Structural, Domestic, and Strategic Interests in the European Union: Negotiation Positions in the Council of Ministers

During his two periods in office, the German Social Democratic chancellor, Gerhard Schröder, earned his nickname “Auto-Kanzler” because of his fierce defense of the German car industry’s interests in two major legislative discussions at the European Union (EU) level — the end-of-life vehicle directive (European Union Council of Ministers 1997) and the takeover directive (European Union Council of Ministers 1995). In the former proposal, the European Commission had suggested a law that would have made vehicle dismantling more environmentally friendly and encouraged car producers to develop cars that were easier to recycle. In the latter law, a legal framework for company takeovers was proposed by the commission that was intended to protect the interests of shareholders. The German opposition caused either the serious delay or the outright failure of these environmentally protective and consumer-friendly commission proposals because of the pressure from lobby groups.

These cases serve as examples that a government can quite predictably behave in the interests of a national industry within the EU, regardless of its partisan preference; in this case, the connection between Schröder and Volkswagen was particularly close because the German car producer is located in Schröder’s home region, where it is a major employer. This behavior by the German government thus suggests that the negotiation positions of governments in the EU Council of Ministers can be explained by structural and domestic interests and that the governments’ partisan orientations play only a secondary role. This is somewhat surprising considering that the left–right cleavage is still the dominant structuring principle in European party politics, and the ideological orientation of parties to a great extent determines whether they are elected for office or not.

Academic interest in preferences, negotiations, and conflict dimensions and their relevance to the activities of the EU has increased markedly in the last decade (Moravcsik 1998; Mattila 2009). Negotiation researchers and others who analyze decision making are particularly interested in what the conflict dimensions in the EU actually are and how the negotiation positions in the EU Council of Ministers negotiations can be explained. Does conflict arise because EU governments have different positions on European integration, because they have different economic or societal interests to defend, or because they are more or less left or right wing? Are their negotiation positions driven by national industries or by the public’s opinion of the EU integration process? Taking into account that the EU Council of Ministers is designed to represent the member states’ interests (in contrast to the European Parliament of the EU where party groups debate legislative proposals and where party ideologies dominate), it could be expected that structural and societal interests would be represented (Hix and Noury 2006).

As for the Council of Ministers, where EU governments negotiate day-to-day legislative issues ranging from the ingredients in honey to environmental standards, opinions diverge as to whether partisan cleavages (Mattila 2004), market-versus-regulation preferences (Thomson, Boerefijn, and Stokman 2004), or financial considerations (Zimmer, Schneider, and Dobbins 2005) are more relevant explanations of negotiation positions.¹ I discuss this in the first part of this study, arguing that examining negotiation positions according to policy areas indicates that it is the interests of national industries and constituencies and not so much the left–right orientation of governments that matter most in Council of Ministers negotiations. Because I concede, however, that some strategic considerations may also influence the choice of a position — especially when particularly important topics from the point of view of a negotiator are discussed — I also consider whether an extreme position can be explained on the basis of strategic considerations. My corresponding results indicate that extreme positions are especially taken by powerful member states and by governments that have a particularly strong interest in the negotiation issue at hand. These findings contribute to ongoing debates about the determinants of preferences (Frieden 1999) and the difficulty of distinguishing between sincere and strategic preferences in international negotiations (Bueno de Mesquita 2004; Achen 2006).

International relations theories and negotiation analyses use the preferences of states as a crucial starting point for the investigation of the relations and interactions of governments in the international sphere. For analytical clarity, I distinguish between interests, preferences, and negotiation positions as follows: by “preferences,” I refer to the underlying interests of a state, which are its long-term goals or, in game-theoretic terms, the way a state would order the nodes of a game tree (Frieden 1999). A negotiation position instead encompasses these preferences as well as the strategic considerations that arise when a negotiator takes the positions of other players into account (Morrow 1999). This might mean that a government strategically “exaggerates” its preference, adopting an extreme position with the intention of dragging the other negotiating states toward its actual position. In the following section, I briefly discuss theories from international relations and the negotiation literature relevant to analyzing negotiation positions.

International relations scholars have formulated several theories to explain the determinants of negotiation positions. Based on the assumption that states behave like “billiard balls” in an international system, “realists” argue that the preferences of states result from their positions in the state system. According to this theory, for example, national military capacity can be seen as a means of signifying the power position of a state (Morgenthau 1967). In the realists’ view, however, neither policy concerns nor domestic interests influence the position of a state in the international system. “Neo-realists” subsequently developed this argument further, arguing that military resources are not the only determining factor of a state’s position, but that economic interdependencies and domestic factors can also exercise a strong influence (Keohane and Nye 1989).

“Liberal” theorists attribute an even stronger influence to domestic variables, asserting that they are the defining elements of a state’s interest. In line with the liberal institutionalist approach, Andrew Moravcsik (1997: 518) defined states’ interests as an aggregate of individual interests and explained that states represent “some subset of domestic society.” Thus, states’ interests are composed, on the one hand, by their fundamental preferences determined by structural economic interests, and on the other hand, by social interests that are defined by their domestic desires. This argument is built on the idea that the more privileged a social group is, the more likely it is to have its interests well represented within the national government (Moravcsik 1997). Thomas Schelling (1960) developed the term “paradox of weakness” to describe the possible effect of domestic groups upon international negotiators, explaining how a government negotiator can jockey for concessions at the international level by convincing the other party how critical and skeptical his domestic constituency is. The history of EU negotiations reveals an abundance of cases in which such concessions were made, for example, the granting of the Danish and British exemptions from the Maastricht Treaty obligations on social policy.

In a similar vein, Simon Hug (2003) criticized intergovernmental analytical approaches in the EU literature for ignoring the link between pivotal actors at the domestic level and preferences declared at the international level, arguing that international bargains often have to be ratified at the domestic level. Likewise, in his study of preference formation in the negotiations regarding the International Treaty on the Seabed Convention, Thomas Bräuninger (2001) was able to show that domestic preferences determined states’ preferences to a large degree.

In addition to structural and social interests, the partisan preferences of governments are also considered relevant for explaining how governments choose policies (Schmidt 1996). Because the ideological positions of parties help distinguish them from their competitors in national elections, we might expect partisan orientation to influence a government’s negotiation position. Philip Manow, Armin Schäfer, and Hendrik Zorn (2008) were able to show that in the context of European Council summits, the left–right orientation of EU governments — in particular, the shift from the political right in the 1950s toward the left in the 1990s — can help explain changes in EU social policy. Similarly, regarding EU environmental policy, Christian Jensen and Jae-Jae Spoon (2011) have shown that Green Party representation in a government predicts that government’s progress toward meeting Kyoto Protocol targets for reducing the production of greenhouse gases. It remains unknown whether this effect holds only for high-level EU summits, however, considering that they are concerned with more general policy guidelines in contrast to the Council of Ministers negotiations under discussion in this study, which are more concerned with discussing day-to-day legislation. But, Michael Bechtel and Gerald Schneider (2010) have shown the importance of economic interests even at the summit level.

As far as the EU Council of Ministers is concerned, opinions diverge regarding the meaning of negotiation positions and what determines them. In their analyses of final votes in the Council of Ministers, both Sara Hagemann (2007) and Mikki Mattila (2004) found the left-wing versus right-wing leaning of EU governments to be an important independent predictor of voting behavior. Sara Hagemann and Bjorn Hoyland (2008) reached a similar conclusion. Their study, however, does not control for the possible impact of other factors. Indeed, other studies have identified a north–south dimension (Mattila 2009); preferences based purely on north–south geography do not seem to support any existing theory and may reflect underlying structural interests, such as the presence of certain industries and populations that are located in the same geographical regions.

Similarly, when examining the original negotiation positions at the beginning of the council discussion from the “Decision Making in the EU” (DEU) data set (Thomson et al. 2006),² other researchers suggested that the political dimensions break down along a north–south dimension (Thomson, Boerefijn, and Stokman 2004; Kaeding and Selck 2005; Zimmer, Schneider, and Dobbins 2005). (The DEU data set includes the negotiation positions of the then fifteen EU member states on sixty-six legislative proposals.) A preference for regulatory versus market-based solutions (Thomson, Boerefijn, and Stokman 2004) and a redistributive dimension (Zimmer, Schneider, and Dobbins 2005) were also identified. These results based on the DEU data set thus suggest that the structural and social interests of the states — to use Moravcsik’s terminology — better explain the negotiation position of an EU government in council negotiations than do partisan preferences.

Finally, scholars who used survey data regarding communication patterns in the council (Elgström et al. 2002; Naurin and Lindahl 2008) found that more communications occurred between countries in the same region than between governments that shared a more left- or right-wing ideology.

None of these analyses has investigated, in detail at least, which domestic preferences are actually instrumental in terms of shaping certain negotiation positions. This study thus adds to these previous analyses by examining which particular domestic interests explain negotiation positions, as well as by investigating the possible role of strategic positions.

One explanation for why governments might choose not to directly represent their sincere structural interests lies in their use of tactics. Strategies reflect an actor’s overall attempts to influence negotiation circumstances; tactics are those actions he or she takes to react to other players’ positions; both are chosen according to the information an actor has about the other players (Morrow 1994: 66). These “signaling games” characterize situations in which an actor with private information sends a signal to his or her counterpart to which that counterpart is then expected to react to (Spence 1973; Gibbons 1997; Morrow 1999). The results of strategic actions consequently depend on the asymmetrical distribution of information and upon the goals and strategies of other players. Such actions could include a bluff (i.e., signaling a more extreme position than is actually intended).

In the case of EU negotiations, Gerald Schneider and Lars-Erik Cederman (1994) demonstrated that adopting an extreme position can be an effective tool for achieving a particular goal in the European Council. Negotiation research has confirmed this; some researchers have found that a strategy of starting tough and then gradually making concessions can be a more effective means of reaching an optimal division of resources, than a strategy in which a softer position is maintained throughout the negotiations (Rubin and Brown 1975). Similarly, negotiation studies have shown that bold strategies can be worth the effort because higher aspirations often lead to higher payoffs (Zetik and Stuhlmacher 2002; Wolfe and McGinn 2005). The effect of taking a hard initial bargaining stance was found to be even stronger when the counterparts adopted a softer and more lenient strategy and were prepared to make concessions (Rubin and Brown 1975). Of course, choosing a more extreme position can be less efficient when all negotiators employ that same strategy because agreements become less likely in such cases (Fisher, Ury, and Patton 1991).

Such a threat failed, for example, during the mad cow disease crisis when the British used a blocking strategy that backfired when it was met by strong resistance, which is not an uncommon response. Out of anger over the embargo against British beef, the British delegation refused to agree to several unrelated legislative proposals in May and June 1996 that required unanimity (Dinan 1999). Because the other fourteen governments showed an even higher degree of determination to continue the beef blockade, the British government was eventually forced to give up its blocking strategy. The high degree of resolve of the fourteen counterparts arose, most likely, from the widespread public fear of contaminated beef in the EU.

As described, actions can reflect sincere preferences or strategic considerations. In the above-mentioned example, the British government treated the export of beef as a salient issue and thus assumed an extreme position with regard to unrelated EU proposals. In particular, this blockade was probably an exaggerated preference and thus not a sincere position. It is hardly ever possible to distinguish exactly between sincere and insincere strategies, and we can thus merely assume that extreme positions may not necessarily reflect a sincere preference only because assuming an extreme position is an established negotiation practice that can reflect strategic as well as sincere considerations by a negotiating government.

Having discussed the variables “structural interests,”“domestic interests” (or social interests in Moravcsik’s terminology), “partisan orientations,” and “strategies” as possible determinants of the preference of a government in Council of Ministers negotiations, I will now use the first part of my empirical analysis to expand the previous analyses of Christina Zimmer, Gerald Schneider, and Michael Dobbins (2005) and Robert Thomson and his colleagues (2004) and show how a more detailed analysis of domestic interests can explain state preferences.

Based on the liberal intergovernmentalist approach, I postulate in the first of my four hypotheses that EU governments assume positions that reflect their structural (economic or redistributive) interests (Moravcsik 1997) in Council of Ministers negotiations:

Apart from economic interests, I also propose that various domestic-level societal interests influence governments’ positions in EU negotiations. Governments may adopt a position that directly represents the opinion of the citizens or of important domestic stakeholders to demonstrate to the domestic constituency that its European-level representative actively defends their interests.

As a control variable, I will account for the partisan orientation of the EU governments at the time of the debate about the proposals.

If one allows for the possibility that bargaining positions reflect negotiation strategies, it should be stated that employing such strategies can come at significant cost. If negotiators exaggerate their positions or fail to honor commitments, they risk losing their credibility and thus losing their threat potential (Hovi 1998). So, while they may be strongly tempted to invent a constrained win set which is the set of all possible agreements that would find a majority among the constituents (Putnam 1988), they also confront the danger a “bluff” will be detected by other parties (Schelling 1960) — especially in an information-rich environment such as the Council of Ministers, where it is not uncommon for negotiators to meet on a daily basis.

Therefore, I postulate that strategies such as assuming an extreme position are only used in exceptionally difficult negotiation situations. For example, in 2000, the British government successfully obtained a number of concessions in the lengthy and tough negotiations over the EU directive on artists’ resale rights by assuming an isolated extreme position at the end of the opinion continuum (European Union Council of Ministers 1996).³ Furthermore, this success could be interpreted as proof of how a tough and intransigent extreme position can pay off in legislative negotiations in the EU: because of the outlying position of the United Kingdom, the final common position of the council supported the notion of a maximum ceiling on artist royalties, thus, seriously distorting the original idea of the proposal, which had wanted to treat artists more favorably at the cost of the auction houses. The British delegation achieved special consideration of its “vital interest” in protecting the future of the British auction houses Sotheby’s and Christie’s where 90 percent of the European auction sales occur.

Thus, my third hypothesis is as follows:

Another possible variable relates to each state’s voting power within the council. Some member states may consider whether they can swing a final vote with their number of votes. In such a situation, large member states might be tempted to assume a position far from the median opinion if they perceive a greater chance of preventing a winning majority in qualified majority voting, in contrast to smaller states whose voices might more easily be ignored. In order to see whether exogenous power resources (Bailer 2004), such as the number of votes, influence the choice of an extreme position, I test the fourth and final hypothesis:

So far, conducting studies on EU Council of Ministers negotiations has been made more difficult because of the extreme secrecy that shrouds Council of Ministers negotiations (Gabel, Hix, and Schneider 2002). A few studies have used the final voting records (Mattila and Lane 2001; Mattila 2004, 2006, 2007; Hayes-Renshaw, Van Aken, and Wallace 2006) and formal statements (Hagemann 2007; Hagemann and De Clerck-Sachsse 2007a, 2007b) to analyze the council voting decisions at the end of negotiations.⁴ To gather information regarding negotiation positions at the beginning of council deliberations, the analysis of negotiation protocols (Bueno de Mesquita and Stokman 1994; König 1997; Hug and König 2001; Hopmann 2002) or the direct interviewing of negotiation participants has been used (Thomson et al. 2006). Whereas text analysis has the advantages of being easily accessible to other researchers, reliable, and economical (Sullivan and Selck 2007), interviews can shed light on actual negotiations not noted in protocols and help researchers identify the most important negotiation issues and determine the salience of these issues for the individual negotiating parties. The downside, however, is that interviewees might have faulty memories or may be self-serving in their responses, portraying situations to their advantage — the so-called ex post rationalization (Bueno de Mesquita 2004).

Both methods of identifying negotiation positions — text analysis of council votes and deliberations and interviews — may not actually reveal how closely official positions reflect the actual preferences of actors; this applies to the qualitative (Moravcsik 1998; Dinan 1999; Hösli 2000) as well as quantitative studies (Bueno de Mesquita and Stokman 1994; König 1997; Thomson et al. 2006). While Christopher Achen (2006) considers it possible to learn a negotiator’s sincere position, Bruce Bueno de Mesquita (2004) has claimed that it is practically impossible to ascertain his or her true opinion. Indeed, he argued that where there is a low probability of finding out negotiators’ real preferences, especially when information is incomplete, it is far too advantageous for a diplomat to assume a strategic position. If such positions reflect only strategic goals, then negotiation analyses could be seriously biased. Consequently, because analyses of negotiations are based on the preferences of actors (as is the case with many other political science studies), the question of which variables determine their positions is highly relevant and important for developing hypotheses about behavior in international negotiations.

The primary source of information for this analysis is the data set “Decision Making in the European Union,” which includes detailed information on sixty-six legislative proposals negotiated by the Council of Ministers from 2000 to 2002 (Thomson et al. 2006). This data set was compiled to create a collection of information on day-to-day negotiations in the Council of Ministers, in order to test bargaining models. Therefore, each legislative proposal needed to raise at least a minimum level of controversy in order to be considered. The selection criterion was whether “Agence Europe,” a daily comprehensive news service reporting about EU activities, mentioned a proposal, and subsequently, whether an EU expert confirmed that the proposal was contentious. Finally, the legislative proposals that were selected were either subject to the consultation or the codecision procedure, which both require unanimity, or to a qualified majority, which required sixty-two of the eighty-seven votes in the Council of Ministers.

Interviews were only conducted with experts who had a chance to witness the whole bargaining process, both between and within the diverse legislative bodies of the EU and thus typically included officials from the European Commission, staff from the Council of Ministers, or the Permanent Representations of the Member States in Brussels. In the interviews, the experts were first asked to identify the controversial issues within a proposal. Based on this, they were subsequently asked to indicate the position that the decisive actors (member states, the commission, and the European Parliament) took after the introduction of the proposal before the council formulated its common position. The experts were then asked to locate the two EU actors holding the two most extreme positions on the two endpoints 0 and 100. This was used as a means of representing all the stakeholders’ opinions on the issue, with the intermediate positions taken up by the remaining EU actors seen as representing less extreme positions or compromise solutions. In cases in which qualitative (as opposed to numerical) issues were negotiated, the creators of the data set relied on the interviewees’ expertise to identify the distance between the positions of the negotiating parties. The majority of issues (109 of 162) reflected a ranked ordering of policy positions; thirty-three of the issues were dichotomous, where the only positions were 0 and 100; and in twenty cases the measurement is relatively scalar, indicating that all points on the dimension represent a substantive position (e.g., financial transfers to a certain EU program).

Other interview questions — apart from those that asked for status quo and outcome — pertained to the salience attributed to the contested issues, which we defined as the importance actors attached to the negotiation topic. Salience was measured on a scale between 0 and 100, with 100 indicating that an issue was of highest importance and 50 marking an average priority.

To explain the negotiation positions assumed, I use several independent variables measuring the economic and societal interests of the EU member states, as well as their ideological orientation. All variables and operationalizations used for this analysis are listed in Table One.

The financial benefit that EU states derive from the EU is measured by the difference between their contribution to and payment received from the EU budget, the net benefit from the EU, the data for which are drawn from the relevant statistics of Eurostat (http://epp.eurostat.ec.europa.eu). The economic interests in the fields of agriculture and fisheries are measured on the basis of the agricultural and fishery workforce in the respective countries.

To measure domestic constraints, I used the attitude of the population toward consumer protection in the EU, as is noted in the Eurobarometer surveys of the EU. To test for attitudes toward the EU and the left–right positions of EU governments, I used the expert survey data from a data set created by Kenneth Benoit and Michael Laver (2005), in which experts gathered the positions of parties on the left–right scale (1 = left, 20 = right) and on the question of whether they favored shifting more authority toward the EU (= 1) or away from it (= 20). Each government represented in the Council of Ministers at the time when the commission initiated one of the sixty proposals is thus assigned a value for its left–right position and EU friendliness. For coalition governments, for example, Germany, I weighed the values according to the size of coalition parties operationalized by the number of ministers in a coalition government.

In this section, I present two methods of analysis for examining the preferences of EU governments. In both, I investigate which variables explain the negotiation positions of member states in the EU Council of Ministers negotiations. For this purpose, I use negotiation issues from just three policy areas — financial, consumer, and institutional policy — in the first analysis. In the second analysis, I consider the latent variables that underlie the negotiation issues as the dependent variable to achieve the same goal, explaining the position of member states in negotiations.

In my first analysis, I identify nine issues from the DEU data set that deal with various types of subsidies and payments, that is, financial issues, as the first dimension. The dependent variable here is the position of an EU government on these issues on a scale from 0 to 100, where a higher scalar position corresponds to spending more money. The issues range from financing measures promoting the European audiovisual industry to the budget of the civil protection program to the amount of money allocated for the seven-year Socrates program supporting educational programs. The two other dimensions, consumer issues and institutional issues, are also scaled so that a higher position on the scale reflects more consumer protection and more EU integration, respectively. Table Two displays the results of the multivariate analyses in which I use a fixed-effects regression analysis technique to control for omitted variables that differ between the issues but are constant over time. This is equivalent to using dummy variables for each of the issues.⁵

In the first column of Table Two, we see that governments’ utilitarian financial considerations allow for the prediction of their positions in the negotiations: the independent variable “net benefit from EU,” which measures the profit of an EU member from the EU budget, has a significant and positive coefficient, suggesting that the more a state has benefited from the EU budget, the more likely it is to spend money on EU policies. Disregarding all other aspects of these negotiation issues (culture, fisheries, etc.), the net variable explains 27 percent of the variance on financial issues.

The variable that controls for the partisan preferences of the EU governments, “Left–right preference of EU governments,” is, however, not significant when tested for its influence on the positions regarding financial issues, unless the financial interests of the EU member states are also accounted for, as I did in column three of the model, where the effect does become significant. Here, the negative coefficient indicates that right-wing governments favor lower EU expenditure than the more left-wing-oriented governments. This is another confirmation of the results of Douglas Hibbs’ (1977) famous study on the influence of parties on public spending in which he found that left-wing governments favor higher state spending.

Regarding the preferences of the EU governments on issues concerning consumer protection, such as the labeling of honey or the strength of a health warning, I could not find a variable that compellingly explains the choice of position. Neither the “Attitude to consumer protection” (operationalized as the percentage of people in a country who think that the level of consumer protection in their own country is high, as measured in Eurobarometer surveys) nor various other variables, such as the post-materialism index (an index compiled of survey answers measuring to which degree postmaterialist values are held in the sample of various populations) created by Ronald Inglehart and Christian Welzel (2005), can explain the positions of the EU governments regarding the labeling of products. Although the variable points in the expected positive direction — indicating that states whose citizens favor stronger consumer protectionism are much more likely to promote stricter consumer laws at the EU level — it is not significant according to conventional standards.

Similarly, the partisan variable is negative as expected, suggesting that right-wing governments are less inclined to support consumer protection than left-wing governments, but fails to reach significance at the conventional level. These results might however be caused by the twentry-seven relatively diverse issues investigated. Indeed, it is quite possible that additional considerations, for example, those related to the agriculture, tobacco, and medical products sectors, are not captured very well by the rather crude measurement of average levels of consumer protectionism in a country.

In the third model, which investigates how we can explain the choice of negotiation positions in institutional negotiation issues, we see that the variable measuring the attitude of EU governments toward EU integration is significant. It indicates that the more skeptical a government is of the EU, the closer to zero it is located on the institutional issue scale. These states seek to guard their rights in interinstitutional disputes concerning, for example, conflicts about the appropriate voting mechanism in the EU implementation (comitology) commissions or debates about the degree of autonomy given to the European Commission. The partisan preference variable indicates that parties further to the right tend to be more careful about shifting rights toward the EU. This confirms Simon Hix’s (2005) findings, which state that in some EU member states, the left is more pro-European than the right.⁶

In sum, I have demonstrated in this first analysis that financial interests and attitudes toward European integration can to some extent explain the positions that EU governments are likely to take in EU negotiations. Controlling for partisan preferences did in two instances help to increase the explanatory power of the models to a small extent. As the F-tests indicate, however, the overall significance of the models is limited; also, some of the coefficients are rather small because many specificities of the individual negotiation issues might not be well captured. Consequently, I have undertaken a second analysis of negotiation dimensions.

My second analysis of the determinants of negotiation positions is based upon data generated by Zimmer, Schneider, and Dobbins (2005), who investigated the underlying dimensions of the 150 negotiation issues in the DEU data set. With a correspondence analysis of council negotiation positions, Zimmer and her coauthors showed that the dominant dimension in the council negotiations was a north–south divide, caused by the states’ differing interests in market regulation, protectionism, and financial contributions to the EU.⁷ Also, by splitting the negotiation issues into the policy domains of internal market and agriculture, they argued that their calculated latent policy dimensions cover between 66 and 86 percent of the variance of positions (Zimmer, Schneider, and Dobbins 2005: 412).⁸

Using correspondence analysis to reduce the issues to these key latent dimensions offers several advantages in contrast to the direct analysis of positions on the actual issues: first, interviewer effects (the discrepancies caused by different interview partners) have a much smaller weight if the dimensions are reduced to a few latent ones. Second, situational effects stemming from particular negotiation situations, such the details of certain policy issues, matter less (Bradburn and Sudman 1979). Thus, this second analysis can show whether my findings from the first analysis also hold for the discussion of latent dimensions and not only for the purely political dimensions, as in Table One. Furthermore, additional policy areas — internal market, agriculture and fishery policy — are also investigated in the second analysis.

Table Three shows the results of my multivariate analyses using ordinary least squares (OLS) regression and addresses the positions of EU member states in the policy sectors internal market, agriculture, and fisheries, using the structural, domestic, and partisan variables. In the first model, I again use the variable “Net benefit from EU budget” to explain the first dimension of positions in internal market negotiations. The reasoning behind this test is the assumption that states receiving a greater share of the EU budget have a different interest in EU internal market issues than other states do. The significant and negative coefficient of the budget variable means that countries benefiting financially from the EU prefer to be located on the low end of the scale of the first latent internal market dimension. The second model refers to the impact of the left–right orientation of EU governments on this dimension and shows that more right-wing EU states are usually located at the other end of the scale. (I refrained from testing the two variables in one model because of the small number of cases.)

By identifying the topics that bear most decisively upon these dimensions, Zimmer, Schneider, and Dobbins (2005) defined the second dimension of the internal market as one that is dominated by consumer protection and producer interests (columns 3 and 4 of Table Three). To test for consumer protectionism, I used the percentage of people who are of the opinion that there is a lower level of consumer protection in other EU countries, using this to measure the “missionary behavior” of citizens who believe that other EU countries are not as advanced as their own country regarding consumer issues. Such behavior can be observed, for example, in internal market negotiations in which countries with high consumer protection standards (such as Denmark and Sweden, which are at one end of this scale) try to impose their ideas on countries with fewer restrictions (such as the United Kingdom or Ireland, which are located at the opposite end of the scale). This can alternatively be interpreted as a “strategy of raising rival’s costs” (Vaubel 2006: 132), which more highly regulated member states may use to raise the costs of others by imposing their own level of regulation on them. Although only an approximate measure of consumer protection, it actually explains a strikingly large percentage (39 percent) of the variance on this second dimension of internal market negotiation positions.

To operationalize producer interests as a counterpoint to those of consumers, I evaluated the effect of a country’s industrial production upon each of the positions. Both variables fail to show a significant effect.⁹ This could indicate that the measurements do not capture producer interests precisely enough; indeed, these are relatively difficult to measure as such interests are very heterogeneous and thus too widespread to be implemented in such an analysis. The partisan preference of the EU governments had a negative effect as expected, indicating that right-wing governments tend to assume positions on the opposite side of the scale as consumer protectionists; however, the variable is not significant by conventional standards.

Domestic as well as structural interests are visible in the explanation of negotiation positions in another crucial policy area of the EU — agricultural policy. Here, the share of people employed in the agricultural sector explains why actors are located on the upper end of the scale of this dimension (column 5 of Table Three); indeed, the larger the workforce employed in the agricultural sector, the further up the scale the country is located on the first agricultural dimension. Interestingly, this relatively crude measure of a state’s interest in the domain of agriculture alone seems to account for a large degree of the variance on this particular dimension. But when I control for partisan preferences, this variable turns out to lack significance. Nonetheless, the positive coefficient illustrates that right-wing governments often enact more farmer-friendly policies and are therefore on the same side of the political dimension as the agricultural sector, which confirms the usual coalition between conservative parties and the agricultural sector. The partisan orientation, however, has a smaller coefficient than does the economic interest of the agricultural workforce.

I obtained comparable results in the area of fishery policy: here, I use the number of fishermen in EU countries, which accounts to some degree (31 percent) for the positions of the governments in discussions pertaining to this policy area, although the coefficient is admittedly quite small. As was true in the case of agricultural and consumer issues in the bivariate models in Table Three, a distinctly smaller percentage of the variance is explained by partisan preferences than by structural and domestic interests. These analyses thus indicate that a country whose employment force is highly dependent on the fishery sector will behave in a predictable way in fisheries negotiations, regardless of whether it has a right-wing or left-wing government.

Looking at the F-tests of the various models, which are significant only for the structural models but not for the partisan models, these analyses illustrate that a large percentage of the positions of EU delegations are an expression of their structural and social interests, confirming that the DEU data actually reflect sincere positions of EU actors, rather than purely strategic ones — a claim made by the DEU research group (Thomson et al. 2006).

The remaining unexplained variance could have several explanations. First, states could have further interests that were not controlled for; this is likely because I used relatively crude measures in the absence of more direct ones. Second, the unexplained variance could be because of measurement errors in the interviews. Third, I must also consider the possibility that there may still be some aspect of strategy included in the choice of a position. Indeed, it is possible that — in spite of the frequency of meetings — the negotiation delegations assume strategic positions that cannot be directly explicated through structural or social interests, which might be particularly germane to cases in which especially important topics are discussed. As noted above, extreme positions at the beginning of a bargaining round can be a popular method for negotiators to demonstrate their determination to the other participants. Therefore, I have analyzed whether an extreme position can be explainedon the basis of strategic considerations. (This analysis also serves as a robustness check of the data.)

Using an OLS regression, shown in Table Four, I investigate which of the independent variables salience and power, introduced in Hypotheses Three and Four, can explain the choice of an extreme position in Council of Ministers negotiations. The measure for an extreme position is the distance of each member state’s opinion from the opinion median of all member states per negotiation issue, the opinion median being the outcome that would result if the negotiators did not bargain (Black 1948). This measure is calculated for the fifteen EU member states on 150 negotiation issues in the DEU data (n = 2,245).¹⁰ For control variables, I use again the partisan orientation of the EU governments, and dummies for the use of the codecision procedure (which involves the European Parliament), qualified majority voting, and the length of a proposal. The variable “Length of proposal” is designated by a 1 if the proposal was debated for more than 900 days and therefore belongs to the top 15 percent of proposals in terms of length. I include this measure to see whether extreme positions were lengthily debated and were thus probably more contested proposals.

From the results in Table Four, we can see that, as expected, extreme positions are assumed when governments care very much about an issue or when they are powerful; indeed, as anticipated in Hypothesis Three, strong interest in a certain negotiation issue motivates governments to choose a position further from the median, indicated by the positive and highly significant coefficient of the salience variable. This confirms negotiation research that has shown that bold strategies are worth the effort because higher aspirations lead to higher payoffs (Zetik and Stuhlmacher 2002; Wolfe and McGinn 2005).

As noted, the analysis also indicates, on the basis of the positive and significant coefficient, that the strategy of commencing a negotiation with a more extreme position is chosen especially by economically powerful EU states. Here, the economic size of EU member states can also be considered as an alternative measure for concern or overall salience of issues, because the economically powerful states have more diverse interests to represent and have thus an overall average salience. This result also shows that, in contrast to some negotiation theories that point out that high aspirations lead to more bargaining success independent of the resources of an actor (Wolfe and McGinn 2005), the delegations in the EU seem to follow the pattern that the large member states — which can afford to be extreme because they can get away with this behavior — end up choosing extreme positions. When I control for the partisan orientation of governments, I find that right-wing governments tend to assume more extreme positions than left-wing-oriented governments as the positive and highly significant coefficient indicates.

In addition, the analysis shows that the choice of position is also influenced by the characteristics of the negotiation topic. The dummy variable controlling for the length of a legislative negotiation indicates that extreme positions are especially likely in negotiations that extend for several years.

As mentioned, I control for the institutional effects of the legislative procedure using two dummy variables, one for the codecision procedure and the other for the existence of majority voting. Theoretically, I would have expected that extreme positions are assumed in cases of unanimity voting, because these are the situations where an actor can turn into a veto player and block a compromise; on the other hand, taking on an extreme position in qualified majority voting situations is theoretically useless because the majority can simply ignore the veto player (Tsebelis 2002). But the significant coefficient of the qualified majority voting dummy could indicate that negotiations in the council are governed by different forms of behavior: even in negotiations in which an outlier could be ignored, actors seem to adopt these positions to signal that they have an increased interest in a negotiation issue and are thus keen to get concessions from the other participants.

Such behavior is reported to be efficient in negotiations when governments overplay their domestic constraints in the sense of the “paradox of weakness,” in order to obtain concessions (Schelling 1960; Putnam 1988; Schneider, Finke, and Bailer 2010). I analyze the positions stated at the beginning of negotiations, however, and it seems likely that extreme positions are abandoned during the course of negotiations so that a compromise with the necessary majority can be found.

The positive coefficient of the codecision procedure suggests that negotiators use the extreme-position strategy especially when they face the more complicated legislative procedure that involves the European Parliament. This coefficient, however, is not significant by conventional standards.

The analysis of extreme positions also serves as a robustness check for my previous analyses. The results concerning the degree of influence achieved by actors taking extreme positions suggest that some negotiation situations are more likely to attract such extreme negotiation behavior. Moreover, these results confirm the earlier findings that indicated a high degree of sincerity in the positions in regular Council of Ministers negotiations. It is quite possible that extreme positions are sometimes chosen in exceptionally salient situations; however, compromise- and result-oriented negotiations in the council appear to be more typical.

In this study, I have demonstrated that negotiation positions in the EU Council of Ministers are most often a reflection of the structural and domestic interests of states. The partisan orientation of governments influences the choice in some instances but less distinctly so. Therefore, knowing the interests of relevant domestic stakeholders might reveal more about the negotiation position of an EU member state than knowing which party is represented in government. After examining the determinants of negotiation positions, I have also investigated whether the negotiation positions collected in the DEU data set are sincere positions. For frequent negotiations in particular — as in the Council of Ministers — positions stated at the beginning of bargaining rounds to a large extent reflect states’ sincere interests and are rarely designed to bluff and hide an actual position.

Nevertheless, I also found evidence of extreme positions being used, at least to some degree, in contested negotiations. Such findings provide a modicum of support for the idea that such a position can reflect strategic and tactical considerations. It is especially when governments care a lot about a certain negotiation issue that they choose extreme positions; furthermore, this strategy is more often chosen by powerful states, possibly because the smaller ones, knowing that they do not stand many chances of winning with such a risky strategy, might choose not to risk such a failure.

The higher frequency of council negotiations, however, in contrast to EU summit negotiations, leads to less use of these kinds of insincere strategies. Here, the “shadow of the future” (Axelrod 1984), the idea that it pays to cooperate when you will interact repeatedly with the same party, ensures that government’s representatives in council negotiations cannot use bluffing and other similarly bold strategies carelessly. Therefore, I assert that both insincere strategic and sincere considerations determine the course of day-to-day legislative negotiations in the EU. Dramatic strategic moves are reserved for exceptional bargains, whereas the daily business of negotiating EU legislation is determined by the sincere structural and societal interests of the governments. It is extremely challenging to conclude, however, whether states are behaving strategically or sincerely by looking only at their negotiation positions and behavior.

Future research could investigate situations in which governments represent domestic interests that are not easily compatible with their ideological and party orientations, for example, when a Social Democratic chancellor defends industrial interests at the expense of the environment and consumers. Future research should examine in more detail who prevails in which situations: the stakeholders or the party ideology. Presumably the salience of negotiation issues — who cares most about and “owns” the issue — and the type of negotiation — whether it is redistributive or regulatory and whether the effect can be directly perceived by stakeholders, as in the case of job losses — are factors that influence these effects, also. Thus, comparative negotiation research interlinking the insights of comparative politics with research on negotiation positions and international politics could further illuminate the negotiation behavior of states.