Literature review

The literature commonly identifies political parties and interest groups as formally organized groups ‘aggregate(ing) individual interests and preferences into collective demands and seek(ing) to influence the form and content of public policy’ (Allern and Bale, Reference Allern and Bale2012; 9–10). They serve as distinct channels of representation but cooperate in the electoral, institutional, and societal arenas in an attempt to influence public policy. However, although parties compete in elections to gain formal political/governmental power and are usually oriented towards a broader range of policy fields, interest organizations do not contest elections and focus on a narrower range of causes/issues/policy fields (Thomas, Reference Thomas2001; Bolleyer, Reference Bolleyer2018).

As several studies have pointed out, there are many conceptualizations of party–group linkages (Römmele et al., Reference Römmele, Farrell and Ignazi2005; Mavrogordatos, Reference Mavrogordatos2009; Allern, Reference Allern2010; Allern and Bale, Reference Allern and Bale2012). Perhaps the most widely used definition of party–group relations comes from Allern (Reference Allern2010), who understands this notion in terms of personal/institutional contacts; it ‘consists of links that connect interest organizations to the party's members, decision-makers and/or decision-making bodies, i.e. links that open up for contact and potentially provide communication about information, know-how, opinions and policy views between parties and interest organizations. Thus, links are those means by which a party and an interest group may communicate, such as, for example, corporate membership, joint committees, leadership overlaps or regular elite contact’ (Allern, Reference Allern2010: 57).

The empirical evidence on party–group linkages is mostly based on case studies and it is characterized by its fragmentation and disconnection (Eising and Cini, Reference Eising, Cini, Heywood, Jones and Rhodes1997; Beyers et al., Reference Beyers, Eising and Maloney2008). Generally speaking, there is broad consensus in the literature that the strong links established between parties and interest groups during the ‘golden age’ of mass parties – the most notable case being the formal affiliation of trade unions and social democratic parties (e.g. Moschonas, Reference Moschonas2001; De Waele et al., Reference De Waele, Escalona and Vieira2013) – have gradually been replaced by looser forms of relations, with no organizational links (Katz and Mair, Reference Katz and Mair1995, Reference Katz and Mair2018; Allern, Reference Allern2010). However, comparative studies have questioned the fact this is a general and linear trend. Thomas (Reference Thomas2001: 271), for example, found no clear patterns in terms of party–group relations in advanced and newer democracies in different regions of the world. Although it is clear that party–group ties are loosening in countries that have had strong left-wing governing parties and historically strong party–group ties (as in Northern Europe, Britain, Spain, and Israel), other countries are characterized by different trends (e.g. Yishai, Reference Yishai2001; Allern, Reference Allern2010; Mach, Reference Mach2015). Although the most common pattern is indeed one of a fairly distant and fragmented relationship, there is substantial diversity, and some parties have quite strong links to interest groups.

In a more methodologically coherent and comprehensive study, Poguntke (Reference Poguntke, Katz and Crotty2006) analyzed party statutes for 11 European countries between 1960 and 1989 to capture the substantive difference and strength of two forms of the organizational access to parties held by collateral organizations: representation in party bodies, and holding the right to send delegates (the latter being considered a stronger linkage). The results show different trends for linkages to external and internal organizations. Although external collateral organizations have become less relevant to political parties in Western Europe, internal organizations have increased at roughly the same rate. The overall picture, then, is one of the declining linkages to external interests offset by the increasing representation of internal interest organizations over the past few decades. Although the author claims that traditional political parties have managed to maintain relatively stable and close ties with their collateral organizations, ‘their capacity to “deliver” votes and interest aggregation has suffered due to ongoing processes of social differentiation” (Poguntke, Reference Poguntke, Katz and Crotty2006: 403). New parties, on the other hand, are concentrated on the core political organization but pay less attention to creating or strengthening peripheral organizations or structures, which results in ‘an increasingly pluralistic system of interaction between organized interests and party politics’ (Poguntke, Reference Poguntke, Katz and Crotty2006: 403).

A recent study based on statutory connections confirms that party–group ties are relatively weak and that there is variation both at the country-level and the party-level (Allern and Verge, Reference Allern, Verge, Scarrow, Webb and Poguntke2017). Based on the comparison between 19 countries (and more than 100 political parties), these authors found that very few parties have formal (statutory) linkages with interest groups (non-party organizations), and that trade unions tend to display a stronger representation – through representation rights in party decision-making bodies – than business groups. On the other hand, parties are more likely to use party sub-organizations, especially women and youth structures, as a linkage mechanism to connect with social groups.

Fraussen and Halpin (Reference Fraussen and Halpin2018) examine political parties and interests as two types of political organizations that share commonalities in terms of internal dynamics (e.g. increasing professionalization and centralization) and their development. Therefore, they argue that both face common challenges and that present a complex relationship because they are inextricably bound. This approach is also important for assessing configurations of party–group relations in a comparative fashion, and to assess the way external factors – for example, party system features and type of policy-making process – interact with organizational features of these political actors.

Extant research shows that no single factor or model is able to account for the variety of party–group links, either within or across countries (Thomas, Reference Thomas2001; Allern and Bale, Reference Allern and Bale2012). Yet, a number of studies, mostly based on Western European countries, have tried to identify the main determinants of party–group connections. Firstly, we need to consider how institutional features can affect the ties between political parties and interest organizations. From this standpoint, the degree of pluralism of the system of interest intermediation may influence these ties. According to extant research, groups have fewer institutionalized contacts in pluralist systems and are therefore expected to establish stronger links to parties than in corporatist systems (Rasmussen and Lindeboom, Reference Rasmussen and Lindeboom2013). Another important institutional variable is institutional strength. In particular, the higher the power of the parliament (vis-à-vis the executive), the higher the incentive for groups to establish contacts with political parties. The second group of variables that help explain the linkage between parties and organized interests is based on group or party characteristics. On the one hand, the way interest organizations were formed, the type of funding, their size and the degree of professionalization are all important variables to take into consideration (Rasmussen, Reference Rasmussen2012; Rasmussen and Lindeboom, Reference Rasmussen and Lindeboom2013; Otjes and Rasmussen, Reference Otjes and Rasmussen2017). On the other, the ideological position of parties (more or less moderate or extreme), their degree of institutionalization, and their participation in governmental offices may also affect party–group interaction.Footnote ² There is also evidence that mainstream parties are the main target of groups in the attempt to influence the policy-making process (Chaqués-Bonafont et al., Reference Chaqués-Bonafont, Cristancho and Muñoz-Márquez2020). Overall, material resources, organizational incentives (e.g. statutory linkages), and mobilization strategies (insider/outsider; issue salience, formal vs. informal contacts) seem to be important explanatory factors that account for party–group interactions (see Allern, Reference Allern2010; Chaqués-Bonafont et al., Reference Chaqués-Bonafont, Cristancho and Muñoz-Márquez2020).

This paper aims at identifying the conditions that are associated with interest group–party interactions. In particular, this study emphasizes the importance of the second bunch of factors, namely party and interest group characteristics, in shaping the strength of these relations. To do so, we focus on Southern European countries, which allow us to control for some key environmental features, as explained more in detail in the following section.

Party–group relations in Southern Europe

The empirical analysis of party–groups interactions focuses on the four main Southern European countries: Greece, Italy, Portugal, and Spain. These cases have been selected not only because there are no recent comparative and empirical studies on this topic, but also because they present a number of similarities in terms of the political and institutional settings. First, these countries can be characterized as ‘mixed-market economies’ (Molina and Rhodes, Reference Molina, Rhodes, Hancké, Rhodes and Thatcher2007), marked by a wide state involvement in the economy and similar models of industrial relations (‘middle’ or ‘soft pluralism’). Second, they share several institutional features, namely a relatively fragmented system of intermediation and social concertation mechanisms (Schmitter, Reference Schmitter, Gunther, Diamandorous and Puhle1995). Third, there has been a high level of politicization of party–group linkages (Morlino, Reference Morlino1998). Finally, they present a wide variety of party families that can be easily compared, ranging from (old and new) communists to extreme-right forces.

Van Biezen's pioneering study (2003) noted that political parties in Southern Europe have not emerged as strong movements of society, but rather as agents of the state. In this context, parties had an institutional, rather than societal, origin and had no real presence on the ground. The social basis of political parties was, therefore, created a posteriori, usually through expansive electoral mobilization (rather than partisan mobilization), leading to low levels of party affiliation and partisan linkages with society (Morlino, Reference Morlino1998; Gunther and Montero, Reference Gunther, Montero, Diamandouros and Gunther2001). As for the relationship between parties and interest groups, this tends to consist of a direct linkage with society and is of a pragmatic (vs. ideological) nature. From the organizational standpoint, most studies confirm the lack of a structured relationship (e.g. Schmitter, Reference Schmitter, Marks and Diamond1992; Puhle, Reference Puhle, Diamandouros and Gunther2001). In particular, the research conducted by Morlino (Reference Morlino1998) emphasizes the weak links between business groups and parties, the politicization but growing autonomy of trade unions, and the lack of interactions between professional associations and political parties. But, in several cases – notably for left-wing parties – the ideological affinity has led to the establishment of a ‘group dependence model’ in which groups (mainly trade unions) are seen as ‘transmission belts’ (Mavrogordatos, Reference Mavrogordatos2009; Tsakatika and Lisi, Reference Tsakatika and Lisi2013; Charalambous and Lamprianou, Reference Charalambous and Lamprianou2016).

To what extent have party–group linkages in Southern Europe followed the general trend observed in advanced democracies towards increasing distance between the two actors? Single case studies have contributed to further qualifying earlier findings. Scholarship tends to agree that links between parties and groups are nowadays driven by contingent and dynamic factors, rather than organizational or ideological ones. In Spain, for example, there has been an open and broad collaboration between parties and civil society organizations, while interest groups are prone to act strategically by targeting mainstream parties (Chaqués and Muñoz, Reference Chaqués-Bonafont and Muñoz Márquez2016; Barberà et al., Reference Barberà, Barrio and Rodríguez-Teruel2019; Chaqués et al., Reference Chaqués-Bonafont, Cristancho and Muñoz-Márquez2020). The pragmatic relation between parties and groups in Southern Europe is also confirmed when we consider patronage practices based on an exchange of material benefits (Lanza and Lavdas, Reference Lanza and Lavdas2000; Jalali et al., Reference Jalali, Silva and Silva2012; Sotiropoulos, Reference Sotiropoulos2019).

Despite these common traits, recent developments during the crisis period show distinct trajectories. Although Greece and Spain have seen the emergence of new actors that have reshaped party system format and dynamics, Portugal has remained relatively stable (Barberà et al., Reference Barberà, Barrio and Rodríguez-Teruel2019; Razzuoli and Raimundo, Reference Razzuoli and Raimundo2019; Sotiropoulos, Reference Sotiropoulos2019). Italy has also displayed high levels of electoral fluidity and new patterns of government formation, leading to much uncertainty and an extremely open party system. From this point of view, the crisis did not lead to convergence in terms of evolution of the party system (e.g. Hutter and Kriesi, Reference Hutter and Kriesi2019; Lisi, Reference Lisi2019a). Indeed, we find as much continuity as change across our cases. A comparative assessment of the evolution of party–group linkages in newer Southern European countries found that in Greece there is an overall decline in the intensity of these links, while both Portugal and Spain show that party genetic models are important in shaping interaction between parties and groups, leading to a variety of interactions, especially between right- and left-wing parties (Lisi, Reference Lisi2019b). Yet, in Spain the degree of innovation is higher due not only to the emergence of new political actors (Barberà et al., Reference Barberà, Barrio and Rodríguez-Teruel2019), but also to the consolidation of stable policy communities that interact with parties more on the basis of information exchange than favouritism (Molins et al., Reference Molins, Muñoz Márquez and Medina2016; Aguilar, Reference Aguilar, Muro and Lago2020). Finally, empirical evidence for the Italian case shows that interest groups have been able to adapt to an increasing fragmentation of the policy-making process, acting as a strategic actor in the policy cycle, thus establishing even more pragmatic and contingent relations with political parties (Lizzi and Pritoni, Reference Lizzi and Pritoni2019). This process was reinforced by the huge changes experienced at the party system level after the onset of the Great Recession.

The general objective of this study is to unveil the diversity of party–group interactions and the main drivers that account for different types of linkages as well as to inspect whether or not the Great Recession has had some homogenization effect. In light of the findings discussed above, we can formulate four broader hypotheses. The first is that new and old parties display different levels of proximity with regards to interest groups (H1). In particular, new parties are more likely to present a lower level of closeness compared to more traditional actors. On the group side, we expect to find closer ties for groups that rely more on formal contacts (at the institutional level), that privilege insider strategies and that have a longer experience (H2). In addition, we expect party–group proximity to be stronger for left-wing than for right-wing parties due to the ideological and historical legacy of the ties between communist and socialist parties on the one hand, and trade unions on the other (H3). Finally, we posit that there are still substantial differences in terms of party–group links across countries, and that the economic crisis has affected this variety given its differential impact on Southern European democracies (H4).

Data and methods

As mentioned in the previous sections, party–group linkages are operationalized differently depending on the approach and conceptualization. By far the most common way of studying this topic is through the analysis of party statutes (see Poguntke, Reference Poguntke, Luther and Müller-Rommel2002; Allern and Verge, Reference Allern, Verge, Scarrow, Webb and Poguntke2017). Another approach seeks to examine party–group ties in the parliamentary arena, for example through MPs' profiles (Celis et al., Reference Celis, Schouteden and Wauters2016; Celis and Mügge, Reference Celis and Mügge2018), or through appearances in parliamentary committees (Chaqués-Bonafont and Muñoz Márquez, Reference Chaqués-Bonafont and Muñoz Márquez2016). While these studies have emphasized the party side, interest group research has relied on surveys of organized interests to analyze their interaction with party actors (e.g. Rasmussen and Lindeboom, Reference Rasmussen and Lindeboom2013; Otjes and Rasmussen, Reference Otjes and Rasmussen2017). Clearly, the main limitation of these approaches is the asymmetrical analysis of party–group ties. In order to consider both actors as a dyad, thus avoiding any bias, we innovate previous research by implementing an expert survey.

To the best of our knowledge, only two studies have employed expert surveys to examine the interaction between interest groups and political parties. The first analyzes Italian interest groups and their influence on policy-making (Pritoni, Reference Pritoni2015). The second work relying on an expert survey is the study conducted by Charalambous and Lamprianou (Reference Charalambous and Lamprianou2016), which examines the linkage between radical left parties and civil society in those European countries most affected by the 2008 economic crisis (Cyprus, Greece, Ireland, and Portugal). Both studies contribute to qualifying conventional wisdom and to identify the conditions that might affect the relations between parties and groups.

In order to collect information on party–group ties, we conducted an expert survey in the main Southern European countries, namely Greece, Italy, Portugal, and Spain. The survey was implemented in four steps. The first was the identification of the most relevant political parties and interest groups. It was quite easy to identify the political parties as we decided to include the five most important (i.e. highest vote shares) parties with parliamentary representation. Therefore, we excluded minor parties as well as regional forces. The selection of interest groups was more problematic given the enormous variety of organized interests that can establish some kind of relation with political parties. For the sake of parsimony and in order to optimize the comparability across countries, we decided to include only the most important economic groups, distinguishing between trade unions and employer associations. Although we acknowledge that the variety of interest groups is much wider and that the number of groups that play a relevant role in the policy-making process is certainly higher, we decided to restrict the choice for two main reasons. First, economic interest groups are the type of organizations that dominate interest systems in these countries (see Molins et al., Reference Molins, Muñoz Márquez and Medina2016; Lizzi and Pritoni, Reference Lizzi and Pritoni2017). Second, it would be impossible and too much demanding for experts to assess party–group relations for a wide variety of group types. From this viewpoint, we chose to survey up to six peak organizations that have traditionally played an important role in their respective political system.Footnote ³ Our choice was validated with the aid of secondary literature and advice from country experts.Footnote ⁴ Therefore, we compare groups that are comparable in terms of ‘party-political relevance’ and organizational resources, yet not in terms of type of membership (corporate vs. individual).

The second step consisted of assembling a list of specialists whom we considered sufficiently qualified to be an expert for our study. The most important requirement was that the expert had published an article or book on interest groups in the past 10 years. The publications did not have to be in English because the surveys were all translated into their original language. In order to assure us that experts had detailed knowledge of a given country, we circulated the master list among the members of the project ‘From Representation to Legitimacy: Political Parties and Interest Groups in Southern Europe’ and canvassed for recommendations. If these recommendations met the first requirement (which they nearly always did), we included them in our master list of experts. In total, we assembled a list of 155 names (see Table 1). For each country, we aimed to have 10 completed questionnaires. Although no fixed number is required for an expert survey, Huber and Inglehart (Reference Huber and Inglehart1995: 76) suggest that one should obtain a minimum of five experts per country; this strategy was empirically supported by Gabel and Huber (Reference Gabel and Huber2000) who show that various expert surveys converge on their estimates for parties' overall ideological orientation. As clearly demonstrated in column 3 of Table 1, we have a sufficient number of experts for each country, with all countries meeting the goal of 10.

Table 1. Population (sample) vs. survey respondents – distribution by country

Source: Own elaboration.

Given the complexity and multidimensionality of the topic, the final list of specialists with proven knowledge on the main objects of analysis – political parties, trade unions, and business associations –, included not only political scientists, but also scholars from different fields (e.g. sociology, history, and economics) and some journalists. It was necessary to extend the academic background given the fact that some Southern European countries have very few political scientists with expertise on interest groups and their relations with parties. We checked for the reliability of experts' responses by calculating the ‘exact agreement index’ (Charalambous and Lamprianou, Reference Charalambous and Lamprianou2016) and we found that no expert was completely out of tune with the rest of the selected specialists. Despite some experts failed to give score to all dyads, no one systematically avoid to answer to the ‘closeness’ question for all parties or groups included in the survey.

The third step involved drafting the questionnaire. Given that it was the first time an expert survey had been administered on such a topic, we could not rely on existing research. However, we were able to benefit from Allern and Bale's project on the links between trade unions and centre-left parties (Allern and Bale, Reference Allern and Bale2017) for the measurement of ‘party–group closeness’, as well as the type of contacts established between these two actors. The survey provided a definition of key concepts, such as ‘closeness’, ‘influence’, etc. to guarantee that all experts interpreted them in the same way. The final questionnaire included two parts. The first asked questions on the relationship between the main economic interest groups (national peak-level business and trade union confederations) and political parties, the interest groups' access to various actors and their capacity to mobilize and influence specific arenas and public decision-making processes. A second set of questions controlled for the respondents' sociodemographic characteristics and specific areas of expertise.

The final step was the survey administration. The surveys were fielded between September and December 2018. The questionnaires were sent out via e-mail to all experts included in the database. The platform used for the filling in of the questionnaires was the 1KA – OneClick Survey (https://www.1ka.si). Two e-mail reminders were sent out to all experts. If we take the number of experts we approached as a base and relate it to the number of completed questionnaires we actually obtained, the response rate ranged between 36% in Spain and 52% in Italy (Table 1).

The empirical analysis follows a two-step approach. First, we evaluate the relations between parties and groups by considering our key dependent variable, that is, the ‘closeness’ of party–group connections, based on the question ‘what is the level of proximity that you believe to exist between the main interest groups and each political party?’. In particular, we examine the score each expert gave to the ‘strength’ of party–group interaction for every party–group dyad, on a 5-point scale (1: not close; 5: very close). Second, we perform a multivariate analysis testing the impact of key independent variables included in different models, regressing party-group closeness (dependent variable) on the main independent variables (see below). Our unit of analysis is each individual party–group dyad. We perform a linear regression analysis (ordinary least squares (OLS)) with country fixed effects, allowing us to control for unobserved country-level variation. The original data collected through the expert survey allow us to include unexplored independent variables related to the mobilization strategy of interest groups, in terms of access (i.e. the target of group activities) and type of contacts (both formal and informal). In the first case, we ask the following question: ‘Please state how important are the following entities for each of the interest groups under analysis, when the latter aim to influence and participate in the processes of public and political decision-making’. We consider in the analysis only the response associated with political parties.Footnote ⁵ This variable (named ‘mobilization parties’) ranges from 1 (not important) to 5 (very important). As for party–group contact, we created a new variable (‘type of contact’) based on an index that measures formal (at the institutional level) vs. informal contacts (ad hoc meetings or exchanges).Footnote ⁶

Findings

Descriptive analysis

We begin the analyses by presenting the aggregate scores for each party–group dyad. Starting with the direct measure of closeness between the two actors of intermediation, we see that there are interesting differences across countries. While Greece displays higher scores for party–group ties (2.57 on a 5-point scale, ranging from 1 – not close to 5 – very close), Spain registers the lowest levels (2.21). This finding may be due to the presence of two recently created parties, Podemos and Ciudadanos (C's), which have had insufficient time and resources to build solid bonds with the main interest groups. On the other hand, Italy and Portugal present roughly the same level of closeness (an average of approximately 2.4), indicating that these ties are rather weak, thus confirming the overall impression given by previous studies. Although these country differences are not statistically significant, there are relevant disparities across distinct types of party and interest group. The general picture suggests that the economic crisis had not a homogenization effect, and this is in line with our fourth hypothesis. Moreover, as argued in more detail below, party system differences and distinct patterns of policy-making processes are important environmental factors that shape interactions between parties and interest organizations.

In order to shed more light on party–group relations, we examine the results country by country. Starting with the Greek case, we can make the overall observation that Pasok and ND have established more widespread links with the main interest groups, while KKE and in particular GD display distant relations with the selected organizations. It is also worth noting that, according to our experts, although Syriza has been able to build close linkages with GSEVEE, its relations with the main business association are very weak. Although this might be largely explained by the clientelistic politics pursued by both Pasok and ND (Afonso et al., Reference Afonso, Zartaloudis and Papadopoulos2015; Sotiropoulos, Reference Sotiropoulos2019), the government experience under Syriza (2015–2019) has probably contributed to fostering similar ties even for this young political actor. In general, these findings resonate with Sotiropoulos' argument (Reference Sotiropoulos, Sotiropoulos and Featherstone2020: 41) that the recent crisis has contribute to diversifying the way interest groups connect to state institutions, going beyond the clientelistic/corporatist model. This was just one step further in the move towards an increasing disengagement of the main interest groups from the political parties. The radical forms of mobilization of civil society against the state that took place during the economic crisis not only enhanced this process, but it also led the main interest groups to legitimize the protest against governing parties.

In the case of Italy, the scores of party–group closeness are generally rather weak, especially for the recently created M5S. Usually considered a ‘populist’ party, the M5S was formed in 2007 and ran for the first time in national elections in 2013 (Bordignon and Ceccarini, Reference Bordignon and Ceccarini2013). Given its strong links with grassroots movements and online forms of mobilization (Mosca, Reference Mosca2014; Tronconi, Reference Tronconi2015), it is not surprising to find very low levels of proximity between the M5S and traditional economic groups. The partial exception is its link with the Confcommercio, given that some of the policies advocated by the party are in line with the interests of autonomous workers. This is also the case of the Lega, which registers the highest level of closeness when we consider the same interest group. As expected, trade unions display distant ties with centre-right parties, but even in the case of the PD the figures always remain below 3, confirming the tension – if not conflicts – between the main centre-left party and labor organizations especially during Renzi's government (Mattina and Carrieri, Reference Mattina, Carrieri, Allern and Bale2017). However, it should be noted that the relationship between trade unions and the PD improved significantly after Renzi withdrew from party leadership and formed a new party. Regardless of the interpretation of our experts, under the new secretary Zingaretti (2019–2021) the leftist party has certainly become closer to CGIL than employer associations.

The results for Portugal provide sound confirmation of previous findings, namely the strong alignment between the radical left and trade unions, as well as the ‘natural convergence’ between right-wing parties and business organizations (Razzuoli and Raimundo, Reference Razzuoli and Raimundo2019). Indeed, the proximity between CGTP and PCP achieves the highest score (4.85), while the ties between the radical left (BE and PCP) and employer associations are the lowest (see Table 2). The PS displays stronger ties with the UGT, but it is also remarkable to find quite close links with business organizations, higher than the score for the main trade union (CGTP). Finally, according to the experts, the PSD is closest to the CIP, followed by the other two business organizations. This alignment has remained rather stable also during the economic crisis and reflects the higher mobilization potential of CIP compared to other employer associations (see Lisi and Loureiro, Reference Lisi and Loureiro2019). It is also worth noting that the fragmentation of the business sector has hindered a more robust right-wing party–group alignment as emerged in the Spanish case during the democratic regime. Overall, these findings seem to confirm the stability of system intermediation, which parallels the resilience of its party system during and after the Great Recession.

Table 2. Closeness between parties and interest groups in Southern Europe (averages)

Source: See Table 1.

Spain is a particularly interesting case because it includes two recently-formed parties of different ideological positions, namely Podemos and C's. As one might expect, the former displays close ties to trade unions, while C's seems to have more links with business organizations, especially CEOE and CEPYME. The same occurs with the right-wing PP, with the exception that this party displays higher scores with respect to ASAJA and COAG. On the other hand, PSOE displays rather weak ties with almost all interest groups, with the partial exception of UGT, with whom there is a stronger convergence. Finally, unsurprisingly IU has a great proximity with CCOO, but displays distant relations with business associations. This picture tells us two things: first, it is still visible the imprint of the cleavage structure in shaping party–group interactions. Indeed, the main political parties have sought to establish and condition social organizations during democratic consolidation (Verge, Reference Verge2012). Second, mainstream parties seem less important in mediating the connection of interest groups to politics, and this can be interpreted as a result of the economic crisis that accelerated a longstanding process of the emergence of lobby-like interaction between economic interests and political power.

To what extent are parties the privileged targets of interest groups' strategies? We asked the respondents of the survey to evaluate the importance of different types of party contacts (government, parliament, central organization, or individual politicians) on a 5-point scale (1 – not important, 5 – very important). Also in this case, we can observe interesting differences both across countries and across organizations. In the Greek case, contacts with the government seem to be the most important type of interactions, followed by party leaders or organizations (Table 3). On the other hand, informal contacts are regarded as less important. The pattern is quite similar in Italy, although the distinctions across different forms of relations are less evident. Contacts with government are particularly important, but in this case other institutional actors seem to play a significant role, even more important than for political parties. As far as Portuguese groups are concerned, the most noteworthy finding is the different patterns displayed by trade unions and employer associations. Although all groups seem to privilege governmental contacts, business organizations seem to rely more on informal contacts than on formal meetings with political parties. Finally, according to our experts informal contacts and government meetings are equally important in Spain, but there are no significant differences among the other categories.

Table 3. Type of contacts between interest groups and parties

Source: See Table 1.

Explanatory analysis

How can we explain the variation in the expert evaluation of the party–group relationship? We use the average score for each party–group dyad as the dependent variable and we regress it on several independent variables deemed relevant in the literature for understanding the closeness between parties and groups. The baseline model includes three variables. First, we distinguish between group types, using a dummy variable that differentiates trade unions (which take the value 0) from business organizations (1). Second, we consider the difference between old and new parties, measuring this item through a dummy variable that takes the value 0 for new parties, and 1 for old parties.Footnote ⁷ Third, we also control for government participation, a dummy variable with a value 1 if the party participated in government before 2018 (and 0 otherwise). Finally, we control for distinct countries and we weight the results according to ideological position of each expert on the left-right scale.Footnote ⁸

The regression analysis displayed in Table 4 reports the main results. The basic model (second column) shows that whether the party is a new or old one matters. In particular, older parties are associated with stronger links to group organizations, thus supporting hypothesis 1. One of our control variables, namely group type, also achieves statistical significance, suggesting that business groups tend to display (according to experts' eyes) higher level of closeness to party actors.

Table 4. Explaining party–group ties (OLS coefficients)

Notes: (1) Dependent variable: ‘closeness’; (2) *p < 0.10; **p < 0.05; ***p < 0.001; (3) Country: Greece as the reference category.

Source: See Table 1.

Moving to the analysis of group-related variables (see model 2), none of them achieves statistical significance. Neither the age of foundation (measured through its logarithm) nor the importance of parties for interest group strategy (gauged through a question included in the expert survey, see the variable ‘mobilization parties’ in the previous section) displays a statistically significant coefficient. Moreover, it is worth noting that our measure of party–group closeness is not associated with the type of contacts these actors establish, even when we include the linkages – established by the statutes – between the central party organization or the parliamentary group, on the one hand, and organized interests, on the other (as measured by Allern and Bale, Reference Allern and Bale2017). This means that the statistical analysis does not support our second hypothesis.

The regression results show that the variance explained is relatively low in the first two models. Consequently, we add specifications by including party-related variables. On the one hand, as noted in the literature review, the age of party organizations seems particularly relevant for explaining the closeness of party–groups ties. We operationalize this variable by considering the foundation year of each party, including its logarithm in the statistical model. On the other, we enter in the equation the vote share for each party (operationalized as the percentage of votes obtained in the lower Chamber), as suggested by the rational approach. According to Allern and Bale (Reference Allern and Bale2017), this is an important dimension that is expected to favour party–group ties. In addition, we include a variable for party ideology, using as indicator the position of parties on the 11-point left-right scale based on the Chapel Hill Expert Survey (CHES).Footnote ⁹ Finally, we add an interaction term between the type of group and the ideological position of the parties (distinguishing between left vs. other parties), with the idea that trade unions display historical linkages to left-wing parties. This variable aims to control for left wing-union dyads vs. not left-wing union dyads.

The results indicate that, although still parsimonious, model 3 does a good job in explaining the variation in party–group linkages. Indeed, several factors account for party–group closeness. Contrary to what we expect from the literature, group type is strongly associated with the intensity of this relationship, suggesting that business groups generally display stronger ties with political parties. As mentioned in the previous section, this is due mainly to the relatively strong alignments between business organizations and right-wing parties. In addition, we also corroborate the closer association between old parties and interest groups, as new parties tend to be more distant from the main peak organizations.

The interaction term between trade unions and left-wing parties indicates that their connection is particularly strong. This seems to confirm a common wisdom in the literature, in particular the importance of the legacy – especially in Southern European countries – of the linkage between left-wing parties and the main trade unions. Our results suggest that this is indeed the case as indicated by the positive coefficient. Finally, the vote share shows a significant association with the degree of closeness between the two types of political actors. This means that the higher the electoral strength of the party, the greater the strength of party–group linkages. Although this runs against the findings of Allern et al. (Reference Allern, Bale, Otjes, Allern and Bale2017: 323), the results are not directly comparable because their analysis is limited to the link between trade unions and left-of-centre parties and relies on a different set of countries. Considering the electoral basis of mainstream parties included in this study and the challenges that have emerged in the electoral arena after the economic crisis, we can speculate whether this finding reflects the broader process identified by Kirchheimer (Reference Kirchheimer, La Palombara and Weiner1966) long time ago towards an increasing interaction between (non-partisan) interest groups and political parties. Indeed, the need to secure access to a variety of interest groups for financial and electoral reasons was one of the main transformations highlighted by this scholar in his seminal contribution. The trans-class appeal of catch-all parties and the longing of organized interests for increasing (or maintaining) their power can be interpreted as powerful incentives to establish connections.

One particular aspect deserves further exploration, namely country difference and, in particular, whether there are statistically significant differences in the way groups interact with political parties in distinct contexts. Our results suggest that the main difference is between Greece and the remaining countries. As mentioned in the descriptive analysis, Greek groups systematically display stronger ties than those shown in other Southern European countries. However, no statistically significant effects can be found when we add interaction terms between our main independent variables and the conditional effect of countries. This is the case, in particular, of the variable ‘party size’ and ideology. Contrary to what we found for the Dutch and the Danish cases (Otjes and Rasmussen, Reference Otjes and Rasmussen2017), party system characteristics – namely the patterns of government and the degree of polarization (or extremism) – do not seem to be relevant intervenient variables that shape the collaboration between parties and groups in Southern European countries. One possible interpretation for these negative findings is the high level of instability and unpredictability of party systems in a region particularly affected by the Great Recession; this had a huge impact on the reshaping of patterns of competition and government through the rise of new parties and the formation of new government coalitions (Bosco and Verney, Reference Bosco and Verney2017; Casal Bertoa, Reference Casal Bertoa2019).