Abstract
The Council of the EU is the primary institution through which ministers of member states can express their positions on Commission proposals and vote upon legislation. This article first examines the extent to which ministers actually attend Council meetings before moving toward investigating the determinants of ministerial participation in over a period covering almost thirteen years between May 2004 and December 2016. It aims to identify determinants of why ministers from the Visegrad countries participate at some meetings and are absent from others. Using an original data set containing information about 940 Council meetings as well as several country-specific characteristics, we show that, on average, at about four out of every six meetings, ministers are absent. The results of our regressions indicate a pattern in which holding the office of the EU’s rotating presidency, the number of b-points on the agenda, and the size of the government increase the likelihood of ministerial participation in meetings. In contrast, high levels of popular support for Eurosceptic parties and holding of national legislative elections decrease the probability of ministerial attendance. Moreover, meetings of Council formations related to policy areas with a low level of EU competence are significantly less likely to be attended by ministers.
Keywords
The Council of the European Union (hereafter “the Council”) is the EU’s main venue for the voices of EU member states in legislative negotiations. While in theory the Council is legally a unitary body, it actually consists of three hierarchical levels: ten different formations of ministers at the top, the working parties at the bottom, and the Committee of Permanent Representatives (Coreper) in between. 1 Dossiers are first discussed in working parties that consist of member state representatives at the level of civil servants. If the working parties cannot resolve all the issues of a dossier, it is sent up the hierarchy to the Coreper, where states are represented by the heads and deputy heads of their permanent representations. Only dossiers for which the working parties and the Coreper fail to reach an agreement are then directly negotiated—as opposed to adopted—by the ministers at the top of Council hierarchy. 2
There is a considerable debate about the amount of dossiers agreed at each level. Early accounts claim that about 15 percent of all dossiers were directly negotiated by the ministers, another 15 percent by the Coreper, and roughly 70 percent by the working parties. 3 Recent research disputes these figures, however, and shows that the ministers directly negotiate approximately 50 percent of all dossiers, another 20 percent of the decisions are reached in the Coreper, and 30 percent are reached in the working parties. 4 The contemporary literature thus challenges the received wisdom that ministers are hardly involved and concludes that “ministers are far more involved in Council decision-making than previously assumed.” 5
The problem is that the ministerial participation is simply assumed: Research does not provide evidence that the ministers are indeed the ones who negotiate dossiers not resolved at the lower levels. However, as we show in this article, this assumption is far from true. While Article 16(2) of the Treaty on European Union provides that “the Council shall consist of a representative of each Member State at ministerial level, who may commit the government of the Member State in question and cast its vote,” the Council’s Rules of Procedure allow member states to be represented by somebody else with the right to commit the government in question. Anecdotal evidence shows that ministers indeed increasingly send non-elected officials to substitute for them. 6
Exact knowledge of when and why ministers participate at the Council’s meetings is, however, important for several reasons. First, the divergence of practices from Treaty stipulations is important from a constitutional point of view. Second, it is often argued that the legitimacy of Council decisions rests on ministerial involvement. A part of the Council’s legitimacy rests on the idea that citizens, acting via national parliaments, can hold ministers accountable for their behaviour in the Council. 7 If non-elected officials decide on the ministers’ behalf, however, democratic accountability is weakened. An underlying argument is that there are differences between the decisions made by ministers and those made by non-elected officials. 8 From the perspective of domestic political processes, non-elected officials are not subject to elections or routine public and/or parliamentary scrutiny; therefore decisions made by ministers are supposedly deemed more transparent as they abide by the norm of democratic accountability, and hence the decisions of ministers are perceived as more legitimate. 9
In this light, the consequences of the initially welcome conclusion that the ministerial level is more involved in the decision-making than is often claimed may be re-evaluated. If non-elected officials decide on the ministers’ behalf, the debate on the impact of the amount of dossiers decided upon by non-elected officials on the Council’s accountability loses some of its attractiveness when those negotiating at the top are still non-elected officials. How do decisions made in working parties differ from those made at the ministerial level if civil servants are making the decisions at both levels? An objection can be raised at this point: that an unelected official representing a minister at a Council meeting is under constant instructions from the minister and/or the cabinet and therefore directly accountable to the national parliament. The problem with this argument is that parliamentary control of EU affairs is still weak, particularly in the countries we investigate, 10 which results in potential accountability problems. Moreover, officials are exposed to different levels of transparency and potential scrutiny from the side of national parliamentarians when they act on behalf of their ministers in ministerial meetings. Finally, even if the non-elected officials perfectly follow instructions from their cabinets and no accountability problems arise, decisions made by non-elected officials instead of ministers encourage citizens’ perception of the EU as an undemocratic, bureaucratic body run by unelected officials, feeding public Euroscepticism. 11
Next, ministerial participation may also matter to the literature on the bargaining success of individual countries. 12 The literature shows that ministers have a greater bargaining power and a greater capacity to navigate politically during negotiations than officials, and the impact of ministerial participation may influence outcomes of the Council’s decision making. 13 Practitioners and scholars alike maintain that ministers are able to bargain in a different and more effective manner than civil servants and have a stronger mandate to negotiate and find the necessary compromise. Since the Council is the main arena where national governments can influence decision-making processes, the relative influence each member state can exert is subsequently of importance for the output legitimacy of the whole EU.
Finally and perhaps most importantly, the issue became relevant also for journalists, parliamentarians, and voters. The report about this issue by the Czech think-tank European Values published in 2015 14 and that by the Danish Think Tank EUROPA published in 2018 15 have received wide media coverage beyond their respective national audiences, reaching citizens across the EU. 16 Both reports were subsequently cited in parliamentary debates and interpellations, indicating that such information may be important for opposition parties, in particular, and for parliamentary scrutiny of EU affairs in individual countries, in general. 17 At the same time, one of the reports was quoted by the Czech State Secretary for European Affairs, indicating that executive actors use such information—if available—to demonstrate governmental activity in a comparative manner. 18 In other words, the media and think tank attention can be seen as evidence of the real-world relevance of and public interest in the type of research we present here. At the same time, the knowledge of ministerial participation is also important in the wider public sphere and may influence citizens’ and parliamentarians’ attitudes towards governmental handling of EU politics.
Therefore, it is important to have knowledge of ministers’ participation rates at meetings and to examine its determinants. To our knowledge, only Grøn and Salomonsen 19 investigated these questions so far. Testing six hypotheses concerning ministers’ participation, they found that the salience of the meeting, the importance of the policy area, and the length of the given country’s EU membership increase the ministerial participation. While their pioneering work has furthered our knowledge of the Council’s decision making, they only focus on a limited period of five years. Moreover, their measurement of meeting salience (operationalized as the number of b-points on the agenda) does not take into account the so-called false b-items. 20 As we show, such measurement significantly distorts the effect of this predictor on ministerial participation. They also omit the potential influence of the given country holding the Council presidency although research indicates that ministers should be more likely to attend meetings during their country’s presidency. 21 The same argument applies to extraordinarily scheduled meetings, although one may expect an increased ministerial presence at these meetings because of their high salience as they are convened as responses to critical political events.
Their analysis also ignores the potential impact of domestic governments’ organizational structures. In organizational theory, the size of an organization has an impact on its capacity. 22 If we apply this to Council meetings, the larger the government is in terms of the number of ministers, the more likely it may be that ministers will participate in the meetings given the increased capacity. Grøn and Salomonsen 23 also leave out the effect of national parliamentary elections. The office of ministers crucially depends on the performance of their parties in national elections. But even in non-election periods ministers have a limited amount of time to devote to the Council and thus usually focus on the most important dossiers. 24 Also, their schedules during a pre-election period may be under even more pressure as they are expected to participate in their party’s campaign. Finally, they only check the robustness of their results by estimating one additional model specification besides their baseline model. In contrast, we estimate four main models and report our regression estimates for a number of alternative specifications in the Appendix.
It is necessary to point out here, however, that while we attempt to improve on the pioneering analysis of Grøn and Salomonsen 25 in terms of the above-mentioned aspects, our framing of the research question, the theoretical framing of the article, and parts of the empirical puzzle depart from and build on their work. Our work thus owes a lot to Grøn and Salomonsen while trying to move beyond their work. At the same, they tested their hypotheses on all EU member states, while we test ours on the subset of EU member states. Against this backdrop, this article describes to what extent ministers from the Visegrad countries (the V4: Czechia, Hungary, Poland, and Slovakia) attended meetings over a period of almost thirteen years, including the determinants of their participation. We show that the average participation is 60.2 percent and that a country’s ministerial presence in the Council is related to factors such as the presidency office, national elections in the given country, meeting salience, government size, and party Euroscepticism. The remainder of this article is organized as follows. We next discuss our theoretical framework concerning the explanation of ministerial participation, including the hypotheses. Afterwards, we describe the data and methodology, after which we present the results of the analysis. We then summarize the findings, discuss their implications, and suggest avenues for future research.
The Theoretical Framework for Ministerial Participation in Council Meetings
We base our explanation of ministerial participation on a general assumption that multiple logics co-exist and influence actors’ behaviour. 26 It is necessary to distinguish between the rational choice institutionalist logic of consequences and the sociological institutionalist logic of appropriateness, both of which guide actors’ behaviour. 27 The two approaches explain behaviour on the basis of two distinct mechanisms. Rational choice institutionalism regards actors as strategic utility-maximizers whose preferences and interests are given and whose behaviour is guided by the logic of consequences. In this view, the goal of an action is to maximize one’s interests and preferences through strategic behaviour, that is, by selecting the best option in the particular institutional setting. 28
Sociological institutionalism, in contrast, regards actors’ interests and preferences as co-constituted by the institutional and social environment in which they find themselves. In this view, the behaviour of actors is guided by the logic of appropriateness. According to this logic, actors follow norms and rules of institutional settings and behave in a way that they suppose is expected from them in these settings. Actors take cues from the institutional setting as they construct their preferences, and select the appropriate behaviour for the given setting. Such rule-guided behaviour is different from behaviour based on the logic of consequences in that in the former actors attempt to do the right thing rather than maximizing their fixed preferences. 29
Although there are at least two distinct logics of action, our theoretical framework tries to echo Risse’s 30 argument that actors “tend to combine various logics of action in their behaviour.” As a consequence, Risse maintains, we do not need theories of EU politics resting on a single logic of social action. Rather, explanations of EU politics “should strive to integrate the various logics of social action and resulting propositions about human behaviour” to see how they compete with and complement each other. 31 Based on these arguments about the existence of multiple logics of action, we develop hypotheses regarding ministerial participation. 32
Determinants of Ministerial Participation in Council Meetings
Ministers only have a fixed amount of time, resources, and capacities that they can devote to domestic and EU political matters. As a result, they have to decide at which level of decision making they will allocate the resources to maximize their interests. Moreover, ministers “think of themselves as, and indeed have been elected as, national politicians.” 33 Ministers are therefore primarily focused on the national level and only secondarily on the EU level. Ministers can therefore devote only a limited amount of time to the Council and may be tempted to instead have officials substitute for them. However, as the stakes and costs of being absent from meetings become high, for example, due to re-election or policy impact concerns, the likelihood of ministers’ participation in the meetings should increase. 34
Research shows that more salient dossiers are indeed more likely to be negotiated at the Council’s ministerial level. 35 Against this backdrop, we expect utility-maximizing ministers to be more likely to attend those meetings where the most substantial dossiers are debated and leave less important ones to non-elected officials. On the Council agenda, dossiers are divided into a-points and b-points. A-points are dossiers that have been already settled by a working party or the Coreper, and ministers merely adopt them en bloc at the beginning of the next meeting. If the working parties and the Coreper fail to reach an agreement on a dossier, it becomes a b-point on the Council’s agenda to be resolved by ministers responsible for the given policy area. 36 The salience of a meeting can therefore be approximated by the number of b-points on the agenda. 37 We thus expect ministers to be more likely to go to Brussels when they have to actually negotiate dossiers rather than merely rubber stamp dossiers already agreed at lower levels.
Hypothesis 1: The more b-points there are on the agenda of the given meeting, the more ministers are likely to attend it.
Member states take turns in chairing the meetings at all Council levels. 38 The Lisbon Treaty changed this system marginally by providing for a permanent president of the Foreign Affairs Council. The presidency of the remaining formations continues to rotate between member states every six months. 39 The representative of the member state holding the presidency is in a unique position that not only provides benefits and duties but also raises expectations deriving from the Council’s norms. 40 While this is not the space to review the literature on functions and normative expectations of the presidency, 41 some of these functions and norms are relevant to ministerial participation, based on both logics of action.
First, the presidency office brings benefits and opportunities. The presidency is in a specific position since it has an information advantage in relation to other member states and possesses information on other EU institutions’ positions. The presidency also manages the Council agenda by introducing, emphasizing, and downplaying issues on the agenda and keeping some issues off the agenda, which allows it to place specific issues of national interest onto the agenda. 42 Moreover, it manages the negotiations between the Council, the Commission, and the Parliament. 43 As a result of its specific position, the presidency-holding country is able to influence policy outcomes in favour of its own national position and promote its national interests. 44 The presidency office thus increases the utility of ministerial participation as it facilitates the adoption of decisions reflecting national interests. 45 Based on the logic of consequences, we expect utility-maximizing ministers to be more likely to come to Brussels when their country holds the presidency.
From the perspective of sociological institutionalism, the literature indicates that holding the presidency evokes structurally driven role expectations. 46 The presidency office is connected with the following roles: agenda manager, mediator, political leader, and representative. These roles are related to certain norms and role expectations according to which holders of the presidency are expected to behave. Two norms stand out as particularly significant here: effectiveness and consensus building. 47 Ministers have arguably greater bargaining power and capacity to manoeuvre politically during negotiations than non-elected officials. In this light, since ministers are to successfully perform the roles of a mediator and a political leader and abide by the norms of effectiveness and consensus building, one can expect ministers to be present at meetings. Research shows that political leadership increases the chances of drawing the given negotiations to a close by providing vision, guiding the negotiations, identifying efficient bargaining outcomes, and taking initiatives to push the negotiations forwards in a given direction, thus serving the norm of effectiveness and consensus. 48 The same argument has been made for its role as a mediator, in which the capacity of ministers to manoeuvre politically during negotiations strengthens the brokerage potential of the presidency and allows it to build consensus and deliver results effectively. 49 Based on these norms and role expectations, we assume that ministers should be present at meetings during their country’s presidency.
Hypothesis 2: Ministers are more likely to attend Council meetings when they represent a country holding the presidency office.
On a more technical level, while upcoming meetings are normally announced in advance on a six-month basis, 50 some meetings are not scheduled in advance. Besides regular meetings there are also extraordinary meetings. Procedurally, they are the same as regular meetings but they differ in not being scheduled in advance as they usually respond to emergency events. As a result, we assume that their agendas are quite salient compared to those of regular meetings, as in extraordinary meetings, the Council discusses a response to an emergency event such as the Russian annexation of Crimea, and the seriousness of the issue increases the costs of non-participation.
Hypothesis 3: Ministers are more likely participate in meetings that are scheduled extraordinarily than at regular meetings.
The Council currently has ten formations that vary in terms of the width of the policy area, the competence division between the national and the EU level, and the frequency of meetings. Most importantly, Council formations comprise policy areas that differ in terms of importance and hence in terms of opportunity costs for ministers. Häge 51 shows that the involvement of the ministerial level in decision making varies considerably across policy sectors because of the higher costs of non-participation in some policy fields compared to others. Given the fixed amount of time ministers can devote to EU issues, we expect utility-maximizing ministers to be selective as to which Council formation meetings they will attend. We expect ministers to be more likely to attend meetings in highly salient policy areas, where the potential costs of sending someone else are higher. 52 We approximate the salience of a policy area based on the degree of EU competence within the given policy field (see Appendix A for a more detailed discussion).
Hypothesis 4: Ministers are more likely to attend meetings in policy areas with a high degree of EU competence than meetings of formations with a lower degree of EU competence.
Based on sociological explanations of EU politics and on the literature on responsiveness of political elites to public opinion, which shows that political elites respond to the electorate’s views and that public opinion shapes government attitudes towards the EU and governmental behaviour during negotiations in EU institutions, 53 we expect the dominant domestic perceptions of the EU to influence ministerial decisions to attend meetings. If the EU is seen as a desirable project by the domestic population, such an environment may increase normative expectations and thus also the ministers’ perception of the EU as a priority decision-making arena.
Before formulating the hypothesis, a more detailed clarification of the mechanism behind it is in order since one may also formulate a contradictory hypothesis, based on rational choice institutionalism. Ministers may decide to attend Council meetings as a result of public Euroscepticism, which motivates them to attend these and defend the national interest in the Council to make sure EU policy outcomes are in line with their national public opinion so as to counter domestic Eurosceptic views and therefore raise their chances of re(s)election. Participation in Council meetings may bring benefits to the ministers in terms of influence on EU policy outcomes or media attention. In the light of an Eurosceptic electorate, it would thus make sense for a minister to attend meetings to influence EU policy outcomes in order to present these achievements at home to gain media attention and votes as a defender of the national interest. This would potentially increase ministers’ chances of being re-selected for governmental office given that their office is dependent on the electoral success of the party that the minister represents, which, in turn, is dependent on the electoral support from voters who can reflect, inter alia, ministers’ behaviour at the EU level.
However, the literature shows that neither national nor European Parliament elections are to a large extent about EU issues. Rather, politicians’ (s)election prospects in elections are still connected to how they behave at the EU level to a very limited extent and are rather related to national/domestic political factors. 54 If politicians want to stay in office, they ought to pay attention to the issues the public cares about. 55 Yet, publics in countries with high levels of public Euroscepticism are often indifferent towards EU issues and focus on national political issues instead. 56 Put differently, ministers from countries with prevailingly positive public attitudes towards European integration can more easily sell the fight for the national interest and achievements in the Council and emphasize their role in the bargaining, and thus potentially gain political credit. In contrast, in countries with a rather Eurosceptic public, ministers would find it much more difficult to sell their Council achievements and get political credit from them, and therefore they may be discouraged from travelling to Brussels.
In sum, from a rational-choice perspective and following research on politicians’ motivations, 57 it is logical that ministers are primarily focused on national politics, as EU politics do not offer many gains in terms of vote- and office-seeking motivations while rather offering benefits in terms of policy-seeking. Ministers may thus prefer to send a substitute to meetings in their place and use their limited capacities to deal with national politics themselves as it has a more direct impact on their re-(s)election prospects. We thus expect ministers to be more likely to participate in the meetings in a more pro-EU environment.
Hypothesis 5: The higher the share of EU support among the citizens in a country, the more likely its ministers are to participate in Council meetings.
A similar mechanism that applies to the H5 applies also to the following hypothesis, and therefore we will not replicate it fully again. Public opinion is not, however, the only way of reflection of domestic perceptions of the EU’s desirability. Perceptions that can produce an EU-friendly or EU-hostile environment can be also reflected in the party system and through electoral processes (in the sense of how many votes pro-EU and Eurosceptic parties receive, respectively). While it is well-established that mainstream parties are significantly more supportive of European integration than the public, 58 Eurosceptic parties are on the rise in many member states, indicating dissatisfaction with how mainstream parties handle EU issues. 59 Moreover, party elites were shown to respond to the public opinion and the views of their supporters on EU issues, and government behaviour in the EU often reflects the partisan composition of the respective national parliaments. 60 Given the increased popular support for Eurosceptic parties, we assume that normative expectations about the EU’s desirability are also formed at the political party level. On the one hand, the above-mentioned normative expectations of citizens will be communicated to ministers via public debate and the media. On the other hand, the normative expectations at the party level will be reflected via political debate, and parliamentary and governmental institutions. As a consequence, we assume that party normative expectations about the EU’s desirability will be embedded as rules of appropriateness in governmental institutions of member states. 61
Following the research showing that activity in EU negotiations is higher for states with lower levels of Euroscepticism 62 and our above-mentioned argument on the embedding of the EU’s desirability in domestic (governmental) institutions, we assume that governments operating in a highly Eurosceptic (party) environment do not motivate their ministers to attend Council meetings and may even discourage them from doing so. This results from the rather low potential to promote policy achievements in the Council at home, which stems from the lack of domestic positive normative expectations of the EU’s desirability. In contrast, governments operating in a pro-EU (party) environment may actively encourage their ministers to travel to Brussels given the higher potential to promote policy achievements from the Council domestically given the existence of positive normative expectations of the EU’s desirability, or simply because of the reputational benefits of attendance at meetings. Therefore, the degree of popular support for Eurosceptic parties can potentially influence whether ministers find it appropriate to attend the meetings.
Hypothesis 6: The higher the popular support for Eurosceptic parties in the given country, the less likely its ministers are to participate in Council meetings.
Because of time, resource, and capacity constraints, ministers may prefer to send civil servants to Brussels in their place and focus themselves on domestic politics, which brings higher benefits in terms of ministers’ re-(s)election and policy influence. 63 However, participation at meetings, while definitionally additional to domestic political duties, can still bring benefits for the vote-seeking, policy-seeking, and office-seeking aspects of ministerial behaviour. 64 Member states also face no limits on which ministers they can send to a particular meeting. 65
Building on organizational theory, we suggest that the number of ministers in a government is relevant when ministers calculate the costs of participating at Council meetings. The concept of size has an important role in organizational theory as a structural characteristic of an organization. 66 By far the most usual measure of organizational size is the number of employees, and it is said to positively increase the capacity and, in turn, the performance of an organization. 67 The idea is that at any point in time an organization (including a government or a government bureau) has a finite number of people available to do its work. The personnel and related capacity constraints thus limit the amount of work the organization can do. In economic organizational literature, this effect is related to economies of scale, whereby under rational, utility-seeking behaviour, personnel size can be conceptualized as an input to a production function. 68 The measure of size as the number of employees, or ministers in our case, is particularly well suited for cross-national comparative research in relation to the same organizational types (governments in our case) that operate under similar organizational contexts. 69
In the context of our research, the (political) size of a government in terms of the number of ministers is thus linked to its capacities and performance. 70 European countries all have roughly similar amounts of political issues to deal with. In multilevel governance systems such as the EU, policy cycles and negotiations take place on the national, the regional, as well as the EU level of the system. 71 This places a high demand on the capacities of actors, particularly if the number of policy proposals on negotiation tables is high. 72
Unlike a lower number of ministers, a higher number of ministers means that one of the ministers can travel to Brussels without this leading to a loss of governmental (political) capacity to deal with national or regional issues. Since member states can send any ministers to any Council meeting, for instance a minister of environment to Justice and Home Affairs Council, the total number of ministers in a government means that member states with a higher number of ministers have a capacity-based comparative advantage over countries with a lower number of ministers. Put differently, a higher number of ministers in a particular government means that the member state is less likely to run into capacity shortages and thus encounter situations in which it does not have enough ministerial manpower to physically attend negotiations and events that are ongoing on the EU, national, and regional level. 73 In sum, the costs of sending ministers to Brussels are comparatively lower for larger governments because of the higher ministerial manpower, while the benefits remain roughly the same for all countries. To minimize the costs of sending an official (and losing some bargaining power), and because there are political actors to send, it is preferable to send ministers to Council meetings. Thus we expect that:
Hypothesis 7: The higher the number of ministers in the government, the more likely ministers are to attend Council meetings.
Building on the rational choice literature on political parties, we assume that ministers’ behaviour, much like party behaviour, is driven by vote-seeking, policy-seeking, and office-seeking motivations. 74 The ministerial office depends, inter alia, on the results of legislative elections and the popular support of the minister’s party. If ministers are not re-(s)elected for the office, they lose their influence on public policies. Their policy influence is thus dependent on their success in terms of vote-seeking and office-seeking, which is, in turn, dependent on the popular support of their party. So while ministers’ capacities are generally stretched, the costs of participation in Council meetings get even higher when their office is at stake due to an upcoming election. We expect that in order to ensure enough popular support for themselves and their re-selection for the ministerial office, ministers devote a part of their limited capacities to campaigning when facing elections. Ministers may therefore have even less of a capacity for Council meetings in pre-election periods. Because of the potential impact of elections on their future reselection, their schedules during pre-election periods are under more pressure than in normal periods. They may thus have even less capacity to attend Council meeting and prefer to send a substitute to Brussels.
Hypothesis 8: Ministers are less likely to participate in the Council meetings if they face upcoming general elections in their country.
We also control for the size of a member state, its administrative system, and for the distance from Brussels as all these variables have been shown to play a role in previous research. Please see Appendix B for the underlining argumentation behind these controls.
Data and Operationalization
We base our analyses on an original data set covering 940 Council meetings of all formations between 1 May 2004 and 31 December 2016. The data on the participants were collected from the official press releases of the meetings, available at the Council website. A total number of 5,024 persons from the V4 countries officially attended these meetings as representatives, including ministers and non-elected officials. Because of space limitation, in what follows we include merely basic information on data sources and operationalization. A more detailed discussion of the data sources and the operationalization for the variables is included in Appendix A.
We have decided to select only a subset of four EU member states for several reasons. First, when collecting data from Council minutes, we found out that it quite often contains mistakes when it comes to positions of individual attendees (whether they are ministers, junior ministers, or civil servants). As a remedy, we decided to focus on countries of which we have a detailed knowledge and/or whose languages we speak, and for which we thus were able to cross-validate the position of each attendee via alternative sources, be they websites and/or online archives of government departments, or personal contacts with government officials. Ideally, we would select a more representative sample of member states. But in such a case, we would not be able to ensure the cross-validation and validity of the data recorded in the minutes, which could, in turn, bias our results. In other words, besides the methodological factors of case selection discussed below, pragmatic case selection considerations stemming from data availability, language, and expertise barriers and resource and access limitation became decisive in our case selection process, although we are aware of the fact that they have no bearing on the external validity of our findings. 75
Second, our selection of the subset of EU member states has consequences for the variation of some of the explanatory variables (e.g., member state size). To mitigate this problem, we applied a diverse case selection strategy whose primary objective was to achieve meaningful variance along relevant dimensions of the analysis. Such a strategy is likely to enhance the representativeness of the sample, limit sampling bias, and thus improve the ability of our regression analyses to produce findings that can be generalized to the larger population. 76 By selecting the Visegrad countries, we tried to ensure a meaningful variation in national-level explanatory variables. Our sample includes one large country, two midsized countries, and a small country. It also includes variation in terms of public and party Euroscepticism, the sizes and types of political leadership, and the membership in the Eurozone that can also influence the participation of ministers, in particular at the ECOFIN meetings since the Eurogroup normally meets before ECOFIN. Finally, we believe that the effects of our explanatory variables can be considered not to be dissimilar across the entire EU since most of them are not culturally or regionally specific.
One caveat is in order. The hypotheses derived from sociological institutionalism may be more culturally or regionally specific as existing (Council) norms can be interpreted and internalized differently in individual member states and thus have different effect on the norm- and rule-based behaviour of ministers. On the other hand, the majority of our hypotheses are based on rational choice institutionalism and its logic of consequences. We maintain that the behaviour of ministers from different countries based on the logic of consequences and cost-benefit analysis should not be highly prone to different regional or cultural interpretations, given the similar institutional settings in EU member states. 77 Put differently, we maintain that even though all the member states in our sample are member states that joined the EU in 2004 and from Central and Eastern Europe, we believe that the chosen cases are to a certain extent representative of the larger population and not wholly idiosyncratic. Having said this, we should be very careful regarding the generalizability of our findings to the entire universe of the EU given the limited coverage of EU countries in our sample.
The Dependent Variable
The dependent variable is coded dichotomously based on whether or not a minister was present at a given Council meeting. We coded the formal titles of all participants for each meeting from the Council press releases. We then coded the participants according to their positions in terms of whether they are ministers or non-elected officials. 78
The Independent Variables
The number of b-points on the agenda
We use the number of b-points on the agenda as a proxy for the salience of the meetings. The variable is continuous and it is operationalized as the number of “items debated” from the minutes of each Council meeting. A-points that are not included in our analysis are defined as “items approved” in the meeting minutes. Unlike in previous research, 79 our measurement of agenda salience is sensitive to the so-called false b-items. While they appear as b-points, in practice they should not be treated as such. These are usually items on which sufficient consensus has been reached at lower Council levels, but one or more member states deemed it necessary to declare its or their dissenting view again in the following ministerial meeting, or cases in which it would be politically incorrect not to debate the issue at the highest level, such as cases of international agreements with third parties. 80 Including these points may lead to underestimation of the impact of true b-points on ministerial participation (see the more detailed discussion in the Appendix). 81
The presidency
This is a dichotomous variable based on whether or not the country held the office of the rotating presidency at the time of a meeting. We coded the related data on the basis of the Council decision on the order of rotation of Council presidencies (Decision 2009/908/EU).
Extraordinary meetings
This variable is coded dichotomously on the basis of the official press releases of the Council meetings. It pertains to whether a meeting was regularly scheduled or was an extraordinary meeting.
Degree of competence
This is an indicator of policy area salience. The degree of competences is coded as a dummy/trichotomous variable distinguishing between Council formations with a high degree of EU competence, those with a low degree of EU competence, and those with a shared competence between the EU and its member states. The operationalization is based on the review of the relevant literature dealing with the relative EU competence in various policy fields (see the more detailed discussion in Appendix A). 82
Public EU support
The level of EU support among citizens is coded on the basis of the following Eurobarometer question: “Generally speaking, do you think that (OUR COUNTRY)’s membership of the European Union is . . . ?” For each year, we calculated the variable as the annual percentage of people giving the answer “a good thing” to this question.
Party Euroscepticism
The level of party Euroscepticism is a continuous variable from the Chapel Hill Expert Surveys. 83 They measure party positions towards European integration, which range from 1 (strongly opposed) to 7 (strongly in favour). On their basis, for each term of office, we operationalized the level of party Euroscepticism as the share of the popular vote of parties that scored less than 4 (which means neutral) in the surveys.
The size of a government
This is a continuous variable operationalized as the number of ministers (excluding junior ministers) in a particular government. The related data were collected via the official governmental websites and cross-validated against the Central Intelligence Agency’s monthly Chiefs of State and Cabinet Members of Foreign Governments reports as well as the European Journal of Political Research Political Data Yearbook.
The pre-election period
Reflecting research on campaign length, this is a dummy variable coded as 1 for Council meetings that took place within 30 days before the national legislative elections in a given country based on information from national electoral commissions. 84 Otherwise it was coded as 0. See Appendix A for information about the data sources.
For the operationalization of the control variables, please see Appendix B.
Results
We first provide descriptive statistics about the extent to which the examined member states send ministers to Council meetings. On average, ministers from the V4 countries were present at 60 percent of the meetings included in the analysis (see Table C1 in the Appendix). There are notable (and statistically significant) differences between the four member states. Czechia tops the rank, sending ministers to 68 percent of all the meetings. It is followed by Poland, which was represented by ministers at 61 percent of the meetings. Finally, at the bottom of the ranking, both Slovakia and Hungary send ministers to the meetings slightly more than half the time (56 percent). These figures underline our theoretical discussion, where we stated that ministers are selective and thus do not attend all Council meetings. They also reveal a variation between member states in terms of the participation of ministers in Council meetings, which we now turn to investigate.
Table 1 reports our regression estimates for four main specifications. 85 Three of the four models are linear probability models (LPM), while the last model uses the logit estimation method as a robustness check. 86 With the exception of the third model, all the specifications use country fixed effects to avoid a potential bias/inconsistency due to unobserved country heterogeneity, while all the models but the first one also include controls for the year of the meeting (year fixed effects) to allow for a potentially time-varying participation rate. In the rest of the text, we focus on the results from the second model, which includes both controls (and at the same time lends itself to easier interpretation than the logit model). .
Determinants of Ministerial Participation in Council Meetings
Note: LPM = linear probability model; FE = fixed effects; N/A = not applicable. The dependent variable is the presence of a minister. For models 1–3, the t statistic is in brackets. For model 4, the z statistic is in brackets. Significance level: *p < 0.05; **p < 0.01; ***p < 0.001. Except for Model 4, the coefficients are multiplied by 100 to simplify interpretation. In all the models, we use robust standard errors clustered at the country level, allowing for correlation between errors across meetings for the minister(s) from the given country.
In line with our expectations, all four models indicate that an increase in the number of b-points increases the probability of ministerial participation. Adding one b-point to the agenda increases the ministerial participation probability by 1.25 percent (model 2). The results are highly statistically significant across all the models. In Table C2 in the Appendix we show that counting as b-points the false b-items as Grøn and Salomonsen 87 do leads to a significantly worse fit of the model as well as a substantially diminished and insignificant coefficient. 88 Moreover, when we test for the effect of the false b-items in addition to the effect of the true b-items, we find that a higher number of false b-items seems to have a statistically significant negative effect. These results highlight the importance of using an operationalization of b-items that excludes false b-items.
The second hypothesis concerns the effect of the presidency. As expected, our models illustrate that the presidency office leads to increased ministerial participation. Of all the variables we include, the presidency has the highest effect on the probability of ministerial participation as it leads to its increase by 30.8 percent. This finding is, again, statistically significant across all the models and indicates that the omission of this variable from the only research on the topic, that by Grøn and Salomonsen, 89 has a substantial impact on the explanatory power of their estimations. Our next hypothesis addresses extraordinary meetings. Contrary to our expectations, extraordinary meetings attract less attention than regular meetings. This result is statistically significant in models 2 and 3, but not in models 1 and 4.
With respect to the effect of the level of EU competence, the results provide some support to our hypothesis. The dummy variable for a low level of EU competence (competence=3 in Table 1) has a statistically significant negative coefficient in all the models, implying that ministers tend to come less frequently to meetings of Council formations with a low degree of EU competence. Also, the effect of this is quite large, as it is at around 15 percent. However, in none of the models is there a statistically significant difference between high competence formations and formations with competences shared between the EU and its member states (competence=2 in Table 1), and in most cases the coefficient has an unexpected sign. Overall, we find a significant tendency for ministers to participate less in formations related to policy areas with a low EU competence than in formations with a shared or high EU competence. 90
The next two hypotheses address domestic perceptions of the normative desirability of the EU as measured on the public and party level. The evidence for a positive effect of public EU support on the probability of the presence of ministers at the meetings is rather mixed. In model 1, without year controls, the coefficient is significant and positive as expected, implying a decrease of the probability by 3.3 percent for every 10 percent drop in public support of the EU. However, the coefficient becomes negative (and remains statistically significant) once we control for year effects (models 2-4). While this could be interpreted as contradicting the hypothesis, we are inclined to see it as an artefact of the structure of the estimation equation: a single coefficient of EU public support imposes a linear relationship between this variable observed at a yearly frequency and the probability of the presence of a minister at a meeting. In contrast, the yearly dummies allow the relationship between annual developments and the probability to be arbitrarily non-linear. Since EU support is a variable observed at a yearly frequency, the more flexible formulation for year effects possibly absorbs the main effect of EU support, leading to a negative estimated coefficient on the linear effect of EU support after accounting for non-linear year effects.
Our view is supported by the fact that the estimated year effects are very tightly correlated with support of the EU. This can be seen in Figures C2 and C3 in the Appendix, which show scatter graphs of year effects and EU support for all countries individually and on average, respectively. All the figures display a positive relationship between these variables—that is, years with larger positive effects are those years when EU support is higher—which is especially tight in the case of Czechia, where the year effects lie almost perfectly on the regression line. Our conclusion therefore is that there is a relationship between the tendency of ministers to attend meetings and general attitudes in society, with EU support possibly being a proxy for those attitudes. This is further supported by the results for the share of the Eurosceptic vote. In all four models, a higher popular support for Eurosceptic parties has a measurable positive and highly statistically significant effect. An increase in the share of the electoral support of Eurosceptic parties by 10 percent leads to a decrease of the probability of ministerial participation by 1.8 percent. This strongly supports our sixth hypothesis.
We also hypothesized that holding a national election would have a negative impact on ministerial participation in Council meetings. We corroborate our hypothesis, showing that during the 30-day period before a national election, ministers are substantially less likely to travel to Brussels. This result is consistent and statistically significant across all four models, and our main model indicates that one month before national elections the probability of ministerial participation decreases by 11.3 percent. Interestingly, while previous research on the topic has omitted this variable from its regressions much like the presidency variable, 91 our findings show that these two variables belong among those with the strongest impact on the probability of presence of ministers at the meetings.
Next, we hypothesized that government size (measured as the number of ministers in a given government) could influence the probability of ministerial participation. In line with our expectations, our findings show that a higher number of ministers in the given government leads to an increased probability of ministerial participation, as it increases it by 1.5 percent for each added minister. This result is not significant in model 1, however, and only at the 0.05 level in model 4. Nevertheless, our baseline model (model 2) and model 3 (as well as all the models in Appendix and the unreported probit model) display a higher level of significance for this variable and we hence deem our hypothesis corroborated.
Table C4 in the Appendix shows that the results for government size, EU support, and party Euroscepticism are predominantly robust across the countries, with only a few exceptions. When controlling for year effects, the effect of government size is negative but insignificant in Poland, while the effect of party Euroscepticism is positive in Slovakia and Poland, a statistically significant result in the former case. As for EU support, its effect is positive and statistically significant for Slovakia and Hungary when we do not include year effects, but negative for Czechia and Poland, significantly so in the former case.
We address our hypotheses and the effect of controls about country-specific factors (i.e., the distance from Brussels, the size of the country, and the administrative type) in two ways. First, we report the results from the pooled OLS estimation method (model 3). Second, we regress the estimated country fixed effects from model 2 on the country-specific factors (Table C5 in the Appendix). The results are consistent across these two methods. First, they indicate that in line with our expectations ministers from countries with politico-administrative systems including junior ministers, are less likely to show up at EU meetings. Second, as expected, ministers are more likely to turn up if their countries are either closer to Brussels or larger. The former finding substantiates the rational choice models showing that larger states have more influence in the Council. 92 Here, our results differ from the only available research on the given topic, which, first, does not find any effect of the given country’s distance from Brussels on the probability of its ministerial participation and, second, finds only a small negative effect of a member state’s size on its ministerial participation. 93
Conclusions
Our aim in this article was to examine to what extent member states are represented by ministers at Council meetings and to account for variation in their participation. Following the assumption that actors combine several modes of action in their behaviour, 94 we proposed eight hypotheses informed by different logics of action and research on the Council. We tested these hypotheses on the basis of a new data set covering 940 meetings in the period 2004-2016 and the V4 countries. We found that the four countries were more often represented by ministers than their substitutes at the meetings. However, the average participation rate across all the counties is only 60.2 percent. In four of ten meetings, these member states are represented by non-elected officials.
Moreover, there are significant differences between the countries. While Czech ministers attended almost 70 percent of the meetings, Slovak and Hungarian ministers were absent from almost half of them. Our results indicate that in the prolonged period we investigated, the participation of ministers from all the V4 countries decreased compared to the first five years of their membership. It thus appears that the post-accession socialization and learning effects proposed in relation to ministerial participation have not materialized in their case. 95 While we have not tested for the effect of the length of the countries’ membership, our estimates of year fixed effects (Figures C2 and C3 in the Appendix) show that the examined countries’ ministerial participation decreased markedly and consistently over time. Future research may examine how internal political developments (political crises, scandals, etc.) and international events (the economic and financial crisis, the migration crisis, etc.) influence the odds of ministerial participation in Council meetings over time.
Concerning our hypotheses, the results lend support for both rational choice and sociological explanations of ministerial participation. We found a strong and robust evidence for the positive impact of the salience of the given meeting, the presidency and the size of the government on the ministerial participation. We also found a strong and negative impact of popular support for Eurosceptic parties and holding of national legislative elections on the ministerial participation. Importantly, the effects of three of these variables have not been previously tested, and two of them (i.e., the presidency and national elections) belong among those with the strongest statistical effect. We therefore importantly complement the previous research by Grøn and Salomonsen. 96
Contrary to our expectations, the findings indicate that extraordinary meetings have a negative effect although it is not robust across all the models. While we are the first to test this relationship, it appears that the dynamics we discussed in the theoretical sections are actually the opposite of what is the case: despite the extraordinary meetings’ high salience connected to their discussions of sudden events, ministers struggle to fit their participation in extraordinary meetings into their busy schedules as they are called at the last minute. It may also be likely that extraordinary meetings scheduled in response to crisis events provide fewer benefits than costs: Ministers can expect that the unfolding events will be discussed at the meeting but no substantial decision will be adopted. Therefore, they rather send substitutes to represent their country and focus on the domestic political arena, which brings more benefits. Scholars can try to substantiate these claims by applying more qualitative methods, for instance.
We also found mixed support for the hypothesis regarding the salience of a policy area being measured by the degree of EU competence. As expected there is a negative, strong, and robust effect of policy fields with a low level of EU competence, suggesting that policy fields where decisions are primarily made at the domestic level, and which thus provide low benefits for ministers, attract their attention much less than other policy areas. In contrast, our findings do not support the expectation that meetings regarding policy areas with a high degree of EU competence would be attended by ministers more often than policy fields with competences shared between the EU and its member states. Future research may look into these dynamics in more detail, potentially developing different measurements of policy area salience than the level of EU competence.
The effect of the size of the political leadership is in line with our expectations: Ministers from countries with junior ministers participate less often in Council meetings as they can be replaced by other political actors (junior ministers).
From a theoretical perspective, this article contributes to the literature by applying the institutionalist approach to Council dynamics. First, we showed that regardless of the treaty-based obligation for ministers to attend meetings, these rules are subject to political interpretation. Ministers interpret them not only based on institutional features of the EU (i.e., the presidency, b-items), but also based on domestic political characteristics (i.e., Euroscepticism, elections). Ministerial behaviour in the Council is thus influenced by both domestic political characteristics and institutional features of the EU. Second, our analysis and findings also indicate a support for Risse’s 97 claim that the two institutionalist approaches are complementary and that actors combine various logics of action in their behaviour: Ministers indeed decide according to logic of appropriateness and logic of expected consequences. Our research design based on regression analysis does not, however, allow us to specify when either logic is more decisive. In the case of the hypotheses that are based on assumptions from both logics (such as the role of the presidency or the role of party and public support for the EU), future research may apply different, perhaps qualitative, methods to figure out which of the two logics is more decisive.
For instance, when a country holds the presidency, its ministers may decide to travel to Council meetings in order to capitalize on the increased bargaining power stemming from the presidency office. Based on the logic of consequences, utility-seeking ministers may thus travel to Brussels more often to promote decisions reflecting national interest. At the same time, holding the presidency office involves role expectations stemming from internalized Council working norms, which make ministers more likely to attend Council meetings. This case illustrates that the logic of appropriateness as well as the logic of consequences may be at play when ministers decide about their participation in meetings. In both cases, it is likely that both logics of action will influence ministerial behaviour, thus pointing to the complementarity of the two intuitionalist approaches. In the case of these complementary and conflicting explanations, elite interviews or other methods may be used to assess which logic of action is more decisive in a particular case and under which conditions. Finally, scholars can use our data set to examine the impact of the extent of ministerial participation on the Council’s legitimacy, accountability, and bargaining success that we posited at the outset of the article.
Supplemental Material
Submission-Who.Governs.in.the.Council-Determinants.of.Ministers.Attendance.to.Council.of.the.EU.Meetings-Online.appendices-v4 – Supplemental material for Who Governs in the Council? Determinants of Ministerial Participation in Council Meetings
Supplemental material, Submission-Who.Governs.in.the.Council-Determinants.of.Ministers.Attendance.to.Council.of.the.EU.Meetings-Online.appendices-v4 for Who Governs in the Council? Determinants of Ministerial Participation in Council Meetings by Jan Kovář and Kamil Kovář in East European Politics & Societies and Cultures
Footnotes
Acknowledgements
Particular thanks go out to our research assistants, Anna Roininen, Matouš Pinkava, and Mateeullah Tareen, who helped collected the data on Council meetings as well as country-specific information. All are or were interns at the Institute of International Relations, Prague.
Notes
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
