In my youth the socialist Mayor and MP was accessible, he would be seen in the neighbourhood, come into my parents’ bar and had office hours. Now to make an appointment with an MP is like trying to get an audience with the Pope (In. 24).

The connections between political parties and civil society are central to parties’ performance of key representative functions (Mair 2013; Panebianco 1988). A party’s links to civil society helps it articulate and aggregate social interests, and the strength of a party’s connections to civil society organisations (CSO’s) has electoral payoffs (Martin et al. 2022a; Poguntke 2002). A revival of scholarly interest in the subject of party relationships with organized civil society has supported earlier findings (Bartolini and Mair 1990; Poguntke 2002) that considerable variation exists in the strength of a party’s connections. The expectation that party ideology and a party’s experience of government office are important explanations for this variation is articulated in several studies. Parties’ government experience is central to calculations made by parties and CSO’s of the mutual benefits of close ties (Allern 2010), while the connections of left-wing parties have endured longer and exhibited greater path dependency than those of right-wing parties (Gunther 2005; Rasmussen and Lindeboom 2013).

Less is known, however, about why parties with shared origins, ideology, goals, and government experience exhibit different levels of, and trends in, connections to CSO’s. Exploring explanations for divergence will add to our understanding of the factors that mediate the relationship between ideology and government experience on the one hand, and connections of parties to civil society on the other hand.

I address this question with a comparative case study of the two members of the socialist party family in Belgium—the francophone Parti Socialiste (PS) and the Flemish Socialistische Partij Anders (SP.a), which changed its name in 2020 and is currently known as Vooruit (Forward in English).Footnote 1 Party connections to civil society are measured by their connective density—a measure drawn from a novel dataset utilizing information from the Comparative Candidates Survey (CCS 2016 and 2019). The two parties had very similar levels of connective density at the 2007 general election (see Table 1).Footnote 2 However, the two parties exhibited divergent trends at the subsequent two elections. While the candidates of the PS sustained and even increased their levels of connection to CSO’s, the connections of SP.a candidates declined.Footnote 3

Table 1 Connective density of PS and SP.a candidates

This divergence in connectivity between two parties that share important characteristics constitutes the research puzzle. I explore this puzzle adopting an inductive approach, conducting a comparative case study, and using a most similar systems research design. The two parties represent a suitable subject for such a design sharing the same origins, membership of the socialist pillar (Van Haute and Wauters 2019), highly similar ideological positions, and 26 years of shared government experience in coalition governments.Footnote 4 They have severed all organisational ties allowing isolation and comparison of the role other factors play in determining the form and closeness of their ties to civil society (Deschouwer et al. 2017).

I deploy two main methods of data collection. First, semi-structured interviews with 28 respondents including current and former officials of both parties, country experts, commentators on Belgian politics, and representatives of CSO’s. Second, triangulation of interview material with a comprehensive review of literature and secondary documentation, and a process tracing analysis of key political events between 1991 and 2014.

I explain the divergence in the parties’ connections to civil society by exploring how the perspectives of party elites influenced their evaluation of the implications of far-reaching changes in the political, discursive and policy contexts of party competition in the Belgian regions. I investigate how decisions based on these perspectives had both direct and indirect impacts on the form and scope of their connections to organized civil society. In determining strategic responses to change, a party does not simply reflect external events but also refracts them through party actors’ evaluations of opportunities, their goals and preferences (Elster 1989). Specifically, party actors have agency over their positioning on issues of importance to voters (Adams, et al. 2004), over whether and in what form they maintain regular or routine contacts with CSO’s (Allern 2010), and whether they select candidates for party lists with strong connections to civil society (Cross 2008). A party also has agency over where and how to focus organizational resources, with consequences for its capacity to connect with and mobilize actors in civil society (Swyngedouw 2012).

I find that changes in the political, discursive and policy contexts in the last two decades presented the parties with challenges in the form of new party competition, the mobilization of discourses based on culture and national identity, and the fracturing of longstanding consensus in policy areas such as the welfare state. The two parties responded to these challenges in different ways, the PS retaining its socialist vision and policy of retaining close links to civil actors, and the SP.a withdrawing progressively from ties to institutions and communities that had connected them to civil society in the past.

This study makes an important contribution to the literature on the relationships between parties and organized civil society. It illustrates how the perspectives of party elites shape the decisions they make on policy positioning, the structure of party organization, and their ties to organized civil society in response to far-reaching change in political contexts, adding to our understanding of why similar parties exhibit different patterns of connection to CSOs. It also sheds new light on how party agency mediates the impact of party ideology and government experience on levels of connection.

Pathways to connectivity

Recent research has suggested that while formal contacts between parties and CSO’s may have declined, informal connections persist (Allern, et al. 2021a; Rasmussen and Lindeboom 2013), and that the connections to civil society of elites in parties with government experience are frequently higher than those with no government experience (Allern et al. 2021a, b). The endurance of connections established by established left-wing parties (Gunther 2005) and the focus of newer parties of the left on articulating neglected issues taken up initially by civil society (Della Porta, et al. 2017), suggest that denser connections to CSO’s may be a defining feature of left-wing parties. Recent empirical studies (Martin et al. 2022b; Lisi 2021) support this contention.

However, even amongst parties with similar ideological backgrounds and identical levels of government experience, connective density may vary (Martin et al. 2022b). I propose to explain the divergence in the extent to which parties are connected to civil society by exploring the role of party agency. Parties are known to respond differently to changes in the competitive environment, and especially to political, discursive and policy changes (Deegan-Krause and Enyedi 2010). This focus enables me to describe how parties assess challenges and opportunities and make decisions that impact on their connections to civil society.

Changes in context

Since the 1990s centre-left parties in Western Europe have been confronted by many similar challenges, in particular de-industrialization and the shrinkage of their traditional base in the industrial working-class (Rennwald 2020), pressures to reform the welfare state, the increased salience of the immigration issue, and the emergence of new party competition. Change, and centre-left parties’ strategic responses to it, have also differed substantially between countries. First, effective competition from the radical right emerged at very different times, between the early 1990’s and the late 2010’s, in Western European polities (Ignazi 2003; Rydgren 2018). Second, the adoption by centre-left parties of ‘left neo-liberalist’ policies such as welfare reform, privatization and de-regulation also occurred at very different times, starting earliest in southern Europe, and differed substantially in scale and scope (Mudge 2018). Third, and perhaps most importantly for the argument here, centre-left parties exhibited very different strategic responses to both competition from the radical right (Bale et al 2010), and to pressures for retrenchment of the welfare state (Schumacher 2012).

In Belgium, the case that I will study, three contextual changes are especially pertinent for the competitive environment faced by socialist parties. First, changes in the structure of party competition, with new challengers for a party’s traditional or core electorates prompting elites to re-evaluate connections and alliances with CSO’s that provide channels for the representation of voters’ policy preferences (Blings 2020). Second, changes in the political articulation of discourses (Koopmans and Statham 1999) comprising ideas about, and representations of, the role and legitimacy of civic and political actors. Third, changes in a high-profile policy domain critical to established relationships between parties and CSO’s, and previously characterized by stability and incremental policymaking (Kingdon 1984). Such changes undermine dominant policy paradigms and destabilise longstanding practices and routines of conflict resolution by political parties and civil society.

Party agency

A party has agency in how it chooses to respond to change. The choices of party elites about connecting to civil society are informed and influenced by their subjective beliefs about the nature and the consequences of these changes, and these beliefs will be informed by a party’s values, goals, and preferences (Elster 1989). I focus in particular on party choices in three key areas—ideological positioning, party organization, and the form of organizational ties to civil society.

First, a party has agency in its positioning on issues of importance to different groups of actors within civil society, and on how to respond politically to new party competition (Bale et al. 2010). A party may adjust its positioning in response to shifts in public opinion, especially when public opinion shifts way from a party’s established ideological position and is perceived as potentially ‘harmful’ (Adams et al. 2004). A party may also adjust positions in response to electoral results, shifting more when electoral losses are larger and the closer the policy adjustment in time to the previous election, (Somer-Topcu 2009), and in response to changes in the policy offers of competing parties (Adams and Somer-Topcu 2010).

Second, a party has agency over how it organizes in pursuit of its goals. The decisions of elites on questions of party organization, such as investment in local branch structures (Swyngedouw 2012), and initiatives to include actors from organized civil society on candidate lists presented at elections (Cross 2008), provide opportunities for different forms of interaction between a party and civil actors at many levels.

Third, a party has agency over the nature and form of any regular or routine contacts it maintains with CSO’s, and whether to sustain links to civil actors with whom they have traditional ties (Allern 2010). They make strategic decisions, for example, about the involvement of CSO’s in the development of party programmes and legislative proposals and output. And electoral incentives and party competition may prompt a party to develop new ‘participatory and environmental linkages’ with civil actors (Tsakakita and Lisi 2017), or to create and sponsor new social organizations, co-opting leaders and activists from CSO’s into public office (Verge 2012).

I assume that party actors are benefit-seeking, aiming to maximise office, policy or vote seeking goals (Müller and Strøm 1990). Party actors make purposeful choices between alternative courses of action but do so under conditions of uncertainty. They may have imperfect information on the distribution of voters’ policy preferences and face uncertain trade-offs when making strategic decisions. As a result, a party may mis-estimate political opportunities and make choices with both intended and unintended consequences for their connections to civil society.

Research design and case description

To explore the research puzzle, I adopt an inductive approach and conduct a comparative case study of the PS and the SP.a, deploying a most similar systems (MSSD) research design. An MSSD design is suitable for this case because of the strong similarity of the two parties on several substantive factors. The PS and SP.a share the same origins, formed in 1978 on the dissolution of the Belgian Socialist Party to create two separate parties competing for the support of Belgium’s main linguistic communities.Footnote 5 The parties are part of the socialist pillar—institutions comprising social insurers, health providers and networks of civic associations—that encapsulate Belgian citizens in ideational subcultures (Van Haute and Wauters 2019). The ideological positions of the PS and the SP.a have remained sufficiently close for them to be considered members of the same party family (Deschouwer et al. 2017). And the parties have 26 years of shared government experience as members of federal government coalitions. Over the last three decades, however, they have severed all organizational ties no longer consulting on strategies and programmes (Deschouwer et al. 2017) and diverged in connections to civil society. An MSSD design allows us to hold these shared factors constant while seeking explanations for the divergence (Anckar 2008).

Data collection

I collected data using two main methods. First, over several months I conducted semi-structured interviews with 28 respondents.Footnote 6 These included senior officials of Belgian left-wing parties, country experts, commentators on Belgian politics, and representatives of CSO’s. I identified interviewees initially by means of information in the literature on country experts and on the websites of the two political parties. Interviews with this group of respondents was supplemented by snowball sampling and through additional secondary research material. The interview transcripts ran to over 18,000 words, and I used initial and secondary coding methods to structure the analysis and group my arguments into theoretical labels. Summary details of the interview respondents are provided in Appendix 2, and a the full list of initial codes and their grouping into thematic categories is described in full in Appendix 3.

Interviews were conducted in English. The interviewees had a clear understanding of the core concept of a party’s connections with CSO’s and engaged with my own operationalization of the concept. This gives me reassurance that the concept of connective density is a meaningful one for political actors and not an artificial academic construct.

Second, I triangulated the interview material with a comprehensive review of literature and other secondary documentation, and by conducting a process tracing analysis of key political events between 1991 and 2014. The review included scholarly literature on both the Belgian case and on the relationships between political parties and civil society. It encompassed articles and commentary in newspapers and blog sites, party documents available on party websites and social media channels, and other political commentary. Original documents in Flemish and French were translated into English using online tools. Details of secondary documentation is provided in Appendix 3, and a summary process tracing analysis of key political events between 1991 and 2014 is provided in Appendix 5.

Analysis of the case

Drawing on data from semi-structured interviews and a process tracing analysis, I analyse how the perspectives of party elites on major change in political context, influenced key strategic decisions in ways that directly and indirectly influenced their connections to organized civil society. I start by analysing the response of both parties to critical changes in political context, and then consider the impact of decisions on party organization and the maintenance of ties to civil society.

Party competition

The two socialist parties were electorally successful until the early 2000’s, obtaining a combined vote share of nearly 28% at the federal election of 2003. However, since then both parties have lost substantial support, the SP.a vote declining by 55% between 2003 and 2019 and the PS vote declining by 27% in the same period, with the combined vote share of the parties falling to 16%. The structure of party competition faced by both parties has changed substantially over the last three decades. While both parties faced new and similar levels of competition from Belgium’s two regional Green parties—Groen (Flemish) and Ecolo (Francophone)—the most significant changes in the structure of party competition differed between the two regions.

The electoral breakthrough at the 1991 election—known as ‘Black Friday’—of the radical right Vlaams Belang (VB) represented a key turning point in Flemish party competition. In French-speaking areas, despite the Front National (FN) securing 5% of the Francophone electorate and two seats in the Belgian Parliament in 1995, competition to the PS from the radical right was short lived and unsuccessful, the FN failing to contest elections after 2010 (De Jonge 2021). From 2008, however, effective competition to the PS emerged from the left in the form of the radical left Workers Party, Partij van de Arbeid-Parti du Travail Belgique (PVDA-PTB). The party’s Renewal Congress in 2008 represented a critical juncture in the party’s political strategy, the party deciding ‘to work with trade unions and not in competition with them’ (In. 12), and ‘to support the demands of new social movements whereas they had previously been critical’ (In. 15). In the past decade the Flemish arm of the Belgian Workers Party, the PVDA, has also opened up new competition from the left to the SP.a. The PVDA’s strategy of strong association with protest activity, organizational penetration of unions and CSO’s, and activism in support of local communities, has attracted a lot of voters disaffected with the policy shifts of the SP.a. The party increased its share of the vote at Flemish regional elections to over 5%, and to over 8% at the last two municipal elections in the city of Antwerp.

In terms of political positioning the two parties’ response to new competition differed substantially. The nationalist appeal of VB, as well as the Nieuw-Vlaamse Alliantie (N-VA),Footnote 7 was perceived by senior figures within the SP.a as attractive to many of the party’s traditional working-class voters in the socialist trade union. A national figure in the SP.a observed, ‘In the 1990’s my husband was in a football team of SP.a supporting workers. Now almost all of them vote VB, they lead difficult lives and feel left behind’ (In. 19). This perception and its implications for positioning was summarized by a former senior figure in the SP.a, ‘On economic issues we were competing with VB for our historical support in the working-class, for union members motivated by racism and migration’ (In. 27). Officials of the SP.a were dismissive of the PVDA because of its opposition to joining government and ‘disowned the administrative coalition between the parties in Zelzate’ (In. 17).Footnote 8

While Flemish nationalists mobilized outside the socialist pillar, the PTB sought directly to infiltrate PS strongholds within civil society. ‘The PTB has complicated things—we notice their presence more in the unions and civil society but especially ‘on the field’ with people in every structure’ (In. 8). For one commentator, the growing strength of PTB in elections for union representatives within companies, ‘presents the PS with a problem because they have limited electoral margins to play with, it becomes increasingly hard to convince defectors to return’ (In. 15). The success of the PTB’s strategy of penetration caused increasing tensions within the socialist union the FGTB, and between the FGTB and the PS, culminating in 2014 in ‘an open letter from the national federation to the PS arguing against their position and urging a leftward shift. The union said that it would invite its members to only support PS candidates who backed union demands and not just because they were on the party list’ (In. 17).

The response of the PS to new party competition was to maintain a strongly left-wing ideological position and close and regular contacts with trade unions, and to seek to bridge the gap between party elites and civil society by trying to ‘build closer links and trying to support alternatives proposed by NGO’s, e.g. the CETA campaign in which PS was very strong’ (In. 18).Footnote 9

In summary, while both parties faced challenges from new party competitors they drew different conclusions about how to respond. The Sp.a moved to accommodate competition from nationalist parties while the PS retained a strong left-wing policy orientation.

Discursive context

The discursive context of party competition during the period of this study was transformed by the articulation of discourses on two issues that introduced tension to relationships between the parties and their civil society allies.

First, the activation of negative discourses on migration. Research over the period 1991–2014 (Billiet et al. 2017) found that levels of anti-migrant sentiment differed very little between voters in Francophone and Flemish regions, and that there had been no unequivocal increase in negative attitudes towards migration over this period (2017, p. 3). From the mid 1980’s parties in both regions attempted to articulate and activate negative discourses at the level of party competition. In Flanders first the VB, and then later the N-VA, sought to frame migration as a threat to the Flemish culture and way of life (De Winter 2004). And in French-speaking Belgium the Front National (FN) mobilized nationalist and xenophobic sentiment (Ignazi 2003).

At first the SP.a refused to engage with the migration issue, ‘Stevaerts [party leader between 2003 and 2005] said “let’s not talk about it”’ (In. 20). Electoral setbacks from 2007, when the SP.a lost 30% of its electorate, marked a turning point in the party’s position. A key feature of the SP.a’s response was the adoption of the VB’s ‘framing of migration as a security issue’ (In. 15), and the adoption of policies that targeted migrant communities in key municipalities under its control (In. 26). Under the Mayorship in Antwerp of former leader Patrick Janssens, which started in 2003, the SP.a introduced a series of measures including a ban on the wearing of headscarves by municipal employees. The SP.a’s urban policies between 2006 and 2012 were perceived as discriminatory by Muslims throughout Flanders (Gaasendam 2021), and were a ‘major reason for [the party’s] weak links with CSO’s in the migration field’ (In. 9). For one commentator, the party’s ambiguity on multiculturalism was, ‘clearly a mistake—they gave up a natural connection to ethnic minority voters assuming that they would never leave them’ (In. 25).Footnote 10

If the SP.a were initially hesitant and subsequently accommodating of discourses hostile to migration, the PS ‘combatted the extreme right from the outset, including actions to try and make the playing field for it as small as possible’ (In. 20). At the same time that Patrick Janssens became Mayor of Antwerp, the PS leadership resisted pressure from within their own ranks for ‘an ethnocentric turn [to] stem losses amongst the native working-class. The PS elite never accepted this argument… In big cities the relative strength of PS is a reflection of their links to new multi-cultural associations which they have regarded as compensation for losses in the traditional pillars’ (In. 15).

A second discursive shift, from the late 1990s onwards, was constituted by a strong anti-union discourse framing trade unions as ‘a relic of a bygone era, opposed to the changes needed to keep Belgium competitive’ (In .9), and articulated by parties in the Flemish government and employers’ associations (Zienkowski and De Cleen 2017). This discourse had a major impact on the ‘influence and high position trade unions enjoyed in civil society’ (In. 7).

A wary and dismissive attitude on behalf of the SP.a towards trade unions was a consistent theme in many interviews. The ‘unpopularity of trade union actions with mainstream publics’ in Flanders (In. 6), were held to be responsible for the unpopularity of ‘brand Socialist’ (In. 11), with trade unionists less present on party lists at the local level because, ‘they don’t get high scores in preference voting’ (In. 10). This negative perception of trade unions is reflected in the rationale party figures provide for infrequent contacts. ‘The trade union brand is less and less helpful, we don’t have to engage with them much apart from say on May 1st—we have to keep up some traditions’ (In. 11.), and ‘you don’t vote for us anymore so why should we listen to you’ (In. 6) were representative of this attitude.

Acknowledging the tensions between themselves and the unions, the PS, in contrast, regards constant dialogue as essential. ‘The best way to maintain good relations is to talk as often as possible’ (In. 8). The party maintained levels of representation of trade unionists on its lists of election candidates,Footnote 11 and, in response to the ultimatum from the FGTB in 2014, the party’s candidates moved sharply to the left.Footnote 12

In summary, the activation of discourses hostile to migration and trade unions, presented fresh challenges to both parties. The PS chose to retain a multicultural discourse and close links to trade unions, while the SP.a distanced itself from trade union allies and adopted policies perceived as discriminatory by migrant voters in Flanders.

Shifts in a key policy domain

In the area of the policy-domain context I highlight reforms of Belgium’s system of social protections. Reforms in two waves between 2001–2005 and 2011–2014 were led by socialist ministers from both the SP.a and PS, had adverse impacts on a diverse set of constituencies represented by CSO’s, and were strongly resisted by shifting coalitions of CSO’s that had traditionally maintained close ties to the two parties (Faniel 2012).

A key figure in the positioning of the SP.a on welfare reform was former party Chairman Frank Vandenbroucke, who returned to Belgian politics in 1999 after a three-year academic sabbatical. The intellectual figure behind the party’s adoption of ‘the ‘activation’ concept of social welfare—focusing on the responsibility of the recipients of social welfare’ (In. 15)—Vandenbroucke ensured that the SP.a was closely associated by civil actors with the promotion of social welfare reform (Faniel 2012). The party held the ministerial positions responsible for the implementation of reforms, and conflict between the party and civil actors resulted in increasing distance between them. As one commentator observed, ‘CSO’s became reluctant to associate with the party because of its low reputation and an anti-politics sentiment in the context of its retreat from the left’ (In. 6).

While the PS was a member of coalition governments that introduced reforms of the welfare state, it might best be described as a silent partner in the activation project. The reforms were deeply unpopular in Wallonia, ‘because in places like Charleroi there was high unemployment but no jobs, making benefits linked to responsibility seen there as an attack on the unemployed’ (In. 20; also In. 7). ‘As a result the PS tended to hide a third way stance from visibility, retaining a left-wing rhetoric’ (In. 13). The strength of grassroots representation on key party committees meant that the PS never formally adopted a third way position. ‘In 2003 there was a strong revolt against a third way positioning, and the party continued with economic polarization at the sub-federal level’ (In. 1).

In summary, both parties were members of governments that adopted ‘third way’ policies on welfare benefits, but the SP.a were more prominent as originators of the policy shifts while the PS sought to distance itself from the welfare reform agenda.

Party organization

The impact of changes in party positioning on connections to civil society were amplified by the parties’ decisions on organizational structure. Organizational choices by both parties have left them with very different capacities to engage with voters on the ground. For the SP.a, without strong local branches, campaigns rely heavily on ‘a central team and not the membership of the party’ (In. 20). The evisceration of the party’s organizational ground force is seen as instrumental in its failure to respond to heavy ground-based election campaigns by nationalist parties (Swyngedouw 2012). Since 2010, by contrast, the PS has reversed a downward trend in its membership providing it with numerous points of contact with local civil actors, and a potent tool for election campaigning. ‘Our membership still constitutes a base that can be mobilized—in Liege alone we have 10,000 members which is more than Ecolo have for the whole of Wallonia’ (In. 15). Party infrastructure remains large and well-resourced. As one senior PS official suggested, ‘We have a lot of staff (around 700 at all levels) and power, and are perhaps like a medium sized business. People are always coming and going between the party and influential organizations’ (In. 2).

In the last two decades the PS and SP.a made very different choices about how to structure party organization. The Sp.a recently ended the right of pillar associations to guaranteed places on it main decision-making board, and several interviewees highlighted how the SP.a has withdrawn resources from its branch structure to centralise its organization and focus on a ‘marketing led’ approach to campaigning (In. 6). Successive party leaders saw its membership base as an obstacle to progress, ‘As long as you have to use the formal democratic systems of parties you only approve those things that are killing you as party’ (In. 27). The party ‘forgot its whole network of local contacts’ (In. 21), ‘that gave important signals to the leadership about the concerns of people on the ground’ (In. 20). For the party’s connections with Flemish citizens the result was that ‘Grassroots links are no longer there’ (In. 20). As one senior SP.a official conceded, ‘It is hard to say what grassroots CSO’s the party is involved with’ (In. 11).

In organizational terms the PS’ response to changes in its political context was one of continuity rather than change. The party ‘retained relatively active local sections’ (In. 17), and ‘kept organizing local meetings to disseminate an ideology’ (In. 7). The party maintained its commitment to an organizing principle of proximité, ‘of links between its elected representatives and the communities they serve’ (In. 20). The PS has retained automatic representation on the party board for the socialist union, the FGTB, and discussions with different ideological currents within the party and with CSO’s are seen as a core element of proximité.

The PS and SP.a have also taken contrasting approaches to recruiting election candidates from civil society. When selecting candidates, the PS has stayed close to, and sought to reward, the associations that comprise the extended socialist pillar. ‘I [PS senior official] find it hard to think of a candidate that had emerged from an independent CSO rather than one of the party’s many socialist societies who we already know is a member of the party’ (In. 8). In contrast the SP.a has consciously sought distance from traditional allies in civil society that it regards as unpopular. In the early 1990s the party centralized candidate selection decisions ‘because it perceived that unions and social insurers had too much influence in candidate lists’ (In. 22). In place of candidates from larger CSO’s, as one national official of the SP.a confided, ‘The different mindset of the twenty-first century … will be reflected in younger candidates that are not from local branches but smaller CSO’s, influencers and even celebrities’ (In. 10).

In summary, the approaches of the two parties to internal organization diverged sharply. The SP.a centralizing control over party decisions at the expense of maintaining local structures, and the PS retaining comparatively strong local party structures and influence for party branches and pillar associations over party decisions.

Decisions on ties to civil society

The two parties also made very different choices about the nature and form of regular contacts with allies in civil society. Despite the enduring strength of trade unions in Flanders, the perceptions of SP.a elites of union unpopularity led the party to, ‘Intentionally move away from the institutions with which it had been associated traditionally’ (In. 25). The exit of the SP.a from the federal government, following the election of 2007, was a shock to the party causing it to double-down on the primacy of its office seeking goals. ‘The SP.a have clearly stated that they want to be a party who decides, for whom opposition is ‘not for us’’ (In. 17). The preoccupation with electoral goals influenced the SP.a’s strongly instrumental approach to links with civil society, ‘If our analysis that the time of membership parties is gone is correct then it would not have helped to retain connections… because it was not bringing in voters anymore’ (In. 27). The SP.a replaced enduring connections to membership based CSO’s with ad-hoc contacts with ‘local volunteers around issue based politics’ (In. 14).

By contrast a close dialogue between the party and CSO’s is important to key party decisions of the PS. ‘When the party is working on its programme before elections, we work with CSO’s in every field… Similarly, when we are in coalition discussions, we meet with CSO’s to consult and inform’ (In. 8). This strategy of drawing CSO’s into important discussions demonstrates the high value that elites in the PS place on maintaining their connections to organized civil society.

Throughout their history both parties have been office-seeking, placing a premium on being able to achieve something for their constituencies by sharing office in multi-party coalitions with parties of the centre-right. ‘We [the PS] pass bills and this is what people remember—we communicate what passes in parliament and how this affects people’s lives’ (In. 4). The two parties have however pursued office goals in different ways. One commentator observed that, ‘The PS knows it needs to also maximise votes or the other parties will form a government without them … In regional government elections however the SP.a has not campaigned so hard in order to remain on good speaking terms with centre and right parties, and so it can enter the regional government’ (In. 20).

In summary, while both parties are office seeking they reached different interpretations about the implications of this goal for their ties to civil society. The PS opted to retain a close dialogue with CSO’s, while the SP.a concluded that close ties did not serve its electoral goals well.


The PS and the SP.a are two office-seeking parties of the left in Belgium who share the same origins, ideology and experience of government, but who have diverged in connections between party elites and organized civil society. To explain this divergence, I conducted a comparative case study of the two parties using a most similar systems design. I find that the perspectives the two parties had on new challenges presented by changes in their political environments led them to make different strategic choices with both direct and indirect impacts on their connection to civil society.

Both parties faced new political competition from opposite ends of the political spectrum, the articulation of new discourses hostile to civil groups with which the parties had enjoyed close relationships, and the fracturing of Belgium’s longstanding policy consensus on social protections.

The SP.a responded to these pressures with a series of choices on party positioning and organization that led to the progressive withdrawal of party elites from the institutions and communities that provided its electoral base. In contrast the PS chose to confront rather than accommodate radical right discourses, retained its socialist vision, and was compelled by new competition from the radical left to draw closer to traditional allies in civil society.

The ‘dilemmas of inclusion’ (Dancygier 2017)—the challenge of balancing appeals to traditional groups of supporters while connecting with and mobilising new voters in growing migrant communities—are illustrated sharply in my case study. The perception of the SP.a leadership of the challenges this posed to its traditional support base meant that the party did not, as some theory expects (Tavits and Potter 2015), increase the emphasis it placed on socio-economic issues. Instead the SP.a pivoted towards the values dimension of party competition introduced by nationalist parties. This pivot, and frequent changes of party leadership, contributed to further electoral decline, the party failing to win back traditional supporters while alienating groups of both liberal-multicultural and migrant voters (Gaasendam et al. 2021).

My study makes important contributions to the literature on how the perspectives of party elites shape the decisions they make on policy positioning, organization, and their ties to organized civil society in response to far-reaching change in political contexts. It adds to our understanding of why similar parties exhibit different patterns of connection to CSOs. Taking a comparative approach to considering how party-based actors refracted far-reaching changes in the context of political competition enables us to bring a number of interesting questions into focus.

First, we can identify the cumulative and compounding effects of strategic decisions by party elites on their connections to civil society. For the SP.a, the impact of the deliberate decision of elites to distance the party from traditional allies in the socialist pillar was compounded by policy shifts in welfare reform and migration that alienated previously loyal groups of civil actor. The PS, by contrast, chose a different course, purposively maintaining a multicultural discourse and clear ideological position, enabling close links to both traditional groups within the socialist pillar and representative organisations of an increasingly diverse civil society.

Second, we can observe how the two parties different approaches to party organization resulted in different levels of decision-making autonomy for party elites. From the mid-1990s the SP.a centralized control over issues of policy, party campaigning and candidate selection, while neglecting its branch structure, enabling elites to set the parties’ strategic direction without meaningful constraint from party members or pillar institutions. The PS by contrast retained a comparatively robust local party infrastructure and the strong influence of ordinary members on the party’s policy committees, as well as concerted pressure from local branches of the socialist trade union, constrained the party from making policy shifts opposed by civil society.

And third, we can observe how strategic miscalculation and the parties’ differing levels of connection to civil society impacted their capacity to respond appropriately to competitive threats. For example, the SP.a policy shifts on migration harmed its relationship with migrant communities that make up a sixth of the Flemish electorate, and the party overlooked evidence that in the 2000’s its supporters were more likely to defect to Green and centre parties than to nationalist competitors (Dassonneville and Baudewyns 2014). The party’s organisational changes also weakened the signals it was receiving about the challenge posed by the PVDA in key municipalities, the SP.a, unlike the PS, failing to draw the implication that it should re-engage with CSO’s and their claims in order to reverse defections to the PVDA. While the PS has also experienced some weakening of its local organization in recent decades, its comparatively strong links to pillar institutions and responsiveness to campaigns of newer social movements, has helped it limit electoral losses to new competitors.

My study also illustrates how parties’ responses to changing competitive contexts mediate the relationship that two left-wing parties with government experience have with organized civil society. Further, it suggests that goal-based accounts of parties’ relationships with civil society may have serious limitations when explaining variation in political outcomes. Neither the PS nor the SP.a are ambiguous about the primacy of their office-seeking goals, requiring sufficient electoral support to maintain their bargaining strength in post-election coalition discussions, and both responded to the threat to their electoral strength that new challenger parties represented. However, the choices of party elites on positioning and organization took them in opposite directions—the former maintaining proximité to civil society and the latter withdrawing from close ties.

My findings suggest a number of fruitful avenues for further research. First, while our case is suitable for a most similar systems design a number of features of the Belgian context such as federal structures of governance and the persistence of consociational practices may limit the transferability of my findings. Future comparative research could explore whether the impact and interaction of a changing competitive context and party agency have similar consequences for party-societal connections in countries characterized by a unitary state and pluralist structures of interest representation.

Second, the structure of civil society in West European societies has itself undergone transformation in recent decades with the emergence of new organizations in the fields of the environment, identity politics, and migration. This diversification of organized civil society presents new challenges, particularly to mainstream parties of the left, who want to retain their traditional ties but also engage with new constituencies. And the growth of digital forms of campaigning and mobilization has provided political parties with new opportunities to communicate directly with citizens rather than through intermediaries in civil society, adding further complexity to party strategies and the relationship between parties and CSO’s. Studying the different strategic responses of social democratic parties to these challenges will help scholars understand the shape and efficacy of the connected left-wing party of the 21st Century.

Third, the measure of connections between party elites and organized civil society—connective density—is restricted to CSO’s that are central to political mobilization (Martin et al. 2022a). The Belgian case suggests however that in societies with traditions of citizen encapsulation into pillars (Van Haute and Wauters 2019), connections to organizations in the cultural field may continue to play a part in the selection of party candidates and in linking those candidates to communities at the local level. Further research could explore the contexts and conditions under which a more extended notion of connective density is relevant for different political and electoral outcomes.

In the autumn of 2021, the SP.a moved into a shared office building with the PS. My comparative case study provides much material to suggest, however, that the PS and SP.a, both under new leadership and, in the case of the SP.a a new party name and identity, are set to continue down different pathways of connection to organized civil society. While the PS has increased its openness to, and dialogue with, the broad field of civil society, the SP.a has further centralized its organization by closing its regional offices and removed the right of the socialist trade union, the ABVV to guaranteed representation on the party board, opting also for an issue based and social media led strategy of connecting directly to citizens. The effectiveness of these two contrasting approaches to relationships with organized civil society seems likely to play an important part in the future political and electoral fortunes of these two members of Belgium’s socialist party family.