This is an open access article distributed under the terms of the Creative Commons
Attribution License (The last years have seen democratic recession among several liberal and electoral democracies (Diamond, 2021), resulting in a third wave of autocratization (Lührmann & Lindberg, 2019). This process seems to be linked with the growing presence of populist leaders which are self-portrayed as the defenders of “democracy by the people” (Mény & Surel, 2002), advocating for a new model of democracy wherein “the people” are the only true object of representation. It is to them only which politicians much explain and justify their actions and only the “people” can hold politicians accountable.
Recent studies on why citizens support this view of democracy (especially anti-elite perceptions) (cf. Hawkins & Rovira Kaltwasser, 2018, p. 7; Rooduijn & Pauwels, 2011, p. 1273) seem to link it to the failure of parties and elites to represent citizens’ preferences and expectations (Castanho Silva & Wratil, 2023), and by extension to the widespread long-term perception of the lack of institutional receptiveness or democratic responsiveness to individual demands1. Populists’ support in many democracies is often linked to a crisis of representation manifested by high levels of distrust in representative political institutions (Lisi et al., 2019).
This begs the question, therefore, of how are individuals that are so critical with the mechanisms of democratic representation willing to accept populist leaders as part of “the people” and not one of the “corrupt elites”? How is it possible that a wealthy real estate developer, a university professor, or even a leader of a former mainstream party could be considered “one of the people”, thereby overcoming the problem of lack of democratic representativeness that seems to be the curse of other political leaders? Finally, what makes populist leaders so appealing given that, populist politicians do not need to rely on a strong populist rhetoric to be ‘recognized’ as such by populist voters (Casiraghi et al., 2023, p. 16)?
While much attention has been given to the characteristics of populist leaders, there has been little that connects the constellation of traits they are perceived to have to their support. Studies that have looked at specific traits of populist leaders have not explored how this relates to their general evaluation (Nai & Martínez i Coma, 2019). Other research has focused on these leaders being seen as “outsiders” and the implications of this once they reach positions of power within institutionalized settings (Castanho Silva, 2019). While promoting themselves as “outsiders” may gain them some support, the literature from social psychology has demonstrated that how individuals perceive others along two dimensions of personality traits can have a major impact on how they evaluate and behave toward them (Cuddy et al., 2008; Fiske et al., 2007). Taking this literature into consideration, we find it useful to think about the strategies that populist leaders use to gain acceptance that goes beyond their populist rhetoric. As has been discussed for the case of Argentina (Annunziata et al., 2018), a country with a long tradition of populism, contemporary populist leaders use a kind of proximity strategy in which they attempt to be perceived as an average person and who, as one of them, has good intentions toward “the people”. We argue here that leaders can be successful in this strategy by fostering an image in line with what the literature in social psychology has deemed “warmth” traits, such as being honest, trustworthy, and compassionate. Although warmth could be present and have an effect for the rest of the leaders in the political system and may be related with individuals’ preexisting political preferences, we intuit that this strategy adopted by populist leaders should be particularly effective among individuals who hold more critical perceptions of existing levels of democratic responsiveness (external political efficacy), as they would be drawn to a leader who would appear to be more of an advocate for “the people” than traditional elites.
In this paper, we attempt to gain a better understanding of the popularity of populist leaders by examining the relationship between how individuals perceive leaders’ traits and their overall evaluation. In particular, we are interested in how the relationship between perceptions of leaders’ warmth traits and their overall evaluation differs for populist leaders, who strive to align themselves with “the people” more so than non-populist leaders, and how that relationship is mediated by citizens’ external political efficacy. While we acknowledge that almost all leaders, to a certain extent, attempt to present themselves as an “average” person, for non-populist leaders this is typically not the center of their political discourse. Populist leaders, on the other hand, make this proximity strategy an integral part of their campaigns and image by playing up the characteristics and intentions that make them part of “the people” thereby having a better chance to be supported among those individuals critical with existing democratic representative institutions. Thus, building upon recent literature that finds perceptions of leaders’ warmth and competence important for individual vote decisions (Bittner, 2011; Costa & Ferreira da Silva, 2015), we argue that perceptions of populist leaders’ warmth contribute to their more positive overall evaluation among citizens displaying lower levels of external political efficacy as a measure of an overall critical view of existing democratic representativeness.
We do not deny the importance of other factors in predicting how individuals evaluate leaders, nor the importance of warmth more generally for all voters, but perceptions of warmth should be particularly relevant in support for the populist leaders among those who display lower external political efficacy. As such, this conditional relationship should be absent for non-populist leaders. Additionally, we argue that among individuals supporting populists, views of leaders’ competence, i.e. the ability to achieve their issue-based goals, although also important for their support, are not equally affected by individuals’ attitudes of external political efficacy.
In the pages to follow, we test this argument in two very different political contexts: Spain and Argentina. Argentina’s electoral competition has long been characterized by the presence of a large populist party in a presidential system (Calvo, 2019), while in Spain populism is a fairly new phenomenon with a new and smaller populist party competing against major programmatic non-populist parties in a parliamentary system (De Cleen et al., 2018; Gómez Reino & Llamazares, 2018). In both contexts, however, at least one leftist populist leader’s party obtained significant electoral results. This ‘most different’ comparative design will allow us to assess the conditional effect of citizens’ external efficacy on the relationship between warmth and the evaluation of populist leaders with varying degrees of electoral support in different institutional contexts and party systems. To approach this comparison, the following study analyzes original online survey data from Argentina collected at the beginning of 20162 and a post-electoral survey following the Spanish general elections in April of 2019.
The Argument and Hypotheses
The literature on populism generally agrees that leadership is essential in framing the conflict underlying populist discourse and that the identification between a populist leader and their supporters is not only essential but also inherently different than the typical leader-voter relationship (Weyland, 2017). While differences exist amongst populist leaders, what sets all of them apart from non-populist candidates is their claim to be the quintessential expression of anti-elite and anti-establishment politics, and that they are the true representatives of “the people” and their interests. However, in general, there is a lack of clarity in the literature on how certain leaders can appeal to citizens whose vision of the political world is highly conditioned by the conviction that there is a “pure people” whose interests differ substantially from those of the “corrupted elite” (cf. Hawkins & Rovira Kaltwasser, 2018, p. 3). According to populist discourse, the “people” is homogeneous and unified, essentially virtuous, and politics should be nothing but the implementation of the popular will; the people, however, are oppressed by the elite made up of a small coalition of powerful and unrepresentative actors that illegitimately controls politics for its own benefit (Canovan, 2002, p. 27). As has been shown, citizens who rage against these elites put their trust in the people’s will (e.g., Van Hauwaert & Van Kessel, 2018).
In addition, populists are strict defenders of “democracy by the people” (Mény & Surel, 2002). As a consequence, the populist model of democracy is characterized by an extreme view of vertical accountability in which only voters can hold politicians accountable, contradicting the delegate model of democratic representation (Taggart, 2002, p. 6). This model thus advocates for extreme responsiveness which attacks mainstream political parties whose actions do not respond to “the people” and the general will (Mair, 2002). This restricted view of representation also applies to the agents of representation: populists are hostile toward intermediaries between the people and decision-makers, especially political parties (Ferrin & Kriesi, 2024).
Hence, a significant part of the recent literature centered on the rise of populism has searched for the roots of populist success in political factors and crises of representation (e.g., Hawkins & Rovira Kaltwasser, 2017; Halikiopoulou & Vasilopoulou, 2018). The more unrepresentative parties appear in voters’ eyes, the more heightened populist attitudes should become (Castanho Silva & Wratil, 2023; Oliver & Rahn, 2016). So while it is true that external political efficacy and populist attitudes constitute different constructs, they are very much related (Geurkink et al., 2020, p. 13). Therefore, the question we derive from this theoretical discussion is how is it possible that a strong anti-elitist discourse that resonates with a general perception of a lack of democratic representativeness could activate support for leaders, who, in many cases, come from the economic, cultural or, even existing political elite? Following previous work from social psychology on how people perceive one another, we draw upon the stereotype content model (SCM) (Cuddy et al., 2008; Fiske et al., 2007) as the basis for our examination of leaders’ traits along two dimensions. These dimensions are given somewhat different names depending on the author and discipline but are often referred to as warmth or morality, and competence or intelligence. Evaluations of warmth include whether an individual is sincere, trustworthy, honest, good-natured, friendly, and kind. It should be noted that this does not mean that the individual is necessarily considered “warm” under all circumstances or to all entities, simply that this individual is perceived to have traits that fall under a broader umbrella that have more to do with intentions and morality than skill. This is particularly important for this study, as populist leaders can often be seen as belligerent or arrogant, especially to those not considered part of what they claim to be their in-group, but they are often seen as having good intentions and acting towards the benefit of those considered part of “the people”. Competence, on the other hand, is attributed to individuals who are perceived to be skillful, capable, intelligent, efficient, and possessing the resources and confidence to achieve their goals. Another way to think about these two dimensions is that warmth, also referred to as communality, is seen as the intent that an actor has, while competence, or “agency”, is the ability of that individual to enact those goals (Formanowicz et al., 2018; Harris & Fiske, 2009). If the intentions of a populist leader are not seen to align with the people, their competence to enact those goals will be of little importance, especially to those skeptical of traditional elites. Although there are other dimensions one could take into consideration when studying traits, scholars have consistently found these two to have effects on individual-level behavior, and that these effects are nearly universal and robust across a variety of contexts and nearly universal (Cuddy et al., 2009).
One way to convince voters that populist leaders are different from the elite establishment is by engaging in a strategy in which the leader tries to resemble to the ‘common man’ through highlighting their traits along the warmth dimension (Ostiguy et al., 2017, p. 73). Populism promotes different leader-supporter dynamics that are based more on ‘people centrism’ and proximity, and less on perceptions of leaders’ competence (Annunziata et al., 2018). These leaders often attempt to portray the traditional elites as cold, corrupt, and lacking other more “humanizing” traits, such as compassion and empathy. Studies have shown that warmth traits take priority in humanization (Chu & Martin, 2021) and that warmth traits are determined by expectations dealing with cooperation with groups in society, while competence is related to one’s socioeconomic status relative to others (Fiske et al., 2002). By improving their image in terms of warmth or communality, it would resonate more with those individuals receptive to anti-elite discourse who would be looking for someone who is more human, more caring and attentive to the needs of the average citizen than they perceive traditional elites to be. This does not mean that non-populist leaders do not attempt to cultivate an image as an average “down-to-earth” person, but this does not take such a prominent role in their discourse in comparison to populist leaders, as non-populist leaders tend to focus more on issue-based programmatic aspects of their platform.
While not all individuals would be drawn to this type of warmth-centric strategy, those with more critical views of democratic representativeness, would be particularly susceptible as it falls in line with their preexisting evaluations of traditional elites, making them more likely to care whether a leader has good intentions and concern for “the people”, than their skills to carry out those intentions, such as if they are ‘intelligent’ or ‘decisive’. Additionally, the appreciation of leaders’ competence traits, i.e., their capacity to achieve goals, should not be something distinctive when it comes to express support for populist leaders and, more importantly, it should be expected to be constant at all levels of external political efficacy. As such, while warmth is a trait dimension that undoubtedly matters for the evaluation of all leaders (Bittner, 2011; Costa & Ferreira da Silva, 2015), we argue that it likely matters more for the evaluation of populist leaders for those citizens who display more critical views of democratic representation (lower external political efficacy).
Based on the preceding discussion we propose the following two hypotheses:
H1: Perceptions of warmth and competence constitute a significant predictor of overall leader evaluation for both populist and non-populist leaders.
H2: The lower an individual’s external political efficacy, the greater the effect their perceptions of populist leaders’ warmth should have on their overall evaluation of that leader.
Data and Case Selection
The data used in this study were obtained from two original surveys in Argentina and Spain. These surveys included the same battery of questions on personality traits for three leaders in Argentina and four in Spain. The data for Spain were collected with an online survey conducted in April 2019, following the national elections, which constitute part of a online survey panel (Torcal, Santana, Carty, & Comellas, 2020).3 Included in the survey were questions regarding four Spanish political leaders: Pablo Casado, the leader of the traditional conservative People’s Party (PP); Pedro Sánchez, incumbent president and leader of the traditional center-left Spanish Worker’s Socialist Party (PSOE); Albert Rivera, who was the leader of center-right Ciudadanos (Citizens) since its creation in 2006 until after the elections in 2019; and, finally, Pablo Iglesias, who was the main leader of Podemos, the new populist party that emerged in the electoral arena in the 2014 European elections.4 The combined vote share of these leaders constituted of more than 78% of the total vote in national elections in April 2019.
The data for Argentina are from an original online survey conducted in September of 2016. The data collection methods and sampling characteristics were very similar between this case and that of Spain5. The same battery of personality questions used in the Spanish survey were also included here for three Argentine political leaders: Cristina Fernández de Kirchner, leader of Frente para La Victoria (FPV – Front for the Victory) and the incumbent president; and two other politicians who challenged her party in the election of 2015—Mauricio Macri, the leader of the large opposition coalition Cambiemos (Coalition for Change), and Sergio Massa, the candidate of another coalition—United for a New Alternative (Unidos para una Nueva Alternativa).
Having cases as distinct as Spain and Argentina provide us leverage in testing the generalizability of our argument. Argentina is a presidential system in which individual leadership, and more concretely, populist leaders and personalism, have long played an important role (Mudde & Rovira Kaltwasser, 2013, p. 156). In this context, vote choice tends to be more related to preferences for leaders and political parties (especially as the two frequently coincide in personalistic parties) than ideology, a tendency that has persisted despite increasing fragmentation of the party system (Calvo & Murillo, 2012). In contrast, Spain is characterized by a very party-centric parliamentary system where the focus of political contestation is on political parties instead of their leaders, and the general effect of individual leader traits on vote choice tends to be lower (Curtice & Lisi, 2014).
Despite their institutional and historical differences, these countries share one characteristic that is critical for the purposes of this study—both have an electorally successful leftist populist party present in the party system. In Spain, the party Podemos emerged following the various social mobilizations that channeled society’s resentment towards austerity policies and the political regime in a context of deep economic recession, high levels of distrust in political institutions and a crisis of representation (Lisi et al., 2019). This party obtained a surprising 8% of the vote in the 2014 Spanish European election, thus becoming the fourth largest political force at the time. While their strength in the parliament has fluctuated since then, in the 2019 Spanish national elections, for which the post-electoral survey used in the present study was conducted, Podemos’s coalition obtained 14.3% of the vote, placing them in the fourth position. Podemos has commonly been cited as a populist party that combines anti-elitist claims with a left-wing discourse (De Cleen et al., 2018) and gained strength by employing fierce populist-based rhetoric that included a critique of all the existing parties and economic and financial elites (‘la casta’), while portraying themselves as the real representative of the interests of the people (‘la gente’). Additionally, despite the fact that populism in Europe seems to predominantly rely on exclusionary discourses (Mudde & Rovira Kaltwasser, 2013, pp. 162-163), as Lisi and colleagues (2019) have more recently shown, the nature of the populist rhetoric used by Podemos is predominantly inclusionary (Marcos-Marne, 2021; Marcos-Marne et al., 2020; Rooduijn et al., 2019). That both parties are leftist inclusionary populist parties allows us to control for the ideological component.
In Argentina, the populist party is the Frente para la Victoria (Front for Victory, or FpV), the party of Cristina Fernández de Kirchner, who had been the President of Argentina for eight years from 2007 to 2015. Her policies were often seen as a continuation of her husband’s, Nestor Kirchner, who was President before her from 2003 to 2007. The couple consolidated their political apparatus under the personalistic FpV party, following the tradition of inclusionary populism that characterized the Peronist Party (Partido Justicialista-PJ) (Annunziata, 2013). In selecting Argentine leaders to compare to Kirchner, we selected the two primary opposition candidates in the 2015 elections—Mauricio Macri and Sergio Massa. Macri ran as the candidate of Cambiemos, a coalition of parties that formed with the explicit goal of removing the Kirchneristas from power. Coming from a very different background, Sergio Massa served in various positions for both Kirchner presidents and later broke with President Fernández de Kirchner to run as the candidate of a new party, the Renewal Front, under the coalition United for a New Alternative (UNA).
The differences across the two cases also provide us with an opportunity to test whether our expectations hold across different characteristics of populist leaders. One advantage that our research design presents is the ability to test across gender of candidates. In Argentina, the populist leader is a woman, whereas in Spain, the populist leader is a man. We consider this to be an important component to consider, as previous research has found that individuals tend to evaluate traits of candidates differently depending on their gender. At least in Western societies that tend to be more individualistic, such as Spain and Argentina, women are typically perceived to have more “communal” traits such as warmth, gentleness, and kindness, whereas men are more likely to be perceived as having instrumental or agentic traits such as assertiveness and aggression (Cuddy et al., 2015), which effects the issues they are seen to be competent to handle (Sanbonmatsu, 2003), as well as on which qualities candidates are evaluated (Fox & Oxley, 2003). By having one female and one male candidate, we can evaluate whether a strategy to emphasize leaders’ warmth in order to appeal to populist-oriented citizens may be effective regardless of the candidate’s gender. Another advantage is that one of the candidates, Cristina Kircher de Fernández, at the time of the survey had been president for 8 years, while the populist party Podemos in Spain had not yet been in government and had typically been the third or fourth most-voted party in the system. While a larger-N study would be necessary to draw conclusions as to the statistical significance, our case selection provides us with at least a preliminary look at whether our expectations still hold in the presence of these differing characteristics among populist leaders.
Operationalization of Variables and Methodological Approach
The Dependent Variable and its Transformation
Our dependent variable of interest is leader evaluation. Respondents were asked for their overall opinion of each leader and provided response options on a scale from 0 to 10. As we display in Figures 1 and 2, candidate evaluation is a variable that presents a very peculiar distribution with a modal response of ‘0’, the lowest possible value, for all leaders included in this study (except for Massa in Argentina). Having a dependent variable with a high concentration of cases in the zeros is very problematic for the estimation of OLS models since it is not normally distributed. It is recommended to convert highly skewed variables using a logarithmic transformation into one with a log-normal distribution whose logarithm is normally distributed – but whose untransformed scale is skewed (Benoit, 2011). The resulting transformation of our dependent variables display a more normal distribution with a range between 0 and 2.30 and median between 1.38 and 1.79 and mean between 1.21 and 1.66 (for the distribution of these variables before and after transformation see Figures A2.1-A2.2 in the Online Appendix). The literal interpretation of a coefficient β in a log-linear model is that a one-unit increase in X will produce an expected increase in log Y of β units. This means that the expected value of Y is multiplied by eβ (Wooldridge, 2010). Thus, for obtaining the expected values according to this parameters and the corresponding confidence intervals, it is necessary to exponentiate the coefficients and variance according to the following formula:
1
where σ2 is the variance of the errors.
Independent Variables
We now turn to the operationalization and measurement of our independent variables of interest. To measure these dimensions of leaders’ warmth and competence, respondents were asked to what extent they agree or disagree with using specific adjectives to describe each leader on a 7-point Likert scale. We included seven traits that follow the existing literature based on the SCM (Collange et al., 2009, pp. 141-142; Sink et al., 2018, pp. 596-597): four warmth traits (‘caring’, ‘trustworthy’, ‘dishonest’ and ‘arrogant’) and three competence traits (‘decisive’, ‘charismatic’ and ‘intelligent’).6 All questions asked explicitly about the leader’s traits and made no reference to the political party with which they are affiliated or any other contextual information besides the leader’s first and last names. The measures of leaders’ trait perceptions are therefore created in a way that attempts to tap into perceptions of the personal traits of the leader and avoid contamination effects from partisan affiliation or incumbency.
The Conditional Variable
Another key variable for our argument is citizens’ perception of democratic responsiveness. Following the existing literature (Linde & Peters, 2020; Torcal, 2014), we measure these by creating a scale using the most traditional indicators of external political efficacy: “…the political system allows people like me to have a say in what the government does…”; “…politicians care about what people like me think…” In addition, for Spain we have the item “…the political system allows people like me to have an influence in politics…”. Using these, we create a measurement model for external political efficacy by estimating saturated measurement models with CFA (with a variance of the factor fixed to 1), obtaining high standardized factor loadings with a very good overall goodness of fit for both countries (see Tables A2.3 and A2.4 in the Online Appendix). We then created normalized variables from the individual factor scores of the latent variables measuring external efficacy and subtract it from 1. The resulting variables have a mean of 0.73 (σ = 0.22) for Spain and 0.45 (σ = 0.29) for Argentina (see Tables A2.1 and A2.2 in the Online Appendix).
Finally, when individuals evaluate political leaders, there are often additional factors that may influence their assessments—particularly, party identification and ideology (Bittner, 2014; Lobo & Curtice, 2014). We therefore include in the models the control variable shared partisan attachment that indicates the strength of identification the respondent feels to the political party of the leader. This variable ranges from 0-3 with ‘0’ indicating no identification with that party, ‘1’ as weak identification, ‘2’ as some identification, and ‘3’ as strong identification.7 While in the Spanish survey we only included in the models one partisan variable that corresponds with the leader’s party (one per leader, and therefore one per model), the party system in Argentina is much more complex. A single leader may be associated with more than one party or coalition, and, at the same time, multiple leaders may be affiliated with a particular party, as is the case with the Peronist Party.8 As such, we included the main umbrella party, coalition, or political party with which the leader is affiliated and ran robustness checks by including different and additional political parties, none of which altered the results of the models.
In a similar manner as partisan identification, we expect ideology to influence how an individual evaluates a leader. We measure ideology as the respondent’s self-placement on a left-right scale where ‘0’ indicates the far left and ‘10’ represents the far-right side of the ideological spectrum. Finally, we also include age, education, gender, income, and religiosity as control variables. All these variables except the categorical and dummy variables were normalized to 0-1.
The Models
To evaluate our hypotheses, we estimate Generalized Linear Models with robust standard errors9 using the natural logarithmic transformation of leader evaluation as the dependent variable. We estimate two models per leader, one with only the direct effects, and, more importantly for our argument, another model including the interaction terms. Thus, we estimated a total number of fourteen models. The final model with the interaction terms estimated for each leader is the following:
where Lnyila is the naral log of the individuali evaluation of leader a. Warmthila and Competenceila are the individuali warmth and competence perceptions of leader a. (External Political Efficacy)i is the scale measuring external political efficacy for individual i. χni is the n vector of control variables for individual i. (Warmthila × External Political Efficacyi) and (Competenceila × External Political Efficacyi) are the interaction terms for individual i between either warmth or competence perceptions of leader a and external political efficacy.
Results
The results for the analyses of the evaluation of the seven leaders in both countries without the interaction effects are displayed in Figures 1 (Spain) and 2 (Argentina) below (for full results see Online Appendix Table A3.1). The results in these figures confirm findings from previous literature and support our predictions in H1 in that the perceptions of leaders’ warmth and competence constitute an important and consistent part of all leaders’ evaluations in both countries regardless of their newness in the political system, size of their party, and experience as incumbent.
Figure 1
Predictors of Leader Evaluation in Spain
Note. Source: E-Dem panel data, Wave 3 (Torcal, Santana, Carty, & Comellas, 2020). Confidence intervals at 95% and 99%.
These results thus point to the importance of these perceptions for support, constituting the baseline of the overall effect of these perceptions on the evaluation of leaders. These results also reveal that ideology and partisanship have a significant relationship with overall leader evaluation, although, as expected, ideology proves to be stronger in Spain.
Figure 2
Predictors of Leader Evaluation in Argentina
Note. Source: 2016 Argentina post-electoral survey (Carty & Torcal, 2023a). Confidence intervals at 95% and 99%.
The key coefficients of interest for our argument are those for the interaction terms between perceptions of leader warmth and competence, and external political efficacy. In the results from the expanded models, the interaction terms including perception of leader warmth are positive and significant for the two populist leaders of both countries, meaning that the effect of warmth on populist leaders’ evaluations increases as external political efficacy decreases. This coefficient of -1.28 is statistically significant at p = .01 for Iglesias, the populist leader of Podemos in Spain, and the coefficient of -0.86 for Kirchner is significant at p = .03 (for full results see Table A3.2 in the Online Appendix).
We display in Figures 3 and 4 the marginal effects of the expected values of warmth and competence on leadership evaluations after reversing their values using the transformation formula [1] discussed above for all the leaders of Spain and Argentina respectively. As we can see in these figures, the effect of warmth as external political efficacy decreases is null for the other leaders in Spain and Argentina, except for Macri (the challenger from the Argentine Cambiemos coalition), for whom the relationship is positive, showing that the effect of warmth on his evaluation decreases as external political efficacy also decreases. This might be a peculiar effect due to the Argentine context, where despite the presence of some proximity strategies adopted by Macri (Annunziata et al., 2018), he was still perceived as very elitist among those displaying lower levels of external political efficacy.
Figure 3
Marginal Effects of Warmth and Competence Perceptions on Overall Leader Evaluation Across Levels of External Political Efficacy for Leaders in Spain
Note. Source. E-Dem panel data, Wave 3 (Torcal, Santana, Carty, & Comellas, 2020). Confidence intervals at 95%.
Figure 4
Marginal Effects of Warmth and Competence Perceptions on Overall Leader Evaluation Across Levels of External Political Efficacy for Leaders in Argentina
Note. Source: 2016 Argentina post-electoral survey (Carty & Torcal, 2023a). Confidence intervals at 95%.
We would also like to point out that this relationship does not appear to be explained by ideological leanings of the leader, as the interaction term is significant for Iglesias, but not for the other leftist leader in Spain, Pedro Sánchez. The same is true with Argentine leftists Kirchner and Massa (an important figure on the left, and even from the same party, the Partido Justicialista in the past). Finally, the results do not reveal a similar conditional effect of external efficacy on the relationship between competence perceptions and leader evaluation for either non-populist or populist leaders and is even negative in the case of Kirchner.
Finally, to support these initial findings, we have estimated two models in each country to observe the effect of this interaction term between warmth and external political efficacy. One model in each country has evaluation of the populist leader as the dependent variable. In the other, we use the average evaluation of the non-populist leaders given by all respondents. As such, we also use the average warmth and competence perceptions for the non-populist leaders for the independent variables in the second models. This interaction term is strong and consistent for populist leaders in both countries, whereas it is not significant for the evaluation of the non-populists (for full results see Table A3.3 in the Online Appendix). Figures 5 and 6, displaying the marginal effects of those interaction terms in Spain and Argentina respectively visually confirm our principal finding when it comes to compare populist and non-populist leaders.
All the preceding findings confirm the role of warmth in explaining populist leaders’ evaluations as individuals' external political efficacy decreases (H2), regardless of their gender, their electoral perspectives, governmental experience, and the historical presence of populism.
Figure 5
Marginal Effects of Warmth on Overall Leader Evaluation Across Levels of External Political Efficacy for Populist and Non-Populist Leaders in Spain
Note. Source: E-Dem panel data, Wave 3 (Torcal, Santana, Carty, & Comellas, 2020). Confidence intervals at 95%.
Figure 6
Marginal Effects of Warmth on Overall Leader Evaluation Across Levels of External Political Efficacy for Populist and Non-Populist Leaders in Argentina
Note. Source: 2016 Argentina post-electoral survey (Carty & Torcal, 2023a). Confidence intervals at 95%.
Robustness Check: Testing the Issue of Endogeneity
Although our attention is not focused on the causality of this argument but on the conditional effect of external efficacy on the relationship between warmth and leaders’ evaluations, we cannot omit the fact that, as several scholars have asserted, it is difficult to disentangle the effects of pre-existing voters’ preferences from leaders’ traits perceptions (Bittner, 2014; Klingler et al., 2019). In an attempt to address this problem, we made use of the panel data in Spain which included respondents’ vote choice for the national elections that took place on April 28th of 2019 in the preceding wave (Wave 2). Using this information, we created a dummy variable with a value of ‘1’ if the respondent’s declared vote intention was for the leader under evaluation and ‘0’ otherwise. Because previous waves in the study we used for our Spanish data did not contain information on leader traits, unfortunately we cannot estimate a longitudinal model with individual fixed effects, but at least we were able to test the robustness of this relationship with a more demanding statistical test to control for some level of endogeneity produced by voters’ existing party preferences. As can be seen in Table A3.4 in the Online Appendix, the coefficient for the interaction term between external political efficacy and perceptions of leader warmth for the populist leader Iglesias remains almost the same (-1.286, se: 0.33), even after including the variable vote choice in the previous national elections (at t-1), confirming the existence of the relationship we observed previously. This interaction term for the rest of the three Spanish leaders continues to not reach statistical significance. Additionally, the interaction term between perceptions of competence and external political efficacy is again non-significant.
Discussion
The growing prevalence of populist leaders seems to be a concerning issue due to the inherently difficult relationship between democracy and populism (Landwehr, 2020). The practices of populist leaders once in power (Haggard & Kaufman, 2021) and tensions derived from these practices (Fitzi et al., 2018) are generally related to bad democratic performance. However, the populist model of democracy, characterized by an anti-elitist discourse and a very critical view of democratic representation, seems to result in a high and durable support for their leaders. So, what makes populist leaders so appealing for a certain sector of the electorate in contemporary democracies?
Although there is a wealth of literature on the rise and electoral success of populist parties, little is known regarding the effect of individual leaders, and more concretely how they are perceived by potential supporters and the effect of those perceptions on leaders’ overall evaluations. Leader evaluations are crucial for understanding support, as they function as shortcuts for voters in understanding and processing more complex and demanding information (Holian & Prysby, 2014). In the preceding pages, we have investigated one part of this issue by focusing on the interactive relationship between perceptions of leader warmth and external political efficacy, which are linked to anti-elitist attitudes (Castanho Silva & Wratil, 2023; Oliver & Rahn, 2016). We have done so by studying original public opinion data in two very different settings: Spain in 2019 and Argentina in 2016, both very distinct in terms of institutional, historical, and cultural characteristics, but have in common the presence of an inclusionary leftist populist leader in electoral competition.
In these two cases we have shown that how people perceive leaders matters for their overall evaluation, but in somewhat different ways for populist versus non-populist leaders. Evaluations of populist leaders tend to increase with warmth perceptions among those individuals with lower external political efficacy, whereas for non-populist leaders this interaction effect is absent and even sometimes negative, suggesting that populists’ success is more dependent on the cultivation of their image along warmth traits in comparison to non-populist leaders.
We believe that these findings constitute an innovative and significant contribution in the discussion of populist leadership by introducing warmth as an important component for the evaluation of populist politicians and the nuances of this relationship across individual-level attitudes of external political efficacy. In general, previous literature has acknowledged the importance of leadership in populist parties, but to our knowledge, beyond the assessment of the direct effect of leadership on voting little has been done to explore what it is about leaders’ traits that has greater resonance among populist-oriented individuals. This is especially noticeable in the contributions made by the ideational approach in the studies of populism. In that respect, we hope the findings presented here will open a door to consider the importance of perceptions of leaders’ warmth traits as an essential mechanism by which populist leaders are accepted as part of ‘true people’ (people-centrism) differentiating them from the ‘corrupted elites.’
These findings have implications in other areas as well. First, while it might be true that populist party leaders do not have a disproportionate direct electoral effect on party preferences (van der Brug & Mughan, 2007, pp. 42-43), as we have shown in the preceding pages, there are certain trait perceptions of populist leaders that might appeal more to citizens with higher levels of populist attitudes. This should have (indirect) electoral consequences, given the existing evidence on how leader evaluation has a strong effect on voting preferences for populist parties (Michel et al., 2020, pp. 283-285). Second, criticism of representative democratic institutions opens the door to populist leaders who take advantage of the lack of external political efficacy and its related political trust (Linde & Peters, 2020; Torcal, 2014). As such, it is essential to understand what makes populist leaders successful in appealing to individuals with such attitudes. As we’ve argued above, these leaders may do so through a “proximity” strategy that emphasizes their warmth traits in an attempt to seem closer to the “people” and therefore better suited than their competitors to be a part of a more “direct” model of democracy, but also their being above institutions of representation and horizontal accountability that limits the “people’s mandate”. Finally, we should also point out that, as recent research has found (Bakker et al., 2021), individuals displaying lower levels of ‘agreeableness’ in their personalities have a greater tendency to support populist leaders. While it may seem like these two sets of findings are at odds and imply that individuals who support populists are less agreeable and therefore should want leaders who are less agreeable, we believe that the conclusions of this literature are not incompatible with our results. Focusing on the importance of in-group linkages between populist leaders and their supporters, we anticipate that individuals who support populist leaders should see them as being warm, displaying such traits as honest and caring, but only towards those in their in-group. It is very likely that they see the ‘toughness’ of some of these leaders as invaluable when dealing with the establishment, elites, or whatever the out-group may be. While this falls outside the scope of our current paper, we believe that this is a research question worthy of future exploration.
In a similar vein, it is possible that right-wing populist leaders may be perceived in different ways than left-wing populist leaders. In this study, our cases have two left-wing populist leaders, which limits our ability to generalize these findings to either centrist or right-wing populist leaders. We suspect that the relationship will be similar, as case studies have noted that center-right candidates have pursued similar proximity strategies (Annunziata, 2013), but we maintain that this is a hypothesis that should be tested in future studies.
While the importance of leaders’ personal traits would seem obvious in cases of presidential systems, and even more so in those with long histories of pervasive personalization like in Argentina, by including the party-centric parliamentary system of Spain we have shown that how people perceive leaders’ personality traits matter in less personalized systems as well. While in well-established democracies particular charismatic leaders may have emerged in times of crisis and gained personalistic notoriety, populist leaders now have more tools at their disposal through modern technology such as television, internet, and social media, to broadcast their brand and make more personal connections with voters. When populist leaders emerge in the party supply, they have been able to cultivate an image that perhaps increases their perceived warmth amongst certain segments of the population through their use of social media and television. Citizens holding populist attitudes and their related negative evaluation of democratic representativeness could be especially receptive to populist discourses, thereby fostering identification with the in-group delineated by these populist leaders that is so powerful that they do not waver despite widespread criticism and whatever gaffs, missteps, and indiscretions that came to light throughout these leaders´ campaigns and mandates. With the rise of new populist parties and candidates in numerous democracies, the findings from this study can help us to understand with more nuance the importance of leader traits, how those differ for populist and non-populist leaders across individuals, thereby shedding light on the dynamics of leader evaluations and party competition in a diverse and ever-growing number of contexts.
This research and the conclusions drawn from them do present some limitations. The first is the restricted nature of the batteries of questions on both leader traits. By developing a more extensive and nuanced of battery of items that could reveal other important traits and better capture the “warmth” dimension, more could be done towards better measurement and understanding of the role of leadership in fostering populism and its electoral consequences. Finally, this study is based just on two country cases and with limited differentiation in the characteristics of the leader. To better test the external validity of our findings would require greater diversity in both the cases and leaders selected.