Clustering and Analysing Relevant Policy Dimensions of Populist, Left-Wing, Centrist, and Right-Wing Parties across Europe

In today’s Europe, we are witnessing an emergence of new forms of governments characterised by populist rhetoric (Kyle & Gultchin, 2018). These new forms do use populist rhetoric in a traditional sense, yet they differ significantly from what we can see in the literature in terms of populist policy preferences and the effects that these can have on an economy. These new forms in fact distance themselves significantly from the concept of economic populism, which has been used to focus on the Latin American context (Dornbusch and Edwards 1989). Since then, the literature mainly focused on rhetoric, and it puts aside the idea of studying populism in terms of policies. The lack of development of this strand of literature for more than 30 years is because of three main reasons: because the concept of economic populism did not apply to the new and right-wing emerging forms of populism, because social sciences abandoned structuralism as a way of thinking and because the study of populism started to focus more on discourse-analytic approaches (Rovira Kaltwasser, Taggart, Ochoa Espejo, & Ostiguy, 2017). These three elements together redefined the study of populism in broader terms and put on the sideline the study of populism in terms of policies.

Nonetheless, the study of populism in terms of policy choices is starting to gain new momentum, especially in terms of measuring and contextualising their consequences (Funke et al., 2020; Guriev & Papaioannou, 2020). Many realised that the new forms of governments characterised by populist rhetoric do have similar patterns and similar outcomes, despite these being very different from the Latin American context and between each other. On the one hand, there is now a consensus that the forms of populism connected to Latin America have theoretical models that do not apply anymore. On the other hand, however, the forms of populism that we are seeing now cannot be connected to any specific theoretical model. This is especially important for Europe, where we have the emergence of new political models connected to populism that do not lead to economic disaster nor are connected to extreme redistribution (Benczes, 2018). Some major examples of this phenomenon are, for example, Hungary or Poland (Toplišek, 2020). These new forms currently need a new theoretical elaboration in terms of policy choices not only because of their novelty but also because they are becoming more and more relevant.

Consequently, this theoretical elaboration is the major contribution of this paper. This paper aims at giving a deductive and theoretical contribution in conceptualising populism in Europe in terms of policies and its consequences. This data-driven deductive contribution is now possible thanks to the relatively recent emergence of harmonised data availability across Europe and because of the rise of new techniques in political science (Cunningham, 2021). Overall, today’s extensive data make it possible to perform a configurational analysis that makes a taxonomy possible. This work uses the two most recent rounds of Chapell Hill Expert Survey (CHES) and the PopuList to perform a comprehensive analysis in terms of policy preferences. We use cluster analysis as a technique to give quantitative and deductive insight on this new configurational model. This, in turn, makes it possible to have a new theoretical elaboration possible where we elaborate on the interaction between populist rhetoric and policy choices in Europe. This new theoretical elaboration based on data is as well the research gap that this work aims to fill.

This paper finds out that populist parties consistently position themselves as economically centrist and identity extremists. At the beginning of the analysis and according to the descriptive statistics, we find that populist parties seem to be a fourth distinct pole compared to centrist, right-wing and left-wing parties. This happens no matter the geographical position of the party, and no matter whether the party defines itself as right-wing or left-wing populist. Based on this result, we perform a cluster analysis on all parties to see whether the data show four clusters that overlap with the four clusters we used to identify parties: right-wing parties, centrist parties, left-wing parties and populist parties. We also split the clusters into four to see as well whether parties in Europe are divided into four geographical clusters linked to specific historical paths: Western Europe, Southern Europe, Northern Europe and Central and Eastern Europe.

Cluster analysis is a machine learning technique that tries to find cluster of similar observations across a dataset. The first cluster analysis in the paper tries to see whether we have a specific populist cluster next to the expected clusters of right-wing parties, centrist parties and left-wing parties. The second cluster analysis tries to see what happens if we cluster parties as Western Europe, Southern Europe, Northern Europe and Central and Eastern Europe: would one of these geographical clusters be particularly linked to populist parties? The idea would be to see if there is a correlation between populist parties and geographical position like for the Latin American case. The null hypothesis behind these two cluster analyses is that the policy preferences are homogeneous and non-statistically different among all parties including populist ones. The alternative hypothesis is that policy preferences do create four distinct, heterogenous, and statistically significant groups among which populist parties represent one. Both hypotheses are the reason why we create four clusters: because we have four expected political clusters in the first case, and because we have four expected geographical divisions in the second. We also repeat the analysis for 2014 and 2019 to see if the results are consistent.

This paper finds that the four clusters represent a mixture of both these expectations. The first cluster, mainly located in Central and Eastern Europe and composed by populist parties, specifically focuses on extreme right-wing identity politics positions and centrist economic positions. The second cluster is located across Western, Northern and Southern Europe, and it portrays a left and moderate centre concerned with the role of the European Union. The third cluster, the liberal centre, is the one more in line with European and liberal values and it is represented all over Europe. The last one, almost uncannily representing identity politics and the intersectional left, is present mainly in Western and Southern Europe.

Overall, this paper looks at how parties divide into the European political space. The main goal for doing so is creating a new configurational model of populism in terms of policy preferences for Europe, as the one created for Latin America does not apply (Dornbusch & Edwards, 1989; Funke et al., 2020). To do so, it analyses the fundamental differences between populists and other sides of the political spectrum. A secondary but equally important objective is to further analyse the relation between politics and policy in populist and non-populist parties in Europe (Kyle & Gultchin, 2018; Toplišek, 2019). Quantitatively analysing policy preferences is useful to understand the link between politics and policymaking. Politics, the use of power to maintain control on a certain territory, and policy, the system of rules to achieve a certain outcome, are strongly interrelated. Policy preferences are in this context an intermediary element between the two. The policy preferences expressed by a party are an integral part of its politics as they serve to get votes and attention. These policy preferences have, in turn, a concrete effect on a country’s policymaking (Dahl, 1961). Once a party is in power, we expect to implement at least part of the policy preferences expressed before.

Of course, the relation between a country’s politics and the policy preferences of its parties is not always straightforward (Matheson, 2016). The policy preferences of a party can be circumstantial and even drastically change when a party is in power. In many instances, politics and policy preferences do not translate into the desired policymaking because of institutional or societal constraints (Kumlin & Stadelmann-Steffen, 2014; Pierson, 1996; Wlezien & Soroka, 2016). For example, in Europe, we know that Central and Eastern Europe tends to have a cronyist approach to policymaking despite strong anti-elite policy preferences (Martin, 2017; Stadelmann-Steffen & Eder, 2021). In this sense, we can say that politics translates into policies via the expressed policy preferences only in specific policy environments without institutional, cultural or political constraints.

Consequently, we do not imply a cause–effect relation between politics, policy preferences and politics. Nonetheless, it is still important to understand policy preferences as they are a strong link between politics and policy (Cooper & Williamson, 1994; Tharanga, 2018). In this case, policy preferences aim at informing how policies could diverge between different parts of Europe, especially if they are fundamentally different as analysed in the cluster analysis. The four clusters identified at the end of this paper, which partially but not fully represent the division between populist, right-wing, left-wing and centrist parties, give an insight on how politics and very likely policies will further diverge in Europe. There is a fundamental policy preference difference among European parties that is given by different ideological stances rather than geographical location that might foster even more divergent policies in the future.

To analyse policy preferences, the paper uses Chapel Hill Expert Survey (CHES) data for 2014 and 2019 for all parties of countries members of the EU mixed with the PopuList database. The paper also uses Gorton (Gorton, Douarin, Davidova, & Latruffe, 2008) as an empirical basis and Dornbusch and Edwards (Dornbusch & Edwards, 1990) as a theoretical one, even if we try to apply them to a different setting and we depart from both of them. The paper concludes that despite the most common definitions of populism (Abromeit, 2017; Mudde, 2004), we should not confuse populism with its host ideologies and how the two mix in terms of policy preferences. We see, in fact, that the parties’ political spectrum in contemporary Europe goes beyond a left–right divide or a populist–non-populist divide. The politics and policy positioning of such parties represent a combination both of libertarian–authoritarian, left–right, populist–nonpopulist and geographical historical paths.

The research question of this paper is about understanding which policy configuration do political parties characterised by populist rhetoric have across Europe, in order to define populism in terms of recurrent policy choices in this context. This paper looks at whether these differences are significant across North, South, Western and Central and Eastern Europe, given the concentration of these parties in some specific areas as shown in Table 2. The following paragraph will briefly elaborate on the importance of looking at the policy preference of parties, the relation between policy preferences and populism and some of the significant differences we can find inside Europe.

Parties’ positions have long been thought of as important determinants of electoral outcomes (Castanheira, Crutzen, & Sahuguet, 2010). However, defining parties’ policy positions is far from being a simple task. As shown in the literature, left and right are often defined as an “empty vessel” that changes according to time and situation (Huber & Inglehart, 2016). This is especially relevant for today’s Europe, where we can see new cleavages emerging in a strictly interrelated waid (Hooghe & Marks, 2017; Welzel & Inglehart, 2016). This said, it is still possible to define a left–right divide based on an economic and cultural dimension (Giebler, Meyer, & Wagner, 2021). On these terms, we can put left-wing and right-wing on the opposite sides of a continuum. On one side, we have little or no redistribution and individual freedom for the right, while on the other, we have more or complete redistribution and collective norms for the left. The centre would position in the middle of this divide. This is as well the scale used in the CHES data, which, however, measures other positions such as identity politics and positions towards the European Union to make the mapping more complex and accurate. This is useful for the first part of our analysis, yet with its cluster analysis, this paper shows how new cleavages in Europe go beyond a simple left–right continuum.

The division of policy positioning in terms of economic and cultural dimensions is nonetheless useful for understanding the previous works on populist parties in terms of policy positions. Populism in terms of policy choices has previously been defined on a traditionally leftist side of the spectrum in the Latin American context (Conniff, 1982; Dornbusch & Edwards, 1990; Kaufman & Stallings, 1991). However, already in that positioning, it started showing a variety of other elements from authoritarianism to neoliberalism. Today, many continue to see populism in relation to its leftist grievances (Mouffe, 2019). Others see populism across the globe more in terms or nativism or economic shocks (Art, 2020; Rodrik, 2017). This led to today’s main definition of populism as a discursive style (Hawkins, Carlin, Littvay, & Rovira Kaltwasser, 2019a) or as a “thin-cantered ideology that considers society to be ultimately separated into two homogenous and antagonistic groups, ‘the pure people’ versus ‘the corrupt elite’, and which argues that politics should be an expression of the volonté générale (general will) of the people” (Mudde 2004; p. 543). All these elements are equally important in defining the European space and leave the open question of which one is more predominant or how these elements exactly configure themselves. This is especially important because the previous theoretical definition of this kind has been exhausted. We map policy positioning of parties with a cluster analysis to create a new definition and to give a new direction to this strand of the literature.

There already have been attempts of mapping parties. One main example deals with the far right (Golder, 2016). For what concern mapping parties and populism, the main work on the topic in the literature is the “demand–supply study of populism” paper by Inglehart and Norris (Inglehart & Norris, 2016). In this work, the authors map the reasons that make people vote for populist parties in order to change that through policymaking. A work more similar to this one is by Meyer and Wagner (Meyer & Wagner, 2020). However, it differs from this one as it focuses on how party positions influence perceived left–right positions. Another one, by Hawkins and Castanho Silva in the book The Ideational Approach to Populism (Hawkins, Carlin, Littvay, & Rovira Kaltwasser, 2019b), relies as well on a mixture of validated experts but focuses on measuring populism from a rhetorical point of view.

Last, we have the works “Measuring Populist Discourse: The Global Populism Database” (Hawkins et al., 2021), “Measuring Populism in Political Parties: Appraisal of a New Approach” (Meijers & Zaslove, 2021) and “Measuring Populism Worldwide” (Norris, 2020). The first differs as it mainly focuses on discourse. However, the second work focuses more on conceptualising populism in a precise and multi-dimensional way rather than comparing existing data to see if we can find enough differences to consider it a different faction. The last one focuses more on conceptualising populism as well, and it has a global perspective instead of a European one. Overall, there are also numerous studies that try to map the drivers of populism using well-qualitative approaches (Akkerman, Mudde, & Zaslove, 2014; Castanho Silva, 2017; Hawkins, Riding, & Mudde, 2012; Schulz et al., 2018; Wuttke, Schimpf, & Schoen, 2020).

This paper emerges out of a wider study on the political divergence between Western and Central and Eastern Europe and seeks more particularly to provide a clearer understanding of the characteristics of populist and anti-establishment parties in Europe, which can provide valuable insights into likely responses to reformed policy environments. CHES data allow for the comparison across 32 EEA Member states. The analysis is divided into three parts based on other relevant works using similar methodology (Gorton et al., 2008).

First, we conduct an exploratory analysis to see common patterns in policy preferences. We divide the data in two ways: a geographical one and “left versus right” one. The geographical one divides between Southern Europe (Portugal, Italy, Greece, Spain, Cyprus, Malta), Western Europe (United Kingdom, France, Belgium, Holland, Germany, Switzerland, Austria, Ireland, Luxembourg), Northern Europe (Sweden, Norway, Denmark, Finland, Iceland) and Central and Eastern Europe (Bulgaria, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia, Slovenia, Croatia). We exclude Turkey from all the classifications because it is outside the European Union. We also exclude Iceland because it is present only in the 2019 dataset. This division is created according to standard divisions of Europe in terms of economic performance and cultural background (Browning, 2020; Sushytska, 2010). The “left versus right” division uses the CHES variable LRGEN, and it is in line with the literature on the left and the right being divided on economic and cultural perspective (Giebler et al., 2021).

Second, descriptive statistics are presented for the whole sample regarding the distribution of the experts’ responses for the Likert scales. Mean scores for the 25 variables are presented with significant differences identified using ANOVA and presenting the F-tests scores and the significance levels. Third, group of parties with similarly held positions in terms of ideological stance are identified using cluster analysis. This is to investigate whether differences in parties’ positions can be discerned according to a right–left–centrist–populist criterion. Since we used 25 variables to describe the parties and in the descriptive part of the study, first we checked the correlations between the indexes. We find that there are multiple indexes which are highly correlated. To deal with the effect of highly correlated variables on the cluster creation, we chose to perform a factor analysis first. In the study, principal components presenting eigenvalue greater than 1.0 were chosen with factor loadings being greater or equal than 0.5 on the least factor. The cluster analysis was performed using the k-means algorithm (Likas, Vlassis, & J. Verbeek, 2003). The factors defined by the factor analysis were used as the basis of the clustering. Instead of choosing the number of clusters based on prior analysis, we chose to have four clusters to compare if there are differences or not in grouping the parties based on to potentially match the left–right–centre–populist classification discussed above. More information on the analysis, including source data and R code, is available online and on GitHub.

Policy positions in the CHES data are measured through secondary survey work conducted during 2014 and 2019. The database maps the policy positions of all parties across Europe by summarising the opinion of 337 experts on each topic for each party. Using these, many experts minimise the individual experts’ biases, and it gives an overall objective measurement for each position. The survey measures the opinions on a 7-point or 10-point Likert scale. These scales drew on previous attempts to capture policy positions and were designed to fit within a cross-national framework. The selected questions provide insights into parties’ opinions regarding: (a) positions towards the European Union, (b) positions towards democracy, (c) positions towards libertarian versus traditional issues, and (d) ideological stance. In addition to the ideological side of the party provided by CHES, we merge data from the PopuList database (Rooduijn et al., 2020). In this database, the parties are classified as populist from 106 experts, and populism is defined as a thin ideology (Mudde 2004).

We use the CHES data from 2019 to have the most recent data available on Europe. We compare it to the 2014 round to have an idea of the consistency of type of parties over Europe in the recent available timeframe. We do not use the 2017 edition of the data as it is based on a smaller number of countries. The CHES data categorise the faction of parties (or ideological stance, as called in the database itself) through a variable called LRGEN. This paper segments the parties in the following way: we classify parties as right with a value between 10 and 6, as centre if the value is between 6 and 4 and as left if it is between 4 and 0.

When we talk about populist parties, we talk about all parties classified as populist by Roooduijn et al. (2020) regardless of being right-wing, left-wing or centre. In Tables 1 and 2, we can see how there are 264 parties in total and 47 classified as populist for the year 2014. Overall, 17% of the parties in the database are populist. For 2019, we have 264 and 57 populists (21% of the total). For what concerns the geographical divide, we can see more right-wing parties in Central and Eastern Europe, probably for their preference for an anti-leftist and therefore anti-communist rhetoric. We also see more left-wing parties in Western Europe, as an answer to liberalism and globalisation. Surprisingly, there is an almost equal distribution of parties by side in Northern and Southern Europe.

The distribution of responses for each Likert scale item for a selected question and the mean scores for each political side is shown in Figures 1–4 and in the supplementary files. Figures 1–4 specifically show the redistribution across the policy dimension improving public services versus reducing taxes for ideological factions and geographical division, both in 2014 and 2019. The figure reveals that left-wing, right-wing, centrist and populist have elements in common specific to the ideological side. However, despite the literature, there are no apparent common patterns that characterise Central and Eastern, Western, Southern and Northern Europe. More significantly, populist parties seem to be centrist in term of economic stance but more extremist than the right-wing for what concerns identity issues.

Fig. 1:

Fig. 2:

Fig. 3:

Fig. 4:

The mean scores for each Likert scale by ideological side for 2014 and 2019 are reported in the supplementary files. To check for significant differences between political sides, ANOVA was performed, and F-test scores and significance levels are reported for a comparison (a) between different political sides and (b) between Central and Eastern, Western, Southern and Northern Europe. Significant differences are uncovered between political sides on nearly all the variables for what concerns the political side and in line with Figures 1–4. Overall, populist parties position themselves on the far-right spectrum for many issues, both in 2014 and 2019. It is especially interesting considering that in these tables, all populist parties are considered together no matter if self-identified as right-wing, left-wing or centre. We find this pattern for the following elements: anti-elite salience, position towards ethnic minorities, position towards the European budget and position towards European integration. Coherently, populist parties are also the only faction to consider that European integration is of no importance (EU salience). In the second categories of extreme values, populist parties position themselves in even more right-wing positions than the right-wing parties themselves. This happens for the following variables: civil liberties versus law and order, EU dissent (only for 2019), social and cultural values, social and cultural value salience, immigration policy, multi-culturalism, nationalism and social lifestyle.

The second category of values for populist parties confirms the hypothesis coherent with the literature that the average values should be centrist, as populist parties are just parties using a specific rhetoric. Such values are corrupt salience, deregulation, economic intervention, EU cohesion, EU foreign and security policy, EU internal market, EU position, ideological stance, ideological stance salience, ideological position of the party, redistribution, decentralisation to regions, religious principle, spending versus collecting taxes, and urban–rural divide. Of course, some of these results are surprising as well. According to the rhetoric they are supposed to apply, we are intrigued to find that populist parties are centrist for what concerns corrupt salience and urban–rural divide. The rhetoric of populist is supposed to defend the “real people” against “the corrupt elite.” If we also consider most populist parties in Europe right-wing as suggested by all the other values, it is also surprising to see populist parties to be on average centrist for what concerns the defence of religious principles in 2014. This result for the economic variable is indeed coherent. Last, we find more significant variables in the division between ideological factions but no clear disparity between Western, Central and Eastern, Northern and Southern Europe. While political sides are interesting to report, it is important to investigate whether political sides are the most important factors in distinguishing groups of parties with similarly held positions. This is investigated in the next subsection, through the application of factor and cluster analysis.

Two tests were applied to assess the validity of the factor analysis. The Keiser–Meyer–Olkin measure of sampling adequacy (Kaiser, 1970) is 0.89, indicating that the data matrix has a very good correlation to justify the application of factor analysis. Bartlett’s test of sphericity is large and statistically significant at the 1% level, and therefore, the hypothesis that the correlation matrix is the identity matrix can be rejected. These measures indicate that the set is appropriate for factor analysis.

A five-factor solution is adopted, choosing the factors that present an absolute eigenvalue greater than 0.5 (Table 3). This solution explains 81% of the variance in the data set, which is more than satisfactory, according to Hair et al. (1998). The first factor is associated with identity values, as it relates to the position towards immigration, ethnic minorities, lifestyle, civil liberties and security. The second factor relates to economics, as the main loadings are statements concerning taxes, deregulation, redistribution and state intervention. The third factor is associated with positions towards the European Union, as it concerns all the statement concerning the EU and whether decentralise power outside the nation state. The fourth factor concerns anti-elite rhetoric and corruption. The last factor concerns more a traditional ideological divide as it focuses on dissent, libertarian versus traditional values and stance of economic issues in the party’s ideology.

These factors form the basis of the cluster analysis. Using the criteria outlined in the methodology section, a four-cluster solution was obtained. The supplementary files present the clusters presenting the mean values for each of the variables included in the factor analysis. It also displays the results for the analysis of variance (ANOVA), conducted to check the statistical significance of differences between clusters. There are significant differences in the comparison between clusters and ideological side, as shown in the supplementary files.

We use political sides and geography to understand the clusters (Table 4). These variables were not use for the cluster analysis itself. These include for each party the geographical division between Northern, Western, Southern and Central and Eastern Europe; and the ideological side divided between right-wing, left-wing, centrist and populist. Most parties initially classified as left are in clusters 2 and 4; in clusters 2 and 3, we can find the centre and the right parties, and in cluster 1, we find the populist parties. Even if the distribution across different geographical sides seems homogeneous, most of the parties that align with clusters 1 and 3 are in Central and Eastern Europe. The derived clusters are first described based on the variables included in the analysis. The description is then refined based on the structural and demographic variables presented in the previous paragraph together with intentions, which improves the profiling and validation of each cluster.

This cluster is second to smallest, yet it comprises 18% of all parties in Europe. Of these parties, the predominant majority is in Central and Eastern Europe, and it identifies as populist. This cluster is especially interesting because it is the most extremist for what concerns most evaluated variables: all the variables concerning Europe, identity values and decentralisation. The only variables on which this cluster is not extremist are the ones related to economic issues. In other words, this cluster is focused on rejecting the identity value politics and the European Integration project. It refuses any sort of economic extremism and as well any sort of identity politics and cooperation. As a cluster mainly located in Central and Eastern Europe, we can see the Fidesz party in Hungary and the PiS party in Poland as two representative examples. The cluster gives quantitative proof to the often cited assumption that populism in Central and Eastern Europe is economically conservative yet focused on an inflammatory rhetoric against liberal values (Benczes, 2018). It proves that populist parties can be mainly identified in terms of economic conservatism and extremist identity politics in Europe.

This cluster mainly identifies as left-wing. The vast majority of its parties are in Western Europe, yet it is the most present one in Northern Europe and its presence is relevant as well in Central and Eastern Europe and Southern Europe. It is the most numerous clusters in terms of total parties being almost 40% of the total. This cluster can be called conservative because even if it identifies with the left, it is the most moderate in almost all the positions considered. However, this cluster is the one debating the most about the role of European Integration and the one debating the least for what concern anti-establishment versus anti-elite rhetoric and reducing political corruption. It represents a moderate and inclusionary form of politics concerned with migration, which can be considered in line with the overall politics in Northern Europe. For this reason, we can hypothesise that in the contemporary landscape, this is the cluster against which the other ones formed, even if for different reasons: cluster 1 for what concerns identity politics and the European project, cluster 3 for what concern economic issues and cluster 4 for what concerns both economic issues and identity politics.

This cluster is the most present cluster in Southern Europe, but its presence is relevant as well in Central and Eastern Europe. In a traditional ideological divide, it would identify as centrist: in favour of the European integration, liberalism and neutral for what concerns identity issues. Cluster 3 is polarised against cluster 4 for what concerns economic issues and against cluster 1 for what concerns European integration. It is a traditionally centrist and libertarian cluster, yet it looks like an extreme one if compared to clusters 1 and 4. It represents the politics of European integration in a nutshell.

This cluster can easily be identified with the left, and it is mainly located in Western and Southern Europe. Even if it is the smallest cluster, it still represents 15% of all parties in Europe. It is extremist and in favour of the state rather than the free market in all the economic issues, with possibly a profound resentment towards liberalism. The strong importance of economic and identity issues in the parties’ stance makes us hypothesise it as a group potentially in line with intersectional politics: identity, race, gender and class are all part of the same problem that needs active redistribution and attention from the community. It is in line with the more redistributive and leftist stances of Southern Europe, and in general with the rise of left focused primarily on how identity issues impact society.

In this paper, we showed that the number of populist parties is rising (Table 1), and that these parties’ positions consistently differ from all other parties in terms of policy positions. These positions also fundamentally differ in terms of evolved left–right divide rather than geographical position. We showed how these differences are significant for most CHES variables across a left–right divide using ANOVA and presenting the F-tests scores and the significance levels in the supplementary files.

With this analysis we empirically proved that the definition of economic populism does not apply to Europe. Even more interestingly, we proved that there is a definition of populism in terms of policy choices that can be applied to Europe. We can see that the European parties position themselves on an axis that comprehend economic positions on one side and identity politics, European Union, libertarian–authoritarian positions and anti-elite rhetoric and corruption on the other side. This says a lot about today’s contemporary political space and as well about populist parties. In these terms, populist parties are a rejection with the problems created by the Europe Union, identity values and decentralisation. Similar views have already been expressed in other works (Lütz & Kranke, 2014). These views are specifically in line with the concerns Central and Eastern Europe. This region is in fact the one that benefitted the least from the European Union in terms of economic convergence (Győrffy, 2021). This made it possible to trigger the widespread use of an “us versus them” rhetoric against the European Union and social lifestyle such as homosexuality. Looking at it in this perspective, it is particularly worrying the fact that the number of parties of this kind is increasing throughout the years and the related underlying problems are remaining substantially the same. The same process applies for the cluster identified in Southern Europe, even if it goes in a more traditional leftist direction. In policy terms, the solution would be to look at the underlying problems.

In this paper, by showing that there are four statistically different clusters in terms of political rather than geographical side, we show how politics much more than geographical position might influence policymaking. Being these four clusters statistically different from one another, they could very likely transform in different sets of policies in case they will not be contained by any other external factor. For example, the parties that identified with cluster 1 would seem more likely to implement policies that impede the democratic process. In terms of relation between politics and policy, these distinct four clusters also inform us of two things. First, that these different clusters are counterposing Central and Eastern Europe and the rest of Europe that despite their difference not being given by geography. Second, this juxtaposition might create further divergence between Central and Eastern Europe and the rest of Europe that if these different groups of policy preferences translated into different policymaking in these two sides of the continent. The cluster mainly located in Central and Eastern Europe has in fact extremist point of views for what concerns economics, immigration, ethnic minorities, lifestyle, civil liberties and security. This might also prove to be critical for the future of the European Union itself.

We also suggest two other elements: that these four clusters seem to represent a more state of party politics in Europe rather than simple left–right division, and that the two more extreme clusters might be born as an opposition to the two centrist ones, which are more widespread across the continent. While in Western Europe we have the rise of both, we see the rise of right-wing populist in Central and Eastern Europe and the rise of an extreme intersectional left in Southern Europe due to local circumstances and events. This matches with the rise of parties like Podemos, Syriza and the Five Star Movement in Southern Europe and PiS and Fidesz in Poland and Hungary respectively. These examples should not be considered as isolated cases but rather as a potential articulation of a new way of identifying across the political spectrum. However, the overall rise in the number of right-wing and anti-European parties has serious future policy implications for the integrity of the European project. Of course, this problem remains particularly complex one to manage in the European Union where it is already hard to coordinate all the member states and the relative national interests. Further research is therefore needed in looking at the relation between the newly identified populist cluster and their related underlying problems.