Political group formation in the European parliament: Negotiating democracy and gender

Political group formation in the European Parliament (EP) after European elections is a core feature of supranational party politics. The research objective of this article is to scrutinise democratic practices and the role of values, more specifically gender equality, in this political group formation. Complementing extant literature, this article engages with gendered political group formation as a dynamic process consisting of three intertwined layers, which are assessed by analysing formal and informal institutions around democracy and gender. The article draws on a unique data set of 130 interviews with MEPs, political group and parliamentary staff conducted in 2018–2020. The key findings of the article show that gender plays a role in each layer of political group formation, yet there are clear differences between political groups. By deciphering the reasons behind these differences, the article enhances the understanding of the political dynamics and struggles within the political groups.


Introduction
Political group formation in the European Parliament (EP) after European elections is a core feature of supranational party politics. With surveys showing that their leaders were unknown figures and their powers and roles are of little interest to media, politicians or citizens (Kelemen, 2020), the groups themselves were traditionally considered of little interest to the general public. This view has changed with the EP gaining considerable powers in EU integration. Furthermore, the rise of radical right populism has led to intense scrutiny and an acute need to understand the political groups as indeed unique EP power players and distinct from parliamentary parties and party groups in national parliaments (Brack, 2018;McDonnell and Werner, 2019).
The current seven political groups had no trouble in fulfilling the formal requirements for political group formation after the 2019 elections: 23 MEPs from 7 different member states and demonstrable 'political affinity' as laid out in the EP Rules of Procedure (2019). These formal requirements are the most well-known and cited characteristics of political group formation in the Parliament (Ripoll Servent, 2018). Political group formation has previously been explained by policy congruence (Hix and Lord 1997;Hix et al., 2007;McElroy and Benoit, 2010), pragmatism with access to resources and office positions as motives (Bressanelli, 2012(Bressanelli, , 2014 or by achieving respectability, especially in domestic politics (Bressanelli and Candia, 2018;McDonnell and Werner, 2019).
The research objective of this article is to scrutinise democratic practices and the role of values, more specifically gender equality, in this political group formation. Complementing extant literature, this article engages with gendered political group formation as a dynamic process consisting of three intertwined layers: (i) formal political group formation after EP elections; (ii) internal political group formation through core functions; and (iii) political group formation and consolidation through policies. Given the minimal formal EP rules on political group formation, we posit that political groups need to develop their own informal rules to decide which national party delegations become members and how to keep the group together by distributing leadership positions, forming joint policy positions and preventing group switching. These informal rules can be assessed regarding their democratic content including the role given to promote gender equality.
The key research questions are as follows: How are the formal and informal practices of political group formation gendered in the European Parliament? What does an analysis of gender equality contribute to understanding democratic practices around political group formation and the clustering of political groups? The article draws on a unique dataset of 130 interviews with MEPs, political group and parliamentary staff conducted in 2018-2020. As semistructured interviews lasting from twenty minutes to two hours this dataset is larger than any in qualitative analyses of EP's political groups. Moreover, it is representative of all political groups, member states and genders enabling a nuanced analysis of differences between political groups. Our data allows us to analyse how political groups engage with gender equality as a broader commitment beyond mere numbers including accounting for equality in informal rules of policy-making, (democratic) practices around leadership allocation or parliamentary organization such as speaking time or issue-related training.
The key findings show that gender plays a role in each of the three layers of political group formation, yet, there are clear differences between political groups. 1 By deciphering these, we enhance the understanding of the political dynamics and struggles within the political groups. First, the political groups can be grouped into three clusters in relation to how gender equality informs political group formation. Gender equality can be: (i) disregarded, (ii) form a benchmarking standard, or (iii) be applied pragmatically. Second, a focus on gender shows how the layers create pathdependencies between them. For example, disregarding gender equality as a normative goal in the first layer of political group formation means that gender equality is unlikely to be attended to in the other layers. At the same time, a commitment to gender equality measures may be consistent across most of the layers for some groups but then disappear showing the precarious position of the norm even for those highly committed to it.

Conditions for forming political groups in the European Parliament
The first direct elections of the EP in 1979 doubled the number of representatives and led to a formalization of political group structures within the parliament (Bartolini, 2005). The current parliament has seven political groups (see Table 1). Over the years, the names, compositions, rules and internal structures of the political groups have fluctuated. Extant literature has explained the different motivations behind political group formation. First, in policy congruence national parties join political groups whose policy stances are closest to theirs (Hix et al., 2007;Kreppel, 2002;McElroy and Benoit, 2010). Second, more pragmatic reasons, such as political group resources and committee positions, were highlighted as motives for joining political groups (Bressanelli, 2012;Whitaker and Lynch, 2014). Third, political group formation has also been shown to be driven by external reasons, such as incentives to increase group size. Larger groups receive more positions (and more powerful ones) in the EP (e.g. Presidency, Committee Chairs) and more resources, including increased plenary speaking time (Bressanelli, 2014: 779;McElroy and Benoit, 2007: 5). At the same time, group cohesion has been shown to remain important for policy influence of the main pro-EU groups and has to be ensured despite increased group size (Hix et al., 2007).
The last two legislatures have been interesting for political group formation as the surge in radical right populism led to the establishment of new political groups, including the ECR, Europe of Freedom and Direct Democracy (EFDD) and Europe of Nations and Freedom (ENF) in the 8th Parliament (2014-2019) and the ID in the 9th Parliament (2019-2024) (Brack, 2018). The different motivations behind the formations of these groups were shown to range from the pragmatic desire to use EP resources to efforts to appear more respectable to national electorate by joining more established conservative parties to attempts to form one big radical right populist group (see McDonnell and Werner, 2019;Whitaker and Lynch, 2014).
Questions about democracy and gender equality in political group formation have been less studied in previous literature. Both are basic values enshrined in Article 2 of the Treaty of the European Union. Liberal democracy is defined by free and fair elections; equal and inclusive citizenship; majority rule; respect of minorities; individual rights; rule of law; freedom of speech, assembly and the press; an active civil society; and constitutionalism in the form of 'separation of powers' and 'checks and balances' (Galston, 2018). As the only directly elected EU institution, the EP is central to the EU's democratic accountability. Open and transparent, participatory and inclusive political processes fulfil democratic principles and provide legitimacy to the EP and its political groups. Translated to political group formation processes, democracy means, for instance, that decisions on leadership appointment and policy positions are formalized and transparent and group members can participate. Alternatively, the democratic legitimacy of political groups could be diminished if decisions are far removed from its members and lack parliamentary scrutiny or are made in small circles involving deals and trade-offs.
Gender equality, in turn, constitutes a core struggle about norms and values of the Union. Without gender equality there can be no democracy. Yet, as a norm, gender equality has become increasingly politicized in the Parliament. Radical right populists actively oppose it as a harmful 'gender ideology' with gender equality advocates strongly upholding it (Kantola and Lombardo, 2021;Kuhar and Paternotte, 2017).
A focus on gender equality creates the need to understand a set of the factors that impact on political group formation. Gender equality comprises institutional norms and practices for political representation and an understanding of parliaments as workplaces (Erikson and Verge, 2022) and this also applies to how political groups are formed. Unquestionably, gender balance is another important aspect and the best researched one. Member state electoral systems, the use of gender quotas and the characteristics of political parties have been shown to impact the share of women in political groups (Aldrich, 2020;Lühiste and Kenny, 2016;Sundström and Stockemer, 2021). Political party ideology matters a great deal for women's political representation in the political groups within the EP (Sundström and Stockemer, 2021, 12). Women MEPs are more likely to come from left-leaning parties, which are more likely to employ gender quotas (Lühiste and Kenny, 2016). Chiva (2014), however, suggested that the left/right division does not hold for women's descriptive representation and Central and Eastern European countries. Building on these debates, Aldrich and Daniel (2020) found that quotas actually help promote women with prior political experience, thereby raising the number of politically experienced representatives at the European level (Kantola, 2022). Table 1 shows how the percentage of women ranges from over 50% in the Greens/EFA and over 40% in S&D, Renew and GUE/NGL, to over 30% in ID, EPP and ECR. Importantly, the percentage of women MEPs in all the political groups is over the threshold of 30%, which is a significant achievement in terms of gender equality and when compared to many national parliaments. Nevertheless, there are large differences between national parties and member states with some member states having comparatively few women MEPs (e.g. Bulgaria, Greece, Romania) or all-male national party delegations (e.g. Cyprus). In sum, whilst there are plenty of competent women available for top positions in the political groups, these women are not equally divided within national party delegations which may create challenges for gender equality norms.
The EP has been traditionally seen as a two-dimensional competitive arena structured along a socio-economic left versus right cleavage, as well as classifying political groups as either pro-or anti-EU integration (Hix et al., 2005;McElroy and Benoit, 2007: 16-17). More recently, scholars emphasised the dimension of GAL (Greens, Alternatives, Libertarians) versus TAN (Traditionalists, Authoritarians, Nationalists) as an important cleavage also potentially capturing the role of norms and values such as democracy and gender equality (Hooghe et al., 2002: 985; see also Brack, 2018: 56, 83). This raises the expectation that gender equality is a core value for political groups on the GAL side of the dimension and informs their decision-making and policy-making and potentially the political group formation too (namely S&D, Renew, Greens/EFA, Left). By contrast, gender equality is more controversial and contested on the TAN side of the dimension with more traditional and family centred notions dominating (for EPP, ECR, EFDD) with some groups challenging its legitimacy all together and speaking of a harmful EU 'gender ideology', which should not be spread into member states (ENF/ID) (Kantola, 2022;Kuhar and Paternotte, 2017). This means that gender equality intertwines deeply with political ideologies such as liberalism, conservatism, nationalism and social democracy of the political groups. Analysing democracy and gender equality in political group formation deepens the understanding of the role of values for political groups and of the distinctions between the groups along and beyond these axes.

Theoretical and methodological framework
Given the article's interest in how democracy and gender play out in different layers of political group formation, we contend that looking into formal and informal institutions is a key. The analytical distinction between formal and informal institutions has been developed in different variants of new institutionalism (Helmke and Levitsky, 2004). Formal institutions have been defined as codified rules with clear enforcement mechanisms and legitimacy. Informal institutions, in turn, signify customary elements, traditions, moral values, religious beliefs and norms of behaviour (Chappell and Waylen, 2013: 605). They are 'hidden and embedded in the everyday practices that are disguised as standard and taken-for-granted' (Chappell and Waylen, 2013: 605). The theoretical framework of Feminist Institutionalism (FI) has generated numerous important empirical findings about how gender shapes the work of political institutions (Bjarnegård and Kenny, 2015;Kenny and Verge, 2016). Feminist institutionalism highlights the significance of studying both formal and informal institutions and especially the interaction between them: the relationship between formal and informal institutions may be competitive or complementary on the basis that informal rules may subvert or reinforce formal ones (Waylen, 2014: 213). Such institutional analysis makes it possible to discern institutional histories too: the path-dependencies created both by formal and informal institutions. By shedding light on informal and formal democratic practices, we explore as key issues the role of the norms of equality, human rights and values in political group formation. Extant literature suggests that changes in group compositionwhether due to enlargement or other processesmake formal rules more important (Bressanelli, 2014: 780). This opens opportunities for gender equality rules too. A commitment to gender equality can be written into the statutes of political groups and thereby strengthened, for instance, on politics on representation (parity, leadership positions) or on tackling sexual harassment. New national party delegations can be asked to sign declarations about adhering to these political group values. Then again, gender equality may be compromised in political group formation revealing the larger tendency to sideline equality and human rights concerns for power gains. For instance, a desire to be big and powerful as a political group may lead into a situation where norms, such as gender equality and non-discrimination, become secondary to the desire to incorporate more MEPs and national party delegations to political groups. In a similar vein, Kelemen (2020) argues that the EPP has compromised its key democratic values by keeping the Hungarian radical right populist Fidesz as its member until 2021.
The research material used in this article allows for analysing such informal institutions of political group formation in a field often characterized by studying the formal rules. Our analysis relies on a large number of semistructured interviews with MEPs and staff (assistants, political group staff, EP staff) from all political groups and also non-attached MEPs. A total of 135 interviews following a semi-structured interview guide were conducted from 2018-2021. The representative sample of all political groups ensured a diversity of member states, genders, seniority and functions. Among the interviewees were eight (deputy) Secretary Generals (SG) of the political groups, who play a pivotal role in political group formation ensuring continuity between legislatures even when MEPs may change as a result of elections. All interviewees shared substantial details about dynamics around political group formation and organisational and political coherence. We also analysed political groups' statutes as primary documents regarding provisions for membership, equality and decision-making procedures.
Building on grounded theory, the interviews were coded using Atlas.ti. Grounded theory enables generating a meta-theoretical explanation which is grounded in data (Creswell, 2013) and provides a 'unified theoretical explanation' (Corbin and Strauss, 2008: 107) shaped by the views of interviewees. Codes were developed deductively as well as inductively; they cover both topical and theoretical codes. For this article, we analysed mainly the code 'PG formation', which was defined as 'How PGs are formed after elections; what are the roles of different national delegations, and SGs in the process, discussions of parties/individuals admitted later, how PGs might fall apart; including political group switching'. We supplemented the main code with the codes 'PG Organisation_statutes' (defined as mentions of political group statutes), 'leadership' (defined as leadership positions, selection procedures, career paths, leadership styles and leadership impact) and 'SG_role' (defined as the jobs and tasks for Secretary Generals). Each code was analysed along the political group lines allowing us to generate a 'thick description'. Our approach is particularly effective in drawing attention to democratic practices and informal institutions and the ways in which they are gendered.

Analysing political group formation through democracy and gender
Building on extant research and our theoretical framework , we discuss the three layers discerned from our research material: (1) formal political group formation after EP elections; (2) internal political group formation through core functions; (3) political group formation and consolidation through policies. Layer 1 comprises the formal steps needed to assemble a political group after EP elections and negotiations within established and newer political groups. Gender equality is either present as a core norm of European integration or it is side-lined. Both Layers 2 and 3 satisfy some goals of national delegations and thus secure their loyalty and continued association. Gender equality is relevantor notboth as a norm relating to the delegation of key positions or as a norm informing policy positions. We show how this focus allows making fine-grained distinctions between the political groups.
Layer 1: Gender equality norms in formal political group formation Forming a political group is a complex negotiation process for new political groups (examples include ECR, EFDD, ENF/ID, Renew) and less so for those which carry on their existence even if they engage in attracting new members too (such as EPP or S&D) (see Table 1). In relation to the first layer of formal political group formation, gender equality has three different places in political group formation for new and old groups.
First, gender equality norms become either unimportant or trumped by the concern for a big group size. We suggest that the conservative, right and radical right populist groups EPP, ECR, EFDD and ID are representative of the first strategy. Against some expectations that gender equality does not matter for conservative and right groups, the EPP and the EFDD political group statutes had provisions for democracy, human rights and rejecting discrimination. The EPP statutes mentioned: 'principles such as freedom and democracy, as well as the rule of law, respect for human rights and subsidiarity' as core values (EPP, 2013). The EFDD statutes stated that the group 'rejects xenophobia, anti-Semitism and any other form of discrimination' (EFDD, 2014). At the same time, extant research showed that political parties and MEPs who do not adhere to these principles were accepted to these groups (Kelemen, 2020). Outright opposition to women's rights as human rights was not uncommon from MEPs from these groups and neither was sexist or racist language (Ahrens et al., 2021). The ECR group statutes, in turn, are based on the so-called Prague declaration of the group, which states 'the importance of the family as the bedrock of the society' and contains no provisions on equality (ECR, 2017). It is evident that gender equality is a value not adhered to by these groups in political group formation.
Second, gender equality is part of the shared values which form a pre-condition for accepting new members in the political group formation process. This was especially the case with the Green/EFA group which appeared most principled in this regard. For negotiations with other parties or individual MEPs, principlesincluding gender equality, LGBTIQ rights and sexual and reproductive rightsmattered for the group (GreensEFA MEP F 130320). For example, according to our interview data a debate about values played a central role in the group's negotiations with the former EFDD member Italian Five Stars Movement (M5S) about them joining the group. One interviewee said that although many recognised the need to 'keep the group as big as possible to have the influence', there was a lot of resistance towards M5S from some delegations because of 'their reputation', especially in relation to refugees (GreensEFA MEP F 250220). Concerned MEPs expressed preference for M5S to join as individual MEPs, not as a party (GreensEFA MEP M 030320). This was eventually the outcome, when four M5S MEPs were to join the Greens/ EFA in 2020.
Finally, some groups mixed the first and second approaches: trying to gain seats but not at any price because of the reputation of the group and the parties involved. This was the case for S&D, GUE/NGL and ALDE/Renew. Gender equality becomes a negotiable norm that can sometimes be sidelined. A staff member from GUE/NGL involved in these processes described to us: After the elections it is a race between political groups because each group has as many votes as you have members. So of course each one of us tries to maximise the impact and influence. It doesn't mean that we would accept anybody into our group just to have numbers but of course we try to. After the last election, our group opened up to new parties and movements, so in that way it was a historical moment. (GUE-NGL Staff F 150519) The political groups that came together to form the Renew group were described as 'natural partners' by some interviewees, which was said to help to constitute a group very quickly (Renew MEP M 131219). Others were more cautious and pointed out the heterogeneity and the new dynamics at play as 'new people coming in' 'always changes the dynamic of any group' (Renew Staff M 050320). Some values are shared, others will be negotiated in the policy-making processes.
For established groups such as the S&D, the unity that has been achieved in this way is of central importance. The interviewees described how group cohesion and thus policy congruence on key issues is prioritised for policy-making purposes (S&D Staff M 290419), which is of great value in Layer 3 of political group formation. S&D, despite strongly committing to gender equality, also has single MEPs that oppose reproductive rights, the Istanbul Convention and EP sexual harassment policies (Berthet, 2022).

Layer 2: Informal gendered rules in internal political group formation through core functions
In the second layer, political group formation takes place through deciding on political leadership positions 2 which is important for forging loyalty of national delegations. It is an additional moment to assess the role of gender equality as a central facet of democracy in political group formation. Women's political representation and leadership positions are for many the most evident face of gender equality. In our analysis, it is central in relation to Layers 2 and 3 which however also reach beyond the numbers of women to look at broader issues such as informal gendered norms too.
The EPRoP provide no rules for political groups on how to distribute positions, except for rule 15: 'When electing the President, Vice-Presidents and Quaestors, account should be taken of the need to ensure an overall fair representation of political views, as well as gender and geographical balance'. Interestingly, we note that political group statutes often lack comparable provisions. Thus, negotiating political leadership positions rely on informal practices and leave space for political negotiations, and differences between political groups. The uneven distribution of women and men in leadership positions is a wellknown gender aspect of political parties and parliaments. The European Parliament is no exception: at the top, out of the ten available political group chair positions, only three were occupied by women (Kantola and Miller, 2022).
We contend that the second layer is characterized by informal rules. We find that the political groups are placed differently in relation to such rules depending on how long political groups had existed: older political groups such as EPP, S&D, Greens/EFA, and to some extent ALDE/Renew and ECR, had informally standardized but not formalized practices while new political groups had to decide quasi adhoc how to proceed. It is thus evident that the historical and institutional memory of the political groups played a role. From a gender perspective, in turn, informal rules are harder to pinpoint 'as they are naturalised as part of the status quo' hence appear as normal 'rules of the game' or 'ways that things are done around here' (Waylen, 2017: 5). Pinpointing the genderedness of such rules is even harder (Waylen, 2017: 5). For instance, in the case of informal rules of political group formation in Layer 2, our interviewees suggested that men are still favoured because of seniority, which is an important factor when key positions are delegated. Women as newcomers rather than incumbents have to first learn the unwritten rules of the game. As one interviewee put it: 'while women are still looking where the toilets are in the parliament, as they don't even know the building, the men already distributed the jobs' (S&D MEP F 060320).
An established practice for all political groups is to distribute leadership along national delegations and their size, although some deviate by offering prestigious positions to small delegations or individual MEPs to secure their membership. The strategy is common in all political groups, yet, differences occur along gendered representation resulting in clusters similar to Layer 1.
The first cluster, the conservative and radical right populist political spectrum (EPP, ECR, ID/ENF, EFDD) disregards gender equality in political leadership negotiations. Negotiating between and satisfying national delegations is the important criterion for these groups. Parity or setting numerical goals for representing women, in contrast, is deemed secondary at best. For example, the ECR recognizes the importance of leadership negotiations for political group formation, but only nationality and not gender is regarded as a hard criterion. One of our interviewees called the process 'difficult' and described how the Secretary General prepared negotiations between the delegations. He also shared an anecdote about a women MEP trying 'to get a place in the bureau on the basis that there weren't enough women' but this was interpreted as a selfish attempt to get power for herself (ECR Staff M 180319). Leadership negotiations are crucial in political group formation and large national delegations may prefer securing leadership positions over forming bigger political groups. The ECR double male leadership shared between the two biggest delegations was explained as a tool to keep the group together (ECR MEP M 191219). One interviewee explained that 'the best person will get the job' but did recognise that two male co-chairmen were a lost opportunity to make progress on gender equality. (ECR MEP M 051218) In Layer 2, GUE/NGL joins Greens/EFA in the second cluster, both representing political groups where political leadership positions are used to promote gender equality. Nevertheless, both groups were willing to recognise remaining gender inequalities and interviewees said there is 'always room for improvement' (GUE/NGL Staff M 210220; Greens/EFA Staff F 100320). Both groups have women Secretary Generals that also have a say in filling leadership positions (Kantola and Miller, 2022).
Especially for GUE-NGL the rules for gender balance were informal as the political group had no political group statutes. Even at this informal level, promoting gender equality overruled the rule of providing large national delegation leaders automatically with leadership positions, thus prioritizing gender parity. Generally, interviewees take parity issues seriously: 'I'm extremely disappointed when a delegation comes, they have two places and they come with two men. I think that's unbearable. (...) We don't have any statutes in our group at all. (...) So in that sense, every time we create the group, every legislature, let's say, we have to be careful (...) And it's very clear that gender is a political priority for a sufficient number of members, so that it's taken onboard when we form the group. (...) I think it would have been very difficult for a man to take the leadership of the whole group, alone. Whereas a woman could have had alone'. (GUE-NGL MEP F 160320) Moreover, several interviewees emphasised parity as a core issue and explained how the political group is actively searching to balance out potential under-representation in some national delegations by recruiting more women from others. This illustrates how commitment to gender balance can work as an informal rule and institutioninformality is not always problematic for gender equality (Waylen, 2017: 5).
The Greens/EFA prescribed gender equality formally in the group statutes as well as referring to it consistently in the interviews. The group statutes lay out precise election procedures for the Bureau's composition 'at least half of the Chairpersons and Vice-Chairpersons posts shall be held by women'. Moreover, an interviewee mentioned 'there is a kind of automatism in the daily life in the decision-making that the gender balance is there. (...) We have a lot of internal rules which are not in the statutes (...) where practically the question of gender balance is completely covered'. (Greens/ EFA Staff F 200319). The difference between the Left and the Greens/EFA group illustrates how gender equality can be achieved in this layer of political group formation through both informal and formal rules. Political commitment to equality is pivotal.
In the third cluster and comparable to Layer 1, ALDE/ Renew and S&D used gender as a flexible criterion.
Although the ALDE group foresaw in their statutes 'fair representation' this was meant for nationality and the two Europarties 3 involved and not for gender parity (ALDE Staff M 050419). For Renew, succeeding ALDE, gender was a pertinent question too, yet ultimately sidelined: 'Actually I think it's also a gender question. I know that some MEPs really say it should be a woman. Also the cochair thing was under discussion but we never did it and I think also this time it won't work. They like to have one group leader'. (ALDE Staff F 130619). Another one commented how the group formation is a slow process where the national delegations have to elect their leaders and the 'rest falls in place slowly, slowly, slowly' and emphasises the 'strategic thinking' between different top jobs (ALDE Staff M 050419). Eventually, Renew elected two male MEPs as president and first vice-president, plus nine vice-presidents, among them four women.
The interviewees also described how despite best intentions the political groups have to work with what the national delegations put forward as candidates; thereby, gender issues were subordinated as criterion: 'When they come after these elections there will be the first meeting of the heads of the lists, the delegation leaders of all the parties to be represented in the group, and supposing that by chance they're all men, I cannot change that. That's the result of national elections. I have to work with what I get. We have to work with what we get'. (ALDE Staff M 050419) 'Too many French guys. (...) It's a macho culture. I can just see there will be ten French male guys. (...) Of course it's good if they come and ALDE will be big, but I don't like it in terms of gender questions, I don't think it will be better. I just have in my mind, these male guys coming in'. (ALDE Staff F 301118) Both citations describe a real challenge and the political differences between national delegations in relation to gender equality that the political group staff must work with. Yet, as Table 1 illustrated all political groups have a representation of women over 30%, hence the pool of women is big. In such a context, claiming that there are not women shifts the responsibility for the lack of gender equality elsewhere: to other levels of decision-making and other cultures, which diverts attention away from how equality could be advanced within the rules that are under the political groups control, such as group statutes.
The S&D group's leadership is reserved for the biggest national delegation and thus switched from the German Udo Bullmann to the Spanish Iratxe Garciá Pérez after the 2019 elections. Other leadership positions were distributed to satisfy power interests of the different national party delegations. While S&D MEPs stated that their group simultaneously aims to satisfy gender parity and a regional factor, some interviewees suggested that this implicitly disadvantaged women, because 'the large delegations with a lot of people have men leading; the smaller the delegation gets, the more women are there' (S&D MEP F 060320; S&D MEP F 200220). Similar to Greens/EFA and GUE-NGL, the Secretary General was considered key in changing women's share in S&D top positions (S&D Staff F 060220; also S&D Staff F 180220).
With the obligatory change of leadership positions after 2.5 years in the European Parliament this second layer continues throughout the legislature. It can become a strategic choice for national delegations to wait for a second round, also in light of bringing women in leadership positions. One of our interviewees indeed described how the second half of the legislature is sometimes used to flexibly factor parity aspects in (ALDE Staff M 050419). While the first set-up for the political groups gives quasi-automatic preference to leaders of bigger delegations, the second half may provide more lee-way for equality considerations, at least for the political groups where gender aspects are flexibly applied as criterion for leadership selection.

Layer 3: Gendered consolidation of political group formation through policies
In Layer 3, political group formation progresses to substantial decisions for policy-making during the parliamentary term and a key issue is policy (as opposed to political) leadership. Much research exists on policy leadership positions including the selection of committee chairs, coordinators and committee members (Daniel and Thierse, 2018;Treib and Schlipphak, 2019) but a fine-grained analysis of different group practices is often lacking in these accounts (see however Chiru, 2020;Mamadouh and Raunio, 2003;Whitaker, 2011). The delicate process requires simultaneously satisfying interests of national delegations for policy fields and related positions andin those groups that strive for group cohesioncommitting them to a group line; gender equality plays a role in all of these processes. Our interviews shed light on how internal rules and selection procedures for policy position intertwine with gender aspects. The EPRoP only stipulate committee membership proportionality in line with group size and regulate cases if political groups leave seats unoccupied (rule 209). No gender parity rules exist for political group nominations at this parliamentary level. 4 Similar to Layer 2, we find that informal institutions in nominating MEPs dominate in Layer 3 and gender plays in differently for the political groups. Interestingly, at the beginning of the 2019 legislature, the 22 committee chairs were for the first time ever equally distributed between women and men (Table 1; Abels, 2020: 417-418). Committee coordinators also became more balanced over time, yet there were significant differences between the political groups ( Table 1). The EPPthe biggest grouphad only 18% women as coordinators (see also Kantola and Miller, 2022).
The selection of coordinators (Daniel and Thierse, 2018) as well as their important role in selecting rapporteurs have been researched (Obholzer et al., 2019;Ripoll-Servent, 2018;Yoshinaka et al., 2011). The formal rules around coordinator selection speak to different gendered norms. ALDE, S&D and EPP had statutory rules to achieve more equal gender representation. An ALDE statute made it the responsibility of the bureau to 'confirm the election of all the Coordinators in order to guarantee a fair representation' (ALDE, 2009: 11). The EPP and S&D mention women and sex composition explicitly but make only general references to parliamentary bodies and posts. The S&D uses a formulation of 'balanced representation' (S&D, 2014: 14.4). The EPP sets the goal at one-third women: 'The Chairman should ensure that, as the result of elections, the overall representation of members holding posts within the Group are composed of at least one third of members belonging to another sex than the majority of members' (EPP, 2013: 19.6). This has not been achieved among EPP coordinators, but there has been a clear improvement in committee chair positions as noted above. Despite the formal rules the practical means by which the groups select coordinators vary pointing to centrality of informal rules. For example, for the EFDD, interviewees described how people 'appeared' and 'emerged' for these positions (Kantola and Miller, 2022). For the EPP, in contrast, they were selected by elections.
Looking at the formation of shared policy positions, for ECR solving policy position related conflicts early on was a tool not to fall apart (ECR Staff M 200219). The radical right populist political groups ID, ENF and EFDD opposed this route and emphasised the importance of accepting differing national positions and open-voting, with no incentive increasing policy congruence. For the 8th Parliament, the EFDD interviewees said that they allowed 'a lot more freedom than maybe the other ones do', because of EFDD MEPs sharing the policy goal of 'changing this place and fighting structures within it' (EFDD MEP M 290119_2) rather than agreeing on shared policy positions. In the new ID group, reaching common policy positions happens at national delegation level rather than at the political group level (ID Staff M 110320). None of the interviewees of these groups mentioned gender equality as an aspect in their policy-making or the positions related to it.
Overall, political groups were divided in their approach to basing the formation of political groups on common policy positions, as noted also in extant literature on policy congruence (McElroy and Benoit, 2010) and pragmatic reasons (Bressanelli, 2012). GUE-NGL specifically described the significance of shared values: 'I mean we always had the social justice factor that unites us. (...) We have been the Nordic Green Left for, for 20 years but we've always been green and feminist'. (GUE-NGL MEP F 160320). Many Green/EFA interviewees also explained negotiations around the M5S joining the group as holding on shared policy issues. For S&D too, the unity achieved in this way is core: 'Our being united, being as a group as, we are not always united, but this is for us very important'. (S&D Staff M 290419). The citation illustrates how group cohesion on key issues is prioritised for policy-making purposes. Extant literature shows, however, that group cohesion on gender equality issues was quite low for political groups of the TAN-dimension (EPP, ECR, EFD, ENF) and high for groups of the GAL-dimension (S&D, Greens/EFA, GUE-NGL) with ALDE holding a medium position (Warasin et al., 2019).
In terms of democracy, negotiated and written policy positions are in many ways democratic: they are agreed upon by the group and the group can then be held accountable to those agreements. There was more flexibility on this in the populist radical right groups and no need to achieve such common policy lines despite the political affinity requirement. These political groups themselves called it 'openness' but it also meant behind the door negotiations each timewith not everyone participating. Despite political groups like Greens/EFA, GUE-NGL and S&D emphasising their commitment to equality issues, they have not introduced rules to secure attention to challenging gendered segregation in policy fields in Layer 3. The remaining political groups do not fare better and their commitment to equality issues remains even more open to debate.

Conclusion
Negotiations about democracy and gender can be understood by engaging with the dynamic model of political group formation. The analysis showed the role of formal and informal institutions in shaping how democratic practices interact with gender in the different layers. Overall, we find that there are surprisingly few formal rules to ensure transparency and equality when forming political groups.
In Layer 1, the EPRoP steer the formal conditions of political group formation, yet, they do not specify any rules regarding gender equality. Whether gender equality plays a role thus often depends on the norms and values (not) included in political group statutes. Already at this stage, negotiations and its gendered aspects happen at the informal level of unwritten norms and practices. In Layers 2 and 3, informal institutions, such as the size of the national delegation, guides which MEPs enter key leadership positions in the parliament, committees and political groups themselves. Gender parity is not formally regulated by the EP and securing parity during their formation depends on the political groups. Next to mere numbers of women and men in leadership positions, it is also important to consider their allocation in line or in contrast with traditional understandings of 'feminine' or 'masculine' policy fields (see also Dingler and Fortin-Rittberger, 2022). Overall, a focus on gender shows how the succeeding four layers create pathdependencies in both directions. Disregarding gender as a normative goal in Layer 1 leads to low attention in later layers, while committing to gender equality measures early on may lead to more consistency across most of the layers.
We can furthermore distinguish different clusters across the political groups. Gender equality can be: (i) disregarded, (ii) form a benchmarking standard, or (iii) be applied pragmatically. The first group comprises political groups originating from the TAN-dimension of political cleavages -EPP, ECR, ID/ENF and EFDDwho are united in their ambition to secure power and resources and where gender aspects are considered unimportant throughout all layers. The second cluster comprising Greens/EFA and in many cases also GUE-NGL explicitly commit to gender equality and use it as their benchmarking standard. Yet, even for this cluster gender equality measures diminish at some layers showing the precarious position of the norm even for those highly committed to it. The third cluster consisting of S&D and ALDE/Renew flexibly factor (informal) gender equality measures in and thus apply them pragmatically in the different layers, thereby falling between the two other clusters. Gender analysis of political group formation thereby nuances the GAL-TAN axis and the role of political ideology pointing to differences within the clusters.