How clearly political parties articulate their positions is a foundational element of party competition (Downs Reference Downs1957; Rovny and Polk Reference Rovny and Polk2020). Positional blurring takes place when political actors present vague or conflicting positions on a given dimension (Lefevere Reference Lefevere2023). The practice of varying the clarity of party positions, commonly referred to as positional blurring, is an important party strategy because it can enable parties to attract a broad base of voters (Nasr Reference Nasr2022; Somer-Topcu Reference Somer-Topcu2015). The subject of extensive research in European party systems, positional blurring is a particularly distinctive strategy among populist right parties that present clear cultural positions while blurring their economic positions (Koedam Reference Koedam2021; Rovny Reference Rovny2012; Reference Rovny2013; Rovny and Polk Reference Rovny and Polk2020).
The Freedom Party of Austria (FPÖ) under Jörg Haider’s leadership exemplifies economic positional blurring: it combined traditionally left and right economic positions by invoking terms like “fair market economy” (faire Marktwirtschaft) while still not fully embracing free-market policies (Mudde Reference Mudde2007, 123–25). Haider made this strategy explicit in the title of this Reference Haider1997 book, Liberated Future beyond Left and Right: Human Alternatives for a Bridge into the New Millennium. In Latin America, sociocultural, rather than economic, blurring has occurred more frequently. Recently, MORENA under López Obrador held unambiguous left-wing economic, positions while deliberately avoiding clear stances on abortion and LGBT rights. That the governing coalition included ultraconservative parties made this ambiguity electorally rational, though it sat uneasily alongside the government’s self-presentation as feminist (Beer Reference Beer2021; Cornejo Reference Cornejo, Kaltwasser, Meléndez, Tanscheit and Zanotti2026; Díaz Domínguez Reference Díaz Domínguez2021).
Despite the wealth of research on positional blurring in European party systems, as well as some limited research in the United States (Bartels Reference Bartels1986; Milita et al. Reference Milita, Simas, Ryan and Krupnikov2017), this research field has yet to be applied more broadly or grappled with the implications of shifting geographic contexts. Do the core insights of existing research apply broadly, or are they context specific? To evaluate this question, we use new expert survey data in Latin America alongside existing data on Europe to expand the empirical scope of this field of study and enable cross-regional comparisons (Martínez-Gallardo et al. Reference Martínez-Gallardo, de la Cerda, Hartlyn, Hooghe, Marks and Bakker2022). This article takes a crucial step to answering these questions by distilling key insights from existing work in the European context and theorizing how those insights are likely to manifest in the Latin American context.
One clear finding from European research is the connection between populism and positional blurring. Radical-right populists, in particular, have engaged in blurring their economic positions, but they are not the only ones to do so. Other populists, like the Five Star Movement who are considered populists without being clearly left or right (Rooduijn et al. Reference Rooduijn, Andrea, Halikiopoulou, Froio, Van Kessel, De Lange, Mudde and Taggart2024), also have blurred economic stances. Populism is a thin-centered ideology, meaning that it has a general orientation toward society without fixed policy content. It is characterized by a Manichean worldview pitting the pure people against the corrupt elite and endorsing the view that politics should be the enactment of the general will (Mudde and Rovira Kaltwasser Reference Mudde and Kaltwasser2017). However, that perspective can justify widely varying electoral and policy programs. Populism is a valuable starting point to understanding cross-regional positional blurring precisely because it is flexible. Does populism in Europe correlate with blurred economic positions because of something inherent to populism, or is that blurring a result of other features?
We argue that these populist blurring patterns are context-dependent and respond to two moderating factors that intercede between the populist thin-centered ideology and party strategy. First, the dimensional structure of political competition matters. In multidimensional party systems, common in much of Western Europe, blurring on the economic dimension is electorally beneficial (Bakker, Jolly, and Polk Reference Bakker, Jolly and Polk2012), allowing populist parties to emphasize competition on cultural issues while being vague about economic policy (De Vries and Hobolt Reference De Vries and Hobolt2020; Mudde Reference Mudde2019; Mudde and Kaltwasser Reference Mudde and Kaltwasser2013; Van de Wardt, De Vries, and Hobolt Reference Van de Wardt, De Vries and Hobolt2014). This strategy is a specific manifestation of the broader insight that parties tend to blur their secondary dimensions while clarifying their primary dimensions (Koedam Reference Koedam2021). However, in party systems where political competition runs along a single ideological dimension, such as in Latin America, and where party politics is generally anchored around economic issues (Martínez-Gallardo et al. Reference Martínez-Gallardo, de la Cerda, Hartlyn, Hooghe, Marks and Bakker2022), ambiguous economic positions can be costly because they would becloud parties’ positions on their primary dimension of competition (Koedam Reference Koedam2021; Rovny Reference Rovny2012). If there is no multidimensional competition, there is no second dimension to blur.
Second, we argue that we can only understand populists’ strategic decisions in light of the interaction between populism and thick ideologies (Hunger and Paxton Reference Hunger and Paxton2022). In this context, we understand thick ideologies as comprehensive sets of policy positions that provide substantive content to a party’s platform beyond the core populist message. How populism and thick ideologies interact can take different forms depending on how the party conceptualizes “the people.” When populism has an exclusionary nature, it adopts a nativist thick ideology that characterizes “the people” as exclusive and distinct from non-native groups like noncitizens, foreigners, and ethnic minorities—those seen as deviating from traditional ways of life. This thick ideology leads to policy stances that emphasize restricting material and political benefits to the pure people. In contrast, an inclusive thick ideology defines “the people” broadly and as structured primarily by economic class (Mudde and Kaltwasser Reference Mudde and Kaltwasser2013). This thick–thin combination generates a politics of expanding material benefits and political inclusion to a broader set of the population. Such inclusive thick ideologies are prevalent among Latin American populists and left-wing populists in Europe.
Although the (multi)dimensionality of party competition makes blurring a viable electoral strategy—downplaying one dimension while making another one salient—we argue it is the connection between populism and the host ideology that determines which dimension populists choose to blur. That is, the type of thick ideology determines which issues are primary and which are not. This insight suggests that exclusive populists generally blur on economic issues to focus their appeals on cultural issues, aligning with their ideological emphasis on nativism and cultural exclusivity. In contrast, inclusionary populism typically leads parties to stake out clear economic positions, reflecting their focus on class-based definitions of “the people.”
We test our argument using data from the Chapel Hill Expert Survey in Europe and Latin America; the latter survey include novel items on party positional blurring on both economic and cultural issues, which we rigorously validate for the first time in Latin America (de la Cerda & Gunderson Reference De la Cerda and Gunderson2026; Jolly et al. Reference Jolly, Bakker, Hooghe, Marks, Polk, Rovny, Steenbergen and Vachudova2022; Martínez-Gallardo et al. Reference Martínez-Gallardo, de la Cerda, Hartlyn, Hooghe, Marks and Bakker2022). Although aggregate levels of blurring between the regions are quite similar, we find significant differences between Europe and Latin on our key moderators. Europe generally has higher-dimensional party systems, and populist parties tend to be more exclusive than in Latin America. We then directly test our moderators—party–system dimensionality and populist type (inclusive vs. exclusive) that captures the interaction between populists’ thick and thin ideology. We find robust results that are consistent with our party-system dimensionality and thick ideology hypotheses.
As the first study to consider the relationship between populism and positional blurring using data beyond Europe, our findings have broad implications for the study of populism and party competition. We provide the first descriptive analysis of positional blurring in Latin America. We also show that the state-of-the-art research in this area, although it draws on the same features, leads to divergent predictions of populist party strategy in different competitive contexts. Crucially, the two moderating factors we highlight reinforce the point that although populism is global, it is not homogeneous, and neither are the strategies of its proponents. To understand the successes of populism (and its strategic articulation), we need to incorporate the structure of party systems and the thick ideology it attaches to into our analysis (Hunger and Paxton Reference Hunger and Paxton2022) because these two factors alter populism’s strategic opportunities and constraints.
In the next section, we briefly summarize our understanding and conceptualization of populism before developing our dimensionality and thick ideology hypotheses.
Populism in Context
Our approach is based on the ideational understanding of populism, which defines it as a discourse or thin-centered ideology with a narrow set of key characteristics (Hawkins and Rovira-Kaltwasser Reference Hawkins, Rovira-Kaltwasser, Hawkins, Carlin, Littvay and Rovira-Kaltwasser2019). Embracing a minimal definition of populism allows us to systematically compare levels of populism cross-nationally and regionally. As has been highlighted elsewhere, the wide variety of ways in which populism has been defined have hindered the ability of cross-national comparisons to effectively build on a shared conceptual foundation (Mudde and Kaltwasser Reference Mudde and Kaltwasser2013).
The ideational definition specifies three characteristics of populism (Hawkins and Rovira-Kaltwasser Reference Hawkins, Rovira-Kaltwasser, Hawkins, Carlin, Littvay and Rovira-Kaltwasser2019; Mudde Reference Mudde2004). First, populist discourse is Manichaean in that it involves a stark dualism in which the world is conceptualized in terms of a cosmic struggle between good and evil. Second, in this cosmic struggle, good is identified with the will of the people; accordingly, politics should be the concrete (reified) version of the will of the people. Third, the elite is the evil side of this radical distinction: it is characterized by populist discourse in conspiratorial terms as a group that consciously works against the people for its own selfish interests. In summary, populism is dualistic, people-centric, and anti-elitist (Hawkins Reference Hawkins2009).
Populist parties combine populism with thick or host ideologies that can differ widely across and within regions (Neuner and Wratil Reference Neuner and Wratil2022). These ideologies provide policy content to otherwise flexible, populist thin ideologies. In this article, we focus on the distinction between inclusive and exclusive thick ideologies. Systematized by Mudde and Kaltwasser (Reference Mudde and Kaltwasser2013, 148–66), this distinction separates two key forms of populism in Latin America and Europe along dimensions of access to material benefits, access to the political system, and membership in the symbolic community. Populism definitionally requires making a distinction between “the pure people” and “the corrupt elite.” Inclusive populism is associated with socioeconomic policies that seek to include and improve the quality of life of the poor and politically marginalized; this type of thick ideology can be found among radical-left and some green parties. In this case, “the people” are characterized as those of lower socioeconomic status, whereas “the elite” are associated with the upper class and international organizations like the IMF. Inclusive populism, notably, approaches this division with the aim of expanding the population of pure people by facilitating their material well-being and political engagement and the strength of the symbolic community; therefore, it is inclusive in that sense.
In contrast, exclusive populism adopts nativist positions that distinguish “the people” from non-native groups including noncitizens, foreigners, ethnic minorities, and the mainstream politicians supportive of those groups. In terms of material dimensions, exclusive populism emphasizes the restriction of material benefits (particularly state benefits) to the native population through policies of welfare chauvinism (Careja and Harris Reference Careja and Harris2022). In terms of political engagement and responsiveness, the dominant theme here is that mainstream politicians are foreclosing key issues, such as the policies of the European Union, from public debate and limiting the people’s ability to influence their own country’s laws; these parties embrace direct democracy measures as ways for the people to directly make policies, commonly in areas related to immigration and international agreements that reinforce the purity and sovereignty of the people (Mudde and Kaltwasser Reference Mudde and Kaltwasser2013, 163). Exclusive populists have also opposed the expansion of civil and political rights to groups, such as immigrants and Muslims, who exist on the margins (163–64).
Although neither type of populism is exclusive to a particular region in the world, inclusive populism has been historically more common and influential in Latin America, whereas exclusive populism has been more prevalent in Europe (Mudde and Rovira Kaltwasser Reference Mudde and Kaltwasser2017; Rooduijn et al. Reference Rooduijn, Andrea, Halikiopoulou, Froio, Van Kessel, De Lange, Mudde and Taggart2024, 974). For example, Chávez and Morales used extensive anti-imperialist rhetoric that emphasized a common Latin American identity, which sharply contrasts with chauvinistic populist programs in Europe that seek to exclude immigrants from welfare provisions (Careja and Harris Reference Careja and Harris2022; Mudde and Rovira-Kaltwasser Reference Mudde and Kaltwasser2013).
These regional patterns are evident, but there is important intra-regional and temporal variation. In Latin America, the latest wave of populism is characterized by a new set of actors who differ significantly from earlier ones like Chávez and Morales. Populist leaders, like Bolsonaro in Brazil (2019–22), Javier Milei in Argentina (2023–), and Nayib Bukele in El Salvador (2019–), lack the inclusive nature common to past populist movements in the region. Milei, for instance, combines a Manichean anti-elite discourse with a libertarian economic program and cultural conservatism that have more in common with European nativist populism than with the class-based, redistributive appeals of his regional predecessors. Intra-regional variation is not unique to Latin America: in Europe, left-wing populist movements such as Podemos in Spain and Syriza in Greece strongly emphasize socioeconomic inclusivity while presenting correspondingly clear economic positions.
Ideology, Dimensionality, and Party Strategy
We distill generalizable principles from existing research to argue that differences in populist electoral strategy across regions stem from two factors: the dimensionality of the party system in which they compete and the ideological character of the movements themselves (inclusionary vs. exclusionary).Footnote 1 In particular, we focus on positional blurring: the extent to which political parties’ electoral appeals on an issue do not indicate a clear position (Koedam Reference Koedam2021; Lefevere Reference Lefevere2023; Nasr Reference Nasr2022; Praprotnik and Ennser-Jedenastik Reference Praprotnik and Ennser-Jedenastik2023; Rovny Reference Rovny2012).Footnote 2 Blurring is a particularly important element of party strategy, given its distinctive use among right-wing populists in Europe (Rovny Reference Rovny2012).
Existing research on the populist radical right in Europe argues that these parties and actors strategically and distinctly obscure their positions on economic issues while presenting very clear positions on cultural issues (Rovny Reference Rovny2013; Rovny and Polk Reference Rovny and Polk2020). Yet, positional blurring is hardly unique or exclusive to populist parties (Breyer Reference Breyer2022; Somer-Topcu Reference Somer-Topcu2015), nor is it universally successful, with studies from different contexts and methods coming to conflicting conclusions (Bartels Reference Bartels1986; Cahill and Stone Reference Cahill and Stone2018; Downs Reference Downs1957; Glazer Reference Glazer1990; Gunderson Reference Gunderson2024; Martin Reference Martin2019; Milita et al. Reference Milita, Simas, Ryan and Krupnikov2017; Tomz and Van Houweling Reference Tomz and Van Houweling2009). However, populist parties in Europe frequently deploy this strategy with a particularly clear and often effective logic tied to how their thin-centered ideology attaches to thicker ideologies and the strategic possibilities of their party system.
Party System Dimensionality
We argue that multidimensional party systems facilitate differential positional blurring. As challenger parties seek to break into parliament and the government, it behooves populist parties to emphasize issues that split mainstream parties’ existing electoral coalitions (Van de Wardt, De Vries, and Hobolt Reference Van de Wardt, De Vries and Hobolt2014). The economic issues and the class cleavage—the primary division separating mainstream parties of the right and left and the conflict that generally led to the construction of the left–right spectrum in most European and Latin American party systems— divide society by economic status (Bornschier Reference Bornschier2009; Lipset and Rokkan Reference Lipset and Rokkan1967).
However, these coalitions often include groups with divergent positions on cultural issues (Hooghe and Marks Reference Hooghe and Marks2018). On the right, for example, those with traditional cultural views are in coalition with mercantile interests favoring the freer movement of people and, in some cases, laissez-faire cultural stances. On the left, highly educated sociocultural professionals are in coalition with manual workers, who generally have less progressive cultural attitudes (Oesch and Rennwald Reference Oesch and Rennwald2018, 795). So long as politics revolves around economic issues and the size of government, these groupings are coherent. When party systems become increasingly multidimensional, as they have in many European countries (Bakker, Jolly, and Polk Reference Bakker, Jolly and Polk2012), the internal divisions within their left and right blocs become vulnerabilities.
For populists, this presents an opportunity to use cultural conflict as a wedge to divide the dominant parties’ coalitions (De Vries and Hobolt Reference De Vries and Hobolt2020; Van de Wardt, De Vries, and Hobolt Reference Van de Wardt, De Vries and Hobolt2014). Indeed, that manual workers and small business owners are two target demographics with starkly divergent economic interests demonstrates that populists’ strategic opportunity lies in shifting the political agenda to cultural issues, immigration, and other issues relevant to their combination of populism and nativism (Oesch and Rennwald Reference Oesch and Rennwald2018; Schattschneider Reference Schattschneider1975).
These examples are grounded in the increasingly prevalent multidimensional European experience. In sharp contrast, party competition in Latin America, as well as in some European party systems, is mostly unidimensional. In these contexts, positions on the economic dimension are strongly related to sociocultural stances: left-wing parties tend to be socioculturally liberal, whereas right-wing parties lean toward social conservatism. This unidimensionality largely stems from the relatively low salience of cultural issues in these contexts: although there is, of course, some discussion of cultural issues, economic issues dominate. Moreover, where cultural issues and ethnic divisions do emerge, they reinforce—rather than crosscut—the dominant economic line of conflict (Martínez-Gallardo et al. Reference Martínez-Gallardo, de la Cerda, Hartlyn, Hooghe, Marks and Bakker2022). The impact of putting cultural issues on the political agenda in multidimensional systems stems from these issues’ abilities to divide mainstream parties’ electorates (Hooghe and Marks Reference Hooghe and Marks2018; Schattschneider Reference Schattschneider1975). In Latin American party systems, however, cultural issues have yet to undo long-standing political coalitions.
Given the importance of left–right economic competition in unidimensional systems, substantial blurring on this dimension is a less viable option. In these party systems, uncertainty about parties’ positions can decrease political support in the same way that spatial distance between candidates and voters does (Enelow and Hinich Reference Enelow and Hinich1989; Rovny Reference Rovny2012). Additionally, one of the primary benefits of positional obfuscation is to downplay the importance of an issue dimension, which is useful to a party if it wants to emphasize another dimension. However, if there is only one dimension at play and the party wants to mobilize voters for whom that dimension is crucial, blurred positions would be counterproductive. A program of aggressive clarity is likely more suitable.
The implication for our argument is that multidimensionality creates a competitive structure in which populists use positional blurring to emphasize a dimension of conflict that crosscuts the existing line of conflict while obscuring its position on the mainstream’s key dimension (normally economics). When only one primary dimension is at play, we argue that positional blurring on the economy becomes a less attractive strategy to cultivate a distinct image. This results in the following two expectations:
Dimensionality Hypothesis (Economy): Populism is positively related to positional blurring on economic issues in multidimensional party systems, but it is negatively related in unidimensional party systems.
Dimensionality Hypothesis (Sociocultural): Populism is positively related to positional blurring on sociocultural issues in multidimensional party systems, but it is negatively (or not related) related in unidimensional party systems.
Notably, whereas multidimensional party systems offer populists greater strategic flexibility, they do not automatically determine what dimension to emphasize or blur. To uncover populism’s full relationship with positional blurring, we argue that it is also essential to consider our second key moderator: the interaction between populism and a thick host ideology.
Thick Ideology
As discussed earlier, populism is a thin-centered ideology that commonly attaches itself to more comprehensive belief systems. We argue that this thick–thin connection plays a crucial role in determining whether populists will blur their economic position. Exclusive populists attach their Manichean worldview to nationalism and nativism, resulting in a definition of “the people” as nationally homogeneous and threatened by immigration and multiculturalism more broadly (Mudde Reference Mudde2019). An ethnonationalist conception of the pure people logically requires that these parties emphasize and build their brand on cultural issues as their primary dimension of competition and distinction (Lupu Reference Lupu2016; Rovny Reference Rovny2013).
However, no such incentive exists on economic issues for exclusive populists (Han Reference Han2020). If anything, economic issues are secondary for these parties, and economic debates have the potential to divide the national community on class lines in a time of globalized capitalism (Robison et al. Reference Robison, Stubager, Thau and Tilley2021). Therefore, we argue that the confluence of populism and nativism pushes populists to a politics that seeks to suppress class divisions in favor of consolidating support on ethnic appeals. As a result, these parties have good incentives to avoid presenting clear positions on economic issues or emphasizing them in their election campaigns (Koedam Reference Koedam2021), lest they lead the electorate to prioritize their economic identities over their national and ethnic identities. Put differently, if populists’ thick ideologies structure their approach to certain issues as we expect, then exclusive populists should attempt to avoid constructing a clear party image on economic issues by blurring their economic positions and making them less salient. For example, the Austrian Freedom Party (FPÖ), which is very populist (8.2/10) and very nativist (9.8/10 on immigration policy) and places very low salience on economic issues (4.2/10, well below the first quartile value of 5.2), has a score of 7 out of 10 on the blurriness of its economic positions, which is one of the highest blurring scores on economic issues in the 2019 CHES Survey (Bakker et al. Reference Bakker, Hooghe, Jolly, Marks, Polk, Rovny, Steenbergen and Vachudova2020).
In contrast, we argue that the combination of populism with an inclusive thick ideology will result in a type of populism that does blur its economic positions. Latin America is significantly less equal both politically and economically than most European countries (Huber and Stephens Reference Huber and Stephens2012), so it is no surprise that anti-elite rhetoric is commonly aimed at the state, multinational corporations, and the global financial system (Mudde and Rovira Kaltwasser Reference Mudde and Kaltwasser2017). Inclusive populism, therefore, defines political conflict in terms of class, emphasizing the material inclusion of the poor and more alternative forms of political participation that can truly represent the voice of those who are left behind (Mudde and Kaltwasser Reference Mudde and Kaltwasser2013).
Indeed, this form of inclusive populism is also a feature of the European populist landscape, found in radical-left parties and some green parties (like the Socialist Party in the Netherlands, the Green Party of England and Wales, and the Red-Green Alliance in Denmark), despite inclusive populism tending to be more successful in the Latin American context. In many instances, these parties adopt a populist approach to politics that emphasizes the incorporation of (economically) marginalized groups into society, rather than the exclusive nativism of the radical right. Because this thick ideology comes with substantive commitments to economic issues, these parties should be particularly unlikely to blur their economic positions. In this ideological context, campaigning on blurred economic positions would be as anomalous as a nativist populist refusing to provide clear positions on immigration. These are the motivating conflicts underlying the types of populism’s host ideologies, so watering down these positions by blurring them would undercut these parties’ raison d’être.
Therefore, we expect that the relationship between populism and the clarity of economic positions will vary by thick ideologies; that is, it should lead to blurrier economic positions when populism is attached to an exclusive or nativist thick ideology but not when it is attached to an inclusive thick ideology:
Thick Ideology Hypothesis (Economic Issues): Populism is positively related to positional blurring on economic issues when it is exclusionary but is negatively related to positional blurring on economic issues when it is inclusionary.
Thick Ideology Hypothesis (Sociocultural Issues): Populism is negatively related to positional blurring on sociocultural issues when it is exclusionary but is positively related to positional blurring on sociocultural issues when it is inclusionary.
Similarly, we expect populists with exclusive thick ideologies to be less likely to blur their sociocultural positions because those are exactly the issues on which they built their identity. In contrast, inclusive populists should blur sociocultural positions more to develop their party reputations based on economic inclusivity.
Empirical Approach
To study the relationship between populism and blurring strategies across both Europe and Latin America and to capture a wide range of political contexts, we turned to the 2019 Chapel Hill Expert Survey (CHES), which contains information on parties’ ideological and policy positions. Importantly, CHES estimates party positions based on several facets of parties’ ideologies and electoral strategies, including the extent of blurring, distribution of issue salience, and the degree of populism. Recently, CHES substantially expanded its regional coverage from Europe to include 12 Latin American countries, making it an ideal dataset to test the proposed hypotheses across the two regions (Bakker et al. Reference Bakker, De Vries, Edwards, Hooghe, Jolly, Marks and Polk2015). The Chapel Hill Expert Survey in Latin America (CHES-LA, fielded in 2020) has been rigorously validated and provides data that are comparable to the European dataset (De la Cerda et al. Reference De la Cerda, Bakker, Jolly, Koedam, Dassonneville, Leslie, Polk, Sheppard and Zur2025; Martínez-Gallardo et al. Reference Martínez-Gallardo, de la Cerda, Hartlyn, Hooghe, Marks and Bakker2022).
One of the main advantages of using the CHES dataset is that it contains direct measures of perceived positional blurring. Experts directly evaluate the extent to which political parties (and presidents in Latin America) blur their economic and sociocultural positions by taking vague, ambiguous, or contradictory stances on economic and sociocultural issues.Footnote 3 This measure has been previously validated in the European context by comparing it to expert uncertainty around parties’ positions, measured as the standard deviation of experts’ placements (Rovny and Polk Reference Rovny and Polk2020). This is a substantial improvement over the standard deviation approach to measuring positional blurring, which is subject to a number of factors that lie outside the core concept (Rovny Reference Rovny2012). For example, the standard deviation could be high because a party truly is blurring its positions or because some experts just do not know much about that party. The direct question (paired with a separate question on internal dissent) helps ensure that it specifically taps the concept of interest.
It is worth noting that even this direct question cannot definitively determine whether blurring behavior is an intentional strategic behavior. No existing measure can provide such insight into the minds of party leaders. Given the number of robust findings relating positional blurring to strategic party interests with implications for voting behavior (Han Reference Han2020; Koedam Reference Koedam2021; Rovny Reference Rovny2013; Rovny and Polk Reference Rovny and Polk2020), we proceed under the assumption that blurring is at least partially a response to strategic incentives.
However, blurring measures have not yet been validated in the Latin American context. Section A3 of the supplementary material contains validation checks of the data along the criteria recommended by Adcock and Collier (Reference Adcock and Collier2001). In summary, we find that CHES’s positional blurring question is positively associated with alternative expert-survey–based measures and with intraparty disagreement among Latin American parliamentarians (Alcántara Sáez, Reference Alcántara Sáez2018).Footnote 4 We also find that the relationships among the direct blurring question, issue salience, and issue position are nearly identical in Latin America and Europe. We interpret these results as strongly indicative of the validity of CHES’s direct blurring question in Latin America.
Our analysis also requires a measure of populism. There have been several recent advances in identifying populist parties and measuring populism. The PopuList and POPPA experts surveys are two prime examples (Meijers and Zaslove Reference Meijers and Zaslove2021; Rooduijn et al. Reference Rooduijn, Andrea, Halikiopoulou, Froio, Van Kessel, De Lange, Mudde and Taggart2024). Unfortunately, we are unable to use either data source because they do not include Latin America.
Therefore, we rely on CHES’s people-versus-the-elite position question. This question is designed to measure the extent to which political parties believe that “the people” are in opposition to “the elite” and that “the people” should make the most important political decisions.Footnote 5 We rely on this single indicator because it captures the central component of populism, and it is the only comparable question in both the European and Latin American surveys. However, in the supplementary material we provide additional validation demonstrating that this measure is strongly associated with more complex multidimensional measures of populism used elsewhere (De la Cerda, Hartlyn, and Martinez-Gallardo Reference De la Cerda, Hartlyn and Martinez-Gallardo2024; De la Cerda, Martínez-Gallardo, and Hartlyn Reference De la Cerda, Martínez-Gallardo and Hartlyn2025).Footnote 6 Moreover, previous research in the European context has shown that CHES’s populism measures are highly correlated with more comprehensive measures of populism from POPPA (Hooghe et al. Reference Hooghe, Marks, Bakker, Jolly, Polk, Rovny, Steenbergen and Vachudova2024; Meijers and Zaslove Reference Meijers and Zaslove2021).
In addition to these variables, we also include several control variables grounded in existing research (Rovny Reference Rovny2012). First, we control for the absolute value of party positions on the economic and social cultural dimensions. As shown in figure A2, there is an inverse-U shaped relationship between positional blurring and a party’s position on both dimensions where parties have clearer positions on the fringes of ideological space. At the center of the distribution, some party positions are very blurry, whereas others are quite clear. Controlling for the absolute value of the position ensures that we are picking up on blurring, rather than just positional differences.
We use a similar rationale to control for the salience of both dimensions. As discussed earlier, the more salient a dimension is, the less parties blur that dimension. Some research also considers low salience itself as a blurring technique (Koedam Reference Koedam2021). Because we want to understand the effect of populism type itself, controlling for issue salience partials out the component of the relationship between populism type and blurring that is due to salience. Salience is also subject to several other factors, including government participation and problem information, like media coverage and expert reports, in society (Greene Reference Greene2016; Green-Pedersen Reference Green-Pedersen2019), which may have their own implications for positional blurring. In models predicting blurring on one dimension, we also control for positional blurring on the other.
Given the nested nature of the data, we use hierarchical linear models to assess the relationship between populism and blurring in Europe and Latin America and thereby test our hypotheses. Although this empirical design is unable to identify causal effects, it is well suited to showing whether our hypothesized relationships arise under the expected circumstances.
In the next section, we proceed by presenting comparative descriptive statistics on blurring in Latin America and Europe. We then discuss the results of a regression model incorporating an interaction between region and our populism variable that shows the anticipated divergence of populism’s effect on positional blurring in the two regions. Following this baseline, we introduce our moderators—party system dimensionality and thick ideology—and incorporate them into the model to test our dimensionality and thick ideology hypotheses.
Analyses
Patterns of Blurring across Regions
We begin with some descriptive statistics on positional blurring at the regional and country levels. As discussed earlier, regions do not perfectly separate our moderating variables, but we present descriptives by region and country to help anchor the reader in the variation we aim to explain, particularly in the Latin American cases where we present these measures for the first time.
Figure 1 contains density plots of positional blurring on economic and sociocultural issues. Average values are quite similar across the two regions, though there are some differences in the shapes of the distribution. On economic issues, average positional blurring is not statistically different across the two regions: 3.602 (95% CI 3.399–3.946) in Europe and 3.893 (95% CI 3.601–4.185) in Latin America. On sociocultural issues, Latin American parties (3.644 [3.342–3.946]) tend to blur more than European parties (2.87 [2.679–3.069]) but only by less than 1 point on an 11-point scale. The sociocultural dimension of conflict is less salient in Latin America (Martínez-Gallardo et al. Reference Martínez-Gallardo, de la Cerda, Hartlyn, Hooghe, Marks and Bakker2022), so we should expect parties to take less clear positions compared to Europe, where it is more central. In both cases, the distributions’ central tendency falls well below the midpoint of the overall scale, suggesting that parties that blur to a significant extent are not conforming to an overall trend in their regions. These distributions also indicate that we should be cautious in assuming the extent of positional blurring at the party level based solely on region. The general tendency toward relatively low levels of blurring also supports our emphasis on the importance of evaluating blurring strategies among the parties we think do so exceptionally, namely populist parties.
Blurring in Europe and Latin America
Note: 95% Confident Intervals.
Figure 1 Long description
A multi-panel density plot with two stacked horizontal panels. The x-axis is labeled Blurring, ranging from 0 (Not Blurred) to 10 (Extremely Blurred). The y-axis represents density from 0.0 to 0.3.
Legend at the bottom:
* Light blue shaded area: Left-Right Economic.
* White area with black outline: Socio-Cultural.
Top Panel: Europe
* Socio-Cultural curve: Peaks early around 2.5 with a density of approximately 0.22, then tapers off.
* Left-Right Economic curve: Peaks later around 4.0 with a density of approximately 0.22.
* Above the curves, two horizontal error bars represent 95 percent Confidence Intervals. The black dot (Socio-Cultural) is centered near 3.2. The blue dot (Economic) is centered higher, near 3.8.
Bottom Panel: Latin America
* Socio-Cultural curve: Peaks around 3.2 with a density of approximately 0.24.
* Left-Right Economic curve: Peaks slightly later around 4.0 with a density of approximately 0.25.
* The 95 percent Confidence Interval dots are much closer together than in the Europe panel. The black dot is near 3.7 and the blue dot is near 3.9, with overlapping error bars.
Focusing on the economic dimension of competition, figure 2 plots blurring by country, ordered from the one with the highest average level of blurring (Bulgaria) to the one with the lowest average level of blurring (Iceland). The key takeaway from this figure is that Latin American countries (in black) and European countries (in blue) are distributed quite evenly. If anything, the Latin American countries tend to cluster in the middle of the scale.
Levels of Blurring on Economic Issues by Country
Note: Darker points with error bars represent the average degree of blurring in the country with 95% confidence intervals. The faded points show the positions of the parties in each country. The dashed line indicates the average level of blurring in the entire sample.
Figure 2 Long description
A dot plot displays the degree of blurring on economic issues by country. The horizontal X axis is labeled Blurring, ranging from 0 (Not Blurred) to 10 (Extremely Blurred). The vertical Y axis lists 42 countries. A vertical dashed line at approximately 3.7 indicates the sample average.
Data points are color-coded: blue for Europe and black for Latin America. Darker central dots represent the country average with horizontal error bars indicating 95 percent confidence intervals. Faded dots represent individual party positions.
From top to bottom, countries with the highest blurring levels (positioned furthest right) include Bulgaria, Cyprus, and Mexico. Countries near the sample average include Germany, Slovakia, and Ireland. Countries with the lowest blurring levels (positioned furthest left) include Switzerland, Belgium, and Iceland, with Iceland being the closest to 0.
Latin American countries (black) like Mexico, Peru, and Colombia generally show higher blurring levels than the average, while European countries (blue) are distributed across the entire spectrum, with a high concentration of Nordic and Western European nations showing lower-than-average blurring.
Patterns of Blurring in Europe and Latin America
Table 1 reports the results of the hierarchical linear models predicting blurring. Blurring on economic issues is the dependent variable in models 1 through 4. Blurring on sociocultural issues is the dependent variable in models 5 through 8. Models 1 and 5 only contain random country intercepts. These models allow us to establish how much variation takes place at the country level by using the conditional
$ {R}^2 $
, which accounts for all variables included in a model; the marginal
$ {R}^2 $
ignores variation explained by the random intercepts. Models 1 and 5 show that around 20% of the variation takes place at the country level and 80% at the party level. In other words, most of the differences in blurring strategies can be attributed to differences across parties, not across countries.
Hierarchical Linear Models Estimates of Region, Dimensionality, and Populist Type on Economic and Sociocultural Blurring
Table 1 Long description
The table presents Hierarchical Linear Models estimates. The dependent variables are split into two main sections: Left-right economic blurring (Models 1 through 4) and Sociocultural blurring (Models 5 through 8).
Key independent variables and their coefficients include:
* Left-right economic (absolute value): Significant negative coefficients around -0.56 in models 2, 3, and 4; non-significant positive values near 0.01 in models 6, 7, and 8.
* Left-right salience: Significant negative coefficients around -0.32 in models 2, 3, and 4; non-significant positive values near 0.03 in models 6, 7, and 8.
* Left-right blurring: Significant positive coefficients around 0.26 in models 6, 7, and 8.
* Sociocultural issues (absolute value): Significant positive coefficients (0.16 to 0.20) for economic blurring and significant negative coefficients (-0.34 to -0.36) for sociocultural blurring.
* Sociocultural salience: Significant negative coefficients around -0.53 for sociocultural blurring.
* Sociocultural blurring: Significant positive coefficients around 0.30 for economic blurring.
* Populism: Significant negative coefficient of -0.108 in model 2 and positive 0.075 in model 3 for economic blurring; significant negative -0.068 in model 7 for sociocultural blurring.
* Exclusionary (ref. Inclusionary): Significant positive coefficients (0.635 and 0.609) in models 2 and 3; and 0.435 and 0.447 in models 6 and 7.
* Europe (ref. Latin America): Significant negative coefficient of -1.543 in model 2.
Interaction terms:
* Populism Europe: 0.269 in model 2 and -0.131 in model 6.
* Populism dimensionality: 0.130 in model 3.
* Populism exclusionary: 0.190 in model 4.
Model statistics at the bottom show Conditional R super 2 values ranging from 0.191 to 0.673 and Marginal R super 2 values ranging from 0 to 0.611. The number of observations is between 383 and 394. Asterisks indicate p < 0.05.
* p<0.05.
Models 2 and 6 add our control variables and our measure of populism interacted with regions. The interaction term between region and populism allows us to evaluate the extent to which the relationship between populism and blurring is specific to the region, rather than common across regions. Models 2 and 6 in table 1 confirm this intuition. The interaction term between populism and Europe is positive and significant for economic blurring and negative and significant for sociocultural blurring. This initial finding provides the first indication that populism’s relationship with blurring strategy responds to contextual factors.
To convey the substantive implications of these results, we plot predicted values and marginal effects of blurring on both dimensions in figure 3 (Brambor, Clark, and Golder Reference Brambor, Clark and Golder2006). Panel A shows predicted values of blurring on economic issues, with a rug plot along the x-axis to display the observed values of populism. In Europe, as the level of populism increases, blurring on the economic dimension increases. The opposite is true in Latin America: parties that have high levels of populism tend to blur less on this dimension. Panel B displays the marginal effect of populism in each region and shows consistent results: the effect of populism on economic blurring is positive in Europe but is negative in Latin America.
Regional Differences in Populist Blurring on Economic and Sociocultural Issues
Note: The left panels (A and C) present the predicted values of populism on economic (A) and sociocultural (C) blurring for Europe and Latin America. The right panels (B and D) present the marginal effect of populism on economic (B) and sociocultural (D) blurring. 95% confidence intervals estimated via Monte Carlo simulation. The rug plot along the x-axis on the left panels (A and C) indicates the observed values of populism in our sample.
Figure 3 Long description
The figure consists of four panels arranged in a two-by-two grid.
Panel A, top-left, is a line graph showing Left-Right Economic Blurring on the y-axis from 0 Low to 10 High, and Populism on the x-axis from 0 Low to 10 High. A blue line for Europe shows a linear increase from approximately 3 to 5. A grey line for Latin America shows a slight linear decrease from approximately 4.5 to 4. Shaded 95 percent confidence intervals surround both lines. A dense rug plot along the x-axis indicates data concentration between 1 and 9.
Panel B, top-right, is a marginal effect plot. The y-axis is Populism Marginal Effect from negative 0.4 to 0.4, with a horizontal reference line at 0. For Europe, a point estimate with error bars is located at approximately positive 0.18, entirely above the zero line. For Latin America, the point estimate is at approximately negative 0.1, with error bars crossing the zero line.
Panel C, bottom-left, is a line graph showing Socio-Cultural Blurring on the y-axis from 0 Low to 10 High, and Populism on the x-axis from 0 Low to 10 High. The blue line for Europe shows a linear decrease from approximately 3.8 to 2.5. The grey line for Latin America is nearly flat, remaining around 3.5. Shaded 95 percent confidence intervals are included.
Panel D, bottom-right, is a marginal effect plot for sociocultural blurring. For Europe, the point estimate is at approximately negative 0.12, with error bars entirely below the zero line. For Latin America, the point estimate is near 0.02, with error bars crossing the zero line.
Panels C and D illustrate the relationship among populism, region, and blurring on the sociocultural dimension. The patterns observed for this dimension diverge from the ones for the economic one. Although higher levels of populism in Europe decrease blurring on the sociocultural dimension, populism does not have a statistically significant effect on sociocultural blurring in Latin America. Panel D clearly shows that the effect of populism in Europe has a negative significant effect, whereas the effect of populism on sociocultural blurring in Latin America is undistinguishable from zero, which is not surprising given the low salience of sociocultural issues in Latin American party competition.
In summary, there are important observable differences in blurring strategies across regions, even when controlling for an important set of covariates. This result already supports our contention that the scholarly understanding of positional blurring may not generalize outside Europe. In the next two sections we evaluate our two potential moderators—party-system dimensionality and populism type—which we argue structure the observed differences between Europe and Latin America. Put differently, our dimensionality and ideology hypotheses aim to draw out the characteristics that lead populism to have differing effects on positional blurring across these two countries and, by doing so, avoid reifying the regions.
Blurring, Populism, and Dimensionality
So far, we have shown that populism in Europe and Latin America have distinct strategic profiles. We argue these result from differences in party-system dimensionality and the inclusive versus exclusive ideological distinction. First, we provide evidence of dimensionality’s role by directly evaluating the relationship between blurring and party-system dimensionality before proceeding to the role of thick ideology in the next section.
To do so, we leverage Martinez-Gallardo et al.’s (Reference Martínez-Gallardo, de la Cerda, Hartlyn, Hooghe, Marks and Bakker2022) estimates of party-system dimensionality. Using a set of nine items on issue positions available in the CHES across both continents, we estimate two models for each country: one that assumes that party competition revolves around a single dimension comprising economic and sociocultural issues and a second one that assesses whether parties compete on a two-dimensional space in which economic and sociocultural issues are related but distinct. Party-system dimensionality is then operationalized as the difference in fit measured by the Comparative Fit Index (CFI) between the two-dimensional and the one-dimensional model, with higher values indicating that party competition is structured around two different dimensions (Bakker, Jolly, and Polk Reference Bakker, Jolly and Polk2012).Footnote 7 We scale this measure to have mean of 0 and a standard deviation of 1.
These estimates are presented in figure 4: lower values indicate that the party system is more unidimensional, whereas higher values indicate a more two-dimensional party system. As we expect, European party systems are substantially more multidimensional than Latin American party systems, although with a non-negligible overlap. The dot and whiskers in figure 4 indicate the average value and 95% confidence interval for both dimensions. The difference here is both statistically significant and substantively large.
Dimensionality Distribution in Europe and Latin America
Note: 95% Confident Intervals.
Figure 4 Long description
The x-axis is labeled Dimensionality and ranges from negative 2 to 4, with specific markers for One Dimension at negative 2 and Two Dimensions at 4. The y-axis represents density from 0.00 to 1.00.
Two distributions are shown:
* Latin America is represented by a white area under a black line. It shows a sharp, narrow peak centered around negative 1 on the x-axis, reaching a density of approximately 0.90. Above this peak is a black dot with a very narrow horizontal error bar.
* Europe is represented by a light blue shaded area. It shows a much broader, lower distribution that peaks around 0.5 on the x-axis with a density of approximately 0.45. Above this peak is a light blue dot with a wider horizontal error bar compared to Latin America.
The Latin America distribution is concentrated heavily toward the One Dimension side, while the Europe distribution is more spread out toward the Two Dimensions side.
We now test our dimensionality hypothesis by substituting the dimensionality measure for the regional dummy in the interaction with populism in our models (see table 1, models 3 and 7). We expect that the more multidimensional the party system, the stronger the positive association between populism and blurring on the economic dimension. This is the prevailing case in European party systems. Conversely, the more unidimensional the party system, the weaker the association between populism and blurring on the economic dimension will be, such as in most Latin American party systems.
The results of the models in table 1 substantiate our expectations. The interaction effects between our measures of dimensionality and populism are statistically significant and correctly signed in models 3 and 7, which predict blurring on economic and sociocultural issues, respectively. The regional dummy is also no longer statistically significant, indicating that the dimensionality interaction substantively identifies one of the important features driving the regional difference. Figure 5 provides a graphical depiction of the relationship among blurring, populism, and party system dimensionality.
Dimensionality, Populism, and Blurring on Economic and Sociocultural Issues
Note: The left panels (A and C) present the predicted values of populism on economic (A) and sociocultural (C) blurring for low and highly dimensional party systems. We operationalize high levels of party-system dimensionality as countries with levels of dimensionality that are one standard deviation above the mean. Countries with dimensionality levels on standard deviation below the mean are conceptualized as mostly unidimensional. The right panels (B and D) present the marginal effect of populism on economic (B) and sociocultural (D) blurring. 95% confidence intervals estimated via Monte Carlo simulation. The rug plots along the x-axis on the left panels (A and C) indicate the observed values of populism in our sample. The rug plots along the x axis on the right-side panels (B and D) indicate the observed values of party-system dimensionality in our sample.
Figure 5 Long description
A four-panel multi-plot arranged in a two-by-two grid.
Panel A (Top-Left): A line graph with Populism on the x-axis (0 to 10) and Left-Right Economic Blurring on the y-axis (0 to 10). Two lines with 95 percent confidence intervals are shown. The line for 1 S D Below Mean dimensionality is nearly flat, starting at 4.5 and ending near 4. The line for 1 S D Above Mean dimensionality shows a positive linear increase, starting at 3 and rising to 5.5. A rug plot along the x-axis shows data density concentrated between 1 and 6.
Panel B (Top-Right): A line graph with Dimensionality on the x-axis (-1 to 3) and Populism Marginal Effect on the y-axis (-1.0 to 1.0). A single line with a shaded confidence interval shows a steady linear increase, crossing the zero-effect line at a dimensionality of approximately -0.5 and rising to a marginal effect of 0.5 at a dimensionality of 3. A rug plot along the x-axis shows high density between -1 and 0.
Panel C (Bottom-Left): A line graph with Populism on the x-axis (0 to 10) and Socio-Cultural Blurring on the y-axis (0 to 10). The line for 1 S D Below Mean dimensionality is nearly horizontal at a value of 3.5. The line for 1 S D Above Mean dimensionality shows a slight negative linear trend, starting at 4 and decreasing to 3. Rug plot density is similar to Panel A.
Panel D (Bottom-Right): A line graph with Dimensionality on the x-axis (-1 to 3) and Populism Marginal Effect on the y-axis (-1.0 to 1.0). The line shows a slight negative linear trend, starting at 0 and descending to -0.3 as dimensionality increases to 3. The confidence interval widens as dimensionality increases.
Legends below the panels indicate that blue dots represent 1 S D Below Mean and grey dots represent 1 S D Above Mean for dimensionality, while line styles in the right panels distinguish between Europe and Latin America regions.
Figure 5, panel A, which plots the predicted values of blurring on economic issues across observed levels of populism at one standard deviation above and below the mean value of our dimensionality measure, shows that populism is positively associated with blurrier economic positions when party systems are dimensionally complex. When party systems approximate unidimensional spaces—that is, lower values on our measure—this relationship between populism and blurring on economic issues decreases. Panel B shows the marginal effect of populism on blurring and how it changes as dimensionality increases. The results are in line with our expectations: as dimensionality increases, the effect of populism on economic blurring becomes stronger and positive.
Panels C and D show the effect of populism and dimensionality on the sociocultural dimension: the results are inversed when compared with the economic dimension. Panel D clearly shows that, as the level of party-system dimensionality increases, the effect of populism on sociocultural blurring becomes stronger but in the negative dimension. In a highly two-dimensional party system, populism has a negative effect on sociocultural blurring. However, as panel C shows, the moderating effect of party-system dimensionality on sociocultural blurring is smaller than on economic blurring. The predicted differences between highly two-dimensional party systems and unidimensional ones are not as stark.
Of course, one concern at this point may be that it is precisely the positions and blurring strategies of radical-right populist parties that lead to the multidimensionality of some European party systems. If this were the case, we would be predicting parties’ blurring strategies based on a contextual variable that is largely the result of their own blurring strategies. Although we do not make causal claims that require strict exogeneity, we do have the ability to directly assess whether party systems remain multidimensional without populists. Of course, radical-right parties are hardly the only challenger parties challenging existing lines of contestation in Europe in times of large social change (Dassonneville Reference Dassonneville2023; De Vries and Hobolt Reference De Vries and Hobolt2020; Inglehart Reference Inglehart1990). We recalculate our measure of dimensionality in Europe excluding all radical-right parties (the parties that we expect might have increased dimensionality) and compare it to the original measures used in table 1. A low correlation between the measures with and without the radical- right parties would lend credence to the concern just raised and substantively weaken the results.
However, the two measures have an extremely high correlation of 0.92, indicating that multidimensionality can be treated as a contextual factor in which these parties act, rather than being the immediate result of their actions. Figure A3 in the supplementary material graphically shows that the two measures are very similar and that the calculated level of dimensionality is not systematically higher or lower with or without radical-right parties. The key takeaway here is that the multidimensionality of party systems in Europe is not substantively changed with the exclusion of radical-right populist parties, suggesting that our measure of party system dimensionality can be treated as an appropriate contextual factor.Footnote 8
Overall, these results indicate substantial support for our dimensionality hypothesis in understanding the relationship between populism and blurring strategies. We take these as evidence that the dimensionality of the system structures the viability of a strategy predicated on obscuring one dimension to highlight another.
Blurring, Populism, and Thick Ideology
In this section we evaluate our thick ideology hypotheses. Although differences between inclusive and exclusive populisms run along several theoretical dimensions (material, symbolic, and political), they can be broadly described by the way in which the term “the people” is described. Inclusive populism emphasizes the inclusion of historically marginalized sectors of the population, whether economically, racially, or democratically. In contrast, exclusionary populism highlights nativism in which the “true people” is defined in opposition from those who are non-natives.
We operationalize the difference between these two types of populism using party positions on immigration policy. Although the distinction between inclusionary and exclusionary populism has been very influential in the literature, to our knowledge, there has not been a systematic effort to categorize populist parties and leaders cross-regionally along this dimension empirically. A question tapping nationalism would have been a better proxy, but it is only available in the European CHES data. Nonetheless, the immigration policy question is very highly correlated with the nationalism question (r = .89), indicating that it should function as a valid proxy of a broader exclusionary, nativist thick ideology.Footnote 9 Figure A3 shows the relationship graphically for both the pooled sample and broken up by country to ensure it is true in both pooled and country-specific samples. The correlations are extremely high in all cases and remarkably consistent. The statistical and conceptual connection among nationalism, immigration, and an understanding of populists as divided between exclusive and inclusive poles is also consistent with Norris’s (Reference Norris2019) use of a cultural dimension ranging from authoritarian to (cultural) libertarian to investigate varieties of populist parties. In line with Mudde and Rovira-Kaltwasser’s (Reference Mudde and Kaltwasser2013) theoretical argument, we distinguish between two types of populism, rather than using a continuous measure. We coded every party that has an average expert score of five or more in the immigration policy question as exclusionary and all others as inclusionary.Footnote 10
Models 4 and 8 in table 1 evaluate the extent to which different types of populist parties have different strategic blurring profiles. In model 4, the interacted coefficient between populism and exclusionary has a positive significant effect, meaning that exclusionary populist parties tend to blur more on the economic dimension. The regional coefficient again falls below conventional levels of statistical significance, as it did when including the dimensionality–populism interaction.
Panel A of figure 6 provides a graphical depiction of the association. The model predicts an increase of economic blurring as populism increases, but only for those populist parties that are of an exclusionary nature. This is consistent with the marginal effects illustrated on panel B, which show that the effect of populism on economic blurring is statistically significant only for exclusionary parties. In contrast, we do not see any differences between exclusionary and inclusionary parties in their sociocultural strategic blurring profiles. Panel D of figure 6 clearly shows that the effect of populism on sociocultural blurring is not statistically significant for any types of populist party. These results support our thick ideology hypothesis.
Type of Populism (Inclusionary vs. Exclusionary) and Blurring on Economic and Sociocultural Issues
Note: The left panels (A and C) present the predicted values of populism on economic (A) and sociocultural (C) blurring for exclusionary and inclusionary parties. We operationalize exclusionary as being opposed to immigration. The right panels (B and D) present the marginal effect of populism on economic (B) and sociocultural (D) blurring. 95% confidence intervals estimated via Monte Carlo simulation. The rug plots along the x-axis on the left panels (A and C) indicate the observed values of populism in our sample.
Figure 6 Long description
Panel A is a line graph with an x-axis for Populism (0 Low to 10 High) and a y-axis for Left-Right Economic Blurring (0 Low to 10 High). Exclusionary populism shows a positive linear increase from approximately 3.5 to 5. Inclusionary populism shows a slight downward trend from 4 to 3.5. Shaded 95 percent confidence intervals surround both lines, and a rug plot indicates data density along the x-axis.
Panel B is a marginal effect plot with Populism Marginal Effect on the y-axis (negative 0.4 to 0.4). For exclusionary parties, the point estimate is positive at approximately 0.15 with error bars above the zero line. For inclusionary parties, the estimate is slightly negative, near negative 0.05, with error bars crossing the zero line.
Panel C is a line graph with an x-axis for Populism and a y-axis for Socio-Cultural Blurring (0 Low to 10 High). Both exclusionary and inclusionary populism show slight downward linear trends. Exclusionary starts near 4 and ends near 3. Inclusionary starts near 3.5 and ends near 2.5. Confidence intervals and a rug plot are present.
Panel D is a marginal effect plot for socio-cultural blurring. Both exclusionary and inclusionary parties show negative point estimates near negative 0.05, with error bars for both crossing the zero line, indicating no significant marginal effect.
Taking together the analysis presented in this and the prior section strongly suggests that differences between populist strategic blurring profiles are driven by both structural differences in party–system dimensionality and the interaction between populism and thick ideologies.
Conclusion
This article begins to untangle the relationship between populism and party blurring strategies. Populism has become an influential political force around the globe, and it is therefore important to understand how populists adapt and tailor their appeals to such varied political contexts. Focusing on the strategy of positional blurring on economic and sociocultural issues, we distilled key party and party-system factors from the Europe-focused research to show that the features that structure expectations of populism’s relationship with blurring in Europe, characterized by blurred economic positions coupled with clear sociocultural positions, lead to different predictions in Latin America. Specifically, we used and validated the recent expansion of the CHES to demonstrate for the first time that populists in Latin America present clearer economic positions than their nonpopulist competitors.
To explain why populism is related to blurring in some cases but not others, we found evidence for two important moderating factors: party-system dimensionality and populist parties’ thick ideology. Multidimensionality gives populists choices. In a unidimensional party system, a party that obscures its position is undermining its own distinctiveness from the rest of the party system. Populist parties seeking to challenge established parties should therefore clarify their positions to stand out from the rest. In a multidimensional setting, however, populists can choose the dimension or dimensions on which they want to build their identity. This means that the party can present clear positions on one dimension while obscuring its position on another so as not to limit its potential appeal.
We also argued that how populism attaches to thick ideologies matters because it structures which dimension ought to be blurred. As a thin-centered ideology based on the opposition between the pure people and the conspiratorial elite, populism can attach itself to diverse thick ideologies that provide concrete policy content. We claimed that exclusionary, nativist thick ideologies would lead populists to present blurred economic positions, because economic policies create social divisions across the native group. Inclusionary thick ideologies, which prioritize bringing the marginalized into the political system, however, should lead populist parties to present clearer economic positions.
This article further underscored the importance of understanding party competition and strategic decision making as fundamentally contextual. Parties sharing similar characteristics, even foundational tenets of how they understand society and its divisions, will adapt their behavior to their competitive environment. We also joined previous contributions in emphasizing that populism must be understood in relation to the thick ideology it clings to (Hunger and Paxton Reference Hunger and Paxton2022). Knowing a party is populist tells us that it sees a bifurcated society between the pure people and a corrupt, illegitimate elite. It does not tell us who those groups are or what policies the party might propose to rebalance politics in favor of the people. All these policies depend on the thick ideology and the strategic opportunities available to a party in its party system.
Though often left unsaid, nonpopulist parties also have thin-centered ideologies (Mudde and Rovira Kaltwasser Reference Mudde and Kaltwasser2017). Usually, this ideology is pluralism (more rarely, elitism)—the belief that society consists of many groups of equal moral standing vying for influence through the political process. We would never presume that just knowing a party is pluralist would enable us to understand the full policy platform or strategic decisions of a nonpopulist party. We would insist on knowing the social background of the party’s voters, its thick ideology (socialist, Christian democrat, liberal, etc.), the nature of the electoral system, and many other factors. Populist parties should be treated similarly both in how we think about their behavior and their normative relationship to democracy.
Future scholarship should adopt these insights as we continue to understand the role of populism in contemporary party politics in Latin America, Europe, and globally. Specifically, it should consider the role of temporal dynamics. This article is cross-sectional in nature because of data availability in Latin America, but there is much to learn about how positional blurring changes over time. This information becomes particularly important as more exclusive populists like Bolsonaro in Brazil, Milei in Argentina, and Kast in Chile become more common.
Future work should also examine how party system institutionalization and party organizational strength condition populist strategies. This is particularly relevant in Latin America, where efforts to build enduring party organizations have been difficult to sustain (Levitsky et al. Reference Levitsky, Loxton, Van Dyck and Dominguez2016) and political parties struggle to effectively aggregate political preferences and coordinate political elites (Luna et al. Reference Luna, Rodríguez, Rosenblatt and Vommaro2021a; Reference Luna, Rodríguez, Rosenblatt and Vommaro2021b)—although some recent works show that this can be successfully done in the region through organizationally mediated appeals via civil society (Poertner Reference Poertner2024). In this context, populism’s natural affinity with charismatic leadership sits uneasily with party-system institutionalization: populism can thrive without partisan organizations, and it may even corrode them (Kitschelt Reference Kitschelt2000; Weyland Reference Weyland2001). This is a challenge to the broader literature on party competition in the region, and we acknowledge it as a boundary condition of the analysis. Future research should assess whether the blurring patterns we identify reflect deliberate party strategy or whether they are better understood as the direct expression of a charismatic leader’s political will, unmediated by party organization.
Moreover, the findings of this analysis should be broadened through the development of more harmonized measures of positional blurring and populism types. This article is the first to introduce blurring as a quantitative concept in Latin American politics. As a result, there is little to validate the CHES measure against or to use as reliable alternative operationalizations. Both concepts are well grounded in the literature, but comparable measures across regions are scarce. The two most recent and most detailed datasets on varieties of populist parties, the PopuList and POPPA (Meijers and Zaslove Reference Meijers and Zaslove2021; Rooduijn et al. Reference Rooduijn, Andrea, Halikiopoulou, Froio, Van Kessel, De Lange, Mudde and Taggart2024), include only European countries, and there are no comparable sources in Latin America. We leveraged a validated approach from the Chapel Hill Expert Study that was included in both the European and Latin American surveys (Hooghe et al. Reference Hooghe, Marks, Bakker, Jolly, Polk, Rovny, Steenbergen and Vachudova2024, 466), but future efforts should incorporate more world regions into the existing populism-focused data collections. This is also true for the exclusive–inclusive divide, which is established in the literature (Mudde and Rovira-Kaltwasser Reference Mudde and Kaltwasser2013) but lacks a designated measurement strategy that can be used across regions. The proxies we use in this article get at key features of this divide, but the field would benefit from established measurement strategies across world regions to operationalize these differences.
We should also consider how the features we identify in this study structure political behavior and public opinion in response to populism at the individual level (Castanho Silva, Fuks, and Tamaki Reference Castanho Silva, Fuks and Tamaki2022), which we bracketed to focus on the parties themselves. How do these types of dimensional structural differences condition patterns of populist support in the electorate, and does the exclusionary or inclusionary nature of populism relate to patterns of negative partisanship (Ridge Reference Ridge2022), affective polarization (Gidron, Adams, and Horne Reference Gidron, Adams and Horne2020; Horne, Adams, and Gidron Reference Horne, Adams and Gidron2022), democratic backsliding (Bauer and Becker Reference Bauer and Becker2020), or other increasingly salient topics and features of contemporary democratic systems.
Supplementary material
To view supplementary material for this article, please visit http://doi.org/10.1017/S1537592726105805.
Data Replication
Data replication sets are available in Harvard Dataverse at: https://doi.org/10.7910/DVN/PCEHWW.
Acknowledgments
This article benefited from comments from a number of friends and colleagues. We would like to particularly thank Jan Rovny, Jonathan Polk, Liesbet Hooghe, Gary Marks, Ann-Kristin Kölln, Jana Schwenk, Lena Huber, Jim Adams, Javier Sajuria, and participants in the Current Perspectives on Party Competition seminar (MZES, Feb. 2024). We are also grateful to the anonymous reviewers and editors of Perspectives on Politics for their constructive feedback.
Open access
