The European Parliament as an Arena and Agent in the Politics of Climate Change: Comparing the External and Internal Dimension

Climate change policy is a key example for globalized multi-level governance, involving the European Parliament (EP) both as legislator of internal European Union (EU) regulation and discursive agent in global climate negotiations. Based on the comparison of decision-making in external and internal climate change policy, the article investigates the link between the role of the EP as a political actor and arena for the interaction of competing party groups: Does EP involvement in negotiations on legally binding legislation prompt or constrain partisan polarization in comparison to declaratory statements about future goals of climate action? Harnessing a discursive institutionalist theoretical framework, the article compares EP resolutions about annual Conferences of the Parties (COPs 20 to 24) with the revision of four legislative acts for Phase IV of EU Climate Action: namely, emissions trading, effort sharing between Member States, the promotion of renewable energies, and energy efficiency standards. Using roll-call voting data and the review of legislative documentation, the contribution tracks the emergence and negotiation of political conflict in these two sets of cases. Overall, the case studies indicate a low level of external politicization, as indicated by low party group polarization, internalization of political conflict at the committee level, and compromise-building between issue dimensions.

Dealing with the problem of climate change is a highly political question, raising strongly contested normative questions concerning the ethical foundations of economic and social order, concepts of justice and equity, and how societies confront questions of risk and uncertainty (Dryzek, Norgaard, & Schlosberg, 2013;Hulme, 2009;O'Brien, Lera St. Clair, & Kristoffersen, 2010;Stevenson & Dryzek, 2014). In this sense, climate change is a key field for research on the politicization of global governance, including the role of legislative institutions.
As a task for empirical research, however, a major challenge for mapping and explaining the political contestation of climate change governance consists of its highly decentralized and multi-level structure: decisionmaking about instruments for mitigating the causes of climate change is not confined to a clearly defined set of democratic institutions within one particular polity. Instead, it evolves within a framework of agreements on the global level, regional policy-making frameworks such as the European Union (EU) as well as national and sub-national constituencies, establishing an idealtypical case of globalized multi-level (or 'polycentric') governance (Gupta, 2014;Jordan, Huitema, van Asselt, & Forster, 2018;Ostrom, 2012;Zürn, 2012).
Against this background, a key puzzle for research about parliamentary institutions in climate governance is their dual role as an actor and arena in a multi-level context: legislative institutions such as the EP combine a role as an arena for the interaction of competing party groups in legislative decision-making with their emergence as a political actor towards other institutions and entities, including at the international level. How are both dimensions linked? Or phrased more specifically, how is decision-making of the EP as an actor of climate governance linked to polarization between its party groups, and what variation do we observe in this regard within different institutional context of EP action? Beyond the specific case of climate governance, this question speaks to key topics of research on the politicization of EU governance, including the links between politics and policy, the role of parliamentary institutions in the contestation of supranational policy, and the party political dimension of EU governance.
This set of questions and topics remains underresearched in the existing literature. So far, most contributions on climate change governance in the EU concentrate on its policy dimension by investigating regulatory approaches and instruments as well as their implementation (Boasson & Wettestad, 2013;Delbeke & Vis, 2015;Dupont & Oberthür, 2015;Ellerman, 2010;Jordan, Huitema, van Asselt, Rayner, & Berkhout, 2010;Oberthür, Pallemaerts, & Roche Kelly, 2010). By comparison, research is still in its early stages concerning the politics dimension, concerning issues such as the framing of climate change in public discourse (Engesser & Brüggemann, 2016), political leadership of EU institutions and Member States and its contestation (Wurzel, Connelly, & Liefferink, 2017), as well as the party-political dimension of climate governance (Carter, Ladrech, Little, & Tsagkroni, 2018). This state of research stands in notable contrast to the literature on climate politics in the US, which (somewhat unsurprisingly) includes a higher number of studies on political conflict and polarization on this issue (Atkinson, 2018;Karapin, 2016;Sussman & Daynes, 2013). Comparative research on the politics of climate change (cp. Harrison & Sundstrom, 2010), however, would obviously benefit very much from more indepth studies that include European cases.
Zooming in on the role of the European Parliament (EP) in this context is interesting for two reasons. First, the EP is generally recognized as an early and progressive actor in the evolution of EU climate change policy (Burns, 2012;Burns & Carter, 2010;Burns, Carter, Davies, & Worsfold, 2013). However, recent assessments have characterized it as a mostly symbolic leader with an ambitious stance in its non-binding declarations on the EU's external climate diplomacy but a much more restrained and pragmatic involvement in internal climate change legislation (Burns, 2017;Biedenkopf, 2015). Addressing the goals of climate action at the global level, the EP appears as much more ambitious and willing to diverge from positions of other EU institutions than in internal climate change legislation, such as emissions trading or promotion of renewable energy. Second, this apparent asymmetry between an ambitious declaratory role of the EP and a cautious legislative one establishes an interesting case for scrutinizing the interrelation between institutional factors and political conflict between party groups: Resolutions by the EP on global climate negotiations (and hence, a crucial part of the EU's external policy) are generally own-initiative procedures (INI) without formalized interactions with the Council and no legally binding effect on forthcoming negotiations. By contrast, EP decision-making on legislative policy-making such as EU emissions trading or renewable energy legislation (internal climate change policy) is conducted through the ordinary legislative procedure (OLP). This procedure establishes a more institutionalized environment for EP action, involving a formal legislative proposal by the Commission, a sequence of formally prescribed interactions with the Council and parliamentary decisionmaking about a legally binding act of EU legislation.
Comparing decision-making on both dimensions therefore appears as a promising case study to understand the interrelation between different forms of parliamentary involvement in climate policy and political conflict as expressed through polarization between EP party groups. Beyond the specific case of climate change policy, this question is also of interest for the broader literature on the emerging role of the EP as an international actor: investigating the interrelation between institutional context and party political conflict speaks to contributions to the literature on the role of the EP that have highlighted a moderating effect of its increased institutional involvement on the stringency of EP demands and party group conflict (or, put more simply, an effect of 'behaving responsibly' due to anticipated effects of legally binding parliamentary decisions; cp. Ripoll Servent, 2015;Trauner & Ripoll Servent, 2016). This discussion provides the point of departure for this article: We investigate how the arena and actor function of the EP in climate change policy relate to each other by asking how decision-making on acts of external and internal policy-making are related to political polarization between EP party groups. The article proceeds in three steps: the next section lays out the theoretical framework for our analysis, before the main part presents a comparison of EP decision-making on external and internal aspects within Phase IV of EU Climate and Energy policy, covering the most recent elective period of the EP from 2014 to 2019. The subsequent conclusion relates the findings back to the main topic of politicization.

Theoretical Framework: The EP as an Actor and Arena of Global Climate Governance
Politicization is broadly accepted in the EU research literature as a term for the emergence of salient and polarized public controversy between political actors involved in European policy-making (cp. De Wilde, Leupold, & Schmidtke, 2016;Hutter, Grande, & Kriesi, 2016). In this sense, most contributions to the literature on politicization focus on mapping and explaining political controversy at the level of public discourse, but usually leave out the analysis of political conflict within or between EU institutions that does not address or reach a broader public. In this sense, politicization should not be equated with political conflict arising from questions of European governance but is generally used as a term for the degree to which political actors communicate it to external actors and audiences.
The approach taken in this contribution relates to the debate on politicization but adopts a more specific and dynamic perspective. By focusing on the polarization between EP party groups on issues of climate change governance as the dependent variable, we single out an important aspect of politicization without, however, considering the public salience and resonance of political debates. Including these latter criteria would require a comprehensive review of public controversy and media coverage of climate governance in at least a selection of EU Member States, beyond the scope of the present article. The question addressed here, by contrast, focuses on the interrelation between the institutional context of decision-making and political conflict between party groups represented in the EP. The rationale of the article is thus to contribute insights how two points of reference of EP activity-namely, the public communication of positions on climate governance and policy influence on specific legislation-drive disagreements between parliamentary party groups as a first stage of potential subsequent, broader politicization in the public sphere of EU Member States.
As mentioned at the outset, this comparison is based on the distinction between the different rules and procedures applied to EP decision-making in external and internal climate change policy. Based on these different procedures, activity of the EP at both levels as such is not in question: for the case studies selected here, EP resolutions on the conventions of the global Conferences of the Parties (COPs) follow their yearly schedule, while recent legislative decision-making on internal legislation was prompted through the forthcoming entry of the EU into the next envisaged stage of its Energy and Climate Action package (dubbed 'Phase IV' for the period 2021-2030). In this sense, EP activity or decision-making as such does not establish a puzzle for explanation but is mostly prescribed by established trajectories, allowing our analysis to focus on the content and political conflict dimension of decision-making.
From this point of departure, we harness Discursive Institutionalism (DI) to establish a theoretical framework for our analysis of EP party group interactions in different institutional contexts of EU climate change governance (Schmidt, 2008(Schmidt, , 2010. This approach is chosen for two reasons: First, DI speaks to essential aspects of the politicization debate by focusing on the dynamic and often conflictual exchange of ideas between political agents through political discourse, and the way these exchanges are structured by institutional frameworks. Second, DI is useful as a heuristic framework to relate two different spheres of parliamentary activity addressed in this article: namely, the public communication of political posi-tions towards a general political public and other entities as captured by the concept of communicative discourse (CMD); and decision-making about specific acts of legislation through parliamentary agents as captured by the concept of coordinative discourse (CRD; cp. Schmidt, 2008Schmidt, , p. 305, 2010). While these two systems of discourse overlap in concrete empirical case studies, they help to distinguish two sets of linkages between agency, institutions and conflictual discourse (see Table 1): (1) Incentives and rationales of discursive agency: By addressing different audiences (namely, the general public versus policy-making agents), both systems of discourse establish different rationales of action for political agents. In the case of CMD, the mobilization of political support through positiontaking on politically salient and polarized topics; in the case of CRD, substantial policy influence through the modification of legislation, independently from its public perception or support; (2) Forms of interaction between discursive agents: While political conflict is present at both levels of discourse, polarizing and adversarial interactions are incentivized more strongly at the level of CMD, where the rationale of mobilization recommends agents to increase the visibility and distinctiveness of their positions in relation to other public speakers; by contrast, the rationale of policy influence at the level of CRD sets incentives for agents to seek ways of accommodating and negotiating competing positions and arguments; (3) Discursive frames of reference: Both discourse systems establish different incentives for the way in which political agents articulate and frame their political positions, namely through the communication of more general, political and relatively simple positions and arguments to maximize public resonance at the level of CMD; and in the form of more specific, technical and more elaborate policy proposals and suggestions for modification of specific regulatory approaches at the level of CRD; (4) Institutional setting and procedures: Finally, the interaction of political agents at the level of CRD is highly structured according to clearly formalized, sequential steps of decision-making as specified in the legal requirements of EU legislative procedure. By contrast, CMD is generally much less formalized and can take place through more spontaneous and less formalized appearances or speeches by political actors; in the same vein, it does not include formally prescribed sequential steps of negotiation but evolves through public statements that are not necessarily tied to a specified procedure of decision-making.
In summation, the concept of EU politicization as commonly used in the literature is virtually synomymous with a shift or expansion of political conflict from the sphere Link to politicization Expansion indicates dynamic of relative Expansion indicates dynamic of relative and political conflict de-politicization politicization (or internalization of political conflict to (or externalization of political conflict to public intra-institutional negotiation) justification and contestation) of coordinative to CMD. However, the advantage of using the present distinction of both levels, however, is to gain insights about how far political conflict remains confined within institutional settings and procedures of EU policymaking (by being internalized within the legislative process of EP decision-making), or if it is communicated to the outside, indicating an externalization of political conflict beyond specific policy-making venues.
Turning to the empirical cases, neither of the decision-making procedures of the EP discussed here can be identified entirely with one of the types of discourse distinguished through DI. However, we assume that EP decisions in the realm of external climate governance resonate more strongly with the logic of CMD (as primarily declaratory decision-making acts), while legislative decision-making on internal climate change mitigation leans more clearly towards the sphere of CRD (as resulting from policy-making negotiation with the Council). Comparing both levels of decision-making, we can therefore provide insights to what degree public communication and policy-influence work as incentives for party group polarization. Furthermore, by going into a more indepth analysis of decision-making in the plenary at the committee level, the subsequent analysis will uncover in how far political disagreement is internalized (i.e., solved within non-public procedures of policy coordination) or externalized (i.e., communicated to the public through public votes and declarations).
To conclude the theoretical discussion, two assumptions about the general context of our case studies must be added. First, we approach the parliamentary dimension of EU climate change policy (or more specifically, its supranational aspect as represented by the EP) as a less likely case of politicization in comparison to other settings and arenas of climate change politics. Several institutional features of the EU suggest a generally low potential for strongly polarized public contention: namely, the absence of a prominent, electorally accountable discursive agent in charge of justifying EU climate action, combined with the absence of a resonant public sphere on a transnational European scale (or CMD in typical DI terms); and the presence of a highly specialized, sequential process of decision-making with a prominent role of legislative mediators (particularly main and shadow rapporteurs) and informal negotiation in the framework of trilogues between Commission, EP and the Council at the level of legislative decision-making (or CRD in DI terms). Second, political conflict on EU climate action is limited from the outset through the fact that its governance framework is well-established with regard to its main approaches and instruments (such as emission trading, effort-sharing, and regulation to set incentives for sustainable energy production and consumption). We expect that this considerably limits the scope of political controversy on EU Climate Action, particularly in comparison to other legislatures struggling with climate-related legislation such as US Congress (Karapin, 2016;Sussman & Daynes, 2013). While not denying the substantial impact of technical details of policy, we expect political controversy to exclude fundamental debates on the reality or severity of climate change, and to focus on the adjustment of specific, existing regulatory instruments. These factors suggest that more political resources are invested into policy negotiation than public communication, and that a dynamic of consensus-oriented negotiation outweighs public communication about general aims and principles of climate change mitigation. We therefore expect to find more intensive party group interaction and conflict within the intra-parliamentary level of policy negotiation over specific pieces of climate change legislation than over external declarations.
In summation, the subsequent empirical section investigates three main hypotheses on the involvement of the EP in European climate change governance: H1: The overall (external) politicization of EP decisionmaking as expressed through party group disagreement is relatively low, particularly by not endangering the overall consent of the EP to decisions tabled in the plenary; H2: Political conflict between EP party groups is expressed more clearly in decisions about external EU climate policy than on internal climate legislation, particulary through the closer proximity of the former to the logic of CMD; H3: Political conflict about climate change legislation is 'internalized' in the EP, that is, expressed more strongly at committee level than in the plenary and accommodated through consensus-oriented negotiation through CRD.

Case Studies: EP Decision-Making on Phase IV of EU Climate Action
The subsequent empirical analysis focuses on key decisions of the EP on external and internal climate change policy during its most recent legislative period (2014)(2015)(2016)(2017)(2018)(2019). With regard to external climate diplomacy, during this period the EP has adopted resolutions on the annual COP, most importantly on COP 21 in Paris where the successor agreement to the Kyoto Protocol was negotiated. As mentioned above, these resolutions were adopted as INI in which the EP responds to a communication by the Commission but is not formally bound to a specific legislative proposal (as in the case of the Paris negotiations). In other cases, it adopted a resolution on a topical subject based on a text proposed by its environmental committee (EP Committee on the Environment, Public Health and Food Safety [ENVI]) without a reference document tabled by the Commission (as is the case in resolutions on COP 20 in Lima and COPs 22-24 in Marrakesh, Bonn and Katowice; cp. EP, 2014EP, , 2015EP, , 2016EP, , 2017EP, , 2018. All resolutions are adopted without any formal involvement of the Council and legally non-binding on negotiating partners at the climate summits (both Member States and the EU delegation). Therefore, these resolutions can be reviewed as acts of decision-making primarily aimed at CMD: an announcement and justification of EP positions without direct effect on or involvement in policy-making.
Within the internal dimension, four particularly relevant legislative procedures conducted in the context of the revision of EU Climate Action for its Phase IV are selected for the analysis of this article: namely, the revision of the Emission Trading System (ETS), the regulation on effort-sharing of Member States for the reduction of carbon emissions, and the directives on the promotion of renewable energies (supply-side), and on targets for energy efficiency (demand-side). All of these legislative acts were negotiated using the OLP and resulted in the adoption of texts agreed between Council and EP after the first reading (including inter-institutional negotiations or trilogue between Commission, Council and EP in the case of the ETS, effort sharing and energy efficiency). With regard to the institutional framework of EP involvement, they obviously differ from resolutions on external climate diplomacy by starting off from a formal legislative proposal by the Commission, involving several EP committees to submit a report for a first vote in the plenary and subsequent negotiation with the Council. Therefore, these procedures are better examples to understand CRD conducted by the EP, establishing more stringent institutional constraints on the mediation between party group preferences to define an EP position, and the projection of EP in its negotiations with the Council. Considering this difference of institutional context, party group interaction and EP decision-making is discussed in the subsequent section.

EP Votes and Decision-Making: External and Internal Dimension
In a first step, we map the decision-making behavior of the eight officially recognized EP party groups as well as non-affiliated members in plenary votes of the EP, assessing party group polarization as expressed by voting behavior of MEPs in a comparison of external and internal EU climate change action. Roll-call voting data exists on all legislative acts on internal climate policy and two resolutions in the external dimension, relating to the COP 21 in Paris and the subsequent COP 22 in Marrakesh (with the rest of resolutions being adopted by a show of hands). The radar plot below depicts an overview of party group votes aggregated from this data, indicating the relative amount of Yes votes in relation to all votes cast in the plenary and specifying this ratio for decisions on external and internal climate policy as well as all votes considered in this analysis (see Figure 1). On the whole, 3,767 votes cast by MEPs in the plenary are included for this analysis (2,569 on the internal, 1,198 on the external dimension).
Two findings result from this overview. First, the plotting of voting behavior confirms that EP decision-making on climate policy is not intensely polarized but broadly supported by four party groups ( . The dent in support by S&D, however, is mostly due to an abstention covering 146 MEPs in the vote on the effort-sharing regulation. Overall, the ratio of Yes/No/Abstain votes on legislative decisions is highly positive for the entire EP (70.3/17.3/12.3). Concerning polarization, the overall pattern of party group voting can be characterized as corresponding to an Inverted U pattern pitting mainstream and environmentally progressive groups against more Eurosceptic ones. Second, another finding is that voting profiles do not significantly differ between votes on legislation and non-binding res- olutions: Resolutions on external climate policy, in this sense, are not more contentious decision-making acts for the EP as decisions on internal legislation.
More specific insights on the dynamic of intraparliamentary decision-making-specifically, the relation between intra-parliamentary negotiation and EP decision-making towards other EU institutions-are gained from comparing voting results on committee reports and subsequent votes in the plenary. The bar chart below presents data on votes at both levels for internal climate change legislation and the one resolution on external negotiations where data is available, based on an overall count of 707 votes at committee level and 2,562 in the plenary for legislation and 251 in committee and 582 in the plenary for the COP 21 resolution (see Figure 2).
The data confirms that political conflict is internalized rather than externalized in EP decision-making about climate change legislation. In all cases of legislative decision-making on internal climate policy except one, the approval rate to decisions is higher in the plenary than at the committee level, suggesting that position differences are fought out more intensely by policy experts within the committees and approved more broadly in the plenary. While this data provides some support to our hypothesis that political conflict is internalized rather than projected by the EP especially in internal climate policy, it is not easily generalizable due to the low number of cases. This recommends a more detailed in-depth qualitative review.

Qualitative Review: EP as an Agent and Arena of Controversy on Climate Change Policy
Based on the review of legislative documentation (committee reports, resolutions adopted by plenary and summaries provided by the European Parliamentary Research Service [EPRS]), a qualitative review of decisionmaking by the EP on Phase IV of EU Climate Action leads two main observations. First, that the EP has generally avoided a strategy of strong partisan politicization; and second, that two distinct forms of involvement emerge that put different emphases on the role of the EP as an actor and arena of negotiation on EU climate action: on the one hand, an emphasis on the communication of strong and rather controversial policy preferences without much internal coordination in external climate policy; and on the other, an almost exclusive concentration on internal policy coordination and negotiation with much more restrained and consensual communication in internal climate policy. Three points are discussed subsequently to clarify this distinction. First, clear differences emerge in a comparison of the scope and intensity of intra-parliamentary policy coordination at both levels. The number of committees involved in the negotiation of reports tabled for the plenary for first reading is lower in decision-making on resolutions about external climate policy: only the resolution on COP 21 in Paris and COP 20 in Lima involved several EP committees (namely, three and five, respectively), whereas the draft text for the subsequent resolutions on COPs 22 through 24 were tabled only by the responsible lead committee ENVI. By contrast, all legislative decision-making procedures on the EU's internal climate policy reviewed here involved two (Energy Efficiency), five (Emission Trading), six (Effort Sharing) and nine (Renewable Energies) committees drafting an opinion on the report tabled in the plenary. The number of amendments tabled at committee level is also higher in internal climate policy, where the amount of amendments tabled ranges between at least 280 in Effort-Sharing and 1,300 in Renewable Energies, as compared to a range between 120 and 202 for those four resolutions on external climate policy where an exact count of amendments is available (excluding the resolution on COP 21 in Paris where at least 230 amendments were tabled and one further report with amendments is not available). CMD as conceptualized by DI, in short, is 'thicker' in internal legislation than in decision-making on external climate policy.
Second, an important qualifying observation about the way the EP negotiates political conflict in relation to climate governance is that contention in this field can hardly be reduced to a single dimension, in which disagreements between party groups and EU institutions could be easily mapped and compared as a measure for politicization. As documented by the high number of amendments tabled and negotiated at the committee stage, decision-making evolves through a highly specialized and complex CRD between policy experts of the involved EP committees and the Council. To systematize this complexity, the distinction of three separate issue dimensions-namely within a socio-economic, territorial and institutional dimension (cp. Hix, Noury, & Roland, 2007)-helps to identify issue dimensions that are balanced with each other to accommodate political conflict between party groups. These issue dimensions can be conceptualized and specified for the given case studies as follows (see also (1) Regulation dimension: This dimension reflects the overall stringency and ambition of action taken to restrict carbon emissions as a cause of climate change, either at the supply-side (Renewable Energies), or at the demand-side (Energy Efficiency) of energy policy, and by establishing a framework for limiting and pricing those emissions (Emissions Trading/Effort Sharing). As a regulatory framework for economic investment, production and consumption, this set of issues is identified with the contrast between market freedom and regulatory intervention ('left/right'); (2) Intergovernmental dimension: A second dimension consists in the accommodation of asymmetries resulting from distributive effects between different economic sectors and by extension, Member States with different constellations of industries and geographical location. This aspect therefore resonates with the representation of territorial interests represented mainly at the intergovernmental level in the institutional framework of the EU; (3) Integration dimension: Finally, a third issue dimension arises from negotiating the degree of authority of the supranational layer of rules and institutions in the governance of climate change, partic-ularly with regard to processes of implementation, oversight and potential sanctioning, as well as the allocation of financial resources to the discretion of supranational institutions; this set of questions therefore relates to a third dimension of political conflict between national sovereignty and supranational integration.
Comparing EP decision-making in internal and external climate policy within these issue dimensions reveals two different patterns of political conflict: EP decision on external climate action are relatively antagonistic and emphasize the stringency of climate action over the other dimensions. Internal legislation, by contrast, combines moderate advances in the market regulation dimension with concessions and flexibility in the territorial and authority or institutional dimension. Resolutions on the external climate policy of the EU establish highly critical stances of the EP towards policies proposed by the Commission as well as the international community and denoting considerable and quantifiable position differences. Within these resolutions, the EP criticizes both the international response to climate change but also the stance of the Commission, whose communication is deemed 'short-sighted and unambitious on a number of levels, specifically as regards the lack of national targets for renewable energy and of any meaningful new action to incentivize energy efficiency' (EP, 2014, p. 11). With regard to substantive policy goals, the most remarkable position expressed in the EP resolutions is the call for more stringent policy goals, calling for the inclusion of a 40% goal for the reduction of carbon emissions, of 40% for the improvement of energy efficiency, and of a 30% share of renewable energy sources into the strategy for 2030 (compared to values of 40, 27, and 27 proposed by the Commission, respectively). Beyond climate action by the EU, the EP expresses concern that action pledged in NDCs under the Paris agreement 'does not bring the world even close to the 2C target', pointing out that 'that additional ambition is necessary' and calling for concerted action by the EU and other major emitters to adjust their action plans (EP, 2016, p. 4). Finally, another important aspect of EP resolutions is the call for more determined financial action, requesting that the EU and its Member States build on existing commitments to bring their contributions to 'their fair share in the overall targeted amount of USD 100 billion a year by 2020' to climate action (EP, 2015, p. 12). Turning to legislation on internal climate policy, much of the ambitious stance of the EP is softened considerably, particularly with regard to the main target values for climate action. Furthermore, an important finding is that the moderately progressive demands raised by the EP in the socio-economic dimension are mostly accommodated and balanced in the other dimensions, defusing political conflict and helping to create broader based coalitions supporting the passage of legislation.

Revision of the EU ETS
The revision of Emission Trading demonstrates the construction of a compromise of positions within the three dimensions outlined above, aiming at the general goal of consolidating and strengthening emission trading. The most high-profile issue of negotiations in this context is the balance between market competition and public regulation concerning the limitation of emissions, particularly through the definition of the linear reduction factor for the emissions cap (hence, concerning the regulation dimension). Concerning this question, the EP accepted to maintain the 2.2% target as in the Commission proposal, securing only an option of an increase to the 2.4% target it initially sought, along with the establishment of a Market Stabilization Reserve to avoid an over-allocation of allowances. Highly important for the achievement of the political compromise achieved with the Council, however, is the combination between an increase in the regulatory stringency of the supranational framework with flexibility mechanisms and targeted recompensation for vulnerable industries: in this sense, the EP achieved more stringent rules for the inclusion of aviation and shipping, reviews of the ETS by the Commission and allocation of revenues to climate action (integration dimension), while providing exemptions for medium-sized enterprises, and supporting companies at risk of carbon leakage with revenues from emissions trading. As a compensation, the agreement increases resources for a 'Just Transition Fund' to assist workers affected by the transition to climate-friendly technologies and providing a Modernisation Fund for countries with a GDP below 60% of the EU average, an aspect considered in the intergovernmental dimension (cp. EPRS, 2018a).

Effort Sharing between EU Member States
A similar combination of progressive and accommodating positions in two different issue dimensions can be observed in the revision of the Regulation on Effort Sharing, the legal framework to establish reductions of carbon emissions for each Member State for sectors not covered by the ETS. Here, the EP sought and achieved a slightly more stringent collective effort of Member States (regulation dimension). This point aims at establishing a base reference and trajectory calculation that results in a lower overall emission budget until 2030 than the one proposed by the Commission, and establishing a longterm linear trajectory for the longer-term period until 2050. These innovations, however, are compensated by more generous flexibility mechanisms advocated by the EP (affecting the intergovernmental dimension). Here the EP pushed to increase the level of 'borrowing' for the first implementation phase from 2021 to 2025 from 5% as proposed by the Commission to 10%, and introducing the right to transfer up to 10% to other Member States in the subsequent phase from 2026 to 2030 (up from 5% in the Commission proposal). Furthermore, the EP proposed an 'early action reserve', under which Member States with lower GDP are allowed to use emission cuts made prior to 2020 in the implementation period from 2026 to 2030 (EPRS, 2018b). The integration dimension, finally, is present mostly in provisions on monitoring cycles and the stringency of reporting duties by Member States to the Commission.

Renewable Energies
Within the comparison of cases presented here, the directive on the promotion of renewable energies appears as the initially most polarized example. Concerning the overall stringency of action (regulation dimension), the EP initially called for a target value of 35% until 2030 in first reading, as compared to 27% in the Commission proposal and only 30% in a previous EP resolution. At the same time, the EP also pushed for a lower threshold for the share of newly created capacity open to cross-border generation, hence limiting transnational exchanges of energy and increasing pressure on Member States to achieve committed targets for renewable en-ergy (intergovernmental dimension). Along with stricter provisions on biofuels (particularly the exclusion of bioliquids produced from palm oil) and setting targets for the transport sector to be realized by each individual Member State, the EP established stringent positions that were however softened in subsequent trilogue negotiations (cp. EPRS, 2019a). In the final agreement, the binding EU target for renewable energies is set at 32% and therefore at a higher level than in the initial proposal, but to be achieved through indicative national contributions without the corrective measures initially pursued by the EP (and hence, a softer approach in the integration dimension).

Energy Efficiency Regulation
As in the previous example, the directive on energy efficiency started out with a clear discrepancy of positions concerning the stringency of action (regulation dimension): here, positions ranged from the target value advocated by the EP as advocated in several resolutions, namely an increase in energy efficiency of 40 % to be reached until 2030 in comparison to a target of 27% target initially adopted by the Council. A second main issue, however, concerned the authority of supranational regulation towards the Member States (integration dimension), with the EP calling for setting individual national targets, and the European Council excluding such targets in favor of indicative national contributions. In addition, the question was raised how to set and operationalize annual savings obligations for Member States. In legislative negotiations with the Council, the EP therefore addressed issues in several dimensions, concerning the overall stringency of regulation, its form of implementation and scope concerning its extension to sectors such as private buildings, transport, and industrial activity (cp. EPRS, 2019b). The agreement reached with the Council represents a compromise with the EP's initial position, setting the EU headline target at the relatively ambitious level of 32.5% but using (non-binding) indicative national contributions and lower annual savings obligations than sought by the EP (0.8% per year for the period through 2030). With regard to its scope, the new directive combines an extension of sectors with a right of Member States to exclude especially sensitive sectors such as transport (intergovernmental dimension). As in other examples, the EP pushed for more stringency in the overall ambition of EU climate policy, but eventually agreed to only moderate advances in combination with flexibility mechanisms aimed at the accommodation of Member State interests, and only moderate advances towards more EU authority in implementation.

Conclusion: Is the EP Resistant to the Politicization of Climate Change Policy?
In summation, this article demonstrates how full legislative involvement of the EP in decision-making about a potentially politicized aspect of global governance works less as a triggering, but more as a constraining factor for partisan polarization of parliamentary decision-making: overall, EP decision-making on legislation passed to revise and advance the policy framework of EU climate action is negotiated effectively and often informally with the Council, is based on broad majorities both at committee and plenary level, and results in policy-making decisions that carefully balance moderate advances in one issue dimension with compensation and flexibility mechanisms in another. While this is not to deny the significant and often very conflictual content of agreements, party group polarization as a first indicator of politicization emerges only to a limited extent (confirming H1). DI offers plausible theoretical arguments to explain this relative restraint: incentives for MEPs to seek policy modifications at the level of CRD in most cases outweigh the incentive to establish more polarizing claims aimed at the wider public at the level of CMD. In this sense, the governance of EU climate action is largely policy-oriented, technical and relatively immune to more fundamental disputes about the severity of climate change or principles of its mitigation.
Nevertheless, a comparison of EP decision-making on the internal and external dimension of EU climate policy reveals variation in the extent of party polarization that equally fits a DI explanation: In the external dimension, EP activity is mostly confined to declaratory, nonbinding resolutions that are relatively stringent and polarizing in their content and based on less internal negotiation and coordination in the EP than legislative acts. EP resolutions at this level are relatively ambitious but do not create strong incentives for intensive partisan contestation or polarization (in this sense, contradicting H2). Within the internal dimension, the EP operates as a highly specialized legislator that uses a sophisticated set of arrangements to conduct a mostly technical legislative bargaining, both between EP party groups and with the Council. Policy coordination at this level is highly specialized and involves substantial political conflict in different issue dimensions, but succeeds in accommodating positions within these dimensions through negotiations that are largely detached from public communication or scrutiny. Taken together, the findings confirm H3, positing that political conflict on climate governance is internalized within the EP rather than projected to external audiences or institutions.
In the bigger picture, and compared to other legislatures, the relative immunity of the EP against broader politicization appears to work in favor of a relatively stringent and progressive development of climate change policy in the EU compared to other political entities, particularly the US (Harrison & Sundstrom, 2010;Skjaerseth, Bang, & Schreurs, 2013). Future research should engage in more comparative studies to investigate the interrelations between the politics and policy of climate change: more specifically, how different issue dimensions and policy-making frames are brought together to build the coalitions needed to pass climate legislation through the EP, and how resulting findings can be set in comparison with other legislatures on the national level, both within and beyond Europe.