Punching Below Its Weight: The Role of the European Parliament in Politicised Consultation Procedures

With Lisbon, the European Parliament formally acquired an equal standing to that of the Council of the EU in the making of policies in the AFSJ (area of freedom, security and justice). However, the growing political salience of policy issues at stake and bottom‐up politicisation in the AFSJ has had the unintended effect of undermining the European Parliament’s internal unity even under consultation procedures. To show how this played out in practice during Europe’s migration and refugee crisis, this article analyses the European Parliament’s role, preferences, and bargaining position in the making of two Refugee Relocation Decisions (Council Decisions 2015/1523 and 2015/1601) under consultation procedure. To do so, this article exploits Putnam’s two‐level framework (level I and II politics throughout the policy‐making process) to explore early agenda‐setting attempts and groups’ positions on issues of refugee relocation and burden‐sharing, as they were for‐ mally stated in their position papers and expressed at the LIBE Committee and at plenary. This article shows that the high domestic salience and politicization of the issues at stake left MEPs torn between competing principals at home andwithin their European Parliament political groups and had the effect of weakening overall unity on the issue of refugee relocation.


Introduction
It is undeniable that, over time, and particularly after the adoption of the Lisbon Treaty, the European Parliament (EP) has seen its power and influence as co-legislator grow remarkably in the AFSJ (area of freedom, security and justice; see Hampshire, 2016;Hix & Høiland, 2011;Trauner & Ripoll Servent, 2016). While growing used to behave more consensually under co-decision, members of the European Parliament (MEPs) remain portrayed as being comparatively more confrontational under consultation procedures, pushing forward 'Christmas wish lists' with 'left-wing, libertyoriented positions' in justice and home affairs (JHA; Ripoll Servent, 2012, p. 67; see also Ripoll Servent, 2015).
In many ways, the entry into force of the Treaty of Lisbon and the gradual move away from consultation provided the EP with a timely opportunity of institutional adaptation (Ripoll Servent, 2012): That is, to close the gap between these two 'schizophrenic' behaviours in co-decision vis-à-vis consultation procedures (Ripoll Servent, 2012, p. 68). Nevertheless, more recent policy developments in the AFSJ-most notably, the use of consultation for formulating key policy responses to the migration crisis (i.e., the Refugee Relocation Decisions under Council Decisions 2015/1523 and 2015/1601 and the EU-Turkey Statement) and the political impasse reached under various co-decision procedures (e.g., the so-called Dublin Regulation Recast and the Asylum Procedure Regulation)-have suggested that member states 'remain privileged policy entrepreneurs in the AFSJ' (Trauner & Ripoll Servent, 2016, p. 1429. The missed change for the EP to transform the Treaty of Lisbon into an opportunity for institutional change became all the more visible in the shaping of emergency responses to the migration and refugee crisis, particularly the two Refugee Relocation Decisions (Council Decision (EU) 2015/1523of 14 September 2015Council Decision (EU) 2015/1601of 22 September 2015, adopted under a special consultation procedure as provided by Article 78(3) of the Treaty on the Functioning of the European Union. Soon after the two Refugee Relocation Decisions were adopted, former EP President Schulz criticized member states for being responsible of unambitious policy responses to the crisis: The European supranational institutions have shown their readiness to act. The European Parliament has supported the European Commission-the EU's executive-in its courageous push for a binding system to help the countries most exposed to the refugee crisis. On the other hand, EU member states often preach solidarity when it suits them and resist it when it does not….It is not the European Unionor Brussels-that is broken. It is the intergovernmental decision-making process jealously guarded by national capitals that has once again proven its ineffectiveness. (Schulz, 2015) It is indeed true that under consultation and faced by conditions of structurally limited power and influence in the legislative procedure, the EP could only 'lobby' other institutions (Hix & Noury, 2009, p. 19) and try to demand radical change in a united manner to be somewhat influential (see, for instance, Kardasheva, 2009;Varela, 2009). At the same time, in view of the heated intergovernmental debate and impasse found at the Council on the refugee relocation decisions (Barigazzi & de La Baume, 2015), this article explores whether the highly salient 'nature of the problem and the absence of a shared "common bad" ' (Ripoll Servent, 2019, p. 307) were reflected in a less united, thereby even less influential EP. In other words, the article analyses whether the high political stakes and domestic salience on the issues at stake cast "the shadow of intergovernmentalism" on intra-EP dynamics too, in such a way that they prevented the EP from pushing forward an ambitious, maximalist agenda on these issues, and undermined any chance for the institution to act united in the formulation of these two refugee relocation decisions.
In order to study this, the article analyses the role, bargaining position and preference formation of the EP in the formulation of the refugee relocation decisions. The article argues that the EP initially tried to strengthen its odds in influencing the negotiations by gathering broad support across political groups and crafting a Report in full respect of the existing competing interests on the issues of irregular migration and asylum.
As shown in this article, between 2013 and early 2015 the EP was arguably successful in its efforts as a whole, pushing for policy change on the issues of refugee relocation and intra-European solidarity. However, as media salience on the issues at stake increased in the Spring of 2015, so did public attention on these matters. Increased salience made MEPs more likely to defect, that is to vote against their EP political group's line when a conflict between their national party and their EP group arose. Bottom-up politicization, much alike the one that undermined consensus in the Council (Barigazzi & de La Baume, 2015), pre-empted the EP from being influential in the formulation of the two Refugee Relocation Decisions. This finding is consistent with existing literature on defection and abstentions in the EP, according to which MEPs are more likely to defect or abstain from their EP group's position 'on specific questions that are of particular importance to them, but not on a general basis (Faas, 2003, p. 860; see also Hix, 2002;Klüver & Spoon, 2015;Koop et al., 2018), particularly when these questions coincide with crucial interests of an MEP's national party which need to be respected in order to stand for re-election (Mühlböck, 2017, p. 60).
In this article, Section 2 presents the methodology and theoretical framework in use. Section 3 reviews the EP's attempts of agenda-setting as from 2013. Section 4 analyses the various EP groups' positions on refugee relocation at the outbreak of the migration and refugee crisis, so as to serve as the basis for comparison between early political objectives in the unfolding of the crisis and the ultimate policy-making outcome. Sections 5 and 6 analyse level I and level II coalition-building efforts and power dynamics within the EP.

Theory and Methods
Seeing as the main preoccupation of this article is to understand whether, and if so in what ways, the entanglement between domestic and international politics played out in undermining the intra-EP unity under consultation, some theories of European integration could be seen as the best fitting theoretical framework. However, rather than linking the findings of this article to a specific theory of integration, this article will critically contribute to the wider debate on European integration by showing how to explain the growing internal divide in the EP in making policies in the AFSJ, even under consultation. This methodological choice was considered most appropriate in view of the inherent weaknesses of existing theories of European integration to explain EU bargaining dynamics. According to Putnam himself, neo-functionalist theories have the crucial shortcoming of disregarding the impact of domestic politics in favour of a more transnational, interdependencebased approach, using as dependent variable 'the hypothesized evolution of new supranational institutions, rather than specific policy developments' in such a way that whenever 'European integration stalled, so did this literature' (Putnam, 1988, p. 431). Even statecentric theories of integration, such as Liberal and new Intergovernmentalism, are too static and narrow to account for the longer 'time horizon preferences' and issue-linkages (Moravcsik, 2018(Moravcsik, , p. 1667) which usually drive the behaviour of policy actors in the AFSJ. While post-functionalism could be somewhat useful to account for the centrality of exclusive identity in shaping MEPs' voting behaviour and ultimate decision to defect (Hooghe & Marks, 2009), as well as the influence of national parties on the EP, it does not tell us much regarding interactions at the EU level of governance other than noting that domestic politics influences the constraints of decision-makers (Schmidt, 2019(Schmidt, , p. 1025. In an attempt to deconstruct political incentives coming from the domestic and the supranational arenas, the article's analytical approach follows Putnam's (1988) two-level game theory. This analytical framework was deemed most suitable for the purposes of this study as it allows for the study of simultaneous interactions, pressures and influences between the EU (level I) and domestic (level II) levels of governance, while keeping a focus on EU politics. Under Putnam's two-level framework, chief EU negotiators respond to, and interact with, different pressures, interests, and diplomatic strategies that contribute to, or ostracize, a tentative agreement among the parties at stake. For the purposes of this article, different EU policy actors on level I (different EP political groups, national delegations and MEPs) are seen as responding to various domestic pressures originating from level II (domestic constituencies, parties, or governments in place).
This article is mainly preoccupied with the definition of win-sets or of 'all possible Level I agreements that would "win," that is gain the necessary majority among the [domestic] constituents, when simply voted up or down' at the EP (Putnam, 1988, p. 437). By providing the right analytical tools for breaking down the win-sets of different EP political groups on the basis of their stated position and internal debate on a mechanism for refugee relocation, as well as for studying how domestic politics influenced MEPs' voting behaviour and the degree of intra-group cohesion and intra-EP overall unity, the two-level game framework constitutes the most suitable model for analysing how levels I and II competing political interests influence policy outcomes at the EP.
As shown in Figure 1, Putnam's two-level game defines clear mechanisms based on which multi-party negotiations take place. In situations of high politicisation at level II, we would expect negotiations at level I to be politicised as well, ultimately rendering the size of the political groups' win-sets smaller across the political spectrum (shrinkage of win-sets or tie-hands). The more EP political groups are united and internally cohesive on the issue at stake, the more we would expect chief negotiators in EP political groups (i.e., rapporteurs and shadow rapporteurs) to try and restructure the Commission's proposal by: a) pushing forward an ambitious wish list of amendments in the EP Report on the second Refugee Relocation Decision (European Parliament, 2015c); and b) demanding more concessions or tying its own hands by pretending that there is no room for manoeuvre at the domestic level. Vice versa, the more decentralized the governance of a political group is, the higher the likelihood will be for MEPs to escape their mandate originating from level II (Ripoll Servent, 2014, p. 372) and to not align with the group's position. In this circumstance, we would expect chief EU negotiators that want to achieve a united position at the EU level to cut slack to their MEPs so as to widen their win-set, ultimately resulting in a less ambitious group position. Another mechanism that we expect to have affected intra-EP negotiations is a change of interests throughout the negotiations, often due to the 'reverberation' of international pressures within domestic politics (Putnam, 1988, p. 456 Moravcsik (1993, p. 32) and Ripoll Servent (2014, p. 572).
In terms of methods, the article is built upon a multi-method research framework rooted in the deductive theory-testing side of process tracing, as defined by Bennett and Checkel (2014, pp. 7-8) as a methodological technique that 'examines the observable implications of hypothesized causal mechanisms within a case to test whether a theory on these mechanisms explains the case.' By operationalising Putnam's two-level game, the article looks at the causal mechanism between the observed outcome, or the unambitious role of the Parliament in the formulation of the two Refugee Relocation Decisions, and the partial explanation of bottom-up politicization of the policy issue at stake. As shown in the following sections, the particular legislative procedure in use-consultation-in the making of the two selected policies was a necessary and sufficient condition for the EP to be a structurally marginal policy actor. However, what rendered the EP's policy agenda even more unambitious was the lack of intra-EP unity, which resulted in unusually conservative opinions in the two procedures.
This framework is integrated with a set of primary and secondary sources, including: six semi-structured elite interviews undertaken between October 2018 and January 2019 (see Annex I in the Supplementary File); vote analysis; and documentary research, spanning from press briefings to European Council conclusions, from interviewees' meeting calendars to EP group position papers on migration. The findings are triangulated with secondary literature on, and media coverage of, the intergovernmental and interinstitutional debate preceding the adoption of the two Decisions.

The EP and Early Agenda-Setting on Refugee Relocation (2013-2015)
From as early as 2013, the EP attempted in different occasions to propose a Union-wide relocation mechanism for refugee redistribution. In a debate in late February 2013 (European Parliament, 2013a), the EP questioned the then JHA Commissioner Cecilia Malström about the Commission's commitment to table a legislative proposal for an intra-European relocation scheme, shaped on the basis of the past pilot projects of intra-EU refugee relocation for the benefit of Malta (the so-called EUREMA I and II). While stating that the Commission was 'very happy with the EUREMA scheme,' Malmström concluded, as based on the limited commitment of member states in implementing the pilot project EUREMA II (only 14 were relocated), that 'there [was] not the [right] political climate… to propose such a scheme' without incurring in 'a robust no…[or] a paper tiger' (European Parliament, 2013a).
Despite the clear political unworkability of relocation, Lampedusa's migrant shipwreck of 3 October 2013 gave the EP enough political momentum to reopen a debate about the situation in the Mediterranean and possible reforms of the AFSJ. A first EP resolution was pub-lished just days after the tragedy in the Mediterranean, encouraging member states and the Commission, among other things, to show 'greater solidarity with Member States facing particular pressure' (European Parliament resolution of 23 October 2013, 2013b). Less than two weeks later, the EP published another resolution, reiterating the legal obligation for member states to assist migrants at sea, as well as the need for responsibilitysharing. It recommended 'creating a mechanism based on objective criteria to reduce the pressure on those Member States receiving higher numbers of asylum seekers and beneficiaries of international protection, in either absolute or proportional terms' (European Parliament, 2013c, p. 6).
While the immediate public and political response to Lampedusa's tragedy was to pledge for 'no more deaths' in the Mediterranean (Muiznieks, 2015), the increased resources invested in search and rescue operations did not have the results expected in the short run. The EP gathered renewed momentum on the issue under the Italian Council Presidency and with the start of a more "political" Commission under Juncker's leadership (Kassim & Laffan, 2019;Nugent & Rhinard, 2019). A critical juncture on the issue was only to come on 18 April 2015, when a shipwreck disaster leaving over 700 migrants dead shook the European Council and prompted a EUCO special summit on 23 April 2015, where member states called on the Commission to 'consider options for organising emergency relocation between all member states on a voluntary basis' (Council of the European Union, 2015, p. 2). This message was reiterated in another EP resolution, which called 'on the Commission to establish a binding quota for the distribution of asylum seekers among all the Member States' (European Parliament resolution of 29 April 2015, 2015d). In other words, in spite of the EP's active efforts in setting the policy agenda on the issue at stake, it was only when the Commission's willingness to table a refugee relocation scheme was echoed by the European Council that a proposal for voluntary relocation was tabled by the Commission.

EP Groups' Positions on Refugee Relocation and Resettlement
While the EP as a whole lobbied member states and other EU institutions with one single voice, all groups came up with a different position paper ahead of the EUCO special summit on migration of April 2015: These papers reflected primarily the groups' sets of acceptable agreements, or win-sets, on the issues of refugee relocation and burden-sharing. Each group's position was evaluated, as summarized in Table 1, in relation to its published position paper (if existing). In so doing, it was possible to define: a) the group's desired "policy focus" for a refugee relocation scheme (refugee oriented; MS capacity; executive control; not defined); b) its preference on the nature of refugee relocation (voluntary; binding; substituting Dublin; permanent; temporary); the group's degree of unity on the issue at stake (low; moderate; high); and c) the aggregate size of the group's win-set (narrow or broad). Seeing as neither the Eurosceptic Europe of Freedom and Direct Democracy Group (EFDD) nor the far-right Europe of Nations and Freedom Group (ENF) came up with a position paper in the outbreak of the migration crisis, it was not possible to determine either of these variables for them.
The most proactive groups were the Greens/EFA and the GUE/NGL. The Greens/EFA argued for 'visa-free travel' for Syrian refugees and for a relocation system focused on the refugees' preferences, language and culture in order to replace the Dublin Regulation (Greens/EFA, 2015, p. 2) and most importantly to reform the way in which the latter assigns responsibility for examining asylum applications to the countries of first entry in the EU, thereby putting enormous pressure on frontline Member States. The GUE/NGL threatened that their MEPs would be opposing any EU budget that would go against the 'activation of Temporary Protection Directive 2001/55/EC… increase [d] sharing of reception of asylum-seekers between Member States, including through relocation programmes that take fully into account family, language and cultural ties, adequate funding and reception conditions and closing down of detention centres' (GUE/NGL, 2016, pp. 1-2). In other words, strong of its internal cohesion on the issue at stake, the group tried to be influential in interinstitutional negotiations by tying its own hands.
Both the European People' Party (EPP) and the Socialists & Democrats (S&D) supported the idea of a binding mechanism for refugee relocation based on a variety of criteria reflecting the member state's capacity. They nonetheless differed in win-sets on a variety of fundamental issues at stake. The S&D 'wanted a legally binding, permanent relocation' system (Interview 1), with a comprehensive basis for the key redistribution criteria, reflecting both the member state's capacity and the individual asylum seeker's preferences. The EPP instead lobbied for a refugee distribution scheme based on solely objective criteria (EPP, 2015, p. 7), such as territorial size, population, economic situation, and the number of migrants already present in the country. In Putnam's terms, the EPP had a relatively narrower win-set on refugee relocation as compared to the other groups, meaning that it had less room for manoeuvre (tied hands). According to Christian Democrat MEP Jeroen Lenaers from the EPP-at the time also shadow rapporteur on the first file-concerns about asylum seekers' preferences were secondary or even tertiary in the discussion, whereas the whole EPP debate was fully on whether there should be a binding or non-binding mechanism for refugee relocation (Interview 2). As from Spring of 2015, the group initiated a discussion relating to the potential costs and benefits of no agreement, as well as the potential benefits which would originate from the ratification of the Scheme. As explained by MEP Lenaers, the group did not like the option of binding relocation, but abode by it in order to maintain pressure on the Council (cut-slack): It was never our favourite solution… but we understand we are in a crisis situation….The only thing we could do [was] delay the procedure really… which is something we really didn't want to do because the Parliament was in a majority in favour for a binding relocation measure so we wanted to keep pressure on the Member States. (Interview 2) In March 2015, the Alliance of Democrats & Liberals (ALDE) Party had already held a seminar on the topic  , 2015), which then resulted in the creation of ALDE's Blueprint for a New European Agenda on Migration on 23 April: In the document, ALDE called for the introduction of a centralized, two-step refugee distribution scheme in replacement of Dublin. This entailed the voluntary offer of relocation spaces by each Member State; if not enough, this had to be integrated with compulsory redistribution, as based 'on both quantitative data (GDP and the Member State's population) and qualitative data (language, cultural ties, family ties of the refugee)' (ALDE, 2015, p. 6). As explained by an ALDE political advisor (Interview 3), the number of political stances comprised within the group-21 member states and 60 national parties-prompted an early formation of the group's win-set on migration.
The main task for the ALDE shadow rapporteurs and political advisors was to clarify and provide reassurance on the content of the legislative texts to their MEPs. This was done so as to ensure that the procedural complexity and negative press coverage for refugee relocation-possibly threatening a change of domestic interests throughout the negotiations, or reverberation effect in Putnam's words (1988, pp. 454-455)-would not get in the way of showing the group's support at plenary (Interview 3). This proved challenging due to the presence within the Group of eight parties that were at the time part of national governing coalitions (reverberation of domestic interests): In the face of increasing popularity of far-right, anti-migrant parties and movements at home, these parties in government would be sending different (i.e., more realist) messages to their domestic audience (level I) than they would within ALDE (level II). On the political debate surrounding the first Commission's proposal, the same advisor stated: Compared to the other groups, the European Conservatives and Reformists (ECR) clarified that, for them, relocation ought to be based only on voluntary contributions made by governments-i.e., executive decision-making centralized at level II-and was sceptical towards the proposed distribution key, insofar as '[s]tatistics, numbers, and graphs rarely reflect the true local and national effects of decisions in the area of migration and asylum' (Kirkhope, 2015, p. 3).
The only two groups not to clarify their win-sets with a position paper were the EFDD and the ENF: the refusal to do so pertained mainly to the great diversity of polit-ical stances at stake in the two groups, as well as the ENF's late foundation (15 June 2015). As a result, both the EFDD and the ENF left their MEPs 'absolute freedom of vote' (Interview 4), with no attempt at consensusseeking (Interview 5).

Level I Politics in the European Parliament
With the EP's Committee on Civil Liberties, JHA (LIBE) in charge for drafting a report on the two legislative files, the main objective for appointed rapporteur Ska Keller (Greens/EFA) was primarily 'to put down in black and white a bit of what we thought about it, and what were our requests' (Interview 5). The results obtained on the first Refugee Relocation Decision in the relative Committee vote-42 members in favour, 13 against, 3 abstentions-were praised by rapporteur Ska Keller as an example for the Council of unity: While member states are muddling through and cannot agree on how to distribute 40,000 refugees, our committee has supported a binding distribution key by a large majority….We are also calling for a permanent distribution mechanism which must go substantially beyond the current proposals….Respecting the interests of refugees is essential for the success of the distribution key. (LIBE Committee, 2015) As explored in the following paragraphs, the cross-group unity achieved by the rapporteur on this and the second Committee votes came at the cost of a less ambitious and maximalist agenda tabled by the EP as a whole on refugee relocation and responsibility-sharing, in order to bring the Conservatives and Centre-Right on board (cut-slack).
When commenting on how the LIBE report was drafted, an ALDE advisor explained how, while political support from the S&D, ALDE, the Greens/EFA and GUE/NGL was almost automatic in view of 'a kind of alliance… on these issues' (Interview 3), this was not the case for the EPP. According to MEP Lenaers-substitute member for the LIBE Committee-the rapporteur knew that 'she needed at least half of the EPP to get a majority in the European Parliament' (Interview 2). For this to be the case, the rapporteur ensured that the preferences of asylum seekers would only be included 'to the extent possible,' whereas the numbers of refugees to be relocated would be kept unchanged, with a possible adjustment accounting for the evolution throughout the Summer of 2015 (Interview 2). The political compromise found was reflected in a swift change in the amendments postulated in the final EP report: Thanks to these adjustments, fourteen out of the fifteen roll-call votes requested by the ECR and the EFDD and tabled in the plenary of 9 September passed by majority.
While the draft report had called for an amendment in numbers of asylum seekers to be relocated from 40,000 to 50,000 (European Parliament, 2015a, pp. 8, 16), in the final report it was only in an explanatory statement that the rapporteur suggested to increase in the future that number 'to 50,000 as a minimum' (European Parliament, 2015c, p. 34).
A similar adjustment was made in relation to the integration of preferences of asylum seekers and to what was politically attainable for the EPP on this policy matter. As suggested in the final report: Neither refugees have a right to choose their preferred Member State nor do Member States have a right to choose their preferred applicants. But their preferences should be taken into account to the extent possible. (European Parliament, 2015c, p. 35) Alongside said effort in wording, the EPP also ensured that its MEPs who attended the LIBE vote would all be aligned to the found compromise: What we did try to do of course is to make sure that in the LIBE vote, so in Committee vote, we would have… people there that would represent the EPP line. And the EPP group line was that we were in favour of a mandatory binding mechanism. So, for instance, you tried to make sure that [when] people were absent from the vote, that they are being replaced by people who follow the EPP group line and not people who… go against the EPP group line. (Interview 2) This was mirrored in the substitutions made for the LIBE vote on the report: Among the substitute members who were present for the vote on 28 July, four out of eleven were MEPs from moderate parties within the EPP group, namely: Sweden's Moderate Party, the Spanish Partido Popular, the Dutch Christian Democratic Appeal or CDA and the Greek New Democracy. Under rule 200(2) of the EP Rules of Procedure, four out of six of the substitutions made followed the same political rationale and comprised substitutes from the following moderate parties: the German CDU, the Bulgarian GERB, the Belgian CD-V, and the Dutch CDA. The ways in which the EPP quickly aligned to the compromise found despite pressures at level II suggests a high degree of centralization and group discipline, as reported in other scholarly literature (Hix, 2002, pp. 690-691).
The legislative procedure for the second Refugee Relocation Decision-a total of thirteen days from the proposal on 9 September and its adoption on 22 September 2015-was much shorter than the first one. As a result of the short timing and the mostly unchanged legislative text, neither the LIBE nor the Budget Committees produced a different opinion. The only remarkable aspect was to see how domestic contestation weighed in preparatory EP debates (European Parliament, 2015b). In a debate on 8 September, a striking amount of references to domestic pressures and national preferences was reported among MEPs. Non-affiliated MEP Korwin-Mikke from Poland highlighted how his country 'has dozens of problems, [and] we have a problem with migrants,' and defined the relocation policy as an 'absurd policy… this will lead to the destruction of Europe.' Slovak MEP Benová (S&D) brought attention to the fact that 'these issues are viewed differently dependent on which Member State you are talking about' and reproached to EU institutions how they 'did not find a manner to get this issue across the population' and communicate it in an effective manner. Czech MEP Mach (EFDD) threatened that 'if you vote a binding quota through in our country [Czechia], then the public would not like to remain in the EU.' As shown in Section 6, domestic politics had heavy repercussions on the degree and nature of coalition-building at the EP, particularly in the second consultation procedure on 14 September 2015.
Looking at the voting dynamics in the two plenaries (see Table 2), the total number of favouring MEPs (pro votes) in the second Refugee Relocation Decisiondiverged from the first one by 25 percentage points. All centrist and liberal-leaning groups lost between 18% up to 38% of "Yes" votes. While the amount of "No" votes (anti votes) went consistently down except for S&D, on a whole, 15 more MEPs abstained (abst. votes) and 143 more MEPs were absent (absent votes) in the second as compared to the first roll-call vote. The increase in absentees by two to four times was visible in all groups.
Commenting on the change in intra-EPP voting dynamics between the first and second plenary votes, MEP Lenaers stated that there was neither a push for MEPs to vote in favour of the second Refugee Relocation Decision nor to be absent for the plenary: [I]n our...parliamentary group, there is always room, if you have a principled issue, to vote according to your conscience, so that [pushing to be absent] was not...not something we tried....We tried of course to ease with as many people as possible… if you are not in favour of this, try at least, if you could at least abstain instead of voting against. (Interview 2) A similar take on absenteeism was shared by an ALDE political advisor, according to whom there is always within ALDE 'a variable of people half for whatever reason they are not present, maybe [due to a] long session of plenary or they make mistakes' (Interview 3). The substantial increase in defections and absentees makes an analysis of level II politics and voting pressures essential in order to either confirm these takes on absenteeism or put it into question in the face of potential level II pressures coming from national parties or governments in place.

The Impact of Level II Politics in the EP
This layer of voting analysis is particularly useful to analyse whether MEPs from the countries that either rejected or abstained from the second Refugee  Table 3 cross-references this information with the voting behaviour of parties in government from the member states selected above. The Polish delegation to the EPP mostly aligned to the group's request to abstain, when unwilling to support the majority. In fact, the number of Polish MEPs abstaining in the second decision increased by five, as followed by a decrease in "Yes" votes (from 11 to 4): All of the abstain votes from Poland came from the then-governing EPP coalition led by the Civic Platform and its junior coalition partner Polish People's Party. These two parties in government not only led the national Abstain front but also the national efforts at the EP to side with the majority (see Table 3). This could suggest that the EPP's request to abstain instead of voting against was respected by Polish MEPs, in a way anticipating Poland's later inclination to side with the majority at the Council level. The then governing party and its PM Ewa Kopacz were in fact steered by President Juncker towards siding with the majority as they had already foreseen losing the upcoming elections regardless of their voting behaviour on these files (Interview 6).
The voting behaviour of the Romanian delegation to the EP seemingly mirrored Romania's rejection of the second Decision at the level of the Council (see Table 3): Within the governing coalition-composed by an EPP-S&D coalition (Partidul National Liberal, or PNL, and Partidul Social Democrat, or PSD)-the only "Yes" vote came from S&D Vice-Chair Bostinaru, whereas the Similarly, the Slovak delegation voted in line with the country's rejection at the level of the Council rather than the S&D's position, with three out of four MEPs belonging to the governing party Direction-Social Democracy voting against.
This was also the case for the Czech parties in government at the time: Česká strana sociálně demokratická, or CSSD (from the S&D), and its junior coalition partner ANO 2011 (from ALDE) either voted against the Council Decision 2015/1601 or did not attend the plenary, including ALDE Vice-Chair Telička. Commenting on the Czech delegation's voting behaviour, an ALDE political advisor suggested that, despite ALDE's affiliation of the then-PM: There was some kind of acceptance and they explained… for national reasons we have a different position. We did have discussions about other ways around this… but when the nature of these consultations is basically saying yes or no, there's not much room for manoeuvre, or discussions, so they were kind of left behind. (Interview 3) Most MEPs from Hungary's and Finland's parties in government were absent at the plenary, suggesting they may have been subject to pressures coming from their own electorate/government and coalition partner respectively to vote in line with their own country's attitudes at the Council.

Conclusion
This article has conceptualized the negotiations leading to the adoption of two Refugee Relocation Decisions (Council Decision (EU) 2015/1523 of 14 September 2015, 2015; Council Decision (EU) 2015/1601 of 22 September 2015, 2015) as a two-level game involving primarily MEPs as responding to different pressures coming from their own EP political group and national party.
The empirical analysis has shown the importance of this analytical approach insofar as, by means of a twolevel conceptual approach, it has shown how bottomup politicization of the policy issues at stake made the EP more internally divided from within, despite working under consultation, and ultimately led to an unambitious policy agenda. Under consultation, the only negotiating leverage left to the EP would have been to be united to be influential and to advance forward-looking amendments. However, the Rapporteur's need to get the Centre-Right on board to come across as united in front of other EU institutions came at the cost of a less far-reaching Opinion. Indeed, the EPP's concerns were successfully integrated into the final text insofar as the group had a relatively narrower win-set (tied hands) compared to other mainstream EP groups and only sided with the majority on the two Council Decisions as the ben-efits from ratifying it surpassed the costs of blocking it (Interview 3).
As shown in this article, the policy issue(s) at stake reverberated so strongly at home, particularly in Central and Eastern Europe, that level II political considerations were as important as, or even trumped, the MEP's loyalty to their EP political group's line, leaving MEPs torn between "competing principals" and compromising the unity of the EP on this issue. This was particularly the case for MEPs from governing parties in countries that voted against Council Decision 2015/1601. The only successful tactic of level I consensus-seeking consisted in asking MEPs who did not comply to the group's position to abstain, or ensuring those attending the voting session at the LIBE Committee would be supportive of the group's majority line (Interview 2).
These findings represent a key contribution to the scholarly debate on politicization in the European Union (see, for instance, Högenauer, 2017;Schmidt, 2019), as they illustrate how increasing pressures from the bottom and polarized debates at the EU level, such as the ones permeating EU policy-making during the migration crisis, are increasingly weakening the EP's transnationalism due to the competing interests on related policy matters (Högenauer, 2017(Högenauer, , pp. 1105(Högenauer, -1106, with the unintended consequence of further undermining the EP's bargaining position in JHA affairs.