United we stood, divided we transform? Exploring coalition transformation divergence in the EU trade policy field

ABSTRACT During the EU-US (TTIP) and EU–Canada (CETA) free trade negotiations, large coalitions of civil society organisations were active not only across borders but also within European member states. In several countries, coalitions saw the opportunity to transform their issue-specific group into a general coalition on EU trade policy in order to achieve more sustained engagement. However, in hindsight, only some of the transformed coalitions remained active and visible with the same organisations, while others experienced a decline in visibility, activities, and membership. This study aims to explore the factors contributing to this divergence in coalition transformation, drawing on the literature from social movement and interest group studies. Based on interviews with trade activists in Belgium and the Netherlands, the analysis points to differences in perception of political and discursive opportunities, resource mobilisation, the degree of ideological and cultural overlap between the coalition’s actors, and organisational structure as important factors.


Introduction
Between 2013 and 2017, several EU free trade agreements (FTAs) were heavily politicised in various EU Member States.Particularly the Transatlantic Trade and Investment Partnership (TTIP) and the Comprehensive Economic and Trade Agreement (CETA) with Canada attracted the attention of a wide range of NGOs, trade unions, farmers' associations, and citizen groups, that formed coalitions.In several countries, broad and comprehensive coalitions of these organisations eventually seized the moment and transformed their issue-specific (Stop TTIP/CETA) group into a general coalition on EU trade policy, to make their engagement more sustainable.In hindsight, however, the transformed coalitions appear to have diverged sharply: in some countries they remained visible, were broadly active, and were supported by a similar group of organisations, while in others, membership and their activity and visibility declined.The purpose of this paper is to explore the possible factors underlying this divergence in coalition transformation in the EU trade policy field.
At the time, TTIP and CETA were prime examples of a new type of EU FTAs, deeper and more comprehensive than previous agreements that focused mainly on tariff liberalisation (Araujo, 2019;De Ville & Siles-Brügge, 2015;Laursen & Roederer-Rynning, 2017).Contemporary FTAs include chapters on issues such as sustainable development, investment protection, intellectual property rights, or regulatory cooperation on technical or food standards.Opponents have challenged TTIP and CETA, either to reform specific parts or to reject them in principle as it was perceived by some as undermining democracy, by giving large corporations additional opportunities to challenge or water down legislation at the European or national level (De Ville & Siles-Brügge, 2017;Verbeek, 2022).
This strong politicisation of EU FTAs has led (trade) scholars to examine the origins and differentiated dynamics of politicisation (De Bièvre & Poletti, 2020;De Ville & Siles-Brügge, 2015;Duina, 2019;Gheyle, 2019;Young, 2016).While several developments conspired to create the intensity of contentious politics surrounding TTIP/CETA (cf.Gheyle & Rone, 2022), it is nevertheless widely acknowledged in this literature that the work of civil society organisations was central.By raising awareness of the dangers of a potential agreement and increasing the relevance of negotiations to a wider audience, CSOs put pressure on politicians to act, punched above their weight, and contributed significantly to the eventual freezing of the TTIP negotiations, and the near derailment of CETA in 2016.
Much of their success stems primarily from the fact that they did not act alone.Several studies have shown how civil society organisations have joined forces at both national and transnational levels, bringing together NGOs, trade unions, farmers organisations or citizen groups (De Ville & Siles-Brügge, 2015;Eliasson & Garcia-Duran, 2018;Gheyle, 2020;Oleart, 2021;Søndergaard, 2020).These (trans)national coalitions facilitated resource pooling, information exchange, strategy coordination, access to policymakers, and often led to a framing of the TTIP debate on their terms.At the height of the politicisation of TTIP and CETA, broad civil society coalitions were active in several countries, and are therefore seen as an important factor in why we have observed a differentiated politicisation of EU trade agreements over time and in different countries (De Bièvre & Poletti, 2020;Gheyle, 2020;Meunier & Czesana, 2019).
In late 2016after the 'freezing' of TTIP and the signing of CETAsome of these broad coalitions sought to make their engagement and contestation more permanent, by transforming their coalition into a general coalition focused on different elements of EU trade policy.Examples include the 'Netzwerk Gerechter Welthandel' (Network Fair World Trade) in Germany, 'Anders handeln!' or 'Handel Anders' (Trade differently!) in Austria and The Netherlands, or the 'Platform Rechtvaardige & Duurzame Handel' (Platform for Sustainable and Fair Trade) in Belgium.In retrospect, however, only some of these coalitions (such as those in Austria or the Netherlands) remained largely visible to the outside world, showed continuous activity on ongoing and emerging policy issues, and had consistent (or growing) membership compared to their issue-specific coalition.In other countries (e.g., in Belgium) the transformed coalition had no outwardly visible identity, its membership declined to pre-TTIP levels, and (viewed from the outside) it did not seem to succeed in focusing attention on (other) targets.
Given its importance for politicisation processes and to better understand the 'ebb and flow' of trade activism, it is therefore important to examine the factors behind this divergence of coalition transformation.While the literature on (the divergence of) coalition transformation itself is scarce, the topic bears many similarities to studies of coalition formation, activity and longevity.Therefore, I draw inspiration from these closely related aspects of coalitions to conduct an exploratory analysis of two divergent (transformed) trade coalitions: the 'Platform Rechtvaardige en Duurzame Handel' in Belgium and 'Handel Anders!' in the Netherlands.Based on semi-structured interviews, the findings point to differences in perceived political and discursive opportunities, resource mobilisation and ideological (and cultural) overlaps, and organisation structure as important factors.In the final section, I summarise these findings and draw overarching conclusions.

Coalition formation, activity and longevity
Coalitions are defined as 'organizational collaborations wherein distinct organizations pool resources to pursue shared goals' (Van Dyke & Amos, 2017).These distinct organisations are individual political advocates, such as NGOs, trade unions, farmer associations, or think tanks, who lobby together on a certain topic (Junk, 2020b).Coalition formation can also be viewed from the perspective of social movement literature, in which these 'organizational collaborations' are seen as part of 'the network of individuals and organizations that comprise a social movement' (Van Dyke & Amos, 2017, p. 17).In terms of activities, coalitions engage in cooperative behaviour, such as information exchange, coordinating strategies to pursue common goals, achieving objectives, or jointly signalling support to policymakers (Junk, 2020b;Van Dyke & Amos, 2017).Coalitions can either be issue-specific, instrumental, event coalitions, or they can take a more general form that involves some degree of shared identity, often with its own organisation (Levi & Murphy, 2006;Van Dyke & Amos, 2017).
In what follows, I draw inspiration from studies of both interest group and social movements.While the former tends to assume a rationalist costbenefit and a resource exchange model to explain coalition formation and activities, social movement studies tend to emphasise cultural aspects such as shared identity or trust relationships.Thus, bringing the two approaches together can lead to complementary insights.Moreover, neither literature devotes much attention to coalition transformation, or the reasons for divergence.
To gain analytical insights into the possible determinants, I turn to literature closely related to this idea: coalition formation, activity, and longevity (cf.Staggenborg, 1986;Van Dyke & Amos, 2017).Van Dyke and Amos (2017) summarise the literature on coalition longevity, arguing that 'many of the same factors that influence social movement coalition formation continue to shape organizational collaboration over time ' (p. 8).The focus of the work summarised, however, is on the fate of untransformed coalitions.Parallel to this observation, therefore, it is useful to draw inspiration from these closely related areas to guide the analysis of coalition transformation divergence.
Coalitions are formed by organisations that act as strategic entities that rely on scarce resources provided and influenced by their environment for their goals (e.g., survival, impact) (Junk, 2020b).Hence, to achieve their goals, they depend on the exchange of resources with other organisations.One response to this dependence is to form coalitions with other groups to exchange resources in order to achieve a higher likelihood of gaining access influencing policymaking (Box-Steffensmeier et al., 2018;Junk, 2019) or to be considered relevant to constituents or the general public (Dür & Mateo, 2014).Typical resources that are exchanged include expert knowledge, dedicated staff, networks, membership bases to strengthen mobilisations, and the reputation of specific organisations (Edwards & McCarthy, 2007).Several studies have shown that particularly diverse coalitions have a positive impact on advocacy success, measured as perceived influence (Dwidar, 2022;Junk, 2020a;Nelson & Yackee, 2012).
However, forming coalitions also comes with a cost.Not only does continued collaboration cost money and time to invest in the coalition, but there is also a potential loss of flexibility in taking positions, and thus a greater difficulty in being responsive to one's constituency, which can have a negative impact on reputation (Holyoke, 2020).The benefits of a coalition therefore have to be weighed against the costs of moderating a position, potentially losing an organisation's uniqueness or niche position.Below, I review four analytical categories that appear to play a role in increasing costs or benefits.These will serve as guideposts later in the semi-structured interviews and in the analysis of empirical cases (Teixeira & Motta, 2022;Van Dyke & Amos, 2017;Zajak et al., 2018).

Political and discursive opportunities
A first aspect that drives coalition formation or activity is the environmental context in terms of (potential) opportunities and threats (Montgomery & Baglioni, 2022;Staggenborg, 1986).In social movement terminology, this is referred to as the political opportunity structure (POS).Because this concept has become a 'sponge that soaks up virtually every aspect of the social movement environment' (Gamson & Meyer, 1996), it is more narrowly defined here as perceived political opportunities that may encourage actors to form coalitions or campaign (Giugni, 2009;Princen & Kerremans, 2008).In other words, institutional or political aspects that trigger groups' motivation and willingness to form or maintain a coalition, depending on their assessment of whether there is a chance of success.Opportunities are what actors make of them and are therefore related to political strategy: where do actors see the opportunity to exert influence?
Based on this interpretation of the incentives for collective action, two dimensions can be identified.The first is the political-institutional dimension, where the perceived opportunity lies in a susceptible and conducive political constellation in parliament and government.If the ruling (government or parliamentary) majority is completely opposed to your ideas and preferences, this is not necessarily an obstacle to collective action, but the chances of success may be perceived as less favourable.Similarly, strong institutional prerogatives for national parliaments (or federal systems) may create additional opportunities and incentives for activism.Second, scholars also emphasise discursive opportunities, which refer to the extent to which demands are consistent with prevailing discourses in the public sphere (De Wilde & Zürn, 2012;Giugni, 2009).This means that the presence of highly mediatised events creates opportunities to put forward claims that resonate with mediatised discourse.

Resource mobilisation
A second aspect is resource mobilisation.Coalition formation, and its subsequent activity and visibility depends on the amount and variety of resources that are pooled in the coalition (Levi & Murphy, 2006).It has been shown, for example, that diverse coalitions (entailing, for example, NGOs and trade unions) have a positive effect on lobby access and influence (Junk, 2019;Madariaga & Allain, 2020;Szabó et al., 2021).To a large extent, this is due to important resources like expert knowledge, campaign experience, or mobilisation power being combined, leading to the utilisation of inside and outside lobbying channels (De Bruycker & Beyers, 2019).In addition, groups may have 'network resources' by being embedded in other (trans)national networks, or social media experience.Ultimately, what can be pooled depends on what each organisation is effectively able to bring to the table.If no time or dedicated staff can be devoted, either because it is lacking, or used for other priorities, coalition resources are hampered.Similarly, mobilisation power remains a potential power source that needs to be activated in order to play an actual role.

Ideological and cultural overlap
Resource pooling is related to the overlap of ideology and culture between organisationsespecially in more general, intensive coalitions.A largescale QCA analysis has shown that this aspect of shared ideology is crucial, and may even be sufficient to build and sustain coalitions (McCammon & Van Dyke, 2010).Of course, even organisations with similar goals and ideology may disagree on certain aspects such as strategy or scope of action (Van Dyke & Amos, 2017).In addition to ideology, a shared culture and identity is also believed to have an impact on coalition building.This refers to a general culture of friendship and trust, as well as working relationships conducive to ongoing collaboration (Hathaway & Meyer, 1993).
In the social movement literature that empirically addresses the issue of coalition diversity and ideological overlap, the link between NGOs and trade unions in particular is seen as a promising (albeit challenging) collaboration (Frege et al., 2003;Schmidt, 2005;Waterman, 2021).Against a backdrop of declining membership and a general diversification of the membership base, union goals have gradually begun to overlap with those of NGOs (Frege et al., 2003;Waterman, 2021).However, although the expected (and demonstrated) benefits for both sides are obvious (Portos & Carvalho, 2022;Szabó et al., 2021), many potential obstacles remain, ideological and cultural.NGO positions are sometimes seen as 'too radical' for union cooperation, in addition to different styles and tone of communication (Schmidt, 2005).Unions, for their part, are sometimes criticised for their slower and more hierarchical decision-making procedures (Heery et al., 2012).
The evaluations that NGOs and trade unions make of capitalism, globalisation, or EU trade policy often differ ideologically (Bieler et al., 2014;Waterman, 2021).While there are also a variety of opinions in the NGO community (cf.Choudry, 2014), many NGOs concerned with EU trade policy are embedded in the Global Justice Movement, and fundamentally challenge the neoliberal logic of (EU) trade policy (Della Porta, 2007;Verbeek, 2022).Trade unions, however, are more divided and often support the EU's trade agenda, while advocating for the inclusion of social conditionalities to mitigate the most damaging effects on labour (Hilary, 2014).As Bieler et al. (2014) argue, unions are divided over how to position themselves to FTAs: the challenge of "free trade" requires trade unions to choose between two competing visions of their identity: as junior partners of capital based on shared interests with their national industries, or as active members of an internationalist movement, the role of which is to contest the power of capital on behalf of working people the world over.(p.7)

Leadership and structure
A final aspect to coalition formation and activity is leadership and structure, to advance the coalition agenda, build compromises, and create space for conflict and deliberation.This means first avoiding the suppression of conflict, and instead using the coalition as a flexible space for deliberation (Hathaway & Meyer, 1993;Levi & Murphy, 2006).It is not necessary that everyone do everything together, or think exactly the same.Organisationally this can be implemented in the practice of subgroups that address different issues (Hertel-Fernandez 2016).Another way to avoid this conflict is to find a broad and inclusive framework that all stakeholders support (Gheyle, 2020).Organisations that are able to move beyond their own specific framing of a policy problem (and solution) and instead create a more inclusive frame remain active together, and are more likely to succeed (Junk & Rasmussen, 2019;Szabó et al., 2021).
A second aspect concerns the presence of so-called bridge-building coalition leaders (Dobbie, 2009) or meso-mobilisation actors (Gerhards & Rucht, 1992).These are (individuals within) organisations who consistently work to unite members' interests, and transform them into something common.Thus, they actively help to defuse conflict within the coalition, but they are also the ones who continually drive the common work forward.They are 'activist-intellectuals' who not only contribute to the policy work, but also take a step back, look at the big picture, and think strategically about who to engage, who to target, and how to frame their messages.Multi-issue organisations are better able to play such a role, because they can bridge the interests of different issue-specific groups (Van Dyke & Amos, 2017).

Research design
This is an exploratory study that seeks to probe the factors contributing to a divergence in coalition transformation in the EU trade policy field.This variation is decomposed into three parts: external visibility and identity of the coalition, activity, and membership dynamics.Visibility and identity mean the continued presence of the coalition (name) on social media, separate web pages, and in published material and outreach to other audiences.Activity refers to the continuous activity and diversity of advocacy goals that a coalition undertakes.Membership dynamics measures whether organisations have left or joined the coalition since transformation, which is an indicator of the coalition's continued viability.These three indicators are assumed to correlate positively with the analytical categories presented in the previous section.
Regarding the selection of cases, it is first important to note that not every coalition active during the politicised TTIP/CETA episode is eligible for analysis in this framework.Rather, I am interested in the cases where the coalitions were already visible and active as issue-specific Stop TTIP/CETA groups, bringing together different types of organisations: trade unions, NGOs, and professional associations (such as farmers or consumer organisations).If this were not the case, the comparison would be between coalitions with different starting points.Crucially, several of these existing broad and comprehensive coalitions seized the moment and transformed their coalition (with a new name and a dedicated structure) to remain active on a permanent basis.After this transformation, however, we see different developments with respect to the three indicators mentioned above.
Case selection in exploratory research often follows a diverse case selection logic (Gerring, 2007).This means that cases are selected that differ as much as possible with respect to the dependent variable, in order to show variation.I thus chose two cases that differ significantly in all three aspects: visibility, activity, and membership dynamics.Specifically, I consider these variables dichotomous (1 or 0): a coalition is visible to the outside world (through various types of media or press) or not; it is or remains active on various trade-related issues, or not; its membership remains the same, or has grown since the transformation, or not.The assessment of these indicators is based on the basis of desk research on the (online) activities of the coalition.Thus, it is these two aspects (similar starting points, but different developments in the transformed coalitions) that ultimately guide the selection (and relevance) of the cases.With these indicators in mind, I chose to compare coalitions in Belgium and the Netherlands.In both countries, a broad and comprehensive Stop TTIP group was present during the height of the politicisation of TTIP/CETA (see Gheyle, 2019), and both have made the transition to a more general coalition, but seem to have strongly diverged since.
First, in terms of visibility, the Dutch 'Handel Anders!' (Trade Differently!-HA) coalition has its own social media and website, where all activities, news, and information about the coalition are centralised.They have their own logo, and in November 2020 an event was organised to mediatise the official launch of the coalition.For the occasion, a publication 'Trade differently!' was distributed, presenting their common principles, demands, and alternatives, with politicians from different progressive parties present.The name 'Handel Anders!' also appeared frequently in press releases, progressive media, and sometimes in the mass media.This contrasts with the Belgian situation, where a general coalition was established under the name 'Platform Rechtvaardige & Duurzame Handel' (Platform for Fair and Sustainable Trade -PRDH) but where these aspects of external visibility are hardly present.A Google search (mid-2022) reveals that the name is hardly mentioned, apart from a joint political statement on the Energy Charter Treaty.
Second, we also see a large difference between the two countries in terms of continued activity on ongoing and emerging policy issues.Based on the newsfeed of the HA website, we can identify 10-12 different issues on which the coalition has been active.The most prominent of these were CETA, ISDS, EU-Mercosur, and the Energy Charter Treaty.A combination of protest and direct-action events, (international) petitions, letters to politicians, and general messaging is used.Smaller issues include the Comprehensive Agreement on Investment (CAI) with China, the Single Entry Point, the TRIPS Waiver, TiSA, EU-Singapore FTA, and the UN Binding Treaty on Business and Human Rights.This again contrasts with PRDH, where only the activity towards ECT can be found online.Importantly, this does not mean that the other topics were not discussed in the Belgian context, but only that the coalition as such has not communicated about them collectively (cf.infra).
Finally, in terms of membership dynamics, we also see opposing movements.Table 1 shows the composition of both coalitions, indicating internal movements since the transformation.It shows that the membership of the HA coalition has remained very similar to the issue-specific TTIP/CETA coalition (with the exception of the consumer organisation no longer involved), and that new organisations (Greenpeace and World Animal Protection) have joined the cause.In Belgium on the other hand, PRDH has seen the departure of the consumer organisation (TestAankoop), the human rights organisation (Liga voor de Mensenrechten) and environmental organisations (Greenpeace, BBL, Inter-Environnement Wallonie) as members.Many farmers' associations are also no longer official members of PRDH, but collaborate on an ad hoc basis.Overall, the Netherlands shows similar (and even increasing) membership since transformation, while the group in Belgium has almost returned to pre-TTIP membership characteristics (Gheyle, 2020).
To explore the factors for this divergence in coalition transformation, I rely on desk research and semi-structured interviews with members of each coalition.Given the exploratory nature and the reliance on two case studies with four variables, it is not possible to draw conclusions in terms of necessary or sufficient variables.However, the comparison of both coalitions still provides qualitative in-depth descriptions and a comparison of coalition politics, potentially indicating where variables overlap.Interviews are preferred, not only because it is the only source of information on the internal politics of coalitions, but also because this study is concerned with actors' perceptions of political opportunities, shared interests, or conflict (cf.Kröger, 2018).Interview questions addressed (i) the origin and motivation for transforming the coalition; (ii) membership dynamics; (iii) internal politics (assessment of resources, functioning, structure); (iv) conflict (resolution); (v) different reasons behind coalition (perceived) success; and (vi) reflections on the coalition existence and future.
Questions were initially asked as open-ended as possible, to avoid directing responses to any of the four analytic categories presented above.When certain categories were not mentioned spontaneously, I asked about them specifically to assess their relevance.I conducted 11 interviews, with representatives of NGOs, trade unions, and professional associations (11.11.11-CNCD,ABVV-FGTB, ACV-CSC, Solidaris, Oxfam Belgium, Handel Anders, TNI, SOMO, FNV, Both Ends and Platform ABC -Landbouwcoalitie Rechtvaardige Handel).Since the analysis deals with issues of internal politics of coalitions (with the presence of conflict) all statements were anonymised and reduced to a number.In the online appendix, an overview and short description of all organisations, and meta-information on the interview, can be found.

Analysis
In the empirical analysis below, I qualitatively compare the material from my interviews on the basis of the four categories outlined in the literature overview: opportunities, resources, ideology and culture, and organisational structures.

Political and discursive opportunities
In the Netherlands, activists were generally positive about the influence they could exert on EU trade policy and even referred to the country's strategic position (2, 3, 10).They cited the closeness of political circles and contacts in a small country (3) and, most importantly, the political constellations in parliament where majorities against EU FTAs have a real possibility of coming about (2, 3, 9, 10).While left-wing parties are typical allies, several people also mentioned the fact that far-right parties such as PVV or Forum voor Democratie strongly oppose trade agreements because of their impact on Dutch sovereignty or agriculture (9, 10).Although there is little cooperation between activists and these parties, the fact that there is opposition of the left and right blocs means that the fate of trade agreements is currently in the hands of centrist parties, which have only a slim majority.This opens up a real avenue for influence when social democratic or Christian democratic parties oppose an issue.From a strategic point of view, activists therefore emphasise the crucial cooperation between more radical groups and trade unions or farmers' associations that have closer ties to these centrist parties (3, 9).Equally important is the political agenda: activists largely depend on what issues are being actively negotiated and at what stage they are (2, 7, 9, 10).Different stages of decision-making can influence activity, as the lingering ratification process and debate over CETA demonstrates.This is a 'mixed' agreement that must be ratified by all national parliaments (cf.Conconi et al., 2021).However, due to opposition from the ChristenUnie (as a governing party), there was a real possibility that the Eerste Kamer (Senate) would reject the agreement.It was not until September 2022 -6 years after the agreement was signed, that the Senate narrowly allowed CETA to pass, after the Social Democratic PvdA voted in favour.Several observers documented that this Social Democratic support was conditioned on the introduction of a special notification procedure for violations of labour and environmental provisions in other FTAs. 1 This episode shows, on the one hand, how 'kicking the can down the road' of CETA was very useful in keeping the coalition active (2, 7, 9, 10) and how the Dutch political constellation provides an avenue for influence in general (2, 3, 10).
The Belgian situation was not found to be similarly advantageous.Here, the federal regions (Flanders, Brussels, Wallonia) have competences in international trade policy according to the principle of 'in foro interno in foro externo' (Bursens & De Bièvre, 2023), which means that they have to cooperate to arrive at a Belgian position.Activists recognise that much is possible in the Brussels and Wallonia region due to predominantly (and historically) left-wing governments, but confronted with the right-wing protrade stance of Flanders, Belgium often has to abstain in the Council, and is therefore no change maker in EU trade policy (1,4,6,11).The Flemish far-right Vlaams Belang does not play a similar role in providing majorities, as it was relatively small before 2019, while it has recently voted inconsistently on EU FTAs 2 , and faces a 'cordon sanitaire' dynamic in which other parties are careful not to vote together with the extreme right (1).Moreover, parliamentary ratification procedures (e.g., on CETA) are less urgent, being postponed to an unknown date to avoid new diplomatic upheavals. 3While several activists argued that the Belgian government is conducive to labour rights arguments (4, 6), there are others who seriously doubt that EU trade policy can be influenced through the Belgian route, and even give higher importance to European civil society umbrella organisations (6, 11).
Finally, activists mentioned discursive opportunities that facilitated sustained mobilisation and coalition activity in the Netherlands, but less so in Belgium.In terms of access to the media, both the Netherlands (9, 10) and Flanders (1, 4, 6) were seen as hostile environments for disseminating alternative trade messages.Nevertheless, Dutch activists were positive about the potential resonance of their messages.For example, given the Europe-wide focus on investment protection as the most controversial element of contemporary trade agreements (Siles-Brügge, 2017), the Netherlands has a unique international position as a party to many bilateral investment treaties with third countries, with so-called 'letterbox companies' using the Dutch link to sue governments (Vervest et al., 2015).In the context of the Energy Charter Treaty, the Netherlands itself has been sued twice by energy companies over its decision to phase out coal. 4 These local and actual lawsuits were heavily mediatised, with activists arguing that this helped to maintain attention on ISDS in general, and ECT in particular (8, 9).

Resource mobilisation
Coalition activity, visibility and membership growth also depend on the pooled resources that an alliance brings together.In Belgium, lack of interest, lack of time, and different priorities of several organisations were often cited as explanations for why membership and activity have declined to pre-TTIP levels (1,4,5,11).While part of the argument is that there are simply too many negotiations and issues to pursue for the often 1-2 people involved in trade policy (1,4,5) others argue that it is simply not their main priority.EU regulations, the health sector, development projects are the main priorities, and only when it is convincingly argued that EU agreements threaten these aspects (as was the case with TTIP, for example) do these organisations want to engage with FTAs (1,4,5,11).Only in the context of EU-Mercosur was this threshold sufficiently met, so that an ad hoc Stop EU-Mercosur coalition was formed, outside the Platform.
After the departure of several organisations, the driving force of the PRDH is 11.11.11-CNCD(which are North-South umbrella organisations that group many smaller organisations) and institutional actors: trade unions and health insurance groups (see Table 1).This has a historical aspect, as this group has been collectively active on trade issues since at least the mid-2000s (Gheyle, 2020).While several activists emphasised the importance of resource diversity, the activities and focus of the Belgian coalition are nonetheless the result of what these resourceful members find important (6, 11).Moreover, unions (and health insurers) still have a strong (representative) position in Belgium, which means that they sometimes rely on their own contacts and (internal) channels, rather than working through the Platform (6,11).
The HA coalition, on the other hand, has a different core group: it consists of a variety of NGOs, farmers' associations, and the largest trade union.This makes the alliance more diverse, but also somewhat less representative compared to Belgium, as large institutional actors (with the exception of FNV) are missing.Dutch activists were generally very positive about the resources brought together in HA.In particular, in terms of expertise, there was agreement that the role of organisations such as TNI and SOMO was crucial, as these activist think tanks produce a steady stream of analysis that is the lifeblood of campaigns and inside lobbying (2,3,7,10).Various organisations also make their own analyses of EU trade policy, providing a multi-perspective framework.Again, the value of resource diversity was emphasised, pointing to organisations such as Milieudefensie, Greenpeace or De Goede Zaak, which are crucial to campaigning, and whose absence can greatly diminish influence on issues (2,3,7,9).
Two other aspects can be highlighted.First, in both cases the importance of motivated activists is stressed (2,3,7,9).Many organisations have only some (or even: one) person focused on EU trade, which means that if this person leaves, the commitment of a whole organisation can disappear.Second, the transnational aspect should not be overlooked.Both coalitions are linked to the transnational Seattle2Brussels coalition, albeit to different degrees.In the Netherlands, the HA coordinator and 3-4 other groups regularly attend meetings, where they talk about strategies, are informed about campaigns in other countries, and meet other activists.In Belgium, there is only one organisation that sends a representative.This could have a signal function towards the transnational level that more is possible in the Netherlands than in Belgium (1).In the early days of the Dutch coalition, there was therefore some funding from S2B to finance a HA coordinator, recognising the potential impact (2, 3).

Ideological and cultural overlap
The observations about the core driving group and the resources that can be mobilised also relate to overlaps in ideological and cultural terms.In Belgium, PRDH organisations share similar principles and values, but friction was mentioned repeatedly, putting a strain on coalition activities and membership, with collaboration perceived as difficult and time-consuming (1,4,5,6).This was already evident during the transformation to the general coalition.Several actors argued that there was a need to move away from the chaotic diversity during the Stop TTIP/CETA coalition, to better reflect the actual realities of power and representation (4,5,6).In other words, the voice of one activist group or individual should not be equated with that of an institutional organisation representing thousands of members.More local, activist, citizen movements or smaller NGOs were therefore de facto excluded from the PRDH.
Moreover, cooperation between NGOs and more institutional actors such as trade unions is not always perceived as smooth (4,6,11).In particular, when it comes to the implementation of Platform policy decisions, there is sometimes frustration with the way messages are presented to the public or to politicians, and it is claimed that carefully considered compromises are sometimes presented with the wrong or specific emphasis.In addition, it is not always clear to some whether an organisation is speaking on its own behalf or on behalf of the Platform.The result is an awkward middle ground where organisations sometimes publish collective and individual papers.
Given the strong role that the unions play in the PRDH, their ideological overlap with other organisations is an important element of the Platform's ability in defining positions or joint actions.For example, there are some ideological differences with farmers' organisations 5 or more radical NGOs (1,4,6,11).These mostly concern analyses of the social impact of trade agreements (1,4,6).Some unions do not oppose FTAs in principle, and rather focus on strengthening the enforceability of the Trade and Sustainable Development (TSD) chapters, where core International Labour Organisation standards are usually inserted.Other organisations (and unions) argue that trade agreements in their entirety kick away the ladder for developing countries, and negatively impact social conditions in third countries, and therefore take a more holistic view in challenging FTAs.
Similar considerations emerge when social and climate policy goals may conflict (1, 6).In Wallonia, for example, the radical left has a strong presence in politics, with strong concerns about the social impact of the climate transition.Trade union positions therefore tend to emphasise these social impacts more strongly.The issue of withdrawal from the Energy Charter Treaty is a case in point.Some Belgian unions were sceptical of the Dutch HA position, which in their eyes does not sufficiently reflect the priorities of the unions.The Belgian position paper is therefore more modest, calling for reform proposals and, if this leads to an exit from the ECT, considering 'the entirety of the consequences of such an exit, in particular in relation to supply and price of energy, and safeguarding and creating qualitative employment in these sectors'. 6 Such ideological and cultural frictions are hardly mentioned in the Dutch coalition.Instead, many of them mentioned the common focus on the underlying EU trade paradigm which in their eyes should be changed rather than focusing on a specific issue or priority (3,8,9,10).This is best reflected in the 'founding publication' of HA (presented in the research design), in which they present their diagnosis of the problem with EU trade and a well-developed alternative trade agenda that serves as a benchmark and a common starting point for joint positions.The organisations pointed out that there is always debate about the timing of actions, tone, language used, and that framing plays an important role in the final message (2, 10).However, this was never perceived as a barrier, as positions were fundamentally shared.In addition, many interviewees emphasised the friendly working relationships with other organisations (2,3,9,10).
The role of the FNV union here is quite different, as several actors said that it has a long history of working with NGOs and doing a more fundamental analysis of the impact of FTAs, in addition to their focus on labour rights (3,7,8).Some confirmed that this is related to the weakening position of unions, and the crises they have gone through in the past 10-15 years, which has led them to continuously collaborate with other organisations (3,8).On the other hand, the structure of the union has also changed with the introduction of a Members Parliament, through which the union rank and file (which may hold more radical positions) can exert pressure on the union leadership (3, 7, 8) (Tamminga, 2017).In general, a historical component can be discerned here, in that experiences in cooperation (whether at the interpersonal or institutional level) have a lasting impact on coalition dynamics.
All in all, there are differences in ideological and cultural overlaps between the organisations in Belgium and in the Netherlands.In Belgium, the strong position of trade unions in general and in the Platform means that only a limited number of their priorities are taken up, highlighting the sometimesdifficult working relationship with other organisations, and ideological divergences of where EU trade policy goes wrong (1,4,6).In the Netherlands, the focus is on an alternative trade paradigm, which makes focusing on other members' priorities less problematic and part of their own struggle.The FNV union, driven in part by a declining role in general, shares these more basic considerations and has a history of working with NGOs.

Leadership and structure
On the surface, both coalitions operate in similar ways, drawing inspiration from the Seattle2Brussels model: a division into subgroups based on policy issues, with not everyone obligated to be everywhere.But this similarity is perceived differently in the two contexts.In the Netherlands, this was seen as a good working model, providing space for those who wanted to be involved in a particular issue rather than everything.Differences were allowed, but were mostly related to tone and timing (when to organise a protest or write a letter).In Belgium, the same structure was generally perceived as involving a lot of work, long agendas, and many meetings (1,5,6).My own interpretation of this difference is that while participation in subgroups is not mandatory, the results of the subgroups must be reviewed by all parties, as this could become the coalition position.Given ideological disagreements, this leads to more work to keep track of everything.Moreover, this could explain why the structure is sometimes perceived as rigid, leading to different constellations with groups outside and beyond the PRDH, again in ad hoc coalitions, such as the 'Belgian Stop EU-Mercosur coalition'.
The biggest differences are in organisational structure and coordination.In the Netherlands, HA is essentially a stand-alone organisation with a semiindependent coordinator.The role of the coordinator is described by almost everyone as extremely important.The coordinator actively seeks for compromises, pushes people to come to meetings, represents HA in meetings with politicians, keeps an eye on the long-term strategy, follows up on international meetings, and basically spurs the organisation onwards (2,3,7,8,9,10).Funding for this position is now provided by the members, but was originally facilitated by S2B, recognising the potential influence in the Netherlands (see above).In addition, the coordinator is employed part-time and is affiliated to TNI, a progressive think tank that focuses on different areas of globalisation and global governance and which can be described as a 'multi-issue organization' that is able to bridge differences between organisations.Several people mentioned that TNI is not a classic membership-driven organisation that needs to work on its visibility or fundraising, so it can focus more on coalition goals than on its own constituency (3,9,10).
In Belgium, there is some confusion about the nature of the platform (1,6,11).Organisationally, it is not a separate organisation, but a working group convened by 11-CNCD.The Platform is somewhere between a decentralised and a centralised organisation: it has some structural working groups but they are mainly dependent on the time and resources of their members.Some interviewees stated that it is not really clear what the Platform is for: informal information exchange, or a springboard for joint activities (6,11).Moreover, the coordinator is not someone who speaks primarily for the Platform, but also for his or her own organisation, which brings its own difficulties.First, trade is not the highest priority for CNCD (which currently holds the coordinating role) which may impact the effectiveness of the role allocation.Second, there is an idiosyncratic peculiarity in the organisation: 11.11.11-CNCD is itself an umbrella organisation with members: many smaller ones, but also the unions themselves.Some activists argued that this puts the Platform (and coordinator) in a slightly awkward position, as it might be difficult to contradict demands from direct members.Nevertheless, discussion during interviews about whether a more independent coordinator would solve all problems was inconclusive (1,6,11).

Conclusion
The purpose of this paper was to explore the factors contributing to the divergence in coalition transformation in the EU trade policy field.To this end, I compared the perceptions of Belgian and Dutch trade coalition activists about the internal workings and fate of their coalitions.Although both have made the transformation from issue-specific to general EU trade policy coalition, they differ greatly in terms of visibility, ongoing (and expanded) activities, and membership dynamics.In my analysis, I draw on four aspects that have been theorised a role in the literature on coalition formation and longevity: perceived political and discursive opportunities, resource mobilisation, overlapping ideology and culture, and leadership and structure.Although the exploratory nature of this study does not allow for conclusions regarding necessary or sufficient variables, the qualitative analysis nevertheless provides a rich empirical description of coalition politics in the EU trade policy field that highlights the interconnection of these variables, rather than clear-cut delineations between them.
In Belgium, respondents indicated that there were few political and discursive opportunities to influence EU trade policy through the Belgian route, given the political-institutional (and media) differences between the Flemish and Walloon regions.This likely played a role in the decision of many organisations previously active in the TTIP/CETA campaigns to withdraw their involvement due to lack of time or interest.Moreover, the remaining organisations that form the driving force of the coalition have worked together in the past: trade unions and North-South (umbrella) organisations.Given the strength of the unions in this coalition and in Belgium in general, their priorities and ideological positions carry weight and are can therefore prove a hamper on collaboration if these positions differ from other groups.This is also reflected in the organisational structure, which occupies a confusing position between a decentralised and centralised coalition: some structural working groups, but organised as a working platform of one of the core organisations and mostly dependent on members' time, resources, and priorities.With the coordinating role tied to a specific organisation, time and priority constraints reappear here in addition to the peculiarity that various organisations (including the unions) are also members of the North-South umbrella.
In the Netherlands, there was considerably more optimism about the possibility of influencing EU trade policy via the Netherlands, primarily because of the fragmented and polarised political constellations.Here, not only left-wing parties, but also (strong) far-right parties oppose FTAs (albeit for very different reasons), opening up paths towards majorities if centrist groups can be convinced.Moreover, differences in the decision-making stages of FTAs (e.g., CETA) and highly mediatised ISDS cases have strengthened actors' engagement against EU trade policies.Internally, many groups remained active, with the large analytical capacity of activist think tanks such as TNI and SOMO found particularly helpful in providing the necessary impetus.The engaged groups also share fundamental analyses of the problematic nature of the current EU trade paradigm, providing a reference point for common struggle across single issues.The role of unions and farmers' organisations is significant in this regard, because they share similar analyses, have a different constituency and institutional role, and reach a large and more diverse audience.Finally, their joint efforts have also become more institutionalised with a semi-independent organisation and a coordinator linked to a multi-issue organisation that actively seeks and attempts to achieve goals and overlaps between members.
Four overarching conclusions can be drawn from this comparison between Belgium and the Netherlands.First, political opportunity clearly plays a role.This is not only mentioned from the bottom-up when deciding to engage further, but is also strategically assessed from the top-down through transnational networks.More attention and resources from transnational groups may flow to national coalitions if they are deemed able and capable of having an impact.Second, the combination of radical and reform-oriented groups is critical to resource diversity and success: combining expertise and mobilising power; grassroots enthusiasm and institutional representation; and outreach to diverse public and political audiences.Shared analyses of social and environmental consequences of EU trade policy is key in this regard.Third, there is a historical, path-dependent component, where past collaboration influences current coalition dynamics.In Belgium, there has long been a core group of organisations working on EU trade policy.While other organisations joined the cause in the issuespecific coalition against TTIP/CETA, the group reverted to its original membership after the importance of these FTAs diminished.In the Netherlands, all of the groups involved have a long history of cooperating, so the transformed coalition is an almost natural continuation of cooperative behaviour.
Despite the divergence described here, we should be wary of concluding that the Belgian situation is 'worse' than the Dutch.First, divergence is defined here in terms of visible activities and membership dynamics; developments that are clearly visible to outsiders.However, this ignores inside lobbying and internal channels of influence, which may be better adapted to a different context.In Belgium, some organisations indicated that their internal channels of influence were just as important to them (and may give them more influence than what is possible through mediatised campaigns).Second, in some cases 'ad hoc' coalitions (such as the Belgian Stop EU-Mercosur coalition) can be formed, and can be just as effective (in their context) as a permanent coalition.Third, the performance of the Dutch coalition is largely due to overlapping preferences between the unions and other organisations, in part because of the declining power of unions in the Netherlands.What is gained in cooperation opportunities, may therefore be associated with losses in autonomous union power.
The paper theoretically also speaks to the dynamics of differentiated politicisation, and the 'ebb-flow' movements of trade activism and mobilisation that has been identified in recent literature (De Bièvre & Poletti, 2020;Meunier & Czesana, 2019).This study shows that variation in mobilisation and politicisation patterns is linked to political opportunities, ideological overlaps between different groups, but also to cultural elements such as personalities, friendships, and trust.Two organisations may theoretically be inclined to work together, but personal relationships, working practices, or different career paths are also mentioned as playing an important role in sustaining coalition activity.This cultural component is an important addition to the more rationalistic interest group perspective that is often deployed in studies looking to explain variation (cf. De Bièvre & Poletti, 2020).In terms of further research, one important avenue is therefore the focus on the union-NGO cooperation in the EU trade policy field, from both an ideological and cultural perspective.

Table 1 .
Membership comparison between issue-specific and general coalition.Cursive: no official member, but ad hoc associated.-: no group members.x: groups present before, but left.