Introduction

For the last decade, there has been a growing literature explaining the bargaining success of member states in EU lawmaking (Aksoy 2012; Arregui 2016; Arregui and Thomson 2009; Bailer 2004; Cross 2013; Golub 2012; Lundgren et al. 2019; Thomson 2011; Thomson et al. 2006; Warntjen 2017). However, this literature contains one limitation that is the rationale for this article. So far, existing studies have not evaluated the determinants of bargaining success of two key European Union’s member states: Germany and Poland. This is surprising given that both these countries are the most powerful actors in the Council in terms of voting power.

This paper seeks to address this lacuna. Its main purpose is to explain how the similarity of Poland’s and Germany’s preferences affects their bargaining success in EU lawmaking. It also examines how their success is influenced by the proximity of their policy positions to pivotal actors, i.e. member states holding the highest voting power in the Council. Drawing from the spatial theory, two hypotheses are delineated. The first expects Poland’s and Germany’s success to increase when both these countries approximate or have similar policy positions on EU legislation. The second stipulates that the similarity of Poland’s and Germany’s preferences provides both these countries with greater success than their closeness to other pivotal member states, namely France, the UK, Italy or Spain. Both hypotheses are tested using a linear regression on the DEUII dataset. The results show that cooperation between Poland and Germany in EU lawmaking is highly beneficial: not only bringing their preferences closer increases both countries’ success, but also it is significantly more profitable than moving their positions towards other member states with the highest voting power, as well as it weakens the negative effect of the preference extremity on their success.

The remainder of the paper is as follows. The second section develops a theoretical framework. It discusses the assumptions of spatial theory and delineates two hypotheses. The third chapter depicts the research design. The fourth part utilizes descriptive statistics tools to analyse the level of Poland’s and Germany’s success in EU lawmaking as well as the similarity of their preferences. In two subsequent sections, a linear regression is conducted to test hypotheses and determine the effect of Polish–German cooperation on the success of Poland (fifth section) and Germany (sixth section). The article concludes by summarizing the results.

Theoretical framework

According to the spatial theory of decision-making (Crombez and Vangerven 2014; Hörl et al. 2005; Tsebelis 2000), legislative negotiations in the EU take place between several actors, namely member states forming the Council, the Parliament and the Commission. Each actor has specific preferences on legislation that can be represented by points in the Euclidean space. In other words, the actor has an ideal policy, i.e. a legislative solution that is the most beneficial for him, and his utility decreases as the final outcome of negotiations moves further away from this ideal point. In this theory, the Parliament and the Commission are represented as unitary actors, since they both have one, own legislative position usually adopted by a simple majority and allow amendments in their internal decision-making. It is assumed that, according to the median voter theorem (Black 1958), their preferences are equal to the median preferences of their members. By contrast, the Council cannot be treated as a unitary actor, because, first, it consists of countries having different preferences, and second, it adopts its positions rather by qualified majority or unanimity than by simple majority.

Such spatial conceptualization of negotiations assumes that actors’ bargaining success is primarily shaped by the positioning of their preferences across issues, in particular its extremity. However, given that states differ in terms of their voting strength, the proximity to certain actors may play a relevant role in determining bargaining satisfaction. Specifically, having a preference close to a pivotal state holding the highest voting power may lead to higher bargaining success, while sharing similar position with less powerful country may be less advantageous. The justification for this statement is twofold. First, since the Council takes a vast majority of decisions by qualified majority, member states need to build winning majorities to realize their preferences. It is argued that a state with a larger number of votes exerts more influence on this process, because it is more likely to be pivotal in turning a losing coalition into a winning one (Shapley and Shubik 1954). Hence, proximity to such a key player could be beneficial to other actors. Second, as Häge (2013) showed, actors in the Council are blocking-minority seekers: they band together with states holding similar positions in order to form coalition that is large enough to block a decision. The reason is that building a blocking minority strengthens a state’s bargaining power by ensuring concessions from other negotiating partners (Nedergaard 2007; Warntjen 2017). Hence, when an actor has a policy position close to that of a state with high voting power, it is more likely to reach the necessary numbers to constitute a blocking minority, thereby increasing its bargaining position. Against this backdrop, I theorize that actors’ bargaining success is influenced by their proximity to the most powerful states.

Assuming that the closeness to pivotal states matters, the question arises: With whom should Poland and Germany cooperate (approximate preferences) in order to be more successful in EU lawmaking? To answer it, two hypotheses are delineated. The first expects Poland’s and Germany’s bargaining success to increase when both these countries approximate or have similar positions on an issue. The rationale for this hypothesis is twofold.

First, Germany and Poland are two of the most powerful actors in the Council. According to power index studies, the former country has the largest mathematical influence on the final shape of legislative acts, as well as on the blocking of negotiations in this institution (Barr and Passarelli 2009; Nurmi et al. 2013). On the other hand, Poland ranks sixth in terms of voting power. Under the former Nice triple-majority rule (2003–2014/2017), its ability to influence legislation and block decisions did not differ significantly from France, the UK, Italy and Spain (Nurmi et al. 2013). The Treaty of Lisbon, by introducing a double majority system (from 2014/2017 until now),Footnote 1 has limited Poland’s bargaining position; however, this country still maintains sixth place (Kóczy 2012). The spatial theory assumes that higher voting power of an actor translates into more influence (Felsenthal and Machover 1998). There are several empirical studies confirming this expectation.Footnote 2 For example, Hosli et al. (2011) found that states with high voting power are less likely to vote against a proposal, though this effect does not hold for the new member states. In turn, Warntjen (2017) discovered a robust positive relationship between the number of votes backing a member state request to change EU legislation and its success probability. On this basis, one can argue that when Poland and Germany approximate or have similar policy positions, their high voting powers combine, increasing their likelihood of success.

Second, the similarity of the Polish and German positions may also increase their chances of forming a blocking minority in the Council which, as mentioned above, provides a state with higher success. Under the Nice system, a blocking minority could be established by countries having 90/321 (in 2004–2006), 91/345 (in 2007–2013) and 93/352 (in 2013–2017) votes. In this period, Poland and Germany had 56 weighted votes; therefore, they only needed 34–37 votes to block the decision. In addition, the Nice system provided an optional condition that the decision could be suspended if states supporting it did not represent 62% of the EU population. Since total population of Poland and Germany constituted approximately 24.8%, both countries needed support from countries representing more than only 13.2% of the EU population to block the decision. A double majority system introduced by the Treaty of Lisbon has changed the conditions for building a blocking minority. Now, it can be established by: (1) four states representing more than 35% of the EU population or (2) 45% (currently 13 countries) or 28% (currently eight countries)Footnote 3 of all Council’s members. Since Poland and Germany represent 23.4% of the total EU population (Decision 2014), their coalition needs only two additional countries representing more than 11.6% of the EU population to block the decision. Hence:

H1:

Poland’s and Germany’s bargaining success is more likely when both these states approximate or have similar positions on the negotiation issue.

While H1 expects the cooperation between Poland and Germany to increase their bargaining success, it says nothing about the contemporaneous impact of their proximity to other specific member states holding a pivotal position in the Council. In fact, both these states may also be more successful when bringing preferences closer to such actors, and not only to each other. Besides Poland and Germany, four additional member states play a pivotal role in the Council, namely France, the UK, Italy and Spain, since they are in the top six in terms of voting power in this institution. While the proximity to these states may also be advantageous for Poland and Germany, I test the different hypothesis expecting the similarity of Poland’s and Germany’s preferences to be more beneficial for both these countries than their closeness to France, the UK, Italy and Spain. This supposition stems from three arguments.

First, Poland and Germany are more capable of building a blocking minority when cooperating together than with other countries. The reason is that a potential Polish–German coalition is likely to obtain—through Poland—the support of the other members of the Visegrad Group (Czechia, Slovakia, Hungary), because Poland is a part of this organization and has relatively convergent preferences with these countries (Tsebelis and Yataganas 2002: 286). The coalition made up of Germany, and V4 accounts for 87 weighted votes (according to the Nice system for 28 countries) and 28.5% of the EU population. As a corollary, under the former Nice system, it was sufficient for this group to team up with only one state holding more than four weighted votes (e.g. Lithuania) in order to achieve a blocking minority. Building such a minority under the contemporary Lisbon system is more difficult, but not impossible, since it requires the support of one or several additional states representing more than 6.5% of the EU population. This can be done through cooperation with only one of the countries with the highest voting power.

Second, several studies using a multidimensional scaling showed a strong proximity between the positions of France, Spain and Italy in the Council (Frantescu 2017; Mattila 2009; Thomson 2009). At the same time, these countries’ preferences are significantly distant from those of Poland or Germany. This suggests that it is easier for France, Spain and Italy to bring preferences closer to each other than to Poland or Germany. Particularly that their potential coalition is very beneficial—since they represent together 34.14% of the EU population (and 85 weighted votes under the former Nice system), they are able to achieve blocking minority easily by getting the support of a fourth country having more than 0.86% share of population (and eight votes under the Nice system).

Third, Poland and Germany, along with the UK, are states that most often vote against legislation in the Council. By contrast, France, Spain and Italy are less likely to contest acts in this institution (Van Aken 2012). Therefore, Poland and Germany have more opportunities to build effective blocking minorities together than with other states holding the highest voting power. Hence:

H2:

The similarity of Poland’s and Germany’s preferences is more beneficial for both these countries than their closeness to other member states holding the highest voting power in the Council.

Research design

The following methodology is used to test the hypotheses. First, a dataset is constructed. Second, the dependent, independent and control variables are operationalized. Third, a statistical test of hypotheses is carried out using a multiple linear regression. Specifically, I conduct two tests, one for Poland and one for Germany, to investigate how the similarity of the Polish and German preferences affects their success.

Dataset

In this study, I use the DEUII dataset (Thomson et al. 2006, 2012). It contains information on 331 controversial issues that arose during negotiations over 125 important EU legislative proposals. Since Poland joined the EU in 2004, the DEUII was reduced to files negotiated after this date. As a result, the hypotheses are tested on a sample of 124 legislative issues nested in 52 proposals. This means that issues are the units of analysis. For each issue, the DEUII contains actors’ policy positions expressed on the 0–100 scale. These positions were specified through semi-structured interviews with key informants who participated in negotiations over these issues. Informants were asked to: (1) identify key controversial issues that arose during negotiations over selected proposals; (2) determine on the 0–100 scale the initial policy positions of all member states, the Parliament and the Commission on each issue; (3) estimate on the same scale the level of salience the actor attached to each issue; and (4) indicate on the same scale the final outcome of negotiations over each issue.

The logic of using the DEUII is presented in Fig. 1. It includes policy positions of member states, EP and Commission as well as the outcome of negotiations on a proposal for a directive on the promotion of clean and energy-efficient road transport vehicles (Directive 2009). During negotiations, several conflicting issues emerged, one of which was concerned with the methodology for calculating the operational lifetime costs of CO2 emissions, in particular determining the monetary value assigned to each unit of CO2 pollution. As Fig. 1 shows, actors were divided on this issue along three lines. The first called for maintaining the previous legal status (status quo), according to which lifetime costs were individually specified by member states. This view was supported by a coalition of six states, including three countries with the highest voting power in the Council: Germany, Italy and the UK (position 0). According to the second position, the monetary value for each CO2 emission unit had to be equal to 2 eurocents per kilogram. This view was backed by the Commission and all other member states, including three countries with the highest voting power: France, Spain and Poland (position 70). The most extreme position was advocated by the Parliament, which proposed introducing the highest value for one unit of CO2 emissions: 4 eurocents per kilogram (position 100). Ultimately, the negotiations ended with a compromise that the operational lifetime cost for one unit of CO2 emissions of a vehicle shall be from 3 to 4 eurocents per kilogram (position 90).

Fig. 1
figure 1

Source: Thomson et al. (2012)

Actors’ policy positions regarding the monetary value assigned to each unit of CO2 pollution. EP European Parliament, COM European Commission, AT Austria, BE Belgium, CY Cyprus, CZ Czechia, DE Germany, DK Denmark, EE Estonia, EL Greece, ES Spain, FI Finland, FR France, HU Hungary, IE Ireland, IT Italy, LT Lithuania, LU Luxembourg, LV Latvia, MT Malta, NL the Netherlands, PL Poland, PT Portugal, SE Sweden, SI Slovenia, SK Slovakia, UK the United Kingdom.

Full size image

The operationalization of the variables

I created two dependent variables: Poland’s success and Germany’s success. They measure Poland’s and Germany’s bargaining success as the absolute distance between their initial policy position and the final outcome weighted by the salience a state attached to an issue:

$${\text{Success}}_{pi} = \frac{{\left| {{\text{Position}}_{pi} - {\text{Outcome}}_{i} } \right|*S_{pi} }}{100}$$

where i is the legislative issue; p, Poland or Germany; Positionpi, policy position of a state p on an issue i on the 0–100 scale; Outcomei, the final outcome of negotiations over an issue i on the 0–100 scale; and Spi, the salience attached by a state p to an issue i on the 0–100 scale.Footnote 4 This salience-weighted formulation of success is preferred in the literature over the unweighted distance between the actor’s position and the outcome (Arregui 2016; Cross 2013; Golub 2012; Thomson 2011). The reason is that states do not care equally about all issues; their success weighs more heavily when it is obtained in highly salient issues to a state, and vice versa. In addition, Schneider et al. (2008) found that models that account for the saliency actors attach to an issue have the best predictive accuracy. According to the equation, Poland’s and Germany’s success can take values from 0 to 100. The value 0 means that a state achieved full success since the outcome is perfectly consistent with its salience-weighted policy position. By contrast, the value 100 means a complete failure—the greatest possible salience-weighted distance between the outcome and a state’s position.

I also created several independent variables. H1 is tested with the POL-GER variable which measures the absolute distance between the policy positions of Poland and Germany on an issue. This predictor ranges from 0 to 100, where the value 0 means that both countries had identical preferences, while the value 100 indicates a total discrepancy between their positions.

To test H2, I created the variables POL-FRA, POL-UK, POL-ITA and POL-ESP. Each measures the absolute distance on the 0–100 scale between the policy positions of, on the one side, Poland and, on the other side, France, the UK, Italy and Spain on an issue. The closer these variables are to 0, the greater is the similarity of their preferences. Likewise, I constructed similar predictors devoted to Germany: GER-FRA, GER-UK, GER-ITA and GER-ESP.

In addition, I created several groups of control variables emphasized in the literature on EU bargaining. The first captures the effect of the Parliament’s position on Poland’s and Germany’s success. The reason is that this institution plays a key role in adopting EU legislation (Kardasheva 2013). Hence, two continuous variables were constructed: POL-EP and GER-EP, measuring the absolute distance between the policy positions of the Parliament and these two countries on an issue.

Besides the EP, an important role in EU lawmaking is played by the Commission. This institution possesses the agenda-setting power by preparing legislative proposals which are then negotiated by the Council and EP (Osnabrügge 2015; Kreppel and Oztas 2017). Several studies found that member states are more successful when they hold a preference close to the Commission (Arregui 2016; Cross 2013; Lundgren et al. 2019). Therefore, two control variables were designed: POL-COM and GER-COM. Both define the level of proximity between the policy positions of Poland, Germany and the Commission on an issue, measured as the absolute distance on the 0–100 scale.

Bargaining success may also be influenced by the extremity of an actor’s policy position. Previous studies discovered that states with more extreme preferences are less successful in achieving their preferred outcomes (Arregui 2016; Arregui and Thomson 2009; Bailer 2004; Cross 2013; Golub 2012; Lundgren et al. 2019). Hence, I created two variables: POLEX and GEREX that measure the absolute distance between Poland’s and Germany’s positions and the average position held by other member states on an issue.

The literature also highlights the relevance of the status quo in EU decision-making. It can be defined as the outcome that would occur in the event of no agreement. Some studies found that states with a position closer to the status quo are more successful (Arregui 2016). Hence, I created two variables: POLSQ and GERSQ. Each is equal to 1 if a state held the status quo position, and 0 if otherwise or there was no status quo during negotiations.

Oftentimes, EU proposals are multidimensional, i.e. they contain multiple issues on which states hold different positions. The literature argues that actors are more successful when such files are negotiated (Aksoy 2012; Kardasheva 2013). The explanation is that multidimensional legislation creates opportunities for package deals: actors can gain a better outcome by exchanging their support across issues to which they attached different preference intensities. Hence, I created the Multidimensional variable. It takes the value 1 for proposals with more than one issue, and 0 for files involving one issue.

Additionally, I controlled for the legislative procedure by creating the Procedure variable: 1 for proposals adopted under co-decision, and 0 for consultation files. Since states’ bargaining success can also vary across different policy areas, I included dummy variables capturing the policy area a file addresses. I used the responsible Commission’s Directorate General which issued the proposal as a proxy.

A linear regression model

To test the hypotheses, I utilized a multiple linear regression model (Fox 2008). However, its use in the present study encounters two problems. First, observations are not completely independent due to the hierarchical structure of the data. Since issues are nested in proposals, it is likely that Poland’s and Germany’s success may strongly depend on the characteristics of the proposal to which an issue belongs. Second, the variance of errors is not constant which violates the assumption of homoscedasticity. To resolve these problems, OLS regression models were estimated with robust standard errors clustered at the proposal level (Cameron and Miller 2015).

Descriptive statistics

Figure 2 reports the actors’ level of success measured by their salience-weighted distance from the outcome.Footnote 5 To remind: the smaller the distance, the greater the actor’s success. In the analysed period, Poland’s policy position was on average 21.42 salience-weighted points away from the final outcome on the 0–100 scale. This means that out of 29 actors Poland occupied only 21st place (ninth from the end) in terms of bargaining success. This result can be perceived as unsatisfactory given that Poland is in the sixth place in terms of voting power. But Germany also punches below its weight. In the years 2004–2009, the position of this state was on average 22.82 salience-weighted points distant from the final outcome. As a corollary, Germany occupied only 25th place (fifth from the end) which is a dramatically poor result given that this state holds the highest voting power in the Council. Thus, Fig. 2 refutes the common opinion that Germany is the most dominant actor in EU lawmaking.

Fig. 2
figure 2

Source: own calculations based on Thomson et al. (2012)

Actors’ salience-weighted average distances from the outcome in 2004–2009.

Full size image

Another issue that is worth examining is the similarity of Poland’s and Germany’s preferences. Figure 3 illustrates average distances on the 0–100 scale between the policy positions of Poland, Germany and other actors on issues negotiated in 2004–2009. It shows a strong divergence between Poland’s and Germany’s preferences. In the analysed period, Poland was on average about 43 points distant from Germany on a scale of 0–100. When excluding the Commission and the EP, this result situates Germany as Poland’s biggest opponent and the last partner having the most discrepant preferences from all member states. However, the same conclusion applies to Poland when looking from the Germany’s perspective. As Fig. 3 shows, Poland is on the fifth position from the end in terms of preference compliance. This result indicates not only the presence of a great divergence of interests between both countries, but also a serious difficulty in reconciling them in EU lawmaking.

Fig. 3
figure 3

Source: own calculations based on Thomson et al. (2012)

Average distances between Poland’s, Germany’s and other actors’ policy positions.

Full size image

Figure 4 illustrates the similarity of Poland’s and Germany’s policy positions across policy areas defined according to the Commission’s DG responsible for the proposal. It shows that over the period considered, both states had the most consistent preferences in the area of external relations where the average distance between their positions was only 8 points on the 0–100 scale. One example of a strong preference compliance in this field is a proposal establishing a European Neighbourhood and Partnership Instrument (ENPI). This file raised two controversial issues on which Poland and Germany had the same position: 1) the involvement of the Parliament in the implementation of the ENPI and 2) the allocation of funds between Southern and Eastern neighbours. On the first issue, both states demanded the EP to only be consulted in some areas of the ENPI instead of having co-decision powers (position 30). On the second issue, they sought to introduce general criteria on which potential recipients of EU external aid would compete for funding (position 0; these criteria were expected to favour the Eastern neighbouring countries when allocating funds) instead of giving specific weights to historical allocations (which would favour Southern neighbours).

Fig. 4
figure 4

Source: own calculations based on Thomson et al. (2012)

Average distances between the policy positions of Poland and Germany according to the policy area.

Full size image

In contrast, Poland and Germany had the most divergent preferences in the area of internal market and services. Their policy positions were on average 94 points distant on the 0–100 scale, indicating that they were completely divided in almost all issues belonging to this policy area. Consider a directive on the internal market of postal services that raised two controversial issues: the protection of existing (usually national) postal service providers that are large employers and the timing of full liberalization. While Poland sought to introduce the protection for existing companies within the framework of EU-wide liberalization (position 33) and wanted the full opening of postal markets to took place later than 2010 (position 0), Germany called for a full liberalization without any exceptions (position 100) and by 2009 (position 100). Figure 4 also leads to the surprising conclusion that, contrary to the conventional wisdom, Poland and Germany did not differ significantly in the field of environment or energy in the years 2004–2009. For instance, both states opposed the introduction of new recycling targets for waste in a directive on waste and the setting of long-term CO2 emissions targets for new passenger cars in a regulation on car emissions. They also took conservative positions on the proportion of carbon credits the aviation industry should be allowed to auction (directive on emission allowances in the aviation sector).

The effect of proximity to Germany on Poland’s success

Table 1 reports the results of linear regression. The dependent variable is Poland’s success. In total, four OLS models were estimated. Model 1 contains only the POL-GER variable. Model 2 complements model 1 with four predictors related to H2: POL-FRA, POL-UK, POL-ITA and POL-ESP. Model 3 adds control variables to model 2, while model 4 supplements model 3 with policy area fixed effects. Figure 5 plots the substantive effects of the variables that are found to be statistically significant in the best-fitting model—model 3. As a robustness check, I estimated other model specifications in the Online Appendix, including those with unweighted bargaining success as the dependent variable and salience as an independent variable. The results are consistent across all models.

Table 1 Regression results—predictors of Poland’s bargaining success in EU lawmaking
Full size table
Fig. 5
figure 5

Substantive effects of significant variables on Poland’s bargaining success (based on model 3)

Full size image

H1 was corroborated in the analysis. As expected, holding preferences close to Germany provides Poland with a higher level of bargaining success in EU lawmaking. This is indicated by the β coefficient of the POL-GER variable which is positive and statistically significant in all models. This result can be interpreted in two ways. First, the greater the divergence of Poland’s and Germany’s preferences, the greater the salience-weighted distance between the Poland’s policy position and the final outcome on an issue, and thus the lower the Poland’s success. Second, the more the positions of both countries are convergent, the smaller this distance is, which translates into a higher Poland’s success. Holding other variables at a constant level, a one point decrease in the distance between the positions of Poland and Germany on the 0–100 scale leads to a 0.1–0.212 points increase in the level of Poland’s success (i.e. a decrease in the salience-weighted distance between Poland’s policy position and the final outcome). According to model 3, when Poland is 100 policy scale points away from Germany, its salience-weighted distance from the outcome increases by 13 points. In addition, in model 1, the POL-GER variable alone explains 13.8% of the variance of Poland’s success, which is about 1/4 of the variance of the model containing all variables (model 4). Hence, the constellation of preferences with Germany is seen to be a relevant predictor of Poland’s bargaining position. This conclusion is also reinforced by the observation that the POL-GER variable is positive and significant even after controlling for other variables (see models 3 and 4). This means that the incremental effect of proximity to Germany on Poland’s success is robust to, inter alia, Poland’s preference extremity, its closeness to the status quo or a policy area the proposal addresses. In summary, the results show that holding a policy position close to that of Germany is extremely beneficial for Poland. Therefore, Poland should cooperate extensively with Germany in order to approximate their positions and work out a common stance on EU legislation. However, as the descriptive analysis has shown, this is seriously hampered by a huge divergence of both countries’ initial preferences.

The analysis also confirmed H2, but with one exception. According to the results, holding a position close to that of Germany guarantees greater bargaining success for Poland than its closeness to the policy positions of three other member states with the highest voting power, namely France, Italy and Spain. This conclusion stems from the observation that while the POL-GER variable is still positive and statistically significant in models 2–4, the POL-FRA, POL-ITA and POL-ESP variables are not. Hence, Poland is not much better off when it holds a preference close to France, Italy or Spain, but at the same time, its closeness to Germany translates into higher success.

However, the analysis revealed one deviation from the effect assumed in H2, related to the UK. The coefficient of the POL-UK variable is found to be positive and significant in models 2–4. Hence, Poland is more successful when its position on the issue is closer to that of the UK, while a greater discrepancy between their preferences provides Poland with less bargaining success. Ceteris paribus, each one-point increase in the distance between the positions of Poland and the UK leads to a 0.186–0.220 increase in the salience-weighted distance between Poland’s preference and the final outcome. Interestingly, the coefficient of the POL-UK variable is even higher than that of the POL-GER measure, suggesting that the proximity to the UK is more beneficial for Poland than holding a position close to Germany. Overall, the obtained result can be explained by the fact that similarity of policy positions of Poland and Germany may produce a greater opportunity to build a blocking minority. Both these states are among actors which most often contest legal acts during the votes in the Council (Van Aken 2012). As a result, after co-opting several smaller member states, they are able to build a blocking minority and force other actors to give concessions (Warntjen 2017). At the same time, the obtained result leads to the conclusion that Brexit may significantly affect Poland’s bargaining capabilities. On the one hand, a strong divergence of preferences with UK would no longer translate into less Poland’s success. But on the other hand, Poland would not be able to significantly strengthen its negotiating position by building a blocking minority with London (Frantescu 2017).

Turning to the control variables, the analysis revealed that the proximity to the Parliament has a strong effect on Poland’s success. The coefficient of the POL-EP variable is positive and significant in models 3 and 4. This means that increasing the distance from the EP decreases on Poland’s bargaining success, while sharing preferences with this institution translates into higher success. This result is consistent with previous studies showing that member states are more successful when their positions are closer to that of the EP (Arregui 2016; Arregui and Thomson 2009; Bailer 2004; Cross 2013). By contrast, the similarity of preferences with the Commission has no visible impact on Poland’s success. The β coefficient of the POL-COM variable is positive, though it does not have statistical significance and its value is small. This result is in contradiction with earlier studies showing that Council’s members are more successful when the Commission supports their policy positions (Arregui 2016; Cross 2013; Lundgren et al. 2019). Hence, the obtained result may suggest that moving preferences closer to the Commission is beneficial only for selected countries.

No significant relationship was found between Poland’s bargaining satisfaction and the extremity of its policy position. This result is surprising in the light of previous studies showing that member states with more extreme preferences are less successful in EU lawmaking (Arregui 2016). A closer inspection of this effect leads to the conclusion that the preference extremity has a negative and statistically significant impact on Poland’s success (see models 7 and 8 in the Online Appendix), but it loses this significance when taking into account the variables measuring the similarity of Poland’s policy positions with that of other member states holding the highest voting power, i.e. POL-GER, POL-FRA, POL-UK, POL-ITA, POL-ESP. This suggests that through adequate coalition strategy and moving preferences towards congruent pivotal member states, Poland can mitigate the negative effect of holding extreme preferences on its bargaining success. Moreover, the coefficient of the POLSQ variable is not significant, indicating that Poland is not better off when it maintains the status quo position. This result is in contradiction with previous findings: while earlier studies found the closeness to the status quo to generally affect member states’ bargaining success (Arregui 2016), this effect is not present in the specific case of Poland.

Also, the multidimensionality of the proposal is not a relevant determinant of Poland’s success. This suggests that this state does not benefit from the possibility of logrolling offered by proposals containing multiple issues in order to extract better negotiation outcomes. Likewise, the legislative procedure is not a significant factor when explaining Poland’s success.

The effect of proximity to Poland on Germany’s success

Table 2 reports the results of regression analysis devoted to Germany’s success. Similarly, model 1 contains only the POL-GER variable, and model 2 supplements model 1 with four predictors related to H2, while model 3 and model 4 add control variables. In Fig. 6, I plot the substantive effects of statistically significant variables in the best-fitting model—model 3. In the Online Appendix, I provided other model specifications (with unweighted success). Once again, the findings are similar.

Table 2 Regression results—predictors of Germany’s bargaining success in EU lawmaking
Full size table
Fig. 6
figure 6

Substantive effects of significant variables on Germany’s bargaining success (based on model 3)

Full size image

Like in the case of Poland, the analysis confirmed H1 expecting Germany’s success to be higher when this state holds a policy position close to that of Poland. This is indicated by the β coefficient of the POL-GER variable which is positive and significant in all models. This means that the more divergent are the preferences of Poland and Germany regarding a legislative issue, the greater is the distance between the position of Germany and the final outcome of negotiations, implying smaller Germany’s success. This finding can also be interpreted in the opposite way—greater similarity of the positions of these countries provides Berlin with a higher level of success. Ceteris paribus, each one-unit increase in the distance between the positions of Poland and Germany increases the distance of Germany’s position from the final outcome by 0.109–0.202 salience-weighted points. Based on model 3, when Germany is 100 points away from Poland’s policy position, its bargaining success decreases by about 13 scale points. Additionally, the proximity to Poland alone explains about 12% of the variance of the dependent variable in model 1, which is about 1/3 of the entire variance generated by the model with all variables (model 4). This implies that, as in the case of Poland, the similarity of both countries’ preferences is a relevant predictor of Germany’s success. Moreover, the coefficient of the POL-GER variable is positive and significant even after including many control variables in models 3 and 4, indicating that the incremental effect of the Polish–German proximity on Germany’s success is independent of other key factors, such as the preference extremity, closeness to the status quo, procedure or policy area. In sum, not only does the similarity of the Polish and German preferences positively affect Poland’s bargaining success as shown in the previous chapter, but also it brings tangible legislative gains to Germany.

H2 was also corroborated in the analysis. This conclusion stems from three observations. First, the coefficients of GER-FRA, GER-ITA and GER-ESP variables are not statistically significant in any model and their values are small. Hence, holding a policy position closer to that of France, Italy and Spain does not increase Germany’s bargaining success. Interestingly, this study shows that, contrary to the conventional wisdom, cooperation with France (often emphasized in the public discourse as the “Berlin-Paris axis”) is not beneficial for Germany. Second, while the GER-UK variable is positive and strongly significant in model 2, it loses its significance after including control variables in models 3–4. Therefore, while a greater compatibility with the UK preferences generally translates into a higher level of Germany’s success, the occurrence of this effect is strongly dependent on other factors, such as the preference extremity or policy area. Third, while GER-FRA, GER-UK, GER-ITA and GER-ESP are not significant in models 2–4, the POL-GER variable is. Importantly, this effect even persists irrespective of other control variables. Hence, as expected by H2, holding or moving a preference closer to that of Poland brings evidently more success to Germany than its proximity to positions of other pivotal member states holding the highest voting power.

Turning to the control variables, the proximity to the European Parliament is a relevant predictor of Germany’s success. This is evidenced by a significant and positive value of the GER-EP variable. Hence, consistent with previous findings (Arregui 2016) and similar to the case of Poland, holding a preference close to the EP translates into a higher level of Germany’s success. By contrast, the proximity to the Commission’s policy position has no effect on Germany’s success. Hence, bringing preferences closer to the Commission—which was generally found to have a positive effect on states’ bargaining satisfaction (Arregui 2016; Lundgren et al. 2019)—is not beneficial for Germany.

Surprisingly, the extremity of Germany’s position was not found to affect its bargaining success, as shown by the insignificant GEREX variable. This result is contradictory to earlier studies showing that member states with extreme preferences are less successful (Arregui 2016). However, after excluding variables capturing the level of preference compliance with pivotal member states, GEREX reveals statistical significance and still has a positive sign (see models 7 and 8 in the Online Appendix). This suggests that, as in the case of Poland, while having an extreme preference is generally disadvantageous in terms of reaching ideal outcomes, Germany is able to mitigate this negative effect by bringing its position closer to selected pivotal actors. Additionally, the coefficient of the GERSQ variable is positive and insignificant, indicating that the maintenance of the status quo position does not increase Germany’s bargaining satisfaction. Given that previous studies discovered that member states with positions closer to the status quo are more successful (Arregui 2016), this study shows that this effect does not apply to Germany. Finally, Germany’s bargaining success is not dependent on the multidimensionality of the proposal and the legislative procedure. None of these variables were found to be statistically significant.

Conclusions

This article explores how the similarity of Poland’s and Germany’s preferences affects their bargaining success in EU lawmaking and whether their cooperation is more beneficial than moving positions closer to other pivotal member states, namely France, the UK, Italy and Spain. Overall, three key conclusions can be drawn from the analysis.

First, despite geographical proximity and shared cultural background, there is a strong discrepancy between Poland’s and Germany’s preferences. Out of all member states, Poland has the most incompatible policy positions with Germany, which places Berlin as its last coalition partner. However, the same conclusion applies to Poland when looking from the Germany’s perspective. These observations suggest that cooperation between both these states through bringing their preferences closer or building a coalition in EU lawmaking is seriously cumbersome.

Second, despite having divergent preferences, mutual cooperation is extremely profitable for Poland and Germany. The analysis shows that both these states are more successful in EU lawmaking when they approximate or hold similar positions on legislation. Importantly, this effect is significant even after controlling for the preference extremity, the maintenance of the status quo position, closeness to the Parliament and the Commission, the proposal’s multidimensionality, legislative procedure or policy area. Hence, by bringing their preferences closer Poland and Germany are more likely to succeed, irrespective of other negotiating conditions, while avoiding cooperation and raising explicitly contradictory demands significantly increases the probability of their failure.

Third, cooperation between Poland and Germany provides both these countries with greater success than bringing preferences closer to other pivotal member states with the highest voting power. In the case of Poland, this country is not much better off when moving positions towards France, Italy or Spain, whereas its proximity to Germany is still advantageous in terms of bargaining success. However, a greater compatibility with the UK preferences also translates into higher success, suggesting that a potential Brexit may negatively affect Poland’s bargaining capabilities. By contrast, the analysis devoted to Germany showed that its success is significantly and positively associated only with the proximity to Poland, while bringing preferences closer to either France, the UK, Italy or Spain is not beneficial.

Apart from that, the article contributes to the literature on EU bargaining in four additional facets. First, it shows that member states can mitigate the negative effect of preference extremity on their success by moving their positions closer to the most powerful actors. While previous studies showed that states with more extreme positions are less successful (Arregui 2016), this relationship was not found in the case of Poland and Germany. A more in-depth analysis reveals that the extremity of preferences decreases these states’ success, but it loses its significance when including the variables measuring the similarity of Polish and German policy positions with those of France, the UK, Italy and Spain. Therefore, future contributions should take into account not only the centrality of actors’ preferences, but also their proximity to pivotal states.

Second, while earlier studies revealed that holding a preference close to the Commission positively influences a state’s bargaining success (Cross 2013; Lundgren et al. 2019), this study has not confirmed this effect for Poland and Germany. This suggests that cooperation with the Commission is not beneficial for all member states.

Third, no link was found between the maintenance of the status quo and Poland’s and Germany’s success, which is inconsistent with previous studies (Arregui 2016). This suggests that either the closeness to the status quo is profitable for selected states or it is not a good determinant of bargaining success as argued elsewhere (Thomson et al. 2006; Thomson 2011).

Fourth, surprisingly, multidimensional proposals were not found to affect Poland’s and Germany’s success. This result indicates that not all states are capable of exploiting the opportunities for an issue linkage to increase their bargaining satisfaction. Hence, future contributions should examine which states are more successful in applying this technique.