”, unpublished paper, Department of political science, University of Oslo.

Trondal, J. 1999b. “On Administrative Integration across Levels of Governance: A Theoretical Account on Principles of Organizations”, unpublished paper, Department of political science, University of Oslo.

Trondal, J. 1999c. “Towards theoretical robustness: On Refutation, Causality and the Comparative Method”, unpublished paper, Department of political science, University of Oslo.

Trondal, J. 1999d. “Mellom politisk styring og faglig uavhengighet: En studie av norske direktoraters europeisering”, Nordiske organisasjonsstudier 1, 3 (forthcoming).

Van Schendelen, M. P. C. M. 1996. “EC Committees: Influence Counts More Than LegalPowers”, in Pedler, R. H. and Sch�fer, G. F. (eds.) Shaping European Law and Policy. The Role of Committees and Comitology in the Political Process, Maastricht: EIPA.

Van Schendelen, M. P. C. M. and Pedler, R. H. 1998. “Are EU Committees Influential?”, in Van Schendelen, M. P. C. M. (ed.) EU Committees as Influential Policymakers, Aldershot: Ashgate.

Weaver, R. K. & Rockman, B. A. 1993. Do Institutions Matter? Government Capabilities in the United States and Abroad. Washington D.C.: The Brookings Institution.

Wessels, W. 1998. “Comitology: fusion in action. Politico-administrative trends in the EU System”, Journal of European Public Policy 5, 2, 209-234.

Wright, V. 1996. “The national co-ordination of European policy-making: Negotiating theQuagmire”, in Richardson, J., ed., European Union. Power and Policy-making. London: Routledge.

Zuna, H. R. 1998. Demografi og beslutninger. Om sammenhengen mellom utvalgte Bakgrunnstrekk ved tjenestemenn i norske departementer og deres beslutningsadferd. Thesis, Department of Political Science, University of Oslo.

Footnotes

* This article is financially supported by the ARENA programme (The Norwegian Research Council), The University of Oslo. An earlier version of this paper was presented at the 12th NOPSA Conference, Uppsala, August 19-21, 1999. The author is indebted to Shrin Ahlb�ck, Svein Andersen, Jan Beyers, Jeffrey T. Checkel, Morten Egeberg, Bengt Jcobsson, Ann-Cathrine Jungar, Tobias Krantz, Per L�greid, G�ran Sundstr�m and Frode Veggeland for valuable comments. Also thanks to Knut A. Christoffersen for technical assistance.

** Department of Political Science, University of Oslo, P.O. Box 1097 Blindern, 0317 Oslo, Norway. E-mail: jarle.trondal@stv.uio.no .Tel.: +47 22 85 51 66, fax: +47 22 85 44 11.

^[1] Inter-sectoral modes of co-ordination, I argue, involve linkage processes across policy sectors as far as actors, problems, solutions, consequences, role conceptions and identities are concerned. In the current study, inter-sectoral co-ordination resembles horizontal inter-ministerial co-ordination processes, whereas intra-sectoral co-ordination resembles both horizontal co-ordination processes within each government institution, and vertical co-ordination processes within each government institution and between the ministry level and the agency level. ^[1] Moreover, my definition of co-ordination is also two-dimensional as regards techniques applied: Co-ordination may involve (i) pro-active processes geared towards outlining written and clear-cut mandates and instructions, or (i) re-active processes and anticipated reaction, not geared towards outlining any written instructions. Whereas the first process tends to foster formally binding mandates, the second process tends to produce unclear mandates. Whereas process towards producing imperative mandates is based upon notions of institutions as coupled, sectorally co-ordinated systems of governance, the notion of unclear mandates rests upon notions of institutions as uncoupled, segmented and inter-sectorally fragmented government systems (Olsen 1988:162-170).

^[2] The current article analyses how different organizational principles – as embedded within the EU Commission and the Council of Ministers – affects modes of co-ordination. The term `Europeanization' resembles in the current study processes whereby domestic co-ordination processes are affected significantly by institutional dynamics within Commission expert committees and/or Council working parties. The function of the Commission and the Council are differently, being located at different phases of the decision-making cycles of the EU machinery. Still, the current article does not theorize on decision-phases. Organizational principles are put to the fore in this study partly because organizational principles are under-researches in general, and partially because on may derive testable hypotheses from organizational principles more easily than from decision-phases.

^[3] Whereas a huge body of literature conceives of the EU system as a novel and partially `mysterious' polity - not resembling anything seen before, I argue that the EU system to a great extent reflects the structures of nation state polities. I thus reason against arguments stressing that “the European Community is a political entity that does not fit into any accepted category of government” (Sbragia 1993, 24) (my emphasis). My argument does not ignore idiosyncrasies of the EU system, being supra-national, multi-national, multi-linguistic and multi-level. The basic arguments laid out here, however, do stress the similarities of the EU polity and other polities. Still, due to this polity being described as unique, novel and `mysterious' (Bartolini 1997), and faced with an assumed lack of appropriate concepts and categories for the understanding of this polity (Coombes 1970, 101; Jachtenfuchs 1997, 40), one vital step in this enquiry must be to carve this polity into empirically meaningful and theoretical fruitful categories.

^[4] Moreover, the existence of formal veto-points may contribute to partially contending principles of organization being biased by one uppermost principle: Veto-points “refer to all stages in the decision-making process at which agreement is required for a policy change” (Haverland 1999, 4). Veto-points, thus, contribute to lock decision-makers into decision situations through processes of anticipating future votes. Referring to EU committees under scrutiny within this study, Commission expert committees only rarely makes formal votes. Council working parties, on the other hand, vote more frequently (Hayes-Renshaw and Wallace1997; Sch�fer 1996).

^[5] Another hypothesis could be that length of service within EU committees reduces the need for co-ordinated instructions due to increased knowledge and `feeling' as regards appropriate solutions and positions. Still, this hypothesis is empirically supported to a lesser extent than the above hypothesis (cf. table 8).

^[6] Due to small N, and owing to few officials participating solely within expert committees, it is not possible to make this distinction within the data set from the EU representations. This argument is valid for all tables presented within this paper.

^[7] Still, 38 (53%) of the Swedish officials, and 6 (14%) of the Danish officials participate solely within EC's. Hence, the distinction made in table two does not correspond perfectly to the distinction between different countries. Moreover, 32 (44%) of the Swedish officials, and 35 (80%) of the Danish officials participate in both EC's and WP's. Finally, only 2 (3%) of the Swedish officials, and 1 (3%) of the Danish officials participate solely within WP's. Thus, to sum up, amongst the Danish and Swedish officials 38% (44) participate solely within EC's, 58% (67) participate within EC's and WP's, while 3% (3) participate solely within WP's.

^[8] 40 percentages of the officials solely participating within expert committees report their decision behaviour to be governed by rules and practises to a fairly great extent, or more. On the other hand, 52 percentages of those officials participating both within EC's and WP's report their decision behaviour to be governed by rules and practises to a fairly great extent, or more.

^[9] Time pressure is reported as one additional hindrance for co-ordination (source: interviews).

^[10] In our sample from the `home administration', 51 percent of the Danish officials entered EC's before 1990, whereas only 5 percentages of the Norwegian officials and 2 percentages of the Swedish officials did enter these committees before 1990. Regarding participation within WP's, all of the Swedish officials entered these committees for the first time after 1994 (when Sweden became EEA-member), whereas most of the Danish officials entered long before 1994.

^[11] Owing to systematic selection of data, significance does not tell us anything about any universe. On the contrary, significance may tell us something about the robustness of the relationships tested. Moreover, owing to the low N, and aiming at studying only fairly robust relationships, Pearsons r not being significant at 95 % level, or being lower than .40 are not shown in the tables presented here. On this basis, the likelihood for random errors is reduced.

^[12] Those officials participating in many EU committees also seem to take part in many formal meetings (Pearson's r = .23**, N = 159). Moreover, those taking part in many committees are also generally senior participants (Pearson's r = .34**, N = 159).

^[13] Pearson's r = .32**, N = 148 within the domestic bureaucracy.

^[14] Still, the strength of some relationships differs along the EC-WP axis, as expected from the compatibility thesis. In example, contacts with other ministries are conducted more frequently amongst officials at the ministry level attending WP's compared to ministry-officials attending EC's (variable 3). The same tendency regards officials attaching importance to other ministries (variable 5), and having contacts with the MFA (variable 7). Finally, officials at the ministry level attending WP's report more frequently to co-ordinate with all relevant ministries than do officials at the ministry level attending EC's. These observations, thus, support the compatibility thesis somewhat. Still, the sign of the relationships are equal along the EC-WP nexus, thus, going counter to the compatibility thesis.

^[15] A test of the independent, non-spurious, explanatory status of EU committee-affiliation would require cross-tabulations controlling for at least ministry-agency affiliation. This test is not done here because is would render only few units in some of the cells (especially amongst ministry-officials).

^[16] Linking law and social sciences into the same variable exhibit one major problem: Whereas the law education is professional in character, which may tend to increase the likelihood for strong professional allegiances, most social science disciplines are cross-disciplinary. Thus, the level of professional allegiance may differ between these disciplines. Economy and technical disciplines, by contrast, both rest upon professional educations.

^[17] In the current sample, technicians loom largest – both within EC's and WP's. Still, whereas technicians and economists dominate as delegates from domestic agencies and ministries (82%), lawyers and social scientists are more frequently representatives at the permanent representations (66%). However, officials educated in law and social sciences represents the ministry level (30%) more frequently than the agency level (7%). At the permanent representations to the EU, lawyers and social scientists tend to represent the diplomatic realm (71%) more frequently than different sectoral realms (67%).

[Date of publication in the ARENA Working Paper series: 15.11.1999]