Representing Globalization Conflict in International Organizations: Cleavage Formation Beyond the State

Transcription

1 Representing Globalization Conflict in International Organizations: Cleavage Formation Beyond the State Pieter de Wilde, Wiebke Junk, Tabea Palmtag WZB Berlin Social Science Center Key words: cosmopolitanism, claims analysis, European Parliament, globalization, representation, UN General Assembly Abstract Recent developments in political theory and political sociology raise new questions about the representation of citizens within international organizations. Given the emergence of a globalization cleavage between winners and losers of open borders, the question is whether both sides are represented in international organizations, and by whom. We analyze plenary debates in the UN General Assembly and the European Parliament about the issues of climate change, human rights, migration, regional integration and trade between 2004 and 2011 to study who delegates claim to represent, and with what effect. Our findings are, first, that claims in favor of open borders are often made by those claiming to represent individuals or other non-territorial constituencies of universal scope. Claims to close borders are, in contrast, made by those claiming to represent a particular territorial community, like their nation state. This leads us to conclude that debates about globalization are embedded within broader ideological perspectives of cosmopolitanism and communitarianism. We subsequently document how debates in the European Parliament feature much stronger representation of communitarians while the UNGA discourse is almost completely cosmopolitan. More direct and more proportional mechanisms of delegation in the case of the EP likely accounts for this observed difference. To be presented at the ECPR General Conference, Glasgow, 3-6 September

2 Introduction In formal global governance as orchestrated through international organizations (IOs), citizens are represented by their national governments. Their delegates participate in decision-making, control the actions of IO secretariats and thereby contribute to making global governance ultimately accountable to citizens (Buchanan and Keohane 2006; Moravcsik 2004). Yet, which citizens are represented, by whom and to what effect has received surprisingly little empirical consideration in the scholarship of global governance. This question gains traction in light of two recent developments in political theory and comparative political sociology, where both are responding to societal challenges brought about by globalization. Firstly, societal interactions increasingly cross state borders, old patterns of representation characteristic to the democratic nation state are compromised. U2 singer Bono claiming to represent Africans in his making poverty history campaign exemplifies an act of representative claims-making that leads political theorists to reconceive representation as a discursive act (Saward 2009). In today s globalized world, a representative is someone who effectively and legitimately claims to represent his or her constituency (Saward 2010; Severs 2010). The demographic characteristics or procedures of election and appointment do not tell us yet whose interests they substantively represent. To know this, we need to empirically study the discursive behavior of representatives. Secondly, globalization has generated considerable societal conflict between winners and losers of open borders. At least the European political spectrum is increasingly characterized by a cleavage between those favoring more international integration and border permeability cosmopolitans and those opposing this, variously labeled sovereigntists (Azmanova 2011), nationalists (Kriesi et al. 2012) or communitarians (Zürn 2014). Clearly, many citizens do not approve of open borders and would rather like to see social, political and economic life restricted to a confined community. Since IOs are prominent arenas and actors in global governance, it becomes increasingly pertinent to ask whether citizens who oppose globalization are represented in these key forums of global governance. The mere presence of government delegates is not enough to conclude adequate representation. It matters what they do, and particularly, what they say. Bringing these recent developments in political theory and political sociology to the study of global governance, the present paper investigates the research question: How and to what extent are cosmopolitan and communitarian citizens represented in the plenaries of international assemblies? In order to tackle this we comparatively assess representative claims making in the United Nations General Assembly (UNGA) and European Parliament (EP) in a most different systems design. We proceed in three analytical steps. First we ask to what extent we find evidence of conflict between representatives of pro-integrationist and pro-demarcationist positions on international integration within these international institutions. Second we investigate if and how representatives within these 2

3 international institutions position themselves to different constituencies. Third we seek to identify similarities and differences in patterns of preference articulation and representation within and between the international institutions and discuss how these may be accounted for, with a special focus on mechanisms of delegation. Our results bear insights, firstly, on the relevance of the cosmopolitan and communitarian concepts in debates about globalization within IO assemblies, as we show that there is a correlation between the constituencies represented and positions taken regarding conflicts on migration, human rights, climate change and regional integration. Secondly, the results have implications for discussions on institutional design of international institutions, as our results suggest that more democratic international plenary bodies (with shorter chains of delegation) seem to favor more communitarian claims making. This paper is structured as follows: The next section locates our paper in the intersection of theories on global governance, representation and globalization cleavage and formulates our hypotheses. Subsequently, the method of data collection, namely a claims analysis in the two plenaries on five globalization issues, is explained. The third section presents our findings regarding cleavage formation. Finally, the conclusion revisits the cosmopolitan and communitarian divide and relates to implications for institutional design. Theory While the debate continues about the degree of influence international organizations have in global governance, the observation that their numbers and powers have increased dramatically in the past century has become widely accepted (Barnett and Finnemore 2004; Hawkins et al. 2006; Hooghe and Marks 2013; Lake 2010). At the global level, the United Nations have expanded as a network of institutions to now include bodies covering many different policy fields, with significant powers in terms of policy-formulation, monitoring of compliance and dispute settlement. At the regional level, the European Union has undergone dramatic changes to establish a common market, a monetary union and a European legal order (Börzel 2005; Burley and Mattli 1993; Dinan 1999). Across the globe, other regional organizations have been established, partially copying features of the European Union, although no single one of them is as encompassing in the scope of policy fields addressed or contains as powerful and independent supranational institutions as the EU (Acharya 2004; Lindberg and Scheingold 1971; Mukhametdinov 2007). In their capacity as arenas to reduce transaction costs of cooperation between states (Keohane 1988) and as agents in their own right, international 3

4 organizations are both manifestations and driving forces of globalization. They are, on the one hand, the result of increasing human interaction across borders and reflect the desire of nation state governments to control and organize this interaction in long-term, stable and predictable ways (Abbott and Snidal 1998). On the other hand, to achieve these tasks, many international organizations have built up significant bureaucratic and legal capacities in secretariats and courts which adopt the organizational self-interest to survive and expand competencies (Barnett and Finnemore 1999; Olsen 2007) leading them to act as entrepreneurs of increasing integration (Burley and Mattli 1993; Haas 2004; Haftel and Thompson 2006). The creation and empowerment of international organizations has always been accompanied by attempts by states to maintain control over their operations through clear mechanisms of delegation and accountability. Moreover there have been frequent attempts to open them up (Tallberg et al. 2013) to other stakeholders such as transnational civil society organizations and the wider public (Hale 2008). Key instruments in this democratization of global governance are international assemblies such as the United Nations General Assembly and the European Parliament. These assemblies combine functions of public deliberation in terms of the open and publicly voiced 1 exchanges of positions on issues by a variety of actors with (limited) decision-making powers. On budgetary matters, for example, assemblies tend to have a veto power and the European Parliament s powers extent considerably beyond that (Rittberger 2005). In this sense of combining public deliberation with decision-making, these institutions can be understood as strong publics (Fraser 1992: 134) that function as platforms of preference formation and aggregation and as intermediaries between governance institutions and the wider audience. With citizens increasingly interested in and divided over globalization issues, assemblies as strong publics present a pivotal arena for the representation of their interests. It is an open empirical question how these constituencies play out in the deliberations within international plenaries. Do representatives conceive of and speak for a constituency and, if so, which one? Recent studies on the discursive acts of representative claims-making reveal that it is far from obvious who represents whom in public deliberations (De Wilde 2012; Erzeel 2011; Saward 2009; 2010). In multi-level structures of international institutions with a global (regional) mandate, which unite complex and diverse cultural, social and economic groups and identities, it is even less clear how representative claims making functions. This paper seeks to address such claims making and identify underlying patterns. The remainder of this section presents the hypotheses that inform our analysis. 1 Transcripts of debates and resolutions are freely available online. 4

5 The first question is whether citizen concerns about globalization are reflected in debates within IO assemblies at all. We assume that they are, given dual conditions of delegation and accountability. First, delegates in assemblies are either elected (i.e. European Parliament) or appointed by member state governments. It is therefore not surprising to find that they consider themselves to perform a representative role and try to perform it (Jacobson 1967), whether they consider it their appropriate duty or fear the consequences if they would do otherwise (March and Olsen 2004). Secondly, formal mechanisms of accountability are complemented by the informal check provided by transparency. As all proceedings are publically available afterwards, assembly members will be stimulated to take into account and cater to the desires of the public (Hale 2008). Given the logics of delegation and accountability and the main discursive role assemblies play in this, we build on Michael Zürn (2014: 65) and hypothesize that: H1: Societal conflicts on political issues in which international organizations have competencies, including those between supporters and opponents of globalization, feature in international assemblies when these have minimally functioning mechanisms of delegation and accountability and of generating public transparency. The literature has documented clear patterns of conflict in the UNGA (Voeten 2000) and the EP (Hix et al. 2006) in terms of voting patterns and roll calls, but much less is known about what is said in the plenary debates (but see Binder and Heupel 2014; Lord and Tamvaki 2013). Given the rising controversy around globalization issues reported by political sociologists, we take special interest in the speech acts that revolve around issues of globalization, defined here as exchanges of goods, services, people and norms across borders. These border-crossing transactions are a major potential subject of social conflict. The terms of cooperation have to be negotiated here and not everyone is convinced that the benefits provided by open borders in terms of consumption, travel or otherwise, outweigh the costs in terms of job insecurity, international crime, loss of cultural distinctiveness and loss of sovereignty. Western Europe, whose countries rank among the most open and internationally interwoven (Dreher et al. 2008), increasingly features a growing cleavage between winners and losers of globalization, or those favoring international integration and those favoring demarcation of the nation state. This cleavage extends far beyond a narrow understanding of globalization to mean free trade and international industrial production, to include transactions, movements and communication more broadly. In particular, the issues of immigration and European integration form the backbone of this cleavage with citizens opposing both clearly distinguishable from those supporting more 5

6 immigration and transfer of sovereignty to the EU (Kriesi et al. 2012; Kriesi et al. 2008). In line with cleavage theory, globalization can be understood as a major societal revolution like industrialization and secularization before it generating structural divisions within societies (Flora et al. 1999; Lipset and Rokkan 1967). Whether these divisions are pre-existing, or created and capitalized on by political entrepreneurs, they foster and solidify subjective identity perceptions with in-group cohesiveness and mobilize power for collective preference aggregation (Bartolini and Mair 1990). Based on this, we expect that preferences about globalization are embedded within a structural societal division, are captured or mobilized by opposing representative organizations and are underpinned by a normative component, including subjective group identity (Deegan-Krause 2013). This latter normative component is directly linked to the philosophical debate between cosmopolitanism and communitarianism: The scope of the conceived group-identity, can be more or less inclusionary, with cosmopolitanism exceeding territorially or socially delimited (i.e. communitarian) groups and emphasizing the equality of individuals under the banner of common humanity (De Wilde and Zürn 2013). It is plausible that these contrasting group-perceptions relate to different demands concerning the openness of borders. From a cosmopolitan moral identity follow preferences for more open borders to allow equal opportunities. If, in contrast, a national or other constitutive community is cognitively internalized as the relevant societal framework for achieving justice, it follows that preferences for closure are more likely to be formed, as one s lifeworld including understandings of justice is limited to that particular group. This assumption is supported by the empirical findings that individuals with exclusive national identity perceptions i.e. I consider myself a citizen of my country, not of the world go together with demarcationist policy preferences (Fligstein 2008; McLaren 2007). We assume that representatives operating in international organizations either appeal to specific groups to support their demands when its suits their needs in a top-down understanding of group mobilization or inform their preferences by the role perception of whom they consider themselves to represent in a more bottom-up understanding of representation. Irrespective of these different causal connections between preferences and representative roles, there would be an empirical correlation between the demands representatives advance and the constituencies they claim to represent. Representation understood as a discursive act (Saward 2010) should therefore feature an empirical linkage between the constituency a representative claims to represent and the policy preference he or she advances. 6

7 H2a: Demands on international issues voiced publically in international organizations for more integration are underpinned by representation claims of individuals or the global community, while demands for demarcation are underpinned by representation claims of constitutive communities. H2b: Manifestation of both cosmopolitanism and communitarians 2 within international organizations depends on group formation within their deliberative forums to coherently mobilize both sides of this conflict. Sociological research shows how high education and affluence lead to more cosmopolitan attitudes (Inglehart and Welzel 2005; Norris and Inglehart 2009). Given their high education, relative wealth and international contacts, it is therefore not surprising to find cosmopolitan dispositions to be disproportionally frequently present among elites (Calhoun 2002; Teney and Helbling 2014). Given that representatives operating within international organizations tend to be highly educated, affluent and socialized in an international setting, we assume them to advance cosmopolitan arguments. This general effect of affluence may also be supported by a socializing effect of the international organization, with its mandate to foster global (or regional) cooperation. Yet, given that citizens tend to hold more communitarian preferences, responsiveness and accountability to an electorate may contravene this cosmopolitan tendency: Representatives facing the prospect of electoral punishment can be expected to deviate less from the preferences of their constituency than those who do not, because the incentives to abide by the preferences of the principals is more forcefully structured. As most representatives in international assemblies are seconded or government ambassadors, however, with such direct accountability lacking, they are likely to disproportionally defend cosmopolitan preferences (Zürn 2014). Even in the only directly elected international assembly the European Parliament elections tend to be second-order (Van der 2 Cosmopolitan arguments are here operationalized as a combination of a call for open borders or maintained international integration together with a claim to represent an individual, global or otherwise non-territorially defined constituency. Communitarian arguments, as direct opposite, then are calls for demarcation for the sake of a national or other territorial constituency. 7

8 Brug and Van der Eijk 2007), meaning that voters orient themselves more on national politics than on what their Member of the European Parliament (MEP) has done in Brussels and Strasbourg. Even in this case, we would, thus, expect the cosmopolitan voice to dominate given weak accountability to citizens. H3a: Cosmopolitan arguments are disproportionally represented in international organizations. H3b: The stronger direct accountability of representatives in international organizations to citizens, the less likely it is that cosmopolitan positions dominate. Data & Method Most Different System Design All theorized hypotheses above with the exception of hypothesis 3b formulate similar expectations concerning political conflict beyond the nation state. In other words, there is no theoretical reason to expect these hypotheses to work in only a subset of international organizations, as long as these have minimally functioning mechanisms for delegation and accountability in place as well as public transparency. This applies to almost all assemblies of major international organizations. Furthermore, given that globalization is a universal phenomenon affecting almost all nation states around the globe, there is no reason to expect that contestation about it is necessarily restricted to Western Europe, where most empirical evidence to date comes from. Given such general expectations, we opt for a most different system comparative research design (Przeworski and Teune 1970) and study public debates on globalization issues within the United Nations General Assembly (UNGA) and the European Parliament (EP). The EP and UNGA may be conceptualized as strong publics (Fraser 1992: 134) beyond the nation state, in that they constitute institutionalized venues for opinion formation and, to some degree, decision making at their respective level of governance. Both fora are the central public deliberative organs of their respective institutional frameworks. Arguably, they are the two most influential international assemblies in today s organization of global governance. Both the UN and the EU, to which these assemblies are attached, have adopted clear mechanisms of 8

9 delegation and accountability of internal decision-making. The proceedings of both institutions are publically available online, including literal transcripts of debates, adopted resolutions and voting behavior of representatives. Finally, the EU and the UN are both general-purpose international organizations dealing with a wide variety of policy issues related to the phenomenon of globalization. Not only would we, thus, expect our hypotheses to hold in these forums, but the study of these two assemblies also carries intrinsic value, given the wide-ranging impact of these two influential strong publics. Yet, besides both being strong publics beyond the state attached to influential organizations, the UNGA and the EP differ strongly. They display ample variance in scope of policy issues which are discussed and voted on, powers vis-à-vis the UN Security Council, the UN Secretary General, the European Council and the European Commission, and delegation mechanisms. Firstly, the regional scope of the EP involves a group of relatively homogenous member states, who are all developed Western advanced industrialized democracies. None of them would suppress globalization conflict from finding democratic expression through authoritarian means, and such standards of democracy likely facilitate or at least do not inhibit the manifestation of globalization conflict at supranational level. The UN s membership is a heterogeneous group of democracies and authoritarian states, rich and poor, deeply internationalized open societies and autarkic states. Even if there is societal conflict about globalization in all UN member states, it is likely that some countries suppress it from manifesting itself fully in public debates. Secondly, the UNGA and EP differ in formal powers. While the EP holds major decision-making power in the European legislative process, even more so after the Treaty of Lisbon entered into force in 2009 (Hofmann 2009), the UNGA s central task is to create an international deliberative space. Its adopted resolutions carry little legal weight. Nevertheless, it also has decision-making power with regard to the UN budget and other formalities, such as the establishment of new UN bodies. While its resolutions have a non-binding character and thus remain recommendations for the member states, the UNGA comes closest to the ideal of a representative democratic institution at the global level (Peterson 2007: 98). Its authority can best be understood in relation to its political impact as a communication forum (Heideking 2000: 182). Thirdly, both institutions differ strongly in their delegation and accountability mechanisms. Essentially, while the EP is composed of representatives directly elected by the European citizens based on the principle of digressive proportionality, the UNGA features one single diplomatic representative per member state, delegated by the respective governments. Consequently, the chains of delegation and, hence, links to citizens are much more distant in case of the UNGA than in case of the EP. Moreover, due to proportional representation there is more heterogeneity in the political spectrum in the EP, with political parties that are domestically in opposition strongly present. 9

10 To further add to the most different systems design, we study debates on five different policy issues. Each signifies a very different aspect of globalization: climate change, human rights, migration, regional integration and trade. While these issues vastly differ in terms of complexity, nature of the societal problem and degrees of established global governance, they each relate to a different commodity that crosses state borders: pollutants, norms, people, political authority and goods respectively. Preferences on these issues can, hence, either support open borders, international integration and application of global norms and responsibility, or support closure of borders, autarky, or a retraction from global norms (e.g. Held et al. 1999; Nol and Thrien 2008; Zürn 1998). Climate change is perhaps together with human rights the most universal policy issue, which clearly affects all human life around the globe, albeit in different degrees and forms. For our purpose, we look at preferences in favor and against coordinated international efforts to combat climate change and mitigate its effects as opposed to national efforts, abdication in favor of the markets or denial of climate change. For human rights, we distinguish preferences for upholding standards of human rights against preferences for deviations or exceptions. Migration understood as the border crossing of people for a mid- or long-term is often understood as a core component of cultural globalization, but clearly opposed by many citizens. Again, we will focus on preferences in favor or against open borders for immigration and emigration, irrespective of the reason or context given. The fourth issue of regional integration has become particularly salient within Europe recently, and students of a globalization cleavage consider it to be a core component thereof, together with the issue of migration (e.g. Kriesi et al. 2012). Yet, it is clearly not a universal issue and we do not study this in the UN general assembly. Finally, trade or the border crossing of goods and services clearly stands at the very heart of the process of globalization and is often most clearly associated with debates about the extent to which borders should be open or permeable. In the logic of our most different systems design, any observed similarity across institutions and issue areas supporting or refuting our hypotheses implies strong evidence in favor or against them as potential interfering explanatory factors are ruled out. The following analysis, therefore, focuses on similarities, more than differences. While it is beyond the present study and its research design to analyze and explain differences in depth, we will address them in the discussion and engage with tentative explanations with the aim of stimulating future research. Method The chosen method of claims analysis provides us with the opportunity to map the globalization debate in the UNGA and EP. The method combines features of discourse analysis and protest event 10

11 analysis and focuses on collective mobilization in the public domain as the key variable uniting the two methods (Koopmans & Statham 1999: 204). A claim is defined as a: unit of strategic action in the public sphere which articulate[s] political demands, decisions, implementations, calls to action, proposals, criticisms, or physical attacks, which, actually or potentially, affect the interests or integrity of the claimants and/or other collective actors in a policy field (Statham 2005: 12). In this analysis a claim is always a verbal utterance made in the plenary sessions in either one of the two institutions. Claims analysis has the advantage that it systematically opens up the possibility of focusing on the relationship between the actors, their position and the respective constituency represented by the claimant (Koopmans 2004: 11). Due to its well-structured approach with a predefined codebook, the method allows researchers to handle large volumes of data. While some information that cannot be transferred into an a-priori existing coding scheme is lost, the method allows to aggregate qualitative information systematically, given a clear research question and respectively chosen codes. In the statistical analysis applied in this paper to assess the relationship between transnational representatives, their constituents and the position they articulate, the following codes addressed the following questions: WHO is making the claim (claimant: function, nationality and party affiliation); WHAT topic is being discussed (issue); HOW the claimants position themselves in the issue area (position). Furthermore, another code measured WHICH scope the claimants attribute to their political demand (scope) and WHOM he claims to represent (object) (compare: de Wilde 2011: 678). In order to ensure the replicability of the analysis and the reliability of the results intercoder reliability tests were performed. Such a test has been performed for a larger project (including, aside from the EP and UN claims analysis, the analysis of newspaper material in five countries). For each of the potential claims, the majority was taken as the norm and the minority as deviation. Of the total 637 decisions, 146 were minority decisions. This corresponds to a unitizing reliability correlation of.88, which clearly meets reliability requirements. All individual codes used in the analysis were above a threshold of agreement of.70 (for a full overview of the codebook, intercoder reliabilty and descriptive statistics, see De Wilde et al. 2014) 11

12 Sampling & Data For the claims analysis, documents were sampled with a keyword search in the online archives of the UNGA and the EP. The issue areas were used as keywords in the searches. Documents were chosen at random from the hits for each year, so as to ensure an equal distribution across time. The result was a stratified sample by year avoiding a bias in the claims which stems from the dominance of a specific event in a particular year. The time frame of the analysis ( ) was selected to include two different legislative periods (before and after the 2009 elections) in the EP. Consequently, it reduces the impact of the specific composition of the EP on the results of the analysis, as well as diversifying partisanship influence. In case of the UNGA the chosen timeframe also has the advantage of including speeches made by different governments of the same country in some cases. This reduces the influence of a single party on the positions of a state uttered in the Assembly. An exception in the sampling was the issue of regional integration, for which claims were coded, but no separate documents were sampled. This made sense, because questions of the degree and nature of regional integration are immanent in many debates of the EP and reveal important information about how national representatives in a transnational institution conceive of its role. While this is not an important topic in the UNGA debate, the research design opted for including this EU-specific element in order to gather more information about representative claims making. When comparing the aggregated claims in the EP and UNGA (including this EU specific issue), the implicit assumption is made that the issue Regional Integration does not vary fundamentally in its nature from the other issues, as these representative claims also concern preferences to integrate or demarcate transnational cooperation. Operationalization To test our hypotheses, we link the uttering of a preference to the characteristics of the claimant who articulates this preference and to a possible constituency this claimant claims to represent. This allows us to map the existence of an integration/demarcation conflict, possible group formation and potential patterns in representation. The position is the central variable to identify the political demand of the claimants with regard to the opening or closure of borders in the particular issue area. It is designed as an ordinal variable with diametrically opposed poles of integration and demarcation as opposite extremes. An integrate position indicates that the claimant is supporting further integration in the issue area and demands either the opening of borders or wants to preserve already open borders. Integration, in this sense, encompasses any kind of interaction facilitating the free flow of goods, people, norms or 12

13 mitigation measures against pollutants; it can encompass legal agreements, the negotiation of common standards and policies or adherence to any kind of internationally accepted institution, as well as the participation in international organizations. Demarcation advocates want to refrain from cross-border cooperation and integration or maintain a non-integrated status quo. This implies a recurrence of sovereignty claims bound to territorially restricted areas, such as the nation state or a specific region. Given that not all political demands on the five issues can clearly be categorized as favoring demarcation or integration, a third intermediary code was used: Problem signifies that the claimant identifies the issue as of major societal importance to be addressed, but does not indicate a clear direction with regard to openness of borders. Hence, the ordinal position variable captures three categories: integrate (pro permeable borders), problem (demand for attention, without direction) and demarcate (pro closed borders). Given that our main interest lies in the analysis of a connection between patterns of representation and these conflict lines, the object variable was used to capture information on who the claimants depicts as constituency. Representation in this understanding is a process in which the actor or claimant defines himself as the representative of a specific constituency by positively evaluating it in a speech act. This is distinct from formalized representation, which is legitimated directly by the constituency. Claims can only capture a self-ascribed role by the claimant, who speaks out for a certain group. For this reason, it is not a constant characteristic, but may vary in different speech acts or be constant if there is a permanent commitment towards an imagined constituency. Following Anderson s (1983) discussion of the imagined community, our understanding of representation also relies on the representative and the represented most likely having no personal connection and the demarcation of the represented group from an imagined other. The underlying narrative is that the constituency the claimant chooses to represent shares a common characteristic on the basis of which it is possible to exclude others from belonging to this group. In this sense the claimant contributes to cleavage formation by constructing or solidifying subjective identity perceptions (Bartolini and Mair 1990). In order to capture those self-ascribed constituencies we analyze only statements including a positive object evaluation. An object in a speech act is defined as a group of people sharing a certain attributed characteristic and in relation to whose interests the claimant is making a statement. The variable itself consists of the object scope and object function. The object function, describes the particular type of object claimed to be represented, for instance migrants, citizens or farmers. This variable defines the claimant s imagined community with respect to a uniting characteristic, such as being part of the working class. The different object functions are regrouped into territorial and a nonterritorial objects on the basis of theoretical considerations referring to key communitarian and cosmopolitan concepts. Thus, the people, citizens or the country are identified as territorial 13

14 objects because they are defined by their spatial limitation, given that the people are constituted by sharing a common territorial origin. Similarly the country and the citizens are bound to the very concept of territorial borders that define the country and its citizens. Non-Territorial objects are characterized by the absence of a defined territory to which they belong, examples would be women or future generations. The second variable important to our analysis is the object scope referring to the territorial dispersion of the object being represented. A dichotomous variable with two characteristics was used: particular and universal. Particular signifies that the scope of the constituency represented by the claimant is delimited. In case of the EU this means that a claim made on the level of the EU itself (or below) is considered particular because the claimant was not willing or able to detach from a national background or the mandate of the institution he was elected for. For example, in case of an MEP claiming to represent a French constituency (national EU member state constituency) the scope would be particular. The same would be applicable if the scope of the object does not exceed the scope of the EU itself, thus EU regional claims, bilateral and subnational claims are also considered as having a particular scope. In case the MEP refers to a constituency with a wider frame than the EU (another region or global) or outside of the EU (a foreign country) the scope would be universal. A foreign scope, in this regard, stands for a concern or demand which is not made in relation to or including also ones own constituency or institutional scope. This inclusion of the other (Habermas 1998) makes it equivalent to a universal claim, even though it is still focused on a particular territory. In case of the UNGA only objects with a global or foreign scope are considered universal due to the global mandate of the institution itself. Given that the institutional scope of the UNGA is global, a claim on the level of the institution itself is also considered universal. The same argument as above applies to claims with a foreign scope. All claims below the global level (EU regional, other regional, bilateral, national and subnational) are regrouped into the particular object scope. When a claimant presents him or herself as the representative of a particular constituency, the claim always includes a specification of the represented constituency (object function) and the scope of the group the claimant is making the claim for. The combination of the two object characteristics helps us to establish a ladder of representation in the two international strong publics. The most delimited form of representation combines a particular scope and a territorial function. On the opposite side of the ladder would be a combination of the non-territorial function with a universal concern for a community not spatially demarcated. This combination (universal non-territorial) is constructed as the most cosmopolitan type of representation. 14

15 Graph 1: The operationalization of constituency Object Scope + Object Function = Combination Particular Territorial Non-Territorial Particular Territorial Particular Non- Territorial Universal Territorial Non-Territorial Universal Territorial Universal Non- Territorial Findings As the results presented below show, there are major parallels in the observed patterns of positioning in both the UNGA and the EP, suggesting that representative claims making in transnational context follows an overarching logic. We present structured in-case comparisons documenting both the existence of conflict over globalization issues (H1) and the predominance of pro-integration arguments over pro-demarcation arguments (H3a). We then investigate the extent to which vocalized preferences of representatives are linked to the constituencies they claim to represent (H2a) and investigate the extent to which this holds in both UNGA and EP and in the five issue areas. Finally, based on the observed linkages, we analyze the extent to which coalition formation within these forums occurs using weighted metric multidimensional unfolding analysis. Existence and Nature of Globalization Conflict in the UNGA and EP A total of 2050 claims made in both institutions are analyzed which include both a position and a constituency claimed to be represented. That globalization issues are debated and contested in both the UNGA and EP is further shown by an average of 17 and 35 claims per plenary transcript on average respectively. The distribution of the claims is unequal with 1293 claims made in the EP and 757 made 15

16 in the UNGA. This discrepancy is owed to the additional issue of regional integration, which appears prominently in EU debates (522 claims), but only marginally in the General Assembly (12 claims). Graph 2: Mean Positions overall and per issue area 1 0,8 0,6 0,4 0,47 0,7 0,89 0,85 0,57 0,56 0,42 0,64 0,46 0,58 EU 0,2 UN 0-0,2 Overall Human Rights Climate Change Migration Regional Integration Trade -0,23-0,4 The distribution of claims in favor and against integration shows overall claims-making in both institutions is clearly pro-integrationist. The mean position (min. -1, max. 1) across all issue areas in the EP is.47 compared to a mean of.70 of all claims made in the UNGA. Not only, therefore, do we find first evidence in support of H1 that conflict about globalization issues also occurs within international organizations, but also that there is a strong cosmopolitan bias in this conflict (H3a). More subtle similarities and differences are revealed, when assessing positions on the issue areas in the sample separately. The debate about human rights is extremely cosmopolitan in both forums, with means of 0.85 and 0.89 meaning that representatives in the UNGA or the EP hardly publically challenge the applicability or need to enforce human rights around the globe. The only exception to the strong cosmopolitan dominance in the debates where the mean position is below 0.0 is the debate on trade within the EP. In a second step we assessed the degree of conflict in the two institutions and the different issue areas. Conflict could be evaluated, firstly, in terms of the mere presence of two opposing preferences voiced 16

17 or, secondly, by the depth of conflict. In this respect, one could either measure the distance between the positions or relate the balance between pro-integrationist and pro-demarcationist positions compared to each other. In our case the distance between the positions is somewhat static, as we work with a fixed threefold indicator for position (demarcate, problem and integrate). For this reason, the measurement of conflict relies on the comparison of the relative frequency with which prointegrationist and pro-demarcationist arguments are made. We operate with a measurement of depth of conflict which relies on a division of the number of the less frequent position (either demarcate or integrate) by the number of the most frequent position, receiving a measure between 0 and 1, with 0 indicating the total absence of conflict and 1 indicating maximum polarization. Table 1: Measure of Conflict Overall and per Issue Area Overall Climate Change Human Rights Migration Regional Integration Trade EU UN Table 1 provides further evidence for H1, while also showing that the issue areas differ markedly regarding the degree of polarization. Trade presents the most polarized issue area in both institutions, followed by migration. Whereas the degree of polarization is much higher in the EP in both issue areas, the overall ordering of the issue areas according to the depth of conflict is comparable. In other words, human rights and climate change are treated as valance issues: There is no consensus on their salience, so conflict exists between those attributing salience to them and those who do not, rather than contest over which policy to pursue (Green 2007). Once representatives accept as a fact that human rights exist and are violated, the argument that something needs to be done about violations almost immediately follows. No one argues that human rights are violated and that this is a good thing. Instead, those who do not want to act against human rights violations either keep silent or argue that there are no human rights violations to begin with. In contrast, positional conflicts over migration, trade or regional integration feature opposing policy demands. Some representatives argue in favor of free migration, more transfer of sovereignty to the supranational level or more free trade, while others demand the exact opposite. Meanwhile, both sides acknowledge that these issues are important and should be of common concern. As evidenced by the mean positions and the degree of polarization on these various policy issues, there is striking similarity between the UNGA and the EP in terms of which issues are valence issues and which are positional. Differences also remain, with stronger polarization in the EP than in the UNGA, to which we will return in the discussion. 17

18 Overall, the analysis shows that the EP is the institution, in which claimants are more likely to engage in conflictual claims-making and take opposing positions, whereas the UNGA is characterized by a large degree of homogeneity in the direction of pro-integration claims. Nevertheless, even this overall relatively homogenous debate structure knows conflictual issue areas, which might be very unbalanced but conflict is still present. On the other hand, we have to acknowledge that even in an institution like the EP, where conflict seems to underpin most of the issue areas, there are exceptions to the rule such as the human rights issue. Here the degree of overall consent toward integration forbids speaking of a true conflict. The distribution of claims clearly shows that even though one may indicate overall trends of existing conflict, it is extremely important to take a closer look at the particular issue structure to understand the role of conflict over integration or demarcation in the two strong publics. Cleavage Formation: Constituencies and Group Coherence This paper is not only interested in mapping positions in favor and against integration, but also in connecting this positioning to patterns of representation beyond nation state borders and to analyze to what extent each side of the debate is represented by different groups or coalitions of representatives. The extent of normative underpinning through subjective identity formation and of coalition formation would point to a process of cleavage formation based on cosmopolitanism and communitarianism beyond the state. Starting with representation, we examine the existence of a pattern between the constituency claimed to be represented and the position taken by a claimant. For this, a combination of object scope and object function serves as the categorization of different types of constituencies. The most inclusive categorization combines a universal scope with a non-territorial function, while the combination of a particular scope and a territorial function is conceptualized as the most delimited constituency being possibly represented by a claimant. In a first step the mean positions of the different combinations of object scope and object function are analyzed. Secondly, the correlation between object scope, object function, the combination of the two and the position with regard to integration is evaluated. Graph 3 reveals a clear overall pattern in both institutions: the more inclusive the represented constituency gets, the more prone is the claimant representing this constituency to make a prointegrationist claim. While delimited constituencies in either scope or function resonate with less 18

19 integrationist claims-making, the reference to inclusive constituencies (in its most extreme form a universal non-territorial constituency) correlates with demands for open borders. Graph 3: Mean position across all issue areas for combination of object scope and object function (per institution) 1 0,8 0,6 0,61 0,5 0,67 0,59 0,75 0,75 0,82 EU 0,4 0,29 UN 0,2 0 Particular Territorial Particular Non- Territorial Universal Territorial Universal Non- Territorial Looking at the patterns in the specific issue areas, we find a very similar connection between the specific constituencies being represented and positioning with respect to permeable borders in the climate change, human rights and migration debate. The only exception to that pattern is the debate on trade. It is the most demarcationist issue area in terms of mean position and the claim to represent a more inclusive constituency does not correspond with a more integrationist position of the respective claimant. In case of the EU there is no traceable pattern with respect to represented constituency and position taken. For trade claims made in the UNGA the most exclusive constituency has the most integrationist position (.66) and the most inclusive possibility of representation presents the most demarcationist positioning (-.20). The descriptive statistics presented in graph 3 above visualize what proves to be a significant positive correlation between the position and the combination of object scope and function as well as with the two object characteristics on their own. 19

20 Table 2: Correlation between position and constituency Object Scope + Function Object Scope Object Function Position UN Pearson Correlation,152**,142**,083* Sig. (2-tailed),000,000,031 N EU Pearson Correlation,202**,171**,109** Sig. (2-tailed),000,000,000 N This strong evidence in favor of H2 implies that we are not facing unstructured, varying conflicts over globalization issues, but that voiced preferences on the permeability of borders are normatively underpinned by group identity perceptions and role perceptions of representatives as the hypothesis that we are facing cleavage formation along cosmopolitan and communitarian lines predicts. It is, in fact, likely that the role perceptions of representatives in terms of the constituencies whose interests they need to take into account lead them to argue in favor or against integration on a variety of issues. The fact that the relation between claimed constituency and preference holds not only in the UNGA and in the EP, but also in four out of our five issue areas supports this strongly. However, additional analysis is needed to investigate whether representatives take up coherent positions arguing consistently in favor of either integration or demarcation related to a consistent role perception. If this is the case, we should find identifiable groups in both forums with cosmopolitan representatives clearly distinguishable from communitarian representatives. If, on the other hand, representatives freely shift from making a pro-integrationist argument for migrants around the world at one moment to making a demarcationist argument to preserve national sovereignty at the next, we should not find clear opposing coalitions of representatives in the debate. To test H2b on the existence of cleavage coalitions in both forums, we conduct weighted metric multidimensional unfolding (WMMDU). Such a technique allows the dimensional mapping of distances between two different units in our case representatives and their claims on the five issues (Borg and Groenen 2005). First, we constructed a cosmopolitan and communitarian index based on 20

21 the mean of position and constituency, both scaled from -1 (communitarian) to +1 (cosmopolitan) 3. Subsequently, we aggregated the data to the level of country in the UNGA and the combination of country and party family in the EP. The subsequent WMMDU analysis was executed using the PREFSCAL algorithm available in SPSS, which in relation to other forms of unfolding successfully limits degeneracy (Busing et al. 2005). Each actor, whether a country in the UNGA or a national party in the EP, is attributed a preference on each issue in the form of the distance of its mean position to both the cosmopolitan pole of +1 and the communitarian pole of -1, where the latter is the inverse of the former. The preferences are weighted by the amount of claims this actor makes on the issue, so that the final solution of the analysis positions an actor more precisely in relation to the issue poles on which it makes many claims than on the issue poles in which it makes few claims, if both cannot be reconciled. The PREFSCAL algorithm has as additional major advantage that the assumption that the order of preferences is linear can be relaxed (Busing et al. 2010), which given the quasi-linear and ordinal nature of the two components of the index more accurately reflects the original data. Given the lack of correlation between position and constituency in the issue of trade, the position values are used in the WMMDU for this issue, instead of the cosmopolitanism/communitarianism index. The results are visualized in graphs 4 and 5. Although Kruskal s Stress-I levels are above the generally accepted levels of.20, this can be explained by the high number of cases plotted against the two dimensional space, which is beyond the factor 10 usually applied (Borg and Groenen 2005: 48). Note further that non-degeneracy levels are remarkably high, well above the usually applied threshold of.70. We therefore accept these results as adequate representations of reality. In the case of the UNGA, we find that the vast majority of state representatives are located in the vicinity of the cosmopolitan poles on human rights (HR cos), migration (M cos) and trade (T cos). A few disperse outliers including Russia, Tajikistan and Trinidad and Tobago are located closely to the communitarian poles on human rights (HR com) and trade (T com) and some, like Israel and Lithuania approach the communitarian migration pole (M com). At first sight, the finding of Fiji located close to the communitarian pole on climate change (CC com) is strange as one would expect such a directly affected island state to be particularly concerned about combating global climate change. However, keep in mind that the index incorporates the represented constituency. The Fiji representative problematized climate change on behalf of his national constituency, rather than on behalf of humanity 3 The variable position was organized in three values: -1 demarcation, 0 problem, +1 integration. The variable constituency was attributed values as follows: -1 particular territorial, -0.5 particular non-territorial, +0.5 universal territorial, +1 universal non-territorial. 21

22 as a whole. In terms of representation, this is a communitarian claim. The fact that most states are located around the cosmopolitan poles visualizes the strong dominance of cosmopolitan discourse in the UNGA. We see furthermore that there is no clear coalition formation, as each of the caucus groups are spread across the space. To a limited extent, we find support for the notion that there is a conflict between the West and the Rest in the UNGA (Voeten 2000), as the West - with the exception of Israel - is unified around the cosmopolitan center while the Rest is widely dispersed across both the cosmopolitan center and communitarian poles. Hence, cleavage formation in terms of opposing coalitions mobilizing each camp clearly remains very limited in the UNGA discourse. Graph 4: Joint plot of WMMDU in the UN General Assembly In the case of the EP, this is clearly different. First of all, we find a one dimensional constellation of issues with the cosmopolitan poles of migration (M cos), regional integration (RI cos) and climate 22

23 change (CC cos) located together and their opposites (M com, RI com and CC com) too. Furthermore, while all party groups are obviously diffuse in terms of their member parties positions, there is a clear pattern of their mean positions being organized along this one dimension. Closest to the cosmopolitan poles on these three issues, we find the Greens followed closely by the social democratic Socialism and Democracy (S&D) group. Moving towards the middle of this dimension, we find the Alliance of Liberals and Democrats for Europe (ALDE) and the socialist Confederal Group of the European United Left/Nordic Green Left (GUE_NGL), then come the conservative European People's Party (EPP) and finally, close to the communitarian poles, we find the far right Europe of Freedom and Direct Democracy (EFDD). All of these parties are close to the human rights cosmopolitan pole, indicating lack of positional conflict on this issue. Trade turns out as an orthogonal and subordinate dimension, delivering additional evidence that this issue does not simply align in the emerging globalization cleavage as was already indicated by the lack of correlation between position and claimed constituency on this issue. Graph 5: Joint plot of WMMDU in the European Parliament 23