Analysing the Links between National Capitals and Brussels in EU Foreign Policy

Abstract

The article contributes to the study of EU foreign policy decision-making processes by analysing the links between national officials working in the committees of the Council of the EU and their capitals. Through an original dataset of 138 questionnaires (and 20 interviews) with national representatives, it explores the micro-foundations of the formulation of EU foreign policy. It first shows how, even in this most intergovernmental field, diplomats in Brussels play a very important role in the policy process: only 30 per cent claim to always have a mandate and half state that they do not feel constrained by their capital. Next, it reveals that if (larger) member states attempt to retain control of CFSP/CSDP negotiations, the effective discretion/autonomy these officials enjoy depends on the experience accumulated in the decision-making process, and knowledge of the (formal and informal) links between Brussels and the home department.

In recent years the study of the committees involved in the decision-making process of the European Union (EU) has proved a very fruitful area of research. Although their contribution to the work of the Council of the EU (hereafter, the Council) may have been slightly exaggerated, ¹ these committees are said to settle approximately two-thirds of all the issues on the Council agenda, and they have been described as the ‘backbone of the European system of integration’ (Hayes-Renshaw and Wallace 1997: 98).

This article contributes to the debates on the committees of the Council by analysing an overlooked aspect of their action (Häge 2007; Panke 2010) – the links between national officials working in these committees and their capitals. In particular, it explores the micro-foundations of national policy – how (and by whom) national positions are prepared. How often do civil servants in Brussels have a mandate? To what extent do they enjoy leeway in their negotiations? What factors explain these dynamics? In spite of some empirical findings that increasingly reveal the important role played by the officials in Brussels (Egeberg et al. 2003; Kassim et al. 2001; Lewis 1998), intergovernmental analysis of the Council has generally ignored this relationship. Somewhat surprisingly, the literatures on socialisation and on the ‘deliberative turn’ in EU studies also have failed to take this relationship into account, and bracketed the domestic level. If EU committee members are socialised to a certain extent, what impact do these (partial) socialisation processes have on the negotiations of the Council? In other words, does a supranational identity held by individual national officials eventually translate into behavioural changes in states, and, if so, how (Zürn and Checkel 2005)? Similarly, how normative suasion can occur within these groups is unclear and often rests on implicit assumptions. How is it possible to argue that Council committees are arenas where diplomats ‘reason, discuss, deliberate and persuade’ (Niemann 2006: 468) without exploring the links between them and their capitals? One may even argue that deliberation between representatives is a contradiction in terms (Neyer 2006). Finally, investigating the control exercised by capitals over their delegates in Brussels is useful from a normative perspective. The growing importance of committees within the Council and the EU raised serious concerns about the transparency and democratic accountability of the decision-making process (Christiansen and Kirchner 2000; Rhinard 2002). As Häge puts it (2008: 556), the more closely national capitals are involved in the elaboration of the mandate, the less the accountability deficit that exists. ²

Relying on an original dataset of 138 questionnaires and 20 interviews with national officials participating in the committees of the Common Foreign and Security Policy (CFSP) and Common Security and Defence Policy (CSDP), this article attempts to partially fill this gap, while also contributing to our knowledge of how EU foreign policy is made. The ‘committee turn’ in EU studies has only partially addressed the area of foreign policy. If some contributions have examined the historical development of CFSP committees (e.g. Duke and Vanhoonacker 2006), other analyses have also revealed that national foreign policies are often defined through social interaction in Brussels, and consensus-building is an essential feature of their negotiations (Howorth 2010; Juncos and Pomorska 2006, 2011; Juncos and Reynolds 2007). This article continues along these lines, by applying the research designs, the methodological instruments and the rigorous testing of hypotheses widely used in the studies of the Council (Quaglia et al. 2008) to the activities of CFSP/CSDP committees.

The article is organised as follows. The first section clarifies and disentangles the European–national nexus in the preparation of the national position. Three dimensions are distinguished: (a) the quantity and quality of the mandate; (b) the relevance of the capital for the activities of diplomats; and (c) the margin of their autonomy. Furthermore, the dataset generated by the research is introduced and methodological issues are considered. The second section presents and discusses the empirical data for the three indicators, which allow us to assess the role of the European and national layers of the national administration in the formulation of the national position. The third section aims at deepening the analysis: the three dimensions are used as dependent variables, and a series of explanatory variables are introduced in order to explain the differences in governmental control over their diplomats in Brussels. Finally, the conclusion summarises the main findings and discusses their implications for wider debates about European integration and EU foreign policy.

Links between Brussels and the National Capital

The article explores the links between officials in Brussels and the national capitals, to determine the characteristics of their coordination, and to assess their power relationship. As mentioned, studies of the Council – which have generated a remarkable amount of data – have rarely investigated these dynamics. Even when this is done, the context remains that of socialisation (Beyers 2005; Beyers and Trondal 2004); or the aim is essentially comparative, showing the many faces of EU committee governance (Egeberg et al. 2003).

At the same time, the literature on EU foreign policy has also paid scarce attention to the European–domestic nexus. Broadly speaking, there are two different perspectives (ideal-types) on the CFSP/CSDP decision-making process. The first approach considers the CFSP as an intergovernmental project (Eliassen 1998; Jones 2007), where member states attempt to promote their own interests (Allen 1996), national concerns take precedence, and a ‘logic of diversity’ prevents relevant agreements being reached (Hoffmann 1966; Krotz 2009). The interplay between the European and domestic level is omitted, and the national capital is – implicitly more than explicitly – assumed to formulate the position, which is then defended and maximised in Brussels by national diplomats. A second, institutional, perspective identifies consensus and problem-solving as key features of EU foreign policy, where solutions to policy problems are found in a collective definition of the issues, and references to European identity and interests and the use of peer pressure are common attributes (Smith 2004). The major role is said to be played by national officials in Brussels, who, socialised to European ideas and values, ‘see themselves … as policy arbiters’ (Tonra 2001: 12) and form a ‘diplomatic republic of Europe’ (Jørgensen 1997). EU foreign policy is increasingly ‘Brusselised’ and formulated by national diplomats in the Belgian capital (Juncos and Pomorska 2011). The interaction between Brussels and the capitals, however, has not been analysed (see Juncos and Pomorska 2006 for an exception). Furthermore, these accounts are mostly based on memoirs, unsystematic interviews and authorial insights.

In other words, the micro-foundations of the formulation of EU (foreign) policy are still largely unexplored. This article intends to analyse some aspects of the skeleton of the vertical coordination between the national and European layers of the national administration, and to provide the literature on committee governance and on the decision-making process of EU foreign policy with evidence at the micro-level.

Three aspects are considered to define the links between Brussels and the capital and to evaluate the practices and logics behind the formulation and negotiation of the national position (Table 1). The first is the quantity and quality of the national mandate. Mandate in this context refers to any formal or informal set of instructions or guidelines from the capital that the European level has to follow, represent and advance in EU negotiations. The characteristics of the mandate can shed light on how national and EU (foreign) policies are made. A mandate which is relatively unclear and not particularly detailed would give diplomats participating in CFSP/CSDP committees greater scope to formulate and represent the national position. Similarly, the absence of a mandate would imply that Brussels diplomats have, at times, a considerable amount of leeway in determining the position and can infuse CFSP/CSDP negotiations with their own perspectives. By contrast, a clearly formulated, detailed and frequent mandate generally ³ reflects a high degree of control exercised by the domestic level over the activities of Brussels diplomats. The coordination of national policies, then, is highly centralised. Most smaller EU states, for instance, are said to have strongly capital-based coordination systems, with little input from Permanent Representations (PRs): Brussels diplomats send documents and information back to the capitals and just ‘wait for their instructions’ (Panke 2010: 773).

table 1. links between national capital and brussels

	Operationalisation	Table
1. Quantity and quality of the mandate	Questionnaire	2,3
2. Capital as source of information	Questionnaire	4
3. Autonomy of the national officials	Questionnaire	4

The second defining feature concerns the source of information when diplomats negotiate. The national level may influence the position to be defended in Brussels well beyond the mandate. Delegates are embedded in their domestic environment and even in the absence of negotiating instructions they can still advance points of view, and use information and insights, elaborated more broadly by the capital. On the other hand, considering the national administration as (only) one of the sources for their activities, and taking into account information coming from the other member states or other actors – such as the European Commission or the High Representative for Foreign Affairs and Security Policy (HR) – reveals a more fluid and dynamic dimension to the CFSP/CSDP negotiations than is often assumed. This could potentially amount to more than a lowest common denominator of member states’ exogenously defined positions, and could mean that they are potentially subject to a more collective definition in Brussels by national officials.

Following these observations, the third element that identifies the links between the capital and the European level is the overall leeway national representatives in Brussels enjoy. Investigating to what extent they feel restricted by the national level in the negotiations they take part in ⁴ sheds further light on the role the domestic and European layers of the national administration play in the formulation of the national position and negotiation of EU foreign and defence policy. Again, an extremely limited (perceived) margin of manoeuvre reveals tight control by the capital and a marginal role for Brussels diplomats. CFSP/CSDP would, in this scenario, still be dictated and controlled by the capitals of member states. During negotiations, diplomats enjoy little strategic leeway while, at best, they are allowed to drift away from the mandate only for tactical reasons, with the explicit approval of the government. Since they are directly accountable to their ministers (and ministries), their potential to act autonomously is limited (Häge 2007: 310). On the other hand, a considerable (perceived) margin of manoeuvre in CFSP/CSDP negotiations would suggest that PRs have acquired a prominent part in EU decision-making processes. The capital has become increasingly dependent on the European level. A process of ‘Europeanisation of time’ may be under way (Ekengren and Sundelius 2004: 119) since the pace of work in the Council is so intense that national representatives often get the necessary documents at the very last minute (Egeberg et al. 2003). The ministry simply does not have the time to process them, and eventually instruct its delegates in Brussels. In this vein, EU foreign and defence policy is increasingly shaped in Brussels by national representatives, more easily permeated by EU perspectives, and potentially formulated according to common definitions of issues and problems.

In order to analyse and test these three dimensions, all ⁵ the national diplomats involved in the CFSP/CSDP decision-making process have been considered and are the population of the research. A closed-ended questionnaire has been administered to the about 30 CFSP and CSDP working groups, Political and Security Committee (PSC), Relex Counsellors, EU Military Committee, etc. The response rate was 36 per cent: 114 questionnaires were received by mail and another 24 were compiled online (for a total number of 138). Data from all 27 member states were collected. In addition, 20 in-depth interviews with national officials were conducted to gain a more in-depth understanding of certain issues and to better interpret the quantitative data. ⁶

Empirical Analysis

Presenting the Data

(a) Quantity and quality of the mandate. The national mandate appears clear: 62.7 per cent assigned a particularly high value (options 5 + 6) and a further 24.8 per cent attributed a score of four. Domestic oversight appears instead more limited in relation to the level of detail: about 40 per cent agreed that the instructions they received were not detailed (options 1 + 2 + 3) and only 7.3 per cent regarded them as extremely detailed (Table 2).

table 2. clearness; detail (%)

	‘The instructions coming from my department are clear’	‘The instructions coming from my department are detailed’
1	3.7	2.2
2	2.9	21.9
3	5.8	16.1
4	24.8	27.0
5	40.1	25.6
6	22.6	7.3
Total	100	100
Notes: Scale from 1 (‘strongly disagree’) to 6 (‘strongly agree’). Because of rounding, some columns -- in this table and in the following ones -- may not sum to 100.

At the same time, if the information flows and contacts between the European and domestic level are intense (71 per cent report that they communicate with their department every day, and a further 25.4 per cent fairly frequently), ⁷ this communication is not one-way: it is highly significant that civil servants do not receive instructions every time they meet to discuss foreign policy issues within the EU. Fewer than one in three respondents state that they always have a national mandate, and 26.3 per cent claim to receive it only a ‘few times’ or ‘never’ (Table 3).

table 3. frequency (%)

	‘Do you have negotiating instructions from your capital?’
1. Yes, always	30.7
2. Often	43.1
3. Few times	24.8
4. Never	1.5
Total	100.0

(b) Relevance of the capital as source of information. A further question explored the importance Brussels diplomats attach to their capital when they negotiate in Council committees. Table 4 shows that the national capital, unsurprisingly, is a very important source of information (the cumulative percentage of the options ‘essential’ and ‘fairly important’ is 90.4 per cent). This is unsurprising given that these officials are required to express the national points of view in those settings.

table 4. capital as source of information; autonomy of national officials (%)

‘How important do you consider your capital as source of information for your negotiations?’		‘To what extent do you feel restricted by a negotiation mandate from your capital?’
1. Essential	44.1	1. Very much	12.4
2. Fairly important	46.3	2. Fairly enough	38.0
3. Little important	8.1	3. Not much, there is enough freedom	49.6
4. Not important at all	1.5	4. There is complete freedom	0.0
Total	100.0	Total	100.0

The option ‘fairly important’, however, was the most frequently selected. The fact that 46.3 per cent chose the second category (and a further 9.6 per cent the third or fourth reply) suggests that, the domestic level notwithstanding, national representatives’ opinions and positions are partly formed in different contexts and as a consequence of different dynamics. For instance, even if the decision-making process of the CFSP/CSDP is centred on the Council, full-timers recognise that the Commission plays an important part in the daily management of the CFSP/CSDP. Nearly 70 per cent (68.7 per cent) consider the role of the Commission ‘crucial’ or ‘important’ for their work; only 7.5 per cent replied saying that this role is negligible. Similar results are recorded regarding the influence of the HR (73.7 per cent and 6 per cent, respectively). ⁸

(c) Autonomy of the national officials. These data indicate that the source of information is not exclusive (the capital). The ideas and positions of other actors can have an impact on the activities of national representatives in Brussels. Margins of manoeuvre thus appear to exist: 49.6 per cent of the interviewees report they do not feel restricted by their capital, and enjoy enough freedom in their negotiations (Table 4). This does not imply that Brussels diplomats play the major part in the CFSP decision-making process. Furthermore, allowing a certain degree of freedom is often a rational strategy for member states to achieve their aims – provided that discretion is used to pursue exogenously given national interests – defined and closely monitored by the capital. What the data, however, reveal is that, at minimum, the control of the national level on PRs is not as strong as some intergovernmentalists may think; they also indicate that these civil servants are able to infuse, and influence, negotiations with their insights and perspectives. Finally, in-depth interviews suggest that, in some circumstances, it is the European branch that determines the negotiating position. Diplomats who work every day in Brussels ‘know the general atmosphere of their own group, and those positions and behaviours that can be accepted and those which are not’. ⁹ They ‘sometimes have to face advice and instructions that are outside of the European logic and also counter-productive’. In these situations, it regularly occurs that ‘colleagues call home to report that the initial position is not realistic’ and that in order to reach a (reasonable) consensus, ‘it is necessary to change it’. ¹⁰ This is more likely to happen when coordination problems within the national administration exist, when the ministries of Defence and Foreign Affairs hold conflicting policy stances, or when the issues are so technical that people at home do not have the necessary expertise or time to analyse them. As a result, around 40 per cent of the respondents indicate that convincing their national administration to modify a negotiating position is a recurrent event; cumulatively, 85.9 per cent acknowledge that it has occurred sometimes in the groups they belong to (Table 5).

table 5. position shifts (%)

	‘Has it ever occurred that any member of your group, disagreeing with the position supported by her own department, acted so as to convince her capital?’
1. Yes, it occurs frequently	39.6
2. Yes, but it occurs rarely	46.3
3. No, it has never occurred	14.2
Total	100

Discussion

All these aspects reveal that the national capital control over CFSP/CSDP full-timers is effective, but not as tight as often assumed. The two levels are interconnected, and the formulation of the national position appears to be a combined activity. How can we interpret this relationship? A strict interpretation of a Principal–Agent (P–A) approach – on which many (intergovernmental) readings of the Council and the CFSP/CSDP rely – does not seem useful in this context. Identifying a clear ‘principal’, who elaborates the national preference, and an ‘agent’, who maximises that preference, is a difficult task (Lewis 1998). The European level, certainly well connected with the domestic bureaucratic structure, nevertheless enjoys a certain margin of (perceived) autonomy and the possibility to advance its own views. This happens in two ways. First, the experience acquired in Brussels is a necessary element to define the content of the national mandate. The data have shown that Brussels diplomats have not been delegated merely the function of representing and defending the position that is prepared by departments in the capital. The information flow is two-way: ‘interaction between the two levels is the key word’, a national official reveals. ¹¹ Only 30 per cent claim to always have a mandate; in some cases it happens that Brussels diplomats write instructions for themselves and then inform the ministry about what they are going to negotiate (Juncos and Pomorska 2011). As a CIVCOM delegate said, ‘instructions from capital are essentially based on proposals sent by the Permanent Representation’. ¹² The subject of the delegation is thus wider and also concerns the elaboration of (part of) the position. Second, Brussels bureaucrats are important players in the negotiation process itself. They know how negotiations are structured and likely to evolve, and what resources are needed in that game. Given their central position and the information asymmetries that emerge between them and their superiors, participants in CFSP/CSDP committees have detailed information on the nature and intensity of other actors’ preferences, and know what is appropriate and possible to achieve in those settings.

These results are largely consistent with the findings of the volume edited by Kassim et al. (2001) on the national coordination of EU policy at the European level. Employing a different methodology – case studies of 11 PRs – and concentrating essentially on the European Community, their analysis demonstrates how PRs participate in the definition of the national position. As a French civil servant said (in Kassim and Peters 2001: 337), ‘I have been surprised, since my arrival [in Brussels], by the degree to which our work involves formulating policy with Paris as opposed to simply executing policies formulated in Paris’. However, if these studies mostly explored information-gathering and de-briefing on the part of officials, this article has further illustrated the degree of discretion the participants in CFSP/CSDP committees enjoy – even occasionally convincing their own capital to change the original position. Persuading the domestic level to accept a better argument raised by another member state would be the most convincing ‘smoking gun’ for a supranational reading of the Council (Lewis 2008).

Not surprisingly, a final query fully confirms the image of a joint decision-making process and blurred boundaries between the national administration and Brussels. Table 6 reveals that 75.6 per cent of the interviewees give a well-balanced evaluation of their power relationship: the links are dynamic and open, and both levels contribute to the formulation of the national policy. Moreover, diplomats who consider the PRs to be the principal actors are more numerous than those who support a strictly intergovernmental image of the Council.

table 6. overall assessment (%)

‘How do you assess the overall power relationships between the officials in Brussels and those in the capital?’
1. It is an open and dynamic relation: both parts contribute to the formulation of the position	75.6
2. It is a basically unilateral relation: the national capital defines the position and Brussels defends it	8.1
3. Brussels plays the major part, as it is at the heart of the CFSP decision-making process	14.1
4. Other	2.2
Total	100.0

Advancing Explanatory Variables

The analysis so far has offered a general representation of the links between capitals and Brussels in EU foreign and defence policy. However, member states develop different institutional arrangements for coordinating European policy. For instance, countries like the United Kingdom, Sweden and France tend to exert tight control on their civil servants, whereas more decentralised coordination systems guarantee a wide margin for manoeuvre for the representatives of Luxembourg, Belgium and Italy (Beyers and Trondal 2004; Kassim and Peters 2001). Even within the same country, ministerial control can vary from issue to issue (Fabbrini and Piattoni 2008).

Consequently, this section introduces and tests some explanations of variation in the experiences of national servants. To evaluate control on the European level, the same three elements that have previously defined the links between the domestic ministry and Brussels – (a) the clearness, detail and frequency of the mandate; (b) the relevance of the capital for diplomats’ activities; and (c) the margin of officials’ autonomy – are used here as outcome variables. First, explanatory elements linked to the institutional context (Working Groups versus PSC; CFSP versus CSDP) are investigated. Next, the analysis statistically tests variables at the micro-level (EU Career; Seniority) and macro-level (Polity; Europeanism; Power).

The Institutional Context: Working Groups versus PSC

A research design widely used by the literature on EU committees examines the activities that take place in those settings through comparative lenses. The sample includes expert groups, Council working groups (CWGs) and comitology committees. This allows researchers to assess the results controlling for contextual factors, as these groups and committees have different degrees of politicisation, network characteristics and rules (Egeberg et al. 2003). If different institutional contexts and affiliations lead to different experiences, ministerial control and bureaucratic leeway are likely to vary in the case of the officials participating in CWGs and of the diplomats sitting in the PSC. Expectations are mixed. On the one hand, CWGs deal with CFSP dossiers in the first stages of the decision-making process, and their participants are often experts in that field. The capital could have fewer opportunities and incentives to control their work. Similarly, senior officials in the PSC are closer to the heart of the decision-making process: national priorities are likely to emerge here with greater force. On the other side, the PSC is a more senior and less transparent body; agreements are taken with minor involvement of external actors, including the capital. PSC representatives are thus expected to enjoy wider margins of manoeuvre.

The picture that emerges from the data ¹³ is relatively clear: oversight by the national capital is stronger in the case of CWGs. All the statistics (but one, where the difference is nevertheless small) lead to the same reading (Table 7). Only 23.8 per cent (26.9 per cent) of the interviewees – compared with 42.3 per cent (40.6 per cent) of CWGs delegates – affirm, for instance, having a mandate when they negotiate in the PSC. Some officials who belong to, or have experience of, both groups confirm that in the PSC ‘there is greater freedom’, and ‘the government hand penetrates less deeply’. Furthermore, PSC officials have a higher level of seniority (ambassadorial rank), which ensures greater autonomy. ¹⁴

table 7. cfsp working groups and psc; cfsp and csdp (%)

	CFSP CWGs		PSC		CFSP		CSDP
	N = 43	N = 27	N = 51	N = 21	N = 94	N = 59	N = 76	N = 54
Instructions are clear^1	95.2	100	80.4	85	87.1	93.3	82.9	87.1
Instructions are detailed^1	52.4	53.8	56.9	55	54.9	56.9	64.5	68.5
I ‘always’ have negotiating instructions^2	40.6	42.3	26.9	23.8	34.4	35.6	27.6	29.7
The negotiating mandate is constraining^3	47.6	53.8	46.2	51.9	46.8	54.2	48.7	51.9
The capital is the ‘essential’ source of information^2	42.9	50	38.5	33.3	42.6	42.4	47.3	48.4
^1Options 4 + 5 + 6 (6-point scale).^2Option 1 (4-point scale).^3Options ‘very much’ and ‘fairly enough’ (4-point scale).

The Institutional Context: CFSP versus CSDP

Another institutional factor that may have an impact on the activities of these committees concerns the different features and rules that characterise the CFSP and CSDP. It may well be that these two regimes demand different coordination and negotiating styles. Only national policy-makers, with their mandate legitimised at the domestic level, can launch a military operation and endanger their own soldiers’ lives. Accordingly, the national hold is likely to be stronger in defence policy. However, the data do not support this claim (Table 7). The statistics show relatively close and/or contrasting values. Instructions are clearer, but less detailed, for CSFP representatives. If CSDP officials pay more attention to the national level, a negotiating mandate is available more frequently in the case of foreign policy. Finally, there are not big differences in the (perceived) margins of freedom between the two groups.

Between Micro and Macro: More Explanatory Variables

The control of the capital over the activities of its officials in Brussels may depend on the individual experiences of the officials themselves: ¹⁵ a greater number of years spent in the CFSP/CSDP can enhance their trust vis-à-vis the capital. Besides, people in PRs can use information and knowledge asymmetries to increase their autonomy (Kassim and Peters 2001: 307). Along similar lines, the same trust and negotiating leeway can be generated by a long overall career in national ministries, foreign embassies and PRs. Junior officials are more likely to be kept on a tight leash by their ministry (Howorth 2010). The hypotheses can be formulated as follows: the longer the career in European affairs (H1: EU Career), the longer the overall career of a national diplomat (H2: Seniority), the less tight the control of the capital on Brussels will be. ¹⁶

On the other hand, some variables at the macro level can be tested to analyse their impact on the domestic oversight of Brussels bureaucrats. First, the level of decentralisation and the existence of a multi-level governance within a country can be positively associated with a greater discretion of national officials (H3: Polity). A more decentralised system of preference formation (with a bigger role assigned to PRs) is more easily accepted in a multi-layered constitutional culture (Beyers and Trondal 2004: 928–29). ¹⁷

Second, states that share a widespread consensus on the benefits of an EU foreign policy tend to rely more on their national servants in Brussels; at the same time, more sceptical countries employ a higher number of officials to better monitor them (Kassim and Peters 2001: 327). To measure these orientations, two proxies are used: first, the attitude of the national elite towards the CFSP/CSDP (H4a: Elite). Research conducted by the Commission in 1996 on top decision-makers’ opinions on EU foreign and defence policy provides the necessary information. These data are available for only 15 countries. Consequently, information on public opinion is used (H4b: Public Opinion) to cover also those countries that joined the EU between 2004 and 2007 (Eurobarometer surveys).

Finally, a last explanatory variable is closely related to the field of international relations and an intergovernmental reading of the CFSP/CSDP. A common (realist) assumption is that the most powerful states have fewer interests in pooling their sovereignty in such a delicate sector, as they are able to conduct independent foreign policies and prefer not to be tied by international/European commitments. As a result, they are expected to give more importance to the control of their own servants in Brussels, whereas smaller states’ representatives are likely to enjoy more leeway in their work (H5: Power). ¹⁸

The five explanatory variables are summarised in Table 8. The three outcome variables (Frequency, Clearness, Detail of the Mandate; Source; Autonomy of the Officials – see Table 1) have a natural ordering, but the distance between levels is unknown. As a result, an ordered probit regression is employed. On each of the outcome variables two different regressions are run: first, with the following explanatory variables (EU Career; Seniority; Public Opinion; Power); next, Polity and Elite are added. ¹⁹ The reason for this lies in the different N of the explanatory variables: as the research selects a listwise approach, ²⁰ considering the data for just 15 countries would seriously limit the analysis.

table 8. between micro and macro: some explanatory variables

Hypotheses	Relation with greater negotiating autonomy	Relation with stricter national control	Operationalisation
(1) EU career	Positive	Negative	Questionnaire
(2) Seniority	Positive	Negative	Questionnaire
(3) Polity	Positive	Negative	Hooghe–Marks index
(4a) Elite (4b) Public opinion	Positive	Negative	Eurobarometer
(5) Power	Negative	Positive	CIMC

Results

This section presents the results of the regression analyses (Table 9). Concerning the quantity and quality of the mandate they show a very clear picture. Only one variable turned out to be statistically significant: the relative power of EU member states. Ceteris paribus, more powerful states are more likely to prepare clearer and more detailed instructions for their delegates, and do it more often than smaller states. The control of larger (in terms of military capabilities) states on their representatives in Brussels is thus stricter. All six regressions yielded statistically significant results in the case of the Power variable. The intergovernmental expectation is confirmed also in the case of the second outcome variable: diplomats from larger member states are more likely to consider the capital as an important source for their negotiations. No other variable can explain the importance national officials attach to their capital. As the interpretation of coefficients of ordered probit models is not straightforward, the marginal effects (of only the variables that are significant) are reported (Table 10): they indicate the effect of a one-unit increase in the independent variables (at their mean) on outcome variables. In particular (first regression), as Power is up by one point (from a mean of 7.5), probabilities of always receiving a mandate are expected to increase by almost 3 per cent. Similarly, probabilities of considering the capital as the essential source rise by 3.6 per cent (from a mean of 7.2). The second regression reports similar results.

table 9. results of ordered probit

	Clearness (1st Regr.)	Clearness (2nd Regr.)	Detail (1st Regr.)	Detail (2nd Regr.)	Frequency (1st Regr.)	Frequency (2nd Regr.)	Source (1st Regr.)	Source (2nd Regr.)	Autonomy (1st Regr.)	Autonomy (2nd regr.)
Threshold	–1.238	–2.292	–1.943	–3.712	–0.586	0.789	–1.211	–0.304	–1.610	–1.362
[=1,00]	(0.569)	(0.660)	(0.668)	(0.731)	(0.799)	(0.889)	(0.706)	(0.527)	(0.669)	(1.128)
	–0.944	–1.985	–0.598	–2.080	0.647	2.082	0.400	1.210	–0.502	–0.138
[=2,00]	(0.546)	(0.646)	(0.594)	(0.615)	(0.793)	(0.918)	(0.736)	(0.633)	(0.661)	(1.043)
	–0.612	–1.622	–0.109	-1.639	2.219	3.500	1.314	–	–	–
[=3,00]	(0.515)	(0.680)	(0.572)	(0.583)	(0.865)	(0.968)	(0.868)
	.236	–0.711	.651	–0.820	–	–	–	–	–	–
[=4,00]	(0.504)	(0.702)	(0.560)	(0.591)
	1.369	0.606	1.633	0.224	–	–	–	–	–	–
[=5,00]	(0.492)	(0.637)	(0.547)	(0.615)
EU Career	0.033	0.033	–0.018	–0.052	–0.011	0.024	0.026	0.051	0.186^∗∗	0.184^∗
	(0.039)	(0.053)	(0.031)	(0.038)	(0.040)	(0.043)	(0.039)	(0.046)	(0.060)	(0.080)
Seniority	–0.001	–0.009	–0.011	–0.011	0.001	0.005	0.003	0.009	0.003	0.017
	(0.009)	(0.011)	(0.009)	(0.012)	(0.010)	(0.014)	.010	(0.012)	(0.013)	(0.019)
Polity	–	0.017	–	–0.027	–	–0.007	–	0.014	–	0.141
		(0.040)		(0.060)		(0.062)		(0.061)		(0.074)
Elite	–	–0.089	–	–0.211	–	0.217	–	–0.067	–	–0.215
		(0.112)		(0.117)		(0.173)		(0.097)		(0.182)
Public Opinion	0.004	–	0.002	–	0.006	–	–0.011	–	–0.016	–
	(0.007)		(0.008)		(0.011)		(0.009)		(0.010)
Power	0.041^∗	0.043^∗	0.043^∗	0.045^∗	–0.083^∗∗∗	–0.071^∗	–0.092^∗∗∗	–0.051^∗∗	0.010	–0.009
	(0.016)	(0.019)	(0.016)	(0.019)	(0.020)	(0.030)	(0.020)	(0.018)	(0.015)	(0.019)
N	125	86	125	85	125	85	124	85	125	85
Notes: Standard errors (adjusted for clusters in Nationality) are reported in parentheses. ^∗ p < 0.05; ^∗∗ p <.0.01; ^∗∗∗ p < 0.001.

table 10. marginal effects

	Clearness (1st Regr.)						Detail (1st Regr.)						Frequency (1st Regr.)				Source (1st Regr.)				Autonomy (1st Regr.)
	1	2	3	4	5	6	1	2	3	4	5	6	1	2	3	4	1	2	3	4	1	2	3	4
CINC	–0.003	–0.002	–0.003	–0.007^∗	0.004	0.012^∗∗	–0.002	–0.011^∗	–0.003^∗	0.002	0.009^∗	0.005^∗∗	0.029^∗∗	–0.004	–0.023^∗∗∗	–0.002	0.036^∗∗∗	–0.024^∗∗∗	–0.010^∗∗	–0.002	–	–	–	–
	(0.002)	(0.001)	(0.002)	(0.003)	(0.002)	(0.004)	(0.001)	(0.004)	(0.003)	(0.002)	(0.004)	(0.002)	(0.007)	(0.004)	(0.006)	(0.002)	(0.008)	(0.006)	(0.003)^∗	(0.002)
EU career	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–0.036^∗∗	–0.038^∗	0.074^∗∗	–
																					(0.011)	(0.018)	(0.024)
	Clearness (2nd Regr.)						Detail (2nd Regr.)						Frequency (2nd Regr.)				Source (2nd Regr.)				Autonomy (2nd Regr.)
CINC	–0.002	–0.002	–0.003	–0.009	0.003	0.012^∗	–0.001	–0.012^∗	–0.004	–0.000	0.010^∗	0.006^∗	0.028^∗	–0.011	–0.016^∗∗	–0.001	0.020^∗∗	–0.015^∗	–0.005^∗	–	–	–	–	–
	(0.001)	(0.001)	(0.003)	(0.005)	(0.002)	(0.005)	(0.001)	(0.005)	(0.002)	(0.002)	(0.005)	(0.002)	(0.011)	(0.006)	(0.006)	(0.001)	(0.007)	(0.006)	(0.002)
EU career	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–0.028^∗ (0.012)	–0.044 (0.024)	0.072^∗ (0.031)	–
^∗ p < 0.05; ^∗∗ p < 0.01; ^∗∗∗ p < 0.001.

Receiving a negotiating mandate more often, with clearer and more detailed instructions, and more frequently regarding the capital as the essential source of negotiations, is expected to show a strong relationship between military power and the representatives’ leeway as well. Being monitored more closely, national servants from larger member states will have less discretion in their activities. The two regressions do not confirm this reading, however: in both cases the relationship is weak and not significant. One other variable explains the relative freedom of CFSP/CSDP diplomats. Running the first regression, it emerges that those diplomats with a long career in the EU are more likely to have more space for manoeuvre than officials with limited experience. As EU career is increased by one point (from a mean of 2.9 years), probabilities of feeling very much restricted by the national mandate decrease by 3.6 per cent. On the other hand, national officials are more likely (by 7.4 per cent) to sustain that they enjoy enough freedom when negotiating in Council committees (from a mean of 4.2 years). Adding Elite and Polity, the relevance of the time spent in EU decision-making is confirmed.

Discussion

The regressions indicate that the relationship between capitals and Brussels can be explained by the military power of member states, and the years spent by diplomats in the EU decision-making process. How do we interpret these apparently contradictory results? In effect, the picture appears quite unambiguous: more powerful countries attempt to impose tighter control on EU foreign and defence policy. The neorealist expectation that larger EU states, being more reluctant to delegate or pool their sovereignty, provide their delegates with more frequent, detailed and clear instructions, is confirmed. However, this stricter control does not seem to translate into the level of discretion enjoyed by national representatives. Once officials reach the stage of negotiation, the dynamics that determine their amount of leeway are explained by another factor. Here, the factors allowing them a significant role in formulating the national position are cumulative experience gained, knowledge of EU and domestic contexts, and knowledge of their relative (formal and informal) codes of conduct. In the words of an official:

at the beginning you don’t realise when and how (and to what extent) you have a certain margin of manoeuvre in your committee … It is something that you learn over time, when you get to know the dynamics between Brussels and your own department. ²¹

Again, this allows them not just to relay back a great deal of information (and even relaying back may be selective), or just to advise the capital on what positions are realistic, but also to formulate the national policy. In other words, if government input can be convincingly explained by the amount of power held by member states, the effective discretion/autonomy of diplomats depends on the time spent in the EU decision-making process.

Conclusion

Through an original database of 138 questionnaires (and 20 interviews) with national officials working in CFSP/CSDP committees, the article has offered a nuanced picture of the dynamics that characterise the formulation of the national position inside those settings. The control of the centre over representatives in Brussels is far from being as tight as a strictly intergovernmental reading of the EU would predict. Assuming a ‘unidirectional causal chain’ – starting with the domestic level ‘and translated through state to the national … positions which are then represented in Brussels negotiations’ (Lewis 2000: 265–66) – and the state as a unitary actor does not describe what happens in the CFSP/CSDP. The relationship of representatives in Council committees with the central administration is open, with both levels formulating the negotiating position. Furthermore, half feel sufficiently free in the activities they take part in. These findings are even more revealing for the study of the Council and international negotiations, as the national control is widely assumed to be at its strongest in foreign policy. Next, a few explanatory variables have been introduced to explain the different coordination efforts of the capital: CWGs are more controlled than the PSC, while there have not been appreciable differences between the CFSP and CSDP. Variables at the micro-level (national officials’ experiences) and macro-level (state characteristics) have been tested: the features of the mandate are explained by states’ military capabilities, and diplomats’ negotiating freedom by the years spent in the Council. Future research should also test these variables in case studies, and explore how they eventually combine in single policy dossiers. Other factors could also be considered: the mandate and representatives’ leeway may vary according to the nature (and salience) of the issue and to the type of CFSP/CSDP operation.

The results of this article are also relevant for theoretical reasons. For instance, they can offer useful insights to better analyse EU foreign and defence policy, and provide implications for the two different perspectives (intergovernmental and institutional) on the CFSP/CSDP decision-making process. In conclusion, three observations can be advanced. First, if a strictly intergovernmental image of the formulation of the national position is contradicted, this does not indicate that intergovernmental studies do not explain the decision-making process of the Council and CFSP/CSDP. For instance, showing that the definition of the position is a joint process does not mean that states do not pursue the national preference. National officials, even when enjoying sufficient leeway, can nevertheless advance their own interpretations of broader national guidelines. ²² However, opening up the black box of the state and admitting the possibility of European inputs informing negotiations pose remarkable challenges to an intergovernmental reading of the Council.

One way to take into account the evidence this article has offered would be to make a more nuanced and explicit use of a P–A approach. The capital delegates a much more significant and sensitive function: (part of) the preparation of the national position, and not only its negotiation in Brussels. The domestic level has indeed several incentives to delegate such a function: reducing information asymmetries, economising on transaction costs, improving the efficiency of the national decision-making process and the quality of the policy, etc. – thus saving the capital’s time and resources. After all, Brussels delegates are national public servants temporarily assigned to national PRs. However, several problems may arise in applying a P–A approach to this intra-bureaucratic delegation. First, the data seem to suggest that the two levels are increasingly blurred: it is often difficult to identify who is the principal and who is the agent. Second, principals should be aware of the act of the delegation – or, at least, of its consequences and the importance of (credible) control mechanisms. Arguably, the mandate is the major control mechanism when the capital delegates the function of negotiating a certain policy. However, the design of delegation and control mechanisms is only loosely present when it comes to the (eventual) delegation of the elaboration of the position. Third, most P–A studies consider asymmetries in preferences, and incentive incompatibility, between principal and agents as one of their core features (Miller 2005: 205) – which does not necessarily occur in the case of the capital and PRs. Finally, the fact that the most important source of agents’ autonomy is not related to the control mechanisms at the disposal of the capital (the mandate, for instance), but to the number of years spent in the Council, challenges profoundly the validity of a P–A approach (Pollack 2007).

Second, in few cases institutional analyses have shown how and where socialisation and/or deliberative processes can impact on national positions inside the Council. As mentioned in the introduction, the mechanisms remain at best implicit; the domestic level, and the national/supranational nexus, are absent, or bracketed. This article has demonstrated that the European level – far from being unconnected with the national capital – nevertheless plays a relevant role in the formulation and negotiation of EU foreign policy, with a certain level of discretion. It has not demonstrated that a problem-solving approach characterises the CFSP/CSDP decision-making process or that specific member states’ policies have changed during negotiations. What has been indicated are the conditions for that to happen in several cases. The findings of this article can therefore provide a micro-foundation for these analyses of the Council and the CFSP/CSDP.

Finally, investigating the relationship between the ministry and Brussels may connect the literature on socialisation with Europeanisation studies (Müller and Alecu de Flers 2009: 19–23). Scholars have traditionally focused on long-term changes, often in the culture and identity of national society and state actors. However, there are two further avenues through which the EU ‘hits’ member states. Socialisation processes can change member states’ policies not only in the medium–long run, but also during the negotiation process itself (Checkel 1999). Again, the role of change agents is played by national officials who are at the crossroads of domestic and European settings. Finally, European input can enter the national decision-making process from the outset: as an Antici counsellor reported (in Lewis 1998: 491), ‘[i]nstructions already contain a big Brussels element in them, and sometimes they are Brussels instructions … sometimes they just copy our reports into instructions’. If the formulation of the national position is shared between the domestic and European branches of the national administration, and if the latter is embedded in a European context with common values and ideas, European elements can mould and Europeanise national (foreign) policy through diplomats in Brussels. Although methodologically difficult to detect, this Europeanisation through national representatives can nonetheless be an important feature of EU decision-making dynamics.

Acknowledgements

I would like to thank Caterina Carta, Mathias Koenig-Archibugi, Benjamin Pohl, Christine Reh, Michael E. Smith and the anonymous referees for their comments on earlier drafts of this article.

Notes

1. See Hayes-Renshaw and Wallace (1997: 40). Using more sophisticated data, Häge (2008) calculated that Council working groups and COREPER account respectively for 43 per cent and 22 per cent of legislative acts, while the remaining 35 per cent is evaluated by ministers. In foreign policy, Juncos and Pomorska (2011) argue that around one-third of the issues are discussed by ministers.

2. Although the distinction between diplomats in Brussels and their counterparts in the national capital is not as strict and clear-cut as assumed (since the former may be relocated to the home department after the experience in the EU), it is therefore analytically necessary, and empirically, theoretically and normatively justified.

3. In order to avoid subsequent criticisms and problems, national delegates in Brussels may occasionally ask to receive an explicit mandate.

4. Leeway/discretion/autonomy are here only loosely defined, and broadly used to evaluate the degree of freedom these officials perceive when negotiating in CFSP/CSDP committees. Although the exact meaning is left to the interviewee, they offer indications of the space of manoeuvre Brussels diplomats may have in the CFSP/CSDP decision-making process in general.

5. There are two exceptions. For logistical reasons, national diplomats who are only partially involved in Council negotiations and who reside in the national capital have been excluded. Only full-timers have therefore been considered. The second exception follows the beginning of field research: two specific parties (COREPER II and the Antici Group) replied that their handling of CFSP/CSDP issues was extremely limited.

6. The questionnaires were sent between April and July 2008, while the interviews were conducted between July and August 2007 and May and July 2008. The research was conducted before the treaty of Lisbon came into force (December 2009). In spite of the changes that have been introduced in the CFSP/CSDP governance (among others, the creation of the new double-hatted High Representative of the Union for Foreign Affairs and Security Policy and the launch of the European External Action Service), the decision-making rules and logics remain the same: decisions are made by the Council, and representatives of EU member states, almost exclusively by unanimity, with a very limited involvement by other actors. The most relevant innovation may concern the diminished role of the rotating presidency in the CFSP/CSDP: the meetings of the Council for Foreign Affairs, PSC and working groups are now chaired almost exclusively by the High Representative and her staff. To what extent, and in what ways, this can affect CFSP/CSDP negotiations, is uncertain and requires further investigation.

7. Table not reported.

8. Tables not reported.

9. Interview (6), Belgian official, May 2008.

10. Interview (8), French official, May 2008. For similar statements, see Juncos and Pomorska (2006) and Juncos and Reynolds (2007).

11. Interview (6), Belgian official, May 2008.

12. Interview (17), CIVCOM official, July 2008.

13. The 138 cases of the research have been divided into three categories: (a) delegates in CFSP working groups; (b) delegates in the PSC (and those working closely with them – Nicolaidis group – or with the capital, Relex Counsellors); (c) delegates who deal with CSDP issues. If national officials participate in more than one category, two strategies have been pursued: first, their scores have been counted for each group (N = 162). Second, to accentuate the variance, only the participants in just one of the three groups have been considered (N = 108).

14. Interview (11), Nicolaidis delegate, June 2008; Interview (2), Italian official, August 2007. However, a partial explanation may lie in the different kind of scrutiny (ex post versus ex ante) the PSC may experience, as its reports (and positions) – dealing with more relevant dossiers – are more likely to be read (ex post) by the capital. On the contrary, the outputs of the CWGs, having less significant and immediate consequences, are more likely to be left unchecked (ex post).

15. Other variables can play a role in explaining variance in the European–domestic nexus. Political salience is certainly one of those. The domestic level may have incentives to leave PRs more room for manoeuvre (both in the formulation of the position, and in the negotiation process) if the issue has little salience (or is highly uncontroversial) for a country’s foreign policy. In this article, the unit of analysis is the single interviewee/diplomat: it can safely be assumed that each of them has handled both more and less politically sensitive dossiers. In other words, political salience can be considered as distributed normally among the population of this research. Furthermore, whereas certain groups of diplomats are constantly biased in one direction (for instance, the PSC is supposed to deal with more sensitive issues than CWGs), the data shown above reveal a different – the PSC has relatively more autonomy – and more complex reality, and would suggest further investigation.

16. Both variables are operationalised through the number of years respondents spent, respectively, in the EU; and in domestic + international + European professional settings.

17. To operationalise this variable, I relied on the regional governance index prepared by Hooghe and Marks (2001). The reference year is 2000, and data are available for the then-15 members of the EU. It follows that this variable was tested only for these countries.

18. Power is here understood in terms of military capabilities. Koenig-Archibugi (2004) measures member states’ power capabilities using the Composite Index of Material Capabilities (CIMC) developed by the Correlates of War Project. What is relevant for his analysis is the percentage of power each state possesses in relation to the other EU member states rather than to all states in the international system. The same purpose is shared in this research. The reference year is 2001. The scale goes from 0.03 (Malta) to 18.82 (Germany).

19. At the same time, Public Opinion – since it measures the same variable as Elite – has been excluded from the second regression.

20. The objective is to make the analysis more robust and accurate; on the other hand, the missing cases are random and limited.

21. Interview (8), French official, May 2008.

22. Interview (19), Italian official, July 2008.