Introduction

Populism is a difficult concept to define. Populist movements embrace the entire spectrum of conventional politics but have a similar core appeal. Dani Rodrik defines them this way: “What all these (populist movements) share is an anti-establishment orientation, a claim to speak for the people against elites, opposition to liberal economics and globalization, and often (but not always) a penchant for authoritarian governance,” (Rodrik, 2018: 1).[1]

Judis (2016) divides such regimes into right-wing and left-wing varieties. However, the difficulties in defining different tendencies in populism are reflected in the difficulties in conducting empirical research on the phenomena. There have been both qualitative (see, for example, Aslandis, 2016a, 2016b; Bánkuti et al., 2012, Mudde & Kaltwasser, 2013) and quantitative (see Fernández-Gracia & Leungo, 2018; Popping, 2018; Hawkins, 2009; Ernst et al., 2017) efforts to distinguish right and left populist regimes from each other. A distinguishing aspect of these studies, however, is that they are descriptive in character, dividing populist regimes generally by reference to a historical inheritance of populism or to particular aspects of the political history of the countries examined. Theoretical explanations for right and left populism are less common.[2] This shortfall creates a problem for research on varieties of populism going forward.

In his papers Rodrik (2018, 2019) has presented a convincing model for how different populist regimes arise in different situations. Their scheme divides society into three main groups: elites, majorities of the middle class and poor, and minorities identified by ethnic, religious, or citizen status differences. This leads to two major potential divisions that populist movements exploit: economic (income/social class) and identity (ethnic-nationalist/cultural) cleavages. He argues that these cleavages shape the anti-establishment politics that Judis (2016) identifies in right- and left-wing populist regimes, but in different ways.

This paper aims to test the cleavages model proposed by Rodrik. The identification of economic and identity cleavages with right- and left-wing populist regimes will constitute a significant theoretical advance if the model proves valid. However, the identification of income/class and ethnic-national identity cleavages with, respectively, left- and right-wing populist regimes has yet to be tested empirically. I will test these hypothetical linkages through an empirical examination of the appeals made by populist regimes themselves.

First, I will provide a short overview of Rodrik’s model. I will then examine the two cases used to develop the analysis in this paper: the Fidesz regime in Hungary as an example of right-wing populism and the United Socialist Party of Venezuela (PSUV) in Venezuela as an example of left-wing populism.

Then, applying the framework suggested by Rodrik, I examine Fidesz and PSUV manifestos in order to develop dictionaries of abstract terms distinguishing right- and left-wing regimes along cleavage lines. I use a classic content analysis of speeches by the leaders of the regimes in Hungary and Venezuela to determine how closely their public discourse tracks the differences proposed by Rodrik.

I then test the model beyond the initial cases by comparing speeches by Barack Obama and Donald Trump with those of Hugo Chávez and Viktor Orbán. Finally, I offer a few remarks about what has been learned and the future direction of the research.

Characterizing Right and Left Populism

The era of “neo-liberalism” emerging in the 1980s is the background change leading to the increasing popularity of both kinds of populist regimes (Rodrik, 2018). The new program, often imposed on developing countries by the International Monetary Fund (IMF) and the World Bank, called for faith in market mechanisms, substantial decreases in the regulation of both domestic and international business, and a decreasing role for the state through the privatization of social services and lower taxes (Ostry et al., 2016). Such policies were aimed at restoring economic growth by reducing inefficiencies created by attempts to control market relationships. When imposed on developing countries by international institutions like the IMF and the World Bank as “structural adjustment” measures required for loan approval, they disrupted the social fabric of developing countries. As the role of the state in regulating economies in developing countries was undermined, wages decreased, regulation of the economy shifted in favor of foreign intervention, privatization and foreign investment increased, and government services were gutted. Economic distress and social unrest often resulted (Gertz & Kharas, 2019).

Mobilization against the causes of these problems was easy for any left-wing populist movement. The distress of the people was the result of an uncaring economic and political elite that was purposefully causing economic hardship by reducing government expenditures and protective regulations while simultaneously opening the economy up to control by foreign businesses and financial institutions that had no democratic accountability. No wonder Rodrik (2018) found that left-wing populism has had significant support in Latin America for the last 50 years (these problems are not new) and that support for left-wing populist movements has sharply increased recently. Again, left populists have targeted the elite and built a political movement by mobilizing majorities and minorities against existing economic and class structures, restoring economic independence, and establishing political accountability (Rodrik, 2018). Thus, left-wing populism can be characterized as:

Majorities + Minorities vs. Elites

and will emphasize economic cleavages.

In advanced capitalist countries, the scenario had a different background but produced the same results. Neo-liberalism was born in the developed world beginning with the elections of Margaret Thatcher and Ronald Reagan. The subsequent “Washington Consensus” was based on the same policies that were forced on developing countries: ending welfare-state supports, deregulation of the economy, reduction of trade and monetary barriers, and ideological shifts toward support for globalization, especially for capital flows (Streeck, 2014). Further, the ties between the various secondary organizations—labor unions, local chambers of commerce, and “fraternal” organizations—that had been the backbone of both social democratic politics and social and economic identification in developed countries became more attenuated (Skocpol, 2003; Streeck, 2017, 2014). Streeck describes the political turmoil that has resulted aptly: “For the domestic politics of a nation-state undergoing neoliberal redefinition, this has profound consequences. Classes struggling over the correction of markets give way to status groups struggling over access to them. At issue are not the terms of exchange and cooperation between conflicting class interests, or the limits of exploitation of one class by another, but status groups with established market access excluding status groups without it from competition” (Streeck, 2017: 6).

The result in this case has been an increasing turn toward status identifications that have become increasingly detached from collective institutions. The relentless centralization of capital and the collapse of connections between, in particular, political parties and mass social organizations have led to an increasing disconnect between national political programs and the preferences of mass publics (Page et al., 2013; Streeck, 2014). In addition, increases in poorly integrated migrant minorities, brought on by labor needs in more prosperous times and by more recent migratory pressures, provide a ready scapegoat for those disoriented by economic distress and cultural change. Here, Rodrik (2018), like Judis, asserts that populist movements connected to elites have rallied the majority against a minority “other” and “globalist influences,” building political power by opposing economic dependence and attacks on ethnic and national identity. Thus, right-wing populism results from:

Populist Elites + Majorities vs. Minorities

and tends to emphasize identity cleavages.

Again, their scheme divides society into three main groups: an elite distinguished by wealth and position, a majority consisting of the middle class and the poor, and a minority identified by ethnic, religious, or citizenship-status differences. This leads to two major cleavages that populist movements can exploit: an economic (income/social class) cleavage and an identity (ethnic-nationalist/cultural) cleavage. These cleavages lead to different characterizations of the anti-establishment politics that Judis (2016) and others (Ernst et al., 2017; Pauwels, 2011; Scheppele, 2020) identify as the central theme of populist movements.

As Rodrik (2018) puts it, the supply of populist politics is dictated by the shocks to societies and the demands these generate for populist solutions. Right-wing populist movements emphasize the loss of control over national economies to international organizations and their domestic allies, as well as assaults on ethnic “cultural” values by influxes of foreign nationals and the adoption of more “cosmopolitan” mores. Hence, factions within the political elite attempt to rally the majority by supplying a populist call to reduce or eliminate the influence of foreign elites, the domestic organizations they control, and the minorities those organizations are allegedly using to undermine national identity.

In contrast, left-wing populism targets the political elite and the uncontrolled foreign corporate power they have unleashed on society as the root of the problems facing majorities and minorities. [3] Left-wing populism also emphasizes the loss of control over the national economy to global organizations and foreign economic interests. In this case, the result of elite control is seen in increases in poverty and social disruption caused by rising unemployment, financial crises, decreased social services due to austerity programs, and ever greater inroads by foreign corporations through reduced barriers to trade.

Tests of the model are possible through the tool of content analysis. If the model is useful, the texts generated by exemplar movements and their leaders can be analyzed to distinguish populist appeals and use them to identify populist tendencies in other leaders in different contexts. By conducting such an analysis in this paper, I can begin to answer a variety of questions.

First, if the proposed economic and identity cleavages allow us to clearly distinguish different populist regimes, there will be grounds for thinking that Rodrik’s theory regarding the divisions of populist regimes is well grounded. This would not constitute a direct test of the model; that will require a much more sophisticated study that includes tests of the basic conditions underlying the demand for populist regimes and the decisions by politicians to supply populist solutions. However, if the cleavages suggested by Rodrik do distinguish right- and left-wing populist regimes empirically, then a first step will have been taken toward a more well-founded definition of their differences. The model will then be subjected to a further initial test in this paper by using it to analyze populist tendencies in recent leaders of the United States.

Second, the model put forth by Rodrik strongly suggests that differences between right- and left-wing populist regimes are arrayed along linear dimensions. Right-wing regimes emphasize identity issues, while left-wing regimes emphasize income and class issues. But obviously, a right-wing populist regime must also have stances on economic issues, and a left-wing one on identity issues as well. Emphasis is not a one-way street, especially since so much of the angst driving populist movements is derived, at least indirectly, from economic insecurity. Right-wing populist regimes should be interested in reassuring majorities that national economies will be bolstered and protected from international interference. By the same token, left-wing populist regimes should be interested in assuring majorities and minorities that citizenship status will be protected and domestic political institutions strengthened. By examining the emphasis placed on the dimensions in the cleavage model, this research should help further efforts to distinguish right- and left-wing populist movements empirically.

The Cases

As the basis for the following analysis, I have chosen two cases that are generally considered to illustrate the poles of right- and left-wing populism: the Fidesz regime in Hungary and the PSUV regime in Venezuela. A short overview of the history of these regimes and the research describing them will show that they fit the cleavage model well.

Shortly before Hungary overthrew communist rule in “The Change” of 1989, a student organization, the Alliance of Young Democrats (abbreviated Fidesz), was formed by a leadership group gathered around Viktor Orbán and espousing a classic liberal program. Fidesz fielded candidates in the first two Hungarian elections after 1989 but saw little success as a result. Early elections in 1990 and 1994 were dominated first by conservatives and then by a reconstituted Socialist Party. As a consequence, Fidesz changed its ideological orientation to the center-right and, in 1998, won control of the unicameral National Assembly with Orbán as Prime Minister. This did not last; the Socialist Party, the inheritor of the Socialist Workers’ Party that had ruled Hungary before 1989, won the 2002 election and, to Fidesz’s chagrin, formed a coalition with the liberal Alliance of Free Democrats to win the 2006 election as well.

In response, Fidesz moved further to the right, particularly on issues of national identity, and awaited events. The Socialist coalition was then left to face the depression of 2008 and the collapse of Hungarian state finances on its own. [4] The 2010 elections were not only a triumph for Fidesz but also returned the party a supermajority of 68% of Assembly seats after the second round of elections. Consequently, Fidesz could, and did, promulgate a new constitution and used it to pass a variety of new laws cementing the party’s power. While the 2014, 2018, and 2022 elections were not as triumphant as the 2010 election, Fidesz maintained its supermajority after each one (Lendvai, 2012, 2017; Krekó & Enyedi, 2018).[5]

Hungary is usually considered a paradigm case of a right-wing populist regime. Since taking power, Fidesz, led by Orbán and a few long-time associates, has taken many actions that bear the hallmarks of right-wing populist politics: extreme nationalism, restrictions on press and academic freedom, attacks on judicial independence, and hostility toward the European Union, international capital, and immigration. Further, empirical research into the public pronouncements of the Fidesz government has shown them to be solidly populist in character (Popping, 2018). But does the Fidesz regime fit the model?

It appears to. Orbán has long been a member of Hungary’s political elite, and Fidesz has been contesting elections since 1990. It started as a conventional center-right party, but swung to the right, especially after it failed to unseat the Socialists in 2006. The unrest in the country, combined with the Great Recession, provided an opportunity to organize the majority to support an elite group—Fidesz—by offering a populist alternative and turning against outside economic influences from the European Union and international agencies. In addition, the threat of large-scale immigration to Hungary and the existence of indigenous minorities fueled antagonisms that Fidesz exploited (Lendvai, 2012, 2017; Krekó & Enyedi, 2018). Further, empirical research into the public pronouncements of the Fidesz government has shown them to be solidly populist in character (Popping, 2018).

The second case is the PSUV (United Socialist Party of Venezuela) regime in Venezuela. In 1998, the presidential election returned Hugo Chávez Frías, a former army officer who had launched an unsuccessful coup attempt in 1992, to office. Chávez had run on a platform of overthrowing the Fourth Republic established after the overthrow of Marcos Pérez Jiménez in 1958. The Fourth Republic had been based on a pact, the “Pact of Punto Fijo,” which had established an orderly sharing of power between the two major political parties of the time, Democratic Action (AD) and the Social Christian Party. The arrangement proved stable for 40 years and cemented Venezuela’s reputation as a functioning democracy.

The Punto Fijo system depended on the distribution of revenues from Venezuela’s oil industry to party elites and, initially, to the people of the country. By the 1998 election, however, the pact arrangement had run its course. Economic inequality and social disruption had increased as the government became increasingly corrupt and unresponsive to domestic concerns. [6]

After Chávez took office, he proceeded to call a “National Constituent Assembly,” elected by the people to write a new constitution. In December 1999, the new structure was put to a vote and won convincingly. There was a subsequent coup attempt against Chávez in 2002, but it failed for lack of either military or civilian support. Chávez was elected president three times—in 2001, 2006, and 2012—by substantial margins. In 2006, he created the PSUV as a coalition of left-wing parties.

Though Chávez died in 2013, the PSUV continues to rule Venezuela today (McCoy, 1999, 2000; Gott, 2011; Cannoni 2004). The new PSUV government under Nicolás Maduro Moro, elected president on his own in 2013 and, in disputed elections, in 2018 and 2024, attempted to continue the “Bolivarian Revolution” started by Chávez, albeit with indifferent results (Smilde, 2015).

The PSUV regime fits Rodrik’s model for left-wing populist movements. As with Fidesz, it might be hard to find a better example. Chávez led a movement based on rallying both the majority and, particularly, indigenous minorities to overthrow rule by entrenched domestic elites and their foreign partners.[7] He promulgated a new constitution aimed at, and successful in, destroying the Punto Fijo parties and creating a new deal for the Venezuelan masses, then used his personal popularity and the executive levers in the new constitution, along with a surge in world oil prices, to deliver on his promises. As the regime matured, Chávez created the PSUV to consolidate left-wing elements in Venezuelan politics behind his government’s efforts.

The PSUV maintained power under Maduro, though the country’s financial problems, international economic sanctions, and increasing internal political opposition led to a continuing trend toward authoritarian policies (Smilde, 2015, Corrales, 2020). Venezuela’s distress reached a new level with the kidnapping—the United States insists this was a “law enforcement operation”—of Maduro and his wife by the United States on January 3, 2026. Despite Maduro’s absence, the PSUV still holds power in Venezuela. Finally, as with the Fidesz regime, empirical evidence also supports the populist nature of the PSUV regime (Hawkins, 2009).

Methods and Data

The question that motivated this paper was whether the cleavage model proposed by Rodrik works. To begin, I must describe how the classification of economic and identity cleavages can be used empirically to distinguish the cases I have selected. Following a long series of research projects on populism, I will use a quantitative content analysis to determine whether the model holds. There are a variety of ways such an analysis can be conducted to describe populism. Texts can be analyzed by identifying the discourses they portray (see, for example, Scholten & van Ostaijen, 2015 or Popping, 2018) or by identifying and quantifying aspects of the texts themselves (again, for example, see Rooduijn, 2014 or Fernandez-Gracia & Luengo, 2018).

In this study, I will be using a computer-assisted dictionary approach, using words in speeches as the unit of analysis (Alexa, 1997). Following Laver et al. (2003), I have drawn the dictionaries from reference texts using both theoretical and inductive methods. In this paper, the reference texts are party programs: the Fidesz manifesto, A Stronger Hungary (2007), and the PSUV Proposal of the Candidate of the Homeland, Commander Hugo Chávez, for the Socialist Bolivarian Government, 2013–2019 (2013).

Both texts were cleaned using the online content analysis program Lexos (Wheaton Lexomics, 2019) and an English “stop word” list (Balet, 2017). [8] The cleaned texts were subjected to a preliminary word count generated by two applications, the open-source content analysis program Yoshikoder (Lowe, 2015) and the “word cloud” visualizer Wordle (Feinberg, 2014).

The word counts involved were sorted a priori into right- and left-wing categories fitting the economic and identity cleavages envisioned by the model. Common terms between the categories were removed, as were those associated with the particular situations of the countries involved. The goal, in other words, was to produce categories that could potentially be used to distinguish the economic and identity cleavages without reference to particular regimes.

The final dictionary terms can be found in Table 1 below. Those with an asterisk beside them are “stems,” i.e., roots representing multiple words. Thus, “Capital*” in the Economy Left category can stand for capital, capitalist, or capitalism.

Table 1: Dictionary Tokens by Category

The resulting dictionary categories were used to analyze a set of “virgin texts” for each regime (Laver et al., 2003). I selected the speeches of Hugo Chávez and Viktor Orbán as an initial test and, as a further comparison, the speeches of Barack Obama and Donald Trump. I divided the material into three categories: annual messages, political speeches to supporters, and speeches to international bodies or representatives. I did this because there is a real chance that the kinds of messages being sent in these different contexts affect the content of the speeches (Choi et al., 2018). One would expect that:

-       annual messages, as formal statements of policy, would be more comprehensive and would be aimed at a wide domestic audience;

-       speeches to political supporters would be more informal and more likely to reflect direct political appeals by the speakers; and

-       speeches to international bodies or representatives would be more explanatory in character and aimed at creating or cementing relations with other regimes.

The number of speeches for each leader and their sources are given in Table 2 below.

Table 2: Number of Speeches: Chávez, Orbán, Obama, Trump

Several aspects of the datasets used in this study should be noted. First, Alvarado’s (2023) data are invaluable. Most of the sites where she collected her archive of Chávez’s speeches have now been removed by the Maduro government. Second, the lack of interviews and press conferences for Chávez forced the entire project to be limited to speeches in order for the data to be comparable.[9] Third, I conducted a systematic sample from a random start of every other speech in Eidenmuller’s (2024) American Rhetoric collection of Obama’s speeches. I did this to ensure that the number of Obama’s speeches was comparable to the other collections.

Finally, for Trump, I had to use a variety of sources: the US Government’s (2024) archive of his speeches, Lillelund’s (2021) collection of Trump’s rally speeches from 2019–2020, and Eidenmuller’s (2024) small collection of speeches at American Rhetoric.[10] All of the speeches for each leader were also cleaned by the Lexos program using Balet’s (2017) stop-word list. This variety of texts in various contexts ought to provide a test of the categories and, by extension, the cleavage model itself.

Yoshikoder provides useful reports that enable a comparison of the likelihood of use of different dictionary categories in texts generated by each leader. The comparisons are based on relative risk ratios.[11] If the ratio is greater than 1, it is interpreted as an increase in the likelihood of the terms being used; if less than 1, as a decrease in that likelihood.

In the tables below, I report the percent increase or decrease in each content category and the level of significance reached based on the 95% confidence intervals of the risk ratios comparing each collection of speeches. (The full results are given in the Appendix below.) If a percentage reading is positive, it indicates that the first collection of speeches has greater content in that category than the second collection, and if the percentage reading is negative, the first collection has less content than the second related to that category (Lowe, 2006, Andrade, 2015).

Hence, a positive percentage for, say, Economy Right items in Orbán’s speeches compared to Obama’s would mean it was more likely that Economy Right content would be found in Orbán’s speeches than in Obama’s. A negative reading would mean that it was less likely that Economy Right content would be found in Orbán’s speeches than in Obama’s. As will be shown below, these comparisons allow a preliminary test of the cleavage model.

Results

Table 3 below presents a preliminary test of the model: a comparison between the combined speeches of Chávez and Orbán for each type of speech and in each category. Again, the table below contains percent differences and their significance levels based on the 95% confidence intervals of the relevant risk ratios (Lowe, 2006, Andrade, 2015). As can be seen, on average, Chávez’s speeches are 147% more likely to contain left populist content than Orbán’s and are, again on average, 186% less likely to contain right populist content.

The contrasts are especially stark for Identity Right content, as might be predicted from Fidesz’s emphasis on national culture, but all these differences are substantial and statistically significant. Further, the pattern of results is what the model would predict: higher percentages for Chávez on every left category compared to Orbán, and even higher negative readings for every right category, especially, as predicted, in right-populist identity content.

It is interesting that the readings track the emphasis of the two styles of populism; Chávez’s speeches are more likely to differ from Orbán’s on left economic content, and Orbán’s speeches differ more from Chávez’s on right identity content. Apparently, a content analysis contrasting the speeches of the two men in a variety of contexts distinguishes their positions exactly as predicted. Not only does this indicate that the dictionaries perform as expected by the model, but also that they could be effective tools to gauge the degree and direction of populism in the content of other collections of speeches.

Table 3: Chávez–Orbán Speech Comparisons with Proportions and Significance Readings

In the tables below, I compare speeches by Barack Obama with those of Chávez and Orbán (Tables 4a and 4b) and speeches by Donald Trump with those of Chávez and Orbán (Tables 5a and 5b), using the same techniques as above. The differences between the two comparisons of speeches are interesting.

First, as can be seen in Table 4a, Obama was no left populist. Chávez’s speeches are, on average, 360% more likely to reflect left populism and, on average, 161% less likely to have right populist content than Obama’s. This difference is especially pronounced in the annual messages and political speeches, but the contrasts are substantial and statistically significant for all categories. This might be expected; the United States is a capitalist democracy, friendly to right-populist themes, and Obama was always careful to present himself as a moderate and distance himself from leftist rhetoric.[12]

It is interesting to note, however, that Chávez was less likely to emphasize Left Identity content in international speeches than Obama (-25%). Since Chávez was, on average (-218%), less likely to use Identity Right content than Obama in international speeches, Chávez may have been trying to show restraint on identity issues in international forums.

Across all speech types, Orbán’s speeches are more likely to reflect Economy Left and Identity Right content than Obama’s. On the other hand, Table 4b shows that Obama’s speeches have interesting variation. Obama does not show any differences from Orbán in either Identity category in his annual messages. As noted above, Obama was always careful not to further inflame partisan differences; he tended to emphasize economic policy instead (Coe & Reitzes, 2010).

Obama’s political speeches also show no significant differences in the Identity Left category, probably for the same reasons. He avoided Identity Right content as well; Orbán’s political speeches have 33% greater Identity Right content than his. It is worth noting that there are no significant differences (in italics) between Orbán and Obama on Economy Right content in international speeches, but no overall pattern emerges as a result.

Again, Obama’s tendency to present himself as a moderate might be the explanation, as well as his ability to choose which themes to emphasize given the context of his remarks (Coe & Reitzes, 2010). As before, the dictionary categories seem to track the divisions foreseen by the cleavage model.

Table 4a and 4b: Obama-Chávez and Obama–Orbán Speech Comparisons with Proportions and Significance Readings

Tables 5a and 5b show a different picture. First, the comparison between Chávez and Trump in Table 5a shows what might be expected. Chávez is more likely to use Economy Left rhetoric (on average 566%), and his speeches are less likely to contain Economy Right content (on average -84%). Readings in the remaining categories across speeches are more interesting. Chávez’s speeches are less likely to contain Economy Right content than Trump’s, again as might be expected. Further, the readings for Identity Right content are much lower—an average of -212%—than for Identity Left. Again, Chávez appears to tone down Left Identity content in international speeches in comparison to Trump (-11%). All these differences are substantial and significant.

Results for the comparison of Orbán and Trump in Table 5b show more interesting differences. Trump’s State of the Union addresses show no substantial or significant difference from the content of Orbán’s annual messages in the Economy Right, Identity Left, or Identity Right categories (in italics). As was the case with Obama, Orbán is more likely to use left-populist rhetoric—104% on average—on economic issues, but Trump clearly does not differentiate himself from Orbán on the very issues that would distinguish him from right-populist stances.

In political speeches, Trump shows significant differences from Orbán on all issues except, again, Identity Right content. Finally, as with Obama, there is no significant or substantial difference from Orbán in the Economy Right content of Trump’s international speeches. Overall, there is evidence that Trump shows identifiable right-populist tendencies, consistently showing no difference from Orbán on right-populist themes across all categories of speeches, but especially in his annual messages.

Table 5a and 5b: Chávez-Trump and Orbán-Trump Speech Comparisons with Proportions and Significance Readings

Discussion and Conclusions

The findings above indicate that the cleavage model shows promise. The content analysis has revealed stark differences between the leaders of the Fidesz and PSUV regimes tied directly to the economic and identity cleavages in their rhetoric. Further, these substantive readings show the direction the model predicted; Chávez shows a much greater tendency to emphasize left economic content, and Orbán to emphasize right identity issues.

When applied to comparisons with Barack Obama and Donald Trump, interesting and counterintuitive findings emerge. Obama cannot be identified as a populist of either variety, given the evidence here. Indeed, he avoids populist rhetoric, especially on identity issues. On the other hand, Trump’s speeches show an affinity for right populism. Like Obama, he is no left populist, as the results show. Chávez’s speeches are, on average, 153% less right-populist than Trump’s and have, on average, 307% more left-populist content.

The comparison with Orbán, however, reveals a pattern of no substantial differences between the speeches of the two men, especially in annual messages. Trump’s State of the Union speeches noticeably tend toward right-wing populist rhetoric, differing little from Orbán’s in most categories. When one remembers that such speeches are wide-ranging policy statements meant for extensive domestic audiences, Trump’s emphasis on right-populist content is revealing.

His insignificant difference from Orbán in Identity Right content in his political speeches indicates that Trump is, again, using right-populist content aimed at domestic audiences of his most fervent supporters. That the readings for Economy Right in international speeches are not significantly different from Orbán’s is also revealing; Trump is presenting himself as a right populist on economic issues to the world as well as at home. The cleavage model has allowed us to empirically identify Trump as a right populist, just as might have been predicted.

The evidence suggests that Rodrik’s model has shown the way toward a new theoretical framework for examining right- and left-wing populism. There is clear evidence that the economic and identity cleavages identified by Rodrik usefully distinguish left- and right-wing populism, both in comparisons between representatives of these tendencies and in comparisons with other leaders.

Further, using the cleavage model allows us to be more specific about how, and in what contexts, leaders show populist tendencies and to what degree. The comparisons made here are still limited in scope; using the same model with approaches capable of capturing more context and tone would expand the range of findings generated. The evidence presented here, however, indicates that there is now a model that could point the way toward a better empirical description of right- and left-wing populism.

Appendix

Table A1: Chávez-Orbán Speech Comparisons with Risk Ratios and Confidence Intervals

Table A2: Chávez-Obama Speech Comparisons with Risk Ratios and Confidence Intervals

Table A3: Orbán-Obama Speech Comparisons with Risk Ratios and Confidence Intervals

Table A4: Chávez-Trump Speech Comparisons with Risk Ratios and Confidence Intervals

Table A5: Orbán-Trump Speech Comparisons with Risk Ratios and Confidence Intervals

[1] The difficulty with defining populism is, I think, a product of the way the politics in these movements works. There is an intentional unwillingness to express any general policy that would allow easy identification of a populist movement with establishment politics. Instead, the concerns of populist adherents are distracted by elite/mass divisions, concern about national economic and political independence, and, in right-wing populism, extreme nationalism and ethnic scapegoating. This is why some scholars (Aslandis, 2016b, Moffitt & Tormey, 2014) have refused to see populism as an ideology at all.

[2] But see Mudde and Kaltwasser, 2013.

[3] It is instructive to note that the leaders of left populist movements tend to be from outside the established political elite.

[4] Admissions at a 2006 party conference by the newly re-elected Socialist prime minister that, despite promises and assertions made during the campaign, the country was basically broke did not help. When the Great Recession struck the coalition government put together by the Socialists had already forfeited the trust of the electorate (Krekó and Enyedi 2018).

[5] This is only a sketch of the actual history of Fidesz’s eventual triumph. For more detail, see the cited works by Lendvai and Krekó and Enyedi.

[6] The 1989 riots in Caracas protesting rises in prices brought on by an austerity program instituted by Andres Perez’s AD government led to unofficial reports of thousands of rioters killed. Trust in Venezuela’s government decreased precipitously afterwards (Ciccariello-Maher 2012).

[7] The new government also specifically targeted policies, especially land reform efforts, that would help Venezuela’s indigenous population (Gott, 2011).

[8] There are literally scores of English stop word lists. The number of words varies from under 100 to over 1000. I chose a middle course by using Balet’s 317 word stop list. His site contains a 988 word stop list as well. I found no substantive difference in the Wordle word clouds by using it. The PSUV manifesto was translated from Spanish using the Lingvanex Translator before further analysis.

[9] The loss of transcripts from Chávez’s Aló Presidente! television program was especially disappointing. Again, Chávez’s speeches were translated from Spanish to English for analysis using the Lingvanex Translator.

[10] I used 8 of Trump’s rally speeches, again selected by a systematic sample from a random start, to avoid oversampling them.

[11] See Lowe’s description of the calculation (Lowe, 2006). For a more sophisticated description, see Andrade 2015.

[12] For an illuminating study of this see Elwanger and Duncan (2014). See also Coe and Rietzes (2010).